Millimeter-Wave RF Sources
for Accelerator Applications1
J. L. Hirshfield
Department of Physics, Yale University, New Haven, CT 06520-8120; and
Omega-P, Inc., 199 Whitney Ave., New Haven, CT06511
Abstract.
The motivation for development of high-power mm-wave sources and test
facilities is reviewed, with emphasis on the need to carry out rf breakdown and surface
fatigue tests on accelerating structures in parameter ranges that go beyond those available
using cm-wave facilities. Developments of mm-wave gyroklystrons and magnicons are
summarized, as these are the most mature of any candidate mm-wave amplifier now under
study. Emerging design and development activities for mm-wave high-power components
and rf pulse compressors are also summarized. Detailed discussions of many of these topics
are in papers found elsewhere in this volume.
INTRODUCTION
The motivation for development of millimeter-wave2 rf sources for a future normal
conducting (i.e., room temperature) collider arises from an expectation of exploiting
the customarily-assumed scaling (in rough proportion to frequency) of the dark current
limit [1] for the maximum accelerating gradient that can be sustained by a copper
accelerating structure. Thus NLC at 11.424 GHz is expected to operate with a gradient
roughly four times that of SLC (2.856 GHz), while yet another factor-of-three or
higher might be possible by operating at 30 GHz or above. The maximum gradient
will be constrained below the dark current limit by rf breakdown [1-4], and further
limited in practice by surface fatigue due to pulsed heating that can affect structure
lifetime [2,4]. Further experimental tests to determine the maximum achievable
accelerating gradient under a variety of conditions must be carried out before the
absolute limits will be known. This is proving to be true even within the context of
NLC, where a wide range of data over a wide range of frequencies on breakdown and
surface fatigue is desired in order to optimize the collider design.
New mm-wave rf amplifiers are required to enable the development of high-power
rf components, accelerator structures, and rf pulse compressors needed to build and
power new test facilities for breakdown and surface fatigue experiments. It even may
not be inconceivable that eventually these devices could serve as rf drivers for an
Sponsored by US Department of Energy.
Here millimeter wavelengths are taken to be 10 mm or less, i.e. frequencies of 30 GHz or greater.
2
CP647, Advanced Accelerator Concepts: Tenth Workshop, edited by C. E. Clayton and P. Muggli
© 2002 American Institute of Physics 0-7354-0102-0/02/$19.00
29
actual collider, such as an up-grade to NLC. Parameters to be judged for a candidate
mm-wave rf amplifier include peak output power, pulse width, average power, gain,
frequency stability, efficiency, bandwidth, and ability to operate into a highlyreflecting resonant load without excitation of spurious modes. For use in a test
facility, operation is required over a wide range of peak power and pulse width.
Few suitable high-power mm-wave rf components exist, so their development must
go hand-in-hand with that of the amplifiers, to allow evaluation of the sources and of
their utility in driving resonant structures. Likewise, rf pulse compressors must also be
developed to generate the high-peak-power, short, rf pulses needed to excite the
accelerator structures. If properly coordinated, such developments should lead to
establishment of test facilities for basic studies of accelerator physics issues at mm
wavelengths, patterned—for example—after NLCTA that operates at cm wavelengths.
This review is intended to set the stage for several expositions found elsewhere in
this volume on rf amplifier development at frequencies between 30 and 91 GHz.
Furthermore, in keeping with the need for coordinated establishment of high-peakpower mm-wave test facilities for accelerator research, development of high-power rf
components and pulse compressors is also discussed, briefly here and in greater detail
in the expositions found elsewhere in this volume.
MM-WAVE AMPLIFIERS
It is useful to take as a frame of reference for a discussion of rf sources for
advanced accelerator R&D the requirements as set down for NLC at 11.424 GHz [5,6].
The amplifiers are to be periodic-permanent-magnet (PPM) klystrons, each providing
75 MW peak power pulses of 3.2 (is duration at a 120 Hz pulse repetition rate. These
tubes operate with an efficiency of 55% [7,8], with improvement to 65% as a goal.
Peak power enhancement of a factor of about 7, to 510 MW in 0.40 jis pulses, is to be
provided by a delay line distribution system (DLDS) that will combine the outputs of 8
klystrons with prescribed phasing; DLDS is expected to operate with an efficiency of
85% [9,10]. A wide variety of mode converters, power combiners, windows, low-loss
transmission lines, waveguide bends, and other components are being developed for
DLDS [11]. For the 0.5 TeV version of NLC, 1,872 klystrons and 234 DLDS
assemblies will be required; these will be installed in the 5-km leading portions of the
10-km electron and positron tunnels. While it is unrealistic to expect individual
sources at mm-wavelengths to have parameters equal to these at cm-wavelengths, the
general values cannot be too much smaller without requiring an unmanageable
increase in the number of individual elements required to achieve a given collider
energy and luminosity.
Round-beam klystrons, such as those developed for NLC, are not considered
practical for high-power mm-wave accelerator applications because of beam optics
issues associated with the small diameter beam apertures. If one simply scales by a
factor-of-three from the NLC klystrons, a 34-GHz klystron would have cavity
diameters of about 9 mm and beam aperture diameters of about 3 mm.
30
A number of alternative amplifier or amplifier-like configurations have been
proposed as potential rf sources for advanced accelerator R&D. These configurations
include gyroklystrons [12], gyroharmonic frequency converters [13], cyclotron
autoresonance masers [14-16], multi-beam klystrons [17], sheet-beam klystrons [18],
and magnicons [19,20]. But among these, so far only the gyroklystron and magnicon
appear sufficiently well-developed to qualify as possible candidates for near-term mmwave accelerator R&D. Accordingly, discussion in this review is limited to those
amplifiers.
Significant technological challenges need to be overcome for successful operation
of any high-power mm-wave amplifier. These challenges include—but are not limited
to—the need for high precision in fabricating the electron gun to keep cathode loading
and emission uniformity within acceptable limits, to minimize edge emission from the
cathode, and to maintain gun field strengths well below breakdown limits; cavity
designs must maintain rf field strengths to below breakdown limits, and cavity
fabrication must be with sufficient precision to yield the desired eigenfrequencies and
quality factors for the operating mode(s); cavity designs must insure suppression of
dangerous self-excited competing modes even in the presence of significant reflections
from the output coupler and/or from a reflecting load (such as an accelerator structure);
and particle dynamics in the output cavity in the presence of strong output reflections
must be such as to avoid current interception, current return towards the cathode,
and/or onset of multipactoring.
MM-Wave Gyroklystrons
A summary of development of mm-wave gyroklystrons for accelerator applications
is provided elsewhere in this volume in a review article by W. Lawson, and additional
contributions are included by L. Ives and G. Nusinovich. So here only a cursory
review is in order. Experimental results obtained with cm-wavelength gyroklystrons,
and accumulating refinements in computational design tools, have established
confidence among gyroklystron researchers that mm-wave gyroklystrons for
accelerator applications can successfully be built. One "data point" that buttresses this
view is the recent successful realization of an industrially-built 94-GHz gyroklystron
of peak and average powers of 100 kW and 10 kW [21], although that peak power is
well below the 10-MW level sought for 3-mm accelerator R&D. Further data points
rest on past results obtained by Granatstein, Lawson and their colleagues at University
of Maryland that include peak power levels in fundamental cyclotron harmonic
gyroklystrons at 9.87 GHz of 30 MW in cylindrical cavity tubes and 8.6 GHz of 80
MW in coaxial cavity tubes [22,23]. In a second-cyclotron harmonic device with
cylindrical cavities at 20 GHz, 30 MW peak power pulses were reported [24,25].
Developments of mm-wave gyroklystrons intended for accelerator applications
that are now underway include a 50-MW fundamental gyroharmonic coaxial cavity
device at 30 GHz designed by M. Blank et al at CPI, Inc.; a 100-MW second
gyroharmonic coaxial cavity device at 30 GHz designed in a CPI/UMd collaboration; a
10 MW cylindrical cavity second gyroharmonic device at 91 GHz designed, and now
31
under construction, by L. Ives et al at Calabazas Creek Research, Inc.; and a 20 MW
coaxial cavity second gyroharmonic device at 91 GHz designed by W. Lawson and
colleagues at U.Md. The 30 GHz device(s) could find application in testing of
accelerator structures for the CERN two-beam accelerator CLIC [26], while the 91GHz devices are being built in response to SLAC interest in novel high-gradient
structures fabricated using LIGA techniques [27]. As examples, we list the parameters
for the devices at 30 GHz, 50 MW (Table 1) and at 91 GHz, 10 MW (Table 2).
Table 1. Design specifications for the 30 GHz gyroklystron [28].
peak output power
bandwidth
total gain (driver + gyroklystron)
pulse repetition rate
dc pulse length
rf pulse length
oscillation level below nominal output signal
harmonic content below nominal output signal
load return loss
nominal operating voltage
nominal operating current
mod anode voltage
average cathode radius
cathode angle
emitter length
peak cathode loading
magnetic compression
average beam radius
space charge limited current
peak field in gun
average alpha
average axial velocity spread
50 MW
30 MHz
80 dB
100 Hz
3800ns
900-1200 ns
-50 dB
-30 dB
-25 dB
500 kV
300 A
145 kV
7cm
42°
0.64 cm
7.75 A/cm2
30
1.3 cm
700 A
85.4 kV/cm
1.5
5.5%
Table 2. Design specifications for the 91 GHz gyroklystron [29].
peak output power
frequency
efficiency
gain
gyroharmonic index
number of cavities
input cavity mode
buncher and output cavity modes
output waveguide
gun voltage
10 MW
9 1.392 GHz
40%
55 dB
2
6
TE-011
TE-021
TE-01/TE-02 hybrid
500 kV
Limits to achievement of design parameters such as those listed for the
gyroklystrons in Tables 1 and 2 can arise from several directions, beyond the general
issues listed in the second paragraph on the previous page. Included among these are
32
larger-than-expected axial velocity spread or azimuthal asymmetry on the beam, and
lack of required accuracy in axial symmetry of the central coaxial conductor (if one is
employed). For example, simulation studies have shown that efficiencies for the 34and 91-GHz gyroklystrons mentioned above would fall from -40% with axial velocity
spreads below 3%, to less than 10% for spreads exceeding 10%. Power reflections
from the load back to the output cavity in the 30-GHz design are taken to be <0.32% (25 dB) for the operating mode and all competing modes, in order to maintain a 20%
margin-of-safety against self-excited oscillations in the output cavity. This margin-ofsafety could shrink or vanish if reflections exceed this limit, and a large enough
reflection could cause self-excited oscillations to occur in the output cavity. Indeed,
operation of high-power gyroklystrons has to date been limited to use of wide-band
diffraction output horns that radiate into anechoic chambers, and thus reflect negligibly
at the operating frequency or at the frequecies of competing modes. It has been
proposed to employ a high-power circulator to isolate the output cavity from a
reflecting load, but there is no evidence that such a circulator can be built at mmwavelengths to handle the required peak power levels and with sufficient bandwidth
and mode acceptance to control all dangerous modes. However, it can be anticipated
that these issues will be vigorously joined once operating tubes are built. Indeed, the
91-GHz tube now under construction by L. Ives et al that is described elsewhere in this
volume is the first of its type with a built-in output coupler, wherein direct coupling
will exist between the load and the output cavity. This feature should allow tests that
can improve understanding as to the performance of gyroklystron amplifiers that
operate directly into reflecting loads.
MM-Wave Magnicon
Accumulated experience in the recent design and construction of three high-power
magnicons for accelerator applications at decimeter and centimeter wavelengths has
provided confidence among magnicon researchers that a successful tube could also be
realized at millimeter wavelengths. Past magnicons include the first at 915 MHz, a
fundamental harmonic tube that produced 2.6 MW peak power with a pulse width of
30 jis, and with an electronic efficiency of 83% [30]. That tube successfully operated
into a standing-wave 9-cavity race-track microtron accelerator at Budker INP
(Novosibirsk) in 1985, and accelerated a 50 mA electron beam by 6 MeV per pass. A
second magnicon at 7.0 GHz, also built at Budker INP, operated as a frequencydoubler, and produced 55 MW, 1.1 jis output pulses at a repetition rate of 3 Hz, with a
gain of 72 dB and an efficiency of 56%. That tube is driven at 3.5 GHz and uses a 430
kV, 230 A beam having an area compression ratio of about 2300:1 [31]. Another
frequency-doubling magnicon at the NLC frequency of 11.424 GHz was designed and
assembled in the USA, in a collaboration involving Yale University Beam Physics
Laboratory, Omega-P, Inc. and US Naval Research Laboratory (NRL) [32]. That tube,
installed at NRL, is designed to produce 60 MW peak power in 1.0 jis output pulses at
a repetition rate of 10 Hz, at 60% efficiency and 59 dB gain. It employs a 470 kV, 220
33
A, 2-mm diameter electron beam [33]. At this writing, the 11-GHz magnicon has been
conditioned up to power levels of 15 and 25 MW for 1.0 and 0.2 (is output pulses,
respectively. It has already found use in high-power tests of active rf pulse
compressors and dielectric-lined accelerating structures, as described in the article in
this volume by S. H. Gold. This has made possible establishment of a facility in the
USA, second only to SLAC, where high-power tests of accelerator structures and
components can be carried out at the NLC frequency.
In the course of designing and building these three magnicons, steady
improvements in computational design tools [34] evolved. Experiments have shown
that all of the challenges listed in the second paragraph three pages earlier have been
successfully met in building and testing these tubes. Based on this experience, a
frequency-tripling magnicon amplifier at 34.272 GHz [35], was recently built by
Omega-P, Inc. at the Yale University Beam Physics Laboratory, with fabrication of
parts by Budker INP (Novosibirsk) and IAP (Nizhny Novgorod). Details of this
magnicon are provided in the article in this volume by O. A. Nezhevenko et al. A list
of its design parameters is given in Table 3.
Table 3. Parameters for 34-GHz magnicon [35].
operating frequency
peak power output
pulse duration
repetition rate
efficiency
drive frequency
drive power
gain
beam voltage
beam current
beam diameter
magnetic field, at deflecting cavities
magnetic field, at output cavity
34.272 GHz
44_48 MW*
1.5ns
10 Hz
41-45%*
11. 424 GHz
150 W
54 dB
500 kV
215 A
0.8-1.0 mm*
13.0 kG
22.5 kG
*Higher power output and efficiency values are for lower beam diameter, and vice-versa.
The gun for this tube has been tested to above its design voltage of 500 kV, and to
its design power of 100 MW, and its measured perveance is indistinguishable from the
design value of 0.61xlO"6 A-V"3/2. The tube is expected to be placed into operation in
the near future, once precise alignment (to <0.2 mm) is realized of the cryomagnetic
field axis with the tube axis.
Magnicons have intrinsic properties that facilitate their operation directly into
resonant loads without self-excitation of spurious modes in the output cavity. These
properties include (i) operation of the output cavity at the fundamental cyclotron
frequency, thus eliminating any lower gyroharmonic competing modes; (//) interaction
with a non-axisymmetric cavity mode that is only strong when the mode azimuthal
index matches the temporal harmonic (e.g., n = 2 or 3 for a frequency doubler or
tripler); (iii) relatively high intrinsic electronic efficiency with a correspondingly low
34
beam current for a given output power (for fixed voltage), thereby providing a good
margin-of-safety against self-excitation of spurious modes; and (iv) the nearimpossibility of causing interception or return of the beam when a high output
reflection is present, since—in contrast to the klystron—the interactions are all
deflecting, rather than axially decelerating or accelerating. These properties have
allowed stable operation of magnicons at 915 MHz and 11.4 GHz, when operating into
highly resonant loads of the sort that are typical of those found universally in
accelerator applications.
MM-WAVE COMPONENTS
As was stated above, development of high-power mm-wave components must go
hand-in-hand with development of the amplifiers themselves. Otherwise, there would
be no means for coupling output power into loads, including resonant loads, to
measure the total output power and the output frequency spectrum and its stability.
The variety of components that are needed will, of course, vary considerably from one
application to the next. But a minimum "shopping list" can be composed that is
probably universal. This list includes the following:
• output window(s), probably in an overmoded structure;
• mode converters for the output window(s);
• phase shifters;
• miter bends;
• power splitters and power combiners;
• low-loss transmission lines.
Development of such components at 34.272 GHz has recently started, as
detailed in the paper by G. Denisov et al that is included in this volume. To date, only
low-power tests of prototypes of several of the components have been carried out.
These tests confirm the validity of the designs, especially for many of the components
that are overmoded—and thus highly sensitive to design errors. High-power tests are
expected to commence once the 34-GHz magnicon is placed into operation.
MM-WAVE PULSE COMPRESSORS
Compression of rf pulses from a high-power amplifier is inevitably required in
accelerator applications, since the peak rf power required at the accelerator structures
far exceeds that available from a single amplifier itself. As stated in the Introduction,
the design for NLC envisions a power gain from the klystrons of a factor-of-7, to 510
MW using DLDS, for the peak power pulse to be applied to the accelerator structure.
A similar requirement exists for mm-wave accelerating structures. For example, a
structure having a shunt impedance of 130 MQ/m will require a power input of over
300 MW/m to achieve an accelerating gradient of 200 MV/m, if breakdown can be
avoided at this gradient level. Therefore, for an amplifier with an output of 50 MW, a
35
factor-of-6 in power gain will be needed if no more than one amplifier is to be required
for each meter of accelerator length.
RF pulse compression is achieved for SLC and NLCTA using the SLED-II
scheme, or a variant thereupon, nowadays referred to as a passive scheme, in that no
elements in the rf circuit have time-dependent properties. In SLED-II, two parallel
delay lines are filled with rf energy during the first portion (c - l)/C of the rf pulse,
whereupon the phase of the amplifier is rapidly switched by 180° to allow a
compressed rf pulse of proportionate width 1/C to emerge from the delay lines. The
power gain is rjC, where 77 is the efficiency of the compressor. A typical value of C
is about 5, beyond which 77 falls below about 50% and the scheme becomes
impractical. It has been shown in theory that, if an active switch could be developed,
then one might operate with C above 10 with high efficiency, thereby providing
greater flexibility in the design of the pulse compressor, and higher compressed power
for a given amplifier peak power.
Both passive and active millimeter-wave rf pulse compressors are under study,
with emphasis on quasi-optical compressor structures to handle high peak powers at 34
GHz and beyond. Details can be found in the paper by A. L. Vikharev et al that is
included in this volume.
CONCLUSIONS
Gyroklystrons and magnicons are the only mm-wave high-power rf amplifiers
mature enough to be judged at this time as to their suitability for accelerator
applications. Parameters to be considered for a candidate mm-wave rf amplifier
include peak output power, pulse width, average power, gain, frequency stability,
efficiency, bandwidth, and ability to operate into a highly-reflecting resonant load
without excitation of spurious modes. For use in a test facility, operation is required
over a wide range of peak power and pulse width.
Magnicons at dm and cm
wavelengths have already shown themselves capable of driving accelerator structures,
and of having operating parameters that are competitive with alternative sources, such
as klystrons. Gyroklystrons at cm wavelengths have provided impressively high peak
power output pulses, and forthcoming tests are expected to provide evidence to judge
their ultimate suitability for accelerator applications. Development of high-power
mm-wave components and rf pulse compressors must proceed hand-in-hand with
development of the rf amplifiers, in order that appropriate measurements can be made
to determine the performance limits of the sources, and in order to construct mm-wave
test facilities for evaluation of accelerator structures—including rf breakdown and
pulsed heating tests. Vigorous progress in this frontier field of accelerator technology
R&D can be anticipated in the near future.
36
ACKNOWLEDGMENTS
Appreciation is extended to the many individuals who provided information and
advice to the author in his preparation of this paper, including O. A. Nezhevenko, V. P.
Yakovlev, A. L. Vikharev, G. G. Denisov, S. G. Tantawi, W. Wuensch, W. Lawson,
M. Blank, and L. Ives. However, errors and/or opinions are solely the responsibility of
the author.
REFERENCES
1. Palmer, R. B., in Pulsed RF Sources for Linear Colliders, edited by R. C. Fernow, AIP Conf. Proc.
No. 337 (AIP, New York, 1995), pp. 1-15.
2. Nezhevenko, O. A., mProc. of the 1997 Particle Accelerator Conference, edited by M. Comyn, M.
K. Craddock, M. Reiser, and J. Thomson (IEEE, NJ, 1998), pp. 3013-3014.
3. Tantawi,S. G., "RF breakdown in high vacuum x-band waveguides," in these Proceedings.
4. Wuensch, W., "Investigations of RF breakdown and pulsed surface heating in 30 GHz CLIC
accelerating structures," in these Proceedings.
5. Eppley, K. R., in Pulsed RF Sources for Linear Colliders, edited by R. C. Fernow, AIP Conf. Proc.
No. 337 (AIP, New York, 1995), pp. 67-81.
6. Sprehn, D., Caryotakis, G., Jongewaard, E., and Phillips, R., in High Energy Density Microwaves,
edited R. M. Phillips, AIP Conf. Proc. No. 474 (AIP, New York, 1999), pp. 31-40.
7. Caryotakis, G., in High Energy Density and High Power RF: Fifth Worlshop, edited B. E. Carlsten,
AIP Conf. Proc. No. 625 (AIP, New York, 2002), pp. 1-13.
8. 2001 Report on the Next Linear Collider, Fermilab-Conf-Ol/075-E, LBNL-Pub-47935, SLAC-R-571,
UCRL-ID-144077, edited by N. Phinney, (SLAC, Stanford, 2001), p. 42.
9. Tantawi, S. G., Bowden, G., Farkas, Z. D., Irwin, J., Ko, K., Kroll, N., Lavine, T., Li, Z., Loewen, R.,
Miller, R., Nantista, C., Ruth, R. D., Rifkin, J., Vlieks, A. E., Wilson, P. B., Adolphsen, C., and
Wang, J., in Advanced Accelerator Concepts: Eighth Workshop, edited by W. Lawson, C. Bellamy,
and D. F. Brosius, AIP Conf. Proc. No. 472 (AIP, New York, 1999), pp. 967-974.
10. Tantawi, S. G., Nantista, C. D., Bowden, G. B., Fant, K. S., Kroll, N. M., Vlieks, A. E., Chin, Y. H., Hayano, H., and Vogel, V. F., Phys. Rev. ST-Accel And Beam 3, 082001 (2000).
11. Tantawi, S. G., "New Developments in RF pulse Compression," SLAC-PUB-8582, August 2000.
12. Gold, S. H., Rev. Sci. Intru. 68, 3945-3974 (1997).
13. Hirshfield, J. L., LaPointe, M. A., and Ganguly, A. K., in High Energy Density Microwaves, edited
R. M. Phillips, AIP Conf. Proc. No. 474 (AIP, New York, 1999), p. 178; LaPointe, M. A., Yoder, R.
B., Wang, M., Ganguly, A. K., Wang, C., Hafizi, B., and Hirshfield, J. L., in Advanced Accelerator
Concepts, edited by S. Chattopadhyay, J. McCullough, and P. Dahl, AIP Conf. Proc. 398 (AIP, New
York, 1996), pp. 887-897; Wang, C., Hirshfield, J. L., Ganguly, A. K., Phys. Rev. Lett. 77, 38193822 (1996).
14. Bekefi, G., DiRienzo, A, Leibovitch, C., and Danly B. G., Appl. Phys. Lett. 54, 1302-1304 (1989).
15. Bratman, V. L., Denisov, G. G., Kol'chugin, B. D., Samsonov, S. V., and Volkov, A. B., Phys. Rev.
Lett. 75, 3102-3105 (1995).
16. Gold, S. H. and Fliflet A. W., in Advanced Accelerator Concepts: Ninth Workshop, edited by P. L.
Colestock and S. Kelley, AIP Conf. Proc. No. 569 (AIP, New York, 2001), pp. 695-701.
17. Smithe, D. N., Bettenhausen, M., Ludeking, L., Caryotakis, G., Sprehn, D., and Scheitrum, G., in
High Energy Density Microwaves, edited R. M. Phillips, AIP Conf. Proc. No. 474 (AIP, New York,
1999), pp. 117-125.
18. Chen, C., "Sheet Beam Klystron," in the Snowmass 2001 Proceedings.
19. Nezhevenko, O. A., IEEE Trans. Plasma Sci. 22, 756-772 (1994).
37
20. Nezhevenko, O. A., Yakovlev, V. P., Ganguly, A. K., and Hirshfield, J. L., in High Energy Density
Microwaves, edited R. M. Phillips, AIP Conf. Proc. No. 474 (AIP, New York, 1999), pp. 195-205.
21. Blank, M., Danly, B. G., Levush, B., Latham, P. E., and Pershing, D. E., Phys. Rev. Lett. 79, 44854488 (1997).
22. Lawson, W., Calame, J. P., Hogan, B., Latham, P. E., Read, M. E., Granatstein, V. L., Reiser, M.,
and Striffler, C. D., Phys. Rev. Lett. 67, 520-523 (1991).
23. Lawson, W., Cheng J., Calame, J. P., Castle, M., Hogan, B., Granatstein, V. L., Reiser M., and
Saraph G. P., Phys. Rev. Lett. 81, 3030-3033 (1998).
24. Lawson, W., Matthews, H. W., Lee, M. K. E., Calame, J. P., Hogan, B., Cheng J., Latham, P. E.,
Granatstein, V. L., and Reiser M., Phys. Rev. Lett. 71, 456-459 (1993).
25. Granatstein, V. L. and Lawson, W., IEEE Trans. Plasma Sci. 24, 648-665 (1996).
26. Corsini, R., in Proc. of the 2001 Particle Accelerator Conference, edited by P. Lucas and S. Webber
(IEEE, NJ, 2001), p. 412-414.
27. Hruby, J., in High Energy Density and High Power RF: Fifth Worlshop, edited B. E. Carlsten, AIP
Conf. Proc. No. 625 (AIP, New York, 2002), pp.55-61.
28. Blank, M., "Design of a 50 MW, 30 GHz Gyroklystron Amplifier," in the Snowmass 2001
Proceedings.
29. Ives, L., "Update on the Development of a 10 MW, 91 GHz Gyroklystron," in these Proceedings.
30. Karliner, M. M., Kozyrev, E. V., Makarov, I. G., Nezhevenko, O. A., Ostreiko, G. N., Persov, B. Z.,
and Serdobintsev, G. V., Nucl. Instrum. Methods. Phys. Res. A 269, 459-473 (1988).
31. Kozyrev, E. V., Nezhevenko, O. A., Nikiforov, A. A., Ostreiko, G. N., Serdobintsev, G. V.,
Schelkunoff, S. V., Tarnetsky, V. V., Yakovlev, V. P., Zapryagaev, I. A., in High Energy Density
Microwaves, edited R. M. Phillips, AIP Conf. Proc. No. 474 (AIP, New York, 1999), pp. 187-194.
32. Nezhevenko, O. A., Yakovlev, V. P., Hirshfield, J. L., Gold, S. H., Fliflet, A. W., Kinkead, A. K., in
Proc. of the 2001 Particle Accelerator Conference, edited by P. Lucas and S. Webber (IEEE, NJ,
2001), pp. 1023-1025.
33. Nezhevenko, O. A., Yakovlev, V. P., Gold, S. H., Kinkead, A. K., IEEE Trans. Plasma Sci., in
press.
34. Yakovlev, V. P., Nezhevenko, O. A., Danilov, O. V., Myakishev, D. G., and Tiunov, M. A., "State
of the Art Simulations of Magnicon," in these Proceedings.
35. Nezhevenko, O. A., LaPointe, M. A., Yakovlev, V. P., Hirshfield, J. L., Serdobintsev, G. V.,
Kuznetsov, G. I., Persov, B. Z., and Fix, A., "34 GHz, 45 MW Pulsed Magnicon," in these
Proceedings.
38