Research and development of a high power Relativistic Klystron Amplifier performed in 1991 at the Naval Research Laboratory is described in this paper.

The operation of the Relativistic Klystron Amplifier (RKA) was extended to higher frequencies. The 5 kA, 4 cm-diameter, 0.5 MeV Intense Relativistic Electron Beam (IREB) was fully modulated at 3.5 GHz. The RF from this IREB was extracted into a coaxial transmission line, in a TEM mode, and subsequently radiated into the atmosphere through a horn sealed with a lucite vacuum window. Radiated power level of 150 Megawatts was achieved. Work is in progress on a similar RKA operating at 9.4 GHz, using a 1.4 cm-diameter electron beam.

The transmission coefficient of a radiating horn is calculated. The calculation imposes a radiation condition at large distances. It employs the boundary integral method developed by Bernstein and Rok.hlin . It is performed in the frequency domain and assumes axial symmetry. Ve discuss the dependence of rf transmission for the horn geometry that is used in the relativistic klystron amplifier experiments at the Naval Research Laboratory.

We have begun a program of repetitive operation of a high-current relativistic klystron on the CLIA facility at Physics International. Potential problems of beam formation have been investigated and found not to be a limitation at repetition rates as high as 200 Hz. We are presently beginning initial testing in repetitive mode of a 5 kA, 500 kV klystron, designed and tested in single shot mode at the Naval Research Laboratory. Initial results at 10 Hz do not indicate any problems that are inherently repetitive in nature.

The characteristics of beam mod ulation in the relativistic klystron amplifier at 1.3 and 3.3 GHz are analyzed. Good agreement is found between our simulation results and the 1.3-GHz RK A experiment by Friedman and Serlin. The unstable window modes for a 3.3-GHz RKA configurat ion are identified, and met hods for suppressing the window modes are discussed.

A self-consistent nonlinear theory of the energy and current modulation in a relativistic electron beam propagating through a klystron amplifier is developed. A closed integro- differential equation for the beam current is obtained, assuming that the beam current is a function of time t and propagation distance z. Properties of the current and energy modulation are investigated from this integro- differential equation for a broad range of system parameters. Magnitudes of the energy and current modulation are determined in terms of the gap voltage, the microwave frequency, geometrical configuration, the beam intensity and initial kinetic energy of the beam. The modulation amplitude increases, reaches peak and decreases slowly, as the beam propagates through the amplifier. The theory is extended to a two-cavity klystron and it is shown that the theoretical results agree remarkably well with previous simulation data.

There is a strong symbiotic relationship between a developing technology and its applications. New technologies can generate applications previously either unrealizable or impractical. Conversely, applications can demand the development of new technological capability. Examples of both types of development can be found in the evolution of HPM. The high power and energy output made possible by HPM have created a technology driven interest in directed energy weapons and short pulse radar. On the other hand, the requirements for heating of fusion plasmas have resulted in an application driven program to develop high average power microwave devices. In this paper we address these and other applications such as RF electron linacs, laser pumping, and beaming of power. Emerging applications, such as ionispheric modification and environmental cleanup, are also touched upon. The approach in this paper will be to review each application separately and then compare the requirements of the applications in terms of the power, frequency and other key requirements necessary for HPM to usefully address the application.

In an idealized model of a traveling wave amplifier operating in the linear regime, it is assumed that all the transients have decayed. This implies that the electromagnetic wave has an amplitude which is constant in time but it may vary in space according to the interaction process. In an idealized model of an oscillator, the situation is reversed. The amplitude of the electromagnetic wave is constant in space and i t may vary in time. We present a generalized formulation of the interaction in a traveling wave tube which includes reflections and spatial and temporal transients. Within this framework i t is shown that, in an amplifier, the reflections cause time variations of the amplitude that are ultimately revealed as broadening of the spectrum. The "transition" from an amplifier to an oscillator is also investigated. In the case of an oscillator, it is shown that in addition to the well known temporal transient (lethargy) there is a spatial transient. This is due to the fact that it takes some distance for the "fresh" electrons entering the oscillator to get bunched. Beyond this transient the amplitude is constant in space, as anticipated by idealized theory.

We have used several different methods to calculate the dispersion relation of the space charge waves on a sheet beam inside a slow wave structure. Pierce's parameters C and QC in his traveling wave tube (TYT) theory are evaluated. Contrary to the conventional theory, we find no unique identification of the space charge fields and the circuit fields in beam-structure interaction. The interdependence between these fields is addressed.

There is a strong symbiotic relationship between a developing technology and its applications. New technologies can generate applications previously either unrealizable or impractical. Conversely, applications can demand the development of new technological capability. Examples of both types of development can be found in the evolution of HPM. The high power and energy output made possible by HPM have created a technology driven interest in directed energy weapons and short pulse radar. On the other hand, the requirements for heating of fusion plasmas have resulted in an application driven program to develop high average power microwave devices. In this paper we address these and other applications such as RF electron linacs, laser pumping, and beaming of power. Emerging applications, such as ionispheric modification and environmental cleanup, are also touched upon. The approach of this paper will be to review each application separately and then compare the requirements of the applications in terms of the power, frequency and other key requirements necessary for HPM to usefully address the application.

The generation of large amplitude, monochromatic current modulation of an intense relativistic electron beam (500 keV, 16 KA) by an external microwave source through a series of cavities has been studied via particle simulation. In the case of two cavity geometry, the nonlinear interaction of the second cavity with the partially modulated beam driven by various strengths of external rf source has been investigated. Various geometric configurations for cavity and gap size and cavity separation have been studied to assess feasibility and prioritized configurations for the efficient operation of relativistic klystron amplifier. A three cavity RKA was proposed for the future extra-high power microwave generation.

We have developed a 60 MW, 60% efficient, 35 Joule/pulse secondary emission magnetron at S-band. We report on experimental results from this moderate voltage (120 kV), repetitively pulsed (10 Hz), injection locked (14 - 15 dB gain) magnetron. Results from particle-in-cell code computer simulations are presented which compare very well with the experiment when space-charge-limited emission is achieved. Experimental results from a proof-of-principle, low power (6 MW) two-magnetron array are also described.

The double stream cyclotron maser is a novel source of millimeter wavelength radiation in which two copropagating electron beams are caused to gyrate in a uniform axial magnetic field. The interaction of the slow cyclotron space charge wave on one beam with the fast cyclotron space charge wave on the other beam leads to high frequency bunching. The desired operating frequency is proportional to the electron cyclotron frequency (or a harmonic thereof) and inversely proportional to the difference in beam velocities, and can be achieved at low beam energies and axial magnetic fields. The linear instability growth rate is calculated from the fully relativistic Vlasov equation for the case of cold electron beams.

The influence of the longitudinal space-charge waves of a coherently gyrophased, helical relativistic electron beam on the cyclotron maser instability is investigated in a cylindrical waveguide configuration using a three-dimensional kinetic theory. A dispersion relation that includes waveguide effects is used to examine detailed stability properties of the cyclotron maser interaction. It is shown that, in general, the effects of space-charge waves on a coherently gyrophased beam are suppressed in a waveguide geometry in comparison with an ideal one-dimensional cyclotron maser with similar beam parameters.

We report on a new regime of free electron laser operation with reversed axial guide magnetic field, in which the rotational motion of the electrons in the helical wiggler field is opposed by the presence of the uniform guide field. The 33.3 GHz free electron laser amplifier is driven by a mildly relativistic electron beam (750 kV, 300 A, 30 ns) and generates 61 MW of radiation with a 27% conversion efficiency. Measurements with the conventional orientation of the axial field show a considerable loss of power and efficiency.

A theoretical analysis is presented which is capable of predicting the starting oscillation condition and the eigenmode temporal growth rate of gyro backward wave oscillators. The theoretical predictions are confirmed by computer simulations. The nonlinear saturation mechanism is due to that in the unstable system, the nonlinear electron phase shift induced by the interaction with the wave field results in a virtual end plane which tends to shorten the effective interaction length and brings the system into the marginal stable state. In the system far above threshold, the virtual end plane first oscillates about and eventually approaches the steady state location. This early oscillating behavior of the end plane strongly correlates to the initial pronounced spiking temporal structure in the output power. The spiking structure could be smoothed out if the magnetic field is tapered.

The linear and nonlinear analysis of backward wave oscillator (BWO) and multiwave Cerenkov generator (MCG) high power microwave (HPM) sources require accurate calculation of quantities that can be derived from the fields of the empty slow-wave structure modes. An efficient accurate method is presented to calculate the TM_0n modes and fields in a corrugated cylindrical waveguide. From the derived field presentation the group velocity, beam-mode coupling coefficient, and scattering matrix for the slow-wave structure termination are calculated and compared with experimental values where possible.

We show that the efficiency enhancement in plasma filled radiation sources and its dependence on the density of the background plasma can be accounted for by a mechanism of self-induced distributed feedback. Two counter propagating waves modulate the plasma density by the ponderomotive potential of their multiplication causing thus a grating in the refractive index of the plasma. The two waves are coupled by Brag reflections. The distributed feedback enhances the interaction between the resonant wave and the beam, and reduces the threshold gain for oscillations. Both effects enhance the efficiency of the oscillator.

Experiments performed in the Soviet Union have demonstrated that the relativistic multiwave Cerenkov generator (MWCG) is capable of producing very high power centimeter and millimeter wavelength radiation at high conversion efficiencies. The MWCG configuration consists of a high power relativistic annular electron beam propagating in close proximity to a highly overmoded cylindrically symmetric cavity region which has two rippled wall sections separated by a constant radius drift space. The design of a highly overmoded MWCG experiment is described. The experiment is designed to operate at a frequency of 35 GHz with output power capability in the 100 MW range.

A high performance, three-cavity gyroklystron has been designed for 33.2 GHz and is currently being constructed. It is an extension of Varian Associates' pioneering 28 GHz gyroklystron. Modifications include an additional buncher cavity, lower beam (alpha) , an improved lossy drift tube, and a slight increase in frequency. For an axial velocity spread of 7% and (alpha) equals 1.5, a self-consistent simulation code predicts an output power of 250 kW with an efficiency of 39% and saturated gain of 52 dB. Mistuning of the penultimate cavity was investigated. For the above velocity spread, mistuning did not improve the efficiency and significantly reduced the gain.

The design of a stable gyro-TWT amplifier that can generate extremely high power at cyclotron harmonic frequencies is presented. To prevent competing backward wave modes from oscillating, an amplifier with a multistage structure is used. Although the interaction strength in a harmonic gyro-TWT is strongest if an axis-encircling electron beam is employed, a more common MIG beam has been used in this design study to make it more practical. Using a 100 kV, 25 A MIG beam, a peak output power of 533 kW, peak efficiency of 21.3% and a saturated bandwidth of 7.4% have been predicted by a nonlinear self-consistent analysis for a three-stage second-harmonic stable gyro-TWT amplifier.

The nonlinear dynamics of a reflex diode microwave source are studied by a two-dimensional (2D) numerical simulation and a 1 dimensional (1D) Duffing equation model. Test particle trajectories in a nonlinear model of the potential between the real and virtual cathode shows distinct classes of electron trajectories in the system. The 1D and 2D models evaluated are similar upon comparison.

We report the studies of a free running virtual cathode oscillator radiating in the S-band. Although this vircator is acting as a nonresonant device, the radiation bandwidth is relatively narrow (< 2%), almost comparable to the results obtained when a high Q cavity is used (<EQ 1%). The customary chirping behavior of the RF pulse is absent in our measurements. The virtual cathode oscillator is radiating in a single mode identified as TM₀₁ mode by using microwave-gas breakdown techniques. The measured peak power is approximately 300 MW. The mode pattern is strongly affected by the presence of electrons inside the waveguide.

Abstract not available.

A design for a launcher of electromagnetic missiles for a target 100 km distant is discussed in terms of the time dependence of a current pulse over an aperture. Two issues in implementing this design are: (1) to provide the required current distribution over a small surface element; and (2) to coordinate many such elements in an array. We consider an energy flux per pulse at the pulse launcher on the order of 3 x 10 exp -8 J/sq cm for a pulse duration of 100 fs. If this pulse generation can be synchronized over a surface aperture of 6-m radius, for a train of 10 exp 6 pulses, one transmits a flux to a target at 100 km on the order of 10 mJ/sq cm per pulse train.

Circular closed-loop arrays have been found to have interesting resonances. There is reason to think that, by deforming these arrays, better directivity can be obtained. In order to search for shapes of deformed closed-loop arrays that are likely to produce a high directivity, a method for estimating the radiation pattern and guiding the search has been developed. The estimation is based on the assumption that the current along the elements of a closed-loop array is close to that of a traveling wave. This assumption has been verified experimentally. Based on the traveling-wave source, some guidelines have been found. By following these guidelines, the radiation patterns have been greatly improved.

Using a FFT algorithm, we study the frequency spectrum of EM ultra-short time duration to determine its influence on the propagation properties of such type EM pulse train. We try to perform these properties by taking into account the energy carried in the high frequency part of the spectrum. We attempt to use the temporal shape of pulses generated by new types of picosecond optoelectronic devices (Photo Conducting Pulsed Switching) in the case of flat wave front. For this study, we also consider the same approach as the one described by WU et al. (Harvard University) by applying the theorem of Bessel-Parseval. We try to correlate the slow decay of EM energy with the spectrum structure presented by the EM wave train.

A beam waveguide (BWG) design suitable for high-power applications is described. The design features a transmit-only, four-port high-gain horn as input to a BWG system with a single parabolic mirror and three flat plates. The use of a single parabolic mirror is such that the highest field concentration is no greater than that caused by the horn itself. The horn is linearly polarized and a grid reflector is used to reflect the orthogonal polarization into the receive feed. A rotatable dual polarizer provides for arbitrary transmit polarization. The dual-reflector system is shaped to provide uniform illumination over the main reflector and therefore maximum gain for the given size aperture. Measured data from a scale model BWG system are presented.

A description is being developed of the spontaneous emission in a uniform magnetic field with a dielectric medium. Considered here are beam energies of up to 10 MeV, indices of refraction (n) of up to 2, and magnetic fields of up to 5 T. As long as n is a slowly varying function of the radiation frequency, the spontaneous emission may go into a large number of harmonics. Compared to a vacuum, a dielectric medium increases the overall energy emission in the 10- to 10⁶-GHz spectral region. In this description, a new effect--a helical Cerenkov effect--is deduced whose power spectrum, unlike the usual Cerenkov effect, depends on the radius of curvature of the electron trajectory which is measured in the plane perpendicular to the direction of the electron guiding center.

A useful expression is derived for the collective bremsstrahlung recoil force on the bare charge of an unperturbed relativistic test particle in a nonequilibrium beam-plasma system. In this part of the bremsstrahlung process, the field of the test particle is scattered by the second- and third-order dynamic polarization current induced by the test particle and acts back on the bare charge of the unperturbed test particle. Attention is focused on the component of the recoil force on the bare particle, arising from self-interaction by the self field scattered in first order off the particle's own second-order dynamic polarization. This work is needed for calculations of collective bremsstrahlung instability in nonequilibrium relativistic beam-plasma systems.

This paper is concerned with a general mathematical framework for the modeling and analysis of pulsed radiation from an extended source distribution. The radiated field is expressed as a continuous superposition of pulsed beam (PB) propagators which emanate from all points in the source plan and in all directions and initiation times. This phase-space distribution of PB is matched rigorously to the given time-depended source distribution via the new 'local Radon transform', which extracts the local space-time spectral (directional) properties of the source distribution. The representation integral emphasizes a priori the local radiation properties of the source distribution, thereby improves numerical efficiency and enhances physical interpretation. The basic concepts have been introduced recently for two-dimensional configurations. The present paper extends the representation to three dimensions, derives expressions for relevant transforms and propagators and discusses the additional phenomena introduced by the three dimensionality. This phase-space formulation should be contrasted with the alternative Hermite pulsed beam expansion scheme presented elsewhere in this issue, which applies only for well collimated radiation.

We present a novel expansion scheme for radiation from well-collimated pulsed aperture distributions. The basis functions are a new set of pulsed beams (PB), termed Hermite pulsed beams (HPB) as they are related to the time-harmonic Hermite Gaussian Beams. They are highly localized space-time wavepackets and form a biorthogonal set of for expansion of time- dependent radiation. This expansion is applicable only for well collimated fields since the HPB are only paraxial solutions of the time-dependent wave equation, but its validity can be extended by using the complex source technique. The expansion is based on a global matching of the HPB set to the entire aperture distribution and therefore should be contrasted with other new PB expansion schemes which are based essentially on local analysis. As expected, the representation is most efficient if the collimation properties of the HPB are matched to those of the entire aperture so that they exhibit the same far field properties (e.g. diffraction angle). We show that the relevant parameter is the collimation (Rayleigh) length and introduce criteria to find this parameter for a given pulsed source distribution. The properties of the expansion are explored via numerical examples.

Antenna arrays are potentially useful for both the transmission and the reception of trains of electromagnetic pulses. In this paper, we formulate an aiming problem pertaining to an RKA- driven launcher of microwave pulses, and draw on a study of quantum measurements to develop an approach to aiming. Appendix A reviews the behavior of microwave radiation focused at a target in the Fresnel zone. Appendix B derives the antenna current for optimal angular sensitivity of a monopulse radar for a target in the Fresnel zone, very close to the axis.

Quasi-optical power-combining offers the most promising method for extracting large amounts of power from solid-state devices in the microwave and millimeter-wave range. This technique can be applied to a variety of devices. The difficulties associated with traditional waveguides power-combiners such as skin-effect losses are eliminated because the signals are combined in free-space. The solid-state devices are embedded in a two-dimensional grid configuration and placed in a Fabry-Perot cavity. In this respect, the quasi-optical power-combiner is analogous to a laser oscillator in which the active medium of the laser is replaced with an array of active devices. The grid presents a reflection coefficient to an incident plane wave which is larger than unity and the resonator provides feedback to couple the devices together. The two-dimensional structure of the grid is amenable to modern photolithographic processing and potentially allows thousands of devices to be integrated monolithically.

SuperIBEX produces 5 MeV electron beams with currents of 5 - 30 kA for atmospheric propagation experiments. Such beams are subject to the resistive hose instability which quickly disrupts propagation unless the beam is properly conditioned prior to injection into the atmosphere. The most successful conditioning scheme to date employs an ion-focused regime (IFR) cell to tailor the beam emittance followed by an active wire cell to center the beam. The emittance tailoring tends to 'detune' the hose instability, while the wire cell damps transverse perturbations which seed the instability. A 17 kA, 1.5 cm radius beam has been propagated in full density air to the end of the 5 m long tank without disruption. The beam has also been propagated in a density channel produced by a laser-guided electric discharge. The channel reduces beam expansion generated by scattering of beam electrons by the atmosphere. Results will be compared with simulation codes and with previous propagation experiments at other laboratories.

The growth rate of the hose instability may be reduced by tapering the beam radius from head to tail. Generation of such a beam has been achieved by a fast-rising magnetic field with a rise time of 30 ns and a peak field of 1.2 kG. The electron beam has an energy of 1.7 MeV, a current of 1 to 2 kA, and a flattop of 12 ns. The beam radius tailoring is indicated by time- resolved beam-radius measurements, from a scintillator in the beam path viewed by a streak camera.

Recent progress in an on-going development program leading to the design of superconducting continuous-wave (cw) linear accelerators for high-brightness ion beams is reviewed. A new spoke-resonator geometry incorporating a half-wavelength resonant line was fabricated and tested. This geometry serves as the basis for the constituent cavities of a superconducting section being designed for high-current testing with a deuterium beam. Considerable progress has been made in the design of this section. A multi-phased program leading to the development of a superconducting radio-frequency quadrupole (SCRFQ) has been initiated. Design considerations and test results from the various activities are presented.

A 6-lens ESQ LEBT system is designed by a detailed computational analysis with the aim of transport a 30 mA, 35 kV high-brightness H^- beam over a length of about 30 cm. The computer simulation studies are made using the beam parameters of two types of sources- -a Penning-Dudnikov source at Los Alamos and a volume ionization source of BNL. In our effort for a waist-to-waist transport mode with the ESQ system, the computer simulation results predict a factor of about 1.8 emittance enhancement of the output beam for the Penning-Dudnikov case, while the emittance growth is insignificant for beams from the volume source. The emittance growth is reduced to a factor of about 1.4 by rejecting approximately 10% of the beam particles in the Penning-Dudnikov case. The lens system is fabricated in-house at Maryland. Experiments on beam transport are planned to understand the physics of emittance growth and establish the confidence limit of the computer codes used.

Modulation of the beam current has been observed during ion focused regime (IFR) transport of a high-power relativistic electron beam propagating through a low-density background plasma In this experiment, a 1.7-MeV, 1-kA, risetime-sharpened electron beam is transported in a KrF excimer laser produced IFR channel in TMA gas. The IFR channel is immersed in a low-density plasma filled transport tube. We present experimental measurements and computer simulations demonstrating modulation of this high-current relativistic electron beam near the low-density background plasma frequency.

J X B drift motion of a relativistic electron beam propagating through an ion channel is investigated for both the flat-top and the gaussian density profiles. Due to finite beam size and due to the accelerator and beam head location, location of the maximum deflection point propagates from the accelerator nozzle to the beam head. The propagation velocity of the maximum deflection point is a fraction of the beam velocity. At the maximum deflection point, the perpendicular velocity of the electron beam segment does not match to the drifting velocity of the ion channel, causing a disruption of the beam propagation through the ion channel. This disruption is particularly troublesome for the later portion of the beam. For example, for a 1 kA beam, the first one microsecond(s) portion of the beam pulse is grabbed by the ion channel, but the later portion of the beam pulse will be lost. The pulse length of the grabbed portion by ion channel is proportional to the square root of the beam current.

Electron cooling of electron (and positron) sources may be important for future linear collider applications. Electron cooling works by superimposing a cooling electron beam on a hot particle beam. Heat is transferred from the hot beam to the cooling electron beam until thermal equilibrium is reached. The cooling electron beam is continuously generated and discarded into a collector, allowing the minimum beam temperature to be obtained. In order to cool electrons with electrons, an intermediary positron beam must be employed, since it is impossible to merge two beams of identical particles into the cooling straight. By adjusting the beta functions of the electron and positron lattices appropriately the final emittance of the stored electron beam can be made less than the emittance of the cooling electron beam. This paper will discuss accelerator physics issues relating to an electron cooled electron beam source.

The equation of transverse motion governing cumulative beam breakup has been solved using Fourier analysis. This technique simplifies the investigation of the effect of finite bunch length and arbitrary current distribution within the bunches on beam breakup, which is an especially relevant consideration for low-velocity, high-current ion linacs. It also simplifies the calculation of the transverse dynamics of unbunched particles which constitute a diffuse longitudinal halo between bunches. If allowed to impinge on the accelerating structures, these particles could cause activation over long periods of continuous-wave operation. Simulations illustrating the salient effects of finite bunch length on steady-state cumulative beam breakup are presented here. The role of focusing in suppressing the transverse displacement of a beam which is misaligned at the accelerator entrance is also explored, and an analytic formula is derived for use in estimating the required focusing strength.

A theoretical investigation of the interaction of electrons with the electromagnetic wave in a microwave amplifier indicates that about 50% of the electrons are in fact accelerated in the amplification process. We analyze a system which consists of an amplifier immediately followed by an accelerator. Although we loose part of the electromagnetic power due to the impedance mismatch, the electrons remain bunched and trapped when entering the acceleration section so that the acceleration process is reasonably efficient. The electrons which lost energy in the amplifier section, practically do not participate in the interaction process, since their velocity mismatch is too large. With 1 kA, 0.85 MV electrons and E_z equals 0.6 MV/m at the input, 40.6 cm of amplification and 43.4 cm of acceleration, calculations show that at the output 6% of the electrons have energies between 8 and 16 MV.

We have tested a high-power 11.4-GHz rf generator which consists of a 5.7-GHz transverse modulating system and two 11.4-Ghz traveling-wave output structures. The device was designed to generate 500 MW of pulsed rf power when driven by a 1-kA, 3-MeV induction accelerator beam. Transverse beam instability due to rf coupling between the two output structures has limited the width of the rf output pulse for currents above 600 amperes. Short rf pulses of total output power of up to 420 MW have been produced. Using a single output structure, rf output pulses with stable phase (< +/- 2 degree(s)) and amplitude (< +/- 2%) have been achieved for widths comparable to the beam width. We have modified and tested an output structure to decrease the growth of fields causing transverse instabilities. During the next year our experimental program will include both studies of rf power extraction and reacceleration of modulated electron beams. In support of reacceleration experiments, we are developing a time dependent computer code for the simulation of transverse instabilities due to dipole modes in the rf structures, and are upgrading the induction beam to 5 MeV.

The Proof-of-Concept (POCE) experiment for the Spiral Line Induction Accelerator (SLIA) involves injecting a 3.5 MeV, 10 kA electron beam into a strong focussing, two-turn, racetrack type spiral magnetic transport line with two passes through each of two 1.5 MV induction accelerating units, giving a nominal 9.5 MeV, 10 kA, 35 nsec output. The design selected for the 80 cm bends of the POCE is a stellarator achromat which gives an identity transfer matrix to first order in the relative momentum mismatch. Experiments at 1 MeV in a straight stellarator field, which simulated the first POCE bend, demonstrated the existence of the stellarator achromat at the predicted magnetic field values. No charge loss was observed for simulated momentum mismatches of up to +25%. The progress of experiments on matching the 3.5 MeV, 10 kA beam from the field-free diode into the nominal 5 kG axial guide field and the transport efficiency through the POCE bends are reported. The status of the remaining hardware fabrication and the schedule for full scale experiments are summarized.

Electron cooling of particle beams should offer a means for variable emittance control. Electron cooling works by superimposing a cold electron beam on a hot particle beam. Heat is transferred from the hot beam to the cold electron beam until an equilibrium is reached. The cold electron beam is continuously generated and discarded into a collector, allowing the minimum beam temperature to be obtained. The rate of cooling is determined by the cold electron beam current. If the cooling rate is too large, equilibrium emittances will be established that lead to instability. A low longitudinal emittance can lead to a susceptibility to the Z/n instability, while low transverse emittances may lead to space charge tune shifts that force a crossing of destructive betatron resonances. Control of the final emittance would enable an avoidance of these problems. In the absence of instabilities and under an ideal vacuum, equilibrium will be established between electron cooling and intrabeam scattering. Varying the electron cooling current will cause a change in the equilibrium emittance. Thus, a variable current electron cooling system can provide effective emittance control, allowing the emittance to be set to a value that is experimentally chosen as the best operating point for a given machine.

A novel, compact S-band LINAC has been designed and is currently under construction at UCLA. It is expected to deliver high brightness, 200 A, 20 MeV electron pulses, less than 4 ps in duration from a device that is about 1 meter long. It comprises; (1) a laser photocathode driven gun that produces 4.5 MeV electron bunches from a 1 1/2 cell cavity operating in the (pi) -mode and (2) an accelerating structure known as a plane wave transformer (PWT) designed by Swenson. The design considerations of the machine and initial operating experience of the gun are discussed. The linac will be used for free electron laser, advanced accelerator research and beam-plasma experiments.

The basic periodicity in the NRL device is twelve-fold. There are twelve toroidal field coils, twelve sectors, and so on. For a long time, the l equals 12 resonance was the dominant resonance in the experiment. To test the importance of the l equals 12 resonance, we have intentionally introduces an l equals 12 field error using twelve resonant coils. By activating these coils when the resonance is observed, the duration of x-rays produced by beam loss was reduced from 900 microsecond(s) to 5.5 microsecond(s) , while the amplitude of the signal increased from 0.5 to approximately 40 volts. The full width of the x-ray pulse at half maximum is inversely proportional to the current through the resonant coils and proportional to the risetime of the pulse. Work is in progress with a set of twelve internal coils that have a risetime of 300 - 400 ns. The results with these coils have confirmed the importance of the l equals 12 resonance. In addition, it has been shown experimentally that by adding a (Delta) B_(theta ) such that the ratio of the total toroidal magnetic field to the vertical field is not an integer, the cyclotron resonance is not excited. These experiments have extended the beam lifetime by more than 100 microsecond(s) , which is approximately equal to the risetime of the applied (Delta) B_(theta ) pulse. In the presence of such a pulse, the beam lifetime is 1 ms and the electron beam energy is 22 MeV, while the trapped current is in excess of 1 kA.

Experiments designed to investigate the beam breakup (BBU) instability have been performed using the long-pulse MELBA electron-beam generator (0.5 - 1.5 microsecond(s) , 0.7 - 0.8 MV, ^ldiode equals 1 - 15 kA, ^lextracted equals 0.1 - 0.5 kA). The experiment consists of 10 identical pillbox cavities each containing a small microwave loop antenna designed to detect the TM₁₁₀ beam breakup mode. For our cavity design the TM₁₁₀ resonant frequency occurs at approximately 2.5 GHz. The cavities are connected by small diameter tubes which attenuate the RF cavity-to-cavity crosstalk. The MELBA diode and subsequent cavity system are immersed in a solenoidal magnetic field (0.8 - 3 kG). Microwaves of 2.5 GHz (1 - 4 kW), whose pulselength exceeds the beam pulse, can be injected into the initial cavity in order to prime the BBU instability. BBU instability growth is measured through the growth of 2.5 GHz RF between the first (or second) and tenth cavities. The BBU growth is compared with predictions made by beam-cavity coupled-mode theory.

The coupling of the accelerating cavities to dummy cavities was recently found to reduce beam breakup growth. This paper analyzes a more sophisticated model, proposed by George Caporaso, that includes the time delay between coupled cavities and covers the possibilities of wave cutoff and resonance in the coupling path. A dispersion relation is obtained. We show that the peak spatial exponentiation rate of the dominant BBU mode is reduced by as much as a factor of two, and this reduction is insensitive to the coupling path length. Other modes are destabilized, however, but they have lower growth rates, in general.

The nonlinear, self-consistent propagation of ultra-intense, subpicosecond laser pulses in plasmas is analyzed. Large amplitude wakefields are generated when the laser pulse length is approximately equal to the plasma wavelength. Relativistic optical guiding is found to be ineffective in preventing diffraction of pulses with lengths less than the plasma wavelength. Short, intense laser pulses can be effectively guided with preformed plasma density channels. Simulations based on a 2D-axisymmetric fluid model are discussed.

A high-power X-band klystron employing a double-gap output cavity has been operating at SLAC. Multi-gap output circuits have lower surface gradients at the interaction gaps than single-gap ones but are prone to self-oscillate due to negative beam loading and trapped higher-order modes. In the double-gap circuit design, considerable attention had been directed to deal with these stability problems. The performance of the present tube appears to be limited by gap breakdown and beam interception particularly at long pulses. A three-gap output cavity is currently under development to further reduce the gap surface gradient. Another new feature of the circuit is an enlarged downstream drift tube to improve on beam clearance. This paper discusses the considerations involved in designing a multi-gap output cavity and presents the cold test measurements on the three-gap circuit. The experimental data is compared with numerical results from the 3-D simulation code ARGUS.

The nonlinear interaction of a continuous electron beam with a traveling electromagnetic wave in the cyclotron resonance laser (CRL) accelerator is analyzed. It is found that the maximum beam energy gain and acceleration distance obey certain scaling laws for optimized systems, and that the maximum energy gain is limited primarily by the degree of wave dispersion and is almost independent of the wave amplitude. As an example, design studies of a 33-GHz, high- current, CRL accelerator based on three-dimensional, self-consistent, multiparticle simulations are presented.

Electrostatic free electron lasers (FELs) operating in the oscillator mode have been demonstrated to be extremely efficient sources of coherent electromagnetic (laser) radiation. Output wavelengths between tens of microns and a few centimeters are achievable. The average power of these devices can be quite large (100s of KW). Linear colliders require high peak power as well as high average power. By storing the laser energy of an electrostatic FEL oscillator in a resonant cavity, and then employing a cavity dump, extremely high peak output power is obtainable. The high efficiency, accessible wavelength range, large average power and high peak power make electrostatic accelerator FELs excellent candidates for linear collider power sources.

In this paper, we introduce two methods of separating electron beam by using the deflection principle of accelerating cavity, according to the concept of the free electron laser master oscillator power amplifier (MOPA) experiment driven by time-sharing an electron beam from a single rf linear accelerator. The features of electron beam deflection at the end of cavity are also presented by simply analyzing interaction between electron and em field in cavity.

The goal of this research effort is to develop a long-pulse relativistic klystron amplifier (RKA) by extending the pulse length of this gigawatt-class device by an order of magnitude beyond the current state-of-the-art (100 ns) to one microsecond. A research approach is described for obtaining kilojoule microwave pulses at 1.3 GHz. Achieving kilojoule microwave pulses requires extending the electron beam pulse duration beyond one microsecond without diode closure, and maximizing the microwave extraction efficiency at the fundamental frequency. Our earliest experiments have produced a modulated electron beam for one microsecond with a peak rf current of 0.9 kA and a voltage of 350 - 400 kV. In some cases we have observed beam modulation in excess of 2 microseconds. The component of beam power at the microwave drive frequency (1.3 GHz) was approximately 350 MW. Although only a small effort has been put forth to address the output coupling issues, approximately 50 - 70 MW was coupled into dominant mode rectangular waveguide. Recently, an electron beam diode has been tested that delivers peak currents in excess of 5 kA for a monotonically increasing current pulse exceeding durations of 1 microsecond(s) at beam kinetic energies above 500 keV. to achieve this result close attention was given to minimizing the current losses from the diode and maximizing the beam current transmission through the RKA.

We discuss basic Relativistic Klystron Amplifier physics. We show that in the intense space- charge regime the maximum power extraction does not coincide with the maximum harmonic bunching. In addition, we show that as the beam is bunched, the additional power stored in the Coulomb fields does not add significantly to the overall power extraction. Because of these effects, the power extraction at 1.3 GHz for a 500 kV, 5 kA beam with reasonable beam-to- wall spacing is limited to around 35%.

At C.E.S.T.A., the 35 GHz FEM ONDINE I experiment is in progress. The work done this year focused on beam creation and propagation on the Euphrosyne generator with both an oxide cathode and a cold cathode: typically, 1 kA at 2.5 MeV is propagated. A lot of calculations have been done on these problems and we have also developed simulation codes on FEM interaction in both Compton and Raman regimes, taking into account 3D wiggler field, guiding field, waveguide modes, electron beam injection to obtain trajectories, gain and efficiency. Good agreement with some previous experiments and codes has been found. A microwave apparatus has already been completed to work in an amplifier configuration: injection magnetron line (100 kW), amplification line and diagnostics (wavenumber analysis for full mode analysis, filters, detectors, etc.). All these components are compatible with the future high repetition rate ONDINE II experiment using the LELIA induction linac. The FEM experiment is now starting at 0.1 Hz with the 1.5 MeV LELIA injector. Beam creation from a thermionic cathode and propagation have been fully studied theoretically and experimentally. An electron beam of 1.3 kA at 1.3 MeV with an emittance of around 200 (pi) -mm-mrad is likely to be compatible with an efficient FEM operation.

A theory of stimulated scattering of electromagnetic waves by high -current relativisitic electron beams is devoleped ,with the transverse inhomogeneity of incident and scattered waves , and the presence of a magnetic field of 20 kG focusing the particles taken into account. The nonlinear results including the cou pling coef ficient versus the longitudinal magnetic field and the radius of electron beam ,the excitation curves .powers and ef ficiency etc. are obtained on a set of input parameters of our recently establishing accelerator and FEL system with an electron energy of 600 keV and a current of 3 kA. It is shown that the electrodynamic system of backward wave oscillator ( BWO ) in the form of a cylindrical waveguide wit h a periodically corrugated wall can have a high Q-factor for oscillations at wavelengths much shorter than the corrugation period of 1. 95 cm and the wavelengths of incident pump waves. The relativistic 3. 2 cm backward wave oscillator provided powerf ul scattered radiation at 3- 8 mm with integral power of 30 M W and efficiency up to 1 3% .

Microwave sources based on backward-wave oscillators driven by relativistic electron beams are capable of producing high power coherent radiation in the cm and mm wavelength regime. Although there have been a number of experiments reported over the last decade on this topic, there are only a few publications providing a theoretical description of these devices. Thus, there is a need for theoretical models which can be compared in detail with the experimental data. This work is devoted to filling this need and applied to the University of Maryland backward wave oscillator experiment. It is shown that the theoretical predictions for the threshold current to start the oscillations, the frequency characteristics, and the efficiency of the device compared favorably with the experimental data.

A unique, high-energy microwave source, called PASOTRON (Plasma-Assisted Slow-wave Oscillator), has been developed. The PASOTRON utilizes a long-pulse E-gun and plasma- filled slow-wave structure (SWS) to produce high-energy pulses from a simple, lightweight device that utilizes no externally produced magnetic fields. Long pulses are obtained from a novel E-gun that employs a low-pressure glow discharge to provide a stable, high current- density electron source. The electron accelerator consists of a high-perveance, multi-aperture array. The E-beam is operated in the ion-focused regime where the plasma filling the SWS space-charge neutralizes the beam, and the self-pinch force compresses the beamlets and increases the beam current density. A scale-model PASOTRON, operating as a backward- wave oscillator in C-band with a 100-kV E-beam, has produced output powers in the 3 to 5 MW range and pulse lengths of over 100 microsecond(s) ec, corresponding to an integrated energy per pulse of up to 500 J. The E-beam to microwave-radiation power conversion efficiency is about 20%.

Abstract not available.

Directed energy flows (DEF), including a High Power ion beams (PIB), are used in different areas of science, engineering and technology. For example, very worth-while is the use of PIB for: the realization of inertial controlled fusion, pumping up gas lasers, the investigations in the area of nuclear physics and energy high density physics, the formation of powerful pulse sources of X-ray and neutron radiation, ion alloying of metals and making surfaces, which improve physical and chemical properties of metals (enlargement of their hardness, corrosion, stability, etc.). The simulation of interaction processes of X-ray radiation with the matter now becomes more actual because of the progress in physics of short length wave laser. High cost and difficulties of the experiments and also the difficulties to get fast changing physical parameters in the area of the DEF--interaction with the target make it necessary to carry out a preliminary computer simulations for the evaluation of the expected physical parameters and the very expediency of such physical experiment. The examples and results of such mathematical simulation on dynamics of intensive pulse actions on metal targets by DEF (high-power ion beams, sharped - charged jets, hypervelocity projectiles, X-ray radiation), are represented in this paper with brief description of used computer models, worked out by High Energy Density Research Center (Russia).

Several different style 11.4 GHz relativistic klystrons, operating with beam pulse widths of 50 ns and using large aperture, tapered phase-velocity TW structures, have recently demonstrated output RF power levels in the range of 100 to 300 MW without breakdown or pulse shortening. To extend this performance into the long pulse regime (1 microsecond(s) ) or to demonstrate a threefold increase in output power by using higher currents, the existing TW circuit designs must be modified (1) to reduce the cavity maximum surface E-fields by a factor of 2 to 3, and (2) to elevate the current threshold values of the beam induced higher order modes (HOM) to ensure avoidance of RF pulse shortening and associated instabilities. A technique for substantially elevating this threshold current is described, and microwave data and photographs are presented showing the degree of HOM damping achieved in a recently constructed 11.4 GHz TW structure.