3
Technical Assessment: Scalability to One-Megawatt Power Levels

HOW TO ACHIEVE 100 KILOWATTS AND ONE MEGAWATT

The current level of FEL power in the near-infrared wavelength range (around 1 micrometer) is 14 kW, established at Jefferson Laboratory. The next step proposed by the Navy program is to demonstrate and study a 100 kW FEL system to establish the technology needed for scaling to the megawatt power level in the infrared. Assuming well-established undulator technology, the undulator period will be in the range of ~3 to ~5 cm, so that the electron beam energy needed to reach infrared wavelengths will be in the range of ~80 to ~120 MeV. The Jefferson Laboratory FEL system, with an energy recovery linac, has established that an energy extraction of a few percent (~2 percent) can be obtained, while inducing an energy spread of ~10 percent. This means that the necessary average current recirculating in the energy recovery system must be ~1 A for a megawatt-class FEL and ~0.1 A for a 100-kW-class FEL. Jefferson Laboratory now achieves its 10 kW operation by recirculating electron pulses of about 0.1 nC at a repetition rate of about 75 MHz. Therefore, the path forward in the new 100 kW FEL will need to achieve average recirculating currents of around 0.1 A, and the path forward in the megawatt-class FEL will need to achieve average recirculating currents of around 1 A.

The increase in average current can be achieved by increasing either the bunch charge or the bunch frequency, or both. To achieve a 0.1 A current for the 100 kW FEL, a bunch charge of 0.1 nC can be produced at an increased frequency of 750 MHz, or the bunch charge can be increased to 1 nC at the same frequency of 75 MHz. Some technical issues depend more on the average current, while others depend more on the peak current in the electron bunches. The lower bunch charge of ~0.1 nC and the associated problems have already been explored at Jefferson Laboratory, while the increased bunch charge of ~1 nC may well lead to new technical issues. The 100 kW FEL will require exploration of the higher average current of around 0.1 A at 750 MHz (“filling every bucket”) but can also involve exploration of the generation and transport of ~1 nC bunches at the lower repetition rate of 75 MHz. To scale the power to the MW class, it will be necessary to increase both the pulse charge to ~1 nC and the pulse repetition frequency to ~750 MHz.

Bunch charges in the nanocoulomb range have been demonstrated and have produced lasing. In the regenerative amplifier FEL demonstration at Los Alamos National Laboratory (LANL), the FEL reached an output power of 140 kW over a timescale of 10 microseconds (μs). This low-duty-factor, high-power FEL demonstration suggested that FEL amplifiers could potentially reach 100 kW continuous average power if high-power radio-frequency (RF) systems and high-duty-factor accelerators were used to provide the electron beams to drive the FEL. The low-duty-factor advanced FEL facility is still available for doing proof-of-principle experiments to test new ideas



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 14
3 Technical Assessment: Scalability to One-Megawatt Power Levels HOW TO ACHIEVE 100 KILOWATTS AND ONE MEGAWATT The current level of FEL power in the near-infrared wavelength range (around 1 micrometer) is 14 kW, established at Jefferson Laboratory. The next step proposed by the Navy program is to demonstrate and study a 100 kW FEL system to establish the technology needed for scaling to the megawatt power level in the infrared. Assuming well-established undulator technology, the undulator period will be in the range of ~3 to ~5 cm, so that the electron beam energy needed to reach infrared wavelengths will be in the range of ~80 to ~120 MeV. The Jefferson Laboratory FEL system, with an energy recovery linac, has established that an energy extraction of a few percent (~2 percent) can be obtained, while inducing an energy spread of ~10 percent. This means that the necessary average current recirculating in the energy recovery system must be ~1 A for a megawatt-class FEL and ~0.1 A for a 100-kW-class FEL. Jefferson Laboratory now achieves its 10 kW operation by recirculating electron pulses of about 0.1 nC at a repetition rate of about 75 MHz. Therefore, the path forward in the new 100 kW FEL will need to achieve average recirculating currents of around 0.1 A, and the path forward in the megawatt-class FEL will need to achieve average recirculating currents of around 1 A. The increase in average current can be achieved by increasing either the bunch charge or the bunch frequency, or both. To achieve a 0.1 A current for the 100 kW FEL, a bunch charge of 0.1 nC can be produced at an increased frequency of 750 MHz, or the bunch charge can be increased to 1 nC at the same frequency of 75 MHz. Some technical issues depend more on the average current, while others depend more on the peak current in the electron bunches. The lower bunch charge of ~0.1 nC and the associated problems have already been explored at Jefferson Laboratory, while the increased bunch charge of ~1 nC may well lead to new technical issues. The 100 kW FEL will require exploration of the higher average current of around 0.1 A at 750 MHz (“filling every bucket”) but can also involve exploration of the generation and transport of ~1 nC bunches at the lower repetition rate of 75 MHz. To scale the power to the MW class, it will be necessary to increase both the pulse charge to ~1 nC and the pulse repetition frequency to ~750 MHz. Bunch charges in the nanocoulomb range have been demonstrated and have produced lasing. In the regenerative amplifier FEL demonstration at Los Alamos National Laboratory (LANL), the FEL reached an output power of 140 kW over a timescale of 10 microseconds (μs). This low-duty-factor, high-power FEL demonstration suggested that FEL amplifiers could potentially reach 100 kW continuous average power if high-power radio-frequency (RF) systems and high-duty-factor accelerators were used to provide the electron beams to drive the FEL. The low-duty-factor advanced FEL facility is still available for doing proof-of-principle experiments to test new ideas 

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS for high-gain FELs. It consists of a 20 MeV integrated photoinjector capable of 5 nC bunch charge at a repetition rate of 108 MHz, or 0.5 ampere average current over 10 μs. At Argonne National Laboratory, the Argonne Wakefield Accelerator with an L-band high-Q injector has demonstrated the world’s highest charge per bunch. In single bunch operation, the system has demonstrated a micropulse charge that is tunable between 1 nC and 100 nC, a current of 10 kA, an energy of 15 MeV (using one 30 MW klystron), and an energy of 30 MeV (by adding a second 30 MW klystron). In bunch train operation, the system has demonstrated four bunches × 50 nC and 64 bunches × 50 nC, 50 ns long (needs a cesium telluride cathode), and a beam power of 1.5 GW. It is important to achieve the 2 to 3 percent extraction mentioned above; otherwise the recirculating beam current will have to be increased further beyond the already impressive 1 A. FEL simulation and experiments indicate that more than 2 percent is possible with sufficiently good electron beam quality. This extraction would require something around a ~20-period undulator for the oscillator design and a ~200-period undulator for the amplifier design. Both the oscillator and the amplifier undulators can be tapered to increase extraction. 1,2,3 In fact, the amplifier must be tapered to reach the 2 percent required extraction. The untapered amplifier will only achieve about 0.5 percent extraction (both simulation and experiment show this) and would therefore require substantially more average beam current to reach the same laser power levels. The motivation for tapering to increase extraction can be seen in the resonance condition, which has already been described. As the average electron beam energy decreases, reducing the Lorentz factor in the denominator of the resonance condition, electrons go out of resonance, beginning the saturation process in strong optical fields. A “trick” to extend resonance is to increase the undulator gap, reducing the undulator magnetic field and hence the value of the undulator parameter, K, in the numerator of the resonance condition. The tapering trick has been dem- onstrated in a number of experiments and many simulations. Often, the tapering does not start until about halfway down the undulator, thereby allowing the FEL to reach strong optical fields near saturation in the first half. While tapering can be used in the amplifier to increase the extraction to the 2-3 percent level, there is also an induced energy spread, as in the untapered case. This induced energy spread cannot be excessive and is con- sidered to be limited to about 10-15 percent because of two important processes in the recirculating FEL. First, bending an electron beam around a 180-degree arc is difficult with a beam containing a large range of energies, and hence bending angles in the dipole magnetic field. The second process is the deceleration of an electron beam with a large energy spread. A fractional momentum spread of 10 percent in the 100 MeV beam becomes roughly 200 percent when the beam is decelerated to the injection energy, typically 5 MeV. Such a large momentum spread can exceed the acceptance of the downstream beam line, causing particle loss. Further, the large energy spread on the decelerating beam causes the longitudinal phase space to be curved, as the particles in the beam occupy a large range of the RF phases of the cavity fields. Beyond a certain limit of the energy spread, these nonlinear distortions of the phase space can cause some of the low-energy particles to get lost in the last RF cavities and not arrive at the exit of the linac. Experience at the Jefferson Laboratory FELs has shown that, for proper energy recovery, the nonlinear distortions must be corrected. So in the Jefferson Laboratory FEL, the optics of the recirculator are set up to impart not only a linear position-energy correlation, but also a quadratic dependence of the fractional momentum spread on the longitudinal position upstream from the linac, which compensates the RF-induced curvature. At the Jefferson Laboratory FEL, these corrections are done with sextupole magnets. The details of the process are too lengthy to include in this report, but are described in Piot et al. 4 Optical sidebands can be generated in high-peak-power FELs. The sideband power can be significant and is a second laser line about ~1/N, or ~1 percent away from the fundamental frequency on the long-wavelength side. It is caused by the mixing of the oscillation frequency of electrons trapped in strong optical fields with the fundamental frequency of the FEL. This sideband generation has been observed in experiments and simulations and is the result of strong optical fields at saturations in each micropulse, not high average power. Both the FEL oscillator and the amplifier may experience sideband generation for the parameters considered in this report. Their presence in the laser beam could be seriously detrimental to propagation through the atmosphere, since windows of low absorption tend to be narrow. If the fundamental FEL wavelength was in such a window, the sideband would experience significant absorption, leading to thermal blooming. Fortunately, the sideband instability can be controlled in a few ways. First, tapering the amplifier, or even the oscillator, in FEL configurations tends

OCR for page 14
 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY to reduce or remove the sidebands. Secondly, adjusting the resonator mirror separation by a small amount (1 part in 107) will desynchronize the bounce time of the optical pulses and the electron pulse repetition frequency so as to remove the sidebands. Also, the output coupling of the FEL oscillator resonator can be increased to remove the sidebands. Finally, the sideband power may be removed by optical means after the FEL interaction. In all, it is not considered a problem to remove the sidebands leaving only the power in the FEL fundamental wavelength. In general, removing the sidebands is accomplished by “turning down” the FEL interaction so that saturation occurs in strong optical fields, but not excessively strong fields. For narrow-band applications, the FEL would be expected to rely on the normal FEL power at saturation. It is important to note that the final obstacle to the achievement of 14 kW at Jefferson Laboratory was thermal distortion of the mirrors. Mirror coatings (absorption and damage) and mirror cooling remain challenging issues. Of particular concern is increased absorption by the optical coatings caused by the UV harmonics of the high-power laser. These issues are more severe when the peak current (and gain) are lower and must be balanced against the injector and beam-transport problems (such as coherent synchrotron radiation) associated with higher peak current. The Jefferson Laboratory FEL is still available for experiments in the oscillator configuration. The average beam current is ~10 mA, with energy of 110 MeV, operating at a wavelength of 1.6 µm. It is important at each level of development to establish a solid connection between experiment and simulation and modeling. Simulations should be established to have a record of predicting experimental observations as well as to explain the observed experimental results. It is only with validated simulations that scaling from the 10 kW level, achieved now, to the 100 kW level and eventually to the megawatt class can be established with adequate confidence. Many codes have been benchmarked with each other as well as to experimental results, but some areas of the system are not modeled well. A discussion of simulation and modeling capabilities and challenges is provided in a separate section, after the following detailed discussion of FEL system blocks. END TO END BY SYSTEM BLOCKS The following sections discuss the FEL from end to end by system blocks. These system blocks are shown in Figure 2.1 in Chapter 2. It should be noted that the path to optimization of the overall system to achieve the parameters specified in the charge to this report is not necessarily that of optimizing each block. This is because there are trade-offs between blocks that allow the requirements on one to be relaxed at the expense of another, and vice versa. For example, by increasing the injector requirements to increase gain, one may use an amplifier and relax the need for high-damage-threshold oscillator mirrors. Other trade-offs exist between the current and the voltage of the accelerator and between repetition rate and charge per pulse. Furthermore, optimization of the system for a weapon application, taking into account the constraints of shipboard operation not considered here, would be different still. Nevertheless, the committee believes this block analysis is a useful way to assess where there are gaps between the current state of the art and the needs of a megawatt FEL. Electron Gun Systems There are three varieties of electron guns—direct current (DC) high-voltage (HV); normal-conducting (NC) radio-frequency (RF); and superconducting RF (SRF)—employing one of three different types of cathodes (thermionic-, field-, and photoemission cathodes). A photoinjector uses a laser-switched photocathode in one of the above electron guns with a booster that accelerates the beam to an energy of several MeV, which allows optimized control of injection of the electron beam into the main accelerator in an energy recovery linac (ERL) configuration. In all cases, a high-average-current electron gun should produce a continuous train of electron pulses. The repetition rate of the electron pulses should be equal to, or a subharmonic of, the RF of the accelera- tor. For optimum acceleration in an RF field, each electron pulse should be much shorter than the RF period. If we assume a nominal RF for the accelerator of 700 MHz, a 1 A average current would require electron bunches that each contain 1.4 nC of charge repeated at a 700 MHz repetition rate. Lower average currents can be achieved either by reducing the repetition rate of the electron bunches, or by reducing the charge per bunch, or both.

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS The committee believes that the most expeditious pathway forward to the 100 kW average-power FEL, scalable to the megawatt class, is a photoinjector, because the FEL requires a very well controlled series of electron pulses matched into the accelerator capable of achieving repetition rates in the gigahertz range.5,6 The electron gun system in block diagram is shown in Figure 2.1. A megawatt-class FEL will require a 1 A average-class electron beam. The Boeing gun of the early 1990s holds the record of 32 mA for approximately 3 hr.7,8 The characteristics of this gun are close to those required for 100 kW FEL operation. The Boeing NC RF photocathode gun, operating with a multialkali photocathode, described below, is the state of the art for the NC RF gun. The next-ranked state of the art is the DC HV gun, with a cesiated GaAs photocathode, also described below. The SRF photocathode guns are ranked as least state of the art. It is difficult to de-couple the photocathode from the electron gun when describing the state of the art as it is truly a system of components that is required to produce the initial beam. Several state-of-the-art laser systems were already used to produce average currents in the NC RF and DC HV systems mentioned above. Laser systems are already being improved in terms of power levels and repetition rates through investments such as in the U.S. Department of Energy Small Business Innovation Research (SBIR) program 9 to accommodate the high-repetition-rate ERL light sources of the future. The specifications they are set to deliver in 2009, based on ongoing work and improvements, are as follows: • Repetition rate: arbitrary; • Pulse duration: 50 picoseconds (ps) is straightforward, 10 ps is in development; • Average power: 60-100 W green; and • Peak power (determined by pulse duration and repetition rate): achieving high peak power becomes more challenging as one goes from 2 kW to 5 kW to 10 kW and higher. 10 To summarize, the drive laser, the photocathode material, and the electron gun work together to ensure a high- quality beam of sufficient energy to lock in the quality. The challenging part is ensuring that the choice of these three components delivers the parameters required by the FEL—the peak and average current, the transverse and longitudinal beam emittances, and the energy spread. Photocathodes Photocathode materials have been researched, developed, and tested extensively over the past 20 years for many free-electron laser systems.11 Progress from 100 kW to megawatt FELs will require short (10 ps) electron pulses from the cathode. To produce such short pulses of high beam quality sufficient for acceleration, transport, and lasing, photoemission driven by a high-quality drive laser pulse with a pulse length on the order of 10 ps, with sufficient energy, and of a certain wavelength is required to overcome the materials’ work function. A suitable cathode will have a high enough quantum efficiency to produce sufficient electron current at a reasonable drive laser power and wavelength and will have a lifetime commensurate with operational requirements. The quantum efficiency is the number of electrons released compared to the number of incident photons. A key issue for the 100 kW FEL scalable to the 1 MW class is the robustness of the cathode—that is, the total charge that can be delivered over a sustained period of operation. The robustness depends on the cathode material, with the most efficient cathodes generally found to be the most fragile. Achieving the 100 kW FEL power requires approximately 100 mA of average electron current (e.g., 0.14 nC per pulse at a 700 MHz repetition rate) and the megawatt-class FEL would require 10 times the charge per bunch, yielding 1 A average current. The committee considers the Boeing photocathode (a multialkali K2CsSb photocathode) to be the state of the art, producing an average cur- rent of 32 mA over 3 hours of operation. In addition, there are several other promising pathways to achieving the 100 mA and 1 A average currents. Currently, the Jefferson Laboratory FEL system is capable of producing an average current of 10 mA for extended periods of time using cesiated GaAs cathodes. While cathodes that meet the quantum efficiency requirement of a megawatt-class system have been developed, their robustness and lifetime are not yet at levels that would make them suitable for long-term use. For the 100 kW and megawatt-class FELs, the cathode is one of the most challenging components of the high-average-power FEL. This is due to both knowledge and technological issues. First, it is difficult to maintain

OCR for page 14
 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY sufficient vacuum pumping in the environment of electron guns that have large electrical gradients at the cathode for acceleration. Second, this gradient leads to electrical breakdown at times that degrades the cathode. Third, it is challenging to continuously refresh or change the cathode due to the ultrahigh vacuum requirements. The knowledge of how to develop an ultrarobust cathode is still being developed. Photocathode Drive Lasers To achieve a 100 mA to 1 A average current beam, the per-pulse charge should be between 0.1 and 1 nC, with pulse length of around 10 ps off the cathode and at a pulse repetition rate between 100 MHz and 1 GHz. The electron beam can be compressed later in a magnetic chicane to increase the peak current per pulse. Drive lasers suitable for driving cathodes to produce 100 mA of average current have been produced; however, no attempt has been made to develop a drive laser suitable for 1 A average current. A photocathode drive laser capable of producing a 1 A average current from a cathode with a quantum efficiency of 2 percent will require approximately 100 W of green laser power to the cathode, which is approximately five times the average power of the current state-of-the-art drive laser at Jefferson Laboratory. The cathode must be adequately cooled to handle such a high incident drive laser power. Electron Guns As was mentioned, there are three varieties of electron guns: direct current (DC) high-voltage (HV); normal- conducting (NC) radio-frequency (RF); and superconducting radio-frequency (SRF). The ideal gun should have excellent vacuum characteristics to preserve the cathode lifetime and a high accelerating gradient to maintain the electron beam quality. DC HV systems are able to achieve excellent vacuum; however, the accelerating gradient is currently limited to less than 6 MV/m, which limits the charge per bunch capability to <1 nC. The Jefferson Labo- ratory DC HV gun has the highest average current (10 mA) of any operating gun. NC RF guns offer the prospect of higher accelerating gradient (close to 10 MV/m) and, consequently, a charge per bunch capability in excess of 1 nC; however, poor vacuum has the potential to limit cathode lifetime. The performance of NC RF electron guns at high duty factor is limited by ohmic heating of the structure. The Boeing NC RF gun holds the record in average power, approaching the requirements for the 100 kW device. SRF guns offer the prospect of both high accelerating gradient (>20 MV/m) and superb vacuum characteristics, with the prospect of excellent photocathode lifetime. To date, SRF guns have only been tested at very low average currents (<1 mA).12 At present, the limiting factor in SRF gun design is the mounting of the cathode in the gun. Several high-average-current SRF guns are in design, and one is under construction.13 Booster The booster is a high-current, non-energy-recovered linac section that boosts the energy of the gun for accel- eration by the ERL. The booster is located between the electron gun and the beam merger of the ERL. Technically, the booster is considered a part of the injector. It may be either a distinct unit or part of the electron gun. The parameters of the booster are an energy gain of a few MeV (between 2 and 8 MeV) at the full current of the linac, which may be up to 1 A. The booster is characterized by a very high RF power input since it is not energy recovered. The present parameters achieved by the Jefferson Laboratory FEL booster are about 10 mA at 7 MeV. The goals of the Cornell booster are 100 mA and an energy gain of 15 MeV; however, this booster is still in the commissioning phase and the parameters have not yet been demonstrated. The objective of the booster is to accelerate the beam rapidly to the energy level at the entrance of the ERL. This is important in order to minimize the emittance growth, both longitudinal and transverse (emittance growth degrades performance). The energy at the end of the booster is determined by the minimum required to achieve efficient energy recovery in the presence of a large energy spread induced by the FEL interaction on the one hand and the energy gain required for stabilizing the emittance growth on the other. The booster design energy gain is limited by the fact that this energy is mostly dumped, thus increasing the power consumption and complicating

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS the beam dump. This consideration places a large premium on reducing the energy of the booster. There is also a penalty for going over 9.91 MeV (the neutron generation threshold in copper). Since some electrons will be dumped at energies exceeding the injector energy, the booster energy should be below 9 MeV. Ongoing research efforts at Jefferson Laboratory and Brookhaven will help determine how low an energy level can be used in the booster. Even though the booster requires a very high RF power input (typically the bulk of the RF power required by the FEL), it is still anticipated to be superconducting RF, primarily to increase the overall efficiency of the FEL, increase the energy rapidly, and reduce the footprint of the FEL, but also to improve the vacuum in the vicinity of the gun. While no booster has been demonstrated at the energy and current required for a high-power FEL, the committee sees and anticipates no physics or knowledge issue associated with achieving the required parameters. Given the freedom to break the booster into a number of smaller cavities (as is done in the Cornell injector), there is no technical issue associated with the coupling of RF power or the handling of higher-order-mode power. A higher-order mode is a cavity mode in the accelerator other than the desired acceleration mode. Higher-order modes are generated by the electron beam and can have undesirable electromagnetic fields that can kick the electron beam and lead to beam disruption. This makes it important to design the cavities such that higher-order-mode power is removed and dissipated quickly. Merger Optics The “merger” is an electron beam optical device composed of magnet beam optical elements. It is an essential element of the same-cell ERL. It serves the function of merging the low-energy beam from the injector with the high-energy beam returning from the FEL, such that both will be directed along the axis of the ERL accelerat- ing (and decelerating) cavities. The merger is located at the entrance of the ERL and is also the last low-energy (arguably the injector) beam element. The main concern with the merger is the minimization of emittance growth. One reason for this emittance growth is that the merger system mixes transverse and longitudinal degrees of freedom and consequently violates emittance compensation conditions. Several merger schemes are in use, such as a reverse bend (Jefferson Labora- tory FEL), a “chicane” (Budker Institute of Nuclear Physics FEL), and a “dog leg” (Japan Atomic Energy Research Institute FEL). All of these mergers introduce some emittance growth. A new merger system, the “zigzag,” is under construction at the Brookhaven National Laboratory (BNL) ERL, which should introduce the least emittance growth. (More information on mergers can be found in Litvinenko et al. 14) The merger does not present any technological or scientific issues. However, one should note that the zigzag merger, which is the only merger presenting a negligible emittance growth, has not been demonstrated experimentally. Energy Recovery Linac The ERL is the element that accelerates the beam from the injection energy to the FEL energy and then recovers (most of the) energy before dumping the beam. Because of this dual function—acceleration and deceleration—the ERL is continually traversed twice by the beam and thus appears in Figure 2.1 between the injection beam merger and the FEL and then again between the FEL and the beam dump. Even though the ERL provides most of the energy for the FEL, its RF power consumption is low thanks to the energy recovery feature. However, the ERL represents the largest load on the liquid helium refrigeration system, which runs continuously. The ERL is composed of one or more cryomodules. A cryomodule is a cryostat containing accelerating cavities and ancillary equipment such as tuners, couplers, and higher-order-mode (HOM) loads. In addition, the ERL requires medium-power RF power units and a cryogenic system. The ERL parameters are up to ~1 A of beam current and ~100 MeV of acceleration. The energy of 100 MeV has been demonstrated (in the Jefferson Laboratory FEL), but the highest current demonstrated so far is 20 mA (in the Budker Institute of Nuclear Physics FEL). There are many technological subjects of interest associated with the ERL. These include stable operation at the design current (beam breakup instability issues), attaining the operational gradient of the superconducting

OCR for page 14
0 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY cavities at a reasonable power consumption level for the refrigerator, dissipating and safely removing HOM power, and controlling vibrations (microphonics). A number of technical issues have been adequately addressed and are not of concern. One of them is the achievement of a high gradient at low cryogenic losses. There are some issues that have not been resolved, but recent developments lead the committee to believe that these issues do not represent a gap in physics or technology. In this category is the beam breakup instability, which has reliable solutions for a beam current of up to a few amperes. Couplers (at the ERL cavity level) and tuners are also not an issue. One should, however, consider microphonics. Mechanical vibrations and helium pressure changes and noise change the resonant frequency of the cavities by up to a few hertz. Since the superconducting cavities in an ERL are narrow-band devices with a Q of about 10 8 (a bandwidth of ~7 Hz), a large mechanical disturbance can drive a cavity away from the RF to the extent that the accelerating fields will either collapse or change enough to prevent the FEL from delivering power. This is considered a solved problem for laboratory-based machines but has not been considered for naval applications. The committee sees and anticipates no physics or technology gaps in this area, but because no ERL has operated at 100 mA (much less at 1 A), there is a certain knowledge gap. Radio-Frequency Couplers and Power Handling RF couplers are used to feed power to the cavities of the ERL and injector, extract higher-order modes from these cavities, and sample the field levels. The sampling (or pickup), as well as the fundamental power couplers (FPCs) of the ERL cavity (but not the injector cavities), are rather routine. Strong, HOM damping of high powers of monopole and dipole modes is essential. The higher-order-mode power can be of significant magnitude (up to kilowatts) and extends over a broad frequency range. The challenge is to ensure adequate damping of HOMs and the extraction of HOM power with good cryogenic efficiency. Several HOM extraction schemes have been proposed for broadband HOM damping, with power dissipated at room or intermediate temperatures (for example, 80 K). For power efficiency, the HOM power should be damped at room temperature without undue increase in the complexity or length of the cryomodules. There is sufficient experience with high-power, HOM damping in high-current storage rings, and a nice adaptation of such devices has been made at Cornell University for incorporation inside a cryomodule, but there is no operational experience with high-power, cryomodule-located, HOM dampers. Ampere-class cryomodule design and fabrication efforts (with appropriate HOM damping) are ongoing at BNL and Jefferson Laboratory. High-power fundamental power couplers for SRF elements, such as RF guns and booster cavities, have been built, but there is no operational experience at the megawatt level. (More information on RF couplers may be found in Rusnak.15) Based on the lack of operational experience with high-power fundamental power couplers and in-cryomodule higher-order-mode dampers, there is a knowledge gap in this area. Energy Recovery Linac Lattice and Peripherals: Transport Challenges The ERL lattice is a system of magnets that serves to transport the electron beam from the output of the linac cavities, through the FEL, and back to the linac for deceleration and energy recovery. The lattice also serves in other functions: matching the beam size and divergence into the wiggler and other components of the ERL; longitudinal phase space manipulations (if necessary); separating the decelerated beam from the accelerated beam to send it to the beam dump; and providing various beam diagnostic functions. The lattice is characterized by machine functions, such as the β function, phase advance, and dispersion. These functions are important for the stability of the ERL and its ability to transport the beam with minimal losses. Another challenge is to preserve the six-dimensional emittance. There are several other considerations, including the effects of coherent synchrotron radiation, halo and beam loss, and ion trapping. At the injection end, longitudinal space charge may lead to beam quality deterioration, particularly for very short bunches. The combination of short bunch lengths and high average currents of high-power FELs presents the tech- nical challenges of beam quality preservation and heat generation. The resistive-wall wakefields created by the

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS electron bunches can generate a high heat deposition in various locations. The FEL wiggler, where the walls are near the bunch and the bunches are at their shortest, is one such place. Bellows or other high-impedance elements must be shielded. Short, high-charge beam bunches interacting with the lattice may cause coherent synchrotron radiation and microbunch instabilities. These lead to emittance growth and severe performance penalties. The coherent radiation from short bunches can deposit significant power on devices where such high power can lead to performance degradation. It is important to control halo and beam loss in ERLs. Beam loss is a serious issue, since it can directly damage equipment, produce unacceptable increases in the vacuum pressure and cryogenic load, and lead to radia- tion hazards for equipment and personnel. Beam losses in the Jefferson Laboratory FEL have been quantified at ~10 mA operation. The total beam loss was <1 μA, <100 nA in the worst locations, and about 10 nA in some other locations. In ERLs that may operate at 10 to 100 times higher current, beam loss must be controlled to better than 1 ppm. Collimation at the injection site, a high-quality photocathode and gun, well-matched beam optics, and excellent machine protection systems will be critical for achieving this goal. Similarly with synchrotron light sources, ERLs need to diagnose and control short bunches at high average beam power, which implies noninvasive diagnostics that allow continuous monitoring of transverse and longitudinal beam properties, synchronization systems, and protection systems. Dipole HOMs in ERLs can pose a beam stability challenge. The beam and the RF cavities can form a positive feedback loop that closes when the beam returns to the same cavity. The feedback loop can lead to a transverse beam breakup (BBU) instability at sufficiently high currents, driven predominantly by the high quality factor of the superconducting cavities. The theoretical models for the BBU instability are mature and in excellent agreement with simulations and experiments. The BBU instability can be significantly ameliorated by specially designed RF cavities, operating at lower frequencies with strong HOM damping. Ionization of residual gas molecules by the electron beam creates an ion column, which can lead to a distortion of the accelerator optics and coupled oscillations of the beam and the ions, both troublesome to the FEL opera- tion. Good vacuum and possibly ion clearing electrodes may be important to minimize deleterious effects of the electron-ion two-stream instability (electron cloud instability). As this long list of potential technical issues indicates, the ERL lattice is far from being a simple system. (Addi- tional information may be found in Merminga.16) While there are no physics gaps, there are technical and knowledge gaps due to the required large increase in beam current from the known territory of 10 mA average beam current to the level of 100 mA and then 1,000 mA, envisaged for the 100-kW-class and 1-MW-class FELs, respectively. Undulator and Associated Pinch for Amplifiers The design and manufacturing techniques for undulators suitable for high-power FEL operation are well advanced. Most of the design features in existing FEL undulators for third- and next-generation light sources can be adapted for use in megawatt-class FELs as needed. There are no significant issues in manufacturing undulators with sufficient magnetic quality to produce lasing at the wavelengths under consideration. 17,18 The undulators for FEL amplifiers will be much longer than those for short Rayleigh-range oscillators. Accord- ingly, the alignment of the amplifier undulators will require more attention than that of the oscillator undulators. The knowledge base for high-power FEL oscillator performance at ~1 μm wavelength is much greater than that for amplifiers. A particular concern for amplifier operation is the need to pinch the electron and optical beam near the exit of the undulator in order to cause the optical beam to expand sufficiently before encountering any optical elements. While this concept has been proposed, it has not been tested experimentally. 19 Optical System Issues Introduction The optical gain of an FEL is strongly related to the peak current of the electron beam in the wiggler. For an oscillator, higher optical gain enables a higher optical out-coupling fraction and reduces the ratio of circulating

OCR for page 14
 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY intracavity power to out-coupled power. For example, the Jefferson Laboratory 14 kW FEL oscillator’s modest optical gain only allowed an out-coupling fraction of about 10 percent. This resulted in 140 kW of power on the cavity resonator optics. For an amplifier, increased optical gain allows the use of a smaller drive laser or a shorter wiggler. However, higher gain demands higher current or smaller emittance, which places greater demands on the injector and may introduce greater problems with coherent synchrotron radiation. In either case, the diameter of the electron beam and the optical beam in the interaction region is only a few millimeters. Diffraction over some distance is then required to allow expansion of the optical beam to the level where the oscillator’s resonator optics or the amplifier’s first relay optics can survive (typically an irradiance of 100 to 200 kW/cm2). Coatings The development of optical coating technology for high-power laser applications in the ultraviolet to near- infrared over the past 35-40 years has been focused largely on the requirements of the national laser fusion and isotope separation (atomic vapor laser isotope separation) programs. These applications involve periodic laser pulse operation: laser-fusion lasers operating at wavelengths from 0.3 to 1 μm produce ~1 ns pulses at much less than 1/s, and the atomic vapor laser isotope separation Cu vapor laser, operating at wavelengths of 511 nm and 578 nm, produces 50 ns pulses at 4 kHz. Unfortunately, the coating materials that have been developed to sur- vive these pulsed laser environments are not optimal for operation in either high-average-power cw lasers or in quasi-cw FELs, where the micropulse fluence (J/cm2) is low. From 1980 to 1990, optical system development for FEL applications was supported by the Strategic Defense Initiative (SDI) for a megawatt-class RF-linac FEL. Substantial progress was made during that time, but coating R&D ended when the SDI program was terminated in about 1991. However, a number of significant results were reported in the open literature and are discussed below. They are important for the development of megawatt-class FEL systems for naval applications. At LANL, Sanders and colleagues (1990) identified hafnium oxide/silicon dioxide (HfO 2/SiO2) multilayer dielectric reflectors, produced by ion-beam sputtered deposition (IBSD), as a prime candidate for meeting the requirements of megawatt-class FEL oscillators.20 The IBSD technology was a direct adaptation from the success- ful methods and experience in producing ultra-low-loss (<10 ppm) mirrors for laser gyros. Collaborations with the industrial scientists who originally developed the IBSD coating technology for laser gyro optical systems resulted in production of IBSD HfO2/SiO2 reflectors with reflectance >0.9999 at 1 μm wavelength.21 Sanders et al. reported that the cw damage resistance of these reflectors at 1 μm far exceeded the FEL system requirement. Two decades later, IBSD multilayer reflectors of HfO2/SiO2 have been installed in the Jefferson Laboratory FEL and have survived cw intracavity average power densities well over 100 kW/cm 2 with the FEL operated at 1.6 μm wavelength with <0.4 ps (FWHM) micropulses at 18.71 MHz repetition frequency. 22 In addition, Menoni reported to the committee that output-coupler mirrors of this type were not damaged by irradiation of up to 200 kW/cm2 with a 1.06 μm cw laser.23 In addition to needing a sufficiently high damage threshold, FEL oscillator optical systems must have very low absorption losses, approximately <10 ppm, to limit mirror substrate thermal loading and the resultant distor- tion that limits the operating power of the FEL. Currently available commercial IBSD reflectors for laser-gyro applications based on optical systems of, for example, TiO2/SiO2 and Ta2O5/SiO2, are advertised to have absorp- tion (and scattering losses) of <10 ppm at 633 nm.24 In her presentation to the committee, Menoni also reported that her laboratory group at Colorado State University had employed IBSD to produce HfO 2/SiO2 optical systems with <10 ppm loss at 1 μm for a high reflector and <14 ppm loss for an antireflection coated surface. 25 She also cited the absorption loss of 10 ppm for a 90 percent HfO2/SiO2 reflector (commercially produced) as measured at Jefferson Laboratory. While attainment of sufficiently high damage resistance and low absorption in candidate reflectors is very encouraging for scaling to 0.1-1 MW-class FEL oscillators, there are other optical system degradation mechanisms particular to FELs that must be overcome. One is the terahertz radiation generated by the short electron bunches interacting with the magnet fields within the oscillator. Absorption of this radiation in the mirror substrates leads

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS to distortion of the reflected optical wavefront, as observed in the Jefferson Laboratory FEL. Effective means to prevent the terahertz radiation from reaching the mirrors were reported. The second and probably most serious optical system degradation mechanism is UV-harmonic-induced infrared absorption in the multilayer dielectric reflectors. FEL oscillators and amplifiers inherently generate substantial power in the optical harmonics of the fundamental lasing wavelength, and the relative strength of these harmonics has been measured in a number of operating FELs. Mirror degradation caused by absorption of the UV harmonic radiation was observed in the Orsay storage-ring FEL in 1984,26 and more recently in the Jefferson Laboratory FEL. In both FELs, the resonator mirrors exhibited increased absorption at the lasing wavelength, which seriously limited the operating power. Sanders et al. measured the magnitude of UV-induced infrared absorption at 1 μm for a number of candidate multilayer reflector materials using excimer radiation at 353 nm and 248 nm.27,28 (These wavelengths corresponded to the third and fourth optical harmonics of the design wavelength.) Not surprisingly, the induced optical system absorption was more severe at 248 nm than at 353 nm. Fortunately, radiation at higher harmonics (shorter wave- lengths) and radiation on even harmonics is generally somewhat weaker than radiation on lower or odd harmonics. The summary of the growth and decay of the induced absorption caused by 248 nm radiation is shown in Figure 3.1. It is noteworthy that the HfO2/SiO2 reflector produced by IBSD using highly pure (0.9997 percent) HfO 2 exhibited the least amount of increased absorption, 0.03 percent (300 ppm). In her April 2008 presentation to the committee,29 Menoni cited a recent measurement in the Jefferson Laboratory FEL: <50 ppm absorption loss in the 90 percent HfO2/SiO2 output reflector coating induced by the optical harmonics. This was significantly less severe than had been experienced using other coating materials for the resonator mirrors. In summary, one coating material system—IBSD HfO2/SiO2 using ultrapure HfO2—has been developed that may well prove resistant to the multiple environmental hazards of FEL oscillators designed to produce output powers in the 100 kW to 1 MW class. Even with this materials system, the most serious limit on FEL circulating FIGURE 3.1 248 nm ultraviolet irradiation at low average intensity (0.35 W/cm2) produced time-dependent increases in absorption (decreased reflectance) of five 1.1 μm multilayer dielectric reflectors (X/SiO2)NX. The magnitude and time dependence of the UV-induced absorption changes were strongly dependent on the high-index oxide layer. Laser irradiation was in 10 ns pulses at 35 Hz with 10 mJ/cm2 pulse fluence. Reflectance was monitored at the design wavelength. SOURCE: V.E. Sanders, J.W. Early, and W. Leamon, “The Response of Multilayer Dielectric Coatings to Low Fluence Ultraviolet Light Exposure,” Laser Induced Damage in Optical Materials: —Proceedings of the Boulder Damage Symposium, Noember -,  (Washington, D.C.: U.S. Government Printing Office, 1990), pp. 561-567.

OCR for page 14
 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY power is UV-induced infrared absorption. Further optimization of the optical system properties should be pur- sued, and other candidate materials should be evaluated. Since no dielectric material will be totally immune from degradation by the FEL harmonics, ways (such as gratings) to redirect the intracavity harmonic power should be explored. In addition, ongoing and future optical-systems work in other programs, such as the Advanced Laser Inter- ferometer Gravitational-Wave Observatory (LIGO) project, in which very low loss mirrors are being pursued for use on test masses, may also have relevance for FEL applications. The project’s interferometer is expected to have cw power levels of ~1 MW at 1.064 μm, and even though UV harmonics will not be involved in this application, the parallel work on coated optics for Advanced LIGO may well reduce the gap between the current state of the art and the needs of the Navy. (Some researchers working in this area are Harry,30 Chalkley et al.,31 Markosyan et al.,32 and Sun et al.33) There are a number of possible “work-arounds” for the optical system problem. Ways to address the threat posed by UV-harmonic-induced laser absorption (and, less likely, laser damage) of the resonator mirrors of an FEL oscillator include (1) ion-beam sputter coating R&D using very pure HfO 2 sputter targets for producing HfO2/SiO2 multilayer mirrors with minimized UV-induced absorption (the level possible to be determined); (2) exploration of other candidate materials for the high-index layers; (3) an FEL resonator design that has a sufficiently large diameter laser beam on the cavity mirrors (a more unstable resonator cavity); and/or (4) incorporation of a silver- coated intracavity grating at a large angle of incidence if such could be shown to survive the intracavity power. However, the most conservative and certain path for bypassing optical system degradation risks associated with an FEL oscillator will be (1) an FEL amplifier configuration driven by a synchronous seed laser or (2) a master-FEL oscillator/power amplifier (MOPA) for producing the high-power beam or (3) a regenerative FEL amplifier (large holes in the mirrors of a recirculation resonator, mirrors that would see very low FEL power, essentially only at laser start-up). The last-mentioned scheme would be the most compact of the alternatives. Oscillators As mentioned earlier, FEL resonators operate at a very low Fresnel number. Provided that hole output coupling is not used, such resonators discriminate strongly against higher-order modes and produce excellent output beam quality. Although the index of refraction and gain of the electron beam in the wiggler can distort the optical mode in the resonator, computer simulations and experimental measurements show that the optical beam has very good quality. Generally the resonator is long and slender, since the optical beam must be long and slender to fit through the undulator. In addition, the resonator is typically much longer than the undulator to give the optical beam the opportunity to expand before it reaches the first optical surface in order to reduce the irradiance incident on the surface. It is found that the Rayleigh range can be much smaller than the length of the undulator to make the beam expand faster.34 This is important, since damage to the first optical surface is the most important issue for FEL resonators. In addition, increasing the gain reduces the burden on the optical system. In the first place, higher gain makes it possible to increase the output coupling, which reduces the recirculating power in the resonator. In the second place, higher gain corresponds to higher peak electron beam current, which reduces the necessary duty factor of the electron beam and, correspondingly, the duty factor of the optical beam on the optical surface. Many options are available for the optical architecture of the resonator. Before the Navy’s FEL average power scale-up program began in 1995, the highest demonstrated average power was ~10 W. The first laser in the Navy’s scale-up program was a 1 kW oscillator built at Jefferson Laboratory. It successfully employed a near-concentric resonator at 6 μm, with a 100 percent reflector on one end and a ~10 percent transmitting out-coupler on the other. This same approach was used by Jefferson Laboratory for the next step (10 kW at 1.6 μm), but this encountered multiple difficulties with optical coating survivability and distortion under thermal load. It was learned that the resonator’s total optical distortion needed to be <λ/10 rms. The approach that ultimately succeeded involved a back-plane-cooled high reflector (with thermal management to control the radius of curvature under thermal load) and a sapphire 10 percent out-coupler with cryogenic edge cooling for thermal and figure control. As the average power of the FEL increases with the 100 kW Innovative Naval Prototype (INP) and follow-on megawatt upgrade, a beam-splitting out-coupler may become unusable due to the requirements to maintain optical

OCR for page 14
 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY essentially single-pulse experiments, they have produced many useful results on saturation of tapered wigglers, beam quality, bunch compression, and other issues. In place of the resonator, the amplifier FEL requires a master oscillator, but 10 W average power (10 kW peak) should be sufficient. The optical quality of the input beam should not be an issue since the beam will be optically guided by the electron beam inside the undulator, and the output beam quality will not be dependent on the quality of the input optical beam. However, the master oscillator must operate at a wavelength determined by the atmo- spheric windows for propagation in a maritime environment. This increases system complexity and will require technology development, as discussed later (see the section on tunability). To achieve the saturated gain necessary for a useful amplifier, it will be necessary to use a long undulator, on the order of 5 m or 10 m.38,39 To propagate the optical beam this far in the narrow dimensions of the undulator, the optical beam will be gain-guided by the electron beam. In the absence of any special measures, the optical beam will arrive at the end of the undulator highly collimated, which will place a very high irradiance on the first optical surface or require it to be widely separated. Sprangle et al.40 have proposed a potential solution, which is to pinch the electron beam near the end of the undulator by placing additional focusing magnets in this region. This will cause the optical beam also to pinch near the end of the undulator, and the smaller-diameter optical beam will expand by diffraction more rapidly when it emerges from the undulator. They calculated that the combination of beam pinching and grazing incidence at the first relay mirror (possibly curved to obtain additional expansion) could make it possible to place the first mirror within ~3 m of the undulator. Their design would require a peak current of 2 kA, which increases the demands on the injector and the degradation of the beam by coherent synchrotron radiation. Other FEL amplifier designs involving pinching the electron beams with lower peak current, e.g., 1 kA, 41 have been proposed. Experiments to test these ideas will be needed to determine their feasibility. The variety of potential solutions enhances the potential for a workable solution. Regenerative Amplifier Alternative to the Master Oscillator Power Amplifier (MOPA) The regenerative amplifier FEL (RAFEL) is a hybrid FEL configuration with the combined features of an oscillator and a high-gain amplifier. The key idea is to feed back a very small fraction (<10 percent) of the optical power generated in the high-gain undulator (~1,000 per pass) to enable the FEL to reach saturation in a few passes. This effectively eliminates the need for a transmissive beam splitter out-coupler/feedback cavity optic. It uses a compact ring resonator (four off-axis reflectors), thereby eliminating the need for a seed input from a master oscillator. This makes it easier to select or tune the operating wavelength. With high single-pass gain, the electron beam causes optical guiding that determines the optical mode, largely independent of the ring resonator. Since the output optical beam quality is only weakly dependent on the input beam quality, a hole in the center of the entrance mirror is used to pass the electron beam on axis. This eliminates the chicane normally used and avoids the concomitant terahertz radiation and electron beam degradation. A special feature of this optical system is that essentially all of the high-power laser radiation is coupled through a large hole in the output mirror. The optical feedback from the output annular mirror is taken from wings of the optical beam, where the intensity is much smaller, so mirror loading is not an issue. This configuration allows a separate output window to be mounted at a suitable distance from the mirror to transmit the beam to the relay optics. Beam-induced thermal distortion of the window must be adequately compensated, but it should not restrict the operation of the laser resonator. A RAFEL operating in the infrared at 16 μm at LANL exhibited a very large, small-signal gain of ~330 per pass and an average output power of 200 kW during each 8 μs macropulse.42 Other notable characteristics included the very large ~1 mm cavity detuning range and large ~1 mrad mirror alignment tolerance, both of which made the resonator cavity relatively insensitive to vibrations. The spectral width of the output was ~6 percent full width at half maximum, as expected for LANL’s low-Q resonator. With sufficient electron beam energy and average current, the RAFEL should be scalable to megawatt output average powers at 1 μm. As with the single-pass amplifier option, an issue for the RAFEL is determining a means to rapidly expand the intense output beam so that its intensity can be handled safely by the first downstream relay reflector. Pinching the electron beam at the exit of the wiggler and use of a grazing-incidence reflector are two

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS approaches being considered. Also, the system will need to be designed to produce a sufficiently narrow laser spectral width for transmission through the selected atmospheric window. Beam Dump The beam dump’s purpose is to stop the electron beam after it has been decelerated by the ERL. It is located at the end of the electron trajectory, which begins at the photocathode. The beam dump has to handle an energy that can somewhat exceed the injector energy (since some electrons are accelerated by the FEL interaction) at a current that is practically the whole injector current. In addition, the beam dump must perform its function over a large range of electron energies and maintain good vacuum conditions. An additional requirement is good ionizing- radiation shielding, given the electron’s high energy and the high currents that are stopped at the beam dump. Beam dumps of 1 to 2 MW are rather routine; for example, the beam dump of a megawatt cw klystron can handle over 1.5 MW. Some adaptation will be needed to account for the higher electron energy, which is more than enough for a 100 kW FEL. There does not appear to be a technical gap to extending beam dump technology to the necessary few megawatts of power (depending on the FEL efficiency) that will be required by a megawatt-class FEL. Tunability FELs are inherently tunable by varying the electron beam energy or certain parameters of the wiggler and have been built to operate from millimeter wavelengths down into the soft x-ray region. Individual FELs have been demonstrated to tune by a factor of 10 in wavelength. However, as the average power of an FEL increases, the limit on its tunability becomes the wavelength range over which the resonator optics retain their high reflectivity. For the Navy, tunability during operation is not an issue, because very precise wavelength control must be maintained to stay at the desired low-absorption spot in any one of the three possible bands. (Figure 3.2 shows typical atmo- spheric molecular absorption and scattering versus wavelength. The absorption curves show that the transmission widths around the desired wavelengths are up to several percent and much broader than the natural bandwidth of the FEL, though sideband production can make absorption an issue; see the discussion in the first section of this chapter.) But it may prove desirable for the Navy to be able to select among the three spectral bands during use, depending on the meteorological conditions at the moment (this remains to be seen). Of the three candidate wavelengths of interest for naval applications (1, 1.6, and 2.2 μm), molecular scattering is somewhat worse at 1 μm than at the other two wavelengths, and the differences in aerosol scattering among them are to be determined. Molecular absorption at 1 μm is a clear winner, but aerosol absorption is to be determined. In addition, the eye damage threshold is substantially lower at 1 μm than at 1.6 or 2.2 μm. Research in propagation is under way to help with the wavelength choice. The ability to select the desired wavelength(s) during the design phase is a fundamental FEL oscillator attri- bute and is quite achievable by proper design of the FEL parameters and resonator coatings. If switching among the various atmospheric windows, which range from 1 to 2.2 μm, becomes essential, the design of the resonator, coatings, and out-coupler substrate will become far more challenging. High-reflectivity dielectric coatings with multiple high-reflectivity bands have been built, but they come at the expense of additional coating layers and higher overall absorptivity. The required coating improvements to allow multiband operation fall into the area of technology issues. For an FEL amplifier, the issues are quite different. This device requires a drive oscillator to properly seed the amplifier with the proper wavelength and pulse format. Overall gain greater than 10 6 is generally difficult to deal with owing to parasitic oscillations. If the amplifier is assumed to have a gain of ~105, the drive laser must supply approximately 10 W of average power in mode-locked pulses with a pulse repetition frequency (prf) of 500-700 MHz synchronized to the RF drive, a pulse width of a few picoseconds, and a peak power on the order of 10 kW. Solid-state (slab or fiber) lasers can approach some of the requirements at 1.045 μm but not easily at 1.6 or 2.2 μm. For the longer wavelengths, it probably will be necessary to shift the wavelength using nonlinear optical techniques such as parametric amplifiers. This would require technology development. In principle, a low-power FEL could be used to drive the amplifier, but this also adds to system complexity.

OCR for page 14
 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY FIGURE 3.2 Typical atmospheric molecular absorption and scattering versus wavelength. SOURCE: Courtesy of Joung Cook and John Albertine. Controls Control of the FEL and its subsystems is important for its operational stability and for personnel and equip- ment safety. The control system is composed of the following: • Operator interfaces; • Equipment set points and readbacks; • Data acquisition, conversion and filtering, and analysis; • Closed-loop control; • Access security; • Sensors (including electron, photon, and radiation diagnostics); • Digital signal processing; • Timing and synchronization; • Equipment protection; • Alarm detection, reporting; • Automatic sequencing; and Operator assistance (save, compare, restore functions of data sets). 43,44 • The complexity and control issues of a high-average-power FEL are similar to those of numerous electron- accelerator-based scientific user facilities. The immediate control needs of a 100-kW-class FEL and a 1-MW-class FEL can be met through this extensive experience base with integrated control systems. Operational needs of

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS weapons-class FELs go beyond the experience base of the accelerator community; however, there is considerable overlap with the needs of some planned facilities, such as light sources based on energy recovery linacs. Active alignment systems have been developed that are better than 10 times what the community thinks will be needed for stability in the ultimate FEL system, but they have not actually been tested at sea with the vibration spectrum appropriate to ships. Brief discussions with engineering companies and experts suggest that this is not a showstopper, but it needs to be seriously addressed in the future and concurrently with the develop- ment of higher-power FELs for Navy applications. Until there is a proven design, or at least one that is closer to it, it would be premature to study the vibration issue in detail and then have to alter the study if the FEL design changes. However, it is clear that the unique operational environment of an FEL on a ship will call for specialized diagnostics that are well integrated with the control system—for example, component motion and vibration sensors and beam position monitors. For high average beam current operation, minimally intercepting or nonintercepting electron beam diagnostics should be developed to monitor and control, for example, coherent synchrotron radiation and related electron microbunching; electron beam halo, transverse emittance, and energy spread; and cathode performance. Existing accelerator facilities have made heavy use of trained operators to interface with the control system. Although there have been efforts to implement automation, the science machines still rely on operators and experts to interface between start-up, shutdown, and many automation routines. Therefore, limited effort has been put into true autonomous operation with real-time decision making and order variation steps based on failure modes. In a megawatt-class FEL, just as in some land-based accelerators, the amount of circulating electron and photon energy is substantial. This means that a fault that results in reduced operational availability or in uncontrolled beam loss could have significant adverse consequences for the safety of personnel and equipment. It will therefore be important to develop automated control systems for turnkey operation, rapid start-up and shutdown, availability on demand from a quiescent state, and recovery from fault conditions that go beyond the current state of the art. SIMULATION AND MODELING The committee recognizes that part of the assessment of current technology is the assessment of numerical simulation capabilities and their ability to predict performance in future experiments. The current state of the art varies from one system element to another, with very good modeling capabilities existing for some and others not being understood or quantitatively predictable. The committee observes that the limitations of particular simula- tions are, in many cases, a by-product of both hardware and modeling. Simulation and modeling associated with major system issues and components are discussed next. Injectors Because high-average-power FELs will most likely operate in a high-charge-per-bunch mode (~1 nC per bunch) and with a low accelerating gradient in the gun, the electron dynamics are likely to be much more dominated by time-dependent space-charge forces than they are in existing electron guns. This presents a greater challenge for simulation and modeling of the electron dynamics in the gun and injector. Short-pulse, space-charge-dominated beams are much more sensitive to the distribution function of the electrons than are less intense beams. In particular, the space-charge forces in a beam from a photoinjector are: • Time-dependent and transversely nonlinear (because of nonuniform beam density)—the beam pulse length is much less than the transit time in the gun or first accelerator cells; • Dependent on the beam distribution (transverse and temporal); and • Dependent on the cathode emission and drive-laser profile (transverse and temporal). Simple extrapolations from low space-charge results are useless. In order to have an accurate model of the self fields, any reasonable simulation must have realistic transverse and temporal beam distributions. Accordingly, there must be close interaction between simulation studies and experiments so that the simulations represent realistic

OCR for page 14
0 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY rather than idealized conditions. The limiting factor at present is the lack of realistic data on the electron beam distribution under the gun conditions required for a megawatt-class FEL electron gun. Coherent Synchrotron Radiation During the last decade, a number of coherent synchrotron radiation (CSR) simulation codes were developed based on different approaches to modeling CSR and different implementations. 45,46,47 The one-dimensional or projected approach incorporates one-dimensional CSR algorithms for the calcula- tion of the longitudinal forces. These algorithms are based on a formula for the energy change by a line-charge distribution in a dipole magnet, with extensions to downstream drift spaces. Codes based on this model tend to be very efficient, but the approach ignores physical effects such as transverse forces and ignores transverse beam dimensions. Examples of this approach include the codes Elegant48 and CSR_calc.49 Elegant’s high longitudinal resolution, which is enabled by the use of one-dimensional models, has led to discovery of the CSR-driven micro- bunching instability in the Linac Coherent Light Source (LCLS),50 and has allowed exploration of the instability involving CSR, longitudinal space charge, and longitudinal wakefields and possible cures such as strong wigglers or laser/undulator beam heaters. The macroparticle approach self-consistently models the self-interaction of the bunch, which consists of a set of macroparticles, by storing away the complete history of the macroparticles as the bunch travels along the beamline. This approach has been used in two-dimensional51 and three-dimensional models, and it can be time consuming. TraFiC4 52 and CSRtrack53 are examples of three-dimensional codes based on the macroparticle approach. CSRtrack incorporates the physics of TraFiC4 and uses new algorithms for the calculation of the CSR fields. Efficient calculation techniques have increased the speed of execution compared to the direct method for a large number of particles. CSR codes also exist that are based on a fully self-consistent Vlasov-Maxwell treat- ment.54,55 This approach has less noise than the standard particle approaches, but it is computationally intensive. Benchmarking among the various codes has been done for 1 nC Gaussian bunches at 5 GeV. 56 These investiga- tions show agreement on the emittance growth but significant differences in the relative energy loss obtained by all one-dimensional codes compared to macroparticle codes. The Vlasov-Maxwell calculation confirms the results of the macroparticle methods. Stronger deviations are noted in the case of 500 MeV, where the interference of the compression process with self-forces is much stronger. It is harder to validate the available CSR codes with experimental data because the experiments are difficult, and experimentally reconstructing the initial and final six-dimensional phase space in order to compare the data with simulation is challenging. Furthermore, other effects, such as space charge and wakefields, also affect the beam phase space significantly. Accordingly, the development and experimental validation of a code that models the coupled system are critical to improve the understanding of these effects and their significance to the design of a high-average-power FEL. Beam Halo Halo particle production in linacs is a consequence of filamentation in the six-dimensional phase space caused by nonlinear and time-dependent forces. The two-dimensional projections of the six-dimensional filaments appear as an extended beam halo. Although there is no consensus on a rigorous definition of “beam halo,” the term usually refers to particles extending beyond the normal beam rms radius, from a few rms beam radii to about 10 rms beam radii or more. Because beam halo increases the risk of beam losses with their multiple deleterious effects, control of halo and beam loss is important in high-average-power ERLs. It is important to measure and reduce the halo significantly for Navy shipboard applications. Computer simulations show that substantial halo is formed in high-current, mismatched linac beams. So far only a limited effort has been put into computer simulations of halo in high-current ERLs. Macroparticle simu- lations using the three-dimensional parallel PIC code IMPACT have been used to compare with experimental measurements of beam halo in the high-current, 67 MeV proton beam in the Low Energy Demonstrator Accelerator (LEDA) facility at LANL.57 These simulations were successful in reproducing the core of the measured matched

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS beam profiles and the trend of emittance growth as a function of the mismatch factor, but they underestimated the growth rate of halo and emittance for mismatched beams. The discrepancy was attributed to a lack of knowledge of the detailed input distribution in the six-dimensional phase space of the injected beam. 58 In an ERL with nanocoulomb bunch charge, the beam is in a highly space-charge-dominated regime, and mismatches can be present because of either beam optics or alignment errors. In order to predict the fraction of the beam that may end up in halo, one needs realistic computer simulations that include machine errors and detailed knowledge of the initial particle distribution from the source. Experimental measurements of beam halo on operat- ing high-power FELs and other electron accelerators are critical for benchmarking the simulations and identifying all of the physical mechanisms for halo generation in high-current electron linacs. Beam Breakup Multipass beam breakup (BBU) in recirculating linacs has been observed and studied for a long time. Many computer simulation codes have been developed in laboratories around the world. There are two main approaches to modeling beam breakup: (1) beam tracking, which calculates the beam position as a function of time and searches the threshold current by changing the beam current, and (2) the eigenvalue method, which converts the beam transport equation to an eigenvalue equation and solves the eigenvalue problem. Codes based on the first approach include TDBBU and ERLBBU, developed at Jefferson Laboratory; BBU-R, developed at JAERI; and BI, developed at Cornell. The code MATBBU, which was also developed at Jefferson Laboratory, is based on the eigenvalue method.59 Cross-benchmarks among the codes show consistent beam behavior and excellent agreement with theoretical models. In a series of comprehensive measurements at the Jefferson Laboratory FEL, the beam breakup threshold current was experimentally determined and found to agree with simulations to within ±10 percent. Overall, beam breakup simulations are mature and have provided reliable predictions. They can confidently be used to derive specifications for the higher-order-mode damping requirements of future high-power-ERL FELs. FEL Simulation Codes FEL simulations have been developed over three decades and compared to many FEL experiments with successful validation. The most sophisticated simulations are now four-dimensional, including all the optical field components in x, y, and z followed in time t for an unlimited number of bounces of the optical pulse between the mirrors or in a single pass through the undulator for the amplifier. Some of these are GENESIS, MEDUSA, WAVEV, GINGER, TDA3D, RON, FRED, and FELIX. The last two, FRED and FELIX, have been inactive for a number of years, while the others have been active, simulating several different experiments and proposed FEL systems. Various types of optical resonators have been modeled. The number of electrons is sampled but can now approach ~1 percent of all the electrons in an electron pulse. All forces from the optical and undulator fields, as well as coulomb forces, can be included, as well as higher-frequency optical harmonics. The basic physics is described with the self-consistent, relativistic Lorentz force and optical wave equations, assuming a slowly varying amplitude and phase as is appropriate for a laser. Even quantum noise and shot noise have been included as well to model start-up of the x-ray FELs. Reviews of several FEL codes describe their attributes and make comparisons.60,61 Nearly all laboratories around the world have some form of FEL simulation. These programs have varying degrees of sophistication, some running on laptops and others running on large clusters. The FEL models have been successfully applied to a remarkably large range of FELs, from centimeter wavelengths to x-ray wavelengths, without a fundamental change other than the input parameters. The simulations are accurate in strong optical fields and with high or low growth rates. The limitation of all FEL simulations is in determining accurate input information describing the initial elec- tron beam and optical fields. The full description of the electron beam’s six-dimensional phase space includes all initial positions and velocities. Electron spin is not included and is typically negligible. Determining the initial electron and optical parameters is often limited by experimental diagnostics.

OCR for page 14
 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY FEL Start-to-End Simulation Codes In the preceding sections, the committee established the importance of reliable and experimentally verified simulations for all the important physical effects that enter into the design of a high-power-ERL FEL. While these individual simulations are necessary for a credible design, they are not sufficient. Effects from one part of the machine can interfere destructively or constructively with effects from another part of the machine. In some cases, this interference may be a design choice used to cancel or partially compensate an effect. For this reason, a care- fully constructed and experimentally validated start-to-end simulation code that takes into account space charge, CSR, resistive wall effects, and RF cavity and other wakefields together is necessary to define the performance of the entire system. This level of modeling will also provide insights into halo production and control. Two FEL system codes have been developed, FELSIM (Advanced Energy Systems) and INEX (Boeing, LANL). NOTES 1. N.M. Kroll, P.L. Morton, and M.N. Rosenbluth, “Variable Parameter Free Electron Laser,” Physics of Quantum Electronics 7: 89-112 (1980). 2. N.M. Kroll, P.L. Morton, and M.N. Rosenbluth, “Free-Electron Lasers with Variable Parameter Wigglers,” IEEE Journal of Quantum Electronics QE-17: 1436-1468 (1981). 3. W.B. Colson, C. Pellegrini, and A. Renieri, eds., Free Electron Laser Handbook (The Netherlands: North- Holland Physics, Elsevier Science Publishing Co. Inc., 1990). 4. P. Piot, D.R. Douglas, and G.A. Krafft, “Longitudinal Phase Space Manipulation in Energy Recovering Linac- Driven Free-Electron Lasers,” Physical Reiew Special Topic—Accelerators and Beams 6: 030702 (2003). 5. A. Todd, “State-of-the-Art Electron Guns and Injector Designs for Energy Recovery Linacs (ERL),” Nuclear Instruments and Methods in Physics Research Section A 557: 36-44 (2006). 6. G.H. Hoffstaettera, V. Litvinenko, and H. Owen, “Optics and Beam Transport in Energy Recovery Linacs,” Nuclear Instruments and Methods in Physics Research Section A 557: 345-353 (2005). 7. D.H. Dowell, K.J. Davis, K.D. Friddell, E.L. Tyson, C.A. Lancaster, L. Milliman, R.E. Rodenburg, T. Aas, M. Bemes, S.Z. Bethel, P.E. Johnson, K. Murphy, C. Whelen, G.E. Busch, and D.K. Remelius, “First Operation of a Photocathode Radio Frequency Gun Injector at High Duty Factor,” Applied Physics Letters 63: 2035-2037 (1993). 8. D.H. Dowell, S.Z. Bethel, and K.D. Friddell, “Results from the Average Power Laser Experiment Photocathode Injector Test,” Nuclear Instruments and Methods in Physics Research Section A 356: 167-176 (1995). 9. Andrew Brown, Aculight Corporation, private communication. 10. Roy Mead and Pratheepan Madasamy, Aculight Corporation, private communication. 11. C. Hernandez-Garcia, M.L. Stutzman, and P.G. O’Shea, “Electron Sources for Accelerators,” Physics Today 61: 44-49 (2008). 12. R. Xiang, H. Buettig, P. Evtushenko, D. Janssen, U. Lehnert, P. Michel, K. Moeller, C. Schneider, R. Schurig, F. Staufenbiel, J. Teichert, J. Stephan, W.-D. Lehmann, T. Kamps, D. Lipka, I. Will, and V. Volkov, “Status of 3½ Cell Superconducting RF Gun Project in Rossendorf,” Proceedings of 00 Particle Accelerator Confer- ence, Knoxille, Tennessee (2005), pp. 1081-1083. Available at http://ieeexplore.ieee.org. 13. I. Ben-Zvi, A. Burrill, R. Calaga, X. Chang, R. Grover, R. Gupta, H. Hahn, L. Hammons, D. Kayran, J. Kewisch, R. Lambiase, V. N. Litvinenko, G. McIntyre, D. Naik, D. Pate, D. Phillips, E. Pozdeyev, T. Rao, J. Smedley, R. Than, R.J. Todd, D. Weiss, Q. Wu, A. Zaltsman, M. Cole, M. Falletta, D. Holmes, J. Rathke, T. Schultheiss, R. Wong, and A. Todd, “Superconducting Photoinjector,” Proceedings of the th International FEL Conference, Noosibirsk, Russia (2007), pp. 290-293. Available at www.JACoW.org. 14. V.N. Litvinenko, R. Hajima, and D. Kayran, “Merger Designs for ERL,” Nuclear Instruments and Methods in Physics Research Section A 557: 165-175 (2006). 15. B. Rusnak, “RF Power and HOM Coupler Tutorial,” Proceedings of the th Workshop on RF Superconductiity, September -, 00, Lubeck, Germany (2003). Available at http://srf2003.desy.de/fap/paper/TuT02.pdf.

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS 16. L. Merminga, “Energy Recovery Linacs,” Proceedings of PAC0, Albuquerque, New Mexico (2007), pp. 22-26. Available at http://ieeexplore.ieee.org. 17. E. Trakhtenberga, V. Tcheskidovb, I. Vassermana, N. Vinokurovb, M. Erdmanna, and J. Pfluger, “Undulator for the LCLS Project—From the Prototype to the Full-Scale Manufacturing,” Nuclear Instruments and Methods in Physics Research Section A 543: 42 (2005). 18. Geoff Pile, Marion White, Emil Trahktenberg, and Stephen Milton, Argonne National Laboratory, private communication. 19. P. Sprangle, B. Hafizi, and J.R. Peñano, “Design of a Compact, Optically Guided, Pinched, Megawatt Class Free-Electron Laser,” IEEE Journal of Quantum Electronics 40: 1739-1743 (2004). 20. V.E. Sanders, J.W. Early, and W. Leamon, “The Response of Multilayer Dielectric Coatings to Low Fluence Ultraviolet Light Exposure,” Laser Induced Damage in Optical Materials: —Proceedings of the Boulder Damage Symposium, Noember -,  (Washington, D.C.: U.S. Government Printing Office, 1990), pp. 561-567. 21. Ibid. 22. M.D. Shinn, S.V. Benson, D. Douglas, P. Evtushenko, C. Gould, J. Gubeli, D. Hardy, K. Jordan, G.R. Neil, D.W. Sexton, and S. Zhang, “Operation of the Jefferson Laboratory Free-Electron Laser at High Irradiances,” Laser-Induced Damage in Optical Materials: 00 (Washington, D.C.: U.S. Government Printing Office, 2008), abstract and relevant Q&A page in abstract section at end of volume. 23. C.S. Menoni, Colorado State University, “Towards 1 MW FELs—Demands on the Resonator Optics,” pre- sented to the committee on April 4, 2008. 24. General Optics, www.generaloptics.com. 25. C.S. Menoni, Colorado State University, “Towards 1 MW FELs—Demands on the Resonator Optics,” pre- sented to the committee on April 4, 2008. 26. P. Elleaume, M. Velghe, M. Billardon, and J.M. Ortega, “Diagnostic Techniques and UV-Induced Degradation of the Mirrors Used in the Orsay Storage Ring Free-Electron Laser,” Applied Optics 24: 2762-2770 (1985). 27. V.E. Sanders, J.W. Early, and W. Leamon, “The Response of Multilayer Dielectric Coatings to Low Fluence Ultraviolet Light Exposure,” Laser Induced Damage in Optical Materials: —Proceedings of the Boulder Damage Symposium, Noember -,  (Washington, D.C.: U.S. Government Printing Office, 1990), pp. 561-567. 28. J. Early, V. Sanders, and W. Leamon, “The Response of Multilayer Dielectric Coatings to Low Fluence UV Light Exposure,” Laser Induced Damage in Optical Materials: —Proceedings of the 0th Symposium on Optical Materials for High-Power Laser,  (Washington, D.C.: U.S. Government Printing Office, 1989), pp. 233-244. 29. C.S. Menoni, Colorado State University, “Towards 1 MW FELs—Demands on the Resonator Optics,” pre- sented to the committee on April 4, 2008. 30. G. Harry, “Progress in LIGO Coating Development,” presented at the Workshop on Optical Coatings in Preci- sion Measurements, California Institute of Technology, Pasadena, Calif., on March 20-21, 2008. Available at www.ge.infn.it/~gemme/virgo/CoatingWorkshop/harry.html. 31. E. Chalkley, R. Bassiri, J. Hough, I. Martin, I. McLaren, S. Reid, S. Rowan, A. Woodcraft, H. Armandula, M.M. Fejer, R. Route, K. Vijayraghavan, P. Wu, A. Gretarsson, S. Penn, and G. Harry, “The Mechanical Loss of Thin-Film Hafnia as a Function of Temperature,” presented at the Workshop on Optical Coatings in Preci- sion Measurements, California Institute of Technology, Pasadena, Calif., on March 20-21, 2008. Available at www.ge.infn.it/~gemme/virgo/CoatingWorkshop/chalkley.html. 32. A. Markosyan, R. Route, and M. Fejer, “Measurement of Sub-ppm Absorption in LIGO Coatings,” presented at the Workshop on Optical Coatings in Precision Measurements, California Institute of Technology, Pasadena, Calif., on March 20-21, 2008. Available at www.ge.infn.it/~gemme/virgo/CoatingWorkshop/markosyan.html. 33. K.-X. Sun, N. Leindecker, A. Markosyan, R. Route, S. Buchman, M.M. Fejer, R.L. Byer, H. Armandula, D. Ugolini, and G. Harry, “Effects of Ultraviolet Irradiation to LIGO Mirror Coatings,” presented at the Work- shop on Optical Coatings in Precision Measurements, California Institute of Technology, Pasadena, Calif., on March 20-21, 2008. Available at www.ge.infn.it/~gemme/virgo/CoatingWorkshop/sun.html.

OCR for page 14
 SCIENTIFIC ASSESSMENT OF HIGH-POWER FREE-ELECTRON LASER TECHNOLOGY 34. W.B. Colson, J. Blau, R.L. Armstead, P.P. Crooker, R. Vigil, T. Voughs, and B.W. Williams, “Short Rayleigh Length Free Electron Lasers,” Physical Reiew Special Topics—Accelerators and Beams 9: 030703 (2006). 35. H.E. Bennett, “Optical Figure Requirements for Laser Mirrors Used at Oblique Incidence,” Laser Induced Damage in Optical Materials:  (Washington, D.C.: U.S. Government Printing Office, 1984), pp. 228-233. 36. D.H. Dowell, M.L. Laucks, A.R. Lowrey, J. Adamski, D. Pistoresi, D.R. Shoffstall, A.H. Lumpkin, S. Bender, D. Byrd, R.L. Tokar, K. Sun, M. Bentz, R. Burns, J. Guha, and W. Tomita, “Final Results of the Boeing and Los Alamos Grazing Incidence Ring-Resonator Free Electron Laser Experiment,” Nuclear Instruments and Methods in Physics Research Section A 318: 74-80 (1992). 37. E. Marshall, “Researchers Sift the Ashes of SDI,” Science 263: 620-623 (1994). 38. A. Murokh, R. Agustsson, M. Babzien, I. Ben-Zvi, L. Bertolini, K. van Bibber, R. Carr, M. Cornacchia, P. Frigola, J. Hill, E. Johnson, L. Klaisner, G. Le Sage, M. Libkind, R. Malone, H-D. Nuhn, C. Pellegrini, S. Reiche, G. Rakowsky, J. Rosenzweig, R. Ruland, J. Skaritka, A. Toor, A. Tremaine, X. Wang, and V. Yakimenko, “Properties of the Ultrashort Gain Length, Self-Amplified Spontaneous Emission Free-Electron Laser in the Linear Regime and Saturation,” Physical Reiew E 67: 066501 (2003). 39. D.C. Nguyen and H.P. Freund, “Possibility of a High-Power, High-Gain FEL Amplifier,” Nuclear Instruments and Methods in Physics Research Section A 507: 120-124 (2003). 40. P. Sprangle, B. Hafizi, and J.R. Peñano, “Design of a Compact, Optically Guided, Pinched, Megawatt Class Free-Electron Laser,” IEEE Journal Quantum Electronics 40: 1739-1743 (2004). 41. D.C. Nguyen, H.P. Freund, and W. Colson, “High-Power Free-Electron Laser Amplifier Using a Scalloped Electron Beam and a Two-Stage Wiggler,” Physics Reiew Special Topics—Accelerators and Beams 9: 050703 (2006). 42. D.C. Nguyen, R.L. Sheffield, C.M. Fortgang, J.C. Goldstein, J.M. Kinross-Wright, and N.A. Ebrahim, “First Lasing of the Regenerative Amplifier FEL,” Nuclear Instruments and Methods in Physics Research Section A 429: 125-130 (1999). 43. J. Carwardine, Argonne National Laboratory, “A Perspective on Control Systems for Ship-board FELs,” pre- sented to the committee on April 4, 2008. 44. Sincrotrone Trieste/INFN/ENEA Timing and Synchronization Workshop, Trieste, Italy, March 26-28, 2008. Available at http://www.elettra.trieste.it/Conferences/2008/TSW/. 45. G. Bassi, T. Agoh, M. Dohlus, L. Giannessi, R. Hajima, A. Kabel, T. Limberg, and M. Ouattromini, “Overview of CSR Codes,” Nuclear Instruments & Methods in Physics Research. Section A 557: 189-204 (2006). 46. M. Dohlus, “Modelling of Space Charge and CSR Effects in Bunch Compressor Systems,” Proceedings of EPAC 00, Edinburgh, Scotland (2006), p. 1897. Available at www.JACoW.org, Edinburgh JACoW. 47. P. Emma, Stanford Linear Accelerator Center, “CSR Effects on the Quality of High Brightness Electron Beams,” presented to the committee on April 4, 2008. 48. M. Borland, “Elegant: A Flexible SDDS-Compliant Code for Accelerator Simulation,” Adanced Photon Source Light Source Notes 287 (2000). Available at http://www.aps.anl.gov/Science/Publications/lsnotes/ls287. pdf. 49. G. Stupakov and P. Emma, “CSR Wake for a Short Magnet in Ultrarelativistic Limit,” Proceedings of EPAC 00, Paris, France (2002). Available at http://accelconf.web.cern.ch/Accelconf/e02/PAPERS/WEPRI029. pdf. 50. M. Borland, Y.-C. Chae, S. Milton, R. Soliday, V. Bharadwaj, P. Emma, P. Krejcik, C. Limborg, H.-D. Nuhn, and M. Woodley, “Start-to-End Jitter Simulations of the Linac Coherent Light Source,” Proceedings of the Particle Accelerator Conference 4: 2707 (2001). 51. R. Li, “Self-consistent Simulation of the CSR Effect,” Nuclear Instruments and Methods in Physics Research Section A 429: 310-314 (1999). 52. T. Limberg, A. Kabel, and M. Dohlus, “Using TraFiC4 to Calculate and Minimize Emittance Growth Due to Coherent Synchrotron Radiation,” Nuclear Instruments and Methods in Physics Research Section A 455: 185-189 (2000).

OCR for page 14
 TECHNICAL ASSESSMENT: SCALABILITY TO ONE-MEGAWATT POWER LEVELS 53. M. Dohlus and T. Limberg, “CSRtrack: Faster Calculation of 3-D CSR Effects,” Proceedings of the 00 FEL Conference (2004), pp. 18-21. Available at www.JACoW.org. 54. R. Warnock, G. Bassi, and J.A. Ellison, “Vlasov Treatment of Coherent Synchrotron Radiation from Arbitrary Planar Orbits,” Nuclear Instruments and Methods in Physics Research Section A 558: 85-89 (2006). 55. G. Bassi, J.A. Ellison, and R. Warnock, “Progress on a Vlasov Treatment of Coherent Synchrotron Radiation from Arbitrary Planar Orbits,” Proceedings of 00 Particle Accelerator Conference, Knoxille, Tennessee (2005), pp. 2699-2701. Available at http://ieeexplore.ieee.org. 56. International Committee for Future Accelerators (ICFA) Beam Dynamics Mini-Workshop on Coherent Synchrotron Radiation, at DESY-Zeuthen, Berlin, Germany, January 14-18, 2002. Available at http://www. desy.de/csr. 57. J. Qiang, P.L. Colestock, D. Gilpatrick, H.V. Smith, T.P. Wangler, and M.E. Schulze, “Macroparticle Simula- tion Studies of a Proton Beam Halo Experiment,” Physical Reiew Special Topics—Accelerators and Beams 5: 124201 (2002). 58. T. Wangler, National Superconducting Cyclotron Laboratory, “Halo in High-Current Beams,” presented to the committee on April 4, 2008. 59. E. Pozdeyev, C. Tennant, J.J. Bisognano, M. Sawamura, R. Hajima, and T.I. Smith, “Multipass Beam Breakup in Energy Recovery Linacs,” Nuclear Instruments and Methods in Physics Research Section A 557: 176-188 (2006). 60. S.G. Biedron, G. Dattoli, W.M. Fawley, H.P. Freund, Z. Huang, K.J. Kim, S.V. Milton, H.-D. Nuhn, P.L. Ottaviani, and A. Renieri, “Impact of Electron Beam Quality on Nonlinear Harmonic Generation in High-Gain Free-Electron Lasers,” Physical Reiew Special Topics—Accelerators and Beams 5: 030701 (2002). 61. H.-D. Nuhn and C. Pellegrini, eds., Proceedings of the X-Ray FEL Theory and Simulation Codes Workshop (Stanford, Calif.: Stanford Linear Accelerator Center, 1999).

OCR for page 14