Upgrades to higher power and shorter wavelength will require

1. higher electron beam energy (implying addition of new module(s) and/or improvements in cavity gradient)
2. higher beam current, demanding
   • source and injector development (particularly in the electron gun and RF windows)
   • accelerator cavity development (improvement of instability thresholds, possibly by using lower frequency cavities; RF window development)
3. FEL development
   • different wigglers
   • improved optical cavities
4. characterization of possibly detrimental effects such as CSR and space charge

The key point in the design of all future upgrades is that FEL output power is not obtained solely at the expense of large increases in single bunch charge and peak current. Instead, the bunch repetition rate is increased (from 37.5 MHz up to as high as the RF frequency), the energy raised and the FEL extraction efficiency raised from the nominal 0.5% of the 42 MeV baseline design. For example, IF a suitable source were available to provide 135pC bunches to the driver at 1500 MHz, the 42 MeV machine could potentially run not at 1 kW, but at 80 kW, by accelerating on each RF cycle (bunch rep rate of 1500 MHz rather than 37.5 MHz) and increasing the extraction efficiency to 1% (with output momentum spread approaching 10%), rather than 0.5%. The trick is to find a 400 mA CW 1500 MHz source and injector RF windows that will handle 2 MW CW forward power!

On this page, we will discuss the more speculative upgrades in the Jefferson Lab FEL family tree. Each "species" will be addressed in turn, and the physical constraints and design issues at each stage discussed.

A. 75 MeV Broad-band FEL(s)

B. 100 MeV 1kW IR FEL

C. The Branch in the Road - UV and High Power IR

D. 1 kW UV Demo

E. 100 kW UV Industrial FEL

F. 1 MW IR Super-Ultra-Mega-Deathray

As noted in our discussion of the 75 MeV upgrade, moving the wavelength reach of the IR FEL down to 1 micron by addition of a second cryomodule requires use of an upgraded optical cavity. This will probably be an R5 based resonator, which will not allow the generation of light at the longer wavelengths available in the 42 MeV system. To reach 3-6.6 microns, either a) the short wavelength FEL will have to be removed, and the longer wavelength system reinstalled, or b) the longer wavelength system will have to be placed in the system at some other location. The former option has significant operational impact, the latter implies design development.

Certain fundamental assumptions hold. It is technically difficult to develop a concept in which both FELs could operate simultaneously, due to phase space degradation at the first laser. We therefore will not pursue this option. As long as CSR is not well characterized, it is assumed an FEL will be downstream of the linac, and on the linac axis. Configuring the system to operate with two wiggler/optical cavity systems in such an arrangement would require either excessive length and/or a complex system of transverse electron beam bypasses for one or both of the lasers. One could, for example, attempt to put the systems side by side, which would be difficult in the available space.

Upgrading to a broad-band, two-laser system will therefore be an attractive option only if the second (now assumed to be longer wavelength) laser can be located in the backleg. Such a system is sketched below; it would be built only after CSR/space-charge studies are completed and if they indicate that emittance dilution effects do not preclude operation of the longer-wavelength FEL.

Given a 75 MeV 1 kW FEL and a characterization of CSR and space charge effects demonstrating a backleg-wiggler option is viable, a 100 MeV ~1 kW upgrade is immediate. The FEL would be moved to the backleg, an additional cryomodule dropped in to yield the desired system. Some optics design work would be required to develop beam handling to and from the wiggler, as a fully symmetric energy recovery arc transport system is no longer used. This is not anticipated to be a fundamental problem, as this issue was addressed in early 42 MeV system designs. The resulting system will be configured as shown below.

Industrial and military applications now begin to part company. To this point, the source performance required for the 42 MeV IR demo has been adequate to drive the upgrades. With only modest improvements in these sources (primarily by insuring emittance conservation), a 1 kW UV FEL becomes feasible. Assuming, once more, the irrelevance of CSR and the desired small improvements in source technology, a UV FEL with 1 kW output can be driven by raising the machine energy to 200 MeV. No increase in average or single bunch current is needed. The configuration of such an upgrade is discussed below.

Military applications of FEL technology, with their vastly higher light power requirements, will require significant research and development of several accelerator systems. A 100 MeV accelerator is adequate for such systems, but the average current must be increased to order 1 A, implying significant source development both in terms of repetition rate (from 37.5 MHz to the full RF frequency) and single bunch current (from 135 pC to ~2 nC for a 500 MHz RF system), must occur. The RF system will need to be more robust, with windows capable of handling much higher power and improved control of collective instabilities such as BBU and wakefield effects. The latter constraint probably will force a need to go to lower frequency cavities. The FEL itself will require development, with optical cavity improvements to handle extremely high power levels and operation at modestly higher extraction efficiencies. The accelerator beam transport may, at these higher single bunch currents, need to accommodate space charge effects.

A evolutionary scenario dealing with these issues is described below. We must note that industrial system concerns are not totally decoupled from military interests. Source technology developments for high power IR applications would in principle be applicable to high power UV applications (with further constraints on beam quality - a better emittance is required). They therefore pave the way to a 100 kW UV industrial source.

A 100 MeV, order 1 kW IR driver can be readily upgraded to a 1 kW UV driver, provided

1. improvements are made in source technology to allow generation of beams with somewhat better emittance,
2. CSR/space charge effects allow placement of an upgraded wiggler/optical cavity system in the accelerator backleg, and
3. RF windows, drive systems and cavity impedances are adequate to allow multipass operation of the system at currents of order 5 mA.

An upgraded system of this type is shown below. The machine concept is that developed as the original Jefferson Lab Industrial UV Demo. The 100 MeV system is used as an "first pass" of a two-pass 200 MeV machine. The beam is accelerated slightly off crest, the first recirculation run isochronously, and the first end loop of the second recirculation used as a bunch compressor to provide an appropriately configured 200 MeV phase space beam to the FEL. The spent beam is energy-compressed on the first decelerating pass, so as to lie within the momentum acceptance of the low energy recirculation transport. The FEL optical system now comprises a wiggler and optical cavity some 60 m in length, and is intended to handle 1 kW UV light power. It therefore represents some development of optical resonator technology beyond that used in the IR devices. General accelerator and FEL system parameters remain otherwise similar (e.g., similar single bunch momentum spread, charge, peak current, somewhat improved emittance, similar extraction efficiency, similar repetition frequency, average and peak power, similar required momentum acceptance) to those present in the IR designs.

We plan development toward the goal of a 100 kW UV production industrial FEL following the successful operation of the 200 MeV 1kW UV Industrial Demo. Such a machine would be used for photo- and chemical- processing applications requiring high intensity UV radiation. A potential parameter set for such a machine is as follows:

• Final energy: 200 MeV
• Beam Current: 50 mA
• Beam Power: 10 MW
• FEL Extraction Efficiency: 1%

These parameters imply significant challenges for each of the accelerator and laser systems; several limitations must be overcome.

1. The electron beam source and injector must be developed:
   • To increase the current from 5 mA to 50 mA, the bunch repetition rate must go from 37.5 MHz to ~350 MHz. This allows use of the fixed 135 pC/bunch charge and peak current developed for IR applications, thereby limiting single bunch space charge and CSR effects to manageable levels. This performance is well beyond the limit of the present Jefferson Lab photo-cathode DC gun, and thus implies significant gun development, perhaps demanding the use of an RF or SRF gun.
   • If an "IR driver class" 10 MeV injector is to be used to limit space charge induced beam quality degradation, RF windows must be improved to accept power levels of 50 mA x 5 MV = 250 kW for each cavity, well beyond the presently implemented 25 kW windows. This must be done for much lower frequency cavities, as discussed below.
2. The SRF cavity systems must be developed:
   • To manage impedance driven collective effects at these high currents, cavities at lower frequency than 1.5 GHz will be needed. The development effort may rely on APT-style 750 MHz cavities.
   • To make the system compact and limit parts count, SRF cavity gradients will need to be increased.
   • To handle the increased RF power required by higher beam current, gradient, and extracted FEL power, more robust RF windows will have to be developed.
3. FEL systems must be developed:
   • Optical cavities capable of handling 100 kW CW UV light will need to be invented.

Depending on gun developments, injector RF window developments, and collective effects lessons learned at 42 MeV, one might envision an RF or SRF gun at a few MeV injecting directly into a linac consisting of some APT cryomodules. We note that the use of lower frequency cavities will allow acceleration of longer bunches within the allowed momentum spread. This acts to reduce the required longitudinal charge density and thus limits space charge effects. Peak currents required for lasing can then be produced by bunch compression at full energy rather than by acceleration of short, high-charge density bunches. To reduce cost, complexity, and size, we expect that a multipass configuration with energy recovery (not unlike the 1 kW UV demo) will be used. The precise "numerology" of the driver will depend on the performance of available SRF cavities, electron sources, and optical cavities available at the time of the project, but a system configuration similar to that shown for the 1 kW demo, in the preceding figure, is likely to be conceptually accurate.

The US Navy holds an interest in the development of a 1 MW IR FEL for shipboard self-defense purposes. Such a device would be used in tactical applications for the "processing at a distance" of anti-ship missiles. Parameters typical of such a system are given below.

• Final energy: 100 MeV
• Beam Current: 1 A
• Beam Power: 100 MW
• FEL Extraction Efficiency: 1%

Development of this device presents several challenges beyond those presented to a 100 kW UV industrial system.

• Generation of a beam with high average and peak current and high single bunch charge
• Operation of SRF cavities in a high current environment
• Improvement of SRF cavity impedances and gradients
• Transport and phase space management of beams though a compact system in the presence of space charge
• Evolution of FEL/optical cavity technology to manage average powers five orders of magnitude beyond present levels
• Development of high-reliability accelerator systems

A system configuration for this machine is shown above. At these very high power levels (in a shipboard environment) energy recovery is vital; the machine therefore has the familiar recirculating footprint. The obstacles are numerous and fascinating. The machine is to be compact and robust. Use of such high currents implies the SRF cavities must be low frequency so as to avoid impedance driven instabilities; compactness requirements demand high gradients (in excess of 10 MV/m), a significant challenge in such cavities. APT 750 MHz cavities may provide the first step toward such an object. The large RF powers required to drive a megawatt output power implies significant development of RF windows must be made.

The desire to use a high average- low peak-power scenario implies the source must produce bunches at the RF frequency, perhaps at 750 MHz. Comparison to the parameter sets of other upgrades suggests therefore that the single bunch charge must be 1.35 nC, an order of magnitude higher than in other applications. This implies significant source development. Injector design is also a problem; use of a "high" energy injector based on an electron gun, buncher and preaccelerator (~10 MeV) is problematic from the point of view of RF windows - a single pair of SRF cavities accelerating 1 A through 10 MV will have to have windows handling 5 MW forward power each - far beyond the 25 kW encountered in the 42 MeV 1 kw IR system. Window R and D is therefore vital, and use (on the advice of young Candidate D. Engwall) of a high energy RF or SRF gun (at perhaps 2-3 MeV output energy) with direct injection into the main accelerator may be necessary.

As this is a low frequency system, the large single bunch charge charge can be distributed through a relatively long bunch length, alleviating some of the obvious concerns one must have about space charge, but not eliminating them entirely. Beam transport and phase space management of the high quality beam required for laser applications in the presence of space charge must be provided. This is to be done, moreover, in a compact footprint, which tends to aggravate CSR and space charge emittance dilution.

FEL and integrated accelerator system issues are of importance as well. The FEL is to produce 100,000 times more average power than has been done to date. Optical cavity performance in such a parameter regime is not well characterized. The system is, in fact, a weapon; it must therefore be robust, easily maintained and readily operated. All three of these properties are not normal attributes of accelerators (though they are features of much lower power medical machines). Opportunities for study of the fundamental physics of beams, machines, FELs, and system designs therefore abound in this project.

Last modified: 10 March 1997
http://www.jlab.org/~douglas/ is maintained by: douglas@jlab.org