The idea of an energy-recovery linac (ERL) traces back to Maury Tigner in 1965. Accelerating two beams, colliding them, and then dumping them is extremely inefficient. After using the high-power particle beam in an experiment, instead of re-circulating the beam and losing brightness or dumping the beam power, the general objective of the ERL technique is to recover the beam power in the accelerating cavities for the next beam to be accelerated without requiring additional power from the grid and in order to achieve optimal brightness with this next beam.
The bacis principle of the ERL method is summarized on the above figure. Like in all accelerators, the first step for an ERL-based accelerator is to produce the particles (electrons or positrons) and to prepare the particle beam to be accelerated. The particle beam is accelerated in RF cavities where energy is given to the particle beam. The particle beam is thereafter used in an experimental area, where typically only a very small fraction of the particles in the beam participate in the experiment. Rather than dumping the remaining beam, in an ERL-based accelerator the remaining particles in the beam are re-directed back to the accelerating cavities but with a phase change of 180 degrees. Accordingly, the particles will experience the decelerating phase of the RF field rather than the accelerating phase. The energy of the particle beam is therefore recovered in the accelerating structure, and is ready to be used to accelerate the next particle beam from the injector without additional power requirements (#1 in the figure). With multiple accelerating turns followed with multiple decelerating turns, higher beam energies can be and have been reached.
Another benefit is that fter the deceleration, the particle beam can be dumped at low energy, equivalent to the injector energy, such as to avoid nuclear activation thresholds in the beam dump system (#2 in the figure).
Important to note is that ERL-based accelerators combine the best of linear and circular accelerators. Like linear accelerators, for ERL-based accelerators the six-dimensional phase space of the beam is largely determined by the properties of the electron source, and the beam brightness is maintained from the injector (#3 in the figure). Like at storage rings, the ERL technique enables the capability for high average currents in the accelerator, and thus promises similar efficiencies.
To unlock the full impact of the ERL concept, it is important to underline the major advances in SRF technology during the last decades enabling high average current operation. Achieving cavity quality factors Q0 above 1010 has been a game changer for ERL.
ERL state-of-the-art
ERLs have reached an advanced state in projects around the world summarized in the bem energy versus average beam current figure below. Jefferson Lab holds several records for superconducting ERLs: highest current recirculated (9 mA) and highestrecirculating power (1.3 MW) at the FEL facility, and the highest energy recirculated (1 GeV) at CEBAF. These records were obtained with single pass energy recovery, but at the s-DALINAC facility in Darmstadt, two-pass energy recovery has been achieved, and at CBETA in Cornell, four-pass energy recovery was demonstrated, albeit at low current due to insufficient time for commissioning.
Elsewhere in the world, the ALICE facility in Daresbury operated successfully for several years, as did the cERL at KEK, which was later converted to industrial applications. In Mainz, the MESA facility is under construction with an ambitious experimental program.
The landscape of past, present, and proposed ERLs. The beam power isolines (dashed) show the need to go from ~1 MW (achieved at Jefferson Lab) to 10 MW at PERLE on the way to 1 GW at LHeC and FCC-eh.
The only room-temperature ERL is the Recuperator at the Budker Institute in Novosibirsk, which has been operating since 2004 with four recirculations of up to 30 mA and 42 MeV, feeding three separate FEL beamlines.
These successful ERL facilities have paved the way for the bERLinPro facility at HZB, Berlin and PERLE at the IJCLab at Orsay, both of which aim to push the state of the art in beam current (100 mA at bERLinPro) and in mutli-turn beam current (20 mA at PERLE) and beam power (10 MW at PERLE). Given the present ERL state of the art, these parameters appear to be well within reach.
The landscape of past, present, and proposed ERLs. The beam power isolines (dashed) show the need to go from ~1 MW (achieved at Jefferson Lab) to 10 MW at PERLE on the way to 1 GW at LHeC and FCC-eh.
