An Energy Recovery Electron Linac-on-proton Ring Collider

Electron-proton colliders with center of mass energies between 14 GeV and 30 GeV and luminosities at the 10 level have been proposed recently as a means for studying hadronic structure [1]. Electron beam polarization appears to be crucial for the majority of experiments. Two accelerator design scenarios have been examined in detail: colliding rings [2] and recirculating linac-on-ring [3]. Although the linac-onring scenario is not as well understood as the ring-ring scenario, comparable luminosities appear feasible, while the linac-on-ring option presents a significant advantage with spin manipulations. Rf power and beam dump requirements make the linac-on-ring option viable only if the electron linac recovers the beam energy, a technology demonstrated at Jefferson Lab’s IR FEL, with cw current up to 5 mA and beam energy up to 50 MeV [4]. We begin with a brief overview of the Jefferson Lab energy recovery FEL and summarize the benefits of energy recovery. The feasibility of an energy recovery linac–ring collider is investigated and two conceptual point designs are shown. Luminosity projections for the linac–ring scenario based on fundamental limitations are presented. Accelerator physics issues are discussed and we conclude with a list of required R&D for the realization of such a design. 1 ENERGY RECOVERY LINACS Energy recovery is the process by which the energy invested in accelerating a beam is returned to the rf cavities by decelerating the beam. To date, energy recovery has been realized in a number of different ways [5], [6], [7]. Same-cell energy recovery with cw current up to 5 mA and energy up to 50 MeV has been demonstrated at Jefferson Lab’s (JLab) IR FEL and it is used routinely for the operation of the FEL as a User Facility [4]. Microbunches with an rms bunch length of ~20 psec are produced in a DC photocathode gun and accelerated to 320 kV. The bunches are compressed by a copper buncher cavity operating at 1497 MHz. They pass through a pair of superconducting rf (srf) cavities operating at an average gradient of 10 MV/m. The output beam at ~10 MeV is injected into an 8-cavity srf cryomodule where it is accelerated up to ~48 MeV. The beam then passes through the wiggler. Afterward it is recirculated through two isochronous, achromatic bends separated by a quadrupole transport line, back through the cryomodule in the decelerating rf phase and dumped at the injection energy of ~10 MeV. The benefits of energy recovery are: 1. The required rf power becomes nearly independent of beam current. 2. The overall system efficiency is increased. 3. The electron beam power to be disposed of at the beam dumps is reduced by the ratio of the final to injected energy. 4. The induced radioactivity (and therefore the shielding problem) is reduced, if the beam is dumped below the neutron production threshold. 2 CONCEPTUAL DESIGNS Next we present the reasoning that allows us to develop a self-consistent set of parameters for an electron linac– proton ring collider. Here we consider only the case of 50 GeV protons colliding with 5 GeV electrons. Conceptual designs at different energies [3], and a design based on the existing RHIC storage ring [8] have also been explored. A schematic representation of the electron-proton collider is shown in Figure 1. Figure 1. Schematic layout of the electron linac – proton ring collider. The linac technology assumed here uses TESLA-style cavities, with shunt impedance R/Q=1036 Ohms per cavity and cavity length equal to 1.038m. The residual resistance is ~3 nΩ , equivalent to a Q of ~10. Considering demonstrated performance from a number of manufacturers, we will assume Q0 of 1x10 at 2K and accelerating gradient of 20MV/m. At these values the refrigeration power is 40 W/cavity. Thus a 5 GeV linac will require 250 cavities with dissipation due to dynamic losses of 10 kW. Two point designs will be explored. In design 1 both the Laslett and beam-beam tuneshifts will remain below the rather conservative and generally agreed upon value of 0.004. To arrive at a self-consistent set of parameters and a luminosity estimate the reasoning proceeds as follows. We first set the electron beam size at the IP based on projected electron source performance. Then the proton beam parameters are set at the Laslett tuneshift limit. The maximum number of electrons per bunch is determined at the beam-beam tuneshift limit of the protons. Finally effects that influence the choice of the bunch collision frequency are discussed and a choice is made. An rms normalized emittance of 60 μm for electrons at a bunch charge of 1.75 nC is assumed, yielding a geometric emittance of 6 nm at the interaction point (IP) at 5 GeV. For a beta function of 10 cm the rms electron beam size at the IP is 25 μm. (Round beams are assumed for electrons and protons, not necessarily equal.) In order for the luminosity not to degrade too much within a collision, the beta function for the proton beam at the IP is set approximately equal to the rms proton bunch length. In this approximation, the Laslett tuneshift can be written as