Effect of Synchrotron Radiation on Staged Plasma Wakefield Accelerators Livio Verra, Alexander Knetsch, Doug Storey Abstract In a staged, beam-driven, plasma wakefield accelerator, electrons are accelerated in a sequence of plasma stages, each powered by a driver electron bunch. Between each stage, a magnetic chicane is used to dispose of the spent driver and to inject the new fresh one, while transporting the witness bunch. We discuss the effect of synchrotron radiation in the interstage section on the accelerating gradient and on the final energy reach of such a machine. 1 Introduction Plasma wakefield acceleration is one of the promising techniques to deliver particles to a high-energy collider at the energy frontier at an affordable cost, exploiting the extremely high accelerating gradients that can be generated in plasmas Dawson 1959 ; Tajima and Dawson 1979 ; Chen et al. 1985 . In the beam-driven plasma wakefield acceleration (PWFA) configuration, a charged particle bunch (the driver ) propagating in plasma drives a plasma electron density perturbation that sustains longitudinal and transverse fields (the wakefields ). A trailing lower-charge bunch (the witness ) traveling in the appropriate phase of the wakefields (accelerating and focusing) can be accelerated with high gradient and efficiency Litos, M. et al. 2014 . In practice, the plasma acts as a transformer, transferring energy from the higher-charge driver to the lower-charge witness bunch Chen et al. 1986 . In the context of the 10 TeV Wakefield Collider Design Study Gessner, S. et al. 2025 , we investigate the feasibility of accelerating electrons ( e^{-}) to 5\, TeV with a PWFA linac for e^{-}e^{-} or \gamma\gamma collisions Barklow, T. et al. 2023a ; Barklow, T. et al. 2023b . A wakefield-based collider could be an energy upgrade of an existing linear collider facility (LCF) H. Abramowicz et al. 2025 . Considering the length of one linac arm of LCF ( \sim 10\, km), the target effective gradient to reach 5 TeV is W_{eff}=E/Z=0.5\, GV/m ( E the final beam energy, Z the length of the linac). There are two ways to reach the target energy of 5 TeV: one is to use one high-energy driver that drives wakefields and accelerates the witness in a single, very long, plasma stage; the other option is to accelerate the witness in a series of shorter plasma stages, each powered by a lower energy (hence, cheaper Foster, B. et al. 2025 ) driver bunch. In the first case, an extremely high energy driver bunch is needed, such as proton bunches from synchrotrons, to drive large-amplitude wakefields over long distances. Even though proton-driven PWFA in a single stage has been demonstrated AWAKE Collaboration 2018 , the acceleration process relies on the self-modulation instability Kumar et al. 2010 , since the driver bunch length is orders of magnitudes longer than the plasma electron wavelength. This process can be controlled with appropriate seeding methods F. Batsch, P. Muggli et al. (AWAKE Coll.) 2021 ; Verra, L. et al. (AWAKE Coll.) 2022 , but it may prevent acceleration of bunches with the extremely high quality required for high-luminosity colliders. One solution would be to compress the proton bunch before injection in plasma Caldwell et al. 2009 , but substantial R \& D is required to demonstrate the feasibility of the scheme Caldwell, A. et al. 2025 . In the case of a staged PWFA linac Lindstrøm 2021 , lower-energy, independent driver electron bunches can be used to drive the wakefields in each plasma stage, whereas the witness bunch is transported from one stage to the next, while preserving (or even improving Lindstrøm 2021 ) its transverse and longitudinal quality. In the following, we investigate the effect of synchrotron radiation in the interstage sections on the accelerating gradient and on the energy reach of such a machine. Figure 1: (a): Schematic of the simplest period of a staged PWFA linac, composed by the plasma stage, a chicane of four dipoles to separate the witness bunch (red arrow) from the spent driver bunch (blue arrow on the left) and to inject the fresh driver (blue arrow on the right) into the following plasma stage; (b): Required length of a magnetic dipole to separate by \Delta_{x} a witness bunch with E_{W}=21\, GeV from a depleted driver bunch with E_{D}=0.5\, GeV, as a function of the magnetic field amplitude B . 2 Staging The basic period of a staged PWFA consists of (see Fig. 1 (a)): the plasma stage, a dipole to separate in energy the spent driver from the witness, a dipole to inject the fresh new driver and the witness in the following plasma stage, and two dipoles in between to bring the witness back on the original propagation axis and to control the transport matrix element R_{56} of the chicane. In general, the interstage sections should be as short as possible to maximize the effective accelerating gradient of the linac. The length of the first dipole is determined by the required transverse separation \Delta x in the dispersive plane between spent driver and witness, and by the amplitude of the magnetic field B , according to Lindstrøm 2021 : \Delta x=BqcL^{2}(\frac{1}{E_{W}}-\frac{1}{E_{D}}), (1) where L is the length of the dipole, and E_{D} and E_{W} are the energies of driver and witness bunches, respectively, at the exit of the prior plasma stage. In the following, we conservatively consider E_{D}=0.5\, GeV, to take into account that some electrons in the driver bunch may have not lost all their energy. Figure 1 (b) shows that the required length of the dipole decreases when increasing B , and for shorter \Delta x . The minimum \Delta x is defined by the dumping system for the depleted driver bunch. For example, the driver could be captured by the magnetic field of a septum magnet, leaving the witness bunch unperturbed Barnes et al. 2010 . As we discuss later, a realistic period would also include optical elements to properly transport and match the witness beam envelope from one plasma stage to the next one. 3 Effect of Synchrotron Radiation As a first approximation, we consider the chicane as made of four identical dipoles, with no empty space between magnets and the plasma stages. We also assume an in-plasma accelerating gradient W_{z}^{+}=5\, GV/m and 3.6-m-long plasma stages (i.e., energy gain \Delta E^{+}=18\, GeV per stage), and a transformer ratio R=W_{z}^{+}/W_{z}^{-}=2 ( W_{z}^{-} the maximum decelerating field within the driver bunch). Since the witness bunch is forced on a curved trajectory by the dipoles in each chicane, it emits synchrotron radiation Jackson 1962 . The average energy loss of an electron through a dipole is calculated classically as: U_{0}=\frac{q^{2}}{6\pi\varepsilon_{0}}\beta^{3}\gamma^{4}L(\frac{B}{3.3pc})^{2}, (2) where p is longitudinal momentum of the witness at the plasma exit and \beta is the beam velocity normalized to c . Figure 2 (a) shows that, for the case of 2-T dipoles with appropriate length to achieve \Delta x=1\, cm, the energy of the witness bunch (green line) increases along the linac but saturates around 3 TeV. This is due to the fact that, as the energy of the witness increases, so does the average energy loss in each interstage section. Therefore the effective W_{eff}(z)=E_{W}(z)/z ( z the propagation distance along the linac) starts decreasing, and the energy saturates when the energy loss in each interstage section becomes comparable to the energy gain in the following plasma stage, preventing a further increase of the witness bunch energy. Figure 2: Energy of the witness bunch (green line, left-hand side vertical axis) and effective gradient (blue line, right-hand side axis vertical axis), along the PWFA linac for 2-T (a) and 0.2-T (b) dipoles. A solution to achieve the target energy of 5 TeV is to decrease the amplitude of the magnetic field in the dipoles (Fig. 2 (b) shows the case with B=0.2\, T), at the expenses of having W_{eff} considerably lower than W_{z}^{+} (5 GV/m in this case). The basic interstage section we have presented would naturally be complicated in a real machine by the presence of optical elements to capture and re-inject the witness from one plasma stage to the following one. If the witness beam is transversely matched to the plasma ion column focusing force, it leaves the plasma at a waist with the matched Twiss \beta function \beta_{m}=\sqrt{2\gamma}c/\omega_{pe} Muggli, P. et al. 2004 ; Verra et al. 2020 , where \gamma is the beam Lorentz factor, c the light speed, \omega_{pe}=\sqrt{n_{pe}q^{2}/m_{e}\varepsilon_{0}} the plasma electron angular frequency ( n_{pe} the plasma electron density, q the elementary charge, m_{e} the electron mass, \varepsilon_{0} the vacuum permittivity). Since n_{pe} is assumed to be the same in all plasma stages, the beam must be re-injected in plasma with the same matched parameters ( \alpha=0 , \beta=\beta_{m} ). Thus, a focusing system is required (drifts plus active elements) to reduce the divergence of the witness beam at the plasma exit, and to re-inject and match it at the entrance of the following stage with 1:1 magnification. Figure 3: Effective gradient W_{eff} of a 5 TeV linac as a function of magnetic field amplitude B , for different length L_{0} of the optics at the first interstage section. For a given focusing device, the length of optics (active elements plus drifts) increases with the energy of the beam as L_{0}\sqrt{\gamma} Lindstrøm 2021 , where L_{0} is the length of the optics at the first interstage. This means that the filling factor of the machine (length of the accelerating sections divided by the length of the transport) decreases along the linac. Figure 3 shows the effective gradient of a 5 TeV linac, as a function of B , for different values of L_{0} . With L_{0}=1\, m (blue line), it is possible to achieve W_{eff}>0.5\, GV/m with a weak magnetic field ( \sim 0.15\, T), whereas the target value cannot be reached for longer optical devices. The only available devices with such high focusing strength in both transverse planes simultaneously are plasma lenses Chen 1987 . In particular, active plasma lenses (APL) van Tilborg, J. et al. 2015 would be more appropriate than passive lenses Barov and Rosenzweig 1994 ; Barov et al. 1998 , as they do not require additional drivers to generate the focusing fields, but only a sufficiently high current pulse. Considering the length of the dipoles as available drift for the optics, it is possible to obtain L_{0}\sim 1\, m employing an APL with length 0.13 m, radius 0.5 mm, and current I=3kA (see Fig. 4 (a)). At the last interstage section (i.e., E_{W}=4.982\, TeV, Fig. 4 (b)), the length of the optical device is increased by \propto\sqrt{\gamma} as expected, for the same current pulse. Figure 4: Root-mean-square of the transverse distribution of the witness beam in the first (a) and last (b) interstage section. Dashed vertical lines indicate the active plasma lens. The emission of synchrotron radiation can also affect the beam quality due to an increase of energy spread in the dispersive sections Raubenheimer, T., Emma, P., Kheifets, S. 1993 . However, this can be mitigated by tapering the amplitude of the magnetic field along the linac C. Lindstrom et al. . Moreover, for the technical design of a real machine, it will be important to take into account the amount of power emitted as radiation, for safety and for machine protection. 4 Conclusions We discussed the effect of synchrotron radiation on the effective accelerating gradient of a staged, beam-driven, plasma wakefield accelerator. We showed that, to achieve the target energy of 5 TeV, it is necessary to reduce the amplitude of the magnetic field in the dipoles separating in energy the spent driver from the witness bunch, so that less energy is lost due to the emission of synchrotron radiation. Additionally, effective accelerating gradients above 0.5 GV/m can be achieved with the use of active plasma lenses as optical elements to transport and match the witness beam from one plasma stage to the next.