WEOBG02
    
    Proceedings of EPAC08, Genoa, Italy
    
    EXPERIMENTAL RESULTS OF A PLASMA WAKEFIELD ACCELERATOR USING MULTIPLE ELECTRON BUNCHES
    E. Kallos, T. Katsouleas, P. Muggli, USC, Los Angeles, CA, USA K. Kusche, J. Park, I. Pogorelsky, D. Stolyarov, V. Yakimenko, BNL, Upton, NY, USA W. D. Kimura , STI Optronics, Inc., Bellevue, WA, USA Abstract
    We present some preliminary experimental results of a plasma wakefield accelerator technique which utilizes multiple electron bunches in order to drive a plasma wave. The experiments were performed at the Accelerator Test Facility of Brookhaven National Laboratory where 5 8 equidistant bunches with a spacing that was varied between 100 250 m were fed into a 6 mm long capillary discharge plasma. By varying the time delay of the bunches with respect to the discharge different plasma densities could be tuned, and the effects of the plasma on the bunches were recorded. Such multiple bunch schemes are of great interest because they can provide increased efficiencies and high transformer ratios for advanced accelerators.
    
    PLASMA & BEAM DIAGNOSTICS Plasma Diagnostics
    The plasma source consists of a 6 mm long polypropylene tube with 1 mm inner diameter, which is filled with hydrogen at a pressure of 150 Torr. A 25 kV electrical discharge breaks down the gas creating a plasma with a peak density of 1.5×1018 cm-3 that occurs approximately 200 ns after the peak of the discharge current. The plasma density is measured through the Stark broadening of the Balmerthat is collected by inserting an optical fiber perpendicularly into the capillary body [5]. The plasma light is directed into an optical spectrometer that is adjusted such that a 1 nm spectral window with tunable central wavelength can be selected. The plasma spectrum (and density) as a function of time is retrieved in successive discharge shots. A specific plasma density can be then selected by delaying the arrival of the electron beam after the initiation of the discharge.
    
    INTRODUCTION
    In the plasma wakefield accelerator (PWFA), a relativistic electron beam is fed into a plasma and excites electron oscillations that can support electric fields (wakefields) that can be orders of magnitude higher than those utilized in present conventional accelerator structures [1], thus providing a promising technology for the design of a future particle collider. Specifically of interest here are schemes where multiple electron bunches are used to drive the wakefield in the plasma [2], due to the possibility of multiplying the energy of an incoming trailing beam in a single PWFA stage. This can occur at a given plasma density if the number of particles in each bunch along with the bunch positions is tuned appropriately, as was theoretically explored previously in reference [3]. In this paper preliminary experimental results of the interaction between a 59 MeV multi-bunched electron beam fed into a 6 mm long capillary discharge plasma are presented, which were recorded during experiments that were performed in the Accelerator Test Facility (ATF) of the Brookhaven National Laboratory (BNL). We have observed increasing energy loss along the length of the beam as indicated by the energy shifts of the individual bunches in an magnetic energy spectrometer, demonstrating a maximum excited wakefield around 22 MV/m near the tail of the bunched beam. The results agree well with simple simulations of the beam plasma interaction at long plasma wavelengths (low plasma densities) [4].
    
    Electron Beam Diagnostics
    The microbunches at ATF are generated by selectively blocking portions of the full 300 pC charge, 5.5 ps long (FWHM) electron beam. The beam is imparted a correlated energy spread (typically around 1.5%) and a metal mask consisting of equidistant solid wires is placed in its path in a high dispersion region of the beamline where the transverse size of the beam is dominated by the energy spread, a process which is described in detailed elsewhere [P. Muggli et al., in these proceedings and to appear in Phys. Rev. Lett.]. In order to experimentally verify the generation of the microbunches, Coherent Transition radiation (CTR) interferometry was employed [6]. The CTR signal that is emitted when the bunches pass though a copper mirror is collected and sent through an interferometer that autocorrelates the signal in time. The auto-correlation timeintegrated trace is recorded for different path lengths inside the interferometer by a liquid helium cooled silicon bolometer detector. A train of N equidistant bunches (typically N is between 5 and 8) yields a symmetric trace with 2N+1 peaks as a function of the path length, while the distance between those peaks is equal to the microbunch period, typically measured around 100 250 spread. This separation corresponds to resonant plasma densities in the 1016 1017 cm-3 range.
    
    03 Linear Colliders, Lepton Accelerators and New Acceleration Techniques
    
    A13 New Acceleration Techniques
    
    1912
    
    Proceedings of EPAC08, Genoa, Italy
    Wake Amplitude [MeV/m]
    
    WEOBG02
    
    150
    
    5 Bunches 150pC No mask 300pC
    
    100
    
    50
    
    0 14 10
    
    10 10 10 Plasma density [cm-3]
    
    15
    
    16
    
    17
    
    10
    
    18
    
    Figure 1: Plasma wakefield amplitude response as a function of the plasma density for a bunched and nonbunched longitudinally square beam. The bunches are separated by one resonant plasma wavelength (250 m) and their width is half a plasma wavelength.
    
    spectrum of 7 bunches after the plasma of the same density is shown in Figure 2, and it is compared to the beam spectrum without a plasma discharge. The first drive bunch has the highest energy, around 59.4 MeV. The different number of particles per bunch reflects the original non-flat charge distribution of the ATF beam. The total charge now is reduced to roughly 150 200 pC, which results in a measured average energy loss of 22 MV/m for the final microbunch. This is still envelope interaction at the low density, as demonstrated by the increasingly larger energy loss of the later drive bunches. The simulated excited wakefield and energy spectra for the bunches is shown in Figure 3 and compared with the experimentally recorded energy shifts, demonstrating good agreement with the data. The bunches are shown to sample the decelerating first half-period of the plasma wave, which explains why the later bunches appear to lose more energy compared to the earlier bunches.
    
    Operation Near the Resonance SIMULATIONS
    The predicted plasma wakefield response as a function of the plasma density is shown in Figure 1. In the case without the mask inserted, the full electron beam is fed into the plasma, and the maximum wakefield amplitude peaks at 40 MV/m assuming a transverse spot size of 100 ope response appears in the 1014 1015 cm-3 density range, while the beating visible at higher densities appears because the model of the beam is a square bunch with sharp cutoffs at the edges. In the case where the wire mask is blocking sections of the beam to create the bunches, the total charge remaining is 150 pC and is assumed to be distributed over N=5 identical bunches which have width equal to 125 (equal to half the resonant plasma wavelength). It is observed that the wakefields are very low for most densities, expect for the resonant density at 1.8×1016 cm-3 (where the wakes of all 5 bunches add in-phase) where the wake is expected to peak at 140 MV/m, even though the total charge is less in this case. Weaker wakefield peaks also appear at the plasma frequency harmonics, as each bunch is located every integer number of periods of the plasma wave. It is important to notice that the resonance width is roughly 20% (FWHM) around the resonant density, approximately inversely proportional to the number of bunches, or ~1/N. This is the accuracy with which the plasma density must be known in order to detect the enhanced wakefield amplitudes. Operation near the resonant density is more challenging because the plasm for the bunches shown here over the desired plasma length in order to observe the enhanced wakefields. Variation in the plasma density larger than this value, nonuniform longitudinally plasmas or slightly non-equidistant microbunches will result in decreased values of the wakefield. In addition, for capillary discharges the density may vary by 3 4 orders of magnitude over a few microseconds, and the beam needs to be timed to arrive
    
    Plasma off
    
    Plasma on
    1.5 Shots at 2.368 s or 7.2e+013 cm-3
    Plasma OFF Plasma ON
    
    Intensity [a.u.]
    
    1
    
    0.5
    
    BEAM
    
    PLASMA INTERACTION
    
    The effect of the plasma on the full 300 pC beam was first recorded at a low density of 7×1013 cm-3 (not shown), indicating an energy loss gradient of 35 MV/m compared to the plasma off case. This was deduced from the energy shift of the beam centroid over the given plasma length. When the mask is inserted the microbunches are fed into the plasma. The experimentally recorded energy
    
    0 58.4
    
    58.6
    
    58.8 59 59.2 59.4 Beam Energy [MeV]
    
    59.6
    
    Figure 2: Recorded energy spectra of the bunched electron beam without a plasma discharge and after a 6 mm long plasma at a density of 7×1013 cm-3. The beam crosses the capillary 2.4 s after the peak of the plasma discharge current.
    A13 New Acceleration Techniques
    
    03 Linear Colliders, Lepton Accelerators and New Acceleration Techniques
    
    1913
    
    WEOBG02
    
    Proceedings of EPAC08, Genoa, Italy
    
    Wakefield [MeV/m]
    
    Wakefield and beam density, 7e+013cm-3 20 0 -20 0 1 Time [ps] Energy Spectrum after 6mm of plasma 2 3 4 5
    
    Head
    
    eBeam Wakefield Data
    
    Beam density [a.u.]
    
    1000
    
    ON
    500
    
    OFF Head
    
    0 58.5
    
    59 eBeam Energy [MeV]
    
    59.5
    
    Figure 3: Simulated wakefield and energy spectra for the experimental data of the beam plasma interaction of Figure 2. The top panel shows the 7 bunches with 200 pC total charge fed into a 7×10 13 cm-3 density 6 mm long plasma. The wakefield under each bunch as deduced from the energy shifts is shown with red dots. The bottom panel shows the simulated energy spectra before and after the plasma. Small Gaussian energy spreads were added to the bunches in the simulation. within this narrow window around the resonance. If the plasma density after some time varies as nmax×exp(- t), where is the decay time constant, and the enhanced wakefield is to be observed in a narrow range of densities 0, then the time 1/ 0). For example, if there is wide resonance and the density drops -1 70 ns. exponentially with a coefficient This is the lowest time resolution with which the density must be scanned in order for the resonance to be detected. Finally, instabilities in the incoming beam energy distribution may also affect the location of the resonance as the bunch period may drift during a run.
    0=20%
    
    a 7×1013 cm-3 density. In the near future these experiments will be extended to the resonant density where a large enhancement of the wakefield amplitude is expected. The authors wish to acknowledge the ATF staff for their valuable support in these experiments. The work was supported by the US Department of Energy.
    
    REFERENCES
    [1] Blumenfeld, I. et al., Nature 445, no.7129, 741 (2007). [2] Maeda, R. et al., Physical Review Special TopicsAccelerators and Beams 7, no.11 (2004). [3] Kallos, E. et al., in Proceedings of the Particle Accelerator Conference (IEEE, 2007), pp. 3070. [4] Kallos, E. et al., in Proceedings of the Particle Accelerator Conference (IEEE, 2005), pp. 3384. [5] Edison, N. et al., Physical Review E 47, no.2, 1305 (1993). [6] Muggli, P. et al., in Proceedings of the Particle Accelerator Conference (IEEE, 2007), pp. 3079.
    
    SUMMARY
    We have presented simulations that demonstrate the enhancement on the plasma wakefield amplitude when a microbunched electron beam is fed into a plasma. We have reported early experimental results of a 150 pC beam indicating that the beam samples a 22 MV/m wakefield at
    
    03 Linear Colliders, Lepton Accelerators and New Acceleration Techniques
    
    A13 New Acceleration Techniques
    
    1914

Add Comment

To add a comment, you need to login or signup (you can signup without leaving this page)