All-optical density downramp injection in electron-driven plasma wakefield accelerators

The properties of electron beams obtainable from plasma wakefield accelerators are largely determined by their injection into the plasma wave. Density downramps can facilitate injection of plasma electrons and are realized routinely in laser wakefield accelerators by gas-dynamic production of longitudinal downramp profiles. We show that density downramp injection is in fact a multidimensional process when operated below the spatiotemporal scales of the plasma wave. In beam-driven plasma waves, unlocking this interaction regime by microscopic plasma torches allows the generation of unique electron beams, such as planar counteroscillating twin beamlets. Our findings generalize density downramp injection and extend its capabilities, which may find applications ranging from radiation production to high-energy-physics research.

Figure 1

Plasma torch injection based on 3D particle-in-cell simulations. The electron drive beam (blue) excites a plasma wave in the blowout regime (green). When the plasma wave traverses the plasma torch (orange), the locally increased density warps the plasma wave, which leads to injection of plasma electrons. Those subsequently form the witness beam (red). Selected trajectories of trapped electrons are displayed in the laboratory frame (black lines). See also [50] for the corresponding video.

Figure 2

Established and novel regimes of density downramp injection using plasma torches with $n_{\mathrm{T},0}=0.5\times n_{\mathrm{ch}}$. Three-dimensional PIC simulations use plasma torches with cores (a) wider and (b) narrower than the blowout. Varying the torch ramp length changes the injected charge $Q$ from the PIC simulation (black) and the model (orange) for PWFAs with plasma wavelengths (c) ${\lambda }_{\mathrm{p}}\approx 43\mu \mathrm{m}$ and (d) ${\lambda }_{\mathrm{p}}\approx 93\mu \mathrm{m}$ (d). The other parameters are (c) $r_{\mathrm{core},\mathrm{wide}}=40\mu \mathrm{m}$ and $r_{\mathrm{core},\mathrm{thin}}=2.5\mu \mathrm{m}$ and (d) $r_{\mathrm{core},\mathrm{wide}}=80\mu \mathrm{m}$ and $r_{\mathrm{core},\mathrm{thin}}=3.5\mu \mathrm{m}$. Gray shades indicate regions $l_{\mathrm{ramp}}\le {\lambda }_{\mathrm{p}}$. For narrow torches, these regions reveal a novel regime of density downramp injection unlocked by volumetric effects.

Figure 3

Replication of the FACET experiment in 3D PIC simulations. Interactions with different torch distributions $n_{\mathrm{T}}$ from (a)–(c) 0.5, (d)–(f) 1, and (g)–(i) 5 mJ torch laser energy are shown at different times (a), (d), and (g) $t_{\mathrm{sim}}=0$; (b), (e), and (h) $t_{\mathrm{sim}}=0.5$; and (c), (f), and (i) $t_{\mathrm{sim}}=12.5$ ps. The torch profiles (red) differ in composition, peak density, and volume as illustrated in (a), (d), and (g). The plasma torch modulates the blowout structure [(b), (e), and (h)] and selected trajectories show electron dynamics, which cause injection and witness beam formation for the cases in (f) and (i).

Figure 4

Impact of transverse variation of volumetric torch-wakefield overlap. (a) Injected charge as a function of transverse offset $\mathrm{\Delta }y$ for the experimental 5-mJ case obtained from the model and PIC simulation, compared to a much wider torch with $r_{\mathrm{core}}=100\mu \mathrm{m}$ and an infinitely wide slab. For improved stability, the two latter are implemented in a much wider channel, and for laser parameters that ionize only ${\mathrm{He}}^{+}$, which reduces the injected charge. (b) Simulation snapshot of the average slice electron density for $\mathrm{\Delta }y=30\mu \mathrm{m}$. The origin of electrons subsequently forming the witness beam is colored in yellow.

Figure 5

Temporal variation of the torch-wave overlap. (a) Charge obtained from experiment, simulations, and model as a function of TOA. For timing TOAs less than $-0.15$ ps, the torch laser arrives before the drive beam and creates a fully formed plasma torch, while for TOAs greater than 0.9 ps the torch forms too late to interact with the wakefield. For such timings, only torch-independent dark current is measured. In the transition region, the torch is partially formed to varying degree. (b) Corresponding electron line spectra as a function of TOA.

Figure 6

Illustration of spatial torch variations on the trapping volume: (a)–(c) radially symmetric trapping volume obtained from a wide torch using $r_{\mathrm{core}}=50\mu \mathrm{m}$ and $l_{\mathrm{ramp}}=10\mu \mathrm{m}$ and (d)–(f) trapping volume from a narrow torch with $r_{\mathrm{core}}=10\mu \mathrm{m}$ and $l_{\mathrm{ramp}}=10\mu \mathrm{m}$. Here the smaller extent in $y$ reduces the volumetric overlap with the plasma wave such that injection originates predominantly from two mirror-symmetric regions within the remaining plane in $z$.

Figure 7

Twin beamlets from reduced volumetric overlap. Shown is the witness beam from the narrow plasma torch in Fig. 6 after $\sim 1.6$ mm of acceleration in the PWFA. Red colors correspond to the beamlet injected from the upper region where $z\ge 0$ and blue corresponds to injection from $z<0$. All presented particle densities are obtained from projections to the respective planes. (a) and (b) Visualization of the witness beam's planar spatial distribution, with arrows indicating the population's oscillation direction. (c) Trace space of the witness beam in the torch plane, displaying two counteroscillating populations. All panels display clear separation of the beamlets along the torch axis. The longitudinal phase space in (d) exhibits a flattened accelerating field profile from beam loading, which reduces the energy chirp. (e) and (f) Evolution of the transverse normalized emittance in both planes along the PWFA, displaying a strong asymmetry between $y$ and $z$.

Figure 8

Generation of planar beams from controlled volumetric overlap. (a) Aspect ratio of witness beam parameters ${\sigma }_z/{\sigma }_y$, ${\sigma }_{pz}/{\sigma }_{py}$, and ${\epsilon }_z/{\epsilon }_y$ integrated over $2.5\text{mm}$ of propagation. Beams are injected on torches with varying core diameters at constant short ramp $l_{\mathrm{ramp}}=10\mu \mathrm{m}$. For small core diameters corresponding to reduced spatial overlap, the beams become wider in $z$ and narrow in $y$ and lead to increasingly planar distributions. (b) Projections to the transverse plane integrated over all PIC snapshots for beams from a narrow and a wide torch injector as well as the limit of an infinitely wide slab.