Meter-Scale Plasma-Wakefield Accelerator Driven by a Matched Electron Beam

A high-gradient, meter-scale plasma-wakefield accelerator module operating in the electron blowout regime is demonstrated experimentally. The beam and plasma parameters are chosen such that the matched beam channels through the plasma over more than 12 beam beta functions without spreading or oscillating over a range of densities optimum for observing both deceleration and acceleration. The wakefield decelerates the bulk of the initially 28.5 GeV beam by up to 155 MeV; however, particles in the back of the same beam are accelerated by up to 280 MeV at a density of $1.9\times {10}^{14}{\mathrm{c}\mathrm{m}}^{-3}$ as the wakefield changes sign.

Figure 1

Schematic of the plasma-wakefield acceleration experimental setup (not to scale).

Figure 2

Transverse size ${\sigma }_x$ of the beam (red points) in the $x$ plane measured on the downstream OTR foil (see illustration inset) as a function of plasma density. The green line is the best fit to the data using a beam envelope model in which ${\sigma }_{x0}=30\mu \mathrm{m}$, ${\epsilon }_x=9\times {10}^{-9}\mathrm{m}\mathrm{r}\mathrm{a}\mathrm{d}$, and ${\beta }_0=0.11\mathrm{m}$.

Figure 3

Streak camera images of a single electron bunch dispersed in time and energy after propagation through (a) low density ($n_e=7\times {10}^{13}{\mathrm{c}\mathrm{m}}^{-3}$) and (b) high density ($n_e=1.9\times {10}^{14}{\mathrm{c}\mathrm{m}}^{-3}$) plasma (see text for details). The blue tick marks show the mean energy of picosecond wide slices of the beam. (c) Relative energy of the mean position of a picosecond wide slice of the beam vs time calculated from the streak camera images for three different plasma densities: $n_e=7\times {10}^{13}{\mathrm{c}\mathrm{m}}^{-3}$ (red diamonds), $n_e=1.5\times {10}^{14}{\mathrm{c}\mathrm{m}}^{-3}$ (green squares), and $n_e=1.9\times {10}^{14}{\mathrm{c}\mathrm{m}}^{-3}$ (blue circles). The curves are an average of $\approx 10$ events such as those shown in (a) and (b). The error bar (rms of the data) shows the vertical jitter in the data.

Figure 4

(a) Relative energy variations experienced by the mean of 1 ps wide slices of the beam fitted with a Gaussian distribution at $n_e=1.5\times {10}^{14}{\mathrm{c}\mathrm{m}}^{-3}$ (green squares) and $n_e=1.9\times {10}^{14}{\mathrm{c}\mathrm{m}}^{-3}$ (blue circles). These curves are obtained by averaging $\approx 10$ individual streaks after lining up the first slice ($-5\mathrm{p}\mathrm{s}$) to remove any vertical pointing jitter of the beam. The uncertainty on the slice timing (not shown) is $<0.5\mathrm{p}\mathrm{s}$. (b) Relative energy distribution of the picosecond slice at $+5.5\mathrm{p}\mathrm{s}$ for $n_e=7\times {10}^{13}{\mathrm{c}\mathrm{m}}^{-3}$ (blue circles) and $n_e=1.9\times {10}^{14}{\mathrm{c}\mathrm{m}}^{-3}$ (red squares, shifted vertically). The horizontal arrows are the full widths at half maximum of the distributions.