Ultralow emittance electron beams from a laser-wakefield accelerator

Using quadrupole scan measurements we show laser-wakefield accelerated electrons to have a normalized transverse emittance of ${0.21}_{-0.02}^{+0.01}\pi \mathrm{mm}\mathrm{mrad}$ at 245 MeV. We demonstrate a multishot and a single-shot method, the mean emittance values for both methods agree well. A simple model of the beam dynamics in the plasma density downramp at the accelerator exit matches the source size and divergence values inferred from the measurement. In the energy range of 245 to 300 MeV the normalized emittance remains constant.

Figure 1

A laser pulse is focused into a 5 mm long hydrogen-filled gas cell and accelerates electrons up to $\sim 350\mathrm{MeV}$. The electrons pass through a pair of magnetic quadrupole lenses $\sim 10\mathrm{cm}$ behind the gas cell which focus the electron beam. The first lens is 25 mm long and the second lens 15 mm; both have a measured field gradient of $\sim 480\mathrm{T}/\mathrm{m}$. The transverse profile of the beam is observed on a removable scintillating screen. A dipole magnet with field strength 1 T disperses the beam and allows the observation of the energy-resolved beam size of a few-MeV bandwidth on a high-resolution YAG crystal screen 2.16 m behind the gas cell.

Figure 2

Lens position scan for 245 MeV electrons. The mean rms beam sizes and their standard error are plotted against the $z$ position of lens two (black squares). The fit curve (blue) neglects space charge and gives a normalized emittance of ${0.21}_{-0.02}^{+0.01}\pi \mathrm{mm}\mathrm{mrad}$ and inferred rms source size and divergence of $0.93\mu \mathrm{m}$ and 0.45 mrad, respectively. The accuracy of the method is illustrated by the expected dependence for a 20% larger emittance by increasing the inferred source size or divergence. The influence of space charge is shown by the yellow line; a particle tracking simulation [33] of 3000 macroparticles including point-to-point space charge for an initially monoenergetic 245 MeV, 50 pC beam with duration 4.5 fs shows a clear deviation from our measurements. The initial source size and divergence in the simulation were chosen to be $0.25\mu \mathrm{m}$ and 0.45 mrad, respectively, to obtain a similar curve as obtained from the measurements. Note the steeper flank on the right-hand side of the space charge curve. A tracking simulation without space charge matches the blue fit curve which confirms the extracted emittance value from the data.

Figure 3

The rms beam size vs beam energy for a single shot (circles). The solid fit line corresponds to a beam with normalized emittance of $0.14\pm 0.01\pi \mathrm{mm}\mathrm{mrad}$. The other lines show the expected functions for a 20% larger emittance by varying the inferred source size or divergence.

Figure 4

Simulations of the electron beam envelope and final divergence after the density downramp at the exit of the accelerator. (a) Evolution of the electron beam envelope (solid blue) in the plasma to vacuum density transition (solid green, based on CFD simulations) according to Eq. (2). The fit line (red dashed) describes an effective electron source with the same emittance as the beam envelope but propagating without the focusing forces of the plasma; the beam waist and divergence are consistent with the values inferred from the lens position scan measurements. The same effective source is obtained with a downramp modeled by an error function of length 3.3 mm (black dotted). The physical exit of the gas cell is at $z=5\mathrm{mm}$. (b) The final divergence after the accelerator downramp as a function of its length, $l$ {where $n_e=n_0[0.5\mathrm{erf}(-z/l)+1]$], for three different beam energies with normalized emittances of $0.2\pi \mathrm{mm}\mathrm{mrad}$. The cross indicates the case in (a).