Electron Bunch Length Measurements from Laser-Accelerated Electrons Using Single-Shot THz Time-Domain Interferometry

Laser-plasma wakefield-based electron accelerators are expected to deliver ultrashort electron bunches with unprecedented peak currents. However, their actual pulse duration has never been directly measured in a single-shot experiment. We present measurements of the ultrashort duration of such electron bunches by means of THz time-domain interferometry. With data obtained using a 0.5 J, 45 fs, 800 nm laser and a ZnTe-based electro-optical setup, we demonstrate the duration of laser-accelerated, quasimonoenergetic electron bunches [best fit of 32 fs (FWHM) with a 90% upper confidence level of 38 fs] to be shorter than the drive laser pulse, but similar to the plasma period.

Figure 1

Laser-accelerated electrons generate coherent transition radiation (CTR) at an aluminum tape target. The CTR is imaged by two off-axis parabolas ($f_1=125\mathrm{mm}$ and $f_2=250\mathrm{mm}$) into an electro-optical ZnTe crystal, which rotates the polarization of a chirped probe ($CP$) beam in the time-domain. This modulation of the $CP$ is analyzed in a cross correlator.

Figure 2

A typical electron spectrum as recorded with the pulse duration measurement. Because of clipping at the spectrometer aperture, the number of low-energy electrons (dashed line) reaching the spectrometer is reduced by more than an order of magnitude.

Figure 3

(a) The angular-spectral intensity of coherent transition radiation (CTR) from a short electron bunch (${\tau }_{\mathrm{short}}=32\mathrm{fs}$) followed after a time $\Delta \tau =356\mathrm{fs}$ by a long, low-energy electron background (${\tau }_{\mathrm{long}}=712\mathrm{fs}$) is constrained by several filter functions (arrows illustrate respective cutoff frequencies), affecting the overall time resolution. (b) In the time domain, the CTR field is a superposition of the low-energy electrons (red curve) trailing in the wake of an ultrashort, subresolution electron bunch (1). The resulting interference (green curve) is measured as an intensity trace in the cross correlator and includes effects of polarization optics (2).

Figure 4

(a) The model of two electron distributions is fitted to the measured cross-correlator intensity trace determining the short pulse duration with a best fit at ${\tau }_{\mathrm{short}}=32\mathrm{fs}$ (FWHM), the temporal offset $\Delta \tau =356\mathrm{fs}$, and the long electron background duration ${\tau }_{\mathrm{long}}=712\mathrm{fs}$ (FWHM). (b)–(d) depicts error margins at the 90% level of ${\tau }_{\mathrm{short}}$, ${\tau }_{\mathrm{long}}$, $\Delta \tau$, and ${\eta }_{\mathrm{el}}$. The dots represent fits to random variations of the best fit ($**$). This deviation is consistent with a non-Gaussian, low-energy tail at later times, which does not affect the short pulse duration measurement.

Figure 5

Left: Bunch duration measurements of several shots including all error contributions. Right: Corresponding measured cross-correlator traces (black) with best fits (colored). For the calculation of the red curves, a measured electron spectrum was available. The best fit electron bunch duration of shot no. 1 (highlighted) is 32 fs and thus comparable to the measured plasma period ${\lambda }_p/c=30\mathrm{fs}$ (dashed line).