High-Resolution $\gamma$-Ray Radiography Produced by a Laser-Plasma Driven Electron Source

An electron beam from a laser-plasma accelerator is converted into a $\gamma$-ray source using bremsstrahlung radiation in a dense material. The $\gamma$-ray beam has a pointlike source size because it is generated by a high quality electron beam with a small source size and a low divergence. Using this $\gamma$-ray source, the radiography of complex and dense objects with submillimeter resolution is performed. It is the first evidence of a $\gamma$-ray source size of a few hundreds micrometers produced with laser-driven accelerators. This size is consistent with results from Monte Carlo simulations.

Figure 1

Electron spectrum, fitted by a Maxwellian distribution with temperature $T=40\mathrm{M}\mathrm{e}\mathrm{V}$, and electron dose profile in the Lanex film in the inset. The dashed curve is the detection threshold.

Figure 2

Experimental setup. The electrons accelerated during the interaction of the laser pulse with the gas jet slow down in the tantalum converter, which leads to $\gamma$-ray generation. The electrons are removed from the laser axis by strong magnets and the aluminum diffuser. The intensity transmitted through the object is imaged on the $\gamma$ camera. A cut of the object is also shown.

Figure 3

Photon spectrum from Monte Carlo simulation averaged between 0° and 1° (with parameters $T=40\mathrm{M}\mathrm{e}\mathrm{V}$ and ${\Theta }_0=1°$ for the electron beam converted in $\gamma$ rays in a 2.5 mm-thick tantalum target followed by a 7.5 mm-thick aluminum foil).

Figure 4

Angular distribution of photons computed by Monte Carlo simulation (with parameters of Fig. 3).

Figure 5

Spot size measurement. This object is a 20 mm-thick steel plate, with 1.0, 0.75, and 0.4 mm-square holes from left to right. The curve represents the derivative of the profile averaged over the rectangular area in the center.

Figure 6

Radiography images of a 2 mm-diameter tungsten object with cylindrical symmetry corresponding to (a) the experimental image and (b) the computed image. Color tables are linear. A simple post processing has been performed, consisting of a division by a reference image without the object.