Ultrafast camera’s early history

For the record, scientists at Los Alamos National Laboratory developed systems that use a streak tube and image compression to record ultrafast two-dimensional movies of transient events some 30 years ago ^{1 , 2} (Physics Today, February 2015, page 12 ). A cylindrical lens focuses the light from a 2D scene into a 1D line, which is actually a tomographic projection integral containing unfiltered information from the entire image. Four such time-dependent projections of a rapidly evolving scene, one every 45 degrees, were fed into coherent fiber-optic ribbons and sent some 25 m to the photocathode of a streak tube. The fiber-optic ribbons, filtered by both wavelength and mode, achieved subnanosecond time resolution. A digital video system recorded the resulting spacetime streak-tube image, with each scan line containing a time sample of the four projections.

In the data analysis, a modified version of G. Minerbo’s maximum entropy tomography algorithm ^{3 , 4} was used to reconstruct a frame of the output movie from each scan line of the streak-tube image. A later variant of the algorithm produced a 3D spacetime tomographic reconstruction using a time-integrated 2D image of the same scene, exposed for the duration of the streak-tube sweep, as a tomographic projection along the time axis. The reconstructed movie solution is then constrained to add up to the time exposure.

More recently, workers at the Dual Axis Radiographic Hydrodynamic Test Facility at Los Alamos successfully used several streak-tomography systems as near-real-time beam diagnostics in the commissioning of the facility’s high-current pulsed electron accelerator. ⁵