Ultrafast imaging at the INRS experimental facility. (Image: Xianglei Liu and Jinyang Liang, INRS)

Recording crystal-clear images of fast movements like falling water droplets or molecular interactions requires ultra-fast cameras that are capable of capturing millions of images per second. The problem? Their use is extremely expensive and their applications are limited.

New research led by Professor Jinyang Liang of the Institut National de la Recherche Scientifique (INRS) unveils a camera that could offer a much cheaper, ultra-fast imaging technique for a wide range of applications such as real-time monitoring, augmented reality, taking medications or even the development of light wave detection and ranging systems (LIDAR) which can be used to prevent collisions with objects or people.

In Optica Professor Liang and his collaborators at Concordia University and Meta Platforms show that their new ultra-fast, real-time diffraction-triggered photography camera (DRUM in English) can record a dynamic event with a single exposure at 4.8 million frames per second. They demonstrate the capabilities of this camera by recording the rapid dynamics of laser pulses on the order of femtoseconds that interact with a liquid, as well as laser ablation in biological samples.

“In the long term, I believe DRUM photography will contribute to advancements in biomedicine and LIDAR-like automation technologies, where a faster recording technique would enable more accurate hazard detection,” says Jinyang Liang. The model of DRUM photography is, however, relatively generic. It can in theory be applied to any CCD or CMOS sensor camera without degrading its other advantages such as high sensitivity.

Record Fast Dynamics

The team created a DRUM camera with a sequence depth of seven frames, meaning it records seven frames in each short movie. After characterizing the spatial and temporal resolutions of the system, the researchers implemented these parameters to record the interactions between a laser and distilled water.

The images obtained in time lapse showed the evolution of a plasma channel and the formation of a bubble in response to the action of a pulsed laser, the measured bubble radii corresponding to those predicted by the theory of cavitation. They also imaged the dynamics of bubbles in a carbonated beverage and recorded transient interactions between an ultrashort laser pulse and a single-layer onion cell sample.

Despite the considerable progress made in the field of ultrafast imaging, current methods remain expensive and complex to implement. Their performance is also limited by the trade-offs users must make between the number of frames recorded in each movie and light output or temporal resolution. To resolve these issues, researchers have developed a new time-division method known as time-varying optical diffraction.

Cameras use these “time gates” to control when light hits the sensor. In a conventional camera, for example, the shutter acts as a kind of door that opens and closes only once. In the time-slicing principle, the door opens and closes rapidly a number of times before the image arrives at the sensor, creating a short, very fast movie of the scene.

Taking into account the space-time duality of light, Jinyang Liang understood how to obtain temporal division using light diffraction. The researcher realized that by quickly changing the tilt angle of periodic facets in a diffraction grating, which can generate multiple replicas of incident light traveling in different directions. In this way, one could scan various spatial positions to synchronize images at different points in time. These could then be assembled to create a high-speed film. To transform this idea into a functional camera, a multidisciplinary team bringing together different expertise in areas such as physical optics.

“It is possible to realize this digital micromirror device (DMD), an optical component commonly used in projectors, from its non-conventional use to produce a scanning diffraction gate of the desired type,” explains Professor Liang. Digital micromirror devices are mass-produced, and no mechanical movement is required to generate the diffraction gate, making the system cost-effective and stable.

The research team is continuing its work to improve the device's performance, particularly in terms of recording speed and sequence depth. They also wish to evaluate the possibility of recording information in color (the images are currently only in black and white) and of applying the system to other areas such as LIDAR techniques.