Exploring the quantum world, the U.S. will build the world's largest atomic interferometer

Recently, Stanford University and the U.S. Department of Energy SLAC National Accelerator Laboratory have collaborated to propose a new light field imaging system that can capture multiple views of an object at once: the design is capable of performing a single tomography scan of an object of size 1 mm3 to reconstruct 3-dimensional (3D) distributions and features that cannot be accessed individually from any single viewpoint. Imaging of cold atomic clouds in atomic interference experiments is a key application of this novel device, where enhanced light collection, high depth of field, and 3D tomographic reconstruction could provide new techniques for characterizing atomic clouds [1].

The research results were published in the Journal of Instrumentation on August 18 under the title "A novel light field imaging device for enhanced light collection of cold atomic clouds" [2].

01World's largest atomic interferometer

In the suburbs of Chicago, about 34 miles west of Lake Michigan, there is a shaft in the ground, about 330 feet deep; it was built many years ago for neutrino experiments: to explore various aspects of quantum physics, scientists put groups of atoms into a vacuum tube, followed by a laser beam.

Today, scientists will repurpose the shaft for Fermilab's MAGIS-100 project [3] - the "Matter Wave Atomic Gradiometer Interferometer Sensor" project - which combines the state-of-the-art 10-meter atomic interferometer proven technology with the world's best atomic clock technology. In addition to enabling new quantum experiments, MAGIS-100 will provide a platform for the development of a future kilometer-scale detector sensitive enough to detect gravitational waves from known sources.

When the MAGIS-100 project is complete, physicists plan to use it to detect hidden treasures: dark matter, which is thought to make up much of the universe and is mysteriously invisible; and gravitational waves, ripples in space-time caused by cosmic shocks such as black hole collisions. They hope to find traces of these previously elusive phenomena by looking at quantum features left on clouds of strontium atoms the size of raindrops.

When built, the MAGIS-100 atomic interferometer will be the world's largest. But it still lacks one key component: a proper camera.

MAGIS-100 Atomic Interferometer

02New imaging technology: Maximizing atom capture, preserving orientation information

But actually observing these atoms is trickier than we thought. To accomplish similar experiments, physicists have so far relied on cameras comparable to those found on smartphones. While such technology may be suitable for sunset or delicious food shots, it limits what physicists can see at the atomic level.

"A cell phone camera/digital SLR camera doesn't care where the light travels from: it captures the intensity of the photon and the color of the wavelength reflection, and that's about it. It's all well and good to take pictures of your family, a city skyline or the Grand Canyon with such cameras. But for studying atoms, it leaves much to be desired. Most importantly, it throws away a lot of light." said Murtaza Safdari, a graduate student in physics at Stanford University and one of the paper's authors.

Conventional imaging technique folding view. The object is placed at a distance x from the center of the lens, off the focal plane of the lens (lens center), which is at a distance F. A plane mirror is introduced between the object and the focal plane, producing a virtual object located on the focal plane, which is focused by the lens onto the sensor. This limits the geometry of the mirror, and preferentially collecting light endangers the imaging system's ability to resolve interference fringe features, creating a challenge for atomic interferometry.

Therefore, the team created a unique camera setup that relies on a "domed mirror". The extra reflections help them see what light is entering the lens and determine which angle a spot is coming from. They hope that this will allow them to look at atomic clouds in a way that has never been done before.

One way to get as much information as possible in one experiment is to set up multiple cameras, allowing them to take pictures from multiple angles and stitch them together to get a more detailed view. Although, this could work well with five cameras; however, some physics experiments require such precise measurements that even a thousand cameras may not be able to do the job. So the Stanford researchers decided to set out to make their own system to solve this problem. "The idea was basically: can we try and completely capture as much information as possible and retain the directional information?" Safdari said.

Their final prototype, made from off-the-shelf and 3D-printed components, looks like a shallow dome with a series of mirror-like dots dotted around the interior. The pattern appears to create an interesting illusion of concentric circles, but it is carefully calculated to maximize the amount of light shining into the camera.

The proof-of-concept 3D model of the imaging demonstrator. (a,b) The black parts on the left and right of the mirror are the "front" and "back" panels, respectively. (c) Enlarged view. (d) Enlarged view without the "front" panel.

A camera prototype (demonstrator) built with 3D printing technology.

03Realization of high-definition imaging, which is expected to be used in the MAGIS-100 project

In the future, the team hopes to apply the above imaging system to the strontium atomic cloud, which is the subject of the MAGIS-100 project. A brief flash from an external laser beam will be scattered away from the mirror point and pass through the cloud at numerous angles. The lens captures the resulting reflections and how they interact with the molecules, and from which "points" they bounce off. Then, based on this information, machine learning algorithms can piece together the three-dimensional structure of the atomic cloud. Currently, this reconstruction takes many seconds; in an ideal world, it would take only a few milliseconds, or even less. But, like the algorithms used to train self-driving cars to adapt to the world around them, the researchers believe the performance of their computer code will improve.

While the authors haven't had a chance to test the camera on an atom, they did try it out by scanning some appropriately sized sample parts: 3D-printed letter-shaped pieces the same size as the strontium droplets they intended to use. The pictures they took were clear enough that they could find defects in the small letters D, O and E that differed from their intended design.

Training view extracted from a 60 megapixel image of the test subject in a single shot.

Comparison of the training (real) view, with the corresponding view reconstructed after machine learning. The results between the real captured images and their generated counterparts are almost identical.

"For an atomic experiment like MAGIS-100, this device is unlike anything else on the market." The state-of-the-art is just cameras, commercial cameras and lenses," said Ariel Schwartzman, a physicist at SLAC National Accelerator Laboratory in California and co-creator of the Stanford device. They looked in the photo equipment catalog for cameras that could see atomic clouds from multiple angles at the same time."

Reference links：

[1]https://iopscience.iop.org/article/10.1088/1748-0221/17/08/P08021

[2]https://www.popsci.com/science/particle-physics-custom-camera/

[3]https://magis.fnal.gov/