A new application of advanced planar optics by researchers at the National Institute of Standards and Technology (NIST; College Park, MD and Boulder, CO) is perhaps a sign of things to come in the wider optics world. The scientists have combined a photonic integrated circuit (PIC), an optical metasurface, and a grating to construct a compact magneto-optical laser trap to, for example, cool and trap atoms for uses such as more-accurate atomic clocks, inertial-guidance sensors, magnetometers, vacuum sensors, quantum simulation, and quantum networking.1 Previous laser cooling and trapping setups were typically constructed from bulk optics on an optical table, making them large and high-maintenance; in contrast, the NIST device is only about 15 cm long and yet cools and traps atoms in a 1-cm-wide region.
Cooling atoms is equivalent to slowing them down, which makes them a lot easier to study. At room temperature, atoms move through the air at nearly the speed of sound, some 343 m/s. The rapid, randomly moving atoms have only fleeting interactions with other particles, and their motion can make it difficult to measure transitions between atomic energy levels. When atoms are slowed to about 0.1 m/s, researchers can measure the atoms’ energy transitions and other quantum properties accurately enough to use as reference standards in a myriad of navigation and other devices.
For more than two decades, scientists have cooled atoms using laser light, a feat for which NIST physicist William Phillips shared the 1997 Nobel Prize in physics. When the frequency and other properties of the laser light are chosen carefully, the photons reduce the atoms’ momentum until they are moving slowly enough to be trapped by a magnetic field. But the table-sized conventional setups are a problem in that they limit the use of these ultracold atoms outside the laboratory.
Although other miniature cooling systems have been built, the NIST device is the first one that relies solely on planar optics, which are easy to mass-produce.
PIC, metasurface, and grating
The NIST setup consists of three optical elements. First, linearly polarized laser light at 780 nm (the transition wavelength for rubidium) is launched from the PIC using an “extreme mode converter,” which enlarges the narrow laser beam, initially about 500 nm in diameter, to 280X that width. The enlarged beam then strikes the metasurface, which further widens the laser beam by another factor of 100 while altering the intensity and polarization of the light. Finally, a diffraction grating splits the single beam into three pairs of equal and oppositely directed beams. Combined with an applied magnetic field, the four beams, focusing on the atoms from opposing directions, serve to trap the cooled atoms (see figure).
The silicon nitride (SiN)-based PIC is made from a 250-nm-thick layer of SiN clad with 3 μm of silicon dioxide (SiO2) on each side; photonic structures are lithographically etched into the SiN. Light in the extreme mode converter is evanescently coupled to a slab mode and sent into free space via a grating; the circular collimated beam has a Gaussian profile and a 1/e2 radius of about 140 μm, an increase in mode area by a factor of 105.
The metasurface is made of dielectric silicon pillars 600 nm in length and 100 nm wide; it produces a flat-top diverging beam at 150 mm away that fills a 10 mm diameter, which is the clear aperture of the final grating chip.
Each component of the optical system—the converter, the metasurface, and the grating—had been developed at NIST but was in operation at separate laboratories on the two NIST campuses, in Gaithersburg, MD and Boulder, CO. NIST researcher William McGehee and his team brought the disparate components together to build the new system. Although the optical system will have to be 10X smaller to laser-cool atoms on an actual chip, the experiment “is proof of principle that it can be done,” says McGehee.
NIST optics: a harbinger?
The replacement of bulk conventional optics by metasurface optics in specialized research setups such as the NIST example could eventually be echoed to a certain extent in the wider optics world. Although lightweight, usually single-element, and able to accomplish exotic optical feats, metasurfaces do have some large disadvantages with respect to conventional optics, including low efficiencies, often-narrow bandwidths, and limited aperture sizes. It may be that metasurface optics will find their place mainly as single, crucial elements for certain types of optical systems, or it could be that metasurfaces will be made so cheaply that (at least with improvements in their optical parameters) they could find use in mass-produced products such as cell-phone cameras.
1. W. McGehee et al., New J. Phys. (Jan. 18, 2021); doi:10.1088/1367-2630/abdce3.