Chip-scale sensor detects light’s orbital angular momentum

Light often has spin angular momentum, more commonly referred to as left or right circular polarization. Only in the past 30 years have researchers been able to impart orbital angular momentum (OAM; see the article by Miles Padgett, Johannes Courtial, and Les Allen, Physics Today, May 2004, page 35 ). Instead of the usual flat wavefront, light with nonzero OAM has a helical wavefront that turns like a corkscrew in the direction of propagation. The quantum number m describes how tight the light’s corkscrew motion is. For example, m = 1 means a full clockwise rotation in a wavelength, and m = −2 means two full counterclockwise rotations in a wavelength.

OAM provides additional degrees of freedom for encoding information in optical communications. Modern optical networks rely on wavelength and intensity modulations to send data; the addition of polarization and OAM could pack in more information without boosting traffic. What’s more, unlike polarization, which has two modes, m can take any integer value. Modes of different m are also orthogonal and wouldn’t interfere with one another during transmission.

Ordinarily, measuring OAM requires a roomful of bulky optics. But Ritesh Agarwal at the University of Pennsylvania and his colleagues have found a compact way to detect light’s OAM through the induced photocurrent in a device that’s just tens of micrometers across. ¹ Their technique could make it easier to read out information encoded in the OAM.

Gaining momentum

To measure m, researchers typically interfere light with a reference beam or send it through a hologram or specially designed aperture. The nature of the interference or diffraction pattern manifests the OAM. But those measurements not only require table-scale setups; they are also limited in the number of modes they can measure. For example, most holograms can detect only one particular value of m, so a new one must be swapped in for each mode.

In CCDs and other solid-state sensors, light is detected through its interactions with matter. A material’s response depends primarily on the local incident number of photons—that is, the intensity. Phase information is usually lost when intensity is converted to photocurrent. Because a clockwise helical wavefront and a counterclockwise one have the same donut-shaped intensity profile, a phase-sensitive measurement is essential for characterizing light with OAM without introducing holograms and apertures.

Agarwal has been contemplating such a measurement for five years. Back in 2015 he and his team found a way to detect a light beam’s circular polarization—its spin angular momentum—with a silicon nanowire. ² They were interested in extending the capability to OAM but weren’t sure how to do it. A couple years later, a breakthrough came when the team was trying to understand the properties of a new group of materials, Weyl semimetals.

Mean-Weyl …

Identified for the first time in 2015, Weyl semimetals have a broken symmetry, generally inversion symmetry, that allows them to host quasiparticles that behave as Weyl fermions—massless, charged, and chiral. (For more on Weyl semimetals, see Physics Today, December 2019, page 24 .) Agarwal and his colleagues found that the photocurrent in Weyl semimetals was sensitive not to the light intensity but to the intensity gradients. ³

Photocurrents in most materials depend only on the light’s local intensity. But when excited at a normal angle, a crystal with mirror symmetry reacts not to the local intensity but to how the local intensity compares with the surrounding intensities, or the intensity gradient. In that case, the photocurrent is nonlocal and results from an imbalance, in both real and momentum space, in the density of excited electrons. Mathematically, it is described by the terms beyond the dipole approximation in the multipole expansion of the light–matter interactions. The symmetry cancels out the dipole response, which would otherwise obscure the other terms. Although Weyl semimetals aren’t the only materials with the requisite symmetry, they have the added advantage of a nonlinear optical response that’s among the largest identified to date.

In their intensity-gradient experiment, the researchers measured a circular photocurrent that could be explained only by the nonuniform distribution of the light. But in addition to a nonuniform intensity, light with nonzero OAM has a circling phase gradient that depends on m. Could a photocurrent sense a phase gradient as well? The group’s calculations indicated that the measurement should be possible.

A new phase

Agarwal and his graduate student Zhurun Ji prepared a 50- to 200-nm-thick layer of the Weyl semimetal tungsten ditelluride and designed U-shaped electrodes, shown in figure 1, to probe the photocurrent generated by light with OAM. The electrode shape needs to distinguish the current due to the OAM phase gradient from the current due to intensity gradients and other effects. Although a few other electrode shapes work, the U shape is the best at measuring a radial current and unambiguously canceling out the other effects.

Figure 1.

Before reaching the WTe₂ detector, a beam with OAM passes through a rotating quarter-wave plate that cycles the polarization from 0° linear to left circular to 90° linear to right circular and back to 0° linear. The measured radial photocurrent, shown in figure 2a for beams with m = +1 and −1, has three components: one that depends on the linear polarization state; one, J_c, that depends on the circular polarization state; and one that is independent of polarization and includes thermal currents. To distinguish between the effect of the intensity gradient and that of the phase gradient, the researchers exploit the components’ differing dependencies by comparing beams with OAM +m and −m, which have opposite helicities but the same intensity profiles.

Figure 2.

The photocurrent is not the same for left and right circular polarization states (the left and right peaks in figure 2a). And the difference, arising from J_c, is around the same magnitude but swapped for OAM +1 and −1. The researchers found the same behavior for all OAM modes that they measured. It turns out that J_c corresponds to the photocurrent response from the phase gradient and grows proportionally with |m|, as shown in figure 2b for OAM up to ±4. It results from light transferring its OAM and energy simultaneously to the electrons in the material. Because the light’s phase isn’t uniform, there is once again a spatial imbalance of excited electrons and a net current either along or perpendicular to the gradient.

To complement the chip-scale detection of OAM, researchers have reduced the production of light with OAM to chip-scale too. Published back to back with Agarwal’s paper, the University of Pennsylvania’s Liang Feng and his colleagues report producing light with five different OAM values with the same microlaser. ⁴ In their scheme, light circling in a microscale semiconducting ring was scattered by ridges on the inner surface of the ring such that the light’s circling became a helical propagation. Control over the OAM chirality came from two waveguide arms, one pumped in clockwise light and the other, counterclockwise.