Nanopatterned silicon shields photonic devices which will increase density of circuits in photonic chips

University of Utah electrical and computer engineering associate professor Rajesh Menon and his team have developed a cloaking device for microscopic photonic integrated devices — the building blocks of photonic computer chips that run on light instead of electrical current — in an effort to make future chips smaller, faster and consume much less power.

The future of computers, data centers and mobile devices will involve photonic chips in which data is shuttled around and processed as light photons instead of electrons. The advantages of photonic chips over today’s silicon-based chips are they will be much faster and consume less power and therefore give off less heat. And inside each chip are potentially billions of photonic devices, each with a specific function in much the same way that billions of transistors have different functions inside today’s silicon chips. For example, one group of devices would perform calculations, another would perform certain processing, and so on.

The problem, however, is if two of these photonic devices are too close to each other, they will not work because the light leakage between them will cause “crosstalk” much like radio interference. If they are spaced far apart to solve this problem, you end up with a chip that is much too large.

So Menon and his team discovered you can put a special nanopatterned silicon-based barrier in between two of the photonic devices, which acts like a “cloak” and tricks one device from not seeing the other.

Nanophotonic cloaks for closely spaced waveguides.

Nature Communications – Increasing the density of passive photonic-integrated circuits via nanophotonic cloaking

“The principle we are using is similar to that of the Harry Potter invisibility cloak,” Menon says. “Any light that comes to one device is redirected back as if to mimic the situation of not having a neighboring device. It’s like a barrier — it pushes the light back into the original device. It is being fooled into thinking there is nothing on the other side.”

Consequently, billions of these photonic devices can be packed into a single chip, and a chip can contain more of these devices for even more functionality. And since these photonic chips use light photons instead of electrons to transfer data, which builds up heat, these chips potentially could consume 10 to 100 times less power, which would be a boon for places like data centers that use tremendous amounts of electricity.

Menon believes the most immediate application for this technology and for photonic chips in general will be for data centers similar to the ones used by services like Google and Facebook. According to a study from the U.S. Department of Energy’s Lawrence Berkeley National Laboratory, data centers just in the U.S. consumed 70 billion kilowatt hours in 2014, or about 1.8 percent of total U.S. electricity consumption. And that power usage is expected to rise another 4 percent by 2020.

“By going from electronics to photonics we can make computers much more efficient and ultimately make a big impact on carbon emissions and energy usage for all kinds of things,” Menon says. “It’s a big impact and a lot of people are trying to solve it.”

Currently, photonic devices are used mostly in high-end military equipment, and he expects full photonic-based chips will be employed in data centers within a few years.

Planar lightwave circuits (PLC) have significant advantages over electronic circuits such as large bandwidth absence of the Joule effect and higher immunity to interference, among many others. However, the main disadvantage of PLC is their considerably lower density compared with integrated electronics. There are several options to increase the integration density of PLC. One can shrink the footprint of the component devices. Various methods have been proposed to decrease device dimensions including the application of plasmonics or of nanophotonics. We have previously demonstrated an integrated nanophotonic polarization beamsplitter with a footprint 2.4 × 2.4 μm2, which is at least an order of magnitude smaller than comparable integrated devices that have been demonstrated experimentally before. A second option to increase integration density is to combine the function of multiple devices into a single compact device. Examples of such multi-functional devices include polarization-splitting grating couplers, mode-converting polarization splitters and a transformation-optics-based beam shifter. A third option for enhancing integration density is to decrease the spacing between the individual devices. Waveguiding of light in the plane of the PLC is one of the most fundamental functions. However, the integration density of waveguiding is limited by the leakage of light from one waveguide to its neighbour (cross-talk), if the spacing between them is too small. Song et al. proposed a method to decrease this spacing without considerably increasing cross-talk. However, a general method to decrease the spacing between various devices has not been demonstrated. On the other hand, cloaking to prevent detection has been proposed using numerous technologies. Zografopoulo and Prokopidis proposed a method for integrated cloaking based on plasmonics, which exhibits considerable parasitic absorption losses due to metal. Integrated all-dielectric cloaks employing conformal mapping were experimentally demonstrated before. However, these cloaks typically exhibit footprints of hundreds of micrometers.

Here they apply cloaking to shield the closely spaced devices so as to enable them to be integrated at a much higher density that is otherwise feasible. An inverse-design algorithm is employed to design the integrated cloak with a footprint of just a few micrometers. They fabricated and characterized cloaked waveguide pairs that exemplify our general approach. Furthermore, their approach is generally applicable to various integrated photonic components.

Abstract

Photonic-integrated devices need to be adequately spaced apart to prevent signal cross-talk. This fundamentally limits their packing density. Here we report the use of nanophotonic cloaking to render neighbouring devices invisible to one another, which allows them to be placed closer together than is otherwise feasible. Specifically, we experimentally demonstrated waveguides that are spaced by a distance of ∼λ 0 / 2 and designed waveguides with centre-to-centre spacing as small as 600 nm (less than λ0 / 2.5). Our experiments show a transmission efficiency over −2 dB and an extinction ratio over 15 dB over a bandwidth larger than 60 nm. This performance can be improved with better design algorithms and industry-standard lithography. The nanophotonic cloak relies on multiple guided-mode resonances, which render such devices very robust to fabrication errors. Our devices are broadly complimentary-metal-oxide-semiconductor compatible, have a minimum pitch of 200 nm and can be fabricated with a single lithography step. The nanophotonic cloaks can be generally applied to all passive integrated photonics.

Earlier related work

Nature Scientific Reports – Chromatic-aberration-corrected diffractive lenses for ultra-broadband focusing

SOURCES – University of Utah, Eurekalert, Nature Communications