Thinnest of Materials Loom Large

HANK HOGAN, CONTRIBUTING EDITOR, hank.hogan@photonics.com

Only a few atomic layers thick, 2D materials such as graphene, the compound semiconductor germanium selenium, the transition metal dichalcogenide molybdenum ditelluride, and others could provide new photonic capabilities — better multispectral imaging and single-photon sources, for example.

As a result, areas ranging from imaging to quantum information processing could benefit.

Graphene is the focus of much research and development, but it’s only one out of as many as a thousand stable 2D materials. That number is far greater when considering possible combinations.

These thinnest of materials could transform the future — if problems of production and integration can be solved.

Yaqing Bie, a postdoctoral associate at the Massachusetts Institute of Technology (MIT), is part of the Jarillo-Herrero research group, which is investigating optically active 2D materials. The researchers have constructed monolayers that work as sources or detectors, a demonstration in principle of on-chip optical interconnects, according to Bie.

Microprocessors constructed using 2D semiconductors show that transistors possibly suitable for backplane display drivers could be made from 2D semiconductors. This could enable display sizes to shrink or resolutions to increase, and allow flexible or curved displays. Courtesy of TU Wien.

“Different 2D materials can have different functionality. There are 2D insulators, semiconductors, semimetals, metals as well as topological materials,” she said. “For example, 2D semiconductors can work as high-yield optical emitters and high-efficiency detectors. 2D superconductors can even become ultrathin single-photon detectors.”

Power in stacking

According to Bie, layers of different materials can be stacked atop one another, either aligned or at a controlled angle. Such an artificial structure has engineered crystal properties, which means that electrical and optical performance of the combined stack can be tuned. That further expands the possible applications.

Ongoing research at MIT is looking into how to create layered semiconductor materials and heterostructures that emit and detect near the standard 1350- or 1500-nm optical telecommunications bands, according to Bie.

The researchers are also looking into how to improve light source and coupling efficiency and how to increase photodetection speeds. Other areas that need work in general involve improving material quality, integration with existing optoelectronic systems, and developing new functional materials.

2D light emitters and detectors can be integrated into silicon CMOS chips. Shown here is a molybdenum ditelluride 2D light source for silicon photonics. Courtesy of Ella Maru Studio.

One of the issues that must be solved for 2D materials to move into mass market applications is the development of standards, such as defining just what is meant by “graphene,” according to Andrew Deakin, chief technical engineer at Versarien PLC of Cheltenham, England. Versarien and its subsidiaries are engaged in graphene development and manufacturing.

“There are a lot of companies that claim to make graphene, but typically the product is graphite or micro graphite,” Deakin said. “ISO [International Organization for Standardization] is in the process of defining graphene with ISO 80004-13:2017, saying that ‘few-layer graphene, FLG, is a 2D material consisting of three to 10 well-defined stacked graphene layers.’ This number of layers is used [because] most of the properties are graphite ... when the particles are greater than 10 layers.”

Graphite flakes. With processing, these flakes can yield graphene, a 2D material made up of a few atomic layers possessing useful optical and electronic properties. Courtesy of Versarien.

When making 2D materials, the thickness must be only a few atomic layers, and the lateral size should be substantial. This is difficult to do in volume production while still maintaining quality, according to Deakin.

He added that the proprietary Versarien process can achieve this, and it has been scaled up from about a gram per day capability three years ago to a kilogram per machine per shift today. With some postprocessing, the yield of few-layer graphene content is better than 90 percent. The same process can be used for other 2D materials, such as hexagonal boron nitride.

Graphene’s optical and electronic properties make it ideal for optical sensing. This fully packaged optical sensor can detect light in the SWIR to visible ranges and was developed by Emberion of Finland for industrial spectroscopy and optical sensing. Courtesy of Emberion.

Versarien is working with commercial partners to incorporate the raw graphene into products because graphene can increase strength or improve thermal, electrical, or photonic properties.

“There are also a large number of layered materials out there that can be processed in the same way to make new 2D materials with different required properties,” Deakin said.

IR transceivers

In the near term, graphene could have an impact on an important photonics application: IR transceivers for data center applications. These devices move data from one location to another by sending and detecting light pulses that travel down an optical fiber. Producing such graphene-based transceivers is a goal of the Graphene Flagship project, the largest research initiative ever undertaken by the European Union in the field of new materials.

The use of graphene promises to cut the power consumption of such transceivers by at least an order of magnitude. This is significant, as the share of overall power consumed by data centers has been growing. By 2020, the goal is to create and begin to deploy a 5G transceiver that works at 336 Gbps, according to Andrea Ferrari, director of the Cambridge Graphene Centre, which is part of the Graphene Flagship. This is faster than today’s state of the art, he said. The increased speed is needed to support 5G networks, which will offer 20× the 1-Gbps maximum download possible with current 4G technology.

Another example of what may be possible with 2D materials in photonics comes from research done by Thomas Mueller, an associate professor who heads the Nanoscale Electronics and Optoelectronics Group at the Vienna Institute of Technology (TU Wien). The group investigates 2D materials such as graphene and layered transition metal dichalcogenides. The researchers recently built a microprocessor out of a 2D semiconductor.

Integration with silicon photonics is essential for graphene applications in telecommunications settings. Germany’s AMO has developed optical photodetectors using graphene with a bandwidth of more than 76 Ghz on silicon wafers using existing wafer processing technologies. Courtesy of AMO.

One application could be in the transistor backplanes in displays. These devices drive the display pixels, switching them on or off. Currently, these transistors, which must be transparent, are made of amorphous silicon, indium gallium zinc oxide, or low-temperature polycrystalline silicon, according to Mueller.

“Two-dimensional semiconductors exhibit higher carrier mobilities than these materials,” he said, “which would allow for shrinking down the transistor size.”

Smaller transistors would reduce power consumption or allow greater resolution displays. Another potential advantage is that 2D semiconductors are very flexible. According to Mueller, these materials could be suited for displays that bend or are curved.

Single-photon emitters

Besides modulating light, 2D materials can also generate it in unique and useful ways. This is a research focus of Rudolf Bratschitsch, a physics professor at the University of Münster. He is investigating light sources that produce one photon at a time. These single-photon emitters would be particularly useful in quantum communications and information processing. Today, single-photon sources are approximated by attenuating a light beam so that only one photon, on average, is found in each time slot.

Single-photon emitters can be positioned in a 2D material because of the strain created by underlying structures. Courtesy of Rudolf Bratschitsch, University of Münster.

In contrast to other single-photon emitters, those found within 2D materials can be precisely positioned with respect to other optical elements, such as waveguides or photonic crystals. This is because a 2D material that is placed on a surface conforms to it, much like a tacked-down carpet. Putting down a sheet of 2D material over a series of microscopic hills and valleys deposited and etched in a surface will create bends and strain. This strain potential, in turn, enables researchers to place a single-photon source where desired.

Bratschitsch and his team have used this technique to couple single-photon emitters to a waveguide on a photonic chip, with the final goal being placing single-photon sources, processing units, and detectors all on one chip.

Although there currently are no technical roadblocks, there are practical barriers to achieving this goal. For one thing, making something that outperforms competing quantum computer technologies in the handling of quantum bits (qubits) will require optimizing every component of source, processing, and detection at once, according to Bratschitsch.

“At the moment, optimization is done mostly separately for the three components, and this is already a challenge,” he said. “Just think about the task of fabricating thousands of qubits for quantum computing on one device.”

Bratschitsch added that also putting all components on the same chip would present additional technical restrictions on component specifications. However, the development of clever schemes could speed up the process of overcoming these challenges.

Finally, the wide spectral response and high sensitivity of graphene or other 2D materials could be an advantage for sensors in such applications as a hyperspectral camera, said Cedric Huyghebaert, R&D manager of the nanoapplications material engineering group at IMEC. A research and development hub based in Leuven, Belgium, IMEC is working to integrate graphene and other 2D materials into a standard CMOS processing flow that will enable the materials to be incorporated into semiconductor devices.

Two-dimensional materials are being incorporated into devices to boost electronic and photonic performance. This is done as processes are developed to be compatible with standard CMOS processing, which is performed in cleanrooms. Courtesy of IMEC.

There are challenges with this, though. The 2D materials are only a few atomic layers thick. They must be grown and then transferred to a CMOS structure. The 2D materials are all surface, which means that control of the interface is important. What’s more, they also tend to be the weakest link — and thus the most likely to shear off — if part of a stack.

According to Huyghebaert, the fact that 2D materials are such good sensors means the conditions of the material and its interaction with the outside world must be carefully managed. If not, the high sensitivity means the 2D materials will act as noise amplifiers, rendering useless the devices they go into.

But, the potential of 2D materials is significant, not just in photonics and electronics but in many other areas, according to Huyghebaert. In part, this is because there are so many different possible materials and combinations.

“You have a portfolio of these materials and you can heterostack them,” he said. “You can do a lot of novel things with these materials. You can really engineer these materials.”