Power/Performance Bits: April 26

An on-chip light source; controlling integrated optical circuits; fabricating fast, flexible transistors.


An on-chip light source

Researchers at the Karlsruhe Institute of Technology (KIT) demonstrated that carbon nanotubes are suited for use as an on-chip light source. By integrating tiny carbon nanotubes into a nanostructured waveguide, the team developed a compact miniaturized switching element that converts electric signals into clearly defined optical signals.

“The nanostructures act like a photonic crystal and allow for customizing the properties of light from the carbon nanotube,” explained Felix Pyatkov and Valentin Fütterling of KIT’s Institute of Nanotechnology. “In this way, we can generate narrow-band light in the desired color on the chip.” Processing of the waveguide precisely defines the wavelength at which the light is transmitted. By engravings using electron beam lithography, the waveguides of several micrometers in length are provided with finest cavities of a few nanometers in size. They determine the waveguide’s optical properties. The resulting photonic crystals reflect the light in certain colors, the same phenomena that results in butterflies’ colorful wings.

Carbon nanotube above a photonic crystal waveguide with electrodes. The structure converts electric signals into light. (Source: WWU)

Carbon nanotube above a photonic crystal waveguide with electrodes. The structure converts electric signals into light. (Source: WWU)

As novel light sources, carbon nanotubes of about 1 micrometer in length and 1 nanometer in diameter are positioned on metal contacts in transverse direction to the waveguide. At KIT, a process was developed where the nanotubes can be integrated specifically into highly complex structures.

This compact electricity/light signal converter meets the requirements of the next generation of computers that combine electronic components with nanophotonic waveguides, according to the researchers. The signal converter bundles the light about as strongly as a laser and responds to variable signals with high speed. Already, the optoelectronic components developed by the researchers can be used to produce light signals in the gigahertz frequency range from electric signals.

Controlling integrated optical circuits

Researchers from the University of Southampton and the Institut d’Optique in Bordeaux devised a new approach for controlling light in a silicon chip by bringing the concept of spatial light modulation to integrated optics.

Photonic chip functionality is usually hard-wired by design, however reconfigurable optical elements would allow light to be routed flexibly, opening up new applications in programmable photonic circuits.

The team made use of multimode interference (MMI) devices, which form a versatile class of integrated optical elements routinely used for splitting and recombining different signals on a chip. The geometry of the MMI predefines its characteristics at the fabrication stage.

The intricate interplay between many modes travelling through the MMI can be dynamically controlled, the researchers showed, by a pattern of local perturbations induced by femtosecond laser which act in concert to effectively shape the transmitted light. Related to wavefront shaping in free-space optics, this allows to freely route light in a static silicon element, thus transforming the device into a much needed building block for field-programmable photonics.

An artistic rendering of a silicon-on-insulator 1×2 multimode interference splitter with a projected pattern of perturbations induced by femtosecond laser. The perturbation pattern achieves routing of light to a single output port with 97 percent efficiency. (Source: University of Southampton)

“This is a potentially disruptive new approach toward field-programmable chips which can enhance and complement existing strategies, or even partially replace current technology,” said Professor Otto Muskens, from Physics and Astronomy at the University of Southampton.

Practical applications of this technology will include all-optical reconfigurable routers, ultrafast optical modulators and switches for optical networks and microwave photonic circuits as well as wafer-scale optical testing of photonic chips. However, more work is needed to develop these ideas into practical applications.

Fabricating fast, flexible transistors

University of Wisconsin-Madison engineers developed a unique method that could allow manufacturers to easily and cheaply fabricate high-performance transistors with wireless capabilities on huge rolls of flexible plastic.

The researchers fabricated a transistor that operates at a record 38 gigahertz, though their simulations show it could be capable of operating at 110 gigahertz. The transistor can transmit data or transfer power wirelessly, a capability that could unlock advances in a whole host of applications ranging from wearable electronics to sensors.

The transistor, consisting of single-crystalline silicon on a polyethylene terephthalate (or PET) substrate, was patterned through blanketing the entire single crystalline silicon with a dopant, rather than selectively doping it.

Using a unique method they developed, a team of UW–Madison engineers has fabricated the world’s fastest silicon-based flexible transistors, shown here on a plastic substrate. (Source: Jung-Hun Seo)

Then, they added a photoresist layer and used e-beam lithography on the photoresist to create a reusable mold of the nanoscale patterns they desired. They applied the mold to an ultrathin, very flexible silicon membrane to create a photoresist pattern. Then they finished with a dry-etching process that cut precise, nanometer-scale trenches in the silicon following the patterns in the mold, and added wide gates, which function as switches, atop the trenches.

Because the researchers’ method enables them to slice much narrower trenches than conventional fabrication processes can, it also could enable semiconductor manufacturers to squeeze an even greater number of transistors onto an electronic device, as well as providing more efficient operation.

According to Zhenqiang (Jack) Ma, professor of electrical and computer engineering at UW-Madison, because the mold can be reused the method could easily scale for use in roll-to-roll processing.