- By John Hewitt
At MWC 2013 in Barcelona, Samsung has announced that it will be funding new research into graphene-based antennae for intra-chip communication in the terahertz band. As part of their Global Outreach Program, Samsung will be giving $120,000 to a multidisciplinary team comprised of researchers from the and the Georgia Institute of Technology. The primary focus of the grant is to develop the coding and modulation schemes necessary for wireless, internal communication at hundreds of gigabytes per second among the thousands of sub-processors of a multi-core chip — an impossible feat with current technology.
Multicore processors can have a large number of subunits that share and execute tasks in parallel. Wireless communication is the next logical step in processor optimization which would relax many of the constraints imposed by traditional planar topologies wired on 2D surfaces. While the terahertz band has been extensively studied for imaging (See: How terahertz laser scanners will spy on you in airports) and communications on scales of meters, reducing things to the nanometer regime presents new challenges. Terahertz radiation (0.1 to 10THz) can be approached from the radio-frequency or the optical perspective, depending on whether you look at it as high frequency waves, or low energy photons. Classical antenna theory predicts that the frequency to efficiently operate a nanoscale antenna would be prohibitively high. However the reduced speed of electrons in graphene suggests that an atomically defined antenna would have a resonant frequency at much more practical values.
Sparked by the 2004 Nobel Prize for the isolation of pure samples of graphene, a breakneck roadmap has been laid out for graphene-based devices. Competing research onplasmonic graphene antennas are ongoing at the Department of Energy’s Oak Ridge National Laboratory and elsewhere. One buzzword that has already emerged in this new field is graphene-enabled wireless network-on-a-chip (GWNoC). There is still much clarification of terms to be made when talking about on-chip (intra) and between-chip (inter) communications. For the most part, nanometer-scale graphene antennae, or graphennae as we like to call them here in the ExtremeTech bunker, will be restricted to operation below a distance of about one centimeter, which might be considered huge by today’s smartphone standards. (See: IBM creates first cheap, commercially viable, electronic-photonic integrated chip for on-chip laser communications.)
Both research groups have ongoing efforts in other graphene-related technologies. Theoriginal grant proposal describes their ambitions to revolutionize the ways processors and memory interact from a multicore perspective, for instance, in terms of data/cache coherence, consistency, and synchronization. They mention that among other technologies, it has recently been shown the emission of photons from nano-structures due to electron-phonon interaction has motivated the study of nanotubes and ribbons as optical emitters or detectors, potentially also in the terahertz range.
With the mention now of phonons we have entered into more esoteric physics — the elemental way in which quanta of heat move through solids as waves or vibrations. Even at the larger scale, however, the value of these types of communications have their place as ants conducting signals more efficiently by drumming on branches rather than emitting into the air, or elephants thumping the ground with seismic infrasound. On board a chip, we are just beginning to imagine the full potential of this technology.
No comments:
Post a Comment