Judges’ Queries and Presenter’s Replies

  • May 21, 2013 | 04:47 p.m.

    The bistability I understand as having consequence for computing applications. But what role does the optical nonlinearity play? Can you explain in layman’s terms please?

  • Icon for: Xiaoliang Zhu

    Xiaoliang Zhu

    Presenter
    May 22, 2013 | 12:30 a.m.

    Optical non-linearity is used for wavelength multiplication and pulse shaping, both of which can improve optical communications.

    For example if you input two colors of light, instead of the original two colors coming out, a third color also comes out containing the same modulated information as the original signals. The third signal is created by the non-linear interaction. Using nonlinearity we can perform the same functions as an electronic mixer in a radio receiver.

    The other possible use is to generate very short optical pulses, which enables us to increase the data-rate in the communication network. In 20 years we could be using optical links operating faster than 100 Gbits/second.

    Optical nonlinearity is the only way to change light on the optical time scale – think the time it takes for a photon to oscillate one cycle. Its discovery yielded the Nobel prize in 1981.

  • Icon for: Qiaobing Xu

    Qiaobing Xu

    Judge
    May 21, 2013 | 05:58 p.m.

    nice work, xiaoliang. The graphene layer coating is interesting to me. is the graphene layer used as the cladding layer? could other materials, e.g. a layer of silicon by CVD, be used to achieve similar function? to my understanding, one of the problem of optical waveguides is the high loss of signal during propagation. is this true? you mentioned your design has a low loss propagation. how did you achieve it?

  • Icon for: Tingyi Gu

    Tingyi Gu

    Co-Presenter
    May 22, 2013 | 12:12 a.m.

    Yes, graphene serves as the top cladding layer for enhancing the optical nonlinearity (Two photon absorption and Kerr nonlinearity). Graphene itself has the third order nonlinearity 5 orders of the silicon, and thus the single layer carbon sitting on the 250nm Silicon ‘bulk’ can enhance Chi(3) of the effective media 20 times – compared to the monolithic silicon device.
    After transferring graphene, We measured the linear propagation loss increases less than 1 dB over the 200 micrometer long photonic crystal waveguide, compared to the control without graphene.

  • Icon for: Xiaoliang Zhu

    Xiaoliang Zhu

    Presenter
    May 22, 2013 | 12:35 a.m.

    One more point about the propagation loss in waveguides – the loss is mostly due to edge roughness from the fabrication process. Over the years this has steadily improved and now we can expect 1-2 dB per centimeter of loss on chip.
    The manufacturing process at the Institute of Microelectronics in Singapore, which was used to manufacture some of the devices, is very good and reproducible.

  • May 21, 2013 | 08:21 p.m.

    Light-based processes significantly enhance the operational speed. However, optical elements require larger sizes, comparable with the wavelength of light. What do you think about the use of plasmonics, which combines advantages of small size and optical speed of operation?

  • Icon for: Tingyi Gu

    Tingyi Gu

    Co-Presenter
    May 22, 2013 | 12:03 a.m.

    Hi Prof. Noginova,
    Yes. Graphene is a zero-bandgap material, and enable optical devices responsing from terahertz to visible light. Its high mobility, high nonlinearity allows high speed, low power operation.
    Plasmonics offers tight confinement in small scale devices, but its realization with conventional metal always face the problem of optical damping/loss. The low linear absorption (2.3%) of graphene, two dimensional electron gas, maybe a key to solve the problem.
    - Tingyi & Xiaoliang

  • Icon for: Xiaoliang Zhu

    Xiaoliang Zhu

    Presenter
    May 22, 2013 | 01:06 a.m.

    I remember seeing a very exciting application of plasmonics as a high-speed absorption based modulator at CLEO 2011 (post deadline) from Prof. Xiang Zhang’s lab at Berkeley. They were able to perform optical modulation using a plasmonic absorption effect in a wavelength size device. So there are definitely applications for plasmonics in next generation optical interconnects.
    http://www.opticsinfobase.org/abstract.cfm?URI=CLEO_AT-2012-CTh5D.1

  • Icon for: Qi-Huo Wei

    Qi-Huo Wei

    Judge
    May 21, 2013 | 08:32 p.m.

    Xiaoliang: Very interest idea. To put graphene on top of 2D photonic crystal, you are using the cavity mode or slow light to enhance the nonlinear effects on graphene, am I right? Could you explain why graphene is a better nonlinear material than other semiconductor material?

  • Icon for: Tingyi Gu

    Tingyi Gu

    Co-Presenter
    May 22, 2013 | 12:12 a.m.

    Yes, we use cavity for enhancing light-matter interaction by circulating photon in a small area for longer time.

    Graphene has delocalized bond electrons for high optical nonlinear response at all photon energies. Its nonlinear coefficient is much higher than the conventional semiconductors and metals, e.g. five order higher than silicon, fourth order of GaAs. Its single layer structure leads to easy integration.

    Tingyi & Xiaoliang

  • May 21, 2013 | 11:54 p.m.

    Very interesting work. But, can you explain how graphene monolayer is readily deposited on the silicone membrane through drying process? Also, how can we confirm the monolayer deposition?

  • Icon for: Tingyi Gu

    Tingyi Gu

    Co-Presenter
    May 22, 2013 | 12:03 a.m.

    Hi Prof. Kong,

    We use CVD to grow single layer graphene on copper, etched away the backside copper by acid, and then transfer the graphene on to the silicon membrane. The monolayer graphene is confirmed by the Raman spectrum. The narrow 2D band indicates the single layer feature. For more details, please refer to the paper:
    http://www.nature.com/nphoton/journal/v6/n8/ful...

    Thanks,
    Tingyi & Xiaoliang

  • Further posting is closed as the competition has ended.

Presentation Discussion

  • Icon for: Brian Drayton

    Brian Drayton

    Faculty
    May 23, 2013 | 11:53 p.m.

    I enjoyed your video, thanks!
    I am curious — as a nonengineer, and non physicist, how “interdisciplinary” your team is? That is, how different are the technical languages/preferred methodologies of the different disciplines represented? What has been the biggest learning curve?

  • Icon for: Xiaoliang Zhu

    Xiaoliang Zhu

    Presenter
    May 24, 2013 | 12:51 p.m.

    Great question. I’m an Electrical Engineer and Tingyi is in the Mechanical Engineering department with a background in physics. I’m interested in the performance of the devices as modulators and switches, and the implications of such performance on a systems and applications level. Tingyi is the expert in the material properties and fabrication requirements of these devices.
    It may seem that these areas are completely different, but once you start working in lab it is easy to learn how different methodologies work. The biggest learning curve for me was understanding the properties of graphene and step into quantum mechanics.

  • Further posting is closed as the competition has ended.

  1. Xiaoliang Zhu
  2. http://www.igert.org/profiles/5383
  3. Graduate Student
  4. Presenter’s IGERT
  5. Columbia University
  1. Tingyi Gu
  2. http://www.igert.org/profiles/5334
  3. Graduate Student
  4. Presenter’s IGERT
  5. Columbia University

Silicon Photonics and Novel Materials for Datacenters and Cloud Computing

Our growing use of the internet and cloud computing challenges datacenters in terms of performance and energy usage. To reduce environmental impact of datacenters and to increase performance, we need a better technology to share data between computers in the datacenter. Silicon photonics, which enables the use of laser light to send massive amounts of data on-chip, is a promising technology. We demonstrate devices combining silicon photonics and novel materials to observe exciting phenomenon useful for next generation high-speed optical processing.