Fall 2008

Monday, October 20, 2008

Microscale manipulation of biological cells and their environment: New technologies for the biologist's toolkit
Nikhil Mittal
nmittal [at] mit.edu

Time: 8:00-9:00 PM
Location: Grier Room B (38-401B)

Abstract: Biological cells exhibit many complex behaviors. MEMS-based technologies can enable new kinds of assays that enhance the repertoire of tools available to biologists to dissect some of these behaviors. In the Voldman BioMEMS lab we have developed several such devices, and are testing their capabilities by using them to explore aspects of stem cell and cancer cell biology. In particular, we have developed devices with applications in cell sorting, fusion, and modulation of intercellular communication. In this talk I will describe the motivation behind, and operation of, some of these devices.

Slides: (pdf)

Wednesday, November 19, 2008

Inside the Playstation 5: How Silicon Photonics May Solve the Chip Bandwidth Challenge
Ben Moss
benm [at] mit.edu

Time: 7:00-8:00 PM
Location: Grier Room B (38-401B)

Abstract: Processor manufacturers have embraced parallelism, and soon we will see systems with dozens, hundreds, or even thousands of cores. Traditional electronic links will not satisfy the huge bandwidth needs of these processors, so we must look at new technologies to solve the chip bandwidth challenge.

What is Silicon Photonics? Can it really outshine electronics in terms of bandwidth density and energy efficiency? Come learn about the benefits and challenges in this multi-disciplinary field, presented by one of our own Course VI graduate students.

Slides: (pdf)

Spring 2008

Tuesday, February 19, 2008

Carbon Nanotubes for Integrated Circuit Interconnects and Devices
Gilbert Nessim
gdnessim [at] mit.edu

Time: 5:30-6:30 PM
Location: Grier Room (38-404)

Abstract: For Integrated Circuit nodes below 32 nm, carbon Nanotubes (CNTs) represent an ideal replacement for copper interconnects as they can carry higher current densities, don't need liners, and don't suffer from electromigration. However, fabrication issues and making electrical contacts remain major technical challenges.

We will show how carbon nanotubes can be 'tuned' during the thermal process. We found that varying the exposure time to the reducing gas can significantly affect the CNT density (number per area), the average tube length (carpet thickness) and the tube diameter (or number of walls). For instance, we could 'tune' the diameters of CNTs by a factor of two, from 14 nm (6-8 walls) to 7 nm (2-4 walls), and the areal densities by more than an order of magnitude, from 3x109 to 5x1010 CNTs/cm2.

To integrate CNTs into interconnect structures we need to growth them directly on metals, to achieve electrical contact. We have successfully grown vertically aligned nanotube mats on a variety of metallic underlayers at temperatures below 500 degrees Celsius. We will discuss the critical role of the metal layer below the catalyst in order to avoid alloying with the catalyst or in pinning the catalyst dots to its grain boundaries.

Slides: (pdf)

Fall 2007

Tuesday, September 11, 2007

Personal Robots
Mikey Siegel
mikeys [at] mit.edu

Time: 5:30-6:30 PM
Location: Kiva/Patil Seminar Room (Room 32-G449)

Abstract: The Personal Robots Group develops a wide range of robotic systems for a diverse range of applications including entertainment, health, education, and more. Its robots are designed to work and learn in partnership with people -- either as members as human-robot teams or in more personal face-to-face collaboration. Its creations range from highly expressive humanoids to embedding robotic technologies into everyday objects such as clothing and desktop artifacts. This is not only a technical endeavor, they also study how people and robots interact with one another, both inside and outside a laboratory setting.

This talk will present the work of the Personal Robots Group (formally the Robotic Life Group) as well as a brief overview of the relevant areas of Human Robot Interaction. Though technical comments and questions are welcome, the talk will be in the spirit of the Media Lab, focusing on showing the product itself using sound, video and images as much as possible.

Slides: (pdf)

Tuesday, October 2, 2007

When (Low) Power Really Matters
Yogesh Ramadass
ryogesh [at] mit.edu

Time: 5:30-6:30 PM
Location: Kiva/Patil Seminar Room (Room 32-G449)

Abstract: Technology scaling has driven portable electronics to new heights with newer, cooler features being added on every passing day. As this explosive growth continues, the power consumed by these portable electronic systems and their battery lifetimes are becoming as important as the new features advertised. Since battery technologies do not scale as well as the circuits, saving power at every stage of the system design becomes highly critical.

In this talk, I will describe the low-power circuit design research being carried on in Prof. Anantha Chandrakasan's lab. I will talk about the main circuit blocks that go into portable electronics systems and what our group does to reduce the power consumed by each of these blocks. I will briefly explain how these same ideas can be extended to devices that can operate out of scavenged energy.

Slides: (pdf)

Tuesday, November 13, 2007

Can Biology Inspire Better Circuit Design? The RF Cochlea as a Case Study
Soumyajit Mandal
soumya [at] mit.edu

Time: 5:30-6:30 PM
Location: Kiva/Patil Seminar Room (Room 32-G449)

Abstract: Circuit designers occasionally hear of biologically-inspired and bio-mimetic integrated circuits. Over the last 20 years or so, silicon cochleas, retinas, muscles and even a complete model of a cortical column have been built. However, the field has not grown as rapidly as most people anticipated, and examples of systems that successfully made the leap from academia into industry are rare.

In this talk, I shall begin by briefly explaining the challenges faced by designers who use algorithms and circuit architectures based on biology to solve an engineering problem. As examples, I will use systems built by Prof. Rahul Sarpeshkar's group at MIT, amongst others. I shall then focus on a particular example: the RF cochlea, which is an attempt to use a model of the mammalian inner ear to perform real-time spectrum analysis at radio frequencies. By illustrating the advantages of this approach, I shall try to show that circuit designers, if they are careful, can indeed use biology to make better integrated systems.

Slides: (pdf)

Spring 2007

Tuesday, February 20, 2007

From Bits to Qubits
Saikat Guha
saikat [at] mit.edu

Time: 5:30-6:30 PM
Location: Room 32-124

Abstract: High-speed modern day computing and information processing systems represent the culmination of years of technological advancements beginning with the early ideas of Charles Babbage (computing) and Claude Shannon (information). Richard Feynman was among the first to recognize the potential in harnessing properties of quantum mechanical systems, such as superposition and entanglement, for solving some computational problems much faster than conventional computers. Quantum computing makes use of 'qubits', which unlike 'bits' can exist as a zero, a one, or both simultaneously, with a numerical coefficient representing the probability for each state.  Peter Shor later came up with his famous factoring algorithm, which had a huge impact on the research on quantum computing becoming immensely popular and exciting.

This talk will introduce the basic ideas behind quantum information processing, leading up to discussing some potential applications. It will also provide interesting insight into complex modern day research on theoretical and experimental quantum information that is underway at MIT, primarily focusing on the work being done at the Optical and Quantum Communications Group, Research Laboratory of Electronics.

Slides: (pdf)

Tuesday, March 6, 2007

How to Wreck a Nice Beach: Theory & Practice
Paul Hsu
bohsu [at] mit.edu

Time: 5:30-6:30 PM
Location: Room 34-401B

Abstract: Communicating with machines via natural spoken interactions has been the dream of science fiction for decades. Yet, commercial speech recognition technology has primarily been relegated to dictation applications for users with repetitive strain injury (RSI) and telephony applications where the system dictates the interaction.

In this talk, we will explore the major elements that make up a speech recognizer and acquire a conceptual understanding for each component. We will then examine some of the most recent applications of speech recognition technologies. Finally, we will present some of the remaining hurdles facing the mainstream adaptation of speech interfaces. By addressing these challenges, we look forward to the day when speech interfaces replaces the keyboard and mouse as the dominant mode of interaction between humans and machines. This talk will incorporate research being done in the Spoken Language Systems Group, Computer Science and Artificial Intelligence Laboratory.

Slides: (pdf)

Tuesday, April 3, 2007

Although We’ve Come to the End of the Road(map): the Future of CMOS
Nicole DiLello
ndilello [at] mit.edu

Time: 5:30-6:30 PM
Location: Room 34-401B

Abstract: Moore’s Law has been the prominent driving force behind the increased performance of Complementary Metal-Oxide-Semiconductor (CMOS) devices for the last 40 years. We are quickly approaching a time when traditional scaling techniques will no longer suffice to make transistors smaller, cheaper, and more powerful. To design faster electronics, engineers are looking for novel methods to continue this scaling.

This talk will focus on the materials and processes that will be introduced in the upcoming years to push CMOS technology further. I will also discuss some larger systems currently being developed at MIT that aim to increase computing speed by integrating electronics and photonics on a single chip. This work is mainly being done in the Microsystems Technology Laboratory and the Research Laboratory of Electronics.

Slides: (pdf)

Tuesday, May 1, 2007

Working at the Interface of Engineering and the Life Sciences (literally)
Rumi Chunara
chunara [at] mit.edu

Time: 5:30-6:30 PM
Location: Room 34-401B

Abstract: Today we are confronted with refined abilities and techniques in engineering that enable applications in a wide realm of disciplines. Through sensitive electrical and mechanical design principles and silicon microfabrication technologies, we are able to probe and understand more about biological systems than previous capabilities have allowed.

How can we use physical and engineering techniques to transduce biological signals to classical realms? We will explore how the two fields intertwine in detection of very small levels of biomolecules and cells, for interesting applications in the laboratory and beyond. This talk will encompass interdisciplinary work being done in the Manalis Lab at MIT.

Slides: (pdf)