- Describe the major research and
education activities of the
project.
The Center for Bits and
Atoms is
studying two core complementary questions: how can a logical
description be embodied in, and abstracted from, a physical form?
These are old questions, dating back to the origins of modern
manufacturing, and before that to the birth of the natural sciences,
but they are profoundly changed by the recognition of the computational
universality of nature.
CBA comprises 15 research groups from across MIT's campus, including
physicists, chemists, biologists, mathematicians, computer scientists,
and electrical and mechanical engineers. All are working at the
boundary between logical and physical descriptions of information, an
essential interface that has historically been divided between hardware
and software, and between physical science and computer science. Their
work is supported by a unique experimental resource enabling input and
output across these discipline boundaries, from nanometers to meters.
A focal event over the last year was a workshop CBA hosted on
"Avogadro-Scale Engineering", gathering collaborators on campus and
from around the world to explore the need for, and prospects for,
engineering practice in a limit of thermodynamic complexity. Two key
ideas from this meeting are at the heart of CBA's research program:
-
Form:
the use of error-correcting
tools to assemble perfect systems from imperfect components. This
promises to have the same impact on fabrication as digital threshold
theorems have had on communications and computation.
- Function:
the use of physical
realizations of mathematical programs as an architecture for organizing
physical resources to solve computational problems, from devices to
circuits to systems.
These are connected by
investigation of transduction at the interface between physical form
and logical function.
CBA's activities are complemented on-campus by a growing sequence of
classes providing training in these interdisciplinary technical
research areas, and collaborations on applications of CBA's
infrastructure. Beyond MIT, CBA is deploying a network of field "fab
labs" around the world, providing access in non-traditional communities
to prototype tools for personal fabrication. And CBA's research is
transferred to industry through joint development projects, including
the
Internet 0 (I0) initiative.
- Describe the major findings
resulting
from these activities.
A central focus that emerged over the
last
year is the role of error correction in fabrication. Macroscopically,
this was studied through the development
of mechanical and electromechanical finite-state machines for
programmed assembly:
[Griffith, 2004]. These have been used to demonstrate linear extrusion of a space-filling
curve, mechanical
assembly of a self-reproducing string, and error-corrected growth of a
crystal:
(images linked to videos).
There are CBA projects seeking comparable control over programmed
assembly across the range of systems and length scales represented. At
molecular scales, new materials are being developed based on molecular self-assembly [Zhang,
2003], and error-corrected
assembly is being applied to de novo
fabrication of large DNA
molecules with a goal of being able to synthesize an entire genome. In
support of this a "green
screen" GFP assay was developed to quantitate successful gene syntheses
through simple colony counts:
This rapid metric
has already enabled the improvement
of gene synthesis protocols to reach an error rate of 0.001 per base
synthesized. While this is better than anything in the current
literature, ongoing work is investigating error-correcting enzymes
and genetic circuits for asymptotic scaling [Carr, to be published].
To provide external control over molecular function, an earlier report
described a CBA project functionalizing proteins with nanocrystal
antennas to electromagnetically control their conformation [Hamad-Schifferli
et al., 2002; Hamad-Schifferli, 2004].
Because this mechanism is poorly understood, self-assembled monolayers on gold
nanoparticles have been used to adsorb DNA oligonucleotides in a radial
configuration in order to monitor the local heat distribution. A
change was observed in the hydrodynamic radius through analysis of gel
electrophoresis
following treatment with mercaptohexanol (MCH) [Park,
2004].
Reversing roles, instead of using a nanoparticle to control a protein,
L-lysine (Lys) monofunctionalized gold
nanoparticles were synthesized
by a solid-phase reaction using
4-hydroxymethylphenoxyacetyl (HMPA) -- polyethylene glycolacrylamide
copolymer (PEGA) resin:
[Sung et al., 2004]. This allows the assembly of functional nanoparticles to be coded for by the attached ligand. The
monofunctionalization was confirmed by HRTEM observation of
dimerization of the nanoparticles:
For larger-scale patterning of functional nanoparticles to assemble
active systems, earlier work reported used CBA's beam tools to
direct-write nanostructures
[Griffith et al., 2002]. This year, an offset liquid embossing process was
developed for printing three-dimensional structures with nanoparticles:
[Wilhelm and Jacobson, 2004]. This process was used to fabricate an
array of electrostatically-actuated mirrors, imaged here with CBA's
confocal microscope and linked to a video taken with its ESEM:
A printing process was also developed for
patterning networks of neurons and glia cells:
[Fuller, 2003].
Finally, on the largest length scales, geometries
were developed to encode the global structure of a building in the
local constraints of it components, shown assembling a dome:
[Seely and Sass, to be pubished]. The
state of the art in modern architecture uses NC tools in
fabrication but still relies on conventional blueprints for
construction; this approach promises to extend the computational
description of a design all the way from conception to construction.
Taken together, the preceeding results provide a framework for the
development of a "logical" manufacturing practice that builds with
intelligence that is internal to the materials rather than contained in
an external controller. The second broad area of CBA activities is
exploring mechanisms and architectural principles to organize such
materials to solve problems.
Two projects considered electromagnetic properties. The first, reported
last year, constructed an artificial "metamaterial" with a negative
index of refraction [Houck et al., 2003]. This was
picked by Science magazine as one of the top 10
scientific highlights of 2003. This year, the internal and
external foci predicted for a rectangule slab of a negative index
material were observed:
[Brock, Houck, and Chuang, to be published]. And the second project
used CBA's laser micromachining tool to create a slot antenna in a
microstripline:
[Maguire, 2004]. This creates an
impedance-matched discontinuity locally converting RF energy to a
strong homogeneous magnetic field, and has shown the
best reported sensitivity for a planar spin-resonance probe.
Here is a two-dimensional spectra obtained from a picomolar sample of
an acetyl-amide peptide:
This project grew out of CBA research seeking to improve the scaling
of molecular quantum computing, but is
likely to have significant implications for molecular
biology by making structural measurements possible on much smaller
samples. Another spinoff from that effort comes from the pulse
sequences that have been developed for experiments such as last year's
report of the first implementation of an adiabatic algorithm [Steffen
et al., 2003]. These composite pulses have been
found to be much more broadly applicable in controlling coherent
quantum information, including restricting
a three-level Josephson junction to a two-dimensional qubit Hilbert
space [Steffen, Martinis, and Chuang, 2003].
CBA's program for providing macroscopic molecular interfaces has
progressed with the development of two new MEMS structures. Last year
we
reported electronic readout of DNA hybridization with single-base
selectivity [Fritz et al., 2002]. This year, a
suspended microchannel resonant mass
sensor was developed for specific biomolecular detection in a
sub-nanoliter fluid volume.
This measures shifts in the
resonance frequency of a suspended microfluidic channel upon
accumulation of molecules on the inside walls of the device,
enabling direct integration with conventional microfluidic systems,
significantly increasing sensitivity by reducing ambient damping and
viscous drag, and allowing the resonator to be actuated by
electrostatic forces. This mechanism has been demonstrated by measuring
the specific binding between avidin and biotinylated Bovine Serum Albumin
(bBSA):
[Burg and Manalis, 2003; Shusteff, 2003; Levy-Tzedek, 2004]. These measurements are expected to
reach changes in surface mass loading
order of 10-19 g/um2.
A related result builds on last year's report of
peptide nanotubes [Santoso et al., 2002], to design
peptide surfactant detergents that can
solubilize, stabilize and crystallize membrane proteins:
[Yang, 2004]. Because all 5 of our senses are based on membrane
proteins, this significantly enlarges the class of candidate molecular
sensors. And for controlling fluids at molecular scales, a capacitative readout was developed for
a flexural valve with nanometer displacement and resolution:
[Ma et al., 2003 (won a student paper award)].
The digital interface to analog electrical signals
was improved by the
development of an A/D converter using a
new successive approximation algorithm for temporal rather than voltage
or current degrees of freedom.
This has a conversion time that
grows linearly with the precision of the converter, and
is expected to scale to sub-uW power consumption:
[Heemin and Sarpeshkar, to be published]. Along with the
analog storage cell reported last year
[O'Halloran and R. Sarpeshkar, 2004],
this is part of a growing family of mixed-signal components using
continuous device degrees of freedom. Two theses in the last year
explored circuit
architectures for such devices. One grew out of the study of fault
tolerance in quantum computation, finding that
classical fault-tolerant constructions
can improve the reliability and
efficiency of conventional circuits as they approach
fundamental scaling limits [Impens, 2004].
And a second applied message-passing algorithms to circuit design,
rolling up a state estimation trellis to obtain a "Noise-Locked Loop" that acquires a
spread-spectrum code using an analog logic representation of bit
probabilities:
[Vigoda, 2003]. The messages passed in this circuit can be understood as
dynamically
solving a mathematical program; current work is extending this insight
to organize large-scale distributed systems. CBA support
contributed to the development of a
prototype 1000 node "paintable" computer that
is programmed by propagating mobile code:
[W. Butera, to be published].
The relationship between physical
dynamics and algorithm dynamics is also being explored in the context
of mechanical control; a passive
dynamical walker was developed that is controlled by an onboard online
reinforcement learning algorithm:
[Tedrake et al., 2004], and a synthetic skin is being
developed as a distributed physical interface [Stiehl et al.,
2004].
- Describe the opportunities for
training and development provided
by your project.
CBA has directly supported about 50 students and indirectly contributed to about 150 students, working with a unique experimental resource developed to provide input and output across 9 orders of
magnitude:
A popular rapid-prototyping class was
developed to provide instruction in its integrated use, MAS.863: How To Make (almost) Anything. This
has led to student research projects such as the production of photonic
band-gap materials:
[J. Walish, to be published].
Rapid-prototyping of microstructures is being taught in 6.151:
Semiconductor Devices
Project Laboratory. The class project this year made a
significant research contribution in developing a microfluidic device
with integrated DNA amplification and detection:
Hands-on training
in quantum computing is provided by a project developed for
8.13: Experimental
Physics:
Other classes that CBA has directly or indirectly contributed to
include:
4.206: Introduction to
Computing
4.212: Design
Fabrication
4.173: Design
Fabrication Workshop
6.971: Engineering
Simple Biological Systems
7.86, BE.481, MAS.866: Fundamental
Limits of Biological Measurement
8.371J, MAS.865J: Quantum
Information Science
BE.442: Molecular
Structure of Biological Materials
MAS.862: The Physics of Information Technology
MAS.864: The Nature of
Mathematical Modeling
MAS.961: How To Make
Something That Makes (almost) Anything
- Describe outreach
activities
your project has undertaken.
The focus of CBA's outreach activities
has been its field "fab lab" program. This is bringing prototype
versions of the on-campus infrastructure beyond campus, to explore
their implications and applications in non-traditional communities.
Capabilities currently being deployed include laser cutting of
three-dimensional structures, sign cutting of interconnect and
electromagnetics, precision machining of circuit boards, and
programming of embedded RISC microcontroller signal chains. Together,
these provide access to engineering in space and time down to microns
and
microseconds. Fab labs have been opened so far in inner-city Boston,
rural India, northern Norway, Costa Rica, and Ghana:
(here an 11-year-old girl is successfully
learning to stuff a surface-mount circuit board in the Boston lab).
Emerging lessons from the fab lab project include the recognition that
beyond the digital divide there are instrumentation and
fabrication divides, and that these can be addressed by bringing IT
development rather than just IT to the masses. These themes are being
explored with a growing
group of institutional partners including the National
Academies, the Indian Department of Science and Technology, and the
Africa-America
Institute. A related project to
produce
eyeglasses in the field
led to a
CBA-funded student, Saul Griffith, winning the Lemelson Foundation
Student Prize for inventiveness:
CBA is participating in a number of joint
development projects with industry for technology transfer; one with
broad economic and social implications is the "Internet 0" (I0)
effort. This grew out of the need to develop embedded IP processors for
interfacing distributed
devices:
and led to the recognition that for low data-rate
devices a near-field time-domain encoding can be used that is the
same across physical transports:
This
picture of an I0 packet looks the same whether it is scanned optically,
capacitatively coupled through a powerline, clicked acoustically, or
sent electromagnetically. Just as the IP protocol first enabled
internetworking across heterogeneous networks, I0 is extending that
insight to "interdevice internetworking" as an alternative to the
proliferation of standards for different devices [Gershenfeld et al.,
2004]. An industry initiative is developing around this project, driven
by embedded networking applications including construction and energy
efficiency, healthcare monitoring, and distributed user interfaces.
Current development work is building on related CBA activities
including the
use of
physical mechanisms
for security [Pappu et al., 2002], and study of scaling and
heterogeneity in large-scale networks [K. Sollins, to be published].
Publications
and Products
- What have you published as a
result of
this work?
Journal publications
T. Burg and S.R. Manalis,
Suspended microchannel resonators for
biomolecular detection,
Applied Physics Letters (83), 2698 (2003)
J. Fritz, E.B. Cooper, S. Gaudet, P.K. Sorger, and S.R. Manalis,
Electronic detection of DNA by its
intrinsic molecular charge,
Proceedings of the National Academy
of Sciences (99), 14142-14146 (2002)
N. Gershenfeld, R. Krikorian, and D.
Cohen, Internet 0: Interdevice
Internetworking, Scientific American, to appear (2004)
Saul Griffith, Mark Mondol, David S. Kong, and Joseph M.
Jacobson, Nanostructure
fabrication by direct electron-beam writing of
nanoparticles, J. Vac. Sci. Technol. (B 20), 2768-2772 (2002)
A.A. Houck, J.B. Brock, I.L. Chuang,
Experimental Observations of a Left-Handed
Material That Obeys Snell's Law,
Phys. Rev. Lett. (90),
137401/1-4 (2003)
K. Hamad-Schifferli, J.J. Schwartz, A.T. Santos, S.G. Zhang, and J.M.
Jacobson,
Remote Electronic Control of DNA Hybridization Through Inductive Coupling to an Attached Metal Nanocrystal Antenna,
Nature (415), 152-155, (2002)
Ma, H., White, J., Paradiso, J., and
Slocum, A.,
Sub-nanometer Displacement Sensing for the Nanogate,
In the Proceedings of the 2003
IEEE International Conference on Sensors, October 21-24, Toronto
Canada, pp. 46-51 (2003)
M. O'Halloran and R. Sarpeshkar,
A 10nW 12-bit
Accurate Analog Storage Cell with 10aA Leakage,
IEEE Journal of Solid State Circuits, to appear (2004)
R. Pappu, B. Recht, J. Taylor, N. Gershenfeld
Physical One-Way Functions,
Science (297), 2026-2030 (2002)
S. Santoso, W. Hwang, H. Hartman, S. Zhang,
Self-Assembly of Surfactant-Like Peptides with Variable Glycine Tails
to Form Nanotubes and Nanovesicles,
NanoLetters (2), 687-691 (2002)
M. Steffen, J. Martinis, I. Chuang,
Accurate control of Josephson phase qubits,
Physical Review B, vol 68, num 224518/1-9, 2003
M. Steffen, W. van Dam, T. Hogg, G. Breyta, and I. Chuang,
Experimental Implementation
of an Adiabatic Quantum Optimization
Algorithm, Phys. Rev. Lett. (90), 067903/1-4 (2003)
Kie-Moon Sung, David W. Mosley, Beau R.
Peelle, Shuguang Zhang, and Joseph M. Jacobson,
Synthesis of Monofunctionalized Gold
Nanoparticles by Fmoc Solid-Phase Reactions, J. Am. Chem. Soc
(126), 5064-5065 (2004)
Russ Tedrake, Teresa Weirui Zhang, Ming-fai Fong, and H. Sebastian
Seung, Actuating a Simple 3D Passive
Dynamic Walker, IEEE International Conference on Robotics and
Automation (ICRA 2004)
David W. Mosley, Mark A. Sellmyer, Erin J. Daida, and Joseph M.
Jacobson., Polymerization of
Diacetylenes by Hydrogen Bond Templated Adlayer, Formation, J.
Amer. Chem. Soc. (125), 10532-33 (2003)
Eric J. Wilhelm and Joseph M. Jacobson,
Direct printing of nanoparticles and
spin-on-glasses by offset liquid embossing, Appl.
Phys. Lett. (84), 3507 (2004)
Zhang, S., Fabrication of novel materials
through molecular
self-assembly,
Nature Biotechnology (21), 1171-1178 (2003)
Books or other non-periodical, one-time publications
S.B. Fuller,
A Fast Flexible Ink-Jet Printing Method for Patterning Networks of Neurons in Culture,
M.S. thesis, MIT (2003)
Saul Griffith, Growing Machines,
Ph.D. thesis, MIT (2004)
K. Hamad-Schifferli,
DNA Hybridization, Electronic Control, in Encyclopedia of
Nanoscience and Nanotechnology, edited by J. A. Schwarz, C. Contescu
and K. Putyera (Marcel Dekker, New York, 2004).
F. Impens,
Fine-Grained Fault-Tolerance: Reliability
as a Fungible Resource,
M.S. thesis, MIT (2004)
S. Levy-Tzedek, Biological Detection by means of Mass Reduction in a
Suspended Microchannel Resonator, M.S. thesis, MIT (2004)
Yael Maguire,
Scalable Electromagnetic Microstructure
Instrumentation, Ph.D. thesis, MIT (2004)
Sunho Park,
Control of
Oligonucleotide
Conformation on Nanoparticle Surfaces for Nanoscale Heat Transfer Study,
M.S. thesis, MIT (2004).
Maxim Shusteff,
A Microfabricated Hollow Cantilever Sensor for Sub-nanoliter Thermal Measurements,
M.S. thesis, MIT (2003).
W.D. Stiehl, L. Lalla, and C. Breazeal, A Somatic
Alphabet Approach to Sensitive Skin, Proceedings of the ICRA
(2004)
Kie-Moon Sung, David W. Mosley,
Beau R. Peelle, Shuguang Zhang and Joseph M. Jacobson,
Nanopearls: A New Synthetic Approach Towards Creating
Monofunctionalized Building Blocks for Programmable Biochemical
Nanocrystal Linear Sequences, in Proceedings
Foundations of Nanoscience, Snowbird, UT, April 21-23,
p. 105 (2004).
B. Vigoda,
Analog Logic: Continuous-Time Analog Circuits for Statistical Signal
Processing, Ph.D. thesis, MIT (2003)
Steve Yang,
Self-Assembly of Surfactant-like Amphiphilic Peptides made of Natural Amino Acids, Ph.D.
thesis, MIT (2004)
Oren Zuckerman,
System Blocks: Learning about Dynamic
Behavior through Hands-on Modeling and Simulation, M.S. thesis, MIT
(2004)
- What Web site or other Internet
site
have you created?
The primary CBA site is http://cba.mit.edu,
containing archives of publications and videos from events including
the weekly Colloquia, http://cba.mit.edu/presentations/,
and the Avogadro-Scale Engineering meeting, http://cba.mit.edu/events/03.11.ASE/.
CBA's fab classes and fab labs on and off of campus are run from http://fab.cba.mit.edu. CBA has also
helped support http://www.thinkcycle.org
for open collaborative design.
- What other specific
products
(databases, physical collections,
educational aids, software, instruments, or the like) have you
developed?
CBA's progress both at MIT and in the
field fab labs has been constrained by the limitations of available CAM
software, including restrictive assumptions about possible workflows,
unreliable algorithms, steep learning curves, and expensive licenses.
This led to the development of a single CAM environment to connect all
of CBA's tools, cam.py:
It can currently read 2D DXF and SVG drawings, 3D DXF shapes, JPG and
TIFF images, Gerber PCBs, and Excellon drill holes, and can write G
codes and RML for NC machining, EPI and UNI for laser cutting, ORD for
waterjet cutting, CAMM for knife cutting, and JPG for beam rasters,
with more formats being added.
A second limitation has been the need for collaborative Web site
development across the field fab labs in order to share projects,
instructions, and files. For this a site authoring tool was written,
site.py:
that exposes a command shell within a Web page.
For teaching kids about core concepts in systems dynamics, "System
Blocks" were developed that let dynamic systems be assembled from
computationally-enhanced blocks:
[Zuckerman, 2004].
|