Computational Nanotechnology: the CCN and the NCN

1
ComputationalNanotechnologythe CCN and the NCN
Univ. of Florida, Univ. of Illinois, Morgan State
Univ., Northwestern Univ. Purdue Univ. Stanford
Univ., UTEP

• Mark Lundstrom
• Purdue University

2
NCN Research

• significant technological promise
• challenging scientific questions
• multi-scale, multi-disciplinary
• - serious computational challenges
• tight links to experimentalists
• clear focus, 3-year duration

systems
integrated nanosystems
devices
chemistry/ materials

• - leadership and networking
• general approaches
• - new software tools
• unique educational resources
• collaborative services
• web-based computing

atoms
3
connecting theory to experiment
styrene
I ---gt
I (nA)--gt
STM
V ---gt
100 Si
3) experimental confirmation M. Hersam,
Northwestern
1) molecules on Si have different and
potentially useful characteristics
Word Line
VDD
styrene on n - Si
molecular NDR
Current (nA)
Bit Line
C
Voltage (V)
2) prediction by simulation Datta et al., Purdue
4) applications in electronic memories K.
Roy, Purdue
4
connecting theory to experiment
High-? gate insulators for CNTFETs
Ohmic contacts / ballistic transport
Javey, Kim, Brink, Wang, Ural, Guo, McIntyre,
McEuen, Lundstrom and Dai, Nature Materisls, 1,
241, 2002
Javey, Guo, Wang, Lundstrom and Dai, Nature, 427,
654, 2003
MOSFET-like CNTFETs
High-field transport in short CNTs
Guo, Javey, Dai, Datta and Lundstrom, submitted,
2003
Javey, Guo, Paulsson, Wang, Mann, Lundstrom and
Dai submitted, 2003
5
more than computation
integrated problem solving
-experiment -theory -simulation -computation
6
connecting computer and physical scientists
H. Dai Group, Stanford

• nanowire transitors are scientifically
• interesting and technologically
• promising.

• Sameh and Polizzi developed
• efficient eigenvalue solvers

arbitrary 2D cross section
2 hrs per bias point on 44 cpus ..becomes 1
hour on 1 cpu
transport
2 hr on 44 1,2GHz processors

• Improvements in algorithms allow
• new problems to be solved.

• quantum scale simulations
• are computationally demanding

7
more than computation
integrated problem solving
cyber- community
-experiment -theory -simulation -computation
-community computing environment -collaborative
environment -unique educational resources
8
nanoHUB web-based, utility computing

• software
•  simulation services
•  educational resources
• data bases

• on-line since 1995
• gt 230,000 simulations since 2000
• gt 2400 users since 2000

PUNCH 5.0 community computing middleware
applications install as is
resources plug in

• USER selects application
• PUNCH identifies available hardware
• PUNCH runs application
• PUNCH displays results to user

9
building a cybercommunity
Read tutorials and work exercises with MolcToy
and FETToy
take a 2-day short course Electronic Device
Simulation at the nano/Molecular Scale
Hückel-IV 50 users / year gt 3500 simulations gt
100 to date
nanoMOS written in MATLAB parallelized in
MATLAB gt 400 downloads 1700 simulations 50
users / year
n
n
i- Si
BOX
Si substrate
take a full semester course for background
Read research papers and reproduce results
Hückel-IV or nanoMOS
join a Users Forum for tool support and
co-development
download source code for inspection or
installation
10
nanoHUB more than simulation

• - ready access to research grade simulation tools
• - easy access to high-performance computing
• on-line user training
• educational resources, full courses, short
  courses
• cutting edge seminars
• - collaborative tools

11
The NCN

• - research and education that makes a difference
• - new theory, approaches, algorithms, software,
  and
• educational resources
• a utility computing infrastructure that delivers
  unique
• simulation services
• a cyberinfrastructure that facilitates
  collaboration
• and education

12
The CCN and the NCN

• industrial strength utility computing middleware
• - expansion of the toolset
• co-development of educational content
• connections to Indiana schools, universities,
  industry