FCRP Logo - International Technology Roadmap for

Download Report

Transcript FCRP Logo - International Technology Roadmap for

Roadmap for Carbon
Nanotubes and Graphene
ITRS Logic Workshop
Tsukuba, Japan
George Bourianoff facilitating
Sept 23, 2008
Workshop and Business Meeting
Objectives (Sept. 22 – 23)

Layout roadmap for Carbon Nanotubes and Graphene
for “Ultimately Scaled CMOS” and “Beyond CMOS”.
Build on results from July ERD meetings.
–
–


Provide information needed to develop new Memory
Table entries for STT RAM (new TE for 2009).
Determine content for 2009 ERD logic section
–
–
–


Review Technology Entries (TEs) from 2007
Review potential TE adds/drops for 2009
Solicit writing volunteers for 2009
Discuss linkage to materials and architecture sections
–

Ultimate CMOS roadmap – potential solution (PIDS Role?)
Beyond CMOS roadmap entry – current ERD format?
How can we improve the integration? (e.g. joint workshops, key
materials properties table, …)
Approximate timeline for 2009
ERD Business meeting
Agenda
• Review objectives, desired outcome and
timeline of roadmapping exercise
• Present and discuss strawman structure
based on ITRS “potential solution format”
• Discuss table contents (Technology
Entries)
• Discuss next steps
Objectives
• Identify infrastructure requirements and
gaps to fabricate industrially relevant,
carbon based prototype devices with
timeline
• Identify key infrastructure requirements
neede to develop new functionalities of
carbon based electronics with timeline
Scope of roadmap discussion




Integrate with other known technology roadmaps to
achieve commercial viability
Identify critical infrastructure requirements to
fabricate industrially relevant prototypes
Identify existing infrastructure & infrastructure gaps
Decide roadmap format – e.g. potential solution format
–
–
Decide major technology entries
Determine approximate timelines
5 ERD WG 9/22-23/08
Work in Progress --- Not for Publication
Carbon-based Nanoelectronics Workshop
Agenda
9:30 Introduction
9:40 “Theory of electronic states and
transport in graphene and nanotube”
10:30 “Graphene conduction control by gate
voltage
11:20 “Epitaxial graphene on Si substrate
mediated by an ultra-thin SiC layer”
12:10 Lunch
13:00 “Evaluation of number of graphene
layers grown on SiC”
13:50 “Beyond-CMOS applications of
graphene based nanoelectronics”
14:40 Summary
15:00 Spin Torque Transfer RAM Workshop
6 ERD WG 9/22/08
Dr. Y. Awano (Fujitsu)
Prof. T. Ando (Tokyo Inst. Tech)
Dr. K. Tsukagoshi (AIST)
Prof. M. Suemitsu (Tohoku U.)
Dr. H. Hibino (NTT)
Prof. P. Kim (Columbia U.)
Dr. Y. Awano (Fujitsu)
Dr. U-In Chung (Samsung)
Work in Progress --- Not for Publication
Proposed roadmap format
• Build on ITRS “potential solution” format
• Separate into FET driven requirements
and novel device infrastructure
requirements
• Tie closely and directly to ERM
ERM table of applications vs TWIG
Table ERM2
MATERIALS
ERD MEMORY
Applications of Emerging Research Materials
ERD LOGIC
LITHOGRAPHY
FEP
INTERCONNECTS
Nanotube
Macromolecules
Nano-mechanical
Memory
Molecular memory
Nanowire
Graphene and
graphitic structures
Molecular devices
Self Assembled
Materials
Nanotubes
High-index
immersion liquids
Resists
Imprint polymers
Metal nanowires
Novel cleans
Selective etches
Low-κ ILD
Selective depositions
Sub- lithographic
patterns
Selective etch
Enhanced
dimensional control
Deterministic doping
Selective deposition
Electrical
applications
Thermal applications
Mechanical
applications
Polymer electrical
and thermal/
mechanical property
control
Selective etch
High performance
Selective deposition capacitors
Semiconductor spin
transport
Spin Materials
MRAM by spin
injection
Ferromagnetic (FM)
semiconductors
FM metals
Tunnel dielectrics
Passivation
dielectrics
Complex Metal
Oxides
Interfaces and
Heterointerfaces
1T Fe FET
Fuse-anti-fuse
Electrical and spin
contacts and
interfaces
Multiferroics (Spin
materials)
High performance
capacitors
Novel phase change
Electrical and spin
contacts and
interfaces
Contacts and
interfaces
Insert new column here??
Low Dimensional
Materials
ASSEMBLY AND
PACKAGE
Proposed structure - table 1
First Year available for Production
2009
2010
45nm
2011
2012
2013
32nm
2014
2015
2016
22nm
2017
2018
2019
16nm
2020
2021
Materials & Processes for Carbonbased Nanoelectronics
CNTs
density (pitch)
growth control
doping
contacting
alignment
variability
Graphene
Manufacturable deposition
patterning
edge effects
contacting
dielectrics
bilayers
variability
This legend indicates the time during which research, development, and qualification/pre-production should be taking place for the solution.
Research Required
Development Underway
Qualification / Pre-Production
Continuous Improvement
2022
11nm
2023
2024
Proposed structure –table 2
Carbon-based Nanoelectronics
Devics
Digital CNTFET
Analog CNTFET
digital GNRFET
Analog GNRFET
bilayer devices
Quantum interference devices
This legend indicates the time during which research, development, and qualification/pre-production should be taking place for the solution.
Research Required
Development Underway
Qualification / Pre-Production
Continuous Improvement
Discussion on table structure
• Recognizes need for improved samples
regardless of specific applications
• Recognizes that some applications will
drive special needs e.g. FETs
• Will this structure accomplish our
objective?
• Is there a better structure
Table contents
General Requirements
CNTs
density (pitch)
growth control
doping
contacting
alignment
variability
Graphene
Manufacturable deposition
patterning
edge effects
contacting
dielectrics
bilayers
variability
Device Specific Requirements
Digital CNTFET
Analog CNTFET
digital GNRFET
Analog GNRFET
bilayer devices
Quantum interference devices
Graphene SETs
Quantum Hall devices
Timeline
•
•
•
•
•
Circulate proposal to entire group
Merge all inputs in December
Solicit volunteers for specific topics
Schedule regular meetings in spring
Output in time for spring meeting
Content of 2009 ERD Logic
section
ITRS Logic Workshop
Tsukuba, Japan
George Bourianoff facilitating
Sept 23. 2008
Workshop and Business Meeting
Objectives (Sept. 22 – 23)

Layout roadmap for Carbon Nanotubes and Graphene for
“Ultimately Scaled CMOS” and “Beyond CMOS”. Build on
results from July ERD meetings.
–
–


Provide information needed to develop new Memory Table
entries for STT RAM (new TE for 2009).
Determine content for 2009 ERD logic section
–
–
–


Review Technology Entries (TEs) from 2007
Review potential TE adds/drops for 2009
Solicit writing volunteers for 2009
Discuss linkage to materials and architecture sections
–

Ultimate CMOS roadmap – potential solution (PIDS Role?)
Beyond CMOS roadmap entry – current ERD format?
How can we improve the integration? (e.g. joint workshops, key
materials properties table, …)
Approximate timeline for 2009
ERD Business meeting
15 ERD WG 9/22-23/08
Work in Progress --- Not for Publication
Objectives
•
•
•
•
•
Determine content for 2009 ERD logic section
Review Technology Entries (TEs) from 2007
Review potential TE adds/drops for 2009
Solicit writing volunteers for 2009
Discuss linkage to materials and architecture
sections
• How can we improve the integration? (e.g. joint
workshops, key materials properties table, …)
• Approximate timeline for 2009
• ERD Business meeting
High performance logic table 2007
Device
FET [A]
Typical example devices
Si CMOS
FET Extension
1D structures
Channel
replacement
CNT FET
III-V compound
semiconductor and
NW FET
Ge channel
NW heteroreplacement
structures
SET
Molecular
SET
Crossbar latch
Molecular
transistor
Molecular QCA
Ferromagnetic
logic
Moving domain
wall
M: QCA
Spin Gain
transistor
Spin FET
Nanoribbon
transistors with
graphene
Cell Size
(spatial
pitch) [B]
Density
2
(device/cm )
Spin transistor
Spin Torque
Transistor
Projected
100 nm
100 nm [D]
300 nm [I]
40 nm [O]
10 nm [U]
140 nm [Y]
100 nm [C]
Demonstrated
590 nm
~1.5 m [E]
1700 nm [J]
~200 nm [K, L]
~2 m [V]
250 nm [Z, AA]
100 m [AB]
Projected
Demonstrated
Projected
Demonstrated
Projected
Demonstrated
1E10
2.8E8
12 THz
1.5 THz
61 GHz
5.6 GHz
4.5E9
4E7
6.3 THz [F]
200 MHz [G]
61 GHz [C]
220 Hz [H]
6.1E9
3.5E7
>1 THz
>300 GHz
61 GHz [C]
Data not available
1E12
2E7
1 THz [W]
100 Hz [V]
1 GHz [U]
100 Hz [V]
5E9
1.6E9
1 GHz [Y]
30 Hz [Z, AA]
10 MHz [Y]
30 Hz [Z]
4.5E9
1E4
40 GHz [AC]
Not known
Not known
Not known
Projected
3E-18
3E-18
3.00E-18
5E-17 [X]
~1E-17 [Z]
3E-18
Demonstrated
1E-16
1E-11 [H]
1E-16 [J]
6E10
~2E9
10 THz [Q]
2 THz [R]
1 GHz [O]
1 MHz [P]
1×10–18 [O]
[>1.5×10–17 ] [S]
–17
8×10
[T]
3E-7 [V]
6E-18 [AA]
Not known
Projected
238
238
61
10
1000
5E-2
Not known
Demonstrated
1.6
1E-8
Data not available
2E-4
2E-9
5E-8
Not known
Operational Temperature
RT
RT
RT [M, N]
RT
RT
RT
Materials System
Si
RT
CNT,
Si, Ge, III-V,
In2O3, ZnO, TiO2,
SiC,
379
InGaAs, InAs,
InSb
III-V, Si, Ge,
Organic
molecules
Ferromagnetic
alloys
Si, III-V,
complex metals
oxides
62
91
244
32
122
Switch Speed
Circuit Speed
Switching
Energy, J
Binary
Throughput,
2
GBit/ns/cm
Research Activity [AD]
[>1.3×10
–14
] [S]
2007 Technology Entries – High
Performance
• Extensions to CMOS - Low dimensional structures previously
included Carbon Nanotube FETs, nanowire FETs, and Nanowire
heterostructures. This edition will also include nanoribbon devices
using graphene..
• 2. Extensions to CMOS - High mobility channel replacement FETs
including III-V and Ge channel replacement
• 3. Single electron devices discussion had similar scope to
previous editions
• 4. Molecular devices had similar scope to previous editions with
primary focus on molecule on CMOS architecture (CMOL) concept
• 5. Ferromagnetic logic devices are based on collective magnetic
effects associated with the magnetic polarity of a nanodomain.
• 6. Spin devices are based on spin dynamics of one or a few
electrons, defects, or nuclei
2007 Alternative Device Table
Resonant
Tunneling
Diodes
Multi-ferroic Tunnel
Junctions
Single Electron
Transistors
Molecular Devices
Ferro-Magnetic
Devices
Frequency
Coherent Spin
Devices
State
Variable
Charge
Dielectric and
magnetic domain
polarization
Charge
Molecular
conformation
Ferromagnetic
polarization
Precession
frequency
Response
Function
Negative
differential
resistance
Four resistive states
Staircase I/V from
Coulomb
blockade
Hysteretic
Nonlinear
Nonlinear
Class—
Example
Mobile
Multi-ferroic tunnel
junction
Voltage tunable
transfer function
CMOL, cross bar
latch
Amplifiers, buses,
switches
Spin torque
oscillator
Architecture
Heterogeneous
Morphic
Heterogeneous,
morphic
MQCA, morphic
Morphic
Application
Elements in hybrid
magneto electric
circuits
Analog pattern
matching
Associative
processing , NP
complete,
Elements in hybrid
magneto-electric
circuits
Microwave
power, tunable
rectifiers
Comments
Additional
functionality
Density,
functionality
Density, cost
functionality
Radiation hard,
environmental
rugged
RF functionality
Status
Demo
Demo
Demo
Demo
Simulation
Material
Issues
Stray charge
RT DMS
2007 Technology Entries
Alternative Information Processing
•
•
•
•
•
•
1.
2.
3.
4.
5.
6.
Resonant Tunneling Diodes
Multi-ferroic Tunnel Junctions
Single Electron Transistors
Molecular Devices
Ferro-Magnetic Devices
Frequency Coherent Spin Devices
Suggestions received
• Include a discussion of hybrid devices
– Not sure that makes sense at the device level
– Proposal – make the point in the intro that some applications
may benefit from a hybrid approach
• Include steep subthreshold devices
– They are in the transition table
– Should we keep them there?
• Include MQCA as a category
– It already is in the alternative device table – no action needed
• Nanocrystal flash devices for logic
– It already is in the alternative device table – no action needed
2007 Transition table
IN/OUT
Rapid Single Flux Quanta (RSFQ)
CMOS extension-III-V channel
replacements
Impact Ionization MOS
Nano Electro Mechanical Systems
(NEMS)
OUT
Reason for IN/OUT
RSFQ devices, systems and
circuits have been developed,
prototyped, and fabricated.
They could become an
important technology if the
correct market driver emerges
Comment
Design and fabrication lines
for RSFQ systems exist.
Cryogenic operation, cost and
material integration issues
limit application space
IN
Low bandgap, compound III-V
semiconductors can
potentially improve transistor
performance
Research on compound III-V
semiconductors on SI
substrates has increased
significantly over the last 2
years
Possible Future
Simulation results showing
very low sub threshold slopes
indicate potential for low
power operation
Reliability remains an issue
may be included in future
editions
Possible Future
Potential for ultra low leakage
device based on nano relay
operation
Issues associated with
stiction, speed, active power
and reliability are being
studied –may be included in
future editions
Lateral interband tunneling
transistor
Possible Future
Floating gate MOS devices
Possible Future
Potential to utilize gate
modulated interband
tunneling to reduce
subthreshold slope
Devices with nanocrystals
embedded in gate allow
circuits with tuneable
thresholds. Potential for low
power circuits
May be included in future
editions
May be included in future
editions
New topics for discussion
– Should we broaden the “high performance” table to
include Low Power” and “Low Standby Power”? (Call
it “High Volume” applications?)
– Pros
• Would align better with ITRS System Drivers
• Would reflect motivation of much research
• Several recent theoretical and simulation papers support this
– Con
• Would be orthogonal to the historical “tracking” function of
the Table
Technology Entries(1)
• FET extensions
– Low dimension Channel replacement
category
• Recommend keeping category perhaps
emphasizing carbon based components
• Discuss CNTFETs with PIDs
– High mobility channel replacements
• Send III-V and Ge to PIDs
• Graphene keep or talk to PIDs
SETs
• Recommend keeping but emphasize low
power application
– Pros
• A lot of recent work supports that. (Very recent
paper by Vishwa Ray at UT Arlington* shows RT
operation self aligned process on Si)
– Cons
• Past efforts have not been fruitful
• Stray charge will always be problem
Nature of Nanotechnology advance online
Molecular
• Recommend moving to transition table
and keeping in the alternate technology
table
• Pros
– Recent progress has been weak
• Cons
– Many people believe passionately that it still
has great potential
Ferromagnetic and spin transistor
• Recommend keeping categories but
merging them together
• Pros
– Line between the two entries is becoming
blurred as we heard yesterday
– Large amount of activity in area (driven
currently by memory including embedded)
• Cons
– ??
New Technology Entries to
consider
• ???
2007 Technology Entries
Alternative Information Processing
•
•
•
•
•
•
1.
2.
3.
4.
5.
6.
Resonant Tunneling Diodes
Multi-ferroic Tunnel Junctions
Single Electron Transistors
Molecular Devices
Ferro-Magnetic Devices
Frequency Coherent Spin Devices
Resonant Tunnel Diodes
• Recommend moving to transition table
• Pros
– Not much progress
• Cons
– It is an interesting device with NDR
– Many people feel passionately about it
Multiferroic tunnel junctions
• Recommendation – change the category to
multiferroic switching elements
• Pros
– Recent demonstration of RT mutiferroic properties in
BFO
– Several applications in low power spin wave circuits
– Main application would be 4 state logic which is not
attractive for other reasons
• Cons
– ???
Single Electron Transistors
• Recommendation- Keep category as it is
• Pros
– Potential Non Boolean logic applications such
as image recognition still receiving attention
– Still an active area
• Cons
– ???
Molecular devices
• Recommendation – Keep the category
• Pros
– Area of intense activity
– Driven by nano bio
– New applications being constantly proposed
• Cons
– ????
Ferromagnetic devices
• Recommendation – Keep the category
– Pros
• Area of intense activity
• Potential applications and devices outside of
Boolean Logic gates
• Include Graphene devices
– Cons
• ????
Coherent spin devices
• Recommendation – Keep the category
– Pros
• Significant activity taking place
• Pseuduospintronic devices on bilayer graphene
would go here
• Spin hall effect devices would go here (also on
graphene most likely)
– Cons
• Single spins at room temperature will never be
robust (?) and therfore should drop