Transcript Ragheb_CNTFET

Carbon Nanotube Field-Effect Transistors
(CNTFETs): Evolution and Applications
for Future Nanoscale ICs
Tamer Ragheb
ELEC 527 Presentation
Rice University
3/15/2007
Conventional Semiconductor
Microelectronics Will Come to an End
Conventional semiconductor
device scaling obstacles:
 Diffusion areas will no longer be
separated by a low doped
channel region
 Equivalent gate oxide thickness
will fall below the tunneling limit
 Lithography costs will increase
exponentially
Vertical
Scaling
Lateral Scaling
Solution:
 Find new technologies such as
molecular electronics and CNT
Hoenlein et al., Materials Science
and Engineering: C, 2003
2/37
Why Carbon Nanotubes (CNTs)?
CNTs exhibit remarkable electronic and mechanical
characteristics due to:
 Extraordinary strength of the carbon-carbon bond
 The small atomic diameter of the carbon atom
 The availability of free π-electrons in the graphitic
configuration
Hoenlein et al., Materials Science and Engineering: C, vol. 23, no. 8, pp. 663-669, 2003
3/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Most of the CNTFETs employ:
 Semiconductor Single-walled carbon nanotube (SWCNT)
as the channel
 The contacts of SWCNT are the source and drain regions
 A gate plate to control the conduction behavior of SWCNT
 Tans et al. reported the first CNTFET (1998)
 Used SWCNT as a channel
 Platinum (Pt) as contacts
 Silicon (Si) as a back-gate
Tans et al., Nature, vol. 393, pp. 49-52, 1998
Hoenlein et al., Materials Science and Engineering: C, 2003
4/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Tans at al.’s CNTFET exhibits p-type FET behavior
Tans et al. succeeded to modulate the conductivity
over more than 5 orders of magnitude
The problem was the thick oxide layer used
Tans et al., Nature, vol. 393, pp. 49-52, 1998
5/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Bachthold et al. replaced:
 The Si-back gate by a patterned Al-gate
 The thick SiO2 layer by a thin Al2O3 layer
 Platinum (Pt) contacts by gold (Au)
The gate biasing can change the
behavior from p-type to n-type
Bachthold at al. succeeded to
build different logic gates using
the p-type behavior
Bachthold et al., Science, vol. 294, pp. 49-52, 2001
n-type
FET
Enhancedmode p-type
FET
6/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Bachthold et al. simulated circuits:
Bachthold et al., Science, vol. 294, pp. 49-52, 2001
7/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Due to difficulty of back gate biasing, Wind et al.
proposed the first CNTFET with top gate
The top gate is divided into 4
gate segments
Each segment is individually
biased for more behavior control
Wind et al., Physical Review Letters, vol. 91, no. 5, 2003
8/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Top-gated CNTFETs allow:
 Local gate biasing at low voltage
 High speed switching
 High integration density
Yang et al. compared the performance of:
 Bottom-gate without top oxide
 Bottom-gate with top oxide
 Top-gate with top oxide
The top oxide used is TiO2 (high-k dielectric)
Yang et al., Applied Physical Letters, vol. 88, p. 113507, 2006
9/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Yang et al. proved that:
 Top gate reduces the hysteresis behavior of CNTFET
 Top gate reduces the needed gate voltage
Yang et al., Applied Physical Letters, vol. 88, p. 113507, 2006
10/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Derycke et al. proposed the first CMOS-like device
by producing n-type CNTFETs by:
 Annealing in a vacuum at 700K
 Doping with potassium (K)
Derycke et al. succeeded to build
the first CMOS-like inverter
Derycke et al., Nano Letters, vol. 1, no. 9, pp. 453-456, 2001
11/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
The inverter fabrication steps:
Derycke et al., Nano Letters, vol. 1, no. 9, pp. 453-456, 2001
12/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Javey et al. proposed converting p-type into n-type
by field manipulation
Javay et al. succeeded to build different logic gates
Javey et al., Nano Letters, vol. 2, no. 9, pp. 929-932, 2002
13/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Javey et al.’s circuits:
Javey et al., Nano Letters, vol. 2, no. 9, pp. 929-932, 2002
14/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Chen et al. proposed a complete integrated logic
circuit assembled on a single CNT
They controlled
the polarities of
the FETs by
using metals
with different
work functions
as the gates
Chen et al., Science, vol. 311, p. 1735, 2006
15/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Chen et al.’s circuit is a voltage controlled (Vdd) ring
oscillator
Vdd=0.5V
Chen et al., Science, vol. 311, p. 1735, 2006
Vdd=0.92V
16/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Hoenlein et al. proposed a vertical CNTFET
(VCNTFET), it consists of:
 1nm diameter 10nm long SWCNT
 A coaxial gate and a gate dielectric with 1nm thickness
Hoenlein et al., Materials Science and Engineering: C, vol. 23, no. 8, pp. 663-669, 2003
17/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
VCNTFET has the advantages of:
 Vertical growth in CNT is much easier and aligned than
horizontal growth
 3D connections can be used in the vertical configuration
Hoenlein et al., Materials Science and Engineering: C, vol. 23, no. 8, pp. 663-669, 2003
18/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
All the previous structures depend on semiconductor
SWCNT.
SWCNT available commercially contain about 3360% metallic CNTs.
For mass production and high yield, methods have to
be found to guarantee that CNTFETs use
semiconductor type SWCNTs.
Chen et al. and Na et al. proposed 2 different
methods to convert metallic CNTs into
semiconductor type.
Chen et al., Japanese Journal of Applied Physics, vol. 45, no. 4B, pp. 3680-3685, 2006
Na et al., Fullerenes, Nanotubes, and Carbon Nanostructures, vol. 14, pp. 141-149, 2006
19/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Chen et al. used plasma treatment to convert
metallic CNT to semiconductor type.
Chen et al., Japanese Journal of Applied Physics, vol. 45, no. 4B, pp. 3680-3685, 2006
20/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Na et al. used protein-coated nanoparticles in the
contact areas to convert metallic CNT to
semiconductor type.
Measured
values
Theoretically
Na et al., Fullerenes, Nanotubes, and Carbon Nanostructures, vol. 14, pp. 141-149, 2006
21/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Liang et al. proposed building CNTFET using a
double-walled CNT (DWCNT)
 The inner-shell is the gate due to its low conductance
 The outer-shell is the channel due to its high conductance
 It is easy to fabricate high-quality DWCNT
Pd contacts
In fabrication:
 Cover the outer-shell partially
by polymer-patterns
 The exposed part can be
etched by H2O or O2 plasma
at room temperature
Router=1.73nm
Rinner=1.39nm
Inter-shell
separation=0.34nm
Liang et al., Physica. E, low-dimentional systems and nanostructures, vol. 23, no. 1-2, pp. 232-236, 2004
22/37
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Liang et al.’s CNTFET simulation results:
Liang et al., Physica. E, low-dimentional systems and nanostructures, vol. 23, no. 1-2, pp. 232-236, 2004
23/37
CNTFET as Memory Devices
Cui et al. employed CNTFET charge storage
behavior to build a non-volatile memory
The memory device is stable to hold the data over a
period of at least 12 days in the ambient conditions
Cui et al., Applied Physics Letters, vol. 81, no. 17, pp. 3260-3262, 2002
24/37
CNTFET as Memory Devices
To avoid the probability of metallic CNT, Cui et al.
used two methods:
 Annealing (to heat at 335K for different periods)
 Controlled oxygen plasma treatment at room temperature
Cui et al., Applied Physics Letters, vol. 81, no. 17, pp. 3260-3262, 2002
25/37
CNTFET as Memory Devices
Lu et al. proposed a non-volatile flash memory
device using:
 CNTs as floating gates
 HfAlO as control/tunneling oxide
 Platinum (Pt) as top electrodes
CNT insertion enhances the memory behavior by
holes trapping
Lu et al., Applied Physics Letters, vol. 88, p. 113104, 2006
26/37
Short Channel CNTFET (Sub-20nm)
Seidel et al. proposed a fabrication method to obtain
CNTFET with sub-20nm long channels
Seidel et al., Nano Letters, vol. 5, no. 1, pp. 147-150, 2005
27/37
Single Electron CNTFET
Cui et al. fabricated single electron
CNTFET (quantum dot) with a length
of 10nm
The observed differential conductance
peaks are a clear signature of single
electron tunneling in the device
Cui et al., Nano Letters, vol. 2, no. 2, pp. 117-120, 2002
28/37
Electro-Chemical CNTFET
Shimotani et al. studied another kind of CNTFET,
which is electro-chemical CNTFET
In this transistor the gate is the electrolyte solution
Shimotani et al., Applied Physics Letters, vol. 88, p. 073104, 2006
29/37
CNTFET as a Chemical Sensor
CNTFETs are very sensitive devices to chemicals.
Zhang et al. studied the sensing mechanism of
CNTFET to NO2 and NH3
 CNT body is more sensitive to ammonia
 CNT contacts are more sensitive to NO2
Zhang et al., Applied Physics Letters, vol. 88, p. 123112, 2006
30/37
CNTFET in RF Circuits
Zhang et al. measured the RF performance of
CNTFETs
RF Measurement
circuitry
Measurement
results
Zhang et al., IEEE Electron Device Letters, vol. 27, no. 8, pp. 668-670, 2006
31/37
CNTFET in RF Circuits
Zhang et al. proposed an RF simple model for
CNTFET
Zhang et al., IEEE Electron Device Letters, vol. 27, no. 8, pp. 668-670, 2006
32/37
CNTFET in RF Circuits
Pesetski et al. employed CNTFET to build RF circuits
that can operate up to 23GHz
Pesetski et al., Applied Physics Letters, vol. 88, p. 113103, 2006
33/37
CNTFET Built on Insulator
Liu et al. succeeded to build a novel nanotube-oninsulator (NOI) CNTFET similar to silicon-oninsulator (SOI) technology
Liu et al., Nano Letters, vol. 6, no. 1, pp. 34-39, 2006
34/37
CNTFET Built on Insulator
Liu et al. built NOI transistors with:
 Top-gated
 Polymer-electrolyte-gated
Liu et al., Nano Letters, vol. 6, no. 1, pp. 34-39, 2006
35/37
Conclusions
CNT is a future replacement for semiconductor
based microelectronics
The evolution of CNTFET is discussed
Employing CNTFET in a lot of applications such as:
 Logic circuits
 Memories
 Chemical sensors
 RF circuits
Integrating CNT based interconnects with devices
can produce a complete future nanoscale ICs
36/37
References (in Order of Appearance)



















Hoenlein et al., Materials Science and Engineering: C, vol. 23, no. 8, pp. 663-669, 2003
Tans et al., Nature, vol. 393, pp. 49-52, 1998
Bachthold et al., Science, vol. 294, pp. 49-52, 2001
Wind et al., Physical Review Letters, vol. 91, no. 5, 2003
Yang et al., Applied Physical Letters, vol. 88, p. 113507, 2006
Derycke et al., Nano Letters, vol. 1, no. 9, pp. 453-456, 2001
Javey et al., Nano Letters, vol. 2, no. 9, pp. 929-932, 2002
Chen et al., Science, vol. 311, p. 1735, 2006
Chen et al., Japanese Journal of Applied Physics, vol. 45, no. 4B, pp. 3680-3685, 2006
Na et al., Fullerenes, Nanotubes, and Carbon Nanostructures, vol. 14, pp. 141-149, 2006
Liang et al., Physica. E, low-dimentional systems and nanostructures, vol. 23, no. 1-2,
pp. 232-236, 2004
Cui et al., Applied Physics Letters, vol. 81, no. 17, pp. 3260-3262, 2002
Lu et al., Applied Physics Letters, vol. 88, p. 113104, 2006
Seidel et al., Nano Letters, vol. 5, no. 1, pp. 147-150, 2005
Cui et al., Nano Letters, vol. 2, no. 2, pp. 117-120, 2002
Shimotani et al., Applied Physics Letters, vol. 88, p. 073104, 2006
Zhang et al., Applied Physics Letters, vol. 88, p. 123112, 2006
Pesetski et al., Applied Physics Letters, vol. 88, p. 113103, 2006
Liu et al., Nano Letters, vol. 6, no. 1, pp. 34-39, 2006
37/37
Thank You
Acknowledgments:
Prof. James M. Tour and Prof. Lin Zhong
Colleagues in RAND group
Colleagues in the ELEC 527 class
Carbon Nanotube Field-Effect
Transistors (CNTFETs): Evolution
Chen et al. used plasma treatment to convert
metallic CNT to semiconductor type.
Not usable CNTs
Chen et al., Japanese Journal of Applied Physics, vol. 45, no. 4B, pp. 3680-3685, 2006
39/37