Sunlin ISSCC 020705 foils
Innovation and Integration in
Innovation and Integration in
the
the
Nanoelectronics
Nanoelectronics
Era
Era
Sunlin Chou
Sunlin Chou
Technology and Manufacturing Group
Technology and Manufacturing Group
Intel Corporation
Intel Corporation
International Solid
International Solid--State Circuits ConferenceState Circuits Conference February 2005
(2)
Key Points
Key Points
y
(3)
“
“Reduced cost is one of the big Reduced cost is one of the big attractions of integrated
attractions of integrated
electronics, and the cost
electronics, and the cost
advantage continues to increase
advantage continues to increase
as the technology evolves
as the technology evolves
toward the production of larger
toward the production of larger
and larger circuit functions on a
and larger circuit functions on a
single semiconductor substrate.
single semiconductor substrate.”” Electronics, Volume 38,
Electronics, Volume 38,
Number 8, April 19, 1965
Number 8, April 19, 1965
1960
1960 19651965 19701970 19751975 19801980 19851985 19901990 19951995 20002000 20052005 20102010
Transistors
Transistors
Per Die
Per Die
10
1088
10
1077
10
1066
10
1055
10
1044
10
1033
10
1022
10
1011
10
1000
10
1099
10
101010
Moore
Moore
’
’
s Law
s Law
-
-
1965
1965
1965 Data (Moore)
1965 Data (Moore)
Source: Intel
(4)
Moore
Moore
’
’
s Law
s Law
-
-
2005
2005
4004 4004
8080
8080 80868086
80286
80286 386™386™ProcessorProcessor 486
486™™ProcessorProcessorPentium
Pentium®® ProcessorProcessor
Pentium
PentiumPentium®®II ProcessorII Processor
Pentium®®III ProcessorIII Processor
Pentium
PentiumItaniumItanium®®4 Processor4 Processor
™
™ProcessorProcessor
Transistors
Transistors
Per Die
Per Die
10
1088
10
1077
10
1066
10
1055
10
1044
10
1033
10
1022
10
1011
10
1000
10
1099
10
101010
8008 8008
Itanium
Itanium™™22 ProcessorProcessor
1K 1K 4K 4K 64K 64K 256K 256K 1M 1M 16M 16M 4M 4M 64M 64M 256M 256M512M512M
1G
1G 2G2G 128M
128M
16K 16K
1965 Data (Moore)
1965 Data (Moore)
Microprocessor
Microprocessor
Memory
Memory
1960
1960 19651965 19701970 19751975 19801980 19851985 19901990 19951995 20002000 20052005 20102010
Source: Intel
(5)
The IC Growth Cycle
The IC Growth Cycle
Density
Density
Performance
Performance
Switching energy
Switching energy
Cost per function
Cost per function
Applications
Applications
Devices
Devices
Users
Users
Revenue
Revenue
Consumption
Consumption
grows
grows
Technology
Technology
advances
(6)
Silicon Scaling Still Improves
Silicon Scaling Still Improves
Density, Performance, Power, Cost
Density, Performance, Power, Cost
130 nm 90 nm 130 nm 90 nm Madison
Madison MontecitoMontecito Cores/Threads
Cores/Threads 1/11/1 2/42/4
Transistors
Transistors 0.410.41 1.721.72 BillionBillion L3 Cache
L3 Cache 66 2424 MByteMByte
Frequency
Frequency 1.51.5 >1.7>1.7 GHzGHz Relative Performance
Relative Performance 11 >1.5x>1.5x Thermal Design Power
Thermal Design Power 130130 ~100~100 WattWatt
Source: Intel
(7)
Key Points
Key Points
y
y MooreMoore’’s Law thriving after 40 yearss Law thriving after 40 years
y
(8)
Convergence Drives Growth
Convergence Drives Growth
Convergence
Convergence
(9)
Convergence
Convergence
Wireless Mobile PC
Wireless Mobile PC
10
10
0
0
50
50
100
100
2003
2003
65
65
90
90 9696
2004
2004 20052005 20062006
% of Notebooks
% of Notebooks
that have Wireless
that have Wireless
Source: IDC
(10)
Source: Goldman Sachs and Co., McKinsey & Company; Dec. 2002
Bits of TrafficBits of Traffic
Convergence
Convergence
Data Phones and Traffic
Data Phones and Traffic
2004: Data Phones
2004: Data Phones
Cross
Cross--over Voiceover Voice
Source: WebFeet Research New Cell Phone Sales by Feature Set
New Cell Phone Sales by Feature Set
0
0
100
100
200
200
300
300
400
400
500
500
600
600
700
700
800
800
Million unitsMillion units
2009
2009
2002
2002 20032003 20042004 20052005 20062006 20072007 20082008
Data
Data
Voice
Voice
Data Traffic: 56% Annual
Data Traffic: 56% Annual
Growth thru 2006
Growth thru 2006
Data
Data
Voice
Voice
2001
(11)
Convergence
Convergence
Digital Consumer Electronics
Digital Consumer Electronics
The Great CrossoverThe Great Crossover
Digital Technologies Surpass Analog Digital Technologies Surpass Analog
Camera Sales
Camera Sales Millions
Millions
0 5 10 15 20 25
0 5 10 15 20 25
‘
‘9595 ‘‘9696 ‘‘9797 ‘‘9898 ‘‘9999 ‘‘0000 ‘‘0101 ‘‘0202 ‘‘0303 ‘‘0404 ‘‘0505
Portable CD Players
Portable CD Players DIGI
TAL
AN
ALOG
Satellites
Satellites
Cell Phones
Cell Phones
DVD Players
DVD Players
Camcorders
Camcorders
Cameras
Cameras
TVs
TVs
’
’9595 ’’9696 ’97’97 ’’9898 ’99’99 ’’0000 ’01’01 ’’0202 ’03’03 ’’0404 ’05’05 ’’0606 ’07’07 ’’0808
Crossover Year
Crossover Year
Sources: Consumer Electronics Association, Photo Marketing Association International, Photofinishing News, SBCA/Sky Trends
(12)
Key Points
Key Points
y
y MooreMoore’’s Law thriving after 40 yearss Law thriving after 40 years
y
y Convergence drives IC industry growthConvergence drives IC industry growth
y
y Integrated platforms optimize user experienceIntegrated platforms optimize user experience
y
(13)
Demand Growth Driven by
Demand Growth Driven by
Better User Experiences
Better User Experiences
Memory
MemoryComputing Computing
Graphics
Graphics
Networking
Networking
Wireless
Wireless
Multimedia
Multimedia
Mobility
Mobility
Compactness
Compactness
Battery Life
Battery Life
Virtualization
Virtualization
Security
Security
Multitasking
Multitasking
Ease of Use
Ease of Use
Manageability
Manageability
User
User
Experience
(14)
Math Math
Coprocessor
Coprocessor SSE2SSE2
Itanium Itanium Architecture Architecture Low Low power power Multi Core Multi Core Hyper
Hyper-
-Threading Threading Processor Processor Capabilities Capabilities Integrated Integrated Cache
Cache PCIPCI
AGP 2X AGP 2X USB USB AGP 4X AGP 4X IAPC
IAPCUSB 2.0USB 2.0 AGP 8X
AGP 8X
Serial ATA
Serial ATA HD AudioHD Audio
Infiniband
Infiniband
PCI
PCI--EE
Graphics Graphics Low Power Low Power Chipset Chipset Capabilities Capabilities User User Experience Experience Multiple Multiple Execution Units Execution Units MMX MMX SSE SSE Quickstart Quickstart Speedstep Speedstep Base Base Performance Performance Virtualization Virtualization ProSet ProSet WiFi
WiFi CommsCommsandand
Software Software Capabilities Capabilities 386 386 processor
processor Pentium®Pentiumprocessorprocessor® Pentium®Pentiumprocessorprocessor®IIII PentiumPentium®processorprocessor®IIIIII Pentium Pentium®®IIIIII
Xeon Xeon™™ processor processor Itanium Itanium™™ processor processor Pentium Pentium®®IIIIII
mobile mobile processor processor
Pentium Pentium®®4 4
processor processor Pentium Pentium®®4 4
Processor Processor
with HT with HT
Pentium Pentium®®4 4
mobile mobile processor processor Xeon Xeon™™ processor processor Itanium Itanium™™22
processor processor Math
Math
Coprocessor
Coprocessor SSE2SSE2
Itanium Itanium Architecture Architecture Low Low power power Multi Core Multi Core Hyper
Hyper-
-Threading Threading Processor Processor Capabilities Capabilities Integrated Integrated Cache
Cache PCIPCI
AGP 2X AGP 2X USB USB AGP 4X AGP 4X IAPC
IAPCUSB 2.0USB 2.0 AGP 8X
AGP 8X
Serial ATA
Serial ATA HD AudioHD Audio
Infiniband
Infiniband
PCI
PCI--EE
Graphics Graphics Low Power Low Power Chipset Chipset Capabilities Capabilities User User Experience Experience Multiple Multiple Execution Units Execution Units MMX MMX SSE SSE Quickstart Quickstart Speedstep Speedstep Base Base Performance Performance Virtualization Virtualization ProSet ProSet WiFi
WiFi CommsCommsandand
Software Software Capabilities Capabilities 386 386 processor
processor Pentium®Pentiumprocessorprocessor® Pentium®Pentiumprocessorprocessor®IIII PentiumPentium®processorprocessor®IIIIII Pentium Pentium®®IIIIII
Xeon Xeon™™ processor processor Itanium Itanium™™ processor processor Pentium Pentium®®IIIIII
mobile mobile processor processor
Pentium Pentium®®4 4
processor processor Pentium Pentium®®4 4
Processor Processor
with HT with HT
Pentium Pentium®®4 4
mobile mobile processor processor Xeon Xeon™™ processor processor Itanium Itanium™™22
processor processor
Platforms Optimize User Experience
Platforms Optimize User Experience
Innovate and Integrate
Innovate and Integrate
Example: Intel platforms Example: Intel platforms
(15)
RMS Applications Growing
RMS Applications Growing
Emerging workloads increase need for
Emerging workloads increase need for
high performance parallel processing
(16)
Performance and Power Efficiency
Performance and Power Efficiency
Increase with Parallel Architecture
Increase with Parallel Architecture
Relative
Relative
processor
processor
performance*
performance* (constant power
(constant power
envelope
envelope))
Source: Intel
Source: Intel
*Average of SPECInt2000 and SPECFP2000 rates for Intel desktop p
*Average of SPECInt2000 and SPECFP2000 rates for Intel desktop processors rocessors vs
vs initial Intelinitial Intel®®PentiumPentium®®4 Processor 4 Processor FORECAST
FORECAST
1 10 100
2000 2000
Single
Single--CoreCore
2008+ 2008+ 2004
2004
10X 10X
Dual / Multi
Dual / Multi--CoreCore
3X 3X
FORECAST
FORECAST
1 10 100
2000 2000
Single
Single--CoreCore
2008+ 2008+ 2004
2004
10X 10X
Dual / Multi
Dual / Multi--CoreCore
3X 3X
(17)
Multi
Multi
-
-
Core Parallelism Going Mainstream
Core Parallelism Going Mainstream
Dual
Dual
-
-
Core Processor Plans (Intel)
Core Processor Plans (Intel)
Servers
Servers DesktopDesktop MobileMobile
90nm 90nm Montecito Montecito
(2005) (2005)
90nm 90nm Smithfield Smithfield
(2005) (2005)
65nm 65nm Yonah Yonah (2005) (2005)
90nm 90nm
Server Processor Server Processor
(2006) (2006)
65nm 65nm
Desktop Processor Desktop Processor
(2006) (2006)
(18)
Key Points
Key Points
y
y MooreMoore’’s Law thriving after 40 yearss Law thriving after 40 years
y
y Convergence drives IC industry growthConvergence drives IC industry growth
y
y Integrated platforms optimize user experienceIntegrated platforms optimize user experience
y
y MultiMulti--core parallelism going mainstreamcore parallelism going mainstream
y
(19)
Silicon Technology Changes to
Silicon Technology Changes to
Increase Power Efficiency
Increase Power Efficiency
1960
1960
’
’
s: Bipolar
s: Bipolar
1970
1970
’
’
s: PMOS, NMOS
s: PMOS, NMOS
1980
1980
’
’
s: CMOS
s: CMOS
1990
1990
’
’
s: Voltage scaling (P = CV
s: Voltage scaling (P = CV
22f)
f)
2000
(20)
Power Efficient 90nm Transistors with
Power Efficient 90nm Transistors with
Strained Silicon
Strained Silicon
HighHigh Stress Stress
Film Film
NMOS
NMOS
SiGe
SiGe SiGeSiGe
PMOS
PMOS
Compressive channel strain Tensile channel strai Compressive channel strain Tensile channel strainn 30% drive current increase 10% drive current increa 30% drive current increase 10% drive current increasese
Innovate and integrate
Innovate and integrate
for cost effective production
for cost effective production
(21)
Strained Silicon Improves Transistor
Strained Silicon Improves Transistor
Performance and/or Reduces Leakage
Performance and/or Reduces Leakage
1 1 10 10 100 100 1000 1000
0.2
0.2 0.40.4 0.60.6 0.80.8 1.01.0 1.21.2 1.41.4 1.61.6
Transistor Drive Current
Transistor Drive Current ((mAmA/um) /um) Std
Std StrainStrain StdStd StrainStrain
+25% I
+25% IONON +10% I+10% IONON
0.04x I 0.04x I OFFOFF
0.20x I 0.20x I OFFOFF
PMOS
PMOS NMOSNMOS
Transistor Transistor Leakage Leakage Current Current (
(nAnA/um)/um)
(22)
Advances in Power Efficient Design
Advances in Power Efficient Design
Using Using prior prior design design techniques techniques
With new With new
power power reduction reduction techniques techniques
ISSCC 2005 P10.1 ISSCC 2005 P10.1 “
“The Implementation of The Implementation of a 2
a 2--core Multicore Multi--ThreadedThreaded Itanium
ItaniumTMTM Family Family
Processor Processor””
Power Power
(W) (W)
(23)
Circuit Techniques Reduce
Circuit Techniques Reduce
Source Drain Leakage
Source Drain Leakage
Body Bias
Body Bias Stack EffectStack Effect Sleep TransistorSleep Transistor
Equal Loading Equal Loading
+ + VeVe Vdd
Vdd VbpVbp
Vbn Vbn
-- VeVe
Logic Logic Block Block
Leakage Leakage Reduction
(24)
Sleep Transistor Reduces SRAM
Sleep Transistor Reduces SRAM
Leakage Power
Leakage Power
7070 MbitMbit SRAM leakage current mapSRAM leakage current map
V VDDDD
SRAM SRAM Cache Cache Block Block NMOS
NMOS Sleep Sleep Transistor
V VSSSS
Transistor
Without sleep transistor
Without sleep transistor With sleep transistorWith sleep transistor
Accessed block
Accessed block
>3x SRAM leakage reduction on inactive blocks
>3x SRAM leakage reduction on inactive blocks
(25)
Sleep Transistors Reduce ALU Leakage
Sleep Transistors Reduce ALU Leakage
Scan Scan FIFO FIFO Scan Scan out out Sleep Sleep ALU ALU Body bias Body bias Control Control
37X leakage reduction
37X leakage reduction
demonstrated on test chip
demonstrated on test chip
Scan
Scan
PMOS
PMOS underdriveunderdriveVVcccc
PMOS overdrive
PMOS overdrive VVssss
Virtual
Virtual VVcccc
Virtual
Virtual VVssss
Scan Scan capture capture control control Dynamic Dynamic ALU
ALU 3232
NMOS overdrive
NMOS overdrive VVcccc
NMOS
NMOS underdriveunderdriveVVssss
ALU core
ALU core
3
3--bit A/Dbit A/D ALU ALU Body Body Bias Bias PMOS PMOS Sleep Sleep Body Body Bias Bias V
Vssssexternalexternal V
Vcccc
external
external
Sleep transistor
Sleep transistor
and body bias
and body bias
control
control
Sleep transistors Sleep transistors
(26)
Key Points
Key Points
y
y MooreMoore’’s Law thriving after 40 yearss Law thriving after 40 years
y
y Convergence drives IC industry growthConvergence drives IC industry growth
y
y Integrated platforms optimize user experienceIntegrated platforms optimize user experience
y
y MultiMulti--core parallelism going mainstreamcore parallelism going mainstream
y
y Holistic solutions deliver power efficiencyHolistic solutions deliver power efficiency
y
y Nanotechnology will extend IC advancesNanotechnology will extend IC advances
y
(27)
Silicon Technology Reaches
Silicon Technology Reaches
Nanoscale
Nanoscale
Micron Micron 10000 10000 1000 1000 100 100 10 10 10 10 1 1 0.1 0.1 0.01 0.01 Nano
Nano- -meter meter Nanotechnology Nanotechnology (< 100nm) (< 100nm) 130nm 130nm 90nm 90nm 70nm 70nm 50nm 50nm Gate Length
Gate Length 65nm65nm
35nm
35nm
1970
1970 19801980 19901990 20002000 20102010 20202020
45nm 45nm 32nm 32nm 22nm 22nm 25nm 25nm 18nm 18nm 12nm 12nm 0.7X every 2 years
Nominal feature size
Nominal feature size
Source: Intel Micron Micron 10000 10000 1000 1000 100 100 10 10 10 10 1 1 0.1 0.1 0.01 0.01 Nano
Nano- -meter meter Nanotechnology Nanotechnology (< 100nm) (< 100nm) 130nm 130nm 90nm 90nm 70nm 70nm 50nm 50nm Gate Length
Gate Length 65nm65nm
35nm
35nm
1970
1970 19801980 19901990 20002000 20102010 20202020
45nm 45nm 32nm 32nm 22nm 22nm 25nm 25nm 18nm 18nm 12nm 12nm 0.7X every 2 years
Nominal feature size
Nominal feature size
(28)
Nanotechnology Hallmarks
Nanotechnology Hallmarks
(For
(For
Nanoelectronics
Nanoelectronics
)
)
y
y Structures measured in nanometersStructures measured in nanometers
–
– Less than 0.1Less than 0.1--micron (100nm)micron (100nm)
y
y New processes, materials, device structuresNew processes, materials, device structures
–
– Incrementally changing silicon technology baseIncrementally changing silicon technology base
y
y Materials manipulated on atomic scaleMaterials manipulated on atomic scale
–
– In one or more dimensionsIn one or more dimensions
y
y Increasing use of selfIncreasing use of self--assemblyassembly
–
– Using chemical properties to form structuresUsing chemical properties to form structures
Nanotechnology innovations will extend
Nanotechnology innovations will extend
silicon technology and Moore
(29)
Design Your Own Film with
Design Your Own Film with
Atomic
Atomic
Layer Deposition (ALD)
Layer Deposition (ALD)
Step 3
Step 3
Step 4
Step 4
Step 2
Step 2
A
A
Step 1
Step 1
A
A
A
A
B
B
A
A
B
B
Atomic level manipulation + Self
(30)
ALD Enables High
ALD Enables High
-
-
k Dielectric
k Dielectric
to Reduce Gate Leakage
to Reduce Gate Leakage
Silicon substrate Gate
3.0nm High-k
Silicon substrate 1.2nm SiO2
Gate
High
High--k vs. SiOk vs. SiO22 BenefitBenefit Gate capacitance
Gate capacitance 60% greater60% greater Faster transistorsFaster transistors Gate dielectric
Gate dielectric leakage
leakage > 100x reduction> 100x reduction LowerLower powerpower
Source: Intel
Process integration is the key challenge
(31)
Nanostructures for the Next Decade
Nanostructures for the Next Decade
(Transistor Research at Intel)
(Transistor Research at Intel)
NonNon--planarplanar Tri
Tri--GateGate
Architecture
Architecture
Si
Si DeviceDevice Miniaturization Miniaturization Carbon Carbon Nanotube Nanotube Transistor Transistor III
III--V DeviceV Device Research Research Gate Source Si body (b) Drain Source CNT Single-wall D = 1.4 nm
Gate (Pt) Drain
(Pd)
Lg = 75 nm
(d) Gate Drain Source Source (c) Multi epitaxial layers Gate Source Si body (b) Drain Gate Source Si body (b) Drain Source CNT Single-wall D = 1.4 nm
Gate (Pt) Drain
(Pd)
Lg = 75 nm
(d)
Source
CNT
Single-wall D = 1.4 nm
Gate (Pt) Drain
(Pd)
Lg = 75 nm
(d) Gate Drain Source Source Multi epitaxial layers Gate Drain Source Source Multi epitaxial layers (c) (a)
L = 10 nmg
Gate Gate Source Source Drain Drain Source: Intel
(32)
0.1 1 10 100
1 10 100 1000 10000
GATE LENGTH LG [ nm]
G
A
T
E
D
E
L
A
Y
C
V
/
I
[
p
s
]
Si MOSFETs CNT FETs I I I - V FETs
NMOS
Benchmarking
Benchmarking
Nanotransistor
Nanotransistor
Progress
Progress
Source: Intel compilation of published data
(33)
Lithography Must Break Through
Lithography Must Break Through
to Shorter Wavelength (EUV* @ 13.5nm)
to Shorter Wavelength (EUV* @ 13.5nm)
* Extreme Ultraviolet
* Extreme Ultraviolet
1000 1000
100 100
10 10
’
’8989 ’’9191 ’’9393 ’’9595 ’’9797 ’’9999 ’’0101 ’’0303 ’’0505 ’’0707 ’’0909 ’’1111
Feature size Feature size
EU V EU V
Lithography Lithography Wavelength Wavelength
193nm & extensions 193nm & extensions 248nm
248nm nm
nm
Gap
Gap
(34)
Lithography Must Break Through
Lithography Must Break Through
to Shorter Wavelength (EUV* @ 13.5nm)
to Shorter Wavelength (EUV* @ 13.5nm)
* Extreme Ultraviolet
* Extreme Ultraviolet
Source: Intel
1000 1000
100 100
10 10
’
’8989 ’’9191 ’’9393 ’’9595 ’’9797 ’’9999 ’’0101 ’’0303 ’’0505 ’’0707 ’’0909 ’’1111
Feature size Feature size
EU V EU V
Lithography Lithography Wavelength Wavelength
193nm & extensions 193nm & extensions 248nm
248nm nm
nm
Gap
Gap
EUV LLC
EUV LLC
Prototype EUV Tool Demonstrated Prototype EUV Tool Demonstrated
EUV LLC invested heavily to spur innovation
(35)
EUV Lithography in
EUV Lithography in
Commercial Development
Commercial Development
EUV Micro exposure tool (MET) EUV Micro exposure tool (MET)
EUV MET Image (8/04) EUV MET Image (8/04)
Integrated development in progress
Integrated development in progress
y
y Source power and lifetimeSource power and lifetime
y
y Defect free mask fabrication and handlingDefect free mask fabrication and handling
y
y Optics lifetimeOptics lifetime
y
y Resist performanceResist performance
(36)
EUV Source Power Increased
EUV Source Power Increased
0.1 0.1 1 1 10 10 100 100 1000 1000
Jul
Jul--0101 JanJan--0202 JulJul--0202 JanJan--0303 JulJul--0303 JanJan--0404 JulJul--0404 JanJan--0505 JulJul--0505 JanJan--0606
EUV Power at Intermediate Focus [W]
EUV Power at Intermediate Focus [W]
115 W Production Requirement
115 W Production Requirement
Exponential fit to data
Exponential fit to data
Average of reported data
Average of reported data
(SEMATECH Source Workshops)
(SEMATECH Source Workshops)
(37)
EUV Mask Blank Defects Reduced
EUV Mask Blank Defects Reduced
Results from SEMATECHResults from SEMATECH
Source: SEMATECH 0.001
0.01 0.1 1
Jan-04 Apr-04 Jul-04 Sep-04 Dec-04
P rocess ad ded d e fect den s it y @ 80nm sensi ti vi ty ( c m -2)
Only 1 added
Only 1 added
defect!
defect! Goal
Target for production @ 32nm Target for production @ 32nm
0.001 0.01 0.1 1
Jan-04 Apr-04 Jul-04 Sep-04 Dec-04
P rocess ad ded d e fect den s it y @ 80nm sensi ti vi ty ( c m -2)
Only 1 added
Only 1 added
defect!
defect! Goal
Target for production @ 32nm Target for production @ 32nm
(38)
Key Points
Key Points
y
y MooreMoore’’s Law thriving after 40 yearss Law thriving after 40 years
y
y Convergence drives IC industry growthConvergence drives IC industry growth
y
y Integrated platforms optimize user experienceIntegrated platforms optimize user experience
y
y MultiMulti--core parallelism going mainstreamcore parallelism going mainstream
y
y Holistic solutions deliver power efficiencyHolistic solutions deliver power efficiency
y
y Nanotechnology will extend IC advancesNanotechnology will extend IC advances
y
y Lithography innovations remain vitalLithography innovations remain vital
y
y MooreMoore’’s Law will outlive CMOSs Law will outlive CMOS
y
(39)
Moore
Moore
’
’
s Law Will Outlive CMOS
s Law Will Outlive CMOS
Transistors/Die
Transistors/Die
Bipolar Bipolar
1960 1970 1980 1990 2000 2010 2020 2030
CMOS CMOS
1µm 100nm 10nm
10µm
PMOS PMOS
10 1088
10 1077
10 1066
10 1055
10 1044
10 1033
10 1022
10 1011
10 1000
10 1099
10
101010 1965 Data (Moore)1965 Data (Moore)
Microprocessor
Microprocessor
Memory
Memory
10 101111
10 101212
10 101313
Voltage
Voltage
Scaling
Scaling ScalingPwrScalingPwrEffEff New New structuresstructuresNanoNano- -NMOS
NMOS
Kilo Xtor
Mega
Xtor GigaXtor TeraXtor
Beyond CMOS? Beyond CMOS? Spin based? Spin based? Molecular? Molecular? Other? Other?
(40)
Innovation and Integration
Innovation and Integration
Will Sustain Moore
Will Sustain Moore
’
’
s Law
s Law
Innovation
Innovation
Integration
Integration
Identify needs and
Identify needs and
create capabilities
create capabilities
that drive growth
that drive growth
Deliver platforms
Deliver platforms
to optimize user
to optimize user
experience
experience
Anticipate barriers
Anticipate barriers
and seek timely
and seek timely
breakthroughs
breakthroughs
Integrate new
Integrate new
materials, devices,
materials, devices,
processes
processes
Make strategic
Make strategic
technology
technology
transitions
transitions
Coordinate
Coordinate
strategic shifts
strategic shifts
across industry
(41)
Key Points
Key Points
y
y MooreMoore’’s Law thriving after 40 yearss Law thriving after 40 years
y
y Convergence drives IC industry growthConvergence drives IC industry growth
y
y Integrated platforms optimize user experienceIntegrated platforms optimize user experience
y
y MultiMulti--core parallelism going mainstreamcore parallelism going mainstream
y
y Holistic solutions deliver power efficiencyHolistic solutions deliver power efficiency
y
y Nanotechnology will extend IC advancesNanotechnology will extend IC advances
y
y Lithography innovations remain vitalLithography innovations remain vital
y
y MooreMoore’’s Law will outlive CMOSs Law will outlive CMOS
y
(1)
EUV Source Power Increased
EUV Source Power Increased
0.1
0.1
1
1
10
10
100
100
1000
1000
Jul
Jul--0101 JanJan--0202 JulJul--0202 JanJan--0303 JulJul--0303 JanJan--0404 JulJul--0404 JanJan--0505 JulJul--0505 JanJan--0606
EUV Power at Intermediate Focus [W]
EUV Power at Intermediate Focus [W]
115 W Production Requirement
115 W Production Requirement
Exponential fit to data
Exponential fit to data
Average of reported data
Average of reported data
(SEMATECH Source Workshops)
(SEMATECH Source Workshops)
(2)
EUV Mask Blank Defects Reduced
EUV Mask Blank Defects Reduced
Results from SEMATECH Results from SEMATECH
Source: SEMATECH
0.001 0.01 0.1 1
Jan-04 Apr-04 Jul-04 Sep-04 Dec-04
P rocess ad ded d e fect den s it y @ 80nm sensi ti vi ty ( c m -2)
Only 1 added
Only 1 added
defect!
defect! Goal
Target for production @ 32nm Target for production @ 32nm
0.001 0.01 0.1 1
Jan-04 Apr-04 Jul-04 Sep-04 Dec-04
P rocess ad ded d e fect den s it y @ 80nm sensi ti vi ty ( c m -2)
Only 1 added
Only 1 added
defect!
defect! Goal
Target for production @ 32nm Target for production @ 32nm
(3)
Key Points
Key Points
y
y MooreMoore’’s Law thriving after 40 yearss Law thriving after 40 years
y
y Convergence drives IC industry growthConvergence drives IC industry growth
y
y Integrated platforms optimize user experienceIntegrated platforms optimize user experience
y
y MultiMulti--core parallelism going mainstreamcore parallelism going mainstream
y
y Holistic solutions deliver power efficiencyHolistic solutions deliver power efficiency
y
y Nanotechnology will extend IC advancesNanotechnology will extend IC advances
y
y Lithography innovations remain vitalLithography innovations remain vital
y
y MooreMoore’’s Law will outlive CMOSs Law will outlive CMOS
y
(4)
Moore
Moore
’
’
s Law Will Outlive CMOS
s Law Will Outlive CMOS
Transistors/Die
Transistors/Die
Bipolar
Bipolar
1960 1970 1980 1990 2000 2010 2020 2030
CMOS
CMOS
1µm 100nm 10nm
10µm
PMOS
PMOS
10
1088
10
1077
10
1066
10
1055
10
1044
10
1033
10
1022
10
1011
10
1000
10
1099
10
101010 1965 Data (Moore)1965 Data (Moore)
Microprocessor
Microprocessor
Memory
Memory
10
101111
10
101212
10
101313
Voltage
Voltage
Scaling
Scaling ScalingPwrScalingPwrEffEff New New structuresstructuresNanoNano-
-NMOS
NMOS
Kilo Xtor
Mega
Xtor GigaXtor TeraXtor
Beyond CMOS? Beyond CMOS? Spin based? Spin based? Molecular? Molecular? Other? Other?
(5)
Innovation and Integration
Innovation and Integration
Will Sustain Moore
Will Sustain Moore
’
’
s Law
s Law
Innovation
Innovation
Integration
Integration
Identify needs and
Identify needs and
create capabilities
create capabilities
that drive growth
that drive growth
Deliver platforms
Deliver platforms
to optimize user
to optimize user
experience
experience
Anticipate barriers
Anticipate barriers
and seek timely
and seek timely
breakthroughs
breakthroughs
Integrate new
Integrate new
materials, devices,
materials, devices,
processes
processes
Make strategic
Make strategic
technology
technology
transitions
transitions
Coordinate
Coordinate
strategic shifts
strategic shifts
across industry
(6)
Key Points
Key Points
y
y MooreMoore’’s Law thriving after 40 yearss Law thriving after 40 years
y
y Convergence drives IC industry growthConvergence drives IC industry growth
y
y Integrated platforms optimize user experienceIntegrated platforms optimize user experience
y
y MultiMulti--core parallelism going mainstreamcore parallelism going mainstream
y
y Holistic solutions deliver power efficiencyHolistic solutions deliver power efficiency
y
y Nanotechnology will extend IC advancesNanotechnology will extend IC advances
y
y Lithography innovations remain vitalLithography innovations remain vital
y
y MooreMoore’’s Law will outlive CMOSs Law will outlive CMOS
y