-Status of Masks - Current Status and Issues for ArF Extension · -ArF will be extended more than...
Transcript of -Status of Masks - Current Status and Issues for ArF Extension · -ArF will be extended more than...
- Status of Masks -
Current Status and Issues for ArF Extension
Nobuhito Toyama, Technical Manager
Member of the eBeam Initiative
2
Outline
• Roadmap & Challenges
• Complex Photomasks
• Approaches
• Trials & Examples
• Further Investigation
• Acknowledgement
3
Roadmap & ChallengesScaling & Scaling
2010 2011 2012 2013 2014
40
30
50
DPL
EUV
Computational Litho.
20
ITRS Roadmap : keep scaling
DRAM HPlogic Metal HPFlash HP
Logic Gate Length [year]
[nm]
Rapidly approaching a brick wall
ITRS Roadmap 2010
4
Roadmap & Challenges
40
30
50
Gate Length at IEDM : CMOS Devices and Technology
2010 2011 2012 2013 2014
20
2008 2009
logic Metal HP
Logic Gate Length
Computational Litho.
DPL or EUV
[year]
[nm]
from IEDM trend
Computational litho opportunities growing
32nm CMOS : still mainstream at IEDM
5
Complex Photomask
Computational Mask Correction more & more complex
conventional early OPC current advanced
more complex : # of features
smaller features
curved features
Computational Litho.
** virtual pattern
6
Complex Photomask
Computational mask correction
complex photomask data
Linearity# of fractured
featuresData
Handing
Inspection
Repair
insufficient
fidelity long writing time
(increasing cost)lower
Tputreduced
capability
Issues
7
Approaches
b) Linearity : small features smaller
a) Complex pattern fracturing : too many features
c) Process bias inconsistency : process bias induced number doubling
Linearity error ideal shape
linearity impacted
issues, and more
** virtual pattern
8
** virtual pattern
Complex Photomaskdesign ideal writing pattern expected fracturing
# of writing features
Courtesy of
Luminescent Technologies.
without correction
- uncontrollable CD
- unstable linearity
- less Depth of Focus
9
Approachesconventional fracturing
optimized fracturing
Fewer featuresoptimal effect
MB-MDP*overlapped shots
reducing writing time
litho-check
litho-checkEB writing check
** virtual pattern
bias
intelligentbias
*MB-MDP : Model-Based
Mask Data Preparation
10
Approaches
simplifying assist features
conventional fracturingMB-MDP
overlapped shots
with circular shape
Fewer features
optimal effect
reducing writing time
** virtual pattern
11
Independent "correction", "MDP" doesn't work
Approaches
- Linearity-aware Correction- Mask Process Compensation
- Litho-aware Correction- MDP-aware Correction
- Bias-aware Correction- Specific bias
correction
MDP
design
writing
a) Linearity : small features smaller
b) Writing time : too many features
c) Process bias inconsistency : process bias induced number doubling
measurement
supposed environment
correcteddata
MDP : Mask Data Preparation
12
1/3-1/4 features
Conventional
Fracturing
MB-MDP
overlapped shots
Trials & Examples
Resist
SEM MFG:75K
Courtesy of
D2S, Inc.
overlapped fracturing
Overlapped fracturing reduces the number of features
13
Trials & Examples
Confirmed : No alignment-error, No dimension error
inspected with conventional
fracturing
written with MB-MDP
overlapped shots
inspection
inspection
14
Trials & Examples
defect
locationdata
clipping
shape
correction
SEM
rendering
Rendering Acquired Image
repair
Courtesy of
Luminescent Technologies.
Repaired-image-reconstruction including
complex ones available
15
Summary
- ArF will be extended more than ITRS expectations
- Photomask data is becoming more complex and intensive
- Successful trials are underway using overlapped shots
with MB-MDP
- Data correction & MDP need to be more closely linked
MDP : Mask Data Preparation
16
Probe further
- Practical process compensation approach
- Effective inspection & repair
- Linearity improvement as well
- Total load estimation
AcknowledgementLeo-san and Young-san : Luminescent Technologies;
Jan-san, Kuwano-san : D2S;
Ohara-san : Nippon Control System;
and DNP Japan related to this topic.