a simulation analysis on defect annihilation in directed ... · a simulation analysis on defect...
TRANSCRIPT
1
A simulation analysis on defect annihilation
in directed self-assembly lithography
Katsuyoshi Kodera, Hideki Kanai, Yuriko Seino, Hironobu Sato, Yusuke Kasahara, Katsutoshi Kobayashi, Hiroshi Kubota, Naoko Kihara, Yoshiaki Kawamonzen, Shinya Minegishi, Ken Miyagi, Toshikatsu Tobana, Masayuki Shiraishi, Satoshi Nomura and Tsukasa Azuma
EUVL Infrastructure Development Center
A part of this work was funded by the New Energy and Industrial Technology Development Organization (NEDO) of Japan under the EIDEC project.
2
Original simple HP 15nm L/S patterning process using PS-b-PMMA
PS short and dislocation defects are observed at the wafer center and edge after SOG etch.
average (3σ )CD (nm) 16.9 (4 .6)LWR (nm) 4.8 (1 .3)LER (nm) 3.7 (1 .5)
85 chips / waferPS short defect
6 chips (7%)
Dislocation defect
8 chips (9%)
71 chips (84%)
500 nm
OK chip
Y. Seino et al., MNE 2014
3
Outline
Grid defects Experimental behavior Simulation results acquired using
self-consistent field theory(SCFT) Simulation results acquired using
dissipative particle dynamics (DPD)
Dislocation defects
Summary
4
Outline
Grid defects Experimental behavior Simulation results acquired using
self-consistent field theory(SCFT) Simulation results acquired using
dissipative particle dynamics (DPD)
Dislocation defects
Summary
5
Characteristic DSA Defect (Grid defect)
grid
defects
short
defects
short
defects L/S with
roughness
fine L/S L/S with
roughness
open
defects
Dry development or SOG RIE Small Large
Although most of these grid defects disappeared after SOG full etching, they
could result in the degradation of etching process margin, and therefore, in line
edge roughness and open/short defects
microphase
separation dry development (PMMA removed)
SOG etched shallowly SOG full etch
PMMA PS
SOG SOC
resist PS
Grid defect
* SOG: Spin On Glass,SOC: Spin On Carbon
Seino, MNE2014
Sato, MRS2014
Kasahara, SPIE2015.
• Guide:
Width=0.5𝐿0
Pitch=3𝐿0
6
Process condition dependence
• Grid defects can be decreased with optimum neutral layer and sufficient
phase separation annealing.
• The subtle adjustment of the neutral layer toward a more neutral condition
results in the decrease of the grid defects.
250℃ 2 min 250℃ 30 min 270℃ 30 min
* contact angle: 84°
200n
m
Annealing condition dependence
Neutral layer: contact angle dependence
200
nm
Contact angle: 80.5° Contact angle: 81.9°
Contact angle: 84.4°
PS affinity PMMA
affinity
Y. Seino, MNE2014
7
Simulation model
top(air)
bottom Pinning
film thickness=1.5𝐿0
pinning height=0.68𝐿0 pinning width=0.5𝐿0 pinning period=3𝐿0
periodic condition
1.7𝐿0
• SCF simulation, masking method • BCP : 𝜒𝑁 = 20 , 𝐿0 = 30 nm (PS-b-PMMA) • top: neutral (𝜒𝑃𝑀𝑀𝐴−𝑡𝑜𝑝 = 2, 𝜒𝑃𝑆−𝑡𝑜𝑝 = 2)
• pinning: PMMA attractive (𝜒𝑃𝑀𝑀𝐴−𝑃𝐼𝑁 = 1, 𝜒𝑃𝑆−𝑃𝐼𝑁 = 2) • bottom: PMMA attractive (𝜒𝑃𝑀𝑀𝐴−𝐵𝑇𝑀 = 1.3, 𝜒𝑃𝑆−𝐵𝑇𝑀 = 2)
8
Grid defect as metastable morphology Vertical Lamellar(V L) Grid Defect (GD) Mixed lamellar (ML)
Side view
Top-down
pinning pinning pinning
3 stable/metastable morphologies including grid defect were acquired using SCFT.
most stable
quasi-hexagonal buried structures
9
Stability of grid defects
𝜒𝑃𝑀𝑀𝐴−𝐵𝑇𝑀
Fre
e e
nerg
y Δ𝐹
-0.001
-0.0005
0
0.0005
0.001
0.0015
0 0.5 1 1.5 2 2.5
GD
ML
VL
Vertical lamellar state is the most stable
Mixed lamellar state is the most stable
PMMA attractive neutral
• These simulation results agree with the experimental behavior.
mixed lamellar
grid defect
vertical lamellar
10
Conformation of BCP chains
(a)PMMA chain is parallel to the bottom.
(b) PMMA chain is vertical to the bottom
𝑛0~𝑛5
𝑛0 > 𝑛5, 𝑛9
terminal
center
junction
𝑛0
𝑛5
𝑛9
Neutral layer(PMMA-attractive bottom)
PMMA
We can acquire the information of the polymer chain conformation by evaluating n0, n5 and n9 immediately above the bottom.
11
Conformation of BCP chain (GD-state)
0
0.2
0.4
0.6
0.8
1
0 0.5 1 1.5 2 2.5 3
n0 n5
n9 nPMMA
Densi
ty
Position (𝐿0) 𝑛0 > 𝑛5, 𝑛9 immediately above the bottom
𝑛0
𝑛5
𝑛9
PMMA
𝑛0
𝑛9
𝑛5
𝑛𝑀𝑀𝐴
12
Conformation of BCP chain (GD-state)
Parallel BCP chains
Vertical PMMA blocks
Parallel PMMA blocks
PMMA blocks are nearly vertical to the bottom in the region where the vertical PMMA lamellar patterns are connected toward the bottom. This characteristic conformation of BCP chains is considered to be related with the origin of the grid defects.
中性化膜に対して垂直 Vertical PMMA blocks
↑Top-down
13
Node density biased MC method
GD-state
terminal segment
junction segment
・・・・・
Field model particle model
Computer Physics Communications 145 267-279.
• The segment density distribution acquired using SCFT was transformed into the atomic representation using the node density biased MC method.
• Using this atomic representation as initial chain conformations, we executed DPD simulations and investigated the defect annihilation dynamics.
14
DPD simulation results
Firstly, grid defects were observed. And finally, the grid defects disappeared and the simulation converged into the equilibrium lamellar state.
15
Defect annihilation dynamics
Polymer chains are partially vertical.
Polymer chains are parallel
Both vertically and horizontally oriented polymer chains appeared randomly.
The polymer chains flipped down and flopped up randomly
(b)1000 steps (Transient state)
(c)10000 steps (Equilibrium lamellar)
(a)Initial state (GD state)
Vertically oriented
horizontally oriented
The defect annihilation dynamics of the grid defects could be understood as the change of the orientation of the BCP chains.
16
Outline
Grid defects Experimental behavior Simulation results acquired using
self-consistent field theory(SCFT) Simulation results acquired using
dissipative particle dynamics (DPD)
Dislocation defects
Summary
17
Polymer conformation in dislocation defects
𝑛𝑐𝑒𝑛𝑡𝑒𝑟 =Segment density distribution of the center node 𝑛𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑙 =Segment density distribution of the terminal node
Center node Terminal node
Single dislocation
Double dislocation
Perfect state
Self-consistent field theory
𝑛𝑐𝑒𝑛𝑡𝑒𝑟 − 𝑛𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑙
18
Polymer conformation in dislocation defects
Terminal segment density is larger than the center segment density.
Terminal segment density is larger than the center segment density.
Terminal segment is localized in a specific region. The delocalization of the terminal segments is effective for defect mitigation.
19
Summary
• We have investigated the conformations of
the polymer chains in “grid defects” and
“dislocation defects”.
• In both defective states, polymer chain
conformations were found to play an
important role in formation of the defective
states.
• For practical application, it is important to
understand the origin of the defective states
more. We need the corporation of the
simulation people more and more…!