Upload
dinhkhanh
View
216
Download
0
Embed Size (px)
Citation preview
Raluca Tiron
DSA: Progress Toward Manufacturing Readiness
Senior scientist, CEA-Leti, France
3 LITHO PROGRAMS IN LETI
DSA
Imprint
ML2
@LETI
ONE FOCUS
DEVELOP LOW COST PATTERNING SOLUTIONS
Graphoepitaxy
for contactGraphoepitaxy
for L/S
Chemoepitaxy
for L/SFirst high chi
evaluation
PS-b-PMMA High -
22nm < L0 < 60nm L0 < 20nm
200 nm200 nm
Technological Flow:
Materials : PS-b-PMMA L0 = [25:55nm]
Neutral layer
High chi BCP L0 < 25nm
300 mm Process Line:Lithography for guding patternes– 193 i or e-beam
DSA dedicated track– Specific bake
– Solvent annealing
– PMMA removal step
Dedicated metrology– CD-SEM
– SP2
Integration:Compact and physical model
Shortloops with ST
Dedicated defectivity tools
DSA at Leti: A large panel of materials and process flows available
Balancing BCP Thickness over Template Density in Graphoepitaxy
All images are taken on the same processed wafer
DSA Planarization: Full control of BCP self-
assembly through pitch
DSA planarization : A full control of the template
density over the wafer
The POR used to implement statistical monitoring
CD mean 17 nm
CDU wafer 3s 1.3 nm
PE-3σ 1.3 nmNb of defects 0 def/cm2
Inspected area 0.01 mm2
Nb of inspected contacts 6 x 105
Champion wafer
SOC
Si-ARC
PS
Balancing BCP Thickness over Template Density in Graphoepitaxy
300 mm Process Stability Monitoring
BCP coating
Natural period L
200nm
HOY & placement error (PE)
CD & CDU
FINGERPRINT PATTERNED SURFACE
POR on SOKUDO DUO Track using C35 from Arkema
Fingerprint: Thickness (NL and BCP); BCP’s L0 and CD
Grapho: CD, CDU, PE, HOY
Inorganic template: a reworkable process
Inorganic templets added value
Rework of DSA litho step
Thermal stability
Better surface affinity control
SIARC/SOC
Organic template
Silicon Oxide
Inorganic template
Guide template:
CD guide = 40.5 nm
CDU-3σ guide = 4nm
DSA:
CD = 17.2nm
CDU-3σ = 1.3nm
Residue ~ 7nm
HOY = 100%
Rework : NO
Planar: OK
Guiding template:
CD guide = 40nm
CDU-3σ guide = 4.6nm
DSA:
CD = 17.6 nm
CDU-3σ = 1.4nm
Residue ~ 7nm
HOY = 100%
Rework: OK
Planar OK
Equivalent DSA performances were achieved with inorganic guiding template
Inorganic vs .organic template
residue
Embedded neutral layer: fingerprintTo accurately control the polymer residue: selective neutral layer (NL) on the template bottom side
Sacrificial layer : concept validation on bulk substrate
BC
P o
n s
acri
fici
alla
yer
BC
P a
fte
rsa
crif
icia
llay
er
rem
ova
l
NL
on
bo
tto
mP
S af
fin
e PS
PS
NL
The sacrificial layer is
successfully integrated in the
new integration flow
SiARC
SOC
Sacrificiallayer NL
Embedded neutral layer: templates
NL residue thickness: 8 nm (3 = 0.6nm)
NL thickness can be reduced by Mw (e.g. 3nm)
Uniformity & thickness control of the residue is achieved !!!
FIB-STEM (view in transmission throught the thin lamellae) Susbtrate
Metal
SOC
Si-ARC
Si-ARC
SOC
NL Sacrificial layer
SiARC
PS PS
Embedded neutral layer: DSA
PoR Embedded NL
CD mean 18nm 22nm
Guiding CD 40nm 50nm
Residue 7nm 8nm*
Residue - 3σ 3.9nm 0.6nm
Hole open yield 100% 100%
Comparative evaluation / PoR vs New integration
Uniformity of the residue is improved using
embedded NL
*Could be reduced by using thinner NL
F.DELACHAT et al., n°10144-23. | SPIE Advanced Lithography 2017
Embedded neutral layer/ Advanced surface affinity control
DefectivityDefectivity analysis
CD-SEM
DSA
pro
cess
ste
p
Detection limit = 10 nmInspected surface = 20 µm2
Guiding pattern
DSA
SEM REVIEW CPIDetection limit = 10 nmInspected surface = 0.01 mm²
Inspected CHs = 1200Defects = 0HOY = 100%Defect density = 0 cm-2
Inspected wafers = 50
Inspected CHs = 7 x 106
Defects = 0HOY = 100%Defect density = 0 cm-2
Inspected wafers = 5
Inspected CHs = 1200Defects = 0HOY = 100%Defect density = 0 cm-2
Inspected wafers = 100
Inspected CHs = 7 x 106
Defects = 0HOY = 100%Defect density = 0 cm-2
Inspected wafers = 3
F. Delachat et. al., DSA Symposium 2016
Defectivity: no more a blocking point for DSA integration
DSA for contact patterning
Addapted from H. Yi,et al., Nano. Letters, 2015
Elliptical templatePeanuts shapeContact shrink
Template by 193iPS-b-PMMA BCP (L0 = 35 or 46nm)
200nm
Pitch = 60nm
200nm
Pitch = 46nm
Fast material & processevaluation
DSA Pitch modulation
by 193i pattern engineering
Pitch densityimprovement
by BCP natural period
Balancing BCP Thickness over Template Density in Graphoepitaxy
Full control of BCP self-assembly through pitch All images are taken on the same processed wafer
Summary
DSA has moved beyond the initial excitement of a new discovery.
DSA can easily meet resolution requirements down to N5.
Defectivity is still the main challenge, but the industry is making steady, order-of-magnitude progress each year.
DSA is on track to be adopted in manufacturing within two to five years.
DSA integration of lamellae-forming PS-b-PMMA (L0=38nm)
Targeted application in Leti:
“tri-gate nanowires”
Intel's 22-nm transistors, Nature 481,
152–153,12 January 2012
Template litho
TrenchDSA
assemblyPMMA etching SiO2 etching DSA - copolymer
removal
25nm TBOX
20nm LG ISPD SiCRSD
Si channel
Planar transistor Trigate Nanowire Stacked nanowires
2D channel 3D channel Multi 3D channels
CMOS architecture evolution
DSA: PS Affine Process
Claveau et al., Journal of Micro/Nanolith, MEMS and MOEMS. 15(3), 031604 (Aug 25, 2016).
1. 193-d L/S pattern
2. PS homopolymergrafting
3. PS-b-PMMA coating and self-assembly (250°C for 5min)
SiARCSOCSiO2SiNSi
BCP @ 120nm
Periodic like variations of residual PS bottom layer
UV expo to selectively tune surface affinity
UV-assisted graphoepitaxy process:
neutral bottom / PMMA sidewall affinity
Post DSA: CDU = 1.4nm / LWR = 2.2nm /
LER = 3.8nm
G.Claveau et al., SPIE 2017, 10144-36,
UV expo improve patterns after etching
Sta
nd
ard
pro
cess
Bridge
at bottom
SiO
2 e
tch
ing
Dry
PM
MA
re
mo
va
l
Str
ip/
lith
o/e
tch
cu
t
UV
assis
ted
pro
cess
Si
etc
hin
g
UV exposure avoid pattern collapse and reduce bridging defects
G.Claveau et al. SPIE 2017, 10144-36,
Process overview
CUT demonstrated on Si nanowire level
Guiding Pattern vs DSA Line Edge Roughness Characterization
Guiding Pattern vs DSA LER Characterization with PSD
Summary
UV-assisted graphoepitaxy approach for a precise control of template affinity.
300mm compatible silicon nanowires patterning with DSA process (litho, etch, cut).
Next:
Electrical performances meas. and comparison to double patterning.
DSA patterning for stacked nanowire devices.
Investigation of BCPs with lower resolution. SiN HMSiSiGeSiSiO2 (SOI BOX)
Si etched~ 45nmCD lam~ 11nmSiO2rest~ 3nm (N1) / 16,5nm (zone)
300mm Chemoepitaxy Process
200 nm 200 nm 200 nm 200 nm
DMF = 2 DMF = 3 DMF = 4 DMF = 5
PS-b-PMMA Nanostrength® EO L28 blends, from Arkema
DMF: Density multiplication factorSEM images w/a removing PMMA
Blends enable to achieve higher multiplication factors L/S patterns with low defectivity.
L. Evangelio Araujo et al. Proc of SPIE 2017, 10146-71
Ebeam cut on DSA patterning
E-beam exposure induces selective PMMA removal
E-beam cut on DSA:DSA only:
G.Claveau et. al., DSA Symposium 2016
20nm x 60nm CUT
DSA technology: from PS-b-PMMA to High-
Nodes
L0 (nm)
N10 N7 N3
FIN Pitch
N14
30nm48nm 36nm -[193i + SADP] [193i + SAQP] [193i + SAOP] or [EUV]
PS-b-PMMA High-
38
30
24
18
24nm
N5
DSA of High- materials: a real competitive
patterning option for sub-7nm nodes
100nm100nm
100nm 100nm
Fin
gerp
rin
tp
erf
orm
ance
sG
rap
ho
ep
itax
yp
erf
orm
ance
s
Standard PS-b-PMMA Organic High-χ
L0 (nm) 22 18
CD (nm)
L/S12.0 / 9.0 9.6 / 8.2
LWR (nm)
L/S2.7 / 1.6 1.5 / 1.1
CD (nm)
L/S11.8 / 9.4 11.5 / 6.0
LWR (nm)
L/S2.0 / 1.1 2.0 / 1.22
Organic High-χ implemented on 300mm pilot line, compatible with the standard PS-b-PMMA process
Silicon free High-
A.Gharbi et al., 10146-31 , Proc. of SPIE 2017
Silicon containing high : PS-b-PDMSB
Detailed lithographic performances under investigation on large areas.Next step: process transfer from samples to the track.
Aft
erp
lan
ari
zati
on
& P
S re
mo
val
Etch
ing
into
Si ~
80
nm
Bio sourced High
Djamila Ouhab PhD
a//=13.9nm a ┴ = 12.5nm
-11
20
qz
(nm
-1)
qy (nm-1)
GISAXS
DSA at LetiPS-b-PMMA High -
22nm < L0 < 60nm L0 < 20nm
Graphoepitaxyfor contact
Shrink and doubling Process stability
monitoring Integration
Graphoepitaxyfor L/S
Benchmark materials Prepare metrology
for high chi
200 nm
Chemoepitaxyfor L/S
Benchmark chemo vs grapho
Prepare high chi
First high chi evaluation
Next generationmaterial and processes
200 nm
A large panel of materials and process flows available.
Acknowledgements
Maxime ArgoudSandra BosShayma BouananiZdenek ChalupaGaelle Chamiot-MaitralGuillaume ClaveauFlorian DelachatAhmed GharbiJerome HazartMayssa Al KharboutlyCeline LapeyreLaurent PainAnne PaquetPatricia Pimenta-BarrosJonathan PradellesIsabelle Servin
Masami AsaiIan CayrefourcqXavier ChevalierLaura Evangelio AraujoMarta Fernandez RegulezGuillaume FleuryDouglas GuerreroMasahiko HarumotoChristophe NavarroCelia NicoletRemi LetiecAntoine LegrainFrancesc Perez MuranoKaumba SakavuyiHarold StokesMarc Zelsmann
YOU ARE WARMLY INVITED TO ATTEND
We hope you will help us do this by sponsoring the 4th International Symposium on DSA.On the behalf of program chairs: Tsukasa Azuma, Geert Vandenburghe, Raluca Tiron, and Joe Kline
Organized by: in collaboration with:
http://dsasymp.org/
| 34
INTEGRATION CONCERNS
POTENTIAL SHOWSTOPPER FOR DSA
P.Pimenta Barros et al. Proc of SPIE 2014, 9054-15
BCP performances depends on guiding patterns densityeMRS | 14th of May 2015 | Raluca Tiron et al
| 35
0
20
40
60
80
100
120
100 200 300 400 500
Ho
le O
pen
Yie
ld (
%)
pitchguiding (nm)
old process
0,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
1,6
1,8
100 200 300 400 500
No
rmal
ize
dB
CP
th
ickn
ess
pitchguiding (nm)
old process
« OLD PROCESS »
Hole Open Yield = f (pitch)BCP thickness= f (pitch)
(a) pitch = 110nm (b) pitch = 250nm (c) pitch = 450nm
(a)
(a)
(b)
(b)
(c)
(c)
old process old process old process
Defectivity over template density induced by BCP film thickness variation
eMRS | 14th of May 2015 | Raluca Tiron et al
| 36
A NEW METHOD TO CONTROL BCP FILM THICKNESS OVER PITCH
DSA PLANARIZATION PROCESS
1/ Guiding pattern 2/ Gap filling with BCP 3/ BCP bake
4/ Etch back5/ Etching transfer
patent pending
eMRS | 14th of May 2015 | Raluca Tiron et al
P.Pimenta-Barros et al, Proc. of SPIE 2015 , 9428-12
| 37
« NEW PROCESS » VS. « OLD PROCESS »
0,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
1,6
1,8
100 200 300 400 500
No
rmal
ize
dB
CP
th
ickn
ess
pitchguiding (nm)
old process
Hole Open Yield = f (pitch)BCP thickness= f (pitch)
(a) pitch = 110nm (b) pitch = 250nm (c) pitch = 450nm
(a)
(a)
(b)
(b)
(c)
(c)
old process old process old process
Old process filling guiding patternes directly with final film thickness
eMRS | 14th of May 2015 | Raluca Tiron et al
Hole Open Yield = f (pitch)BCP thickness= f (pitch)
(a) pitch = 110nm (b) pitch = 250nm (c) pitch = 450nm
0,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
1,6
1,8
100 200 300 400 500
No
rmal
ize
dB
CP
th
ickn
ess
pitchguiding (nm)
old process
new process
(a)
(b)
(c) (a)
(b)
(c)
old process old process old processnew process new process new process
Old process filling guiding patternes directly with final film thickness
New process: using DSA planarization