CEA-LETI - NIKON Joint Development Program ... - Nikon Precision
CEA-LETI - NIKON Joint Development Program ... - Nikon Precision
CEA-LETI - NIKON Joint Development Program ... - Nikon Precision
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>CEA</strong>-<strong>LETI</strong> - <strong>NIKON</strong> <strong>Joint</strong><br />
<strong>Development</strong> <strong>Program</strong> Update<br />
C. Lapeyre, S. Barnola, I. Servin,<br />
S. Gaugiran, S. Tedesco,<br />
L. Pain, A.J. Hazelton,<br />
V. Salvetat, M. McCallum
Outline<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec background<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec & <strong>NIKON</strong> <strong>Joint</strong> <strong>Development</strong> <strong>Program</strong><br />
• Double Patterning Process development and control<br />
• Topography impact on DP technology<br />
• Reticle & Pellicle impact on CDU & overlay<br />
• Experimental <strong>Nikon</strong> error model validation<br />
• DP processing to achieve k 1 = 0.14
<strong>LETI</strong> Background<br />
• 1,600 people on research programs (1000 <strong>CEA</strong> headcount)<br />
• 8000m 2 clean room<br />
• 160 industrial partners<br />
• 30 startups<br />
• 1600 patents portfolio<br />
• 250 patents in 2008<br />
3
<strong>LETI</strong> Background<br />
• 1,600 people on research programs (1000 <strong>CEA</strong> headcount)<br />
• 8000m 2 clean room<br />
• 160 industrial partners<br />
• 30 startups<br />
• 1600 patents portfolio<br />
• 250 patents in 2008<br />
Nanoelectronic 200mm and 300mm<br />
•200mm clean room (2500 m 2 )<br />
•300mm ball room (1500 m 2 )<br />
• State of the art tools<br />
• Full research line<br />
• Continuous operation<br />
7 days per week/ 24 hours per day<br />
4
<strong>Nikon</strong> and <strong>LETI</strong> <strong>Program</strong><br />
• <strong>Nikon</strong> and <strong>CEA</strong>-<strong>LETI</strong>-Minatec are partnering through a<br />
<strong>Joint</strong> <strong>Development</strong> <strong>Program</strong> to understand the sensitivities<br />
of various DE/DP techniques since early 2007.<br />
• Placing special emphasis on ensuring next generation<br />
exposure tools satisfy DP requirements.<br />
• Using a 0.85 NA dry ArF <strong>Nikon</strong> scanner in <strong>LETI</strong>’s Nanotec<br />
300 research facility in Grenoble, France.<br />
NSR-S307E – 0.85 NA Dry ArF
Outline<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec background<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec & <strong>NIKON</strong> <strong>Joint</strong> <strong>Development</strong> <strong>Program</strong><br />
• Double Patterning Process development and control<br />
• Topography impact on DP technology<br />
• Reticle & Pellicle impact on CDU & overlay<br />
• Experimental <strong>Nikon</strong> error model validation<br />
• DP processing to achieve k 1 = 0.14
45nm Double Patterning Process<br />
• NSR-S307E 0.85 NA ArF<br />
• Sokudo RF3<br />
• Lam Versys<br />
• 45 nm L/S pattern<br />
(k 1<br />
=0.20)<br />
Stack<br />
Resist ArF<br />
177nm<br />
Organic Barc<br />
32 nm<br />
HM1: Nitride<br />
30 nm<br />
HM2:Oxide<br />
20 nm
DP Process Control (1/3)<br />
• Integration stack requirements :<br />
– Good selectivity HM1/HM2<br />
– Reduce thickness to minimize<br />
topography<br />
– Low temperature deposition to avoid<br />
carbon mask damage<br />
• Etch Process optimization :<br />
– Ensure good resist budget margin to<br />
avoid faceting<br />
– Control of resist trimming<br />
– Etching chemistry selectivity<br />
BARC/HM1/HM2<br />
– Conservation of CD and profile of<br />
pattern 1 through etch2<br />
Line 2 :<br />
Oxide<br />
Line 1 : Nitride<br />
on Oxide<br />
Good oxide/nitride selectivity<br />
Good CD control with trimming
DP Process Control (2/3)<br />
70<br />
Impact of each step on CD mean<br />
60<br />
50<br />
66.0<br />
64.0<br />
CD (nm)<br />
40<br />
30<br />
48.0<br />
-3nm<br />
44.5 46.5<br />
20<br />
10<br />
0<br />
litho1 etch1 (line1) litho2 etch2 (line1) etch2 (line2)<br />
CD1 is well maintained through etch2 : CD bias ~-3nm<br />
Etch process very robust : CD1 and CD2 controlled independently
DP Process Control (3/3)<br />
4.5<br />
ITRS 45nm requirement<br />
CD uniformity 3σ (nm)<br />
4.0<br />
3.5<br />
3.0<br />
2.5<br />
2.0<br />
1.5<br />
1.0<br />
ITRS 32nm requirement<br />
2.6<br />
3.0<br />
3.3 3.2<br />
3.4<br />
0.5<br />
0.0<br />
litho 1 etch1 litho2 etch2 (line1) etch2 (line2)<br />
CD uniformity remains around 3nm throughout the DP<br />
process in agreement with ITRS 32nm requirements
Outline<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec background<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec & <strong>NIKON</strong> <strong>Joint</strong> <strong>Development</strong> <strong>Program</strong><br />
• Double Patterning Process development and control<br />
• Topography impact on DP technology<br />
• Reticle & Pellicle impact on CDU & overlay<br />
• Experimental <strong>Nikon</strong> error model validation<br />
• DP processing to achieve k 1 = 0.14
Topography Impact on Litho (1/3)<br />
180nm<br />
180nm<br />
65nm<br />
Si<br />
45nm<br />
Si<br />
Si<br />
Resist, 177nm<br />
Barc, 32nm<br />
15nm<br />
30nm<br />
45nm<br />
180nm 65nm Resist, 177nm<br />
Barc, 32nm<br />
Litho1<br />
Etch1<br />
Litho2<br />
•Substrate topography BARC thickness<br />
variation on features<br />
Impact on reflectivity control<br />
Impact on line1 etch2 dependent upon<br />
BARC/nitride selectivity<br />
12.3nm<br />
7.7nm<br />
2.6nm<br />
15nm height 30nm height 45nm height
Topography Impact on Litho (2/3)<br />
15nm height<br />
30nm height<br />
45nm height<br />
High topography alters resist profiles
Topography Impact on Litho (3/3)<br />
69<br />
10<br />
CD mean (nm)<br />
68.5<br />
68<br />
67.5<br />
67<br />
66.5<br />
66<br />
65.5<br />
+1.1nm<br />
+1.3nm<br />
9<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
CD uniformity 3σ (nm)<br />
65<br />
0<br />
0 15nm 30nm 45nm<br />
substrate topography<br />
Higher topography induces: Larger CD and larger dispersion<br />
Difference between 3σ(L2) and 3σ(L1)<br />
Due to topography, DP requires :<br />
<strong>Development</strong> of planarizing BARC to improve DP process<br />
Thinnest stack as possible : 30nm Si 3 N 4 / 20nm SiO 2
Outline<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec background<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec & <strong>NIKON</strong> <strong>Joint</strong> <strong>Development</strong> <strong>Program</strong><br />
• Double Patterning Process development and control<br />
• Topography impact on DP technology<br />
• Reticle & Pellicle impact on CDU & overlay<br />
• Experimental <strong>Nikon</strong> error model validation<br />
• DP processing to achieve k 1 = 0.14
Reticle Impact on CD and Overlay<br />
• 6% att-PSM in bright field tonality, no pellicle<br />
• Quartz flatness 0.5µm (T)<br />
• Designed with e-beam writer E5000<br />
• Mask : CD tool : NANOSEM3, Registration : IPRO4 (VISTEC)<br />
• Wafer : CD Hitachi CG4000, Overlay : CD-SEM & KLA<br />
32 nm reticle specifications<br />
• CD 0.5nm<br />
• OVL 0.9nm<br />
Reticle data fits ITRS<br />
Mask CDU < ¼ of wafer CDU<br />
Mask OVL < 10% of wafer OVL<br />
budget
Pellicle Induced Reticle Deformation<br />
In collaboration with Toppan Photomasks<br />
- Mechanical performance of pellicle may introduce pattern distortions<br />
- Image placement has been registered before and after pellicle mounting<br />
on same relevant structures spread out over the layout (10µm)<br />
Mapping measurements<br />
-Mask registration measured in X and Y axis<br />
-Distortions after isotropic and orthotropic corrections<br />
-Orthotropic corrected by exposure tool<br />
Pellicle distortions 3σ (nm) after corrections<br />
• X axis 3.2 nm<br />
• Y axis 2.4 nm<br />
Pellicle induced distortion represents 15% additional error in the<br />
image placement budget for the 45nm node
Outline<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec background<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec & <strong>NIKON</strong> <strong>Joint</strong> <strong>Development</strong> <strong>Program</strong><br />
• Double Patterning Process development and control<br />
• Topography impact on DP technology<br />
• Reticle & Pellicle impact on CDU & overlay<br />
• Experimental <strong>Nikon</strong> error model validation<br />
• DP processing to achieve k 1 = 0.14
<strong>NIKON</strong> Error Model Validation<br />
<strong>Nikon</strong> developed a theoretical model to establish the required CD and<br />
overlay requirements for DP that could be used in tool design. This was<br />
targeted for experimental verification at Leti.<br />
[A.J. Hazelton et. al, LithoVision 2008]
L & S Error Model Validation (1/3)<br />
•Standard LELE process : 19 wafers with programmed x-trans offsets<br />
•CD measurements : line1, line2, space1 and space2, and 3σ<br />
•Overlay measurements<br />
CD & SPACE<br />
70<br />
65<br />
CD1<br />
CD2<br />
X mean raw data (nm)<br />
14<br />
10<br />
6<br />
2<br />
-2<br />
OVERLAY DATA<br />
CD-SEM data (nm)<br />
60<br />
55<br />
50<br />
45<br />
40<br />
35<br />
30<br />
SP1<br />
SP2<br />
-8 -6 -4 -2 0 2 4 6 8 10<br />
programmed X-translation offsets (nm)<br />
-6<br />
-8 -6 -4 -2 0 2 4 6 8 10<br />
programmed X-translation offsets (nm)
L & S Error Model Validation (2/3)<br />
∆<br />
Global dispersion<br />
on the two line<br />
populations<br />
⎡3<br />
⎢<br />
⎣2<br />
( )<br />
⎤<br />
2<br />
L − L<br />
2<br />
2<br />
CD Line<br />
= ∆CD<br />
+<br />
1 2<br />
Dispersion mean<br />
of L1 and L2<br />
25<br />
⎥<br />
⎦<br />
∆CD<br />
= ∆CD 1<br />
= ∆CD<br />
Difference of<br />
line CDs<br />
20<br />
• Line error model is validated<br />
experimentally by varying both<br />
CD line and CD space<br />
∆CDline<br />
15<br />
10<br />
5<br />
y = 0.9974x + 0.0135<br />
R 2 = 1<br />
0<br />
0 5 10 15 20 25<br />
root sqr[ ∆CD² + [3/2(L1-L2)]² ]
∆CD<br />
space<br />
=<br />
Global dispersion on<br />
the two space<br />
populations<br />
∆CDspace<br />
L & S Error Model Validation (3/3)<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
∆CD<br />
2<br />
overlay<br />
y = 1.1671x - 3.3029<br />
R 2 = 0.9774<br />
2<br />
+ ∆OL<br />
2<br />
Line<br />
Dispersion mean<br />
of L1 and L2<br />
CD-SEM<br />
y = 0.9692x + 1.5178<br />
R 2 = 0.998<br />
0 10 20 30 40 50 60<br />
root sqr [∆CD²/2 + ∆OL² + [3(m1-m2)]² ]<br />
+<br />
overlay KLA<br />
CD-SEM<br />
[ 3( m − m )] 2<br />
∆<br />
1<br />
2<br />
∆ = ∆OL<br />
+ ∆OL<br />
= 2∆OL<br />
OL SO<br />
2<br />
2<br />
2<br />
1<br />
∆CD<br />
= ∆CD 1<br />
= ∆CD 2<br />
Overlay term : (3σ)² + [3mean]²<br />
with m 1 -m 2 = ½ x (CDspace 1 -CDspace 2 )<br />
CD space<br />
=<br />
∆CD<br />
( ) 2<br />
SP − SP<br />
2<br />
2 ⎡<br />
⎤<br />
+ ∆OL<br />
+<br />
1 2<br />
2 ⎢2<br />
⎥ ⎦<br />
Space error model<br />
validated experimentally<br />
⎣<br />
3
Outline<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec background<br />
• <strong>CEA</strong>-<strong>LETI</strong>-Minatec & <strong>NIKON</strong> <strong>Joint</strong> <strong>Development</strong> <strong>Program</strong><br />
• Double Patterning Process development and control<br />
• Topography impact on DP technology<br />
• Reticle & Pellicle impact on CDU & overlay<br />
• Experimental <strong>Nikon</strong> error model validation<br />
• DP processing to achieve k 1 = 0.14
Practical Illumination Solutions<br />
• The existing Dipole Optical Element was combined with a new aperture<br />
to produce the desired illumination profile.<br />
+ =<br />
ID4 ( dipole )<br />
Simulation result using NSR<br />
illumination data and SolidE<br />
simulator<br />
4x improvement using new<br />
aperture<br />
Focus offset (µm)<br />
**See V. Salvetat et al poster later tonight
32 nm Using Double Patterning<br />
• Successful DP process developed at <strong>CEA</strong>-<strong>LETI</strong>-Minatec<br />
coupled with custom illumination solution enabled 32 nm<br />
patterning with k 1 = 0.14 using a 0.85 NA dry ArF scanner.<br />
k 1 = 0.14<br />
32 nm with pitch 64nm<br />
obtained with DP<br />
Demonstrated k 1 = 0.14 while showing potential<br />
extendibility of dry ArF tools
Summary and Outlook<br />
•Double patterning process developed at <strong>CEA</strong>-<strong>LETI</strong>-Minatec is<br />
step by step controlled<br />
•Topography impact on DP process has been demonstrated<br />
•Reticle and pellicle contributions on CD & overlay budgets<br />
have been determined experimentally on DP structures<br />
•<strong>Nikon</strong> line & space error model validated experimentally<br />
•Successful partnership between <strong>CEA</strong>-<strong>LETI</strong>-Minatec and <strong>Nikon</strong><br />
demonstrating a Double Patterning process down to k 1 of 0.14<br />
is possible.<br />
•<strong>Nikon</strong> / <strong>CEA</strong>-<strong>LETI</strong> partnership will continue
Acknowledgements<br />
• Brid Connolly, Ralf Ploss, Toppan Photomasks<br />
• Amandine Pikon, Christophe Brault, Rohm&Haas<br />
Electronic Materials