CEA-LETI - NIKON Joint Development Program ... - Nikon Precision

CEA-LETI - NIKON Joint 

Development Program Update 

C. Lapeyre, S. Barnola, I. Servin, 

S. Gaugiran, S. Tedesco, 

L. Pain, A.J. Hazelton, 

V. Salvetat, M. McCallum

Outline 

• CEA-LETI-Minatec background 

• CEA-LETI-Minatec & NIKON Joint Development Program 

• Double Patterning Process development and control 

• Topography impact on DP technology 

• Reticle & Pellicle impact on CDU & overlay 

• Experimental Nikon error model validation 

• DP processing to achieve k 1 = 0.14

LETI Background 

• 1,600 people on research programs (1000 CEA headcount) 

• 8000m 2 clean room 

• 160 industrial partners 

• 30 startups 

• 1600 patents portfolio 

• 250 patents in 2008 

3

LETI Background 

• 1,600 people on research programs (1000 CEA headcount) 

• 8000m 2 clean room 

• 160 industrial partners 

• 30 startups 

• 1600 patents portfolio 

• 250 patents in 2008 

Nanoelectronic 200mm and 300mm 

•200mm clean room (2500 m 2 ) 

•300mm ball room (1500 m 2 ) 

• State of the art tools 

• Full research line 

• Continuous operation 

7 days per week/ 24 hours per day 

4

Nikon and LETI Program 

• Nikon and CEA-LETI-Minatec are partnering through a 

Joint Development Program to understand the sensitivities 

of various DE/DP techniques since early 2007. 

• Placing special emphasis on ensuring next generation 

exposure tools satisfy DP requirements. 

• Using a 0.85 NA dry ArF Nikon scanner in LETI’s Nanotec 

300 research facility in Grenoble, France. 

NSR-S307E – 0.85 NA Dry ArF

Outline 








45nm Double Patterning Process 

• NSR-S307E 0.85 NA ArF 

• Sokudo RF3 

• Lam Versys 

• 45 nm L/S pattern 

(k 1 

=0.20) 

Stack 

Resist ArF 

177nm 

Organic Barc 

32 nm 

HM1: Nitride 

30 nm 

HM2:Oxide 

20 nm

DP Process Control (1/3) 

• Integration stack requirements : 

– Good selectivity HM1/HM2 

– Reduce thickness to minimize 

topography 

– Low temperature deposition to avoid 

carbon mask damage 

• Etch Process optimization : 

– Ensure good resist budget margin to 

avoid faceting 

– Control of resist trimming 

– Etching chemistry selectivity 

BARC/HM1/HM2 

– Conservation of CD and profile of 

pattern 1 through etch2 

Line 2 : 

Oxide 

Line 1 : Nitride 

on Oxide 

Good oxide/nitride selectivity 

Good CD control with trimming


70 

Impact of each step on CD mean 

60 

50 

66.0 

64.0 

CD (nm) 

40 

30 

48.0 

-3nm 

44.5 46.5 

20 

10 

0 

litho1 etch1 (line1) litho2 etch2 (line1) etch2 (line2) 

CD1 is well maintained through etch2 : CD bias ~-3nm 

Etch process very robust : CD1 and CD2 controlled independently


4.5 

ITRS 45nm requirement 

CD uniformity 3σ (nm) 

4.0 

3.5 

3.0 

2.5 

2.0 

1.5 

1.0 

ITRS 32nm requirement 

2.6 

3.0 

3.3 3.2 

3.4 

0.5 

0.0 

litho 1 etch1 litho2 etch2 (line1) etch2 (line2) 

CD uniformity remains around 3nm throughout the DP 

process in agreement with ITRS 32nm requirements

Outline 








Topography Impact on Litho (1/3) 

180nm 

180nm 

65nm 

Si 

45nm 

Si 

Si 

Resist, 177nm 

Barc, 32nm 

15nm 

30nm 

45nm 

180nm 65nm Resist, 177nm 

Barc, 32nm 

Litho1 

Etch1 

Litho2 

•Substrate topography BARC thickness 

variation on features 

Impact on reflectivity control 

Impact on line1 etch2 dependent upon 

BARC/nitride selectivity 

12.3nm 

7.7nm 

2.6nm 

15nm height 30nm height 45nm height


15nm height 

30nm height 

45nm height 

High topography alters resist profiles


69 

10 

CD mean (nm) 

68.5 

68 

67.5 

67 

66.5 

66 

65.5 

+1.1nm 

+1.3nm 

9 

8 

7 

6 

5 

4 

3 

2 

1 

CD uniformity 3σ (nm) 

65 

0 

0 15nm 30nm 45nm 

substrate topography 

Higher topography induces: Larger CD and larger dispersion 

Difference between 3σ(L2) and 3σ(L1) 

Due to topography, DP requires : 

Development of planarizing BARC to improve DP process 

Thinnest stack as possible : 30nm Si 3 N 4 / 20nm SiO 2

Outline 








Reticle Impact on CD and Overlay 

• 6% att-PSM in bright field tonality, no pellicle 

• Quartz flatness 0.5µm (T) 

• Designed with e-beam writer E5000 

• Mask : CD tool : NANOSEM3, Registration : IPRO4 (VISTEC) 

• Wafer : CD Hitachi CG4000, Overlay : CD-SEM & KLA 

32 nm reticle specifications 

• CD 0.5nm 

• OVL 0.9nm 

Reticle data fits ITRS 

Mask CDU < ¼ of wafer CDU 

Mask OVL < 10% of wafer OVL 

budget

Pellicle Induced Reticle Deformation 

In collaboration with Toppan Photomasks 

- Mechanical performance of pellicle may introduce pattern distortions 

- Image placement has been registered before and after pellicle mounting 

on same relevant structures spread out over the layout (10µm) 

Mapping measurements 

-Mask registration measured in X and Y axis 

-Distortions after isotropic and orthotropic corrections 

-Orthotropic corrected by exposure tool 

Pellicle distortions 3σ (nm) after corrections 

• X axis 3.2 nm 

• Y axis 2.4 nm 

Pellicle induced distortion represents 15% additional error in the 

image placement budget for the 45nm node

Outline 








NIKON Error Model Validation 

Nikon developed a theoretical model to establish the required CD and 

overlay requirements for DP that could be used in tool design. This was 

targeted for experimental verification at Leti. 

[A.J. Hazelton et. al, LithoVision 2008]

L & S Error Model Validation (1/3) 

•Standard LELE process : 19 wafers with programmed x-trans offsets 

•CD measurements : line1, line2, space1 and space2, and 3σ 

•Overlay measurements 

CD & SPACE 

70 

65 

CD1 

CD2 

X mean raw data (nm) 

14 

10 

6 

2 

-2 

OVERLAY DATA 

CD-SEM data (nm) 

60 

55 

50 

45 

40 

35 

30 

SP1 

SP2 

-8 -6 -4 -2 0 2 4 6 8 10 

programmed X-translation offsets (nm) 

-6 

-8 -6 -4 -2 0 2 4 6 8 10 

programmed X-translation offsets (nm)


∆ 

Global dispersion 

on the two line 

populations 

⎡3 

⎢ 

⎣2 

( ) 

⎤ 

2 

L − L 

2 

2 

CD Line 

= ∆CD 

+ 

1 2 

Dispersion mean 

of L1 and L2 

25 

⎥ 

⎦ 

∆CD 

= ∆CD 1 

= ∆CD 

Difference of 

line CDs 

20 

• Line error model is validated 

experimentally by varying both 

CD line and CD space 

∆CDline 

15 

10 

5 

y = 0.9974x + 0.0135 

R 2 = 1 

0 

0 5 10 15 20 25 

root sqr[ ∆CD² + [3/2(L1-L2)]² ]

∆CD 

space 

= 

Global dispersion on 

the two space 

populations 

∆CDspace 


60 

50 

40 

30 

20 

10 

0 

∆CD 

2 

overlay 

y = 1.1671x - 3.3029 

R 2 = 0.9774 

2 

+ ∆OL 

2 

Line 

Dispersion mean 

of L1 and L2 

CD-SEM 

y = 0.9692x + 1.5178 

R 2 = 0.998 

0 10 20 30 40 50 60 

root sqr [∆CD²/2 + ∆OL² + [3(m1-m2)]² ] 

+ 

overlay KLA 

CD-SEM 

[ 3( m − m )] 2 

∆ 

1 

2 

∆ = ∆OL 

+ ∆OL 

= 2∆OL 

OL SO 

2 

2 

2 

1 

∆CD 

= ∆CD 1 

= ∆CD 2 

Overlay term : (3σ)² + [3mean]² 

with m 1 -m 2 = ½ x (CDspace 1 -CDspace 2 ) 

CD space 

= 

∆CD 

( ) 2 

SP − SP 

2 

2 ⎡ 

⎤ 

+ ∆OL 

+ 

1 2 

2 ⎢2 

⎥ ⎦ 

Space error model 

validated experimentally 

⎣ 

3

Outline 








Practical Illumination Solutions 

• The existing Dipole Optical Element was combined with a new aperture 

to produce the desired illumination profile. 

+ = 

ID4 ( dipole ) 

Simulation result using NSR 

illumination data and SolidE 

simulator 

4x improvement using new 

aperture 

Focus offset (µm) 

**See V. Salvetat et al poster later tonight

32 nm Using Double Patterning 

• Successful DP process developed at CEA-LETI-Minatec 

coupled with custom illumination solution enabled 32 nm 

patterning with k 1 = 0.14 using a 0.85 NA dry ArF scanner. 

k 1 = 0.14 

32 nm with pitch 64nm 

obtained with DP 

Demonstrated k 1 = 0.14 while showing potential 

extendibility of dry ArF tools

Summary and Outlook 

•Double patterning process developed at CEA-LETI-Minatec is 

step by step controlled 

•Topography impact on DP process has been demonstrated 

•Reticle and pellicle contributions on CD & overlay budgets 

have been determined experimentally on DP structures 

•Nikon line & space error model validated experimentally 

•Successful partnership between CEA-LETI-Minatec and Nikon 

demonstrating a Double Patterning process down to k 1 of 0.14 

is possible. 

•Nikon / CEA-LETI partnership will continue

Acknowledgements 

• Brid Connolly, Ralf Ploss, Toppan Photomasks 

• Amandine Pikon, Christophe Brault, Rohm&Haas 

Electronic Materials

CEA-LETI - NIKON Joint Development Program ... - Nikon Precision

Create successful ePaper yourself

Delete template?

Save as template?