© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Metrics and Methods for Reducing Rinse Water
Consumption in an Immersion System
Steve NelsonFSI International
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Outline
• Introduction• Overflow rinse uniformity• Overflow rinse chemical carryover• Dump rinse uniformity• Dump rinse chemical carryover• Conclusions
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
• 7570 liters UPW per 200mm wafer (SEMI, 2001)
• 80% for wet etching and cleaning (SEMI, 2001)
• 60% is for rinse steps (Micro, May 1995)
• 5-7 liters/cm2 Fab UPW use in 2003 (ITRS 2001)( = 3500-5000 liters per 300mm wafer)
• 0.075 l/cm2 Tool UPW use in 2003 (ITRS 2001)( = 53 liters per 300mm wafer per run, or 2650 liters/50 wafer run)
• ITRS requires significant reductions in future
Water Usage in IC Manufacturing
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Cleaning OptimizationMultiple Goal Optimization
• Process Performance– uniformity, defects, surface termination
• Productivity– COO, footprint, cycle time, throughput
• Chemical and DI Water Usage– continuous improvement
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Small Footprint, Short Cycle Time• In Batch Immersion Reduced cycle time achieved with:
– Muli-use tanks– Dilute, single-pass chemistry– Chemical injection in final rinse tank
• Challenge: Minimize Chemical and Water Usage
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Overflow Rinsing
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Plug Flow Rinsing versus High Velocity Water Jet Stir Rinsing
−
= Vqt
oeCC
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Overflow Rinse Uniformity200:1 HF thermal oxide etch,
followed by in-situ rinse, 300-mm wafers
Original Rinse Designstdev = 7.8 Årange = 33.6 Å
New Rinse Designstdev = 1.2 Årange = 4.4 Å
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Overflow Rinse Carryover
• In-situ resistivity probe installed in overflow catch weir
• Standard practice = rinse to 14 Mohm
• Not representative of water near wafer boundary layer.
– residual chemical in weir
– CO2 dissolution.DI Flow
14 MΩ-cmResistivity
Meter
Mendicino et al., SEMI Technical Symposium, SEMICON West 2001
Traditional “Rinse-to-Resistivity”
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Measuring at the Wafer
• Wafer Gap Conductivity CellLindquist et al, p.55, MRS Proc. Vol 386, 1995
• Oxide Trench w/ Buried PolyShadman et al, p.189, Solid State Phenomena Vols. 76-77 (2001) [proceedings of UCPSS 2000]
• Measure pH of Carry Over LayerThis work
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Carry Over Layer Method
DI waterwet wafer
Drain
N2 blanket
Measure pH
Orbital Shaker, 10 minutes
Fill tank with diluteHCl solution andsoak wafer.
Place wet wafer intray with DI waterto collect chemicalfor measurement.
Pull Philic waferFrom tank withsome liquid
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Carryover Layer Calibration
* R. Spearow, J. Rosato and C. R. Helms, “Studies of Rinse Efficiencies in Wet Cleaning Tools,” Proc. 3rd Intl. Symp. on Cleaning Technology inSemiconductor Device Manufacturing, J. Ruzyllo and R. Novak, Eds., The Electrochemical Society, Pennington, NJ, PV 94-7, 1994, pp. 140-152.
= −
−
k
collected
pH
pH
htan10
107.1414
Surface area of waferand dilution with waterIn collection tray.
300-mm wafer pulled from 0.05 M HCl
Wafer Pull Speed from Bath by Robot (mm/sec)0 5 10 15 20 25 30 35 40 45 50
Liqu
id L
ayer
Thi
ckne
ss (µ
m)
0
5
10
15
20
Liquid drops offwafer bottomreducing collectedamount
* h=1.51 v2/3
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Original Rinse Design Results
Rinse Time (minutes)0 1 2 3 4 5 6 7 8 9 10
[H+ ] (
mol
es/li
ter)
1e-6
1e-5
1e-4
1e-3
1e-2
1e-1
Overflow liquidTheoretical for CSTRFit (0-6 min and 6-10 min)
Carryover layer
Fit = 4.98x10-2e-0.84t
84% efficiency
Fit = 5.012x10-2e-1.002t
Fit = 1.43x10-2e-0.64t
64 % efficiency
Fit = 5.6x10-2e-0.669t
67% efficiency
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
New Rinse Design Results
Rinse Time (minutes)0 1 2 3 4 5 6 7 8 9 10
[H+]
(mol
es/li
ter)
1e-6
1e-5
1e-4
1e-3
1e-2
1e-1Overflow liquidCarryover layerFits CSTR model
Fit =7.5x10-2e-0.99t 88.8 % efficient
Fit =5.01x10-2e-1.115t
Fit =5.01x10-2e-1.048t
94.0 % efficient
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Etch Uniformity and Particle Results after Rinse Optimization
Rinse Time (seconds)0 60 120 180 240 300C
ross
waf
er u
nifo
rmity
(1 s
igm
a)
0.40.60.81.01.21.41.61.82.0
Rinse Time (seconds)90 120 150 180 210 240 270 300 330
Part
icle
s A
dded
>0.
09 µ
m
0
100
200
300
400
500
600
700
200:1 HF with in-situ overflow rinse using new rinsehardware and settings. 180-second rinse time.Followed by IPA dry.
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Benefits for 300-mm Production
120-Liters400-LitersWater Used
180-seconds600-secondsRinse Time
New Rinse MethodOld Rinse Method
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Dump Rinsing
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Dump Rinse Uniformity
Side spray bars Overhead spray bars
Shadowingfrom tank
Etch Data
Ave. = 20.7 Å
Min. = 17.7 Å
Max. = 36.2 Å
Each color = 1Å
Etch Data
Ave. = 23.8 Å
Min. = 20.6 Å
Max. = 26.0 Å
Each color = 1Å
200:1 HF thermal oxide etchfollowed by chemical dump and spray
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Tank Bottom and Wafer Surface Carryover
Tank and waferswith dilute HCl.
Dump tank, wafersand tank coatedwith chemical.
Refill tank with water.Use megasonics to mixchemical with waterto collect and measure.
830-mL chemical remainedin the tank after 15 second dump.
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Accounting on Liquid when Tanks Dumps
• Liquid remaining on wafer = πr2h x 2 x # wafers, h = 1.51v2/3 = 170-mL
• Liquid remaining on tank walls = 10-mL
• 600+170+10 = 780-mL
Liquid remaining in bottom of tank,collected with sponge and measured
Quick Dump Time (seconds)0 5 10 15 20 25 30 35
Liqu
id o
n Ta
nk B
otto
m (L
iters
)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
15 sec, 600-mL
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Dump and Spray Time (seconds)0 15 30 45 60 75 90
[H+]
(mol
es/li
ter)
1e-5
1e-4
1e-3
1e-2
1e-1
Overflow rinseNo inlet flush, with sprayWith inlet flush and spray
Spray Rinse During Dump Step
• Tank inlet flush needed to rinse chemical out of tubing or it ends up in tank in next fill.
• With good coverage spray, chemical is removed faster with dump time up to 60 seconds.
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Dump Rinse Recipe Changes
3,000,00040,000Dilution
145-Liters305-LitersTotal Water Used
270-sec425-secTotal Rinse Time
25Number dump cycles
New Dump Rinse Recipe and Hardware
Old Dump Rinse Recipe and Hardware
© 2005 FSI International, Inc. All rights reserved. always thinking // better®
Conclusions• Oxide etch uniformity and pH measurements can be
effective rinse characterization tools.• Rinsing chemical from the wafer surface behaves
different than rinsing the bulk of the liquid.
• Good spray coverage recommended for dump rinsing.
• Tank inlet flush recommended for dump rinsing.• Longer dump times with fewer dumps save time and
water and rinse better.
Top Related