1Metryx Copyright ©
2Metryx Copyright ©
Using Mass Metrology for Process Monitoring and Control During 3D Stacking of IC’s
SEMATECH 3D Interconnect Workshop
11th July 2012
3Metryx Copyright ©
Mass Metrology
More MassLess Mass
� All process steps create a wafer mass change
� This mass change reflects all aspects of process performance within a wafer
� Metryx Mass Metrology provides passive data collection on product wafers to assess process performance
4Metryx Copyright ©
Mass Metrology
More MassLess Mass
� All process steps create a wafer mass change
� This mass change reflects all aspects of process performance within a wafer
� Metryx Mass Metrology provides passive data collection on product wafers to assess process performance
Etch
5Metryx Copyright ©
Mass Metrology
More MassLess Mass
� All process steps create a wafer mass change
� This mass change reflects all aspects of process performance within a wafer
� Metryx Mass Metrology provides passive data collection on product wafers to assess process performance
Deposition
6Metryx Copyright ©
Measurement Capability
� 1σ thickness repeatability for blanket films on 200mm & 300mm wafers� Thickness sensitivity improves with patterned wafers and increased surface area� Atomic level accuracy maintained (or improved) with smaller feature sizes and more
complex structures
H He
Li Be B C N O F Ne
Na Mg Al Si P S Cl Ar
K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr
Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe
Cs Ba La Hf Ta W Re Os Ir Pt Au Hg Tl Pb Bi Po At Rn
Ti 2.5ǺTiN 2.5Ǻ
Cr 1.6Ǻ
Co 1.3ǺCoSi 2.6ǺCoWP 1.2Ǻ
Ni 1.3ǺNiSi 5.0Ǻ
Cu 1.3Ǻ
Al 4.2ǺAlCu 3.8ǺAlN 3.5Ǻ
Si 4.9ǺSiO2 4.9ǺSiN 4.5ǺSiON 4.7ǺSiOC 7.5Ǻ
Zr 1.7Ǻ
Hf 0.9ǺHfO2 1.2Ǻ
Ta 0.7ǺTaN 1.3Ǻ
W 0.6Ǻ
Mo 0.6Ǻ
Pt 0.5Ǻ Au 0.6Ǻ
Ag 1.1Ǻ
7Metryx Copyright ©
Mass ≠ Weight
� Weight Measurement
� Unstable, irreproducible, not designed for
semiconductor measurement use
� Mass Measurement� Load-cell utilising complex
force measurement
� Real-time corrections for internal and external forces influencing measurement
� Fully automatic wafer handling and host communication compliant
8Metryx Copyright ©
3D IC Metrology Challenges
TSV Formation M & I Challenges
• Via CD, depth & profile (optical/mass/interferometry/)
• Etch defects/residues (BF- DF pattern inspection)
Via Fill M & I Challenges
• Barrier and Seed continuity (Electrical)
• Void detection in filled TSV’s (Mass/Electrical)
Temporary Carrier Processing M & I Challenges
• Glue layer void/defect & thickness variations (optical (IR/SAM))
Thinning process M & I Challenges
• Wafer thickness and shape (Optical(IR)/capacitance)
• Post thinning/recess etch backside defects ( Optical BF/DF)
• Cu TSV co-planarity (Interferometry/confocal)
• Edge defects (optical )
Debonding I challenges
• Thin wafer inspection on FFC(BF/DF inspection)
• Thin die inspection on FFC (BF/DF inspection)
Dicing and Stacking M& I challenges
• Die alignment (Optical (IR) X-ray )
• Underfill integrity (X-ray)
• Edge boundary violation (BF- pattern recognition software)
TSV module
Wafer thinning module
Debonding and stacking module
9Metryx Copyright ©
3D IC Metrology Challenges
TSV Module
� Via CD, depth & profile
� Etch defects/residues
� Barrier and Seed
� Via filling
� Void detection in filled TSV’s
Wafer Thinning Module
� Glue layer void & thickness
� Wafer thickness & shape
� Post thinning/recess etch
� Cu TSV co-planarity
� Edge defects
Debonding & Stacking Module
� Thin wafer inspection
� Thin die inspection
� Die alignment
� Underfill integrity
� Edge boundary violation
10Metryx Copyright ©
TSV Etch – Mass vs Depth
y = -0.0003x2
+ 0.2343x - 0.1731
R2 = 0.9992
0
5
10
15
20
25
30
0 20 40 60 80 100 120 140 160 180
Mass Loss (mg)
Via
Dep
th (
um
s)
Via Depth
5 um Diameter Vias
PoR 25 µm Depth
y = -0.0003x2
+ 0.2343x - 0.1731
R2 = 0.9992
0
5
10
15
20
25
30
0 20 40 60 80 100 120 140 160 180
Mass Loss (mg)
Via
Dep
th (
um
s)
Via Depth
5 um Diameter Vias
PoR 25 µm Depth
11Metryx Copyright ©
TSV Etch
� Mass is used as a monitor of etch process stability
� Out of spec lots usually have deeper TSV’s etch
� Depth variations in TSV has implications in the TSV grinding and reveal process
12Metryx Copyright ©
Liner/Barrier/Seed
� Highly topographical layer
� Difficult to measure optically as most of film is on sidewalls, or bottom of features
� Surface area increase improves sensitivity of mass change.
� Etch variations produce surface area variations so mass after TSV etch and mass of L/B/S are related.
0
100
200
300
400
500
600
700
800
1 2 3 4 5
Aspect Ratio (AR)
Ma
ss
(u
g)
Sidewall
Top Surface
1:1 5:1 10:1 20:1 50:10
100
200
300
400
500
600
700
800
1 2 3 4 5
Aspect Ratio (AR)
Ma
ss
(u
g)
Sidewall
Top Surface
1:1 5:1 10:1 20:1 50:1
13Metryx Copyright ©
Via Filling and Voids
� Theoretical calculations reveal possibility to detect voids in Cu vias
� Assuming void width = via diameter
� More than 6600 TSVs with 100nm height voids need to be present to be detected (50µ x 5µ TSV)
14Metryx Copyright ©
Via Filling and Voids
Wafer TSV etch Liner B/S Cu fill Anneal
MetryxMetryx Metryx MetryxMetryx Metryx
CMP
Metryx
� Mass of wafers with voids is
less than mass of wafers without voids
� Difficult to separate “mass”of voids from mass of over etch, or over burden
15Metryx Copyright ©
y = 0.9882x + 0.0141
R2 = 0.9994
-18
-15
-12
-9
-6
-3
0
3
6
9
12
15
18
-18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18
Mass Deposited Deviation (mg)
Mass R
em
oved
Devia
tio
n (
mg
)
CMP (PoR)
Outliers
Linear (CMP (PoR))
CMP Stability
�The correlation between the mass added (Barrier + Fill) is very well correlated to the mass removed during CMP.
�Therefore the CMP is well controlled. Over or under-polish in the CMP results in data scatter.
�The over and under polish splits are clearly distinguished.
Under Polish
Over Polish
(M n+4 - M n+3)
(M n+3 - M n+1)
16Metryx Copyright ©
Via Filling and Voids
Wafer TSV etch Liner B/S Cu fill Anneal
MetryxMetryx Metryx MetryxMetryx Metryx
CMP
Metryx
� Mass change of wafers with voids is less than mass change of wafers without voids
� The difference is related to the loss of plating liquid from the void during the anneal process
� Mass can therefore indicate the presence of voids in the TSV
17Metryx Copyright ©
Wafer Thickness Measurement
� Wafer thickness by micrometer
� Correlates well with mass data
� Mass metrology provides average wafer thickness quickly (<60s/wafer) compared to micrometer (manual) and acoustic reflection probing (time consuming) methods
Carrier Mass vs Thickness
y = 0.0414x + 0.6173
R2 = 0.9764
28.3
28.4
28.5
28.6
28.7
28.8
28.9
29.0
670 672 674 676 678 680 682 684
Thickness (um)
Mas
s (
g)
� Wafer thickness by acoustic reflection data
� Correlates well with mass calculated thickness
18Metryx Copyright ©
Summary
� Mass Metrology is an effective technique for monitoring process excursions during
� TSV etch
� Barrier/Seed Deposition
� Cu plating
� Wafer grinding
� Acknowledgements
� Co Authors
� Sandip Halder, Peter leunissen, Andy Miller, Mirielle Maenhoudt, Eric BeyneIMEC, Kapeldreef 75, B-3001 Leuven, Belgium
� Adrian Kiermasz, Gary DitmerMetryx, 1240 Park Avenue, Aztec West, Bristol, UK
� Additional thanks to
� Augusto Redolfi, Eddy Kunnen, Sarasvathi Thangaraju, Harold Philipson
19Metryx Copyright ©
Metryx