chemistry and electrochemistry of copper cmp – capturing...
TRANSCRIPT
October 17, 2005
1
FLCC
Chemistry and Electrochemistry of Copper CMP – Capturing their Influence at the
Feature and Bulk Levels
Chemistry and Electrochemistry of Copper Chemistry and Electrochemistry of Copper CMP CMP –– Capturing their Influence at the Capturing their Influence at the
Feature and Bulk LevelsFeature and Bulk Levels
Ling Wang and Fiona M. DoyleLing Wang and Fiona M. DoyleUniversity of California at BerkeleyUniversity of California at Berkeley
Department of Materials Science and EngineeringDepartment of Materials Science and Engineering210 Hearst Mining Building # 1760210 Hearst Mining Building # 1760
Berkeley, CA 94720Berkeley, CA 94720--17601760
[email protected]@berkeley.edu
October 17, 2005
2
FLCC
OutlineOutlineOutline
•• Introduction Introduction –– objective of FLCC effort in CMPobjective of FLCC effort in CMP•• Passive layers and their importancePassive layers and their importance
–– Anomalous passivity induced by hydrogen Anomalous passivity induced by hydrogen peroxideperoxide
•• EQCM studies of passive filmsEQCM studies of passive films–– Relationship to other literatureRelationship to other literature–– Capturing behavior in modelsCapturing behavior in models
•• Galvanic interactionsGalvanic interactions–– How best to apply in feature scale modelsHow best to apply in feature scale models
•• ConclusionsConclusions
October 17, 2005
3
FLCC
AimAimAim
Insure uniform, global Insure uniform, global planarizationplanarization with no with no defects by means of optimized process defects by means of optimized process recipes and consumablesrecipes and consumablesIdealized singleIdealized single--phase CMP processes are phase CMP processes are now well understood in terms of the now well understood in terms of the fundamental physicalfundamental physical--chemical phenomena chemical phenomena controlling material removalcontrolling material removalThe challenge is to extend this to The challenge is to extend this to heterogeneous structures that are heterogeneous structures that are encountered when processing product encountered when processing product wafers with device featureswafers with device features
October 17, 2005
4
FLCC
ApproachApproachApproach
•• Our approach is to develop integrated Our approach is to develop integrated featurefeature--level process models linked to basic level process models linked to basic process mechanicsprocess mechanics
•• These models will drive process optimization These models will drive process optimization and the development of novel consumables and the development of novel consumables to minimize featureto minimize feature--level defects and pattern level defects and pattern sensitivitysensitivity
•• For this, we will require the capability of For this, we will require the capability of faithfully predicting defects and nonfaithfully predicting defects and non--idealities at feature boundaries. idealities at feature boundaries.
October 17, 2005
5
FLCC
Mechanical Phenomena
Chemical Phenomena
Interfacial and Colloid
Phenomena
Chemical Mechanical
Planarization
October 17, 2005
6
FLCC
OutlineOutlineOutline
•• Introduction Introduction –– objective of FLCC effort in CMPobjective of FLCC effort in CMP•• Passive layers and their importancePassive layers and their importance
–– Anomalous passivity induced by hydrogen Anomalous passivity induced by hydrogen peroxideperoxide
•• EQCM studies of passive filmsEQCM studies of passive films–– Relationship to other literatureRelationship to other literature–– Capturing behavior in modelsCapturing behavior in models
•• Galvanic interactionsGalvanic interactions–– How best to apply in feature scale modelsHow best to apply in feature scale models
•• ConclusionsConclusions
October 17, 2005
7
FLCC
Passive films, or corrosion inhibitors, are key to attaining planarization
Passive films, or corrosion inhibitors, are key Passive films, or corrosion inhibitors, are key to attaining to attaining planarizationplanarization
Kaufman’s Model for PlanarizationFor effective planarization, must maintain higher
removal at protruding regions and lower removal at recessed regions on the wafer
1- removal of passivatingfilm by mechanical action
at protruding areas
3- planarization by repetitivecycles of (1) and (2)
Metal Passivatingfilm
2- wet etch of unprotected metal by chemical action.passivating film reforms
October 17, 2005
8
FLCC
Chemistry of CopperChemistry of Copper--Water SystemWater System
-0.8-0.6-0.4-0.20.00.20.40.60.8
0 2 4 6 8 10 12 14 16pH
E, V
vs.
SHE
Cu2+
CuO
22-
Cu
Cu2O
CuO
Potential-pH diagram, with {CuT} = 10-5
Potential-pH diagram, with {CuT} = 10-4 or 10-6, 4% H2O2 or 0.5M hydroxylamineTamilmani, Huang, Raghavan & Small, JECS, 2002
October 17, 2005
9
FLCC
Chemistry of Chemistry of GlycineGlycine--Water SystemWater SystempKa1=2.350 pKa2=9.778
+H3NCH2COOH ↔ +H3NCH2COO- ↔ H2NCH2COO-
Cation: H2L+ Zwitterion: HL Anion: L-
Cu(II) glycinate complexes•Cu(H3NCH2COO)2+ : CuHL2+
•Cu(H2NCH2COO)+ : CuL+
•Cu(H2NCH2COO)2 : CuL2
Cu (I) glycinate complexes•Cu(H2NCH2COO)-
2 : CuL2-
H-O-H
H-O-H
N-H2H2-C
OO=C
C-H2H2-N
C=OOCu2+
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12 14 16pH
E, V
vs.
SHE
Cu2+
CuL2CuL+
CuO
22-
CuO
Cu2OCu
Potential-pH diagram, with {CuT} = 10-5, {LT} = 10-2
October 17, 2005
10
FLCC
Mechanisms for copper passivation and removal
Mechanisms for copper Mechanisms for copper passivationpassivation and and removalremoval
•• The The ““wet etchwet etch”” of copper is an electrochemical of copper is an electrochemical corrosion processcorrosion process–– RedoxRedox reaction in which oxidation occurs at sites separated reaction in which oxidation occurs at sites separated
(albeit at microscopic or (albeit at microscopic or nanoscopicnanoscopic scale) from sites where scale) from sites where reduction occursreduction occurs
–– Cu = CuCu = Cu2+2+ + 2e + 2e -- oxidation, anodic reactionoxidation, anodic reaction–– Ox + Ox + nene = R = R -- reduction, reduction, cathodiccathodic reactionreaction
•• The The Cu(IICu(II) may form a solid phase that ) may form a solid phase that passivatespassivates the the underlying metal. This would be the layer that is underlying metal. This would be the layer that is removed mechanically removed mechanically –– copper itself is very ductilecopper itself is very ductile
•• Regardless, slurry must contain some oxidant, Ox, Regardless, slurry must contain some oxidant, Ox, that is reduced while oxidizing copperthat is reduced while oxidizing copper
October 17, 2005
11
FLCC
•• Consider the dissolution of Consider the dissolution of copper in a slurry copper in a slurry containing hydrogen containing hydrogen peroxide:peroxide:
Cu + HCu + H22OO22 + 2H+ 2H++ = Cu= Cu2+2+ + 2H+ 2H22OO•• Reaction occurs as two Reaction occurs as two
separate reactions, at separate reactions, at different locations:different locations:–– Oxidation of the metal Oxidation of the metal -- called called
anodic dissolution:anodic dissolution:Cu = CuCu = Cu2+2+ + 2e+ 2e
–– Reduction of the hydrogen Reduction of the hydrogen peroxide peroxide -- called called cathodiccathodicreduction: reduction: HH22OO22 + 2H+ 2H++ + 2e = 2H+ 2e = 2H22OO
Polycrystalline copper
Cu =Cu =
CuCu2+2+ + 2e+ 2e
Electrochemical mechanisms of metal Electrochemical mechanisms of metal dissolutiondissolution
HH22OO22 + 2H+ 2H++ + 2e + 2e = 2H= 2H22OO
October 17, 2005
12
FLCC
•• Oxidation is more favorable at high energy Oxidation is more favorable at high energy surface defects, such as emergent surface defects, such as emergent dislocations and grain boundariesdislocations and grain boundaries–– The electrons lower the electrical potential The electrons lower the electrical potential
of these sitesof these sites•• Reduction is Reduction is ““forcedforced”” to dominate at other to dominate at other
areas, where surface sites are available for areas, where surface sites are available for sorption of hydrogen peroxide and sorption of hydrogen peroxide and hydrogen ions, and any reaction hydrogen ions, and any reaction intermediatesintermediates–– The utilization of electrons in the reduction The utilization of electrons in the reduction
process tends to raise the electrical process tends to raise the electrical potential of potential of cathodiccathodic sitessites
•• The potential difference draws electrons The potential difference draws electrons from anodic sites to from anodic sites to cathodiccathodic sites to sites to maintain the reactionmaintain the reaction
Polycrystalline copper
CuCu2+2+HH22OO22
2e2e
2H2H2H2H22OO
Electrochemical mechanisms of metal Electrochemical mechanisms of metal dissolutiondissolution
October 17, 2005
13
FLCC
•• For both the anodic and For both the anodic and cathodiccathodicreactions, the rate of reaction depends reactions, the rate of reaction depends upon the potential relative to the upon the potential relative to the equilibrium potentialequilibrium potential–– potential is a measure of the chemical potential is a measure of the chemical
activity of electrons, one of the activity of electrons, one of the participants in the reactionparticipants in the reaction
•• When the kinetics are dominated by the When the kinetics are dominated by the electron transfer process, the Butlerelectron transfer process, the Butler--VolmerVolmer equation applies:equation applies:
•• May also see kinetics affected by May also see kinetics affected by transport and the availability of reactantstransport and the availability of reactants
∴ i = io { exp -αnFη
RT⎛ ⎝
⎞ ⎠ - exp
1 - α( ) nFηRT
⎛ ⎝ ⎜ ⎞
⎠
Pote
ntia
l, V
(SH
E)
Anodic
Cathodic
Eoc
Eoc
Ecorr
log icorr
Kinetics of Electrochemical Reactions Kinetics of Electrochemical Reactions –– Crucial for Reaction ControlCrucial for Reaction Control
System settles at point where anodic and cathodic currents are equal
(electroneutrality), with distinctive corrosion current density and corrosion potential
log i
October 17, 2005
14
FLCC
Magnetic stirrer
Pt counter electrode
Luggin probe & reference electrode
Fritted glassgas bubbler
Rotating Cudisk electrode
Pote
ntia
l, V
(SH
E)log i
Anodic
Cathodic
Eoc
Eoc
Ecorr
Measure current passing between counter electrode and working copper electrode. This is the net difference between the anodic and cathodic curves, as shown
Obtain information on electrochemical Obtain information on electrochemical processes with polarization studiesprocesses with polarization studies
October 17, 2005
15
FLCC
More complex polarization behaviorMore complex polarization behaviorMore complex polarization behavior
•• In aerated solutions used in some In aerated solutions used in some of our experiments, of our experiments, cathodiccathodicreaction principally reduction of reaction principally reduction of oxygen:oxygen:
OO22 + 2H+ 2H22O + 4e = 4OHO + 4e = 4OH--
current limited by transport of current limited by transport of oxygenoxygen
•• Anodic reactions are oxidation of Anodic reactions are oxidation of copper to Cucopper to Cu2+2+, Cu, Cu22O, Cu(OH)O, Cu(OH)22 or or CuOCuO –– current may be limited by current may be limited by transport, blockage of sites or transport, blockage of sites or presence of presence of passivatingpassivating films films
•• Behavior complex if there are Behavior complex if there are multiple anodic or multiple anodic or cathodiccathodicreactions
Pote
ntia
l, V
(SH
E)Anodic
Cathodic
Net
Increasing stirring
Eoc
log ireactions
October 17, 2005
16
FLCC
Polarization Curves in CuPolarization Curves in Cu--GlycineGlycine--HH22OO
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12 14 16pH
E, V
vs.
SHE
Cu2+
CuL2CuL+
CuO
22-
CuO
Cu2OCu
i, A/m2
10-4 10-3 10-2 10-1 100 101 102 103
E m
V v
s. S
HE
-800
-600
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
pH 4pH 9pH 12
{CuT} = 10-5, {LT} = 10-2
{LT} = 10-2
October 17, 2005
17
FLCC
InIn--situsitu Electrochemical Polarization While Polishing Electrochemical Polarization While Polishing
Magnetic stirrer
Pt Counter Electrodes
Luggin Probe & Reference Electrode
Polish pad
Copper Working Electrode
Slurry poolP
Rotator Frame
ω
October 17, 2005
18
FLCC
InIn--situsitu PolarizationPolarization
i, A/m2
10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1
E m
V vs
. SH
E
-800
-600
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
No abrasionPolishing with pad onlyPolishing with pad and5 % alumina particles
i, A/m2
10-4 10-3 10-2 10-1 100 101 102 103
E m
V vs
. SHE
-800
-600
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
No abrasionPolishing with pad onlyPolishing with pad and5 % alumina particles
i, A/m2
10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1
E m
V vs
. SH
E
-800
-600
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
No abrasionPolishing with pad onlyPolishing with pad and5 % alumina particles
pH 4
pH 9
pH 12
Aqueous 10-2 M glycine, 27.6 kPa, 200 rpm
October 17, 2005
19
FLCC
Kinetic Experiments Using HKinetic Experiments Using H22OO22
DissolutionDissolution•• Cleaned, weighed, Cleaned, weighed,
copper coupons (50 copper coupons (50 x 25 x 1 mm, x 25 x 1 mm, 99.999%) suspended 99.999%) suspended in stirred solutionsin stirred solutions
•• After tests, dried After tests, dried and weighedand weighed
•• Copper removal rate Copper removal rate determined by determined by weight lossweight loss
PolishingPolishingUsed same Used same equipment used for equipment used for inin--situsitu polarization polarization teststestsElectrochemical Electrochemical information unstable information unstable with peroxidewith peroxidePolishing rates Polishing rates determined from determined from weight loss weight loss measurementsmeasurements
October 17, 2005
20
FLCC
Effect of HEffect of H22OO22 on Dissolution and on Dissolution and Polish RatesPolish Rates
0
20
40
60
80
100
120
140
160
0 1 2 3 4 5 6H2O2 , wt%
Rem
oval
Rat
e, n
m/m
in Dissolution Rate
Polish Rate
pH 40
102030405060708090
100
0 1 2 3 4 5 6H2O2, wt%
Rem
oval
Rat
e, n
m/m
in
Dissolution Rate
Polish Rate
pH 9
Aqueous 10-2 M glycine, 27.6 kPa, 200 rpm
October 17, 2005
21
FLCC
Effect of HEffect of H22OO22 on Open Circuit on Open Circuit Potential in Aqueous, 10Potential in Aqueous, 10--22 M M GlycineGlycine
0
100
200
300
400
500
600
700
0 1 2 3 4 5 6
H2O2, wt%
EO
C , m
V v
s. S
HE pH 9
-100
0
100
200
300
400
500
0 1 2 3 4 5 6
H2O2 wt%
EOC ,
V v
s. S
HEpH 4
Nernst Equation: E = Eo + 2.303RT/nF log Πaox/ Π ared
October 17, 2005
22
FLCC
Effect of HEffect of H22OO22 on Open Circuit on Open Circuit Potential in Aqueous, 10Potential in Aqueous, 10--22 M M GlycineGlycine
pH 4
pH 9y = 166.03log(x)+518.76
R2 = 0.948
0100200300400500600700
0.1 1 10
H2O2 wt%
EOC ,
mV
vs.
SH
E
y = 59.31log(x)+395.801R2 = 0.996
050
100150200250300350400450500
0.1 1 10
H2O2 wt%
EOC ,
mV
vs. S
HE
October 17, 2005
23
FLCC
Equivalent Polarization CurvesEquivalent Polarization Curves
i, A/m2
10-4 10-3 10-2 10-1 100 101 102 103
E, m
V vs
. SH
E
-800
-600
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
Polarization in the absence of H2O2
Dissolution in the presence of H2O2
Polishing in the presence of H2O2
i, A/m2
10-4 10-3 10-2 10-1 100 101 102 103E,
mV
vs. S
HE
-800
-600
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
Polarization in the absence of H2O2
Dissolution in the presence of H2O2
Polishing in the presence of H2O2
pH 4
pH 9
October 17, 2005
24
FLCC
Polarization Curves in CuPolarization Curves in Cu--GlycineGlycine--HH22OO
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12 14 16pH
E, V
vs.
SHE
Cu2+
CuL2CuL+
CuO
22-
CuO
Cu2OCu
i, A/m2
10-4 10-3 10-2 10-1 100 101 102 103
E m
V v
s. S
HE
-800
-600
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
pH 4pH 9pH 12
{CuT} = 10-5, {LT} = 10-2
{LT} = 10-2
October 17, 2005
25
FLCC
Why the unexpected Why the unexpected passivationpassivation??
PassivationPassivation induced by Hinduced by H22OO2 2 at moderate at moderate concentationsconcentations cannot be due to colloidal cannot be due to colloidal effects altering abrasiveeffects altering abrasive--substrate substrate interactions, because this effect is seen for interactions, because this effect is seen for dissolution, in the absence of abrasivesdissolution, in the absence of abrasivesNot due to a thermodynamically stable phaseNot due to a thermodynamically stable phaseCuCu22O (O (Hernandez, Hernandez, WrschkaWrschka, and , and OehleinOehlein, J. , J. ElectrochemElectrochem. Soc., 148, G389, 2001). Soc., 148, G389, 2001)CuCu22OO33 or CuOor CuO22 ((PourbaixPourbaix, 1965) , 1965) –– unlikelyunlikelyThe pH at the surface, is much higher than in The pH at the surface, is much higher than in the bulk: Hthe bulk: H22OO22 + 2H+ 2H++ + 2e = 2H+ 2e = 2H22O O –– difficult to difficult to envisage such a dramatic differenceenvisage such a dramatic difference
October 17, 2005
26
FLCC
OutlineOutlineOutline
•• Introduction Introduction –– objective of FLCC effort in CMPobjective of FLCC effort in CMP•• Passive layers and their importancePassive layers and their importance
–– Anomalous passivity induced by hydrogen Anomalous passivity induced by hydrogen peroxideperoxide
•• EQCM studies of passive filmsEQCM studies of passive films–– Relationship to other literatureRelationship to other literature–– Capturing behavior in modelsCapturing behavior in models
•• Galvanic interactionsGalvanic interactions–– How best to apply in feature scale modelsHow best to apply in feature scale models
•• ConclusionsConclusions
October 17, 2005
27
FLCC
Quartz Crystal MicrobalanceQuartz Crystal MicrobalanceQuartz Crystal Microbalance
• Sauerbrey equation:
where µq is the shear modulus of the quartz crystal, ρq the density, and f0 the resonant frequency
• for an AT-cut quartz crystal with a resonant frequency of 5 MHz gives that ∆m/∆f is –1.77 x 10-8
g/cm2Hz
( )( )2
0
21
2 ffm qqρµ
−=∆∆
• The changes in frequency of a piezoelectric quartz crystal, ∆f, are related to changes in mass, ∆m, of a substrate (e.g. Cu) that is attached to the quartz crystal:
October 17, 2005
28
FLCC
Teflon holder1 inch AT-cut polished
5 MHz Cr/Cu crystal
MaxTek, Inc., PM-700 Series Etching Products
Measurement resolution:
Frequency:0.5Hz
Weight: 0.01 µg/cm2
Thickness: 0.001 µm/min ( 0.1 Å/s)
EQCM
October 17, 2005
29
FLCC
pH 4, OCP, 0.01 M glycine premixed in acetate buffer
pH 4, OCP, 0.01 M pH 4, OCP, 0.01 M glycineglycine premixed in premixed in acetate bufferacetate buffer
Temporary loss in Temporary loss in weight, followed by weight, followed by significant gain in significant gain in weight, more weight, more pronounced at higher pronounced at higher concentration of Hconcentration of H22OO22..
October 17, 2005
30
FLCC
pH 9, OCP, 0.01 M glycine added to carbonate buffer after stabilization
pH 9, OCP, 0.01 M pH 9, OCP, 0.01 M glycineglycine added to added to carbonate buffer after stabilizationcarbonate buffer after stabilization
Slow loss in weight Slow loss in weight upon adding upon adding glycineglycine. . Temporary sharp loss Temporary sharp loss in weight after adding in weight after adding peroxide, followed by peroxide, followed by significant gain in significant gain in weight.weight.
October 17, 2005
31
FLCC
Effect of adding additional glycine, afteradding 2.09% hydrogen peroxide
Effect of adding additional Effect of adding additional glycineglycine, , afterafteradding 2.09% hydrogen peroxideadding 2.09% hydrogen peroxide
DeionizedDeionizedwaterwater
pH 9pH 9
October 17, 2005
32
FLCC
Open circuit potential of copper, pH 9, 0.01 M glycine and 2.09% hydrogen peroxide
Open circuit potential of copper, pH 9, 0.01 M Open circuit potential of copper, pH 9, 0.01 M glycineglycine and 2.09% hydrogen peroxideand 2.09% hydrogen peroxide
No No passivationpassivation without without HH22OO2. 2. See that behavior See that behavior is strongly dependent is strongly dependent on history of on history of glycineglycineadditions; oxidized additions; oxidized layers must resist layers must resist dissolutiondissolution
No HNo H22OO22. . Potential Potential same as that same as that induced by induced by HH22OO22
October 17, 2005
33
FLCC
Potential application in CMPPotential application in CMPPotential application in CMP
Polarize copper sample to oxidizing potential, to allow protective film to develop, then add
glycine - little effect
1- removal of protectivefilm by mechanical action
at protruding areasMetal Passivating
film2- freshly created
surface will dissolve actively in
glycine solution. Recessed regions remain protected
October 17, 2005
34
FLCC
Effect of glycine and H2O2 additions at different potentials, pH 9, 0.01 M glycineEffect of Effect of glycineglycine and Hand H22OO22 additions at additions at
different potentials, pH 9, 0.01 M different potentials, pH 9, 0.01 M glycineglycine
Iron diskIron disk--Au ring Au ring electrode. Helectrode. H22OO22produced during produced during reduction of Oreduction of O22is rapidly is rapidly reduced at high reduced at high and low and low potentials, but potentials, but can escape can escape electrode at electrode at intermediate intermediate potentialspotentials
S. S. ZeZeččevievićć, D.M. , D.M. DraDražžiićć, S. , S. GojkiviGojkivićć; ; J. J. ElectroanalElectroanal. . ChemChem, , 265 (1989) 179265 (1989) 179
At controlled At controlled potentials, either potentials, either oxidizing or reducing, oxidizing or reducing, HH22OO22 does NOT lead to does NOT lead to weight increase. weight increase. Protective film must be Protective film must be sensitive to potentialsensitive to potential
However, this is not consistent However, this is not consistent with with passivationpassivation at high at high concentrations of Hconcentrations of H22OO22
October 17, 2005
35
FLCC
Capturing Chemical/Electrochemical Behavior into Models for CMP
Capturing Chemical/Electrochemical Behavior Capturing Chemical/Electrochemical Behavior into Models for CMPinto Models for CMP
•• We know that chemical/electrochemical material We know that chemical/electrochemical material removal rates are relatively slow compared to rates removal rates are relatively slow compared to rates measured while polishingmeasured while polishing–– So is chemistry important at all?So is chemistry important at all?
•• Dornfeld group has successfully modeled various Dornfeld group has successfully modeled various aspects of mechanical behavioraspects of mechanical behavior
•• Talbot group has added colloidal behavior to the Talbot group has added colloidal behavior to the LuoLuoand Dornfeld model for material removal rates; and Dornfeld model for material removal rates; accounts for effect of solution chemistry on accounts for effect of solution chemistry on agglomeration of abrasive particlesagglomeration of abrasive particles–– Would there be large errors from ignoring chemistry?Would there be large errors from ignoring chemistry?
•• Yes!Yes!
October 17, 2005
36
FLCC
Chemistry also interacts synergistically with mechanical/colloidal phenomena
Chemistry also interacts synergistically with Chemistry also interacts synergistically with mechanical/colloidal phenomenamechanical/colloidal phenomena
Mechanical forces on copper introduce defects, increasing reactivity
Mechanical properties of films appear to be strongly dependent on chemistry, and probably potential
Chemistry affects degree of aggregation of abrasive particles.
Copper nanoparticleshave dramatic effect
October 17, 2005
37
FLCC
Adding chemistry/electrochemistry to modelsAdding chemistry/electrochemistry to modelsAdding chemistry/electrochemistry to models
•• Obtain quantitative information on Obtain quantitative information on mechanical properties of surface films under mechanical properties of surface films under different chemical conditionsdifferent chemical conditions–– Add to modelsAdd to models
•• Use curveUse curve--fitting to capture major chemical fitting to capture major chemical detailsdetails–– add to models as additive material removal rate add to models as additive material removal rate
over and above mechanical phenomenaover and above mechanical phenomena•• Model material removal with Model material removal with ““chemical chemical
toothtooth”” approach, and add kinetic data to approach, and add kinetic data to estimate rates for reforming passive filmsestimate rates for reforming passive films
October 17, 2005
38
FLCC
OutlineOutlineOutline
•• Introduction Introduction –– objective of FLCC effort in CMPobjective of FLCC effort in CMP•• Passive layers and their importancePassive layers and their importance
–– Anomalous passivity induced by hydrogen Anomalous passivity induced by hydrogen peroxideperoxide
•• EQCM studies of passive filmsEQCM studies of passive films–– Relationship to other literatureRelationship to other literature–– Capturing behavior in modelsCapturing behavior in models
•• Galvanic interactionsGalvanic interactions–– How best to apply in feature scale modelsHow best to apply in feature scale models
•• ConclusionsConclusions
October 17, 2005
39
FLCC
How are electrode kinetics affected by multiple half-cell reactions?
How are electrode kinetics affected by How are electrode kinetics affected by multiple halfmultiple half--cell reactions?cell reactions?
Pote
ntia
l, V
(SH
E)Anodic
CathodicEcorr
• Metal A, which is more active, corrodes at E’corr
• Metal N, which is more noble, corrodes at a higher potential, Ecorr
• When metals A and N are in electrical contact, they develop a common corrosion potential, E”corr
• At E”corr, the corrosion current of N is lower than for the isolated metal, but the corrosion current of A is greater than for the isolated metal
• This effect is called galvanic corrosion
E’corr
E”corr
A
N
log i
October 17, 2005
40
FLCC
• Galvanic cells may form when polished patterned copper wafers with Ta or Ti-based barriers are immersed in CMP slurries; either copper or the barrier layer is corroded preferentially by galvanic corrosion
• Galvanic corrosion may impair the quality and reliability of copper interconnects. This effect will become increasingly pronounced with smaller-scale features
• Galvanic corrosion can be observed microscopically, measured by electrochemical techniques and controlled by modifying the slurry chemistry, temperature etc.
• Therefore, it is becoming increasingly important to understand the galvanic cells that develop during copper CMP, and incorporate this understanding into models
MotivationMotivation
October 17, 2005
41
FLCC
What is the effect of different electrode What is the effect of different electrode areas?areas?
•Total anodic CURRENT must equal total cathodic current•A small anode:cathode area ratio causes more severe galvanic corrosion, because• The total anodic current is concentrated over a smaller area • Hence the anodic dissolution rate increases
Jones, Denny A., Principles and prevention of corrosion, 2nd ed.
•This effect could be really problematic in CMP if diffusion barriers are more active than copper
October 17, 2005
42
FLCC
Galvanic Series and Sensitivity to Galvanic Series and Sensitivity to Solution ChemistrySolution Chemistry
•• A galvanic series is a series in which the corrosion A galvanic series is a series in which the corrosion potential of different metals and alloys is placed in potential of different metals and alloys is placed in ascending orderascending order
•• Comprehensive galvanic series have been published Comprehensive galvanic series have been published for a large number of metallic materials in sea water for a large number of metallic materials in sea water and aerated fresh waterand aerated fresh water
•• Much less information is available for copper and Much less information is available for copper and barrier materials in solution chemistries barrier materials in solution chemistries representative of CMP slurriesrepresentative of CMP slurries
•• Anticipate significant variations in potentials of Anticipate significant variations in potentials of copper and barrier materials according to the copper and barrier materials according to the precise precise complexingcomplexing agents and oxidantsagents and oxidants
October 17, 2005
43
FLCC
Polarization tests in solutions containing H2O2 and glycine
Polarization tests in solutions containing Polarization tests in solutions containing HH22OO22 and and glycineglycine
• Electrodes: Cu, Ti, Ta polished down to 0.25 polished down to 0.25 µµmmusing diamond pasteusing diamond paste• Electrolytes:
–Solution containing H2O2 and 0.01M glycine–Solution pH controlled using buffer solution
• Acetate buffer pH=3• Carbonate buffer pH=9• KOH solution pH=12
–Varying H2O2 concentration•• PotentiodynamicPotentiodynamic polarization measurements were polarization measurements were made on the Ta and Cu rotating disk electrodes, made on the Ta and Cu rotating disk electrodes, scanning at 1mV/s, rotating at 650 rpmscanning at 1mV/s, rotating at 650 rpm
October 17, 2005
44
FLCC
Ta-Cu, Ti-Cu, pH 3 buffer + 0.01M glycine+ 0.65 M H2O2
TaTa--Cu, TiCu, Ti--Cu, pH 3 buffer + 0.01M Cu, pH 3 buffer + 0.01M glycineglycine+ 0.65 M H+ 0.65 M H22OO22
-600
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
2000
-10 -8 -6 -4 -2 0
LogI, A/cm^2
E vs
SC
E, m
V
Ta-0.65%wt H2O2
Cu-0.65%wt H2O2
-400
-200
0
200
400
600
800
1000
1200
1400
1600
1800
2000
-10 -8 -6 -4 -2 0
LogI, A/cm^2
E vs
SC
E, m
V
Ti-0.65%wt H2O2
Cu-0.65%wt H2O2
October 17, 2005
45
FLCC
Reminder – what you are seeing with polarization curves
Reminder Reminder –– what you are seeing with what you are seeing with polarization curvespolarization curves
Pote
ntia
l, V
(SH
E)
Anodic
Cathodic
Eoc
Eoc
Ecorr
Pote
ntia
l, V
(SH
E)
Net cathodic
Net anodic
Ecorr
log i log i
October 17, 2005
46
FLCC
Polarization tests Polarization tests –– commercial slurrycommercial slurry
•• ElectrodesElectrodes: : Ta rotating disk electrode (6.25mm in Ta rotating disk electrode (6.25mm in diameter) and Cu rotating disk (10mm in diameter), diameter) and Cu rotating disk (10mm in diameter), polished down to 0.25 polished down to 0.25 µµmm using diamond pasteusing diamond paste
•• ElectrolytesElectrolytes::–– Mixture of commercial slurry A + HMixture of commercial slurry A + H22OO22 (final concentration: (final concentration:
1.0%wt H1.0%wt H22OO22 or 6.0%wt Hor 6.0%wt H22OO22), referred to as Cu), referred to as Cu--1.0%H1.0%H22OO22, , CuCu--6.0%H6.0%H22OO22, Ta, Ta--1.0% H1.0% H22OO22 ,Ta,Ta--6.0% H6.0% H22OO22
–– 50 times dilution of the above mixture, referred to as Cu50 times dilution of the above mixture, referred to as Cu--1.0%1.0%--diluted, Cudiluted, Cu--6.0 %6.0 %--diluted, Tadiluted, Ta--1.0 %1.0 %--diluted,Tadiluted,Ta--6.0 %6.0 %--diluteddiluted
–– The above electrolytes had a measured pH of 4.00The above electrolytes had a measured pH of 4.00•• PotentiodynamicPotentiodynamic polarization measurements were polarization measurements were
made on the Ta and Cu rotating disk electrodes, made on the Ta and Cu rotating disk electrodes, scanning at 1mV/s, rotating at 650 rpmscanning at 1mV/s, rotating at 650 rpm
October 17, 2005
47
FLCC
TaTa--Cu: Commercial slurry + 1% HCu: Commercial slurry + 1% H22OO22
-1000-800-600-400-200
0200400600800
10001200140016001800
-10 -8 -6 -4 -2 0
LogI, A/cm^2
E vs
SC
E, m
V
Cu-1.0%-diluted
Ta-1.0%-diluted
October 17, 2005
48
FLCC
TaTa--Cu: Commercial slurry + 6% HCu: Commercial slurry + 6% H22OO22
-800-600-400-200
0200400600800
100012001400160018002000
-10 -8 -6 -4 -2 0
LogI, A/cm^2
Evs
SCE,
mV
Cu-0.6%-dilutedTa-6.0%-diluted
October 17, 2005
49
FLCC
Results of electrochemical measurements
Results of electrochemical Results of electrochemical measurementsmeasurements
0 200 400 600 800 1000-2.0x10-6
0.0
2.0x10-6
4.0x10-6
6.0x10-6
8.0x10-6
1.0x10-5
1.2x10-5
1.4x10-5
1.6x10-5
H2O2
Gal
vani
c C
urre
nt,A
Time,s
Ta:Cu=2.6:6.0 (cm2) Ta:Cu=3.9:6.0 (cm2) Ta:Cu=3.9:4.5 (cm2)
0 200 400 600 800 1000-0.300
-0.250
-0.200
-0.150
-0.100
-0.050
0.000
0.050
0.100
0.150
0.200
Pote
ntia
l, V
vs S
CE
Time, s
Ta:Cu=3.9:6.0 (cm2) Ta:Cu=2.6:6.0 (cm2) Ta:Cu=3.9:4.5 (cm2)
Area ratio effect, without polishing
Start with DI water, then add buffer, then glycine, then H2O2 to create pH=9 carbonate buffer, 0.01M glycine and 0.3%H2O2
October 17, 2005
50
FLCC
Results of electrochemical measurementResults of electrochemical measurementResults of electrochemical measurementEffects of [H2O2], pH=9, without polishing
0 300 600 900 1200
0.0
5.0x10-6
1.0x10-5
1.5x10-5
2.0x10-5
Gal
vani
c C
urre
nt, A
Time,s
0.3% 0.1% 0.6% 1.0%
0 300 600 900 1200-0.300
-0.200
-0.100
0.000
0.100
0.200H2O2, 30 wt%
stock glycine sltn
carbonate buffer
Mix
ed P
oten
tial,
V vs
SC
E
0.3% 0.1% 0.6% 1.0%
in pH=9 carbonate buffer, 0.01M glycine and varying [H2O2]
October 17, 2005
51
FLCCsample assembly capable electrochemical control and measurement
Cu Ta
Pt
Cu wires underneath for electrode connection
Galvanic currents and potentials will be measured with or without polishing, to investigate the effects of surface modification by chemicals and mechanical force on the polarity and galvanic current; area ratio of Cu:Ta is to be varied using different sample assemblies with different area ratios
Experimental setup and samplesExperimental setup and samplesExperimental setup and samples
•sample carrier
Potentiostat
Polishing pad
Rotating platen
Slurry film
Sample assemblyconditioner
slurrypressure
rotation
Electrical connector to Potentiostat leads
Reference electrode /Pseudo RE connecting to Potentiostat
October 17, 2005
52
FLCC
How to model galvanic effectsHow to model galvanic effects• Modest galvanic currents can have devastating
effects on quality if localized• Need to focus on feature level
• Basic electrochemical theory suggests that the corrosion current in a galvanic couple can be predicted• Relative areas of anodic and cathodic metals can
then be invoked to predict damage due to galvanic interactions
• But the corrosion potentials of Cu, Ta and Ti differ in different solution chemistries; in some cases the relative nobility changes• Hence the polarization behavior of each material
must be determined for different slurries to provide input for modeling copper CMP
October 17, 2005
53
FLCC
OutlineOutlineOutline
•• Introduction Introduction –– objective of FLCC effort in CMPobjective of FLCC effort in CMP•• Passive layers and their importancePassive layers and their importance
–– Anomalous passivity induced by hydrogen Anomalous passivity induced by hydrogen peroxideperoxide
•• EQCM studies of passive filmsEQCM studies of passive films–– Relationship to other literatureRelationship to other literature–– Capturing behavior in modelsCapturing behavior in models
•• Galvanic interactionsGalvanic interactions–– How best to apply in feature scale modelsHow best to apply in feature scale models
•• ConclusionsConclusions
October 17, 2005
54
FLCC
ConclusionsConclusionsConclusions
•• The chemical and electrochemical mechanisms in The chemical and electrochemical mechanisms in copper CMP are complexcopper CMP are complex–– Sometimes the behavior is not fully understoodSometimes the behavior is not fully understood
•• However, the subtleties of chemical behavior alone However, the subtleties of chemical behavior alone may be of second order, compared to synergism may be of second order, compared to synergism between mechanical, chemical and colloidal between mechanical, chemical and colloidal behaviorbehavior–– A mechanistic understanding is invaluable in tackling A mechanistic understanding is invaluable in tackling
modeling of this synergismmodeling of this synergism•• Galvanic effects are different; here models must Galvanic effects are different; here models must
focus on feature scale effectsfocus on feature scale effects