flcc feature-level compensation control plasma technology
DESCRIPTION
Plasma Technology FLCC Workshop & Review April 5, 2006 Professors Jane P. Chang (UCLA), Michael A. Lieberman, David B. Graves (UCB) and Allan J. Lichtenberg, John P. Verboncoeur, Zhuwen Zhou, Sungjin Kim, Alan Wu, Emi Kawamura, Jon T. Gudmundsson, Chengche Hsu, Joe Vegh, Insook Lee (UCB), and John Hoang (UCLA) FLCC Workshop & Review April 5, 2006 FLCC - Plasma 04/05/2006TRANSCRIPT
FLCC Feature-level Compensation & Control Plasma
Technology
April 5, 2006 A UC Discovery Project Plasma Technology FLCC
Workshop & Review April 5, 2006
Professors Jane P. Chang (UCLA), Michael A. Lieberman, David B.
Graves (UCB) and Allan J. Lichtenberg, John P. Verboncoeur, Zhuwen
Zhou, Sungjin Kim, Alan Wu, Emi Kawamura, Jon T. Gudmundsson,
Chengche Hsu, Joe Vegh, Insook Lee (UCB), and John Hoang (UCLA)
FLCC Workshop & Review April 5, 2006 FLCC - Plasma 04/05/2006
Dual/Triple Frequency Capacitive Discharge
for Dielectric Etch Coordinated research involving three PIs
Michael A. Lieberman (UCB) - Theory and kinetic (PIC-MCC)
simulations - Global model and experiment (supported by NSF &
Lam Research) David Graves (UCB) - Chemistry, plasma and neutral
transport, and transient effects - Fluid simulations (FEMLAB) and
molecular dynamicssimulations of fluorocarbon chemistries Jane P.
Chang (UCLA) - Profile evolution in SiO2, porous dielectrics,
high-k dielectrics - Feature scale simulations (DSMC) and
experiments (SEM) FLCC - Plasma 04/05/2006 Relationships Among the
Plasma Projects
Lieberman (Theory, PIC-MCC) Reactor-scale experiments Graves (Fluid
and MD) Reactor-scale experiments Surface-scale experiments
Electron energy deposition Ion and neutral fluxes Ion energy
distribution Dielectric etch molecular dynamics Chang (DSMC)
Feature-scale experiments Feature level profile evolution and
control FLCC - Plasma 04/05/2006 Plasma Sources for Feature Level
Compensation and Control
FLCC Workshop & Review April 5, 2006 Michael A. Lieberman,
Allan J. Lichtenberg, John P. Verboncoeur, Zhu-wen Zhou, Sungjin
Kim, Alan Wu, Emi Kawamura, Jon T. Gudmundsson UC Berkeley FLCC -
Plasma 04/05/2006 Summary of Research (Lieberman)
Develop kinetic simulation models of multiple frequency capacitive
discharge tools for dielectric etch and deposition Focus on
electron energy depositions and ion energy distributions FLCC -
Plasma 04/05/2006 Voltage Ions See and Predicted IEDF (Alan
Wu)
Using the inverse slope of the voltage-time graph, predict the
IEDF.Notice the peaks are around voltages where the slope of the
previous graph is nearly flat. FLCC - Plasma 04/05/2006 Dual
Frequency Sheath Model (Emi Kawamura)
Normalized Stochastic Heating Upper Bound High f controls n and ion
flux while low f controls: wh2/wl2 >>Vl/Vh>>1 Low f
coupling enhances by widening sheath and transporting oscillating
e- to lower sheath density and hence higher sheath velocity. Hl = a
normalized low f bulk oscillation amplitude. PIC result for
normalized stochastic heating agrees well with Kawamura et al
(2006) which uses a hard wall model which subtracts out the bulk
oscillation . FLCC - Plasma 04/05/2006 Comparison of global model
with experimental results (Sungjin Kim)
( experimental data from T. Kimura at Nagoya Institute of
Technology) By introducing effective volume of ionization and
energy losses, simulated results shows good agreement with the
experimental data. FLCC - Plasma 04/05/2006 Confined and unconfined
dischargewith instabilities at transition (Sungjin Kim)
Main discharge only (150 mTorr) 80 W absorbed High frequency (43.3
kHz)relaxation oscillation Main & peripheral discharge (280
mTorr) 202 W absorbed Low frequency (4.21 Hz) relaxation
oscillation FLCC - Plasma 04/05/2006 Future Milestones Perform
particle-in-cell simulations with dual and/or triple frequency
source power to determine ion energy distributions at substrate
Complete experiments and theory on instabilities in dual frequency
discharges FLCC - Plasma 04/05/2006 Plasma Sources for Feature
Level Compensation and Control
FLCC Workshop & Review April 5, 2006 David B. Graves, Chengche
Hsu, Joe Vegh and Insook Lee UC Berkeley FLCC - Plasma 04/05/2006
Summary of Research (Graves)
Develop fluid simulation models of multiple frequency capacitive
discharge and inductive tools for dielectric etch and deposition
Coupling neutral and plasma in the reactor-scale modeling for DPS
tools. Focus on chemical composition and plasma-surface
interactions FLCC - Plasma 04/05/2006 Reactor Scale Model
Experimental Validation
Experimental System Diagnostic ICP Multiple Diagnostics Fits 6-in
wafer Well-defined boundaries Stainless steel walls and dielectric
top plate Axisymmetric Model 2D, Fluid Model Ar/O2 ICP as the
preliminary test FemlabTM and MatlabTM Coupled neutral and plasma
model Easy to share / access Easy to extend to different
chemistries/systems FLCC - Plasma 04/05/2006 Equation System
Iteration Scheme Plasma Model Details Iteration scheme
Able to handle over 9 neutral species, and 8 charged species (22
equations) with PC (~1GB memory). 6 neutral ,4 ions species and 15
equations in Ar/O2 model. Convergence: one iteration <
20min.Need < 10 iterations. Robust, and easy to converge. Model
assumptions: Ambipolar, quasineutral, isothermal ions, and
Maxwellian electrons Model solves for: Neutrals: overall continuity
equation, mass balance equation for each species, momentum balance
and energy balance Plasma: ion continuity and electron energy. EM
field: Helmholtz Wave Equation FLCC - Plasma 04/05/2006 Modeling
DPS Tools (Chengche Hsu) 500W, Ar/O2:20/20sccm, 40mT.
Neutral velocity field ne plot: peak ne=2.3*1011 cm-3 FLCC - Plasma
04/05/2006 Neutral Flow Configuration: Top-fed and Side-fed
(Chengche Hsu) Gas feeding system directly impact the process.
Investigate the significance of neutral flow pattern and how it
influences the chemistry and the plasma characteristics.
Preliminary work for extending the current model to Dual flow/Duel
powerDPS II system. FLCC - Plasma 04/05/2006 Neutral Flow
Configuration: Top-fed and Side-fed
(Chengche Hsu) 500W, Ar/O2:20/20sccm, 40mT.O radial mass fraction
plots. Side-fed Top-fed Chemistry strongly affected by the feed gas
configuration. FLCC - Plasma 04/05/2006 MD Results (Joe Vegh) FC
SiC SiF Bulk Si
Sideviews of Si layers etched with CxFy/F/Ar+; demonstrates FC film
thickness fluctuations at surface FLCC - Plasma 04/05/2006 MD
Results (Joe Vegh) For all FC/F/Ar+ ratios examined, a significant
fraction of the C leaving the surface does so in clusters of 6 or
more C: implications for FC etch plasma chemistry models at tool
and feature scales FLCC - Plasma 04/05/2006 MD Results (Joe Vegh)
On a relative carbon weighted basis, the products show similar
distribution across all simulations. FLCC - Plasma 04/05/2006
Future Milestones Coupling the fluid model with multi-frequency
driven plasmas, including EM effects. Modeling DPS II tools:
capture the characteristics of dual flow and dual power system. Use
surface simulations to improve reactor scale and feature scale
models FLCC - Plasma 04/05/2006 Plasma Sources for Feature Level
Compensation and Control
Feature Profile Evolution during Shallow Trench Isolation (STI)
Etch in Chlorine-based Plasmas FLCC Workshop & Review April 5,
2006 Jane P. Chang and John Hoang UCLA FLCC - Plasma 04/05/2006
Summary of Research (Chang)
Feature Scale Modeling Combine accurate descriptions of plasma
fluxes to quantitatively predict the feature profile evolution
during etching/deposition processes Enable process development by
shortening experimental time and cost Feature scale model can be
coupled to tool scale (e.g. Prof. Graves, UCB) Feature scale model
can be coupled with PIC/MC model (Prof. Lieberman, UCB) Shallow
Trench Isolation (STI) An enabling technology over local oxidation
of Si (LOCOS) since the 0.18 m node A lower temperature process
avoiding annealing used for thermal oxidation A promising
technology for even smaller dimensions with properly developed
lithography, etch, and gap-fill technology FLCC - Plasma 04/05/2006
STI Process AMAT DPSII Reactor Cl2 N2 O2 Ws Wbias Coil Power
Substrate Bias Iouter Iinner Pressure Shallow trench isolation
(STI) replaced LOCOS since the 0.18 m node Rounded bottom corners
to minimize stress and allow void-free trench fill Positive trench
tapering of 75 to 89 to avoid a sharp corner in active silicon
Parameters examined for STI etch Chamber Pressure (mTorr) Source
Power (Ws) Wafer bias (Wbias) DC ratio = Iouter/Iinner Cl2 flowrate
(sccm) N2 flowrate (sccm) O2 flowrate (sccm) FLCC - Plasma
04/05/2006 Correlation between Process and Simulation
Parameters
Process Parameters Simulation Parameters Chamber Pressure (mTorr)
Ion Angle Distribution (IAD) Source Power (Ws) Ion Energy
Distribution (IED) Wafer bias (Wbias) Mean Ion Energy DC ratio =
Iouter/Iinner Cl Neutral to Ion Ratio Cl2 flowrate (sccm) N to Ion
Ratio (in development) N2 flowrate (sccm) O to Ion Ratio (in
development) O2 flowrate (sccm) E-Field lines (future plans?) Other
simulation parameters defined by elemental assignment of initial
profile Additional simulation parameters defined by different
plasma compositions FLCC - Plasma 04/05/2006 Monte Carlo Simulation
with Elemental Balance in Cells
Source plane + n Periodic boundary conditions Si N Mask (SiNx) O Cl
Silicon Si provides available sites Etchant such as Cl leads to the
formation of volatile products Reactant such as O leads to the
oxidation thus changing the etching characteristics Deposition of
SiCl2 and Si will add sites FLCC - Plasma 04/05/2006 Fractional
Factorial DOE for Si Etch
Pressure (mT) Ws (W) Wb (W) DC ratio Cl2 (sccm) N2 (sccm) O2 (sccm)
Mechanisms considered in simulation Chlorination: Sorption of
Chlorine ion: Ion-enhanced etching: SiCl2 Deposition: Oxygenation:
Sputtering: Sorption of sputtered Si: Recombination of chlorine: 7
factors, 2 levels, and 16 experiments Pressure (plasma density) and
DC ratio had statistically significant effects Need to quantify the
effect of oxygen addition Cho, H.S. et al. Mat. Sci. in Semi.
Process. 8 (2005) 239 Ulal, S.J et al. J. Vac. Sci. Technol. A
20(2) 2002 FLCC - Plasma 04/05/2006 Comparison of Simulation with
Experiments
Similar plasma densities Substrate bias governs the etch depth High
density versus low density plasmas Plasma composition controls
profile evolution Simulation on-going (significantly different
sidewall slope could be due to a change in plasma composition) FLCC
- Plasma 04/05/2006 Refined Surface Normal Determination
Previous Approach Current Approach - 82 218 348 Mask Silicon 5
cells included in normal computation Number of cells included
increase as long as R2 approaches unity The discrete nature of cell
surface representation results in artificial changes in surface
normal at a sloped surface and requires a refined approach bumps in
sloped side walls removed FLCC - Plasma 04/05/2006 Outcome of
Design of Experiments
Low density plasma High density plasma, with O2 in Cl2 With O2 in
Cl2 Without O2 in Cl2 Low DC ratio High DC ratio Without oxygen,
microtrenching formation and less tapering More hard mask erosion,
resulting in slight bowing High density plasma High density plasma,
with O2 in Cl2, low DC ratio With O2 in Cl2 Without O2 in Cl2 Low
substrate bias High substrate bias Without oxygen, much less
tapering Higher etch rate, more hard mask erosion, resulting in
slight bowing FLCC - Plasma 04/05/2006 Future Milestones Quantify
the effect of O2 addition to the etch profile evolution during STI
etch Predict feature profile evolution during STI etch and confirm
simulation with experimental measurements Special Acknowledgements:
Helena Stadniychuk and Andrey Zagrebelny at Cypress FLCC - Plasma
04/05/2006