PREDICTABILITY IN LOW TEMPERATURE PLASMAS: FROM LABORATORY TO TECHNOLOGY* Mark J. Kushner University of Michigan Dept. of Electrical Engineering and Computer Science Ann Arbor, MI 48109 USA
[email protected] http://uigelz.eecs.umich.edu November 2008 * Work supported by NSF, SRC, Applied Materials, TEL Inc., 3M Inc., AFOSR
AGENDA • Low Temperature Plasma Science – Providing Societal Benefit • LTPS – Its role in the IT Infrastructure • Two Convergent Science Based Technologies • Electric Discharge Excimer Lasers • Plasma Materials Processing • Where to from here? Low Temperature Plasma Science Workshop
University of Michigan Institute for Plasma Science and Engineering
COLLISIONAL LOW TEMPERATURE PLASMAS • Low temperature plasmas are a power transfer medium. WALL PLUG ENERGETIC ELECTRONS POWER CONDITIONING
COLLISIONS WITH ATOMS/MOLECULES
e
PHOTONS RADICALS IONS
A
E ELECTRIC FIELDS
LAMPS LASERS ETCHING DEPOSITION
EXCITATION, IONIZATION, DISSOCIAITON (CHEMISTRY)
• Electrons transfer power from the "wall plug" to internal modes of atoms / molecules to "make a product”, very much like combustion. • The electrons are “hot” (several eV to 10 eV) while the gas and ions are cool, creating “non-equilibrium” plasmas. University of Michigan Institute for Plasma Science and Engineering
SOCIETAL BENEFIT – SCIENCE BASED TECHNOLOGIES • Thrusters
• Lighting
• Healthcare
• Microelectronics Processing
• Displays
• Jet Engine Spray Coatings
IMPACT OF LTPS ON DAILY LIFE…INCREDIBLE
• Ref: Plasma 2010 Decadal Study
University of Michigan Institute for Plasma Science and Engineering
PLASMA LIGHTING AND THE ENERGY ECONOMY
• Annual US electrical power consumption: 3.5 x 1012 kW-Hr • Electrical power expended in lighting: 22% (7.6 x 1011 kW-Hr) • Expended in fluorescent lamps: 9% (3.1 x 1011 kW-Hr) • 35 1-GWe power plants are used to excite a single multiplet of Hg states in fluorescent lamps. http://www.eia.doe.gov/cneaf/electricity/epa/epates.html http://antwrp.gsfc.nasa.gov/apod/ap970830.html http://www.eere.energy.gov/buildings/info/documents/ pdfs/lmc_vol1_final.pdf
University of Michigan Institute for Plasma Science and Engineering
PLASMA LIGHTING AND THE ENERGY ECONOMY • Conversion of incandescent to plasma lighting would save 2.2 x 1011 kW-Hr a year…the equivalent of 24 1-GWe power plants.
• GE-A19 • GE-T3 Plasma • GE BD-17 Incandescent Fluorescent High Intensity 17 Lumens/W 52 Lumens/W Discharge
87 Lumens/W • Optimizing the electron f(ε) in plasma lighting by 0.1 eV translates into three 1-GWe plants….and this has been done. • This is an incredible accomplishment and mastery of discharge physics. http://www.gelighting.com/na/
University of Michigan Institute for Plasma Science and Engineering
r r ∂f qE ⎛ ∂f ⎞ = −v ⋅ ∇ x f − ⋅ ∇v f − ⎜ ⎟ ∂t m ⎝ ∂t ⎠ collisions • Boltzmann’s Equation
ON THE GROUND DISCHARGE PHYSICS • Optimizing Ar/Hg fluorescent lamps.
• Electron Energy Distributions
• Reaction mechanism
• Cross Sections
• Power Flow to Excited States University of Michigan Institute for Plasma Science and Engineering
LTPS – A VERY DIVERSE FIELD • The diversity of field makes leveraging advance in science to produce society benefiting technologies challenging. • Microplasma arrays (Ref: J. G. Eden)
• Applied Materials PECVD for LCD panels and solar cells.
• Example - Size: Large, stable plasmas (5 m2 plasmas) for LCD television panels to tiny (100 µm2) plasmas so intense that the plasma electrons merge with solid electrodes. University of Michigan Institute for Plasma Science and Engineering
• Excimer Laser Lithography
• Roadrunner Supercomputer
• Plasma Materials Processing
TWO LTP TECHNOLOGIES HAVE ENABLED THE WORLDWIDE IT INFRASTRUCTURE • Electric discharge excited excimer lasers… • RF discharge plasma etching, deposition and sputtering systems… • Two low temperature plasma systems singlehandedly responsible for the worldwide information technology infrastructure. • Moore’s law would not exist in the absence of LTPs.
• www.intel.com University of Michigan Institute for Plasma Science and Engineering
TWO LTP TECHNOLOGIES HAVE ENABLED THE WORLDWIDE IT INFRASTRUCTURE
• ….And it can all be traced to our mastering the application of Boltzmann’s equation to technological plasmas. • Understanding the coupling of electron and ion velocity distributions to photon-generation and radical production have created our high technology infrastructure.
University of Michigan Institute for Plasma Science and Engineering
CHAPTER 1 – ELECTRIC DISCHARGE EXCIMER LASERS FOR PHOTOLITHOGRAPHY
ELECTRIC DISCHARGE EXCIMER LASERS FOR PHOTOLITHOGRAPHY • Microelectronics features are defined by photolithography – transferring a pattern from a mask to the silicon wafer. • Feature sizes are limited by the wavelength - excimer UV lasers have enabled sub 0.1 µm features
• www.nature.com University of Michigan Institute for Plasma Science and Engineering
EXCIMER LASERS FOR PHOTOLITHOGRAPHY • …A triumph of applying incredibly complex plasma chemistry to development of 24/7 dependable technology.
University of Michigan Institute for Plasma Science and Engineering
TIGHT COUPLING OF PRODUCTS AND f(ε) • Tight coupling between atomic-molecular properties, ionization fraction and excitation rates provides the means to customize plasmas to produce desire end products…this was known early on. r r r r r ∂f (v , r , t ) = −ν ⋅ ∇ x f (v , r , t ) ∂t r r r r r ⎛ ∂f (v , r , t ) ⎞ − a ⋅ ∇ v f (v , r , t ) + ⎜ ⎟ r ∂t ⎝ ⎠ collisions ∂N (r , t ) = ∂t r r r − ∑ ∫ f (v , r , t )ν σ ij (v )N j N i d 3v i, j
University of Michigan Institute for Plasma Science and Engineering
EARLY WORK ON BOLTZMANN’S EQUATION • Early approximate methods for solving Boltzmann’s equation involved 2-term spherical harmonic expansions. • First introduced by Morse, Allis and Lamar (1935) and popularized by Holstein (1946) in context of low pressure plasmas.
r r r f (v , r ,t ) = ∑ f i (v , r ,t )Pi (cos(θ )) i
r r ≈ f 0 (v , r ,t ) + f i (v , r ,t ) cos(θ )
• Deviation from Maxwellian is in part the anisotropy of f(v). • Capture anisotropy in spherical harmonic expansion (SHE).
University of Michigan Institute for Plasma Science and Engineering
IMPROVEMENTS IN EXPANSION THEORY • In absence of statistical methods (e.g., Monte Carlo simulations), SHE of Boltzmann’s equation used for non-isotropic transport. • Higher order terms enabled subtleties to be revealed and application to highly nonequilibrium situations.
University of Michigan Institute for Plasma Science and Engineering
LEVERAGING f(ε) TO POWER FLOW • Leveraging connection between f(ε) and power flow to specific excited states enabled a revolution in technology development. • …Fostered by numerical solutions of Boltzmann’s equation.
• EEDs in CO2 vs E/N Nighan, PRA 2, 1989 (1970) University of Michigan Institute for Plasma Science and Engineering
OPTIMIZING f(ε) FOR LASER TECHNOLOGIES • These works provided fundamental guidance for developing highly efficient discharge excited lasers by optimizing f(ε) as the excited state manifold evolved. • Became clear that optimizing rates of excitation of CO2(v) was incompatible with self sustained discharges. • Motivated development e-beam sustained discharges.
• EEDs in CO2 vs T(vib)
• Fractional energy in CO2 vs E/N.
• Nighan, PRA 2, 1989 (1970)
University of Michigan Institute for Plasma Science and Engineering
E-BEAM SUSTAINED CO(v) LASER - INSTABILITIES
• E-beam sustained discharges optimized f(ε) to excite CO(v) while providing background ionization. Multi-kJ pulses were realized – but instabilities terminated the pulses. University of Michigan Institute for Plasma Science and Engineering
BOLTZMANN KINETICS AND DIAGNOSTICS LEAD TO SCIENCE BASED DESIGN • High power laser development was put on science basis by experimentally tracking power flow through excited states.
University of Michigan Institute for Plasma Science and Engineering
BOLTZMANN KINETICS AND DIAGNOSTICS LEAD TO SCIENCE BASED DESIGN • Complex, validated, design capable computer models were developed based on Boltzmann kinetics.
University of Michigan Institute for Plasma Science and Engineering
GLOW-TO-ARC TRANSITIONS TERMINATE LASER • Streamers originating from cathode fall perturbations imprinted by low pre-ionization densities were diagnosed…..
University of Michigan Institute for Plasma Science and Engineering
REMEDYING DISCHARGE INSTABILITIES IN EXCIMER LASERS • Kinetics were understood but discharge instabilities prematurely terminated laser pulses…Realization of importance of pre-ionization revolutionized the field.
University of Michigan Institute for Plasma Science and Engineering
MODEL BASED SCALING LAWS • Models using Boltzmann analysis verified need for critically large pre-ionization density to prevent micro-arcs.
University of Michigan Institute for Plasma Science and Engineering
TECHNOLOGY’S RESPONSE • An entire new generation of laser technology was developed built upon the fundamental understanding of micro-arc generation – and how to prevent it – using x-ray preionization.
University of Michigan Institute for Plasma Science and Engineering
INDUSTRY QUALIFIED ArF EXCIMER LASER PHOTOLITHOGRAPHY – NEXT GENERATION EUV • ArF discharge lasers enable 32 nm photolithography. • Discharge laser produced plasmas are basis for next generation EUV sources.
• Cymer 50 W EUV demonstration – Semi. Intl., Nov. 2008
• Cymer, Inc.
University of Michigan Institute for Plasma Science and Engineering
CHAPTER 2 – PLASMA MATERIALS PROCESSING
PLASMAS FOR NANOSCALE FABRICATION • Plasmas are and will continue to be indispensable for etching, deposition and cleaning in microelectronics fabrication. • Control of dimensions at 22 and 35 nm nodes requires resolution of a few nm or less.
• Required: Unprecedented control of reactant fluxes from the plasma onto the wafer: Uniformity, Composition, Energies
• http://www.intel.com http://realworldtech.com
University of Michigan Institute for Plasma Science and Engineering
ACTIVATION ENERGY: SUB-eV, SUB-DEGREE CONTROL • Activation energy is largely delivered through ion bombardment. • Distinguishing between materials will be determined by sub-eV and sub-degree control of ion energies.
• …and with excimer laser photolithography
• Intel Fin-FET
University of Michigan Institute for Plasma Science and Engineering
MULTI-FREQUENCY CAPACITIVELY COUPLED PLASMA ETCHER:
APPLIED MATERIALS CENTURA ENABLER • Plasma etching of dielectric materials for logic contacts and interconnect – 300 mm wafers at the 45 nm node.
• Ref: S. Rauf, AMAT
University of Michigan Institute for Plasma Science and Engineering
TYPICAL PLASMA ETCHING REACTOR
• Hitachi XT ECR http://www.hitachi-hta.com
University of Michigan Institute for Plasma Science and Engineering
ION NEUTRAL SYNERGISM • The basis of plasma etching is the synergism between neutral surface chemistry and ion activation.
• Separately controlling composition and energies of radicals and ions will enable precise feature evolution. University of Michigan Institute for Plasma Science and Engineering
SELECTIVITY IN PLASMA ETCHING mainetch_ied
• Fabricating microelectronics devices requires preferential etching a material – selectivity. • Selective etching occurs by controlling radical fluxes and the ion energy and angular distribution (IEAD) to wafer. • Sheath physics will dominate
University of Michigan Institute for Plasma Science and Engineering
EARLY PROGRESS IN SHEATH DYNAMICS • With the realization that sheath physics would dominate the ability to selectively etch material, early work addressed their dynamics.
• Analytic approaches provided keen insights into scaling. • Analytic models were later incorporated into large scale computer models (which did not resolve sheaths) as “jump” boundary conditions.
University of Michigan Institute for Plasma Science and Engineering
DIAGNOSTICS: SOPHISTICATED PROBES • The rf plasma environment probe measurements difficult. Mastery of the technique enabled in depth analysis of rf discharge kinetics.
LASER ELECTRIC FIELD MEASUREMENTS • Stark shift laser-induced-fluorescence spectroscopy of increasing greater sophistication enabled measurements of sheath properties.
University of Michigan Institute for Plasma Science and Engineering
80 mTorr SF6, 200 W
Ref: E. Aydil
Ref: R. Dorai
Ref: W. Holber
DATA BASES OF ATOMIC AND MOLECULAR PROPERTIES • The allied science areas [AMO (Atomic, Molecular, Optical), Surface Science] were and continue to be critical in developing knowledge bases of fundamental parameters. • Many of the diverse molecular gases relevant to the field had no prior database. • At first stifled industrially relevant modeling – later remedied by improved databases.
University of Michigan Institute for Plasma Science and Engineering
MODELING AND SIMULATION – FIRST PROVIDED INSIGHTS • The computationally harsh environment of rf discharges in complex gas mixtures at first greatly challenged the community. • Early efforts addressed fundamental transport properties, and transition between stochastic and resistive heating.
University of Michigan Institute for Plasma Science and Engineering
ANOMALOUS SKIN EFFECT AND POWER DEPOSITION IN ICP • Collisional heating:
λ mfp < δ skin ,
r r r r r J e ( r , t ) = σ (r , t )E (r , t )
• Anomalous skin effect:
r v r λ mfp > δ skin , F = v × B r r r r r r r J e ( r ,t ) = ∫∫ σ (r , r ' ,t ,t' )E (r ' ,t' )dr ' dt'
• Regions of time averaged positive and negative power deposition; and nonmonotonic E-field
University of Michigan Institute for Plasma Science and Engineering
ADVANCES IN MODELING CAPTURE ANOMALOUS BEHAVIOR • E-field
• Power
• Ionization
• 2-d models with non-local, kinetic solutions of Boltzmann’s equation coupled to electromagnetics capture experimentally observed anomalous sheath behavior – and show importance axial magnetic forces.
ANIMATION SLIDE
• Ar, 10 mTorr, 7 MHz, 100 W
University of Michigan Institute for Plasma Science and Engineering
FINITE WAVELENGTH EFFECTS
• With increasing wafer size and rf frequency, and plasma shortened wavelengths CCPs become inductive with finite wavelengths.
Design Capable Models: Effect of B Field on [e] – 13.5, 162 MHz Electrostatic edge effects dominate at low frequency while at high frequency plasma is center peaked due to the standing electromagnetic wave. HF w/B-Field: E×B drift shears plasma in opposite directions (top, bottom). LF w/B-Field: Asymmetry produced dc bias produces drift in one direction.
S. Rauf, J. Kenney, and K. Collins, Applied Materials AVS Symposium, Nov. 2008 Applied Materials External
• Science based design of plasma processing tools: • Diagnostics • Modeling • Experience….
Refs: AMAT, AIST, Japan; Intel
University of Michigan Institute for Plasma Science and Engineering
A CONVERGENCE OF SCIENCE BASED TECHNOLOGIES • Excimer Discharge Laser Lithography
• Plasma Materials Processing
• Could Boltzmann have predicted that conservation of fluxes in phase space would one day produce supercomputers?
r r r r r ∂f (v , r , t ) ( = −ν ⋅ ∇ x f v , r , t ) ∂t r r r r r ⎛ ∂f (v , r , t ) ⎞ − a ⋅ ∇ v f (v , r , t ) + ⎜ ⎟ ∂t ⎝ ⎠ cols
LOW TEMPERATURE PLASMA SCIENCE WORKSHOP • Convened by DOE-Office of Fusion Energy Science. • Summarize status of research in LTPS. • Identify outstanding major scientific questions. • Articulate their importance – science and relation to technology. • Describe basic research to address questions. • Develop a prioritized roadmap. • Report: Low Temperature Plasma Science: Not only the Fourth State of Matter but All of Them • Published: September 2008 University of Michigan Institute for Plasma Science and Engineering
LTPS – PRIORITIES • 1 – Predictive Control of Plasma Kinetics • Plasma kinetics underlie the means of utilizing LTPs and the generation of chemically reactive species. • Crafting and controlling the velocity distributions electrons and ions are key to optimizing the end product. • 2 – Collective Behavior and Non-linear Transport • LTPs produce unique collective behavior & nonlinear transport. • With a broad array of positive and negative ions there is a rich possibility of waves and instabilities. • 3 - Interfaces and Multiple Phases in Plasmas • LTPs uniquely interact with all phases: solid, liquid and gas. • Plasmas in liquids are now surgical instruments while low pressure plasmas create nano-crystals of unique composition. University of Michigan Institute for Plasma Science and Engineering