PLASMA ETCHING OF EXTREMELY HIGH ASPECT RATIO FEATURES

PLASMA ETCHING OF EXTREMELY HIGH ASPECT RATIO FEATURES: TWISTING EFFECTS* Mingmei Wanga), Ankur Agarwalb), Yanga) and Mark J. Kushnera) a)Iowa State University, Ames, IA 50011, USA mmwang@iastate. edu mjk@iastate. edu b)University of Illinois, Urbana, IL 61801, USA http: //uigelz. ece. iastate. edu 60 th Gaseous Electronics Conference, October 2007 *Work supported by Micron Technology Inc. , SRC and NSF

AGENDA · High Aspect Ratio Contact (HARC) Etching · Approach and Methodology · Charging of features · Fluorocarbon etching of HARC · Si. O 2 -over-Si etching · Potential · Effect of open field · Concluding Remarks MINGMEI_GEC 07_AGENDA Iowa State University Optical and Discharge Physics

HARC ETCHING: ISSUES · As aspect ratio (AR) of features increases, complexity of plasma etching increases. · Aspect Ratio Dependent Etching · Etch rate decreases with increasing AR. · Charging of features due to ion and electron bombardment. · Electric field variations affect ion trajectories; deviation from ideal profile. · Non-uniform ion flux despite uniform bulk plasma. · As AR increases, the cross-sectional area of each via is smaller. · Increasingly random nature of incident ions and radicals. MINGMEI_GEC 07_01 Ref: Micron Technology, Inc. Iowa State University Optical and Discharge Physics

OBJECTIVES AND APPROACH · Objectives · Computationally investigate consequences of charging of high aspect ratio features in Si. O 2. · Approach · Reactor scale: Hybrid Plasma Equipment Model. · Feature scale: Monte Carlo Feature Profile Model. · Poisson’s equation is solved for electric potentials. · Acceleration of ions and electrons due to electric fields in feature. · Dissipation of charge through material conductivity. MINGMEI_GEC 07_02 Iowa State University Optical and Discharge Physics

HYBRID PLASMA EQUIPMENT MODEL (HPEM) · Electromagnetics Module: Antenna generated electric and magnetic fields · Electron Energy Transport Module: Beam and bulk generated sources and transport coefficients. · Fluid Kinetics Module: Electron and Heavy Particle Transport, Poisson’s equation · Plasma Chemistry Monte Carlo Module: · Ion and Neutral Energy and Angular Distributions · Fluxes for feature profile model MINGMEI_GEC 07_04 Iowa State University Optical and Discharge Physics

MONTE CARLO FEATURE PROFILE MODEL · Monte Carlo techniques address plasma surface interactions and evolution of surface morphology and profiles. · Inputs: · Initial material mesh · Surface reaction mechanism · Ion and neutral energy and angular distributions. · Ion and radical fluxes at selected wafer locations. · Maxwellian electron fluxes with Lambertian distribution · Fluxes and distributions from HPEM. MINGMEI_GEC 07_05 Iowa State University Optical and Discharge Physics

MCFPM: CHARGING ALGORITHMS · The electric potential is solved using the method of Successive Over Relaxation (SOR). · Large mesh sizes pose computational challenges to solve for potential after launch of each particle. · Electric field is being updated after the launch of every 30 charged particles. Charged particle Mask + Si. O 2 · Particles are a few nm on a side. · Total particles launched (ions and radicals): 150, 000 -300, 000. · The charge of pseudo-particles mesh is adjusted to account for finite sized particles. MINGMEI_GEC 07_03 - - + + + + + Si Iowa State University Optical and Discharge Physics

FLUOROCARBON PLASMA ETCHING OF Si. O 2/Si · CFx radicals produce polymeric passivation layers which regulate delivery of precursors and activation energy. · Chemisorption of CFx produces a complex at the oxide-polymer interface · Low energy ion activation of the complex produces polymer. · Polymer complex sputtered by energetic ions etching. · As Si. O 2 consumes the polymer, thicker layers on Si slow etch rates enabling selectivity. MINGMEI_GEC 07_06 Iowa State University Optical and Discharge Physics

FLUOROCARBON ETCH OF HARC · Dual frequency capacitivelycoupled (CCP) reactor geometry. · Base conditions: · Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm · 40 m. Torr · 500 W at 25 MHz · 4000 W at 10 MHz · Low frequency: Substrate High Frequency: Showerhead MINGMEI_GEC 07_07 Iowa State University Optical and Discharge Physics

REACTANT FLUXES · 10 m. Torr, HF 500 W, LF 4 k. W, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm · Dominant Ions: Ar+, CF 2+, C 2 F 4+, CF+ · Dominant Neutrals: CF 2, C 2 F 4, CF 3, F · Polymer clearing fluxes · O = 3 1016 cm-2. s-1 · O+ = 3 1014 cm-2. s-1 MINGMEI_GEC 07_08 Iowa State University Optical and Discharge Physics

ION ENERGY ANGULAR DISTRIBUTIONS (IEADs) · IEADs for sum of all ions. · Peak in ion energy increase with increasing bias power. · High ion energies required for etching of HAR features. · Narrow angular distribution reduce sidewall impacts. · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, LF 10 MHz, HF 500 W. MINGMEI_GEC 07_09 Iowa State University Optical and Discharge Physics

Si. O 2 -over-Si HARC ETCH: NO CHARGING · Etch profile evolution without charging. · Etch rate higher at higher bias powers owing to high ion energies. · No charging: · Generally straight profiles. · High ion energies low polymer coverages. · Some evidence of randomness due to small contact area · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, 10 MHz, HF 500 W. Aspect Ratio = 1: 10 MINGMEI_GEC 07_10 Iowa State University Optical and Discharge Physics

Si. O 2 -over-Si HARC ETCH: EFFECT OF CHARGING · Charging effects are considered: · Charge buildup on polymer affects plasma potential. · Ion trajectories influenced by electric-field. · Electrons neutralize charge deep in trench. · Lower ion energies (due to buildup of charge) · Lower etch rates. · Deviation from “ideal” anisotropic etch profiles. · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, 10 MHz, HF 500 W. Animation Slide Aspect Ratio = 1: 10 MINGMEI_GEC 07_11 a Iowa State University Optical and Discharge Physics

Si. O 2 -over-Si HARC ETCH: EFFECT OF CHARGING · Charging effects are considered: · Charge buildup on polymer affects plasma potential. · Ion trajectories influenced by electric-field. · Electrons neutralize charge deep in trench. · Lower ion energies (due to buildup of charge) · Lower etch rates. · Deviation from “ideal” anisotropic etch profiles. · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, 10 MHz, HF 500 W. Aspect Ratio = 1: 10 MINGMEI_GEC 07_11 b Iowa State University Optical and Discharge Physics

Si. O 2/Si HARC ETCH: PLASMA POTENTIAL Max Min 213 V 116 V 151 V 110 V · Charge deposition on polymer affects plasma potential. -5 -3 -6 -4 · Small depths: · Electrons effectively neutralize charge buildup. · Potential essentially maintained at zero. · Large depths: · Trapping of charge in polymer perturbs ion trajectories. · Electrons are “pulled” into bottom of trench by large positive potential and neutralizes. AR = 1: 10 Increasing Power MINGMEI_GEC 07_12 a Animation Slide -6 0 213 Iowa State University Optical and Discharge Physics

Si. O 2/Si HARC ETCH: PLASMA POTENTIAL Max Min 213 V 116 V 151 V 110 V · Charge deposition on polymer affects plasma potential. -5 -3 -6 -4 · Small depths: · Electrons effectively neutralize charge buildup. · Potential essentially maintained at zero. · Large depths: · Trapping of charge in polymer perturbs ion trajectories. · Electrons are “pulled” into bottom of trench by large positive potential and neutralizes. AR = 1: 10 Increasing Power MINGMEI_GEC 07_12 b -6 0 213 Iowa State University Optical and Discharge Physics

Si. O 2 -over-Si HARC ETCH: RANDOMNESS? · Monte Carlo modeling utilizes random number generator to simulate a physical process. · Different seed numbers · All other conditions are same. · Is it reproducible? · No charging effects: · Etch profiles vary little · Anisotropic etch · No anomalies observed · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, LF 4 k. W, HF 500 W. Different seed numbers Aspect Ratio = 1: 10 MINGMEI_GEC 07_13 Iowa State University Optical and Discharge Physics

Si. O 2/Si HARC ETCH: RANDOMNESS OF CHARGING? · Different seed numbers · All other conditions are same. · Is it reproducible? · Charging effects: · Stochastic nature of incident ion fluxes reflected in profiles. · Twisting observed · Etch direction shifts which reinforces anomoly. · Some unphysical behavior also observed (last trench) Different seed numbers · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, LF 4 k. W, HF 500 W. Aspect Ratio = 1: 10 MINGMEI_GEC 07_14 Iowa State University Optical and Discharge Physics

Si. O 2/Si HARC ETCH: RANDOMNESS OF CHARGING? · 6 Trenches receiving “same fluxes. · Stochastic nature of fluxes produces random twisting. · Similar behavior observed experimentally. · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, LF 4 k. W, HF 500 W. Ref: Micron Technology, Inc. Aspect Ratio = 1: 10 MINGMEI_GEC 07_15 Iowa State University Optical and Discharge Physics

EFFECT OF OPEN FIELD: NO CHARGING · 4 trenches followed by a “plasma-only” region with hard mask. · Open field has large sidewall polymerization. · Charging not considered · Trenches have some randomness in profiles owing to non-uniform ion fluxes. · No effect due to plasma-only region. Aspect Ratio = 1: 10 MINGMEI_GEC 07_16 a Animation Slide Iowa State University Optical and Discharge Physics

EFFECT OF OPEN FIELD: NO CHARGING · 4 trenches followed by a “plasma-only” region with hard mask. I II · Open field has large sidewall polymerization. · Charging not considered · Trenches have some randomness in profiles owing to non-uniform ion fluxes. III Aspect Ratio = 1: 10 MINGMEI_GEC 07_16 b IV · No effect due to plasma-only region. Iowa State University Optical and Discharge Physics

EFFECT OF OPEN FIELD: EFFECT OF CHARGING · Open field can impact etch of adjacent trenches by trapping of charge in polymer. · Transverse electric fields from external charge significantly affects adjacent trenches. · Inner trenches less affected by charging. · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, LF 4 k. W, HF 500 W. Animation Slide Aspect Ratio = 1: 10 MINGMEI_GEC 07_17 a Iowa State University Optical and Discharge Physics

EFFECT OF OPEN FIELD: EFFECT OF CHARGING I II · Open field can impact etch of adjacent trenches by trapping of charge in polymer. · Transverse electric fields from external charge significantly affects adjacent trenches. · Inner trenches less affected by charging. · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, III Aspect Ratio = 1: 10 MINGMEI_GEC 07_17 b IV 300 sccm, LF 4 k. W, HF 500 W. Iowa State University Optical and Discharge Physics

OPEN FIELD EFFECT ON CHARGING · Open field impacts adjacent trenches by transverse electric field from trapped charged in polymer. I · Isolating open field by making spacer of Si. O 2 thicker reduces transverse fields and perturbation of etch profiles. · Smaller deviation for the adjacent trench. II · Note effect of stochastic ion fluxes in second trench. · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, LF 4 k. W, HF 500 W. MINGMEI_GEC 07_18 Aspect Ratio = 1: 10 Iowa State University Optical and Discharge Physics

COMPUTATIONAL ASPECTS: DISSIPATION OF CHARGE · Dissipation of charge accounted for through material conductivity · I: Static charge · II: Only electron charges move · III: Both ion and electron charges move · Positive charges inside materials leads to high potentials inside the trench · Lower ion energies polymer deposition · Etch stop observed · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, LF 4 k. W, HF 500 W. MINGMEI_GEC 07_20 Aspect Ratio = 1: 10 Iowa State University Optical and Discharge Physics

ELECTRIC FIELDS: BOUNDARY CONDITIONS · Boundary conditions for Poisson’s equation: Zero potential at mesh boundaries. · Both electron and ion charges move · Small mesh: · Unphysical high gradients in fields · Leads to etch stops · Wide mesh: · Gradients in fields relaxed · Etch progresses to completion · Higher conductivity less effect of charging · 10 m. Torr, Ar/C 4 F 8/O 2 = 80/15/5, 300 sccm, LF 4 k. W, HF 500 W. MINGMEI_GEC 07_21 Aspect Ratio = 1: 10 Iowa State University Optical and Discharge Physics

CONCLUDING REMARKS · Etching of high aspect ratio contacts (HARC) has been computationally investigated in fluorocarbon plasma. · Charging of features has been included to investigate anomalies such as twisting observed during etching of HARCs. · Charge buildup in/on polymer layer decreases etch rates and deviates the etching profile. · Ultimately a stochastic process for small features. · Various factors affect etching profiles: · Special structures like open field. · High energy ions may mitigate the effect of charging. · Charge dissipation due to material conductivity. MINGMEI_GEC 07_22 Iowa State University Optical and Discharge Physics