Plasma Etching Theory Guide

Comprehensive Theory and Mathematical Foundations

1. Plasma Etching Fundamentals

1.1 Plasma Generation and Characterization

Plasma etching relies on the generation of partially ionized gas containing reactive species. The fundamental parameter describing plasma behavior is the degree of ionization, which typically ranges from 10^-6 to 10^-2 in processing plasmas. Understanding the basic physics requires knowledge of electron energy distribution functions (EEDF) and collision processes.

Equation 1.1: Plasma Frequency
ω_pe = √(n_e * e^2 / (ε_0 * m_e))

Where n_e is electron density, e is electron charge, ε_0 is permittivity of free space, m_e is electron mass

Equation 1.2: Debye Length
λ_D = √(ε_0 * k_B * T_e / (n_e * e^2))

Where k_B is Boltzmann constant, T_e is electron temperature

1.2 Sheath Physics

The plasma sheath is a critical region where ions are accelerated toward surfaces. The sheath voltage drop determines ion bombardment energy, which directly affects etching anisotropy and selectivity. The Child-Langmuir law describes current flow in the sheath region.

Equation 1.3: Bohm Criterion
v_i ≥ √(k_B * T_e / m_i)

Ions must enter the sheath with velocity exceeding the Bohm velocity

Equation 1.4: Sheath Thickness
s = λ_D * (2/3) * (2 * e * V_s / (k_B * T_e))^(3/4)

Where V_s is the sheath voltage

1.3 Power Coupling Mechanisms

RF power couples to the plasma through capacitive (CCP) or inductive (ICP) mechanisms. The power absorption determines plasma density and electron temperature, which in turn control radical and ion production rates.

Equation 1.5: Power Absorption (CCP)
P_abs = (1/2) * ω * ε_0 * ε'' * E_0^2 * V_plasma

Where ω is RF frequency, ε'' is imaginary part of permittivity, E_0 is electric field amplitude, V_plasma is plasma volume

Key Insight: Plasma Uniformity

Maintaining uniform plasma density across the wafer is crucial for uniform etching. The skin depth in ICP systems limits power penetration and affects uniformity. For CCP systems, edge effects and standing waves can create non-uniformities.

1.4 Electron Energy Distribution

The electron energy distribution function (EEDF) determines reaction rates through collision cross-sections. In most processing plasmas, the EEDF deviates from Maxwellian due to RF heating and transport effects.

Equation 1.6: Maxwellian EEDF
f(E) = 2π * n_e * (1/(π * k_B * T_e))^(3/2) * √E * exp(-E / (k_B * T_e))

Idealized Maxwellian distribution for electron energies

Equation 1.7: Reaction Rate Coefficient
k = ∫ σ(E) * v(E) * f(E) dE

Where σ(E) is collision cross-section, v(E) is electron velocity

1.5 Collisional Processes

Collisions between electrons and neutral species drive plasma chemistry. The collision frequency depends on electron density, neutral density, and collision cross-sections. Elastic collisions transfer energy while inelastic collisions create ions, radicals, and excited species.

Equation 1.8: Collision Frequency
ν_c = n_g * σ_c * v_e

Where σ_c is collision cross-section, v_e is electron velocity

Equation 1.9: Mean Free Path
λ_mfp = 1 / (n_g * σ_c)

Average distance traveled between collisions

1.6 Plasma Parameters

Key plasma parameters include electron density (10^10 - 10^12 cm^-3), electron temperature (2-10 eV), ion energy (20-500 eV), and neutral gas temperature (300-400 K). These parameters are controlled through pressure, power, and gas composition to achieve desired etching characteristics.

Parameter Symbol Typical Range Primary Control
Electron Density n_e 10^10 - 10^12 cm^-3 RF Power
Electron Temperature T_e 2 - 10 eV Pressure, Power
Ion Energy E_i 20 - 500 eV Bias Power
Pressure P 1 - 100 mTorr Flow/Pump Speed

2. Plasma Kinetics and Chemistry

2.1 Electron Impact Processes

Electron impact reactions are the primary mechanism for generating reactive species in plasma etching. These include ionization, dissociation, excitation, and attachment processes. The rates depend strongly on electron energy and gas composition.

Equation 2.1: Ionization Rate
R_ion = n_e * n_g * k_ion(T_e)

Where n_g is neutral gas density, k_ion is ionization rate coefficient

Equation 2.2: Dissociation Rate
R_diss = n_e * n_mol * k_diss(T_e)

Where n_mol is molecular density, k_diss is dissociation rate coefficient

2.2 Gas Phase Chemistry

Complex gas phase chemistry occurs in etching plasmas, involving hundreds of reactions. For fluorocarbon plasmas used in dielectric etching, polymerization competes with etching. The balance between these processes determines selectivity and profile control.

Equation 2.3: Species Conservation
∂n_i/∂t + ∇·Γ_i = S_i - L_i

Where Γ_i is flux, S_i is source term, L_i is loss term for species i

Equation 2.4: Radical Production Rate
R_rad = α * P_abs / (E_diss * V_plasma)

Where α is dissociation efficiency, E_diss is dissociation energy

2.3 Fluorocarbon Plasma Chemistry

Fluorocarbon gases (CF4, C4F8, CHF3) are widely used for etching SiO2 and Si3N4. The F/C ratio in the gas determines the balance between etching (F-rich) and polymer deposition (C-rich). Understanding this chemistry is essential for process optimization.

Equation 2.5: Polymer Formation Rate
R_poly = k_poly * [CF_x] * [surface_sites]

Where [CF_x] represents polymerizing radical concentration

Gas F/C Ratio Primary Use Etching/Deposition
CF4 4.0 Silicon etching Etching dominant
CHF3 3.0 Oxide etching Balanced
C4F8 2.0 DRIE, passivation Deposition dominant
C4F6 1.5 Polymer deposition Strong deposition

2.4 Halogen Chemistry

Chlorine and bromine chemistries are primarily used for silicon, metal, and III-V semiconductor etching. These halogens provide high etch rates but typically lower selectivity compared to fluorine-based chemistry.

Equation 2.6: Cl2 Dissociation
Cl2 + e^- → 2Cl + e^-

Primary dissociation mechanism producing reactive Cl atoms

Equation 2.7: Silicon Chloride Formation
Si + 4Cl → SiCl4 (volatile product)

Etch product formation and removal

2.5 Oxygen Addition Effects

Adding oxygen to fluorocarbon plasmas increases the F/C ratio by forming CO and CO2, promoting etching over deposition. Oxygen also affects photoresist erosion and can be used to clean polymers from chamber surfaces.

Equation 2.8: Oxygen Scavenging
CF_x + O → COF_y + F

Oxygen removes carbon from fluorocarbon radicals

Equation 2.9: Effective F/C Ratio
(F/C)_eff = (F/C)_gas + k_O * [O] / [CF_x]

Oxygen addition increases effective F/C ratio

2.6 Ion-Molecule Reactions

Ion-molecule reactions in the plasma create complex ion distributions. These reactions affect the composition of ions arriving at surfaces and influence etching chemistry. Charge transfer and dissociative recombination are important ion loss mechanisms.

Equation 2.10: Charge Transfer
A^+ + B → A + B^+

Ion charge transfers to neutral with lower ionization potential

3. Transport Phenomena

3.1 Neutral Species Transport

Neutral radical transport to the wafer surface occurs primarily through diffusion in low-pressure plasmas. The neutral mean free path is typically larger than reactor dimensions, leading to molecular flow conditions at very low pressures or transitional flow at moderate pressures.

Equation 3.1: Diffusion Coefficient
D = (1/3) * v_th * λ_mfp

Where v_th is thermal velocity, λ_mfp is mean free path

Equation 3.2: Neutral Flux (Diffusion)
Γ_n = -D * ∇n + n * v_gas

Combined diffusion and convection flux

3.2 Ion Transport

Ions are accelerated through the sheath by the electric field, achieving directed energy at the substrate. The ion energy distribution function (IEDF) and ion angular distribution function (IADF) critically affect etch profiles and damage.

Equation 3.3: Ion Energy at Substrate
E_i = (1/2) * m_i * v_i^2 + e * V_s

Ion energy from Bohm velocity plus sheath acceleration

Equation 3.4: Ion Flux
Γ_i = (1/4) * n_i * √(8 * k_B * T_i / (π * m_i))

Ion flux at sheath edge (thermal component)

3.3 Angular Distribution

The ion angular distribution affects sidewall etching and profile evolution. Collisions in the sheath scatter ions, broadening the angular distribution. At higher pressures, this leads to increased sidewall attack and profile bowing.

Equation 3.5: Angular Spread
θ_rms = √(2 * k_B * T_i / E_i)

RMS angular spread due to thermal motion

Equation 3.6: Collision Probability in Sheath
P_coll = 1 - exp(-s / λ_mfp)

Probability of ion-neutral collision in sheath of thickness s

3.4 Gas Flow and Pumping

Reactor pressure and residence time are controlled by the balance between gas flow rate and pumping speed. Proper gas flow distribution ensures uniform radical delivery and efficient removal of etch products.

Equation 3.7: Residence Time
τ_res = V_reactor / Q_pump

Where V_reactor is reactor volume, Q_pump is pumping speed

Equation 3.8: Pressure-Flow Relationship
P = (Q_gas * k_B * T) / Q_pump

Where Q_gas is gas flow rate in molecules/s

3.5 Knudsen Number Regimes

The Knudsen number (Kn = λ_mfp / L) determines the flow regime. Kn << 1 indicates continuum flow, Kn >> 1 indicates molecular flow, and intermediate values indicate transitional flow. Most plasma etching occurs in the transitional regime.

Equation 3.9: Knudsen Number
Kn = λ_mfp / L_characteristic

Ratio of mean free path to characteristic dimension

Equation 3.10: Molecular Flow Conductance
C_mol = (1/4) * A * v_th

Conductance in molecular flow regime

4. Surface Reactions

4.1 Adsorption Kinetics

Reactive species must first adsorb on the surface before reaction can occur. The coverage of adsorbed species depends on the balance between adsorption and desorption rates, described by Langmuir isotherm or more complex models for interacting adsorbates.

Equation 4.1: Langmuir Isotherm
θ = (K * P) / (1 + K * P)

Where θ is coverage, K is equilibrium constant, P is partial pressure

Equation 4.2: Adsorption Rate
R_ads = S * Γ * (1 - θ)

Where S is sticking coefficient, Γ is incident flux, θ is surface coverage

4.2 Ion-Enhanced Etching

The synergy between ions and neutrals is fundamental to plasma etching. Ion bombardment enhances chemical etching through several mechanisms: damage creation, breaking of surface bonds, heating, and enhanced desorption of products. This ion enhancement enables anisotropic etching.

Equation 4.3: Total Etch Rate
ER_total = ER_chem + ER_phys + ER_ion-enh

Chemical, physical, and ion-enhanced contributions

Equation 4.4: Ion Enhancement Factor
η = (ER_total - ER_chem) / ER_phys = f(E_i, Γ_n/Γ_i, θ)

Enhancement factor depends on ion energy, flux ratio, and coverage

4.3 Sputtering Mechanisms

Physical sputtering contributes to etching, especially at high ion energies or for materials with low chemical reactivity. The sputter yield depends on ion energy, mass, angle of incidence, and target material properties.

Equation 4.5: Sputter Yield
Y(E,θ) = Y_norm(E) * [cos(θ)]^(-f)

Where Y_norm is normal incidence yield, θ is angle from normal, f ≈ 1-2

Equation 4.6: Energy-Dependent Yield
Y(E) = Y_max * (E - E_th) / (E_opt - E_th) * exp[-(E - E_opt) / E_c]

Where E_th is threshold energy, E_opt is optimum energy, E_c is characteristic energy

4.4 Product Desorption

Etch products must desorb and be pumped away for etching to continue. Product desorption can be thermally activated, ion-enhanced, or limited by surface reaction. Volatile products desorb readily, while involatile products can inhibit etching.

Equation 4.7: Desorption Rate
R_des = ν * θ_p * exp(-E_des / (k_B * T_s))

Where ν is attempt frequency, θ_p is product coverage, E_des is desorption energy, T_s is surface temperature

Equation 4.8: Ion-Enhanced Desorption
R_des,ion = σ_des * Γ_i * θ_p

Where σ_des is ion-induced desorption cross-section

4.5 Surface Temperature Effects

Surface temperature affects adsorption, reaction rates, and desorption. Temperature control is crucial for process repeatability and selectivity. The wafer temperature results from the balance of heating (ion bombardment, recombination) and cooling (backside gas, conduction).

Equation 4.9: Power Balance
P_ion + P_rec + P_rxn = P_cond + P_rad

Ion heating, recombination, reaction heat = conduction + radiation cooling

Equation 4.10: Temperature-Dependent Rate
k(T) = A * exp(-E_a / (k_B * T))

Arrhenius temperature dependence of reaction rates

5. Profile Evolution

5.1 Feature Scale Modeling

Profile evolution during etching is determined by the local balance of species fluxes and surface reactions. Features evolve differently based on aspect ratio, charging effects, and transport limitations. String and cell-based methods are used to track moving boundaries.

Equation 5.1: Local Etch Rate
v_n = f(Γ_i(x), Γ_n(x), E_i(x), θ_i(x)) / ρ

Normal velocity depends on local fluxes, ion energy, coverage, and material density ρ

Equation 5.2: Surface Evolution
∂r/∂t = v_n * n̂

Where r is position vector, n̂ is surface normal

5.2 Aspect Ratio Dependent Etching (ARDE)

Etch rate decreases with increasing aspect ratio due to transport limitations of neutrals (neutral shadowing), ions (angular distribution), or charging effects. ARDE is also known as RIE lag and significantly affects uniformity across features of different sizes.

Equation 5.3: Neutral Transport Factor
F_n(AR) = 1 / (1 + AR/AR_c)

Where AR is aspect ratio, AR_c is characteristic aspect ratio for neutral transport

Equation 5.4: ARDE Factor
ER(AR) = ER_0 * F_n(AR) * F_i(AR) * F_c(AR)

Combined neutral, ion, and charging contributions to ARDE

5.3 Charging Effects

Dielectric surfaces charge due to imbalance between electron and ion currents. Charging deflects ions, causing profile distortion, notching, and twisting. Pulsed plasmas can reduce charging by allowing charge dissipation during the off-time.

Equation 5.5: Surface Potential
dV_s/dt = (Γ_i - Γ_e) * e / C_s

Where C_s is surface capacitance per unit area

Equation 5.6: Ion Deflection
Δθ = (e * V_s * L) / (2 * E_i * d)

Where L is feature length scale, d is sheath thickness

5.4 Microloading and Macroloading

Loading effects occur when etch rate depends on pattern density (microloading) or total open area (macroloading). These effects arise from local or global depletion of reactive species and affect across-wafer and within-die uniformity.

Equation 5.7: Microloading Factor
ER_local = ER_isolated / (1 + α * ρ_local)

Where ρ_local is local pattern density, α is loading coefficient

Equation 5.8: Radical Depletion
n(x) = n_0 * exp(-∫ k_etch * ρ(x') dx')

Exponential depletion of radicals across pattern

5.5 Selectivity and Mask Erosion

Selectivity between materials determines the achievable etch depth and profile control. Mask erosion reduces critical dimension and can limit maximum achievable aspect ratio. Selectivity depends on the ratio of etch rates and is typically enhanced by proper chemistry selection.

Equation 5.9: Selectivity
S = ER_target / ER_mask

Ratio of target material to mask material etch rates

Equation 5.10: Maximum Aspect Ratio
AR_max = S * (t_mask / CD_target)

Where t_mask is initial mask thickness, CD_target is target critical dimension

6. Plasma Damage Mechanisms

6.1 Physical Damage

Energetic ion bombardment causes lattice damage through atomic displacement and defect creation. Damage depth depends on ion energy and mass. For silicon, damage extends 5-20 nm for typical plasma etch conditions. This damage can affect device electrical properties if not properly removed.

Equation 6.1: Damage Depth
d_damage = λ * √(E_i / E_d)

Where λ is lattice constant, E_d is displacement energy (15-25 eV for Si)

Equation 6.2: Defect Density
N_defect = Y_d * Γ_i * t

Where Y_d is defect creation yield per ion, t is exposure time

6.2 Charging Damage

Antenna structures can collect charge during plasma processing, leading to gate oxide damage when current flows through thin dielectrics. The damage threshold depends on oxide thickness and quality. Modern processes use protection diodes or plasma doping to mitigate charging damage.

Equation 6.3: Antenna Ratio
AR = A_collector / A_gate

Ratio of charge collection area to gate area

Equation 6.4: Oxide Stress
E_ox = V_plasma / t_ox

Electric field across oxide of thickness t_ox

6.3 Contamination

Metal contamination from chamber materials or etch products can incorporate into devices. Mobile ions (Na, K) are particularly harmful to gate oxides. Chamber cleaning, proper material selection, and seasoning protocols minimize contamination.

Equation 6.5: Metal Incorporation Rate
R_cont = Γ_metal * S_metal * (1 - θ)

Where Γ_metal is metal flux, S_metal is sticking coefficient

6.4 UV and VUV Radiation

Plasma emits UV and VUV radiation from excited species. This radiation can damage photoresists, create interface states in oxides, and modify material properties. Radiation effects are especially important for advanced nodes with thin gate oxides.

Equation 6.6: Photon Flux
Φ_photon = β * n_e * n_i * k_rec

Where β is photon yield per recombination, k_rec is recombination coefficient

Equation 6.7: Interface Trap Generation
dN_it/dt = σ_UV * Φ_photon

Where σ_UV is UV damage cross-section, N_it is interface trap density

7. Advanced Topics

7.1 Atomic Layer Etching (ALE)

ALE provides atomic-scale control by separating the process into sequential self-limiting steps: surface modification and modified layer removal. This eliminates damage and enables ultra-precise etching for advanced nodes. Thermal and plasma ALE variants exist.

Equation 7.1: ALE Etch Per Cycle
EPC = d_modified * f_saturation

Where d_modified is modified layer thickness, f_saturation is removal fraction

Equation 7.2: ALE Saturation Condition
θ_sat = (S * τ_dose * Γ) / (1 + S * τ_dose * Γ * σ_site)

Surface saturation depends on dose time τ_dose and site density σ_site

7.2 Cryogenic Etching

Cryogenic etching uses very low temperatures (-100°C to -140°C) to enhance selectivity and reduce sidewall attack. At cryogenic temperatures, polymer sticking coefficients increase dramatically, and chemical reactions are suppressed. This enables highly anisotropic silicon etching.

Equation 7.3: Temperature-Dependent Sticking
S(T) = S_0 * exp(E_ads / (k_B * T))

Sticking coefficient increases exponentially at low temperature

Equation 7.4: Sidewall Passivation Rate
R_pass = Γ_polymer * S(T) * cos(θ)

Angular-dependent passivation enables anisotropy

7.3 Pulsed Plasma Etching

Pulsed plasmas modulate power at frequencies from Hz to MHz, allowing independent control of ion energy and flux. During the off-time, surfaces can cool, charge dissipates, and radicals decay. This enables new process windows and damage reduction.

Equation 7.5: Time-Averaged Ion Energy
E_i,avg = E_i,on * DC + E_i,off * (1 - DC)

Where DC is duty cycle (on-time / total time)

Equation 7.6: Radical Decay During Off-Time
n_rad(t) = n_0 * exp(-t / τ_rad)

Exponential decay with time constant τ_rad during pulse off-time

7.4 Multi-Frequency Processing

Applying multiple RF frequencies simultaneously allows independent control of plasma density and ion energy. Typically, high frequency (27-60 MHz) controls density while low frequency (2-13 MHz) controls bias. This provides additional process knobs for optimization.

Equation 7.7: Dual Frequency Sheath Voltage
V_dc = V_hf + V_lf

DC bias is superposition of high and low frequency components

Equation 7.8: Frequency-Dependent Skin Depth
δ = c / ω_pe = √(m_e * ε_0 / (n_e * e^2))

Skin depth determines power penetration in ICP plasmas

7.5 Machine Learning in Process Control

Modern plasma etching increasingly uses machine learning for endpoint detection, fault detection, and process optimization. Neural networks can identify subtle patterns in optical emission or impedance data that correlate with etch performance.

Equation 7.9: Neural Network Prediction
y = f(W_n * σ(W_{n-1} * ... σ(W_1 * x + b_1)... + b_{n-1}) + b_n)

Multi-layer neural network with weights W_i, biases b_i, activation σ

Future Directions

As semiconductor devices continue scaling, plasma etching faces new challenges: higher aspect ratios (greater than 100:1), atomic-scale precision, damage-free processing, and new materials (2D materials, wide bandgap semiconductors). Advanced techniques like ALE, cryogenic etching, and AI-driven process control will be essential for next-generation manufacturing.

7.6 Modeling and Simulation

Comprehensive plasma etching simulation requires coupling multiple physics domains: plasma kinetics, electromagnetic fields, neutral and charged particle transport, surface chemistry, and feature-scale profile evolution. Multi-scale modeling from reactor to feature scale enables predictive process development.

Equation 7.10: Coupled System
{∇·Γ_i + S_i - L_i = 0, ∇²φ = -e(n_i - n_e)/ε_0, ∂r/∂t = v_n(Γ_i, Γ_n)}

Coupled species conservation, Poisson equation, and surface evolution

Summary

This guide has covered the fundamental theory of plasma etching from basic plasma physics through surface reactions to advanced techniques. The key principles are:

  • Plasma generation creates reactive species through electron impact processes
  • Ion-enhanced etching provides anisotropy through synergy of ions and radicals
  • Transport phenomena determine species delivery to surfaces
  • Feature-scale effects (ARDE, charging, loading) affect profile evolution
  • Damage mechanisms must be understood and controlled
  • Advanced techniques (ALE, cryogenic, pulsed) enable new capabilities

Understanding these principles and their mathematical descriptions enables rational process development and troubleshooting for next-generation semiconductor manufacturing.

Back to Main Project