2D SF₆/O₂ Plasma Etching with SF6O2Etching

This tutorial runs a 2D physical plasma etching model using SF6O2Etching. It covers:

  1. Create a masked trench geometry
  2. Set physical units (required for physical models)
  3. Configure model parameters (fluxes, ion energy distribution)
  4. Configure advection and ray tracing settings
  5. Configure coverage steady-state initialization
  6. Enable logging and run the process
  7. Save results

Download Jupyter Notebook: 09_plasma_etching.ipynb

1. Import ViennaPS 

import viennaps as ps

2. Helper: Create a Masked Trench Domain

We use a helper function so the geometry setup stays in one place.

def createTrenchMask():
    extent = 30
    gridDelta = 0.3

    domain = ps.Domain(xExtent=extent, gridDelta=gridDelta)

    ps.MakeTrench(
        domain,
        trenchWidth=10.0,
        trenchDepth=0.0,
        maskHeight=10.0,
        maskTaperAngle=10
    ).apply()

    return domain

What this creates

  • 2D domain with width 30 and resolution 0.3
  • A trench opening of width 10
  • Flat bottom (trenchDepth=0.0) so the focus is on mask + surface evolution
  • A thick mask (maskHeight=10.0) with a small taper (maskTaperAngle=10°)

3. Set Units

Physical plasma models require consistent units.

ps.Length.setUnit("um")  # micrometers
ps.Time.setUnit("min")   # minutes

This means:

  • All geometry dimensions are interpreted in µm
  • All process durations are interpreted in minutes

4. Set Up SF6O2Etching Model Parameters

Start from the default parameter set and override the key values.

params = ps.SF6O2Etching.defaultParameters()

params.ionFlux = 10.0
params.etchantFlux = 4500.0
params.passivationFlux = 800.0

params.Ions.meanEnergy = 100.0
params.Ions.sigmaEnergy = 10.0
params.Ions.exponent = 500

model = ps.SF6O2Etching(params)

What these parameters mean (high level)

  • ionFlux: flux of ions that cause physical sputtering and ion-enhanced etching
  • etchantFlux: flux of reactive species that chemically etch the surface
  • passivationFlux: flux responsible for passivation

Ion energy distribution:

  • meanEnergy: average ion energy
  • sigmaEnergy: energy spread (Gaussian distribution width)
  • exponent: controls angular spread in the initial ion directions (higher values → more directional)

The resulting model computes local etch/passivation behavior based on:

  • ray-traced fluxes
  • local surface orientation
  • evolving surface coverages (reaction state)

5. Configure Advection and Ray Tracing Settings

Advection (surface evolution scheme) 

processParams = ps.AdvectionParameters()
processParams.integrationScheme = ps.IntegrationScheme.LOCAL_LAX_FRIEDRICHS_1ST_ORDER

This selects a robust first-order scheme for the level-set advection.

Ray tracing (flux sampling quality) 

rayTracingParams = ps.RayTracingParameters()
rayTracingParams.raysPerPoint = 1000
  • raysPerPoint controls Monte Carlo sampling noise vs. runtime
  • Higher values reduce noise but increase compute cost

6. Assemble the Process 

domain = createTrenchMask()
domain.saveVolumeMesh("SF6O2Etching_1")

process = ps.Process()
process.setDomain(domain)
process.setProcessModel(model)
process.setProcessDuration(10)  # 10 minutes
process.setParameters(processParams)
process.setParameters(rayTracingParams)

At this point you have:

  • geometry (domain)
  • physical etch model (model)
  • numerical settings (advection + ray tracing)
  • a fixed process duration of 10 min

7. Coverage Steady-State Initialization

Physical plasma models often include surface coverages (e.g., passivation fraction, reactive coverage). Before advancing the surface, the coverages are initialized to a steady state so reaction rates are stable.

coverageParams = ps.CoverageParameters()
coverageParams.tolerance = 1e-4
coverageParams.maxIterations = 10

process.setParameters(coverageParams)

Meaning

  • tolerance: convergence criterion for relative coverage change
  • maxIterations: hard cap on iterations if convergence is slow

So the solver will:

  • iterate coverages until changes are small enough, or
  • stop after 10 iterations even if the threshold is not reached

8. Enable Logging (Optional) 

ps.Logger.setLogLevel(ps.LogLevel.INTERMEDIATE)

This enables additional intermediate output during the run. Useful for debugging convergence and checking that the model is behaving as expected.

9. Run the Process and Save Result 

process.apply()

domain.saveVolumeMesh("SF6O2Etching_2")
domain.show()
  • process.apply() performs the etch simulation for the configured duration
  • saveVolumeMesh() exports the final state to VTU for visualization

Download Jupyter Notebook: 09_plasma_etching.ipynb