Particle-in-Cell Monte-Carlo Simulation

A parallel Particle-in-Cell Monte-Carlo (PIC-MC) simulation software was developed and implemented at Fraunhofer IST, supporting the modelling of gas flows and gas discharges in all system geometries with spatial and time resolution. The PIC-MC simulation method is based on reproducing the movements and collisions of representative simulation particles and is particularly well suited for the low-pressure range. Various wall types such as pumps, chemically active surfaces, particle sources and so on are available for taking wall collisions into account. This permits the realistic representation of complex systems with their external, physical parameters in the simulation model.

Relevant results include:

  • Pressure and charge density distributions
  • Electrical potential
  • Particle flow densities
  • Energy distribution on the substrate

These values are determined by means of ensemble and time averaging across the simulation particles, and reflect the growth conditions in detail depending on the process parameters.

Sample application I: Optimization of magnetron targets

Simulation electron distribution
© Fraunhofer IST

Local plasma density aggregations that occur primarily in the proximity of linear magnetron targets can significantly reduce the lifespan of the targets. Such effects are usually compensated by adjusting the magnetron and/or the magnetic enclosure. This applies correspondingly for optimizing the target utilisation, for example by widening the target racetrack.

Iterations of the BEM magnetic field calculation and PIC-MC simulation were performed in order to optimize planar magnetron targets without intermediate experimental steps. Various magnetron configurations were calculated for this purpose with BEM, checking whether the magnetic field met the required specifications. Three-dimensional PIC-MC simulations were performed with the corresponding magnetic fields in the next step. From the stationary state of the simulation, the flow of ions onto the target surface was extracted, permitting conclusions to be drawn about the empirical erosion profile. This enabled the validation and targeted further development of optimization approaches. After multiple (numerous) iterations, a magnetron configuration was ultimately found that exhibited a homogeneous erosion profile across the entire target racetrack when used for industrial applications.

Sample application II: Investigation of film defects in DC sputtering

© Fraunhofer IST

Magnetron discharges in a sputtering system for the Ag reverse side coating of LEDs were simulated in three dimensions with PIC-MC. The objective was to investigate experimentally measurable LED defects, apparently caused by high-energy Ar+ ion bombardment in the downstream sputtering process. Here the PIC-MC simulation among other things supplied energy distribution functions of the ion bombardment on the surfaces in the simulation space. The process conditions could in fact be adequately reproduced in the PIC-MC simulation: the ion bombardment of the substrate surface exhibited a high-energy component that correlated with the measured extent of the defects.

A further evaluation of the simulation data showed that propagating plasma density fluctuations occur within the electrodeless discharge. These were more or less highly defined depending on the process conditions. These plasmas density fluctuations appeared to have a direct influence on the energy distribution function of the ions.

Several experimental studies on the phenomenon of plasma fluctuations in magnetron discharges have been published in recent years. PIC-MC simulation permitted detailed insights into the cause and effect of the phenomenon for the first time. It was therefore possible to show how electropositive field enhancements accompany the plasma density fluctuations and contribute to the generation of high-energy Ar+ ions and high-energy ion bombardment.