The magnetic field is a key feature that distinguishes magnetron sputtering from simple diode sputtering. It effectively increases the residence time of electrons close to the cathode surface and by that increases the energy efficiency of the discharge. This becomes apparent in high power impulse magnetron sputtering (HiPIMS) discharges, as small changes in the magnetic field can result in large variations in the discharge characteristics, notably the peak discharge current and/or the discharge voltage during a pulse. Here, we analyze the influence of the magnetic field on the electron density and temperature, how the discharge voltage is split between the cathode sheath and the ionization region, and the electron heating mechanism in a HiPIMS discharge. We relate the results to the energy efficiency of the discharge and discuss them in terms of the probability of target species ionization. The energy efficiency of the discharge is related to the fraction of pulse power absorbed by the electrons. Ohmic heating of electrons in the ionization region leads to higher energy efficiency than electron energization in the sheath. We find that the electron density and ionization probability of the sputtered species depend largely on the discharge current. The results suggest ways to adjust electron density and electron temperature using the discharge current and the magnetic field, respectively, and how they influence the ionization probability.

The ionization region model (IRM) is applied to model a high power impulse magnetron sputtering discharge in argon with a graphite target. Using the IRM, the temporal variation of the various species and the average electron energy, as well as internal parameters such as the ionization probability, back-attraction probability, and the ionized flux fraction of the sputtered species, is determined. It is found that thedischarge develops into working gas recycling and most of the discharge current at the cathode target surface is composed of Ar⁺ ions, which constitute over 90% of the discharge current, while the contribution of the C⁺ ions is always small ( < 5 %), even for peak current densities close to 3 A cm⁻². For the target species, the time-averaged ionization probability ⟨α _t,pulse⟩ is low, or 13–27%, the ion back-attraction probability during the pulse β _t,pulse is high ( > 92 %), and the ionized flux fraction is about 2%. It is concluded that in the operation range studied here it is a challenge to ionize carbon atoms, that are sputtered off of a graphite target in a magnetron sputtering discharge, when depositing amorphous carbon films.

High-power impulse magnetron sputtering (HiPIMS) is an ionized physical vapor deposition technique that provides a high flux of ionized target species for thin film growth. Optimization of HiPIMS processes is, however, often difficult, since the influence of external process parameters, such as working gas pressure, magnetic field strength, and pulse configuration, on the deposition process characteristics is not well understood. The reason is that these external parameters are only indirectly connected to the two key flux parameters, the deposition rate and ionized flux fraction, via two internal discharge parameters: the target atom ionization probability α_t and the target ion back-attraction probability β_t. Until now, it has been difficult to assess α_t and β_t without resorting to computational modeling, which has hampered knowledge-based optimization. Here, we present a simple method to deduce α_t and β_t based on measured deposition rates of neutrals and ions. The core of the method is a refined analytical model, which is described in detail. This approach is furthermore validated by independent calculations of α_t and β_t using the considerably more complex ionization region model, which is a plasma-chemical global discharge model.

We apply the ionization region model (IRM) and the Orsay Boltzmann equation for electrons coupled with ionization and excited states kinetics (OBELIX) model to study the electron kinetics of a high power impulse magnetron sputtering (HiPIMS) discharge. In the IRM the bulk (cold) electrons are assumed to exhibit a Maxwellian energy distribution and the secondary (hot) electrons, emitted from the target surface upon ion bombardment, are treated as a high energy tail, while in the OBELIX the electron energy distribution is calculated self-consistently using an isotropic Boltzmann equation. The two models are merged in the sense that the output from the IRM is used as an input for OBELIX. The temporal evolutions of the particle densities are found to agree very well between the two models. Furthermore, a very good agreement is demonstrated between the bi-Maxwellian electron energy distribution assumed by the IRM and the electron energy distribution calculated by the OBELIX model. It can therefore be concluded that assuming a bi-Maxwellian electron energy distribution, constituting a cold bulk electron group and a hot secondary electron group, is a good approximation for modeling the HiPIMS discharge.

In magnetron sputtering, only a fraction of the sputtered target material leaving the ionization region is directed toward the substrate. This fraction may be different for ions and neutrals of the target material as the neutrals and ions can exhibit a different spread as they travel from the target surface toward the substrate. This difference can be significant in high power impulse magnetron sputtering (HiPIMS) where a substantial fraction of the sputtered material is known to be ionized. Geometrical factors or transport parameters that account for the loss of produced film-forming species to the chamber walls are needed for experimental characterization and modeling of the magnetron sputtering discharge. Here, we experimentally determine transport parameters for ions and neutral atoms in a HiPIMS discharge with a titanium target for various magnet configurations. Transport parameters are determined to a typical substrate, with the same diameter (100 mm) as the cathode target, and located at a distance 70 mm from the target surface. As the magnet configuration and/or the discharge current are changed, the transport parameter for neutral atoms ξtn remains roughly the same, while transport parameters for ions ξti vary greatly. Furthermore, the relative ion-to-neutral transport factors, ξti/ξtn, that describe the relative deposited fractions of target material ions and neutrals onto the substrate, are determined to be in the range from 0.4 to 1.1.