Modeling and Experimental Studies of High Power Impulse Magnetron Sputtering Discharges

Sammanfattning: HiPIMS, high power impulse magnetron sputtering, is a promising technology that has attracted a lot of attention, ever since it was introduced in 1999. A time-dependent plasma discharge model has been developed for the ionization region (IRM) in HiPIMS discharges. As a flexible modeling tool, it can be used to explore the temporal variations of the ionized fractions of the working gas and the sputtered vapor, the electron density and temperature, the gas rarefaction and refill processes, the heating mechanisms, and the self-sputtering process etc.. The model development has proceeded in steps. A basic version IRM I is fitted to the experimental data from a HiPIMS discharge with 100 μs long pulses and an aluminum target (Paper I). A close fit to the experimental current waveform, and values of density, temperature, gas rarefaction, as well as the degree of ionization shows the general validity of the model. An improved version, IRM II is first used for an investigation of reasons for deposition rate loss in the same discharge (Paper II). This work contains a preliminary analysis of the potential distribution and its evolution as well as the possibility of a high deposition rate window to optimize the HiPIMS discharge. IRM II is then fitted to another HiPIMS discharge with 400 μs long pulses and an aluminum target and used to investigate gas rarefaction, degree of ionization, degree of self-sputtering, and the loss in deposition rate (Paper III). The most complete version, IRM III is also applied to these 400 μs long pulse discharges but in a larger power density range, from the pulsed dcMS range 0.026 kW/up to 3.6 kW/where gas rarefaction and self-sputtering are important processes. It is in Paper IV used to study the Ohmic heating mechanism in the bulk plasma, couple to the potential distribution in the ionization region, and compare the efficiencies of different mechanisms for electron heating and their resulting relative contributions to ionization. Then, in Paper V, the particle balance and discharge characteristics on the road to self-sputtering are studied. We find that a transition to a discharge mode where self-sputtering dominates always happens early, typically one third into the rising flank of an initial current peak. It is not driven by process gas rarefaction, instead gas rarefaction develops when the discharge already is in the self-sputtering regime. The degree of self-sputtering increases with power: at low powers mainly due to an increasing probability of ionization of the sputtered material, and at high powers mainly due to an increasing self-sputter yield in the sheath.Besides this IRM modeling, the transport of charged particles has been investigated byiv measuring current distributions in HiPIMS discharges with 200 μs long pulses and a copper target (Paper VI). A description, based on three different types of current systems during the ignition, transition and steady state phase, is used to analyze the evolution of the current density distribution in the pulsed plasma. The internal current density ratio (Hall current density divided by discharge current density) is a key transport parameter. It is reported how it varies with space and time, governing the cross-B resistivity and the mobility of the charged particles. From the current ratio, the electron cross-B (Pedersen) conductivity can be obtained and used as essential input when modeling the axial electric field that was the subject of Papers II and IV, and which governs the back-attraction of ions.