In 20 years of working with various plasma sources this is by far my favorite picture. This is an early iteration of the Thermal Plasma Deposition (TPD) source that I was developing to deposit high purity materials directly onto the outside of cylindrical backing tubes to make cylindrical sputtering targets.

Years of working to optimize sputtering processes and solve customer sputtering problems helped me identify a key problem when reactively sputtering aluminum oxide from aluminum targets, the grain size. High purity aluminum likes to form really large grains with most target manufacturing processes. If the cylindrical target material is cut down to size on a lathe the mechanical cutting process tends to break up the grains within the first milimeter or so of the target surface. This initial finer grain structure would enable fantastic arc free control of the reactive sputtering process but as soon as the target eroded away and exposed the larger grains the aluminum would sputter preferentially from the exposed grain boundaries and oxides would redeposit on the centers of the grains leading to arcing on the target surface. This lead me to ask, how can I control the grain size when making a target? Of course, there are several solutions that already exist like adding silicon and other dopants to disrupt the grain coalescence, cold working, and cold spraying but all of these processes can add impurities, trapped gas, and incorporate oxides. I wanted a no compromise solution to control the grain size without sacrificing purity and thus I came up with the TPD process.

In the picture, a copper induction coil is surrounding a ceramic crucible filled with high purity aluminum metal. The induction current is used to melt and evaporate the aluminum metal in the crucible, the voltage differential from coil to coil helps form a plasma from the evaporated aluminum flux and accelerate the ionized species out the end of the coil. The magnetic field generated by the coil help confine the plasma primarily within the opening of the crucible. This process operates only from evaporated flux and no other gas is introduced into the vacuum system.

Unfortunately, aluminum was one of the worst materials to work with because in it's liquid state it aggressively attacks and destroys all high temperatures ceramics making the process extremely difficult to sustain. Ultimately we abandoned aluminum and created a refined implementation of this process that was able to successfully deposit ITO directly onto cylindrical backing tubes with a controlled grain size and oxide reduction.