Metal interconnects and features are a critical component of modern silicon circuits. In space, NASA and other space agencies have prototyped new ion engine technologies which promise more affordable and faster propulsion to distant targets. What both technologies have in common is the need to create ions to drive their key processes.
Researchers at the U.S. Department of Energy Lawrence Berkley National Laboratory have devised an improved method to produce more metal ions, allowing it to create better circuits, and unlock other applications.
Metal ion creation in the semiconductor industry relies on a technique called sputtering. Traditional sputtering relies on a gas such as argon being heated to plasma and then contained by a magnetic field between a layer of metal and a target circuit. The plasma knocks metal ions off the metal source, creating a current of metal ions which flows towards the circuit, depositing metal on the disk.
High Power Impulse Magnetron Sputtering (HIPIMS) was invented in the 1990s as a means of improving this process. It uses a more powerful magnetic field to accelerate the plasma to higher speeds and to allow some metal ions to return to the metal source, knocking off more metal ions in a chain reaction of sorts. They key limitation to this process was power. More power means better performance, but in commercial semiconductor production typically only 1 kW magnets can be used, and they require water cooling. The result is a sputtering process that is not self-sustaining, though it last slightly longer.
Researchers at LBNL believe they have created the world's first self-sustained sputtering process. Their key is to use high power impulses, rather than a steady higher current, which could melt the magnet.
Andre Anders, a senior scientist in Berkeley Lab’s Accelerator and Fusion Research Division, describes, "Three quantities determine the self-sputtering threshold. One is the probability that a sputtered atom gets ionized. Another is the probability that the new ion returns to the target. Finally, there’s the actual yield of atoms from self-sputtering. Multiply these together and you get the self-sputtering parameter, which is symbolized by the Greek letter pi. When pi equals unity, you reach a new steady state (provided) that the power supply can keep up. We use a special power supply, up to 500 kilowatts peak power. If the system wants power, we give it power!"
The process is also unique in that the power is high enough that it can create a thick plasma of pure metal ions, eliminating the need for argon or other gases in the sputtering process. The result high power continuous sputter has many benefits including cost cuts in chemicals, better circuits, and less mechanical parts (by removing the need for gas injection).
For very small circuits, that will soon arise as die shrinks continue, depositing metal using previous methods might be infeasible as they leave regular voids that on a nanoscale could break connections. With the new approach, the thicker metal ion plasma yields in essence void less deposition, allowing for nanoscale designs with excellent electrical character.
Another potential use of the new sputterer is in spacecraft. Bottles of gas or liquids are bulky and ultimately increase weight by requiring more metal to enclose their greater volume. A metal ion source, using the new method would be self sustaining and much more compact, lowering the weight and cost of launch for ion engine powered spacecraft.
The method also works in a vacuum, so it could also be used for metal ion sputtering in spacing, aiding orbiting construction platforms one day. The method could also be applied on Earth to allow for the first ever successful sputtering of niobium, a tough metal to sputter. This would allow for superconducting cavities of future particle accelerators to be coated with this metal for improved performance, unlock a plethora of new research possibilities.
In short, the new self-sustained sputtering method is a breakthrough which will help advance a number of fields, and if properly implemented, should become an integral technical advance of the new century.
quote: Doing it on an autonomous spacecraft [with] no humans around to monitor or maintain anything for years or even decades is a whole other matter
quote: Even a 1970s-era nuclear reactor doesn't require a large amount of human oversight, except for operations like refueling. With new advanced reactor designs and, even more importantly, advances in computer monitoring, there really isn't any need for humans in the loop at all.
quote: Furthermore, I'm not sure why you're assuming an unmanned flight in the first place. We can already build deep-space probes using ultra slow Hohmann transfers and gravity slings...you need nuclear propulsion for fast manned missions.
quote: Unmanned [missions] will also need to be nuclear powered if we don't want to wait decades or even centuries for a probe to get anywhere.
quote: In space, computers can fail(even permanently due to radiation)
quote: Sounds like you're admitting the need for nuclear propulsion.