From Wikipedia, the free encyclopedia
Ion scultping is a term coined by Golovchenko at Harvard and co-workers. This term describes a two-step process to make nanopores. The first step is make either a through hole or a blind hole, most commonly using an focused ion beam (FIB). The holes are commonly ~100 nm, but can be made much smaller. This step may or may not be done at room temperature, with a low temperature of -120 C. Next, there are three common techniques to now 'sculpt' the hole: broad area ion exposure, TEM exposure, and FIB exposure. Holes can be closed completely, but also they can be left open at a lower limit of 1 - 10 nm.
Broad area ion exposure
This technique uses a broad area argon ion source beam. If the hole is blind (a blind hole is a hole that has not broken through on the backside yet) the wafer (often SiN or silicon oxide) is then turned upside down, and exposed with the argon beam. A detector counts the amount of ions passing through the membrane (which should be zero). The process stops when ions begin to be detected. This enables for a much smaller hole to be opened than if using an FIB alone.
Alternatively, if the hole is through, the argon beam is aimed at the wafer, and by effects not fully understood, atoms from elsewhere on the wafer move to close the hole. The process is stopped when the detector is only receiving a small current. If the current drops to zero, then the hole is closed. The idea is that the smaller the hole, the less ions get through to the detector. This is the process used by J. Li and J. Golovchenko. They published a paper in July 2006 saying they can now use all the noble gases for this process, not just argon.
A through hole in a wafer can be closed down by a transmission electron microscope. Due to hydrocarbon buildup, the electrons stimulate hole closure. This method is very slow (taking over an hour to close a 100 nm hole). The slow method allows for great control of the hole size (since you can watch the hole decrease), but its drawback is that it takes a long time.
This is the easiest of the techniques, but the least useful. After a hole is milled with an FIB, one can just image the hole (analagous to the TEM technique). The ions stimulate movement on the wafer, and also implant themselves to help close the hole. Unlike the other two methods, the holes closed in this technique are not very circular and smooth. The holes appear jagged under TEM photos. Also, it is much hard to control the size of the hole to the single nanometer regime. Another drawback to this technique is that while imaging the hole, the ion beam is continually sputtering membrane material away. If the beam scan area is large enough, the rate of atoms moving to close the hole will be greater than the rate of sputtering, so the hole will close. If the membrane is too thin or the scan area too small, then the rate of sputtering will win, and the hole will open up.