Mitchell W. Larsen, Physics and Astronomy
With the expansion of x-ray technology and recent development of laser x-rays, development of a practical, high precision x-ray mirror has become more urgent. One example of a practical and current need for such a mirror is the x-ray microscope at Duke University. The above planned microscope would function by focusing x-rays on a tiny section of a cell frozen during a specific metabolic stage. The molecules in the cell would absorb these x-rays, and then radiate x-rays off. Every element would radiate a unique wavelength of x-rays, and we would thus be able to determine which elements are present in the celL
There are currently no methods to mass produce x-ray lenses, Every one is hand made and very, very expensive. Current x-ray lenses are lacking in accuracy because after the mirror or lens is ground into the correct shape, a reflecting layer must be placed on top of it. The process of applying the reflecting layer deforms the mirror and makes it less precise.
I have been working on a project to develop a method by which x-ray mirrors can be constructed cheaply and efficiently. A mirror constructed in the manner we propose uses a wafer of silicon as a substrate, Rather than grind the silicon wafer to the desired shape, we wish to bend it into shape. Bending occurs when two substances are bound to each other at very high temperatures. As the substances cool, they shrink at different rates, This causes them to bend. Rarely does just applying a film to a silicon wafer cause it to bend as much as needed. We must first weaken the silicon so it bends more easily.
To accomplish this, one must etch deep, narrow grooves into the crystal structure of the silicon. These grooves must have the desired dimensions and must be placed in a specific place on the wafer. Forming the desired grooves is very difficult and has been the main focus of my research.
In etching, the substrate is immersed in a chemical which erodes the silicon which would erode the entire wafer. To etch grooves, a protective substance must be placed on the surface of the silicon which will protect the areas which should not be etched. As the solution begins to etch, it etches down, but not sideways into the silicon. The solution will eat directly into the substrate, and you will have grooves arranged in your desired pattern.
Obtaining this pattern is difficult. First, I had to get my desired pattern onto a glass slide which could then be used to expose this pattern onto the silicon. This was done by making a 3′ by 3′ pattern on a large piece of plastic. This plastic was hung on a semitransparent window with lights behind it. This pattern was then exposed onto some glass slides with photosensitive material on one side. This glass slide could then be used as a mask to expose the pattern onto the silicon.
I then acquired two silicon wafers to be used as substrates. A protective film had to be applied over the entire surface of the silicon. This substance is usually silicon dioxide, Si02. On top of this Si02 one applies a surface of photosensitive material called photoresist. When certain parts of this photoresist are exposed to light, it will change chemical composition. The desired pattern was then transferred to the photoresist by placing the glass slide on the photoresist and exposing it to light. The exposed portion of the photoresist and subsequent layer of silicon dioxide were removed leaving the desired areas of the wafer exposed.
The wafer was then placed into a solution of KOH, which etches down into the exposed areas. When the solution has etched the desired depth into our substrate, we stop etching, and then remove the protective layer of Si02 by placing it in a solution of HF. These grooves facilitate bending as calculated by a computer.
I have etched some wafers as described above. However, as I was studying them do discover how deep the grooves were, they shattered. I then acquired ten new wafers and while repeating the process I ran into a large problem. The pattern I etch into the wafer must be aligned with the molecular structure of the crystaL I had been using a diffractometer on campus to align my pattern with the molecular structure, but the owner of the diffractometer dismantled it. He then could not locate the instruction manuals he needed to put it back together. I worked over the summer on new ways to find the orientation of the crystal lattice in my wafers. Using a pattern from Livermore Laboratories, I etched patterns at every angle into the wafer. The patterns which were not on the correct angle eroded the entire wafer, while the pattern which was oriented with the structure of the silicon left clean grooves and ridges. By inspection under a microscope I could tell which angle was the best to orient my pattern.
After etching the pattern into these new wafers, I will need to measure the dimensions of the grooves, and then apply the film which will warp the wafer. This is done by chemical vapor deposition. The wafer is placed face down in a special oven which vaporizes silicon nitride. The silicon nitride fills the entire oven and begins to deposit itself on the wafer and react with it. It slowly forms a thin layer which is bonded to the wafer. As the wafer then cools, it shrinks at a different rate than the silicon nitride, and the wafer will bend.
These curved silicon crystals are the base for possible x-ray mirrors. With more research and knowledge about the subject, functional mirrors could be built in a short time. This research project would not have-beenpossible-without tht; sGholarship from the departmenLof researchand_creatiYe _ works, and I am very grateful for your help. Hopefully, my research will lead to further funding for the department and especially this project.