Spencer E. Olson and Dr. R. Steven Turley, Physics
Ellipsometry, a common technique used with visible light, measures the change of the state of polarization as it reflect off a material. By so doing, the way that light interacts with materials can be understood more clearly. Also, optical properties, known as the index of refraction and absorption coefficient, can be found from this technique. With these known optical properties of various materials, elements, such as mirrors and polarizers, can be optimized for optical instruments. The name ellipsometry comes from the fact that the technique is based on the analysis of what is called ellipsometrically polarized light. For the simple case, linearly polarized light reflected by a material could become what one terms as ellipsometrically polarized. By analyzing how the polarization state of a light beam changes as it reflects off of a material, one can determine the optical properties of the material, known specifically as the index of refraction and the absorption coefficient of the material. In turn, once the optical properties of a material have been determined, elements can be created to construct instruments such as telescopes or even an ellipsometer. [The physics can also be greater understood through this method .]
The goal of this project is to design an ellipsometer for the extreme ultraviolet (XUV or EUV) range of light. This energy range of light, unlike visible light, is difficult to work with because it is easily absorbed by a material it may pass through. To compensate for part loss of beam intensity by air, one uses vacuum monochromator, which is also used to select a specific energy of light. Still, a major obstacle of this project is to find a way to analyze the polarization state of the light without drastically decreasing the intensity. It was first thought that the diffraction grating of the monochromator would provide partially polarized light sufficient to perform ellipsometric methods. While attending a research conference in Breckenridge, Colorado, I met Dr. Masaki Yamamoto of Tohoku University, Japan. Dr. Yamamoto is doing similar experiments in Japan. In his experiment, it was found that the diffraction grating of his system modified the polarization state of the incident light so little that the idea of using a diffraction grating as the only source of polarized light needed to be abandoned[1].
Another presentation at the conference in Breckenridge showed the use of thin-film multilayers as polarizers in the soft x rays, an energy region between the x rays and XUV[2]. With this, we are trying to design multilayers that will provide sufficiently polarized light and enable us to analyze the polarization state change due to the reflecting material. To do this we are using a computer that will follow a specific algorithm to find the best multilayers for our ellipsometer. Recently, I have been focused on modifying code of an existing program to optimize the polarizer design. The program, originally written by Dr. R. Steven of the Department of Physics, is based on the genetic algorithm. The genetic algorithm functions on a survival-of-the-fittest idea. The program begins by randomly creating a population of multilayers. Each individual, or multilayer, carries its own design information in what are called genes and chromosomes. Each multilayer is then rated according to its computed polarization and the population is sorted according to those ratings. Parents are then chosen and allowed to replicate with a probability of replication proportional to their calculated rating. During the replication, mutations and crossovers of genes are introduced into the population. With a now larger population of multilayers, ratings are redone and the members with the least rating are eliminated, thus keeping the number in the population of multilayers a constant from generation to generation. The population evolves until the improvement from one generation to the next is minimal[3]. Currently, the program is being checked against other numerical methods to find whether it is performing the computations correctly.
Also on the original plan of the project, was the use of a CCD camera as a detection device. Due to leakage in the vacuum system, micro wires of the CCD camera were broken and the detector had to be replaced. It has been replaced by a channeltron detector. The new detector has many advantages. Data from the channeltron is sent via TTL pulses which are then gathered and counted by a computer. This method of data acquisition does not require very much of the computer resources and therefore depends little on the computer. Data can therefore be taken at a much faster rate.
By utilizing Labview, from National instruments, as the programming environment, I was able to incorporate a program I wrote earlier to control a stepper motor inside the monochromator, thus making data acquisition much more automated.
Other improvements include the light source and the vacuum system. Leaks in the vacuum system are a constant problem of any vacuum environment. We have spent a lot of time and improved our system greatly. We have also worked on improving the stability of our light source, with some success. There are future plans to build a much more stable and useful source of XUV and soft x rays.
References
- Masaki Yamamoto, private communication, March 1998.
- Werner Jark, “Theory and Practice of Soft X Ray Transmission Multilayer Polarizers”, Fourth International Conference on the Physics of X Ray Multilayer Structures, Breckenridge, Colorado, March 1998.
- Shannon Lunt, R. S. Turley; “Using the Genetic Algorithm in design of VUV Multilayers”; Fourth International Conference on the Physics of X Ray Multilayer Structures; Breckenridge, Colorado; March 1998.