In Situ Optical Sensing and State Estimation for Control of Surface Processing

Abstract

Many industrial processes involve the modification of surfaces. The semiconductor and microelectronics industries are extremely reliant upon surface processing [1], but other applications include solar cells, MEMS and microfluidics, thermal barrier coatings, and fuel cells. Optical measurements are commonly used to indirectly quantify the surface and bulk properties of materials. Unlike direct imaging methods, in which the local surface topography is visualized, indirect methods provide information on surface and bulk properties which are averaged over the entire area being sampled. Typically, a beam of light is directed at the surface to be measured, and the light interacts with the surface to produce the reflected or diffracted light that is then measured by a photodetector. One advantage of indirect measurement techniques is that they are non-invasive, and therefore can be used in extreme or harsh processing environments such as vacuum chambers and reaction vessels by passing the light through windows. These in situ optical sensors have been used for process control of film thickness and chemical composition. However, many more microscopic and nanoscale surface properties can be measured and potentially controlled using optical sensors. The barriers include the difficulty of modeling the optical response, the difficulty of inverting these relationships to robustly infer surface properties, and the practicality of including windows on the chamber to enable optical access. However, to create precise nanostructures in a high-throughput manufacturing setting, real-time measurement and control will be required, and optical sensors are expected to play a significant role.