Modeling and Optimization of Laser Cleaning

Research project P 14700 Modeling and Optimization of Laser Cleaning N. ARNOLD 27.11.2000 Laser cleaning is a new technique to get rid of tiny dust particles, which produce killer defects in many modem nano-devices, such as microprocessors, memory chips, microlenses, and micromechanical devices. Cleaning of polymers was not widely studied despite their importance in future technologies. Small particles block ink-jet printer nozzles, hinder quality of such prototype polymeric devices as solar cells and displays. They stick to the surface with such tremendous forces (billion times stronger than gravity), that conventional methods (rinsing, ultrasound) fail to remove them. Laser cleaning is done with nanosecond or shorter pulses. Heated material thermally expands by an almost unnoticeable distance, but over a very short time and pushes the particle away. In steam laser cleaning a thin liquid layer vaporizes so fast, that it almost explodes and removes the particle. The smaller the particles the more difficult is it to remove them. For modem applications one has to remove particles that are actually invisible and to avoid damage of the underlying material. How exactly the particles are removed is not yet clear. It can be that the particle is initially compressed by the expanding substrate and then bounces off the surface like a squeezed rubber ball. If the substrate is transparent, thermal expansion of absorbing particle may lead to the same effect. Small particles act like lenses -- they focus light onto the surface underneath and thus increase the temperature there. Moisture may condense between the particle and the substrate. This will glue the particle by capillary forces, but may also result in fast evaporation of water and better cleaning. One can try to avoid damage and improve cleaning using even shorter pulses. There exists however a limit -- thermal expansion cannot propagate faster than sound. To understand and improve laser cleaning, one has to model it. Simulations should describe the influence of laser- pulse duration and energy; relation between the minimal particle size and transparency and elasticity of the particle and substrate. Using polymers, one can test different possibilities due to broad variation in their properties. The project should improve the quality and understanding of laser cleaning that is being applied in many areas of nano-technology. It will also clarify how small particles stick to the surface and help to micro-manipulate them.

Tiny dust particles are everywhere. Coming into the wrong place, they may kill modern nano-devices, such as microprocessors, memory chips, ink-jet printer nozzles, displays, during the production stage. Small dust sticks to the surface with forces billion times stronger than gravity. Laser cleaning helps to remove it. How does it work? Surface is illuminated with the laser pulses shorter than billionth of a second. Then it heats, expands, and stops very fast. The particle is then shrugged off the surface. This process is not instantaneous. We modeled how the particle is squeezed by the expanding substrate and then transforms this compression energy into motion. Dust particle lying on the substrate may behave like a "spring". If one excites this spring in properly, cleaning is more efficient. For example if the laser pulse is shorter, "shaking" due to expansion and stopping is faster, and the particles are removed better. But if the pulse is too short, then the material does not have the time to expand (it cannot expand faster than sound). Besides, too short pulses can damage the material. Small dust particles may heat the material around them. They either absorb more light like a soot, or scatter light, so that more is absorbed nearby, or even focus it if they are large enough and spherically shaped. We found that laser cleaning often happens not because of "shaking", but because the material under the particles evaporates and flies away together with the particles. This is unacceptable, because one should not damage the device that one cleans. Therefore we had to understand how to suppress excessive heating. One possibility is to use the light where the substrate is more transparent. Then heating is weaker, but thermal expansion may be kept the same, because larger volume is heated. Another is to let laser-excited sound wave propagate out of the beam area. Evaporation can be turned into advantage, if one ablates water instead of substrate. One can deposit very thin layer of water on the substrate, but this is difficult and may contaminate all surface. We put the samples into the chamber with high humidity, and then water condenses only in the right places - near the dust, like dew in the pores. This results in better cleaning of small particles, when their size is comparable with the water droplets. Employing small microspheres as model dust particles, we noticed that one can use them to make many nano-holes or nano-dots on the surface. One cannot cheaply manufacture millions of micro-lenses, but millions of microspheres can be cheaply produced because many small objects tend to become spherical, like tiny droplets. Such microspheres easily arrange themselves into honeycomb structure (like billiard balls). They act like lenses with exotic properties - instead of focus they produce "hot" focal line. To use them for the processing we developed a theory which explains how such microspheres focus light. This can find applications in colloid and aerosol science.