Removal of small particles from solid surfaces is of critical importance for the microelectronic industry where ~50 % of yield losses are due to particle contamination. Laser cleaning is a technique developed in the late 1980s to remove micro and sub-micro scale particles from surfaces. In this study, a two-dimensional molecular dynamics approach is used to simulate the cleaning process. The model approximates laser energy heating the system that includes the particle, a substrate, and an energy transfer medium (ETM), which is a thin liquid film. The particles are removed through the explosive boiling of the ETM.
Three methods of heating are tested: (1) heating only the particle, (2) heating both the particle and the substrate, (3) heating the ETM layer. These cases will be compared with a previously analyzed case, that of the substrate absorbing the laser light.
Molecular Dynamics
Initialize the system with a set of initial points and parameters.
Calculate the forces on each atom. The Lennard-Jones 12-6 potential function is used along with neighbor lists.
Integrate the equations of motion. Movement of atoms in a time interval are calculated using initial positions, velocities, and forces.
Update the position of each atom.
Repeat the process for the next time interval until finished.
The Lennard-Jones 12-6 potential is used to model the interaction potential between a pair of molecules.
“r” is the distance between two molecules. “σ” is a measure of the molecule’s diameter (the distance where the potential is zero) and “ε” is the depth of the potential well, a measure of the strength of interaction.
The parameters σ and ε are chosen to fit the physical properties of the materials.
Lennard-Jones Potential
The 1/r12 term models the repulsion of the molecules, especially at short distances.
The -1/r6 term constitutes the attractive part, dominating at long distances.
Neighbor Lists
The goal of neighbor lists is to improve the speed of the program by maintaining a list of neighbors of the molecules and updating them at intervals.
If molecules are separated by distances greater than the potential cutoff (known as the cutoff radius), then the program skips those expensive calculations.
Cleaning efficiency is defined as the percentage of particles that are removed from the substrate for a given configuration.
The simulation is run for 30,000 time steps (0.33 ns) for 10 different initial configurations.
The film thicknesses vary from 3σ (1.02 nm) to 70σ (23.8 nm). The particle’s diameter is 19σ (6.46 nm).
The temperatures vary from 1.0 (121 K, -152°C) to 5.0 (605 K, 332°C). 0.1 reduced units correspond to 12.1 K.
There will be three methods of heating that are tested in this study: (1) heating only the particle, (2) heating both the particle and the substrate, and (3) heating the ETM layer.
Substrate heating keeps the particles intact at higher temperatures and seems to work best between layers 10σ and 50σ.
Particle heating works more effectively for thinner film layers than thicker film layers at lower temperatures.
Heating of both the particle and the substrate is very efficient for removing particles especially at lower temperatures, but deforms the particles at a faster rate.
Heating the ETM completely appears very efficient, but would only be feasible for thinner films.
Heating the top 2.5σ of the ETM is not efficient at all, though may be a realistic configuration.
S. Shukla: Optimization of Thickness of Energy Transfer Medium for Laser Particle Removal Process (2003) (Available: http://etd.lib.fsu.edu/theses/available/etd-11242003-113245/unrestricted/manuscript_final_24Nov.pdf)
