When a laser with a large enough fluency is focused on a surface, it generates a plasma. The light heats up the electrons of the target that couple with the phonons within a few tens of ps leading to a phase transition. The dense and hot plasma then expands and cools down with a few tens of s in air. Under liquids, the cooling is much quicker because of the liquid’s evaporation. Shock waves can also be observed during the first few hundreds of nanoseconds, followed by a bubble during the first ms. Space and time evolution of the plasma is fast and complex. In this work, I explore a new way to describe nucleation out of equilibrium based on statistical physics. I developed spectroscopic techniques to measure the time and space evolution of the plasma in order to determine its composition and temperatures.

Understanding and modeling the nucleation of clusters and nanoparticles in the plasma is the key to master their synthesis. I spent a part of my first year of my PhD developing a growth model for clusters. It is an activated complex-like model using Weisskopf’s microcanonical approach to handle the kinetics of the growth when the transient nature of the processes disregards the canonical ensemble. This model is applied to size distributions’ calculation of aluminum oxide clusters and confronted to measurements I performed on a cluster source in Tokyo. It efficiently reproduces experimental size distributions of clusters formed with different buffer gas densities. The activation energy in the model is addressed using DFT calculations and experimental measurements. These results, compared to the model, show that the thermodynamic equilibrium is not reached during the growth and that there is no equipartition of the energy in the clusters. The bounding energy measured is ten times lower than the energy given by DFT calculations and lower than the bulk cohesive energy.

One of the key parameters in the description of laser-generated plasma is the pressure. As a thermodynamics parameter, it is important for nucleation processes and for the plasma description. As a mechanical parameter, it is interesting for phase transition in the target and damage to the surface. The ablation generates very large pressure and shock-waves. I measured the dynamics of the generated shock-wave as a function of the pulse energy using a shadowgraph setup, which gives the jump pressure using Rankine–Hugoniot relations. Generated pressure can reach few tens of MPa for atmospheric pressure ablation. I show that the dynamics of the shock-wave is fully determined by Taylor’s blast model if one takes into account anisotropy in the energy spreading. The main interest of using Taylor’s model for laser ablation, is that it is extremely simple and universal and depends on the laser pulse energy only. Using this model, one can determine the pressure from 20 ns after the laser pulse only from the pulse energy and the density of the surrounding atmosphere. Considering Saha’s law, I show that one can determine the average temperature of a plasma from measurement of the electrons density and energy of the pulse only. This temperature appears to be lower than the temperature measured from atomic emission, which could be explained by temperature gradient in the plasma or partial failure of Local Thermodynamic Equilibrium (LTE).

A large part of my PhD was devoted to the development of the time and space resolved spectroscopy of the plasma. By fitting the recorded spectra of AlO molecules, I was able to determine the rotational and vibrational temperatures of AlO molecules and the rotational temperature of TiO molecules, depending on their position in the plasma and time. The formation of the molecules occurs mainly in the outer part of the plasma, where the temperature is low enough. The vibrational temperature of AlO appears much larger than its rotational temperature, proving that the thermodynamics equilibrium is not reached in the excited state of the molecules. Because of a short lifetime of the excited state, molecules formed have not enough time to be thermalized. The rotational temperature of the molecules is likely to cool down quicker due to large coupling with the gas. The rotational temperature of TiO molecules is similar to the one of AlO but slightly lower, probably due to larger collision cross-section and better thermalization. Rotational temperature of the molecules appears to be a good probe of the kinetic temperature of the gas, especially when the excited state lifetime is long and coupling with the gas is large.

Molecules emits light only when they are excited. If one wants to probe the temperature of the ground state molecules, emission spectroscopy is not enough… One need to excite these molecules first and use fluorescence spectroscopy. Fluorescence spectra can be decomposed in two contributions: a first one which is direct fluorescence, conserving an initial population in excited state, a second one which is thermalized, populating many other quantum levels in excited state. The emission of the relaxed contribution can be approached by calculations considering population distributions following these vibrational and rotational “relaxation” processes. These populations can be modeled with Boltzmann distributions taking very different vibrational and rotational temperatures, which strongly depend on the excited transitions. The direct fluorescence can be described considering a population distribution following resonant absorption of the laser. This population is directly related to the ground state population that can be modeled with a Boltzmann distribution. By fitting fluorescence spectra, one can measure the rotational temperature of the ground state molecules.