Professor Themis Matsoukas | Research
Colloid Chemistry of Sol-Gel Nanoparticles
Nanoparticles from various materials can be synthesized in a number of ways. However, maintaining a stable suspension of such small particles is, however, a difficult task as such particles have a high tendency to aggregate. Our work in this area focuses on the aggregation and deaggreation of nanocolloids and their stabilization. We synthesize colloidal suspensions of particles as small as 10 nm in diameter and use light scattering to study the rate of aggregation and deggregation under varying processing parameters.
We are actively working on the development of kinetic models to understand -and quantify- the dispersion of aggregated nanoparticles.
Gas-phase processing of materials has several advantages but some important limitations as well. Particles in the gas phase are almost invariably obtained in agglomerated form. This is compounded the lack of tools such as pH agents, ionic strength, solvation forces, or steric interactions to control the growth process. Low-pressure plasmas offer the promise of breaking away from such limitations. Particles in these systems become electrostatically charged and this offers an opportunity to effect interactions at the microscopic level.
Our recent work in this area has quantified several similarities between liquid-phase colloids and particle clouds in ionized gases. This fundamental analogy drives our theoretical and experimental efforts in developing what could be rightfully called "science of plasma colloids".
Population Balances by Monte Carlo
The growth and disintegration of dispersed systems and the evolution of their size distribution is a problem that arises in several processes. In the presence of coagulation or fragmentation the population balance that governs the evolution of the size distribution is an integro-differential equation. Several difficulties associated with the solution of such equations are circumvented through the use of Monte Carlo (MC) methods. Monte Carlo utilizes probabilistic tools to sample a finite subset of a system in order to infer its properties.
We have recently developed a new MC algorithm which we have shown to be both accurate and efficient. Most significantly, the simulation time is not limited by the finite number of particles and the error propagation has been shown to be much slower than conventiona techniques.