University Park, Pa. --- Penn State engineers have developed a new simpler computer simulation for ultrafine particle size growth and distribution that is potentially applicable to processes ranging from powdered milk production to ceramic membrane development to air pollution control.
Dr. Themis Matsoukas, assistant professor of chemical engineering and leader of the project, says the new simulation is not only simpler than the others available, but is also fast, accurate and uses only modest amounts of computing power.
Matsoukas and his research group are using the simulation to understand the grouping and breakup processes that ultrafine titanium dioxide particles undergo before they reach their final size. Formation of ultrafine, nanometer-sized titanium dioxide particles, about the size of a virus, is characterized by rapid aggregation of large particle groups followed by slow breakup of these groups. The breakup or de-aggregation can take hours or days depending on the processing conditions.
The new simulation can accurately predict how the size and distribution of the particle groups occur over time. Some results from the simulation have already been confirmed and verified by comparison with experimental results.
"We haven't yet applied the simulation to other particulate materials. However, it could potentially be done for any powder of interest, powdered milk, for example. The simulation could also be used to model the formation and behavior of aerosols, sol/gel particle producing systems or even aggregates of pollutants in the atmosphere," says the Penn State researcher.
Matsoukas chose to model titanium dioxide nanoparticle formation because of current interest in using metal oxides in ceramic membranes for gas separation. The size of the particles is important to the activity of the membrane in this application and nanosize particles are difficult to obtain in non-aggregated form.
He and his group have shown that stirring induces aggregation of titanium dioxide nanoparticles and affects their stability but has little effect on the rate of breakup. The presence of alcohols also promotes aggregation as does coating the nanoparticles with polymers. They have found breakup to be promoted by the type of starting material (alkoxide type) and acidity.
The new simulation is based on a standard Monte Carlo method, a technique for estimating the solution of a mathematical problem by artificial sampling. In the new simulation, however, the Penn State researchers simplify the solution by using a sample of fixed size, regardless of whether the actual growth process results in a net loss via aggregation or gain via breakup.
They have shown that, by using this technique, they can obtain accurate results for very long growth times in about four minutes on a mainframe computer.
The Penn State research team's results on titanium dioxide were detailed in a series of papers at the '98 American Institute of Chemical Engineering meeting in November. The simulation method and comparison with numerical and theoretical solutions is detailed in the paper, "Constant-number Monte Carlo Simulation of Population Balances," in a recent issue of the journal, Chemical Engineering Science.
EDITORS: Dr. Matsoukas is at txm11@psu.edu by email.
Journal
Chemical Engineering Science