Eric Halpenny and Dr. Larry Baxter, Chemical Engineering

Aerosols are small (sub-micron) particles formed mainly in combustion processes by the vaporization and subsequent re-condensation of volatile, inorganic compounds. Cloud formation is an example of an aerosol process. However, unlike clouds, most aerosol-generating processes produce negative effects on both the process that forms them and the environment. Aerosols represent major process and emissions issues for essentially all systems employing low-grade fuel (coal, biomass, and black liquor). An accurate, predictive understanding of aerosol formation and chemistry would substantially improve the efficiency and environmental performance of many processes.

The purpose of my research was to develop a computer simulation capable of predicting aerosol formation and chemistry. This computer simulation effectively describes the formation of sub-micron, liquid-phase particles (or “aerosols”) based on the temperature and composition of a gas phase. The results of the code were compared with previously performed experimental results tested in both the laboratory and the field. The simulation is based on mathematical theory derived by experts in the field of aerosol dynamics.

This simulation would be most effectively used in conjunction with computational fluid dynamics (CFD) codes for black liquor combustion or gasification to predict fume formation and subsequent deposition as a function of operating conditions, fuel properties, and boiler design. Additionally, the code predicts important aspects of sulfur chemistry in recovery boilers, specifically the scrubbing of sulfur dioxide from the gas phase and formation of sulfates in aerosol phases. This tool predicts the amounts, sizes and composition of aerosols and the composition and concentration of remaining vapors under arbitrary boiler conditions.

The data made available by this simulation, specifically applied to black-liquor-fired recovery boilers, which lend themselves to investigation by this method, could prove beneficial in counter-acting degrading effects produced by aerosol particles. Predictions of this type are also useful for all other low-grade fuel systems, i.e. biomass- or coal-fired boilers, as well as in advanced materials systems (such as particle-assisted chemical vapor deposition processes).

One of the major challenges involved in completing this project was applying theoretical mathematical equations based on ideal conditions to a non-ideal system. Most of the chemical interactions involved in aerosol formation occur under conditions that cannot be described by well-developed theory. Instead, they are either described by empirical correlations or are not well understood. Working to better understand and then apply these situations in a practical manner was of great educational value to me.

This project is perhaps the first of its kind in that the simulation can predict composition, mass and number concentration, and the shape of the particle size distribution for aerosol species in black-liquor-fired recovery boilers. The biggest improvement over codes of the past is its ability to deal with multi-component vapor and condensed phases. It marks a significant step in understanding aerosol dynamics, with possible implications in improving the efficiency and life of recovery boilers. Additionally, this research verifies the current hypothesis that the initial vapor concentration of alkali salts and sulfates, and gas temperature are the principle factors affecting aerosol formation.

My work would be measurably enhanced by further research in three areas. First, work in non-ideal thermodynamic interactions for liquid-salt mixtures would improve the reliability of aerosol-phase predictions, including sulfate scrubbing from the vapor phase. Secondly, it is important that the vapor condensation theory be researched further such that it accurately models the diameter and number of condensed particles. Finally, accounting for non-ideal mixing, either through further research or by incorporating the model into a CFD code, would resolve the major discrepancies between predicted and measured size distributions.

This project was an educational experience for me and proved to be an opportunity for me to grow intellectually. I was able to apply much of the engineering theory I had learned during my coursework to a complex problem. From that experience, I gained a far greater perspective of the scope of my education. I was also able to closely interact with one of my professors in discussion material in his field of expertise. This helped me gain insight and knowledge not available in a classroom setting, or even in normal one-to-one sessions.

However, perhaps the most fulfilling part of this project was in the final result. I was able to bring together the results of my work into a presentation for an international group of experts in the field I was researching. The final project and results of the simulation were presented in poster format at the International Chemical Recovery Conference (ICRC) held June, 2004 in Raleigh, North Carolina. A paper describing the theory and results of the simulation was also published in the Technical Association for the worldwide Pulp and Paper converting Industry Journal.