Date of Award

Spring 2011

Document Type


Degree Name

Doctor of Philosophy in Chemical Engineering - (Ph.D.)


Chemical, Biological and Pharmaceutical Engineering

First Advisor

Edward L. Dreyzin

Second Advisor

Rajesh N. Dave

Third Advisor

Norman W. Loney

Fourth Advisor

Robert Benedict Barat

Fifth Advisor

Richard J. Lee


The objective of this work was to understand the processes that lead to ESD stimulated ignition events in metal powders. Several metals were studied: Mg, Al, and Ti. A commercially available ESD test device was adapted for the present experiments. The ESD current and voltage were measured using inductance coils; the data were used to determine the electric circuit impedance, and energy transferred to the powder. A mode of the powder ignition (individual particles vs. dust cloud), ignition delays, burn times, and velocities of ignited and ejected particles were determined from optical measurements.

It was determined that the ESD energy is delivered to the powder primarily via its Joule heating. Powder heating in ESD occurs adiabatically. Powder ejection from the sample holder is due to a reflected shockwave generated by the spark. Among many ejected particles, only some are heated to ignition. The spatial distribution of energy in the powder sample was evaluated from spark imprint measurements.

For Mg, the ESD-ejected powder produces a dust cloud flame. At the minimum ignition energy, ignition can be described by simple heat transfer analysis combined with the Mg thermal initiation kinetics. For Al powder, there is a threshold ESD energy. Above the threshold, ESD ignition results in a dust cloud flame and below the threshold, ESD ignites individual particles producing distinct luminous streaks. A correlation of the longest burn time (for the largest ignited particle) as a function of Joule energy was observed for Al powder with nominal particle sizes in the range of 10-14 μm; however, the correlation was not detected for a finer, 3-4.5 μm powder. A simplified model proposed to describe ignition, considered the Joule energy distributed among the particles based on their surface area. For Ti, powder particles struck by ESD readily fuse together resulting in drastic reduction in the associated Joule heating and reduced ESD ignition sensitivity.

For all metals, it was found that the powder layer thickness affects significantly their ESD ignition. Both powder layer resistance and Joule energy increase with increasing layer thickness; however, the energy density increases substantially for thinner layers. Velocities of ejected particles increase for thinner layers suggesting a greater fire hazard caused by ignited particles traveling longer distances. For powders placed in a monolayer, particles struck by ESD are fragmented, resulting in enhanced ignition, a more readily formed dust cloud flame, and reduced durations of individual particle burn events due to the reduced particle dimensions.

A detailed ESD ignition model for metal powder layers was developed simulating packing of polydisperse powders using Discrete Element Modeling. In the model, the electrical resistance network created by particle contacts is analyzed to determine the powder impedance and energy distribution among heated particles. The model is calibrated for Al powders by matching predicted and measured powder resistance and Joule energy for different layer thicknesses and particle size distributions. The model predicts the temperatures of individual particles heated by ESD. It was observed that the number of particles predicted to be heated to the boiling point compares well to the number of particles ignited in respective experiments.