Microgravity Experiments on Polymer Spheres

A series of microgravity experiments to investigate combustion of supported thermoplastic spheres were performed aboard the NASA DC9 and KC135 Reduced Gravity Aircraft by team led by Jiann Yang [6], [7]. Burning histories were recorded on videotape at 35 frames per second for polymethyl-methacrylate (PMMA), polypropylene (PP), and polystyrene (PS) under various conditions of pressure and oxygen concentration. Events observed during combustion include bubbling, sputtering, soot shell formation and breakup, and ejection of material from the burning spheres. This author was privileged to observe some of these experiments in person.

The behavior of combusting polymer sphere proceeds as follows. The sphere is covered with large number of bubbles almost immediately upon ignition. As the bubbles grow, the population appears to be monodisperse in size until bubble bursting is observed. For the first few seconds of burning, as shown in Figure 1, the flame front is relatively quiescent and nearly spherical except for minor asymmetry caused by g-jitter in the aircraft. This is followed by the sudden onset of violent ejection events, which continue at an average frequency of 3 Hz for PMMA and 5 Hz for PP and PS until the fuel is gone [7].

Figure 1: A PMMA sphere in the first few seconds after ignition, with a relatively undisturbed spherical flame front.

Two distinct types of ejection events are observed [18]. In the first, observed during combustion of every polymer sample, the flame front shows large disturbance whose length is on the order of the flame thickness. In the video, this flamelet appears suddenly from one video frame to the next, and decays over few tenths of second, as shown in Figure 2. The structures displayed in these two sequences are suggestive of vortex-flame front interactions such as those studied by Roberts and Driscoll [20] and Renard et al. [21].

In the second type of ejection event, burning particle is emitted from the sphere. Unlike the previously-described gaseous events, this event does not significantly distort the flame front, and the burning droplet travels in straight line away from the sphere, as shown in Figure 3. Unfortunately, the luminosity of the flame front surrounding the droplet prevents the measurement of its size. During the microgravity experiments, PP samples were observed to eject considerable number of particles, while none were observed for PMMA.

The sequence of events for bursting bubble has been photographed by Newitt et al. [22]. As the bubble reaches the surface from within the fluid, the outer surface forms dome while the internal bubble pressure maintains depression at the inner interface. Liquid drains from the dome until it breaks into cloud of droplets on the order of ten microns in size. The bubble gases are released under pressure, likely generating vortices in the quiescent environment and transporting the tiny droplets. The depression left by the escaping gases collapses into central jet, which may break up into one or more relatively large drops (≈ 0.1 mm to 1 mm in diameter) travelling at speed which may initially exceed thousand cm per second [23].

The gas expelled from the bursting bubble and the droplets produced by jet breakup provide good explanations for the two types of ejection events from combusting polymer spheres. Droplet size measurements by Tomaides and Whitby [23] suggest a wave-form disturbance in the bubble film during film rupture. The breakup of the resulting concentric toroids into nearly monodisperse droplets could account for the concentric rings that frequently appear in flamelet events such as that shown in the righthand sequence of Figure 2. The breakup of the central jet into droplets is highly sensitive to the physical properties of the liquid. Since the viscosity of molten polymer is dependent on molecular weight, it is not surprising that jet breakup occurs more readily for PP, which degrades by random scission, than for PMMA, which unzips to monomer during thermal degradation [24] and therefore remains a high-viscosity melt.

Figure 2: Images from combustion of thermoplastic sphere in microgravity. One the left is a sequence of frames showing a developing flamelet in side view. In the previous frame, the lower surface of the flamefront was undisturbed. On the right is a sequence showing a flamelet apparently emitted toward the camera.

Measurements of thermoplastic sphere diameter with time showed that the average burning rates increase with initial sphere diameter and oxygen concentration. The burning rate of PP is slower than that of PMMA and PS. Due to swelling, the measured sphere diameter remains approximately steady for about half of the total burn duration, after which time the square of the diameter linearly decreases with time.