Challenges to predict and assess flammability of materials intended for use in microgravity environments: a case study of polydimethylsiloxane (PDMS) membrane sheets
View/ Open
Rojas-AlvaU_2022.pdf (4.287Mb)
Date
09/06/2022Item status
Restricted AccessEmbargo end date
09/06/2023Author
Rojas-Alva, Ulises
Metadata
Abstract
Experimental work with different setups was carried out to study the flammability behaviour of polydimethylsiloxane (PDMS) membranes. The flammability was evaluated through thermal degradation analysis, ignition, flame growth, flame spread rates and extinction limits (near-limit phenomena). The experimental results were then compared with microgravity results from the Saffire II experiments that NASA carried out on board the Cygnus spacecraft. The environmental parameters used for the current study were forced flow, flow direction with respect to flame spread (opposed or concurrent), oxygen concentration, ambient pressure, normoxic condition (constant oxygen partial pressure).
The pyrolysis behaviour was evaluated by means of a thermogravimetric analysis in a standard TGA (ASTM E1131). A standard Cone calorimeter (ASTM E 1354) was used to assess the ignition behaviour under stagnant flow conditions where incident heat fluxes were required to derive fundamental properties. Then, for ignition, flame spread and near-limit behaviour under forced flows (concurrent and opposed) coupled with other environmental conditions (oxygen concentration, ambient pressure), three ad-hoc rigs were employed. The first rig was a customised flow tunnel where opposed and concurrent flame spread and near-limit phenomena were studied. The second rig was a modification of the FPA apparatus (ASTM E2058), where an inner flow reducer was custom-designed to improve and straighten the laminar forced flow. In both rigs, forced flows up to 150-200 cm/s could be attained. The third rig was a pressurised chamber, the TOPOFLAME (ZARM, Bremen), which was used to evaluate the effect of oxygen concentration and ambient pressure on the opposed flame spread, near-limit and the ignition behaviour. Lastly, it was also equally important to use a Scanning Electron Microscope (SEM) to quantify silica ash deposition (a PDMS by-product of pyrolysis and combustion) on post-burnt samples.
A complete theoretical evaluation of the flammability behaviour of PDMS was challenging as the literature does not offer reliable information or readily available fundamental properties. It was also found that the deposition of silica ash during any measurement pollutes any reading (such as a mass loss rate). Therefore, some fundamental properties were empirically determined and compared against literature values whenever possible. All the PDMS samples have a critical heat flux of about 30 kW/m2, with small noticeable differences between sample thicknesses. The PDMS used in the current investigation behaves in a complex manner since silica is also formed via solid-phase. The TGA study and mass loss rate
measurements showed that the formation of silica in the solid-phase depends on the heating rate.
Under the effect of the forced flow and flow direction, the forced flow virtually affects the ignition delay times through cooling (pyrolysis) and mixing (increased flow time). The effect of flow direction, concerning the ignition location, affects the ignition delay times as a boundary layer is formed. Thus, the ignition coil for opposed flow (at the top of the sample) experiences a reduced flow, whereas the ignition coils face the forced flow in the concurrent case. The limiting flow condition for piloted ignition was found to be between 50 and 70 cm/s for opposed and concurrent forced flows, respectively. The linear dependency of ignition delay time on oxygen concentration and ambient pressure is via chemistry. For normoxic conditions, the ignition delay time behaves non-monotonically, and the regime change occurs at around 30% oxygen concentration and 0.7 bar ambient pressure. The mechanisms behind such change in behaviour are not well understood, but it is believed to be due silica-ash formation in the gas-phase and silica formation via the solid-phase.
The flame spread behaviour largely depends on the transport and deposition of silica-ash (SiO2) and on the formation of silica in the solid-phase. Under opposed forced flows, if the chemistry is not changed (air under atmospheric pressure), the flame spread rate is independent of the forced flow, and silica-ash formation seems unimportant. Kinetics dominate extinction for high opposed forced flows, and the formation of silica in the solid-phase dominates near-limit as the sample increases in thickness. In the case of flame spread and near-limit, the SiO2 ash deposition can be the dominant mechanism as it affects heat and mass transport significantly for the concurrent case. For high concurrent forced flows, enough silica ash is dispersed ahead of the flame front, and flame spread can reach a steady condition. There is a boundary condition at which concurrent traverses from extinction to steady flame spread.
As expected, flame spread rates were found to depend linearly on the oxygen concentration via chemistry. Flame lengths exhibited almost exponential dependency due to soot oxidation and increased fuel supply. For conditions where pressure was varied, it was found that two mechanisms control opposed flame spread. For relatively higher pressure, the flame spread was dominated by heat and mass diffusivity, whereas for lower pressures, radiative losses through the solid-phase dominated the flame until extinction occurred. Extinction at low pressure is believed to be due to heat losses and kinetics.
For normoxic conditions, the oxygen concentration dominated the flame spread more than the change in pressure. The change of regimes for flame lengths as a function of ambient pressure was found to occur near the normoxic limit coincidentally. That is, the largest flame length occurs at the normoxic limit. The flame spread rates were found to depend linearly on the gas-phase characteristic length (flame length) up to a limiting length (at around 20 mm); after that, flame spread rates were independent. Characteristic solid-phase lengths (pre-heat and pyrolysis) results were not conclusive, as the IR picked up the silica ash particles before the solid-phase.
Using a simplified extended theory (for opposed flame spread), the gas-phase kinetic parameters were empirically derived to evaluate the near-limit phenomena. The opposed flame spread was found to reach a kinetic regime for high forced flows where the effect of silica ash deposition was drastically reduced. Nonetheless, for low opposed forced flows, including in microgravity conditions, the silica formation through the solid-phase and silica-ash through the gas-phase dominated the heat and mass transfer. Pressure modelling for PDMS does not seem to reproduce the same behaviour as the one observed in microgravity, especially for near-limit conditions. In microgravity, for the sample thicknesses tested under 20 cm/s, it seems that the silica ash deposition is the dominant mechanism for flame extinction, but there should be a limiting case where under high forced flow, a steady flame spread might be viable.
A series of properties could be extrapolated from all the experimental results on the PDMS membranes.... The results for the LOI vary according to the sample thickness, which agrees with the literature, due to the nature of the PDMS related to the thermal decomposition and formation of silica via solid-phase. Also, the pre-exponential factor found for the gas-phase varies on the sample thickness and should not be the case, but it is also a consequence of the thermal decomposition of this particular siloxane.
A positive aspect shown by the PDMS membranes is their lower flammability results as they have higher LOI and higher critical heat flux. They’re less likely to ignite compared to other thermopolymers (such as PMMA, PP and others). The formation of silica via solid-phase and silica-ash via gas-phase, along with the transport and deposition on the unburnt fuel, can hinder self-sustained burning. Thus, such distinctive behaviour is desirable for a range of thermoplastic and polymers. Siloxanes, as coatings or silica as fillers on other materials, might help in reducing their flammability. However, it was observed that opposed flame extinction (kinetic regime) requires very high forced flows. The depressurisation strategy on spacecraft (fire-fighting) might not be ideal for an established diffusion flame over a PDMS or similar silicone material.
Overall, the current study found that the flammability criteria might fail to address the flammability behaviour of PDMS fundamentally. Such methods do not gather the effect of other environmental parameters, mainly affecting the silica ash deposition. Therefore, it is recommended that any future flammability assessment method for material selection should not be based solely on a criterion or a fundamental limit that depends on a limited environmental scenario. Instead, the flammability assessment of solid materials should be rated based on a robust array of methods based on normal gravity and microgravity along with fundamental properties inherent to the material that is being studied.
It is recommended to continue working on investigating the flammability behaviour of PDMS and other silicone-based membranes. The formation of silica or silica-ash via solid- and gas-phase, and the transport and deposition of the silica-ash need further understanding. Another important aspect is the effect of normoxic conditions on all aspects of flammability, and it is especially relevant on spacecraft. Experiments in microgravity should are encouraged to further expand the data, and thus a better understanding of microgravity can be attained. Given the PDMS's lower flammability, siloxanes and silica fillers or coatings should be explored and studied to evaluate reducing the fire risk of materials intended for spacecraft applications.