The effect of different chemical classes on the swelling of NBR O-rings in blends with synthetic paraffinic kerosene

Thesis / Dissertation


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title
Synthetic jet fuels provide a number of benefits over petroleum-derived fuels. They have thus been considered as alternative fuels. Compatibility issues, however, are of concern; specifically the interaction of synthetic fuels with polymeric materials which are commonly used to seal fuel systems. This is because of differences between the composition of synthetic fuels and petroleum-derived fuels. Synthetic fuel streams contain no or very low aromatics, unlike petroleum-derived fuels. This investigation was consequently initiated to gain a greater understanding of the factors affecting seal swell in the aviation industry. The study focussed interactions between fuel components from various fuel classes and nitrile rubber (NBR). Currently fully and semi-synthetic jet fuels are required to contain a minimum of 8% aromatic components by volume in order to minimise any changes in polymeric swell when switching between synthetic jet fuels and petroleum-derived jet fuels. In this study procedures were refined to allow seal swell to be assessed. ASTM 01414 and 0471 were used as base methods together with the use of an elastomer compression rig that had been designed and built at the Sasol Advanced Fuel Laboratory (SAFL). A method was developed to remove plasticiser for the O-ring seals to provide samples that are more representative of O-rings in service. Experiments were conducted on both new and conditioned (deplasticised) NBR O-ring samples. Materials tested in this investigation included petroleum-derived Jet A-1, Fischer Tropsch-derived synthetic paraffinic kerosene (SPK) and pure compounds including isomers of n-, iso- and cyclic paraffins, aromatics and oxygenates. The paraffins were tested as neat components. With the aromatics and oxygenates, 8% (v/v) blends with coal-to-liquid (CTL) SPK were prepared to simulate fully synthetic jet fuels that meet the minimum 8% aromatic specification. The data which were collected were primarily changes in mass and volume of the O-rings which underwent fuel exposure. An assessment was made of both the kinetics of fuel uptake and the extent of swell achieved at equilibrium. Initial experiments focussed on method refinement. This included the measurement techniques (gravimetric, volumetric and seal swell rig), O-rlng conditioning (i.e. plasticiser removal), the effects of temperature and the solvent to polymer ratio used. The greatest level of repeatability was achieved with gravimetric measurements. Similar swelling trends were observed with volumetric measurements, although with lower repeatability. Solubility parameters and molar volume were shown to be key determinants of seal swell. It was demonstrated that the extent of swell was not significantly different between n- and iso- isomers of octane and dodecane. What was different was that n-paraffins initially swell faster than their iso-equivalents. A significant difference was observed with cycloalkane isomers which swell to a much greater extent. It is suggested that this difference is the result of a combination of molar volume and solubility parameter differences. A study on a series of n-alkanes showed that although solubility parameters increase with increasing carbon number, seal swell decreases. Molar volume, by contrast, decreases. Thus for n-alkanes molar volume is the key determinant. A statistically significant correlation between the density of n-alkanes and the extent of swell was observed. Investigations into blends of coal-to-liquid (CTL) SPK showed that the seal swell was highly dependent on the hydrocarbon aromatic used. Some blends swelled more than petroleum-derived Jet A-1 and some less than Jet A-1. Again molar volume was demonstrated to be important but also the ability of aromatics to form hydrogen-bond like interactions with NBR. Lower molar volumes and higher Bh values produced more favourable swelling. The importance of the presence of multiple rings was also demonstrated by the increased swelling of the C10, tetralin, over other C10 aromatics such as n-butylbenzene. Aromatic ethers caused significantly more swell than hydrocarbon aromatics. Some of these aromatic oxygenates swelled even more than petroleum-derived Jet A-1 even when used at the minimum 8% aromatic level. This was explained in terms of their even stronger polar and hydrogen bond interactions. The effects of temperature were an increase in the rate of swelling, as expected, and a slight decrease in the extent of swelling at elevated temperatures. However, the influence of temperature on the aromatic oxygenate, benzyl alcohol, which has previously been reported as producing large swell even when added at concentrations as low as 0.5%, was significant. It is suggested that this is because of the very strong hydrogen bond between benzyl alcohol and NBR. Because of the exothermic nature of hydrogen bonding, the equilibrium constant for this interaction decreases at elevated temperatures leading to less swell. The results from this study suggest that lighter (lower carbon number) SPKs and SPKs containing cycloparaffins would promote swelling. The addition of lighter, higher Sh and/or multi-ring aromatics to SPK would produce the most swell of the aromatic species added. This could be even more enhanced by the presence of aromatic oxygenates. Such compounds could be targeted for blending with synthetic paraffins to produce fully synthetic jet fuels but their impact on other properties such as flash point and distillation behaviour cannot be ignored.