1 Energy Engineering, Department of Mechanical Engineering, Technical University of Denmark2 Department of Mechanical Engineering, Technical University of Denmark
Dimethyl ether (DME) has been recognised as an excellent fuel for diesel engines for over one decade now. DME fueled engines emit virtually no particulate matter even at low NOx levels. DME has thereby the potential of reducing the diesel engine emissions without filters or other devices that jeopardise the high efficiency of the engine and increase the manufacturing costs. DME has a low toxicity and can be made from anything containing carbon including biomass. If DME is produced from cheap natural gas from remote locations, the price of this new fuel could even become lower than that of diesel oil. Fueling diesel engines with DME presents two significant problems: The injection equipment can break down due to extensive wear and DME attacks nearly all known elastomers. The latter problem renders dynamic sealing diƣult whereas the first one involves the poor lubrication qualities of DME which are the main concerns of the present study. The volatile fuel tribotester (VFTT) was developed, capable of testing material compatibility with DME. This apparatus has the potential of selecting new materials for future DME pumps. Two properties are important for describing these lubrication qualities: The viscosity which is important in the hydrodynamic lubrication regime and the lubricity which is a measure of the lubrication performance in the boundary regime. These properties of DME were not easily established because in addition to the dissolving power of the fuel, it has to be pressurised in order to become liquid. The medium frequency pressurised reciprocating rig (MFPRR) has established the lubricity of DME to be very low but it is easily increased by additives in reasonable dosages. The viscosity of DME was established for the first time by the volatile fuel viscometer (VFVM). It is also very low for DME and can only be increased significantly by unrealisticly high dosages of additives. The literature indicates that the viscosity has a significant effect on the outcome of lubricity tests revealing a dominant in of hydrodynamic lubrication on this outcome. As the wear tests should predominately reflect boundary lubrication, this result indicates that they could only be used for fuels with a viscosity similar to that of diesel oil. If this is correct, the currently used lubricity tests cannot be used for low-viscosity fuels such as DME. The present study combines theoretical molecular dynamics (MD) calculations and practical lubricity tests to clarify the above. Linear alkanes of varying length were used as lubricants in MD calculations and the results revealed that longer alkanes are better lubricants than shorter ones when surfaces are separated by molecular thin films. These results were confirmed by MFPRR tests of alkanes. By inspection of the contacts in the MD calculations, it can be concluded that the length of the alkane is the primary property governing the wear amount. The viscosity is a secondary property as it is a function of the length of the molecule. This conclusion is supported by MFPRR tests of branched isomers of alkanes. At similar viscosity levels, branched alkanes have significant lower lubricity than the linear ones. This would not be possible if the viscosity was the primary wear controlling property. The above adds some very important knowledge to wear testing in general and in particular to the future of DME as a fuel. The results indicate that a new lubrication quasi-boundary regime may be present between the known boundary and hydrodynamic regimes. The film forming capacity of the lubricant seems to play an important role in the new regime as this property controls the squeeze out of the last lubricant layers in asperity contacts.