Mechanisms of sulfur dioxide and sulfuric acid neutralization in lube oil for marine diesel engines

Abstract

The maritime sector has seen notable changes in recent years, the most important being sailing at reduced speeds to save fuel, which has been possible due to the abundance of large vessels. This, combined with new engine designs and tunings for further fuel savings, has resulted in increased water and acid condensation onto the cylinder liner, promoting a combination of corrosion and wear, significantly reducing the liner lifetime. The condensing sulfuric acid (H2SO4) originates from the fuel bound sulfur. To counteract corrosion, the lubrication (lube) oil is formulated with additives (CaCO3-based reverse micelles). Nowadays, further engine improvements for better fuel efficiencies are desired, however, the effects on corrosive wear are difficult to predict. In addition, switching fuels frequently is being introduced driven by availability, price, and stricter regulations on emissions. To cope with legislation, either sulfur-poor fuel can be combusted or sulfur-rich fuel with subsequent treatment with an exhaust after-treatment technology. These improvements/changes put additional demand on lubricating the engine optimally because defective lubrication has severe consequences. To understand how to optimally lubricate the engines, knowledge on what happens to the sulfur-related species (SO2 and H2SO4) in the lube oil film is required. The work presented in this thesis is an investigation of the neutralization of SO2(g) and H2SO4(aq) by reaction with CaCO3(s) reverse micelles in a fully formulated lube oil. The reactions are studied individually and by experiments and mathematical modeling. The first part of the thesis investigates experimentally the reaction between H2SO4 and CaCO3 in lube oil in a mixed flow reactor (MFR) setup by varying the Ca/S molar ratio, H2SO4(aq) inlet concentration, residence time, and stirrer speed. The analysis methods applied were Fourier Transform Infrared (FTIR) spectroscopy and a titration method to quantify the conversion of CaCO3 and H2SO4 at specific conditions. The results revealed that the first step of the reaction was emulsification of the H2SO4 into the lube oil followed by reaction between the solubilized H2SO4 droplets and CaCO3 reverse micelles. For the residence times investigated, it was observed that the reaction between H2SO4 and CaCO3 was significantly reduced when reaching a critically low Ca/S molar ratio. A certain degree of stirring was found to initiate and maintain the reaction. Also, no apparent effect of varying the residence time was observed. Diluting the inlet H2SO4 concentration led to a decreased conversion of CaCO3, probably due to the introduction of a large amount of water, leading to poor solubilization of the H2SO4 droplets. The second part of the thesis concerns mathematical modeling of the experimental MFR data for the reaction between H2SO4 droplets and CaCO3 reverse micelles. It was difficult to identify conclusively the limiting step of the neutralization reaction, however, a quantitative description of the reaction rate and its temperature-dependency was determined. The validated mathematical model was used to predict conversion of H2SO4 in a lube oil for conditions relevant for a full-scale application. The calculations show that H2SO4 may interact with the cylinder liner surface regardless of how well-wetted the surface is. The third and last part of this thesis describes a study of the mechanism underlying the reaction between gaseous SO2 and CaCO3 reverse micelles in lube oil in a pressurized stirred batch reactor setup. The first step of the mechanism is the absorption of gaseous SO2 into the lube oil emulsion, followed by reaction with CaCO3 reverse micelles. The reaction shows a dependence on the initial water concentration due to increased SO2 absorption in the lube oil emulsion. The overall temperature dependence on the reaction was observed to be weak because the absorption of SO2 decreases at increased temperature. It was also observed that CaSO3 was initially formed, followed by formation of CaSO4 at extended residence times and increased temperature. A mathematical model was derived and kinetic parameters were determined by fitting of the model to experimental data. The batch reactor model was used to predict the CaCO3 conversion in a lube oil emulsion from SO2 for worst-case conditions relevant for a full-scale application. The simulations revealed that consumption of CaCO3 by SO2 is insignificant in a two-stroke marine diesel engine application; the H2SO4-CaCO3 reaction is much faster than the SO2-CaCO3 reaction. The work contributes with insight on how H2SO4 is neutralized by CaCO3 reverse micelles in the lube oil film and that SO2 is not a concern with respect to CaCO3 consumption. The tools developed can be used in models of complete two-stroke marine diesel engines, including estimation of the condensation rate of H2SO4 and the corrosion rate of the cylinder liner surface by unreacted H2SO4 in the lube oil emulsion, with the aim of determining optimal lubrication strategies.