Blending Behavior of Hydrocarbon and Oxygenate Molecules to Optimize RON and MON for Modern Spark-Ignition Engines (SI)

<div class="section abstract"><div class="htmlview paragraph">Gasoline blending is known to be complicated, because individual gasoline fractions with different octane numbers, Research Octane Number (RON) or Motor Octane Number (MON) do not always blend linearly. Instead, they may blend non-linearly, in a synergistic or antagonistic manner. Even though RON and MON are regulated properties, linear and non-linear octane blending is not a broadly understood topic. The target in the developing process of a modern SI engine is to have 100% combustion efficiency which would lead to the reduction of hydrocarbon and carbon monoxide emissions. Therefore, the properties of gasoline, especially RON and MON, need to be optimized to ensure proper ignition in the engine and prevent harmful autoignition reactions.</div><div class="htmlview paragraph">There are hundreds of hydrocarbons in gasoline which have different octane numbers (ON). The explanations for these variations are the structural differences in hydrocarbon molecules that influence on their reactivity. For instance, longer n-paraffins have lower octane numbers compared to aromatics where electrons are delocalized around their ring which increases stability of aromatics and thus, ON. In this paper, we report and visualize qualitatively the octane blending behaviour of different hydrocarbon and oxygenate molecules to facilitate gasoline components mixing to produce high quality gasoline for clean combustion. The present study shows ethanol to blend non-linearly, but synergistically with paraffins and olefins, while the blending with aromatics is antagonistic. We also conclude that oxygenate molecules such as furans and cyclic ketones, blend synergistically with hydrocarbons. However, predicting the ON of end gasoline is challenging, as gasoline is not a blend of two components, but rather a blend of many isomers and functional groups. Therefore, in this study we highlight the need for more complex blending models than binary ones.</div></div>