(87b) Determination of Diesel Properties through Comprehensive Multi-Dimensional GC×GC and SFC-TwinGC×GC

Refining processes produce varieties of petroleum streams which have to be characterized in order to evaluate their quality, to monitor the process or to ensure that the products meet the required specifications. For middle distillates, global properties (density, refractive index) are thus routinely measured in refining products.

The cetane number (CN), which characterizes the combustion behaviour of middle distillates in diesel engines, is one of the most stringent specifications for the European market (EN 590). The reference method (ASTM D613), which consists in the measurement of the fuel auto-ignition delay, is time consuming and requires a large amount of sample (ca. 1L).

Several predictive models for cetane number have been developed based on correlations with fuel global properties (density, simulated distillation). The main drawback of these global models is the lack of explicit link between the molecular composition of fuels and the cetane index which can be explained by the poor separation power of conventional analytical techniques.

Comprehensive two-dimensional gas chromatography (GC×GC) enables within one single injection the characterisation of diesel components by families (paraffins, naphthenes, mono-, di- and tri-aromatics) and within each class of components by groups of isomers. The integration of 2D-Chromatograms by means of a dedicated software (2DChromTM, developed by IFP and commercialised by Thermo Scientific) yielded approximately 150 groups of compounds. An average cetane number was then attributed to each group, based on experimental cetane numbers of model compounds obtained from literature. A linear regression was carried out in order to optimize this cetane number for each group of compounds taking into account chemical constraints (cetane number must be positive, normal paraffins have greater cetane numbers than iso-paraffins...) The cetane number of each diesel could thus be obtained from the molecular composition provided by GC×GC analysis, using a simple linear correlation. For most of the diesels investigated, excellent agreement was reached between the cetane numbers calculated from GC×GC and the reference engine values, which validates the strategy followed.

However, significant differences were observed for diesels enriched in polynaphthenes (3- and 4-ring naphthenes) and olefins. These differences can be attributed to an incomplete separation in GC×GC, due to co-elutions between polynaphthenes/mono-aromatics on one hand, and olefins/naphthenes on the other hand. To solve this problem, the analysis of such diesels was performed using Supercritical Fluid Chromatography (SFC) hyphenated on-line to GC×GC. Using this powerful technique, saturated and unsaturated compounds were separated and recovered in two different fractions, each of which could be analysed on-line by GC×GC. Therefore, polynaphthenes and olefins could be quantified by families (3- and 4-ring naphthenes, n-, iso and cyclo-olefins) and within each class of components by groups of isomers. A cetane number was attributed to each group following the same strategy as described for GC×GC. The cetane index obtained from SFC-twinGC×GC analysis of these diesels was found to be in excellent agreement with the engine ASTM method. Therefore, SFC hyphenated on-line to GC×GC enables to predict the cetane number of olefinic and naphthenic diesels, unlike single GC×GC.

In conclusion, a simple composition-based model was successfully developed for predicting the cetane number of diesel fuels. Due to the explicit link between the cetane number and the detailed composition obtained from GC×GC or SFC-twinGC×GC, cetane numbers can also be estimated for narrow cuts of diesel samples. This is very useful for blending studies.

To our knowledge, this is the first cetane number predictive model based on the molecular composition of middle distillates obtained through a single analysis. This strategy is currently being applied to the prediction of various diesel fuel properties (aromatic carbon content, cold flow properties...).