TRAPPING HYDROPYROLYSATES ON SILICA AND THEIR SUBSEQUENT THERMAL DESORPTION TO FACILITATE RAPID FINGERPRINTING BY GC-MS

Abstract

Analytical hydropyrolysis performed under high hydrogen gas pressure (>10 MPa) has been demonstrated to possess the unique ability to release high yields of biomarker hydrocarbons covalently bound within the non-hydrocarbon macromolecular fraction of crude oils and source rocks. This study describes the development of the experimental procedure for trapping the product oils (hydropyrolysates) on silica to facilitate more convenient recovery than conventional collection and to allow analysis by thermal desorption-GC-MS without any prior work-up. Conventionally, the trap has consisted of a stainless steel coil, cooled with dry ice from which the products are recovered in organic solvents. Replacing this with a system in which the hydropyrolysates are adsorbed on a small mass of silica greatly reduces the turn-around time between tests, and aids the recovery and separation of the products. This method has been developed using an oil shale and an oil asphaltene fraction, with the silica trap producing very similar biomarker profiles to that from the conventional trap. The quantitative recovery of hydrocarbons from a light crude oil desorbed from silica under hydropyrolysis conditions demonstrates no significant loss of the high molecular weight n-alkanes (>n-C10) for both trapping methods. The use of liquid nitrogen as the trap coolant results in significantly improved recovery of the lower molecular mass constituents. The silica trapping method allows for the hydropyrolysates to be characterised by thermal desorption-GC-MS, which has been investigated both on- and off-line. The oils undergo relatively little cracking during desorption, with similar n-alkane and biomarker profiles being obtained as with normal work-up and GC-MS analysis. Thus, in terms of fingerprinting geomacromolecules, ''hypy-thermal desorption-GC-MS'' appears to have the potential to be developed as an attractive alternative to traditional py-GC-MS.