Application of light hydrocarbons (C₄–C₁₃) to oil/source rock correlations:

Abstract

Relatively little work has been published on the correlation between the light hydrocarbon distributions in reservoir fluids and their proposed source rocks [Philippi, G. T. (1981)]. The aim of our work was to study this relationship in detail for samples from Mid-Norway. The main source rocks offshore Mid-Norway are the marine shales of the Late Jurassic Spekk Formation and the coals and paralic shales of the Early Jurassic Åre Formation. Reliable light hydrocarbon (C₄–C₁₃) data for source rock samples were acquired by thermal extraction-GC of the source rocks. Of these, notably the hydrocarbons in the C₆–C₈ range (routinely measured in test fluids) were used to discriminate between the Spekk and Åre Formation samples. A total of twenty-six samples from the Spekk Formation and twenty-four samples from the Åre Formation at different maturity levels and facies were analyzed. In general, the two source rock types differ in their light hydrocarbon composition by the presence of relatively more aromatics and cyclohexanes in the Åre samples, while the Spekk samples are richer in cyclopentanes and acyclic hydrocarbons. We show that source rock facies is a more important indicator of light hydrocarbon composition than maturity variation. Differences in the chemical composition, which can be used to discriminate between the two source rocks, were supported by differences in the carbon isotope composition of individual components of the same fraction, as determined by GHM-IR-MS analysis of eleven samples. Further, the light hydrocarbon compositions of reservoir fluids (oils and condensates) were compared with those for the source rock(s). Sixty-six gas chromatograms of oils and condensates, representing most of the known petroleum accumulations in Mid-Norway, were collected. Of these, five oil samples were selected for detailed isotopic analysis of individual components (GC-IR-MS). When using a classification scheme based on data from sediment samples, data for the light hydrocarbon fraction of oils and condensates indicate that the Spekk Formation is the dominant source for most of the fields from Mid-Norway, with a significant contribution from the Åre Formation being detected principally in one field. Differences in the chemical composition of the C₆–C₈ fractions were supported by differences in the carbon isotope composition of individual components, which also discriminate between the oils. Although the classification diagrams used in this study are based on source rock data from Mid-Norway, the method can be applied to other areas, providing that the diagrams are calibrated with source rock data from the area of interest.

Introduction

Two main source rocks are present offshore Mid-Norway: the marine shales of the Late Jurassic Spekk Formation and the coals and paralic shales of the Early Jurassic Åre Formation (Fig. 1). The principal reservoir units of this region are the sands of the Middle and Early Jurassic Fangst and Båt Groups, as well as the Late Jurassic Rogn Formation in the Draugen Field. A location map of the main petroleum accumulations offshore Mid-Norway is illustrated in Fig. 2.

Documentation of the importance of the Spekk and Åre Formations as sources of the petroleum has been published by several authors (Elvsborg et al., 1984; Ellenor and Mozetic, 1986; Cohen and Dunn, 1987; Hvoslef et al., 1988; Khorasani, 1989; Whitley, 1992). The Spekk Formation is of marine origin and contains kerogen which varies from type II to type III (Whitley, 1992). Johannesen (1995) divided the Spekk Formation chronostratigraphically into the Upper Spekk Unit, which contains dominantly type II to II/III kerogen and the Lower Spekk Unit with type III to IV kerogen. The Lower Jurassic Åre Formation consists of alternating sandstones and claystones which are interbedded with coals. The organic matter is dominated by terrestrially-derived humic material (vitrinite and inertinite) and is classified as a type III to type IV kerogen with a high potential for generation and expulsion of gas/condensate (Odden, 1986; Cohen and Dunn, 1987; Hvoslef et al., 1988; Khorasani, 1989). The petrographical composition of the Åre Formation coals and shales is rather similar. Vitrinite is the major component, inertinite varies between 10 and 50% and the exinite from minor to 10%. The most abundant exinite submaceral is sporinite with a variable input of cutinite and resinite (Odden, 1986). However, to explain the volumetrics of the petroleum accumulations in Mid-Norway, some authors have used the coals and shales of the Åre Formation as a supplementary source rock to the Spekk Formation, or even as the main oil contributor (Hollander, 1984; Hagevang and Rønnevik, 1986; Heum et al., 1986; Ekern, 1987, Forbes et al., 1991; Espitalié et al., 1991), based primarily on basin modelling and kinetic studies. Odden (1986), Cohen and Dunn (1987), Khorasani (1989) and Karlsen et al. (1995) did not agree with this assertion and claimed that it is debatable whether Åre Formation coals and shales can act as an oil source rock based on correlation of geochemical data for source rock extracts with oils.

The published geochemical data of the oils and condensates from Mid-Norway (Ellenor and Mozetic, 1986; Cohen and Dunn, 1987; Goesten and Nelson, 1992; Provan, 1992; Karlsen et al., 1995) focus on the C₁₅+fraction of the petroleums, with little or no attention to the intermediate fraction (C₆–C₁₄), which is more abundant in condensates and oils. Differences between individual oils have been explained by Karlsen et al. (1995) as reflecting variations in the extent of the terrestrial influence in the Spekk Formation, rather than contributions from the Åre Formation. Further, they concluded that the C₁₅+fractions of all reservoired oils and condensates from Mid-Norway originate predominantly from the Spekk Formation. Where the origin of the light “condensate-range” material has been discussed, the general opinion appears to be that the Åre coals have contributed significant amounts of condensate (Elvsborg et al., 1984; Cohen and Dunn, 1987; Hvoslef et al., 1988; Whitley, 1992), but these opinions are not supported by analytical data on the light hydrocarbon fraction.

Hunt et al. (1980) noted that the light hydrocarbons make up about 30% of a crude oil, and therefore information derived from these components is more representative of the bulk of an oil than that from the biological markers, which may make up 1%. Classification by light hydrocarbons may also be used to gain information from samples where biological marker data are unavailable or unobtainable.

Light hydrocarbon data from oils and condensates from other locations have been studied by several authors. Thompson, 1979, Thompson, 1983, Thompson, 1987, Thompson, 1988 published a number of papers which provide source and maturity information on the light fraction (C₆–C₈) of unaltered petroleums, and which interpret transformation processes in the reservoir such as bacterial degradation and phase fractionation. Mango, 1987, Mango, 1990a, Mango, 1990b, Mango, 1994, Dai Jinxing (1992) and Ten Haven (1996) claimed that some components of the C₇ hydrocarbon fraction in petroleum can be used to deduce the organic matter type in the source rock. Halpern (1995) used ratios of C₇ compounds to correlate oils and condensates and to assess relative degrees of post-generative transformation among related hydrocarbons.

A variety of analytical techniques have been developed to extract light hydrocarbons from source rock samples prior to GC analysis. Durand and Espitalié (1972) transferred the C₁–C₅ fraction to a mercury container after vacuum crushing of the sample and the C₆–C₁₅ hydrocarbons were extracted by CFCl₃ in a sealed crusher. Jonathan et al. (1975) employed a process of thermovaporisation at 220°C to extract C₆–C₁₅ hydrocarbons from sedimentary rocks. Leythaeuser et al., 1979a, Leythaeuser et al., 1979b analyzed the C₂–C₇ hydrocarbon fraction by a hydrogen stripping/capillary gas chromatographic technique. However, of these papers, only Leythaeuser et al., 1979a, Leythaeuser et al., 1979b distinguished between different source rock types. They concluded that organic matter derived from higher land plants contained a relatively higher proportion of benzene and toluene than organic matter of marine origin. These authors also mentioned the evidence of loss of light hydrocarbons due to improper sample storage.

Philippi (1975), Jonathan et al. (1975) and Thompson, 1979, Thompson, 1983, Thompson, 1987 used the C₆–C₇ hydrocarbon fraction to interpret maturity and showed that the alkane/cycloalkane ratios increased with increasing maturity. Thompson (1983) proposed that the C₇ cycloalkanes progressively undergo ring opening with increasing temperature. Thus, the ratio nC₇/MCC₆ was offered as a maturity index, presumably increasing with subsurface temperature as MCC₆ undergoes thermal ring opening. Mango, 1990a, Mango, 1997 questioned this explanation and, in opposition to Thompson, claimed that cycloalkanes are more stable than open chain alkanes. However, Mango stated that the alkane/cycloalkane ratios increase with maturity, but the reasons are not clear.

Relatively little work has been published on the correlation between light hydrocarbon distributions in reservoir fluids and source rocks. Philippi (1981) compared the C₆–C₇ hydrocarbon fractions of crude oils with those for source rocks through ten “similarity coefficients”. The latter were calculated to vary from 0 to 1 with increasing correlation. Kornacki (1993) analyzed the C₇ hydrocarbon fraction in oils and source rocks and used the sum of methylhexanes vs the ratio of 1,3-dimethylcyclopentanes to dimethylpentanes (after Mango, 1994) for interpretation of the data.

In our study, the light hydrocarbon compositions of source rocks and test fluids from Mid-Norway were examined. We selected Mid-Norway as the study area, since a comprehensive data base of light hydrocarbon and other fluid data already exists. In order to facilitate the correlation with possible source rocks, the initial part of the project focussed on obtaining reliable light hydrocarbon data for source rock samples from the Spekk and Åre Formations at different maturity levels and facies. The light hydrocarbon fractions of source rock samples were thus analyzed by thermal extraction-GC, using the method described by Bjorøy and Solli (1984) and Bjorøy et al. (1984), with quantification of the individual components in the C₄–C₁₃ range. Of these, notably the hydrocarbons in the C₆–C₈ range (routinely measured in test fluids), were used to discriminate between the Åre and Spekk Formation samples. The light hydrocarbon compositions of reservoir fluids (oils and condensates) from Mid-Norway were interpreted with regard to their source rock(s). For this purpose, gas chromatograms of oils and condensates were used with quantification of those components of the C₆–C₈ fraction which are present in relatively higher proportions.

The calibration of classification diagrams for distinguishing between the Spekk and Åre Formation samples is based upon results from multivariate modelling (PCA) and the previous work by Mango, ten Haven and Dai Jinxing referred to above. Further, the source of the light hydrocarbons in test fluids from Mid-Norway was assessed by the parameters and plots which distinguished the Spekk from the Åre Formation samples.

Multivariate modelling indicates that the light hydrocarbon fraction is most strongly influenced by source rock type, and less so by maturity. However, the light hydrocarbon distributions in source rocks can be potentially affected by losses due to evaporation of these volatile components. The storage effect was studied here.

Interpretation of the compound distribution was augmented by compound specific carbon isotope compositions of selected Åre and Spekk Formation samples. These were measured by coupling a Geofina Hydrocarbon Meter (Bjorøy et al., 1984) with an isotope ratio mass spectrometer (GHM-IR-MS), which made it possible to analyze the individual components of thermal extracts without prior sample preparation. The carbon isotope composition of individual components of five oil samples was measured by GC-IR-MS (Bjorøy et al., 1990, Bjorøy et al., 1991, Bjorøy et al., 1992; Hansen et al., 1995) and these data have also been compared with corresponding data for source rock thermal extracts.