Compositional Alterations of Petroleum As a Result of Expulsion and Migration in the North Sea Central Graben Petroleum System

Compositional alterations of petroleum as a result of expulsion and migration in the North Sea Central Graben petroleum system

vorgelegt von M.Sc. Geologe Volker Ziegs geb. in Karl-Marx-Stadt (heute: Chemnitz)

von der Fakultät VI – Planen Bauen Umwelt der Technischen Universität Berlin zur Erlangung des akademischen Grades

Doktor der Naturwissenschaften - Dr. rer. nat. -

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender: Prof. Dr. Wilhelm Dominik Gutachter: Prof. Dr. Brian Horsfield Gutachter: Prof. Dr. Reinhard Sachsenhofer (MU Leoben)

Tag der wissenschaftlichen Aussprache: 20. Juli 2018

Berlin 2018

Dedicated to science

“Cogito ergo sum – I think therefore I am.” - Sean-Paul Descartes -

Table of Contents

Table of Contents ...... I

Acknowledgements ...... V

Abstract...... VII

Zusammenfassung ...... XI

List of Publications ...... XV

List of Figures ...... XVII

List of Tables ...... XXIII

List of Abbreviations ...... XXV

1 Introduction ...... 1 1.1 The history of organic geochemistry in petroleum system evaluation ...... 1 1.2 Origin of organic matter ...... 4 1.2.1 Characterization of organic matter ...... 4 1.2.1.1 Optical analysis ...... 4 1.2.1.2 Elemental analysis and Rock-Eval© ...... 6 1.2.1.3 Pyrolysis-gas chromatography ...... 8 1.2.1.4 Kinetic characterization ...... 10 1.2.2 Natural origin of NSO compounds ...... 11 1.3 Maturation of organic matter ...... 13 1.3.1 Compositional evolution of organic matter ...... 13 1.3.2 Common maturity markers ...... 16 1.3.3 Timing of petroleum formation ...... 17 1.4 Primary migration & expulsion ...... 18 1.4.1 Mechanisms of primary migration and expulsion ...... 19 1.4.2 Migration fractionation ...... 22 1.5 Secondary migration from source to reservoir ...... 23 1.5.1 Mechanisms of secondary migration ...... 23 1.5.2 Processes of chemical fractionation ...... 24 1.5.3 Phase behaviour during up-rise ...... 26 1.6 The scope of this thesis ...... 28 1.6.1 Sample selection ...... 30 1.6.2 The structure of this dissertation ...... 31 2 Petroleum retention in the Mandal Formation, Central Graben, Norway ...... 33 2.1 Abstract ...... 33 2.2 Introduction...... 34 2.3 Databases, samples and methods ...... 38

2.3.1 Regional geochemistry database ...... 38 2.3.2 Choosing samples for compositional analysis ...... 39 2.3.3 Sample preparation and screening...... 39 2.3.4 Thermovaporisation-gas chromatography ...... 40 2.3.5 Pyrolysis-gas chromatography ...... 40 2.3.6 Bulk kinetics ...... 41 2.4 Bulk source rock characteristics ...... 41 2.5 Petroleum-type organofacies ...... 47 2.6 Generation characteristics ...... 48 2.7 Petroleum expulsion ...... 52 2.7.1 Hydrocarbon retention profiles ...... 52 2.7.2 Mass balance and total liquid petroleum profiles ...... 56 2.7.3 Linking retention to organic matter characteristics ...... 60 2.7.3.1 Gas retention ...... 60 2.7.3.2 Total oil retention ...... 65 2.8 Conclusions ...... 68 2.9 Acknowledgement ...... 69

3 Unravelling maturity- and migration-related carbazole and phenol distributions in Central Graben crude oils ...... 71 3.1 Abstract ...... 71 3.2 Introduction ...... 72 3.3 Geological setting & hydrocarbon habitat ...... 74 3.4 Samples & methods ...... 76 3.4.1 Sample set ...... 76 3.4.2 Analytical methods ...... 77 3.5 Results & discussion ...... 79 3.5.1 Oil and condensate properties using conventional parameters ...... 79 3.5.1.1 Variations of API with depth ...... 79 3.5.1.2 Gross chemical composition ...... 82 3.5.1.3 Molecular geochemistry of hydrocarbons ...... 83 3.5.2 The polar NSO fraction of petroleum ...... 86 3.5.2.1 Gross composition ...... 88 3.5.2.2 Compound class distribution ...... 88

3.5.2.3 N1 compounds & maturity assessment ...... 89

3.5.2.4 Correlating GC-MS maturity and degree of N1 annulation ...... 91 3.5.2.5 Detailed variations due to reservoir lithology ...... 93

3.5.2.6 Other influences on N1 chain length distributions...... 96

3.5.2.7 Factors controlling O1 compounds ...... 99 3.6 Conclusion ...... 104 3.7 Acknowledgements ...... 106

3.8 Appendix ...... 106 3.8.1 Maturity correlations: background information ...... 106

3.8.2 Maturity correlations of O1 DBE 4 and 5 compounds ...... 107 4 Deeper insights into oxygen-containing compounds of the Mandal Formation, Central Graben, Norway ...... 109 4.1 Abstract ...... 109 4.2 Introduction...... 110 4.3 Geological evolution of the relevant petroleum system elements ...... 112 4.4 Samples & methods ...... 115 4.4.1 The sample set ...... 115 4.4.2 Analytical methods ...... 116 4.5 Results ...... 119 4.5.1 Environmental information from the hydrocarbon fraction ...... 119 4.5.2 Gross composition of source rock extracts ...... 120 4.5.3 Elemental class distribution ...... 121 4.5.4 Characterization of selected compound classes ...... 124 4.6 Discussion ...... 132 4.6.1 On the origin of elevated oxygen content in Mandal extracts ...... 132 4.6.2 Broad inferences regarding petroleum expulsion efficiency ...... 136 4.7 Conclusion ...... 138 4.8 Acknowledgements ...... 139

5 Summary & perspectives ...... 141 5.1 Summary ...... 141 5.1.1 The Mandal source rock ...... 141 5.1.2 Central Graben crude oils ...... 144 5.1.3 Migratability of high-molecular weight generated petroleum ...... 146 5.2 Perspectives ...... 148 5.2.1 Retention in and expulsion from the source rock ...... 148 5.2.2 Crude oils ...... 150 6 References ...... 151

III

Acknowledgements

The first contact with organic geochemistry can be traced back to Thomas Schwarzkopf’s and Jürgen Rückheim’s petroleum geochemistry course at TU Göttingen. Therefore, I want to thank those gentlemen for introducing this interesting research field to me. But it was a lucky coincidence to meet Prof. Dr. Brian Horsfield at Oslo airport discussing about gas shales and finally offered me the chance to conduct exciting research projects at the GFZ Potsdam. Thus, my very gratitude goes to him and also to Prof. Dr. Rolando di Primio for guiding me through the deep waters of organic geochemistry. Providing your experience, expertise, and patience at every working and sleeping hour enabled me to grow scientifically and personally over the past years.

My sincere gratitude is extended to Drs. Nicolaj Mahlstedt, Mareike Noah, Stefanie Pötz and to Dr. PD Hans-Martin Schulz for the fruitful and inspiring discussions. Furthermore, it would have been impossible to carry out my research without the professional and reliable technical support of Ferdinand Perssen, Cornelia Karger, Anke Kaminsky, Kristin Günther and Claudia Engelhardt.

Aker BP ASA, in particular Drs. Jon Erik Skeie, Joachim Rinna and Alexander Hartwig, are acknowledged for financial support, providing the data and showing great interest in this project.

Scientific results and discussions sometimes turn in multiple cycles which might be frustrating. Thus, I am very grateful of having been part of a highly volatile structure called “Cyclopentane” consisting of Janina Stapel, Sascha Kuske, Seyed Hossein Hosseini Baghsangani, Shengyu Yang, and my humble self. Your company balanced life in science and beyond in the accurate level: entertainment, friendship, and trust.

A special thank goes to my family and friends for their constant support and encouraging words since the past years at university and during the doctorate. Robby Gwizdz is greatly appreciated for lending his courage to me when needed.

Finally, I want to express my highest gratitude to my dear sweetheart Maria Bade who always supported me, suffered in bad times and enjoyed the good times. You have been the most patient with me and I am thankful that you are by my side.

Abstract

Chemical and physical processes affect the composition of petroleum during its generation, expulsion and migration within a sedimentary basin. The understanding of these processes is crucial for correct assessment of the prospectivity of a given petroleum system. The well-explored North Sea Central Graben developed with a relatively simple burial history and contains numerous oil and gas/condensate fields which were charged from differently mature Upper Jurassic marine shale source rocks. The organic matter-rich Mandal Formation, the principal source rock in that basin, has been identified by geochemical screening and mass balance modelling as a locally very inefficient expeller of its generated products, different to other prolific marine shales. This makes the Mandal Formation an interesting natural laboratory to re-assess the factors controlling generated, retained and expelled fluid compositions as a function of maturity and kerogen type, and to unravel processes affecting compositional fractionation during expulsion of different evaporative fractions.

The present dissertation addresses structural characteristics of initial sedimentary organic matter (OM), retained and expelled portions of generated petroleum and draws correlations to amounts of retained products. Using source rock samples, their solvent extracts and expelled crude oils, the approach combines bulk chemical methods, such as Rock-Eval pyrolysis, and compositionally resolving analyses, such as pyrolysis-gas chromatography-flame ionization detection (Py-GC- FID), coupled gas chromatography-mass spectrometry (GC-MS) and the ultra-high resolution Fourier Transform-Ion Cyclotron Resonance-MS (FT-ICR-MS). This enables the evaluation of the hydrocarbon fraction and the polar heteroelemental inventory containing low- to medium- to high-molecular weight compounds.

Source rock evaluation was conducted on different stratigraphic levels and locations of the Mandal Formation within the Central Graben. Containing mainly Type II OM of marine algal origin with up to 12 % TOC content and up to 647 mg HC/g TOC, the Mandal Formation generates a Paraffinic-Naphthenic- Aromatic Low Wax oil. While the source rock’s richness shows two modes at 2 and 5.5 % TOC, generation potentials are uniformly distributed. Aromaticity and aliphatic chain length distributions of source rock pyrolysates are mainly a function

VII of maturity, but can as well be attributed to varying degrees of oxygen deficiency during deposition or to variable input of different organic matter types. Bulk kinetic investigations revealed narrower activation energy distributions, thus more homogenous OM at marginal basin locations, while they are broader in axial basin sites, leading to differential generation and expulsion behaviour at different basin locations.

The Mandal Formation retains elevated amounts of free hydrocarbons in its high-molecular weight organic matter (HMW OM). Retention capacity increases as function of thermal maturity. Being able to separate the complex OM fraction into the true generative kerogen and generated heavy and immovable petroleum rich in

NSO compounds (S2bitumen), we were successful in attributing different retention modes to these fractions. While the true kerogen portion controls retention of gaseous hydrocarbons by the amount of its cross-linked monoaromatic sites, generally addressed as aromaticity, oil retention is strongly dependent on the relative proportion of the HMW bituminous OM fractions, affecting physical properties such as viscosity und thus movability. This is an important constituent of OM in the Mandal Formation constituting 24 – 63 % of its conventional S2 and possessing a distinct generation potential. Its compositional characterization is conducted using ESI negative FT-ICR-MS analysing medium- to high-molecular weight (m/z > 200 Da) polar compounds that contain acidic heteroatoms. However, the bulk compositional comparison with retained extracts of the very efficiently expelling marine Posidonia Shale suggests that it is not the amount of the NSO fraction but its composition that affects retention and expulsion behaviour of the free, mobile fluid phase, the S1. Retained petroleum in the badly expelling Mandal Formation is enriched in highly polar compounds containing 2 to 6 functional oxygen atoms incorporated in aliphatic chains and/or aromatic ring systems. Possible reasons for the high oxygen content are discussed on a local basis and might be attributed to the input of autochthonous bacterial and/or allochthonous terrigenous OM, as indicated by locally varying abundances of C20 aromatic and C16 and C18 saturated acids, or due to variations in the paleogeography and climate specific for the Central Graben and affecting the composition of the algal body and other microorganisms. Composed of highly aromatic, polar oxygen compounds in general, Mandal Formation extracts contain long aliphatic chains attached to the aromatic core structures, thus reducing their aromatic character. As a consequence

VIII of their high polarity as well as sheer size and awkward shape, interaction with other polar sites in the source rock, such as hydrous clay minerals, residual organic matter or molecular water, is strong. Migration of such molecules in the Mandal Formation shales is restricted until effective bulk fluid migration sets in during the main generation phase.

Selected crude oils in the Central Graben contain less amounts of polar compounds than source rock extracts and are composed of lower polar constituents

(N1 > O1 >> O2, N1O1). Although varying in API vs. depth gradients, crude oils from clastic and carbonate reservoir lithologies do neither show significant variations in their SARA gross composition (saturates / aromatics / resins & asphaltenes) nor in their hydrocarbon inventory. However, carbazole distributions as part of the polar NSO fraction show subtle changes. Relative distributions of carbazole homologues serve as maturity indicators and correlate well with hopane maturity ratios (29Ts / [29Ts + NH]). While variations in lower-fused homologues can be attributed to subtle maturity differences, the loss of short-chained benzocarbazoles in oils from carbonate reservoirs at the eastern basin margin might be a result of secondary migration effects. Polar compounds in the migrating phase interact with other polar phases or surfaces such as water or clay minerals. This effect is reduced by longer aliphatic side chains attached to the core structures. Similar trends have been observed in the water-soluble and surface-adhesive phenolic species. Size and shape of migrating molecules, determined by the number of aromatic rings and length of aliphatic side chains, but mainly the activity of polar sites affect the chemical composition of migrating phase. The preferential removal of small, aromatic but very polar species during secondary migration in initially water-filled pore spaces of different carrier lithologies have an impact on density and phase behaviour of crude oils and hence significantly influence the in-reservoir oil quality.

Zusammenfassung

Während der Genese von Erdöl, dessen Ausscheidung aus (Expulsion) bzw. Rückhaltung (Retention) in dem Muttergestein und der Migration innerhalb eines Sedimentbeckens verändern chemische und physikalische Prozesse die molekulare Zusammensetzung des Öls. Das Verständnis um diese Prozesse ist für die korrekte Beurteilung der Höffigkeit eines Erdölsystems entscheidend. Der Zentralgraben in der Nordsee ist ein sehr gut untersuchtes Sedimentbecken mit einfacher Versenkungsgeschichte, in welchem zahlreiche Erdöl- und Gas/Kondensat-Felder erschlossen sind. Diese sind von Oberjurassischen Schiefergesteinen gespeist, die reich an organischer Materie marinen Ursprungs sind und unterschiedliche thermische Reifegrade erreicht haben. Die Mandal Formation als Hauptmuttergestein im Zentralgraben wurde mittels Massenbilanzierung, basierend auf zahlreichen Rock-Eval-Analysen ihres Genesepotentials und der gebildeten Mengen, als ein Muttergestein mit lokal hohem Retentionspotential charakterisiert. Dieses Verhalten ist für derartige marine Schiefer untypisch. Aufgrund dessen ist die Mandal Formation ein geeigneter Kandidat, um physikalisch-chemische Einflussfaktoren auf die Erdöl-Zusammensetzung in Reservoiren und im Muttergestein zu differenzieren sowie in einzelne Prozesse aufzutrennen. Die chemische Fraktionierung während der Expulsion verschiedener chemischer Komponentengruppen wird in Abhängigkeit der thermischen Reife und des Kerogentyps des organischen Ausgangsmaterials analysiert.

Die vorliegende Dissertation befasst sich mit den chemischen Struktur der sedimentären organischen Materie (Kerogen), den daraus gebildeten, zurückgehaltenen Anteilen sowie dem abgeschiedenen Erdöl und korreliert Qualität und Quantität des generierten Petroleums. Muttergesteinsproben, deren extrahierbare organische Materie und produzierte Erdöle werden mittels quantitativer sowie qualitativ hochauflösender Methoden untersucht. Es wurden Rock-Eval-Pyrolyse, Pyrolyse-Gaschromatographie-Flammenionisierungsdetektor (Py-GC-FID), GC-gekoppelter Massenspektrometrie (GC-MS) und hochauflösende Fourier-Transform-Ionenbeschleunigungsresonanz-Spektrometrie mit gekoppelter negativer Elektrospray-Ionisierung (ESI negativ FT-ICR-MS) eingesetzt. Dies ermöglicht die Bewertung der Kohlenwasserstoff-Fraktionen sowie des Anteils niedrig- bis hochpolarer (NSO)-Heteroelemente.

Die Evaluierung des Muttergesteinspotentials der Mandal Formation beinhaltet marginale und zentrale Beckenlokationen im Zentralgraben und erfolgt innerhalb verschiedener stratigraphischer Level. Mit einem TOC-Anteil bis zu 12 % und einem Genesepotential von bis zu 647 mg HC/g TOC lässt sich die Mandal Formation dem Kerogentyp II zuordnen und generiert ein Paraffinisch- Naphthenisch-Aromatisches Öl mit niedrigem Wachs-Gehalt. Während der Kohlenstoffgehalt zwei Modi bei 2 und 5,5 % TOC zeigt, ist das Genesepotential unimodal verteilt. Aromatizität und Alkylkettenlängenverteilung in Pyrolysaten des Muttergesteins folgen hauptsächlich Reifetrends. Dies kann einer Variabilität des anoxischen Milieus oder des Eintrags organischer Materie während der Ablagerung zugeschrieben werden. Gesamtkinetische Untersuchungen des Kerogens zeigen engere und breitere Verteilungen der Aktivierungsenergien, welche einem homogeneren Eintrag organischer Materie, vorzüglich am Rand des Sedimentbeckens, bzw. einem heterogeneren organischen Eintrag im Beckenzentrum gleichgestellt werden können. Die beobachtete Variabilität führt zu verschiedenem Genese- und Expulsionsverhalten an unterschiedlichen Beckenpositionen.

Die Mandal Formation hält erhöhte Mengen generierter, freier Kohlenwasserstoffe in seiner hochmolekularen organischen Materie zurück, wobei zurückgehaltene Mengen und thermische Reife positiv korrelieren. Die Unterteilung der komplexen organische Materie in makromolekulares Kerogen und in schwere, NSO-reiche und schlecht migrierbare Bestandteile seines

Geneseprodukts (S2bitumen) ermöglichte es, unterschiedliche Retentionsprozesse zwischen den Fraktionen auszuarbeiten. Gasförmige Kohlenwasserstoffe werden an der vernetzten, monoaromatischen Oberfläche des Kerogens, gemeinhin als Aromatizität bezeichnet, adsorbiert, während die Retention flüssiger Komponenten stark mit den relativen Anteilen der hochmolekularen, bituminösen Komponenten des Erdöls zusammenhängt. Hierdurch werden seine physikalischen Eigenschaften wie Viskosität und schließlich die Migrationseffizienz bestimmt. Diese wichtige Fraktion der hochmolekularen organischen Materie beträgt 24 – 63 % des konventionellen Rock-Eval S2-Wertes eines nicht-extrahierten Muttergesteines der Mandal Formation und besitzt ein nicht weiter quantifiziertes Genesepotential für Erdöl. Die hochmolekulare, polare Heterofraktion mit Massen ab 200 Da und einem aziden Wasserstoffatom wird mit Hilfe von FT-ICR-MS im negativen ESI-

XII

Modus kompositionell aufgelöst. Zusammensetzungen der generierten Produkte wird mit dem Norddeutschen Posidonienschiefer verglichen, welcher das generierte Erdöl sehr effizient abscheidet. Die Retentionskapazität des Muttergesteins wird nicht durch die Quantität, jedoch die Zusammensetzung seiner NSO-Fraktion bestimmt. Zurückgehaltene mobile oder freie Geneseprodukte (Rock-Eval S1) werden in der Mandal Formation von hochpolaren Sauerstoffverbindungen dominiert, welche 2 bis 6 funktionelle O-Atome enthalten und in aliphatische Ketten und/oder aromatische Ringsysteme eingebunden sind. Mögliche Gründe für den hohen Sauerstoffanteil der organischen Materie werden anhand lokaler Fallbeispiele diskutiert und sind möglicherweise durch den Eintrag bakteriellen (autochthon) und/oder terrigenen organischen Materials (allochthon) verursacht, was sich durch variierende relative Konzentrationen an aromatischen C20 und gesättigten C16 und C18 Carbonsäuren ausdrückt, oder kann Klimavariationen sowie paläogeographischen Gegebenheiten zugeordnet werden, die spezifisch für den Zentralgraben im Oberen Jura sind und Auswirkungen auf die Zellstruktur von Algen und anderen Mikroorganismen hat. Obwohl hocharomatische, polare Sauerstoffkomponentengruppen die Zusammensetzung dominieren, sind sie durch lange aliphatische Ketten an den aromatischen Kernstrukturen gekennzeichnet. Dadurch wird der aromatische Charakter reduziert. Aufgrund ihrer hohen Polarität, schieren Größe und bizarren Form ist eine Interaktion mit anderen polaren Muttergesteinskomponenten, wie z.B. kristallwasserhaltigen Tonmineralen, remanenter organischer Materie oder molekulares Wasser, stark ausgeprägt. Bis zum Beginn des Ölfensters mit extrem effektiver Erdölmigration aufgrund der gebildeten Mengen ist die Migrationseffizienz in der Mandal Formation stark eingeschränkt.

Natürliche Erdöle in Reservoiren enthalten weniger polare Bestandteile als Muttergesteinsextrakte, und heterozyklische Verbindungen sind zudem von niedrigerer Polarität (N1 > O1 >> O2, N1O1) gekennzeichnet. Ausgesuchte Öle aus dem Zentralgraben wurden von siliziklastischen und karbonatischen Reservoiren produziert und weisen in den verschiedenen Lithologien unterschiedliche Tiefen- Gradienten für die Dichte auf. Jedoch gibt es weder signifikante Unterschiede in deren Aliphaten-Aromaten-NSO-Verhältnis noch in den Zusammensetzungen der Kohlenwasserstoff-Fraktion. Lediglich die NSO-Fraktion zeigt interessante Variationen auf. Relative Häufigkeiten von Carbazolen und ihren höher-

XIII annelierten Homologen innerhalb der N1-Komponentenklasse dienen als Reifeindikatoren und korrelieren mit jenen Indikatoren aus der Kohlenwasserstofffraktion (hier: 29Ts/[29Ts+NH]). Während Variationen der relativen Konzentrationen niedriger-annelierter Carbazol-Homologe leichten Reifeunterschieden zu zuordnen sind, kann der Verlust kurzkettiger Benzocarbazole in Karbonat-Reservoiren am östlichen Beckenrand eine Folge von sekundären Migrationseffekten sein. Polare Komponenten der migrierenden Phase interagieren mit anderen polaren, organischen oder anorganischen Stoffen bzw. Oberflächen, wie z.B. Tonmineralen oder Wasser. Lange aliphatische Seitenketten an den heterozyklischen N-Verbindungen verringern diese Interaktion. Ähnliche Charakteristika wurden innerhalb der wesentlich höher polaren phenolischen Komponenten beobachtet. Dies impliziert, dass die Größe und Form der migrierenden Moleküle, determiniert durch die Anzahl aromatischer Ringe und durch aliphatische Kettenlängenverteilungen, aber hauptsächlich die Aktivität der funktionellen, polaren Atome Auswirkungen auf die chemische Zusammensetzung und demzufolge physikalischen Eigenschaften der Erdöle haben. Während der sekundären Migration aus dem Muttergestein durch wassergefüllte Porenräume verschiedener Transportgesteine in das Reservoir werden bevorzugt kleine, nicht zu aromatische und sehr polare Moleküle aus der migrierenden Phase entfernt. Dadurch verändern sich Dichte und Phasenverhalten der Öle, was sich signifikant auf deren Qualität im Reservoir und während der Förderung auswirken kann.

XIV

List of Publications

The scientific outreach of this project was attained by various publications in journals as well as contributions at national and international conferences.

Scientific journals

Ziegs, V., Horsfield, B., Skeie, J.E., Rinna, J. (2017) Petroleum retention in the Mandal Formation, Central Graben, Norway. Marine & Petroleum Geology 83, 195-214. DOI: 10.1016/j.marpetgeo.2017.03.005

Ziegs, V., Horsfield, B., Noah, M., Poetz, S., Hartwig, A., Rinna, J., Skeie, J.E. (2018) Unravelling maturity- and migration-related carbazole and phenol distributions in Central Graben crude oils. Marine & Petroleum Geology 94, 114-130. DOI: 10.1016/j.marpetgeo.2018.03.039.

Ziegs, V., Horsfield, B., Poetz, S., Hartwig, A., Rinna, J., Skeie, J.E., Deeper insights into oxygen-containing compounds of the Mandal Formation, Central Graben, Norway (pre-print). Submitted to Energy & Fuels on May, 9th 2018.

Conference contributions

Ziegs, V., Horsfield, B., di Primio, R., Rinna, J., Skeie, J.E. (2015) Petroleum expulsion from the Mandal Formation, Norwegian North Sea. 27th International Meeting on Organic Geochemistry - IMOG Sept, 13-18, Prague, Czech Republic. (Abstract + Poster)

Ziegs, V., Horsfield, B., Rinna, J., Skeie, J.E. (2016) Control of kerogen structure on petroleum expulsion from the Upper Jurassic Mandal Formation, Central Graben, North Sea. Tagungsband der DGMK/ÖGEW- Frühjahrstagung, 21./22. April 2016, Celle, Deutschland, p. 9. (extended Abstract + Poster)

Ziegs, V., Horsfield, B., Rinna, J., Hartwig, A., Skeie, J.E. (2017) Factors controlling the composition of nitrogen- and oxygen-containing compounds in Central Graben crude oils, 28th International Meeting on Organic Geochemistry – IMOG Sept, 17-22. Florence, Italy. (Abstract + Poster)

XVI

List of Figures

Fig. 1.1. Evolution of sedimentary organic matter fractions at depth (modified after Durand et al., 1980)...... 14

Fig. 2.1. Structural map of the investigation area. The general stratigraphy of the Central Graben area in the Norwegian sector is shown...... 35

Fig. 2.2. Analytical workflow and overview about conventional and novel parameters that will subsequently be used in the text. Blue boxes represent sample types; the bullet points indicate analyses and parameters conducted on and obtained from the distinct samples. . 39

Fig. 2.3. Bulk chemical characterisation of the Mandal Formation in the Central Graben using Rock-Eval pyrolysis obtained from a comprehensive geochemical data base of the Norwegian Continental Shelf including well info, stratigraphy and Rock-Eval parameter. Red circles indicate samples from the Søgne Basin, the blue line discriminates between kerogen types. The orange line is HI depletion trend of Northern areas, purple line

represents same for southern area. The attached HI histogram to the HI-Tmax plot represents the distribution of immature source rock strata of the Mandal Formation in different structural regions...... 43

Fig. 2.4. Bulk chemical characterisation of selected samples of the Mandal Formation using Rock-Eval pyrolysis of solvent-extracted samples representing the true nature of organic matter. Although samples represent a medium and lower quality interval of the Mandal Formation, the natural maturity series of the upper quality interval follows the kerogen degradation of a Type II kerogen (after Cornford et al., 1998)...... 46

Fig. 2.5. Bulk compositional characterisation of selected, extracted Mandal Formation samples obtained by open pyrolysis-GC-FID and using ternary plots of (a) the chain length distribution (Horsfield 1989, 1997) and (b) kerogen typing (Eglinton et al., 1990)...... 47

Fig. 2.6. Bulk kinetic parameter of the most immature Mandal Formation samples in a marginal (upper left) and central (upper right) basin position and comparison to literature data as indicated in Table 2.2 (a); and bulk kinetic parameter of 6 representative out of 24 samples from both kerogen types at different early mature levels (b)...... 49

Fig. 2.7. Transformation ratio curves of two kerogen types of the Mandal Formation calculated for an average geological heating rate for the Central Graben of 1 °C/Ma as means of comparison of kinetic stability with literature data...... 50

Fig. 2.8. Retention characteristics of Mandal Formation for bulk petroleum using Rock-Eval parameter of whole rock samples provided by the geochemical database. Coloured lines in depth plots represent maximum retained yields in marine source rocks of the Norwegian North Sea, excluding outliers in single horizons (<5 % of total number of samples, similar to as shown for Mandal Formation). Transparent circles represent generation and expulsion characteristics of Posidonia Shale (Germany) at different maturity stages. Dotted circles allow correlation of interesting samples between plots. In S1/TOC plot, Jarvie (2012)’s “oil cross over effect” with OSI values >10 % (>100 mg/g TOC) is indicated...... 53

XVII

Fig. 2.9. Bulk petroleum retention characteristics of selected samples of the Mandal Formation using Rock-Eval parameter of whole rock and extracted rock samples. For comparison, the retention series of the homogeneous and excellently expelling Posidonia Shale is plotted in shaded areas...... 54

Fig. 2.10. Petroleum Expulsion Efficiency of the Upper Jurassic Mandal Formation as obtained from mass balance modelling using a modified approach of Cooles et al. (1986). By means of comparison, an area representing the trend of excellent expellers (Upper Jurassic Draupne Formation in the Norwegian North Sea and the Northern German Posidonia Shale) is plotted in grey...... 56

Fig. 2.11. Characterisation of bitumen portions generated by the Mandal Formation. Upper

diagrams show absolute bitumen content denoted against extracted S2ker (a), and the

proportion of reactive organic matter as part of the measured, unextracted S2wr yield (b). Lower diagrams illustrate properties of generated petroleum: (c) Total Oil yields = volatile hydrocarbons (S1) + involatile, heavy bitumen, and (c) determination of retained petroleum quality (= S1/Total Oil). Green circles indicate kerogen type II source rocks, red circles

represent type II/III intervals. Gradients in (a) are colour-coded the same as maturity Tmax. The background data set comprises the Northern German Posidonia Shale, a typical marine source rock containing Type II organic matter, and unconventional U.S. oil and gas shale formations characterized as hybrid shale plays...... 58

S2bitumen yields. Saturation threshold until start of expulsion is significantly increased to 35 – 50 % conversion (PGI) for kerogen type II and II/III, respectively...... 59

Fig. 2.13. Comparison of retention characteristics of selected Mandal Formation samples using extracted samples (a), and correlation of aromatic compounds of reactive carbon (kerogen) with gas retention capacity (b)...... 61

Fig. 2.14. Cross plot of mono-, di- and tri-aromatic ring systems and gas sorption capacity of reactive carbon for the total sample set (a–c), and as a function of alkylation (d–f). Decreasing correlation with increasing aromaticity (ring number) of the compound class. Very good correlation for mono-aromatic ring systems with two fused side chains...... 63

Fig. 2.15. Cross plot of gas sorption capacity of reactive carbon versus aromatic compound group showing the best correlation. The correlation factor R² for both kerogen types (dark blue) is significantly higher than for individual kerogen types (grey)...... 64

Fig. 2.16. Retention of generated petroleum fractions in kerogen as a function of bitumen proportion of S2 indicating that bitumen is the major retention site of products from the Mandal Formation. Worldwide dataset (see Fig. 2.11) shows higher yields of retained products for low-S2 intervals of peak to late oil window which are thus dominated by the heavy bitumen portion...... 66

Fig. 2.17. Correlation of Total Oil retention on reactive organic matter with compositional parameter of the kerogen from extracted Mandal Formation samples showing no or weak correlation but a distinct influence of S2 yields (see Fig. 2.11). When excluding the two outliers a “pseudo-correlation” with the kerogen’s aliphaticity (R² = 0.51) can be observed...... 67

XVIII

Fig. 3.1. Investigation area showing the distribution of crude oil samples in the central and marginal area of the Central Graben rift system. Attached is a chronostratigraphic chart of the Mandal-Ekofisk petroleum system (after Cornford, 1994) illustrating oil-bearing reservoir intervals (oil well symbol after PPDM and FGDC). Circles represent oils from different lithologies and groups: green and blue: carbonate reservoirs in eastern and western basin; brownish-green and dark orange: clastic reservoirs of the norther margin (so-called “extended Ula Trend”) and from pre-Jurassic reservoirs, respectively...... 75

Fig. 3.2. API gravity variations with depth for different reservoir lithologies and stratigraphic intervals. Crude oils qualities of Cretaceous reservoir rocks are independent of reservoir depth, whereas Jurassic and older reservoirs in clastic strata become gradually lighter with increasing depth with similar gradient as in the southern Viking Graben (Justwan et al., 2006). Circles represent eastern (green) and western (blue) carbonate reservoirs investigated for their polar inventory...... 80

Fig. 3.3. Gross chemical composition of collected Central Graben crude oils showing common trend towards more saturate composition with increasing API gravity, independent of reservoir lithology. No regional features can be observed; the green circle identifies oils accumulated in the eastern carbonate reservoirs and the blue circles represent oils from the western carbonate reservoirs (Fig. 3.1)...... 82

Fig. 3.4. Biomarker cross plots indicating (a) the organic matter input and (b) the depositional environment with chemical variations in the water column. Colour coding of circles that represent particular basin areas applies as before...... 83

Fig. 3.5. Thermal maturity of the analysed oil and condensate samples based on (a) sterane

isomerization, and (b) C27 vs. C29 Ts/(Ts+Tm). The inset of (b) shows carbonate oils analysed for their polar, high molecular weight inventory using ESI negative FT-ICR-MS. All of the maturity biomarkers are specific for a distinct maturity range but provide similar results. The vitrinite reflectances for chemical equilibrium of sterane isomerizations are estimated after Waples and Machihara (1990) and Peters et al. (2005)...... 84

Fig. 3.6. Cross plot of benzene and toluene ratios against their water-insoluble structural equivalents. Benzene is more hydrophilic than is toluene...... 86

Fig. 3.7. Rudimentary oil family discrimination using elemental data from FT-ICR-MS. Interferences with maturity can be expected. Combined with maturity assessment, compositional variations within the same maturity stage can be used for classification. Circles represent regions as in previous illustrations...... 88

Fig. 3.8. DBE class distribution of the N1 compound class. Dotted and dashed lines indicate oils from eastern and western carbonate reservoirs, respectively...... 90

Fig. 3.9. Maturity assessment based on the distribution of carbazole and its higher fused homologues (Oldenburg et al., 2014). Maturities of the polar and hydrocarbon components

of petroleum fit quite well, but show a higher contribution of N1 DBE 12 benzocarbazoles for carbonate oils and thus an apparent maturity retardation of those oils...... 91

Fig. 3.10. Correlation of maturity indicator 29Ts/(29Ts+NH) to the total monoisotopic

abundances of N1 DBE 9, 12 and 15 compounds illustrating that individual abundances (TMIA) react in the same manner as their relative TMIA’s in a ternary plot. R² has been calculated, excluding 3/7-4 because of too strong facies variations...... 92

XIX

Fig. 3.11. Chain length distribution of aliphatic carbon attached to benzocarbazole core

structures (N1 DBE 12). Earliest mature oils contain longest side chains (a). Peak oil window mature oils stored in Upper Cretaceous carbonate reservoirs (dashed) contain aliphatic chains with higher contribution of intermediate carbon numbers than clastic oils,

depending on regional occurrence (b). Crude oils at the eastern rim contain N1 components with significantly longer chained substituents than at the western margin (compare Fig. 3.1)...... 94

Fig. 3.12. Maturity correlation of relative abundances of individual N1 DBE 9, 12, 15 and 18 compound classes representing carbazoles and its higher-fused homologues. Correlation (a) shows oils produced from clastic intervals interbedded with Upper Jurassic OM-rich, marine black shales, and (b) illustrates oils from carbonate reservoirs that lie stratigraphically higher than the principal source rocks. 29Ts/(29Ts+NH) is used as maturity indicator. Colours indicate the strength of the correlation as used in natural sciences. Dotted lines represent the average relative abundances (% TMIA) of the particular compounds in the data set to illustrate the significance of correlations...... 96

Fig. 3.13. DBE class distribution of the O1 compound class. Dashed lines indicate oils from

carbonate reservoirs. Inset shows oils with lower O1 contribution (cf. Fig. 3.7)...... 100

Fig. 3.14. Carbon number distribution of O1 DBE 4 (a) and DBE 5 (b) compounds of marine oils. Oils in carbonate reservoirs are represented by dashed lines with 1/9 oils belong to the western basin centre and 2/5 and 2/6 oils are structurally related to the Steinbit Terrace and

Søgne Basin. The insets show samples with increased O1 contribution due to variations in depositional environment...... 101

Fig. 3.15. Ternary plot of three DBE classes correlating linearly with 29Ts/(29Ts+NH) maturity indicator. Using different compound classes excludes the influence of rock-fluid interactions...... 102

Fig. 4.1. Structural map and the general post- and syn-rift stratigraphy of the Norwegian Central Graben showing the well locations of selected source rock samples as well as oil and gas fields and discoveries. Post-print from Ziegs et al. (2017)...... 112

Fig. 4.2. Chemical gross composition of the maltene fraction of extracts from the Mandal Formation (squares for Type II and triangles for Type II/III) and Posidonia Shale (circles)

illustrated as a function of maturity, Rock-Eval Tmax and %RO...... 121

Fig. 4.3. Distribution of major compound groups of the Mandal Formation (Type II and II/III) and Posidonia Shale (Type II-A)...... 121

Fig. 4.4. Major compound classes of Mandal Formation extracts ([a] Type II intervals, [b] Type II/III intervals) and (c) Posidonia Shale extracts evolving as a function of maturity. Dominance of oxygen-containing heterocompounds in the Mandal Formation and nitrogen- containing constituents in the Posidonia Shale are clearly illustrated...... 123

Fig. 4.5. Relative distributions of O1 to O5 compounds within the Ox class of extracts from the

Mandal Formation and Posidonia Shale that illustrate the dominance of O3+ compounds in the Mandal Formation...... 123

Fig. 4.6. DBE class distribution of the O2 compound class comparing source rock extracts of the Mandal Formation and Posidonia Shale at three maturity stages...... 125

Fig. 4.7. Carbon number distributions of selected O2 DBE classes addressing detailed compositional differences of the Mandal Formation and Posidonia Shale extracts at three maturity stages...... 126

Fig. 4.8. DBE class distribution of the O3 compound class comparing source rock extracts of the Mandal Formation and Posidonia Shale at three maturity stages...... 128

Fig. 4.9. Carbon number distributions of selected O3 DBE classes addressing detailed compositional differences of the Mandal Formation and Posidonia Shale extracts at three maturity stages...... 129

Fig. 4.10. DBE class distribution of the O4 compound class comparing source rock extracts of the Mandal Formation and Posidonia Shale at three maturity stages...... 130

Fig. 4.11. Carbon number distributions of selected O4 DBE classes addressing detailed compositional differences of the Mandal Formation and Posidonia Shale extracts at three maturity stages...... 131

Fig. 4.12. Oxygen contents of the macromolecular organic matter obtained using Rock-Eval pyrolysis and FT-ICR-MS correlate well with each other suggest genetic relationship between kerogen and extracts of the source rock...... 133

Fig. 4.13. Carbon number distributions of the O2 DBE 7 class of immature New Zealand coal

extracts are clearly dominated by the C20 homologue representing a diterpenoid structure typical for resinous material...... 133

XXI

XXII

List of Tables

Table 1.1. Genetic stages of petroleum formation for a Type II kerogen which is an intermediate between Type I and III OM (after Espitalié, 1986)...... 15

Table 2.1. Levels of thermal maturity in the Central Graben (Spencer et al., 1986b; Cornford, 1994)...... 41

Table 2.2. Bulk kinetic parameters of most immature Mandal sections (marginal and central basin position) and comparison to literature data from Viking Graben and U.K. Central Graben...... 51

Table 3.1. Specifications of analysed oil and condensate reservoir samples originating from the Central Graben and ordered according to regional occurrence and structural affiliation within the basin...... 81

Table 3.2. Relative abundances of compound classes obtained from ESI (-) FT-ICR-MS for 16 oil samples...... 87

Table 4.1. Selected sample set of the Mandal Formation in the Central Graben containing information on sampled wells, bulk source rock characteristics (Rock-Eval) and retained petroleum...... 118

Table 4.2. GC-FID and GC-MS data of Mandal Formation extracts characterizing the organic matter input and depositional setting...... 119

Table 4.3. Relative abundances of major NSO compound classes all samples investigated by ESI (-) FT-ICR-MS...... 122

XXIII

XXIV

List of Abbreviations

A Frequency factor

API Measure of oil density = (141.5 / ρrel) – 131.5 ASA Norwegian corporate form; stock company BCU Base-Cretaceous unconformity BHT Bottom hole temperature BI Bitumen Index = S1∙100/TOC CN / C# Carbon number Da Dalton, is equal to "unified atomic mass unit (u)" DBE Double bond equivalent DST Drill stem test

EA Activation energy Eh Redox potential EI Electron ionization (mode)

Emean Mean activation energy EOM Extractable organic matter upon solvent extraction ESI Electro-spray ionization mode ex Extracted rock Fm Formation FT-ICR-MS Fourier transform-ion cyclotron resonance-mass spectrometry GC-FID Gas-chromatography-flame ionization detector GC-MS Gas chromatography-mass spectrometry GIP Gas-in-place GOR Gas-to-oil ratio GR Gamma ray H# Hydrogen number HC Hydrocarbon HI Hydrogen Index = S2∙100/TOC HMW High-molecular weight KCF Kimmeridge Clay Formation ker Kerogen LMW Low-molecular weight m- meta-position of substituents attached to aromatic rings Ma Million years MeOH Methanol

Mn Number-average molecular weight MPLC Medium-pressure liquid chromatography MRM Multi reaction monitoring mode of GC-MS MSSV Micro-Scaled Sealed Vessel pyrolysis method (Horsfield et al., 1989)

Mw Weight-average molecular weight NE Northeast NIGOGA Norwegian Industry Guide to Organic Geochemical Analyses NPD Norwegian Petroleum Directorate

XXV

NSO Heterocompounds containing nitrogen, sulphur and/or oxygen NW Northwest o- ortho-position of substituents attached to aromatic rings O# Number of oxygen atoms, Oxygen number OI Oxygen Index = S3∙100/TOC OIP Oil-in-place OM Organic matter OSI Oil saturation index = S1∙100/TOC p- para-position of substituents attached to aromatic rings PCA Principal component analysis PEE Petroleum expulsion efficiency PGI Petroleum generation index + pH Scale for acidity or basicity of an aqueous solution = -log10 [a(H30 )] PI Production Index = S1/(S1+S2) PNA / P-N-A Paraffinic-Naphthenic-Aliphatic (petroleum generating facies) Py-GC Pyrolysis at 300 – 600 °C upon GC-measurement R Correlation coefficient: measure of the strength and direction of the linear relationship R² Coefficient of determination: proportion of the variance in the dependent variable that is predictable from the independent variable(s)

Rc Calculated vitrinite reflectance res Resolved compounds separated by the GC-column

RO Vitrinite reflectance S1 Volatizable organic matter upon Rock-Eval pyrolysis at 300 °C S2 Volatizable organic matter upon Rock-Eval pyrolysis at 300 – 850 °C SE Southeast SIR Single ion recording mode of GC-MS Sm³ Standard cubic metres SW South-west

Tmax Temperature of maximum S2 upon Rock-Eval pyrolysis in °C TMIA Total monoisotopic abundance TOC Total organic carbon Tvap-GC Thermovaporization at 300 °C prior to pyrolysis upon GC measurement U.K. United Kingdom U.S. United States (of America) VR Vitrinite reflectance wr Whole rock, unextracted rock WSW West-southwest

XXVI

1: INTRODUCTION

1 Introduction

1.1 The history of organic geochemistry in petroleum system evaluation

Organic geochemistry is, in the broad sense, the science of organic substances and the exchange of carbon within the biosphere, pedosphere, geosphere, hydrosphere, and atmosphere of the Earth. The global carbon cycle addresses biochemical and geochemical processes of organic carbon interacting with organic and inorganic solids, fluids and gases, e.g. soils, surface and subsurface waters or

CO2. While the biochemical cycle involves living organisms and its oxidized remains, the geochemical cycle concerns dead organic matter deposited and preserved in recent and ancient sediments or coals. Petroleum geochemistry as particular field of this branch of science seeks to understand the composition and origin of sedimentary organic matter, mechanisms of formation, modes of deposition and distribution within the subsurface, as well as their interaction with incorporated solid, fluid and gaseous phases.

The origin of petroleum was for many years a matter of controversial debate. Early scientific theories were based on assumptions that organic carbides as iron or aluminium carbide are abundant in the earth’s crust forming petroleum during interaction with water (Berthelot, 1866) and that petroleum is formed in depths where no living organisms occur (Mendeleiev, 1878). Other concepts concerning the abiogenic origin of petroleum became popular reaching their peak between the 1950’s and 1980’s (Glasby, 2006) when major oil and gas discoveries in crystalline basement rocks had to be explained in the Soviet Union. Two principal theories on the abiotic origin of petroleum were developed based on the assumption that methane is abundant within the Earth’s mantle as primordial material. The Russian-Ukrainian theory of deep, abiogenic origin (Kudryavtsev, 1951) believed methane to be erupted to the Earth’s surface and converted to higher hydrocarbons at higher temperatures and pressures, similar to upper mantle conditions. Based on the Russian theory, Gold (1985) assumed a Fischer-Tropsch-like reaction, forming long-chained hydrocarbons from inorganic reactants in the presence of hydrogen gas. However, thermodynamic conditions in the upper crust are not favourable for such reactions (Kenney et al., 2002) and both theories were lacking

1: INTRODUCTION in explaining the complexity of petroleum composition which has been determined using more sophisticated analytical methods, such as GC-MS, developed in the 1970’s.

However, even Louis Pasteur (1822 – 1895) is said to have supported the biogenic or organic theory (Durand, 2003) stating that petroleum was formed from bacterial action on organic debris originating from animals or land plants (e.g. Hunt, 1863). The biotic origin gained major recognition especially after Treibs (1934) discovered porphyrins in crude oil, and noted them to be structurally similar to chlorophyll. Although suggested in 1924 by the Russian geologist Vernadsky, the thermal origin of organic matter present in sedimentary rocks and their incorporation into the geochemical cycle of elements between living biosphere and ancient geosphere, the concept of “source rock” was introduced by American petroleum geologists (Trask and Wu, 1930; Trask et al., 1932; Trask and Patnode, 1942). The precise formation mechanism of petroleum was under debate until the 1960’s, but some alternative theories could be successfully applied to unconventional petroleum plays, i.e. the hydrocarbon formation from bacteria (Pasteur-like, ZoBell, 1947; Stone and ZoBell, 1952) contributing to some natural gas occurrences (e.g. Antrim Shale, U.S.; Puchkirchen Fm, Molasse basin, Germany; Po plain, Italy; Venezuelan offshore fans; pers. comm.: Hans-Martin Schulz, 2017, Curtis (2002)); or the radioactive decay altering organic matter and causing hydrocarbon formation (Lind, 1938; Yang, 2017) that may have contributed significantly in e.g. the Alum Shale but cannot seen as a sole mechanism. Other theories like the direct accumulation from recent organic-rich sediments by compaction (Baker, 1959) can be regarded as precursor/parent concept to modern expulsion and migration theories of two-phase flow of water and oil.

Accompanying the rapid elaboration of new analytical instruments since 1960, enormous progress has been made in the understanding of the complex nature of organic matter. The development of gas chromatography and mass spectrometry coupled together (GC-MS) allowed separation and identification of molecules giving the undeniable evidence of petroleum originating from biological precursors, e.g. long chained n-alkanes from cuticular waxes, isoprenoids from side chains of chlorophyll, fatty acids from marine and lacustrine plankton as well as hopanes and steranes from cell membranes of prokaryotic and eukaryotic

1: INTRODUCTION microorganisms, respectively (e.g. Peters et al., 2005; Vandenbroucke and Largeau, 2007).

Following on Verdnasky’s concept of thermal origin of petroleum from organic matter, pyrolysis gas-chromatography coupled with a flame ionization detector (GC-FID) having had intended to artificially generate petroleum from immature organic-rich sediments (based on Engler, 1888) illustrated the link of laboratory pyrolysis and natural formation in sedimentary basins. Varying composition from products in the natural system (more resins/asphaltenes, formation of unsaturated hydrocarbons, hydrogen deficiency) stopped attempts to reproduce natural petroleum formation (e.g. Lewan et al., 1979). Alternatively, the characterisation of the organic matter structure (Larter, 1984; Horsfield, 1989; Eglinton et al., 1990; Horsfield, 1997) and its maturation level using the Rock-Eval© instrument (Durand and Espitalié, 1975; Espitalié et al., 1977) has emerged, combined in maturity-dependent structural descriptions using flash pyrolysis- or MSSV- pyrolysis-GC-FID (e.g. Eglinton et al., 1988; Horsfield et al., 1989; Horsfield and Düppenbecker, 1991). Using pyrolysis methods, the artificial maturation of organic matter is mathematically described using kinetic parameters A and EA of its bulk inventory (Burnham et al., 1987; Quigley et al., 1987; Ungerer and Pelet, 1987; Ungerer et al., 1988) or using defined pseudo-fractions (Dieckmann et al., 2000; di Primio and Skeie, 2004; Horsfield et al., 2011; Ziegs, 2013) to serve as calibration parameter in 3D basin modelling.

Some of the present and future challenges of petroleum geochemistry address heavy oils and solid- or semisolid organic phases in reservoirs (Huc, 2003). Such occurrences may be a result of (1) primary factors as source rock facies generating and expelling long-chained hydrocarbons and high-molecular weight compounds, e.g. Type I and III organofacies, or secondary effects as (2) phase behaviour or (3) in-reservoir petroleum cracking to lighter hydrocarbons and a carbon-rich residue, the pyrobitumen, forming a major risk during production and long-term field development. Being a matter of petroleum generating facies (source rock), such risks can be identified incorporating the polar and/or high-molecular weight fraction into bulk or compositional predictions on a basin scale. Improving the understanding of the role of the heavy fraction in expulsion- and migration-related processes, thresholds in different migration steps, quantitative losses and

1: INTRODUCTION compositional modifications along the migration pathways (kerogen-mineral-fluid interactions) may be predicted more precisely.

1.2 Origin of organic matter

In the assessment of crude oil accumulations, initial factors determining oil quality are (1) the source rock depositional environment and facies variations as well as (2) the maturation stage of the source rock package at the time of expulsion. Both processes directly affect the kerogen composition that determines the oil quality. Secondary factors having an impact on the crude oil quality are a function of the initial petroleum composition: (3) chemical fractionation during expulsion and migration, (4) phase separation after expulsion (Silverman, 1965; England et al., 1987; England and Mackenzie, 1989; Larter and Mills, 1991), (5) biodegradation and water washing (e.g. Connan, 1984), (6) in-reservoir maturation of accumulated petroleum (Gabrielsen et al., 1985; Horsfield et al., 1992), (7) contamination during migration through other organic matter-rich rocks (Curiale and Bromley, 1996), and (8) mixing of petroleum charges of different source and maturity.

1.2.1 Characterization of organic matter

Organic matter (OM) in petroleum source rocks encompasses kerogen, a geopolymer formed during diagenesis of disseminated OM which is insoluble in non-oxidizing acids, bases and organic solvents (Forsman and Hunt, 1958), as well as petroleum and bitumen, the soluble fraction representing the kerogen’s generation products upon increased thermal stress. Throughout the historical evolution of organic geochemistry, different physical, chemical and pyrolytical methods have been applied for characterizing kerogen whose structural features determine quantity and quality of petroleum as well as its timing of generation and expulsion from the source rock.

1.2.1.1 Optical analysis

The optical evaluation of kerogen is closely linked with coal petrography. Preparing thin slices and polished thick sections made from organic-rich sediments or concentrated OM and observed under transmitted or reflected light, microscopic

1: INTRODUCTION methods reveal information on debris from organisms and microfossils (palynology) or allows classification based on so-called macerals (Stopes, 1935).

Macerals are optically homogenous aggregates of OM with distinct physical and chemical properties (Spackman, 1958) that are found concentrated in coals, and in a finely dispersed state in other sedimentary rocks. Maceral analysis is conducted under reflected light using 546 nm wavelength, normal incidence to a perfectly polished section and a lens with immersion in oil of refraction index 1.517. According to their similar colour, reflectance and fluorescence, macerals can be grouped into three families (Stach et al., 1982). With increasing order of reflectivity at a given maturity or rank, the following groups can be distinguished:

• Liptinite maceral group with low reflectance but fluorescent under UV light. Liptinite macerals, or as they were originally termed, exinites, originate from the “outer parts” of higher plants that are lipid-rich, e.g. spore cases, pollens, leaf cuticles (sporinite, cutinite). Alginites are formed from unicellular algae, phytoplankton (telalignite and lamalginite). Miscellaneous amorphous substances that are derived from intra- or extra-cellular resins (resinite), organic aggregates or bacteria can be very abundant.

• Vitrinite maceral group with medium reflectance. Being derived from lignocellulose debris (wood, cortex) of higher plants (telinite) or from their decomposition products (collinite which is amorphous), these macerals may be textured, amorphous or occur as gels. The reflectance of low-gray or

primary vitrinite macerals (VR) is commonly used as a maturity indicator of organic matter.

• Inertinite maceral group with high reflectance. Derived from fungal or plant debris (sclerotinite and fusinites), the latter macerals may be oxidized before or during sedimentation, e.g. due to varying redox conditions or forest fires, or represent reworked or bacterially altered material.

Observations under transmitted light allow distinction of the paleo-microflora and –fauna, inferences of their depositional environment and life conditions, and evaluation of bio- and chronostratigraphic relationships. Identification of the palynofacies is conducted on amorphous OM and recognizable fragments of identifiable biological origin (Raynaud and Robert, 1976) after dissolving inorganic

1: INTRODUCTION constituents using hydrochloric and hydrofluoric acid. Fragments of vegetable tissue, lignitic fibers, epidermises, cuticles, resins etc. but especially microfossils like pollens, spores and algae (Dinoflagellate, Tasmanaceae, Botryococcacaeae) and their subgroups have occurred during particular time intervals and allow a small- scaled analysis of OM-rich intervals and spatial correlation of basin-scaled events (e.g. Dybkjaer, 1998 for the principal source rock in the Central Graben, North Sea).

The optical appearance of spores and pollens under visible transmitted light and vitrinite macerals under UV reflected light is strongly dependent on thermal stress having experienced since their deposition and subsequent burial. A discussion on these maturity parameters will be presented in chapter 1.3.2.

1.2.1.2 Elemental analysis and Rock-Eval©

Following on from Hunt’s and Verdnasky’s early theories, a clear definition of kerogen composition has ensued, thanks to the development of a variety of analytical techniques. Kerogen is mainly composed of C, H, O, N and S (Durand and Monin, 1980). Based on elemental composition attained by combustion of different aliquots for each element, four different kerogen types can be classified using their H/C and O/C ratios plotted in the van-Krevelen diagram (Van Krevelen, 1950; Durand and Espitalié, 1973). Parameters obtained from Rock-Eval pyrolysis (Espitalié et al., 1977; Behar et al., 2001) can be conceptually incorporated into a plot of this type as well. Rock-Eval pyrolysis is a cheap and rapid geochemical method heating whole rock samples or kerogen extracts in a standardized temperature program and quantifying the peaks using a flame ionization detector (FID). The first peak (S1) indicates thermovaporized free compounds at 300 °C and the second peak (S2) shows pyrolyzed products until 800 °C (Vandenbroucke and Largeau, 2007) or 850 °C (Lafargue et al., 1998) that represent the remaining

generative potential. CO2 measured until 390 °C and up to 850 °C represent S3 and S4 peaks, respectively. Total organic carbon content (TOC) is the sum of all peaks and is in accordance with TOC from LECO SC-444 (Lafargue et al., 1998; Behar et al., 2001). Important parameters obtained for source rock evaluation are the

Hydrogen Index, HI = S2/TOC, Oxygen Index, OI = S3/TOC, and Tmax which is the

1: INTRODUCTION temperature at maximum S2 yields. HI and OI correlate very well to H/C and O/C ratios, respectively (Espitalié et al., 1977; Espitalie et al., 1985).

Biota from different sources and depositional environments distinguish their initial elemental ratios and consequently the petroleum generation characteristics,

HI and Tmax. Organic matter deposited in anoxic lakes or marine basins is mainly composed of algae, phytoplankton and bacteria. These extremely hydrogen-rich materials are oil-prone and constitute kerogen type I or II. Type III kerogen is derived from terrigenous OM and land plants.

Type I OM which is characterized by H/C > 1.5 and HI > 700 mg HC/g TOC originates largely from lacustrine algae being transformed to alginite or saporpelinite macerals. Examples are the Green River and Wealden Shales, or torbanite (boghead coal).

Type II OM is mainly deposited in marine environments containing marine planktonic material, spores, pollen, cuticles or higher plant secretions as precursor to liptinitic or exinitic macerals. HI ranges between 200 and 650 mg HC/g TOC (H/C = 0.8 – 1.5) with pure marine planktonic subtype II-A constituting HI’s above ~400 mg HC/g TOC generating mainly oil and minor gas, and mixed marine- terrestrial subtype II-B below that boundary, generating oil and gas in different quantities (after Delvaux et al., 1990). Source rocks containing marine Type II OM have charged most of the global petroleum reservoirs due to the sheer abundance of this depositional environment. Prolific examples are the Upper Jurassic Kimmeridge Clay Formation in the North Sea, the Toarcian Posidonia Shale in Western Europe (Germany, France), Permian Irati oil shale (Brazil) but as well shale gas plays in the U.S., e.g. Barnett, Eagle Ford, Bakken and Marcellus shales at high maturity levels.

Type III OM is mainly gas-prone but may generate wax-rich crude oils (Horsfield, 1984; Hunt, 1991). Composed of vitrinitic macerals, these diagenetic products of higher land plants are poor in hydrogen (H/C = 0.5 – 0.8 and HI < 150 mg HC/g TOC) and may be enriched in oxygen. Constituting shales and coals deposited in fluvio-deltaic environments, marshs and swamps, conventional gas/condensate and waxy oil fields can be found in the Mahakam delta, Indonesia, Nile delta (Hunt, 1991).

1: INTRODUCTION

Type IV OM (Harwood, 1977) is dominated by inertinite consisting of oxygenated, reworked or thermally residual OM and has no petroleum generating potential with HI < 50 mg HC/g TOC.

Mixing of different kerogen types contributing to the same organic matter sink may be identified by incorporation of different approaches to kerogen assessment and linking with sedimentary facies analysis.

1.2.1.3 Pyrolysis-gas chromatography

The structural inventory of kerogen can be tracked down to the molecular level using pyrolysis gas-chromatography (Py-GC-FID). A programmable pyrolysis unit, chemically degrading organic matter by thermal energy, coupled with a GC-FID allows classification of both, free hydrocarbons (similar to Rock-Eval S1 yield) and of kerogen (similar to the S2 peak) based on the composition of their pyrolysis products. Having shown that pyrolysis products are structurally similar to the parent kerogen structure and yields are quantitatively comparable to the amount of petroleum generated under natural conditions (Horsfield, 1997), compositional ‘fingerprinting’ provides not only insights to the kerogen structure, its type and maturity (Horsfield et al., 1983; Larter, 1984; Horsfield, 1989; Eglinton et al., 1990) but further refines the Rock-Eval approach relying on bulk volumes only. On the premise that kerogen composition directly controls the types and yields of volatile products generated during natural maturation, it further offers exploration- and production-related information like gas-oil-ratio, aromaticity, sulphur content and wax content of potentially generated natural oils.

Upon pyrolysis, organic matter yields many compound types including hydrocarbons, ketones, alcohols, nitrils and thiols that are incorporated into saturated and unsaturated, cyclic and acyclic carbon structures (Dembicki Jr. et al., 1983). The most commonly occurring identifiable compounds are aliphatic hydrocarbons reflected by doublets of n-alkanes and n-alk-1-enes, aromatic hydrocarbons such as alkylbenzenes and alkylnaphthalenes, and other aromatic compounds such as alkylphenols as well as alkylthiophenes (Horsfield, 1997). Comparison of pyrolysate compounds and structural parameters with other geochemical methods provides evidence that structural information from kerogen is well represented by readily identifiable, GC-amenable pyrolysis products.

1: INTRODUCTION

Aromaticity determined using relative proportions of major aromatic and aliphatic compounds coincides with aromaticity determined from 13C NMR spectrometry (Horsfield, 1989). The hydroxyl-oxygen content determined by wet chemical methods (Senftle et al., 1986) has been shown to be directly proportional to the alkylphenol content of Carboniferous coal pyrolysates in the United States (Larter and Horsfield, 1993) indicating that phenolic compounds in pyrolysates might be representable of oxygen-substituted benzoic structures in kerogens. Eglinton et al. (1990) have proven that pyrolytic sulphur species are representative of kerogen- bond sulphur molecules by comparing relative abundances of alkylthiophenes versus aromatic plus aliphatic hydrocarbons to the atomic S/C ratio.

Fingerprinting using open-pyrolysis products is now a widely accepted approach for kerogen typing and characterization of depositional environments by relating compositional parameters as paraffinicity and aromaticity to phenol (Larter, 1984) or sulphur contents (Eglinton et al., 1990; di Primio and Horsfield, 1996). Pyrolysates of organic matter types deposited in aquatic systems and enriched in hydrogen to different degrees (H/C > 1.5 for Type I and H/C = 0.8 – 1.5 for Type II kerogens) are dominated by intermediate to long-chained, saturated hydrocarbons composed of CH2 homologues. Type III OM contain higher contributions of aromatic compounds (combinations of C6H6 base structures) linked with variably long aliphatic side chains and oxygen compounds (elevated O/C). Phenol is a major indicator of these initial features in thermal degradation products (Larter, 1984). Mixed contributions that may be categorized as Type II or II/III OM contain features of all end member kerogens.

Horsfield (1989, 1997) has outlined that not only compound types, but rather the chain length distribution of n-alkane/n-alkene doublets in pyrolysates representing the linking structures in kerogen (Tissot and Welte, 1984) can be applied for prediction of petroleum types. Five major Petroleum Type Organofacies have been identified and can be related to marine, lacustrine, deltaic and terrestrial depositional environments of kerogen: Gas and Condensate, Paraffinic- Naphthenic-Aromatic (PNA) oil with high/low wax content, and Paraffinic Oil with high/low wax content. n-Alkyl pyrolysate compositions can be directly related to land plants (odd carbon numbers [CN] in C22+ range of Type II kerogens) or alginites (even CN’s in C9-18).

1: INTRODUCTION

1.2.1.4 Kinetic characterization

Already the fathers of organic geochemistry (e.g. Vernadsky, 1934) have noticed that petroleum formation is dependent on burial depth and thus must obey rules of chemical kinetics (Tissot and Espitalie, 1975). Kinetic calculations describe the thermal degradation of organic matter as a variable of heating time, temperature and/or pressure (Hunt, 1979; Tissot and Welte, 1984; Vandenbroucke et al., 1999). Kinetic parameters, consisting of a series of activation energies (EA) and a single frequency factor, are derived using a parallel reaction model that follows a quasi-first-order reaction scheme and the Arrhenius law (Jüntgen and van Heek, 1968; Burnham et al., 1987; Braun et al., 1991; Schenk et al., 1997). The frequency factor is related to the vibration energy of the reaction and correlates to the activation energies (Stainforth, 2009).

Kerogen degradation reactions can be characterized as bulk kinetics using the bulk kerogen as a whole describing primary formation of petroleum (Quigley et al., 1987; Braun et al., 1991). Increasing thermal energy causes the petroleum to crack down to gaseous hydrocarbons which can be described by secondary kinetics in the high-maturity source rock (Ungerer et al., 1988; Pepper and Dodd, 1995; Dieckmann et al., 1998) or in deep reservoirs with complex geological history (Horsfield et al., 1992; Schenk et al., 1997). Compositional kinetics of primary or secondary cracking processes (Düppenbecker and Horsfield, 1990; Behar et al., 1997; Dieckmann et al., 2000; di Primio and Skeie, 2004; di Primio and Horsfield, 2006; Ziegs, 2013) assembles (pseudo)-fractions with kinetic information and provides more detailed insights into petroleum type prediction.

Compositional features of kerogen are as well represented by their thermal degradation during petroleum generation. The chemical variance of the kerogen inventory is characteristic for particular types (Tegelaar and Noble, 1994) and is determined by the stability of their chemical bonds. In general, sulphur incorporated into the kerogen structure reduces its thermal stability (Orr, 1986), and thus dissociation energies are ordered in the following manner: C–S < C–O < C–H < C–C. Stability is increased with unsaturation of the covalent bond (C=C or

C≡C), while aromatic rings with 3 π-double bonds are most stable. The EA distribution of a Type I kerogen is typically extremely narrow and mostly

dominated by one principal EA which is a result of its extremely homogenous

1: INTRODUCTION

composition. Differently, Type III OM possesses a broad EA distribution (Ungerer and Pelet, 1987) representing the variety of covalent bonds with numerous heteroatoms incorporated as well as aromatic structures linked by aliphatic chains. Marine type II kerogen shows an intermediate shape which is normally Gaussian- distributed but may vary in dependence of its precursor structures and sedimentary facies.

1.2.2 Natural origin of NSO compounds

Since the discovery of porphyrins in petroleum proving their biological origin (Treibs, 1934) heteroelements have been attributed as geochemical markers but their understanding is still incomplete (Tissot and Welte, 1984). Porphyrins are important components in the plant and animal kingdom as constituents of pigments like chlorophyll, the photoactive green plant pigment, and hemin, the red blood pigment and contain N in its chelate structure and O as linkage to the isoprenoid side chains (Tissot and Welte, 1984). In general, living organic matter comprises abundant macromolecular compounds that contain N, S and O heteroatoms bound in or at aromatic ring systems or aliphatic chains, e.g. as part of their storage materials (polysaccharids: starch, glycogen, triglycerides), or as structural components (cellulose, alginic acids, chitins, proteins, glycolipids, polyisoprenoids, polyterpenoids, cutins, lignins, sporopollenins, algeanans, etc.) (de Leeuw and Largeau, 1993). Heteroatomic compounds may occur as “simple molecules”, such as (fatty) acids, esters, waxes and porphyrins which are formed and restructured during early diagenesis, isolated using chemical procedures and identified using open pyrolysis-GC-FID, GC-MS and spectroscopic analyses (Vandenbroucke, 1980). On the other hand, heavy ends like asphaltenes and resins contain abundant NSO compounds and are analysed using global methods similar to kerogen identification (elemental analysis, spectroscopic ones: X-ray, UV, IR, NMR; Yen, 1974) but i.e. FT-ICR-MS for their most polar moieties (Hughey et al., 2002; Marshall and Hendrickson, 2002; Marshall and Rodgers, 2004). However, the majority of the structural components possess a high potential to be selectively preserved during the early diagenetic transformation of deposited organic matter into geobiopolymers (Tegelaar et al., 1989a) and are better known as macerals or their subgroups making up the kerogen structure. While kerogen type I is mainly composed of hydrogen-rich, aliphatic material stemming from e.g. cell walls of algae

1: INTRODUCTION or bacteria, terrigenous Type III and aquatic Type II organic matter comprises variably abundant polyaromatic structural components cross-linked by heteroatoms (Durand-Souron, 1980).

Oxygen is by far the most abundant and diverse functional or linking atom in such compounds (Behar and Vandenbroucke, 1987) It occurs mainly in functional groups, either terminal (carboxyl and hydroxyl groups) or in cross-linking position (ester and methoxy groups) and may be present in heterocycles (Vitorovic, 1980; Pelet et al., 1986). High oxygen contents can be found in vitrinitic macerals or coals, but may be a function of individual preservation efficiencies. While carbonylic and carboxylic acids or alcoholic and phenolic structures constitute the majority of the total oxygen content and increase from Type I < II < III, no esters are present in Type III OM (Rouxhet et al., 1980; Vandenbroucke, 1980) as these may possibly originate from marine OM (Vitorovic, 1980). However, the loss of oxygenated functional groups in humic acids during early diagenesis/selective preservation pathways (Tissot and Welte, 1984) promotes their insolubility in water and incorporation into the kerogen structure (Vandenbroucke and Largeau, 2007).

Nitrogen amounts for 2 – 3 wt.-% in sedimentary organic matter (Tissot and Welte, 1984). Nitrogenous compounds, for example amino acids from protein biodegradation, can easily be incorporated during oxidative polymerisation of phenols (Vandenbroucke and Largeau, 2007) and are the most abundant cellular constituent in humic substances. They become rapidly degraded during early diagenesis by exoenzymes that cleave peptide bonds. The released amino acids are then used for microbial metabolism (Huc and Durand, 1973; Vandenbroucke and Largeau, 2007). It was considered for some time that nitrogen mostly occurs as heterocyclic aromatic structures in source rocks and crude oils (Tissot and Welte, 1984), i.e. pyridines, quinolones and carbazoles which may be skeletal fragments of plant alkaloids (Whitehead, 1973). The [b]-configuration of benzocarbazole is seen indicative of terrigenous organic matter contributions (Oldenburg et al., 1999; Steen et al., 2000; Bakr and Wilkes, 2002), while the ratio of the [a/(a+c)] configurations and their alkylated homologues might be sensitive to facies changes of marine depositional environments (Clegg et al., 1997; Bakr and Wilkes, 2002). However, the influence of maturity and migration fractionation can dominate these distributions in other petroleum systems and will be discussed in detail in chapter 3.

1: INTRODUCTION

Sulphur is largely introduced in-situ during early diagenesis by “natural vulcanization” rather than through the selective preservation of parent sulphur- containing organic material (e.g. Sinninghe Damsté et al., 1989; de Leeuw and Largeau, 1993). Multiple mechanisms have been postulated which are indicative for carbonate and hypersaline environments (Hughes, 1984), but may occur during other conditions where hydrogen sulphide (or possibly polysulphides) is present (Sinninghe Damsté et al., 1989). Incorporation of inorganic sulphur via functionalized lipids into the macromolecular matrix and/or intermolecular cross- linking of functionalized lipids via sulphide and disulphide bonds (Tissot and Welte, 1984; Sinninghe Damsté et al., 1988; Sinninghe Damsté et al., 1989; Sinninghe Damste and De Leeuw, 1990) seems to be widely accepted. S-rich kerogens, mostly of marine Type II OM, break at weak sulphur sites and generate smaller but intact fragments of kerogen, asphaltenes and resins as well as low- molecular weight products, e.g. alkylated thiophenes and higher fused homologues, which can be identified in source rock pyrolysates and crude oils (Sinninghe Damsté et al., 1989).

1.3 Maturation of organic matter

1.3.1 Compositional evolution of organic matter

The increase of thermal energy due to burial of source rock horizons within a sedimentary basin causes structural re-organization and alteration of organic matter, changing its chemical and physical properties and ultimately resulting in the transformation of the solid kerogen to liquid and gaseous phase, the petroleum. The stages of this thermal maturation process are defined as diagenesis, catagenesis and metagenesis (Tissot and Welte, 1984). During diagenesis, mostly in depths less than 1000 m, the sedimentary organic matter undergoes biological and geochemical alteration processes to form kerogen by polycondensation, insolubilization and selective preservation of carbohydrates, amino, fulvic and humic acids and lipids (Tissot and Welte, 1984; Rullkötter and Michaelis, 1990).

During this stage, mainly (biogenic) CH4, CO2, H2O and heavy heteroatomic compounds are formed (Fig. 1.1). The changing physicochemical conditions upon progressive burial cause the kerogen structure to adapt to these new conditions by

1: INTRODUCTION the separation of smaller molecules from the macromolecular structure according to bond strength with weaker ones breaking before stronger ones. The most important kerogen degradation processes for generation of liquid petroleum and wet gases occur during catagenesis, while mainly dry, thermogenic methane is formed during metagenesis. As the kerogen structure develops towards higher aromatization and forms an ordered carbon structure, leaving behind a hydrogen-depleted residue and ultimately resulting in the potential formation of graphite (Radke et al., 1997), the generated volatile products structurally resemble the kerogen structure at the time of their separation (Tissot and Welte, 1984). General observations made on kerogen and coals throughout the whole maturation process is the defunctionalisation of acyclic and cyclic compounds (Wilson et al., 1984) in most cases, saturating with hydrogen or forming radical sites available for cross-linking reactions and polycondensation.

During diagenesis and early catagenesis, compounds with lowest thermal stability, e.g. peptidic, carbonyl and carboxyl bonds, get depleted from the kerogen (Fig. 1.1). Generated products, called initial bitumen, are characterized by heteroatomic, higher molecular weight fragments resembling the kerogen structure (resins, asphaltenes, heavy bitumen extracts), and linear, branched and cyclic aromatic and aliphatic

hydrocarbons as well as volatile gases (CO2, H2O,

N2, CH4). In elemental ratios, this is expressed in

Fig. 1.1. Evolution of sedimentary an O/C decrease, while pyrolytic products contain organic matter fractions at depth (modified after Durand et al., 1980). long-chained aliphatic and aromatic hydrocarbons and a high GC-unresolvable complex mixture, the hump representing oligomeric or functionalized, polar NSO compounds (Horsfield, 1997). These meta-products from kerogen conversion undergo secondary cracking as soon as they are formed and thus contribute to the hydrocarbon potential of the organic matter for ~30 % in the Type II OM-containing Toarcian Shale of the Paris Basin (Behar et al., 2008). With increasing maturation, catagenetic products are dominated by aliphatic and aromatic hydrocarbons, swamping the diagenetic products (Radke et al., 1997), while high-molecular weight NSO compounds are

1: INTRODUCTION progressively less abundant in bitumen (Durand, 1980). Having eliminated most functional groups from various bond types, residual NSO heteroatoms are transformed via cyclisation and aromatization reactions to heterocycles (Sinninghe Damsté et al., 1989 for sulphur-containing moieties) and released into the fluid state upon higher thermal stress. The generation of hydrogen-rich extractable products is expressed by a relative carbon increase, thus decreasing H/C and generation potential HI. Due to thermal instability of longer homologues, the average chain length of n-alkanes in coals and sediments decreases (Bray and Evans, 1961; Philippi, 1965; Leythaeuser and Welte, 1969). In pyrolysates, these are represented by doublets of n-alkanes/n-alkenes containing biological precursor information. Odd carbon numbers resulting e.g. from plant cuticular membranes (Tegelaar et al., 1989b) may survive to high levels of thermal stress (Horsfield, 1989). Until a kerogen has reached its maximum petroleum generation at the peak oil window being then called mature, generated fragments become smaller, devoid in oxygen and more aliphatic and aromatic, which is expressed by further chain length shortening of n-alkyls and higher proportions of low-molecular weight aromatics in pyrolysates (Larter and Douglas, 1980; Horsfield, 1989). Depending on kinetic stability of generated products, thus on kerogen type, the secondary cracking of non-expelled moieties starts during the late catagenesis resulting in formation of wet gas (C2-5) and ultimately of methane during metagenesis or the dry gas window (Table 1.1). The cracking of sulphur containing heterocycles from kerogen may form H2S (Tissot and Welte, 1984). Having transformed to a carbon- rich residue, the pyrobitumen, kerogen has no significant hydrocarbon generation potential left. However, a late gas potential in excess of 2.0 %RO is determined by the type of OM, i.e. the portion of methylaromatics, rather than the retention capacity of the source rock (Mahlstedt, 2012).

Table 1.1. Genetic stages of petroleum formation for a Type II kerogen which is an intermediate between Type I and III OM (after Espitalié, 1986). genetical or maturity stage VR (%RO) Tmax (°C) diagenesis immature < 0.5 - 0.7 % < 420 - 435 beginning of oil window 0.5 - 0.7 % 420 - 435 catagenesis beginning of wet gas window ~ 1.3 % 450 - 470 metagenesis beginning of dry gas window 2.0 % 470 - 520

1: INTRODUCTION

1.3.2 Common maturity markers

The described maturity stages can be identified using multiple optical, pyrolytical and chemical approaches based on kerogen and generated bitumen and including kinetics, bulk compositions, stereochemical and isotopic changes of specific compounds of which the screening methods of bulk properties will be presented here.

Having applied maceral analysis for identifying the origin of organic matter in coal petrography, describing coal ranks, the vitrinite reflectance (Vr) has been extended to particles of disseminated OM in petroleum source rocks (Teichmüller, 1971). Due to the minuscule particle size, a distinction of reflectivity of different macerals is difficult, thus measuring all identifiable reflecting macerals. Upon maturation, kerogen loses its volatile products, H2O, CO2, organic acids and hydrocarbons (Littke et al., 1989; van Krevelen, 1993), resulting in increased reflectances. Boundaries for different genetic stages of a Type II kerogen are given in Table 1.1. In that context, it must be noted that boundaries are floating for different kerogen types depended on their structural inventory starting oil generation earlier when composed of weaker bonds to heteroatoms. Similar observations have been made for pyrolytical approaches, e.g. the Rock-Eval pyrolysis (Espitalié et al., 1977). Measuring bulk volumes of generated petroleum (S1) and residual reactive organic matter (S2) as well as the temperature of maximum (artificial) generation (Tmax), the shape of the S2 peak mimics the kinetic stability of a kerogen. Depending on kerogen type and thermal stability of their inventory, the onset and maximum of petroleum generation varies between Type I generating at higher artificial and geological temperatures within a narrow temperature range and Type III kerogen starting generation at lower temperatures within a broader range. With ongoing maturation, the more instable kerogen moieties are converted to petroleum (S1) and thus S2 decreases while Tmax increases. The ratio between generated and residual pyrolysis yields, the so-called Production Index PI = S1 / (S1 + S2), can be used as maturity indicator but is highly sensitive to primary migration efficiency and could only be used in absence

of expulsion. Thus, Tmax is a well-suited maturity indicator for screening purposes that can be correlated with VR (Espitalié, 1986) but must be validated with yields (Schenk et al., 1990) and compositional information.

1: INTRODUCTION

Useful tools for maturity assessment based on the composition are undoubtedly biomarkers obtained from GC-MS on soluble bitumen extracts, such as stereoisomeric ratios and abundances of steranes, tri- and pentacyclic terpanes, but as well ratios obtained from aromatic (e.g. methylphenantrene index or naphthalene indices) and low-molecular weight NSO compounds (e.g. thiophenes, phenols and carbazoles). A comprehensive overview can be found in Peters et al. (2005). The advantage of extract-derived information is the comparability to crude oil compositions, however contamination of i.e. early products may be a risk. Thus, information from the kerogen structure itself is offered by Py-GC, e.g. kinetic stability of phenol (Larter, 1989), an abundant constituent of pyrolysates (Horsfield, 1997), or the increasing branching of alkylated thiophenes with maturity (Eglinton et al., 1990). Nevertheless, most of these ratios are influenced by source input (Larter, 1984; Eglinton et al., 1992) and must be adjusted for each source rock.

1.3.3 Timing of petroleum formation

Examining the prolificness of a petroleum system, the properties of accumulated petroleum are dependent on chemical variations of OM (source), temperature (its maturation history) and physicochemical changes during migration from source to carrier into reservoir rocks. The physicochemical conditions during the time of expulsion including temperature but as well pressure are strongly dependent on depth. Thus, the kinetic stability of OM interacts with burial depth, rate and heat flow. The latter factors are a function of the thickness and type of sedimentary record within a basin, the basin type, its formation mechanisms and history, and thus tectonic regime, e.g. intracontinental basins might subside slower than extensional marine or marginal-oceanic basins.

Subsidence rates of basins, or in particular heating rates of the organic matter, have a strong control on formation and cracking of petroleum fractions during artificial heating experiments – van Heek (1982) for coals and Horsfield (1997) for petroleum source rocks – with slower heating rates causing petroleum to form at lower temperatures (Schenk and Dieckmann, 2004). Dieckmann et al. (2000) have found that proportions of bulk yields of generated fractions remain constant for different laboratory heating rates (0.013 – 5 K/min), but compositional

1: INTRODUCTION changes occur with enrichment of n-alkanes and depletion of NSO compounds (the GC-hump) observed for decreasing heating rates. While in nature (1 – 5 °C/Ma) these compositional changes are less intense, they might still slightly alter the initial gas-to-oil-ratio. The depth of generation has a major influence not only on volumes and phase behaviour of generated fractions, but as well on migration efficiency and flow mechanisms as petroleum migrates as one saturated or two separated gas and oil phases through carrier systems when pressure and temperature gradually fall. Hereby, physical and chemical properties of individual petroleum compounds gradually change towards less intermolecular activity affecting solubility, sorption and partitioning between inorganic, organic, solid, fluid and gaseous phases of the source-carrier-reservoir system, i.e. minerals, kerogen, (heavy) bitumen, water, oil and gas.

1.4 Primary migration & expulsion

After generation of petroleum by proceeding maturation of organic matter, generated products move from the source rock via carrier beds into reservoir horizons which is called migration. The release of petroleum compounds from solid OM, the kerogen, in source rocks and their transport within and through the capillaries and narrow pores of a fine-grained source rock is defined as primary migration. Oil that is expelled from a source rock passing through wider pores of more permeable and porous rock units is called secondary migration. Migrating through various porous carrier beds, petroleum phases are stopped when reaching a significantly lower permeable rock unit, called seal rock, then forming a reservoir.

On their pathway through the subsurface, oil and gas mainly occur as a single fluid phase due to high pressure and temperature at the location of generation, and, as being the lightest moveable phase, generally tend to move vertically and interact with different solid and fluid phases. In the source rock, varying amounts of kerogen are embedded in a fine-grained, micro- and mesoporous, mostly impermeable matrix of siliciclastic or carbonate grains. On the other hand, (macro)porous carrier rocks are dominated by various minerals and a pore-filling water phase bound on these surfaces. The migration conduits are characterized by their petrophysical properties such as porosity, permeability and pore pressure which are determined by lithology and fabric of the sediment and influenced by

1: INTRODUCTION tectonic events. Based on the different surrounding conditions, various migration mechanisms prevail for primary and secondary migration experiencing a change during expulsion from the source.

1.4.1 Mechanisms of primary migration and expulsion

The most important form of subsurface petroleum movement is migration as a discrete hydrocarbon phase occurring during the main generation phase (Dickey, 1975; Momper, 1978; Hunt, 1979; Tissot and Welte, 1984; Welte, 1987; Durand, 1988; Ungerer, 1990; Mann, 1994). Until having reached the macropores, some less effective mechanisms as diffusion are likely seen as initial mechanisms until pressure-driven bulk flow takes over (Mann et al., 1997).

Pressure-driven bulk flow. Initiation of petroleum flow within macropores or fractures in a source rock requires a pressure gradient which is generated by compaction due to burial (Athy, 1930; Meissner, 1978; Hunt, 1979; du Rouchet, 1981) or by generation of excess amounts of petroleum (Mann, 1994). Compaction causes increasing pressure acting upon mineral grains and transfers the stress to pore fluids and the load-bearing kerogen (Kroos et al., 1993) resulting in overpressure by trapped pore water. If pore water is trapped in the source rock, e.g. by oil droplets blocking pore throats or in unconnected pores, water expands with increasing temperature (aquathermal pressuring, Barker, 1972). Although most water has left the source rock upon beginning petroleum generation (Burst, 1969; Perry Jr and Hower, 1972), additional water may be formed from clay mineral dehydration due to thermal conversion, e.g. kaolinite-illite reactions at ~130 °C (Bjørlykke, 1998) or upon conversion of kerogen into petroleum (Lewan, 1997), promoting migration by hydrodynamic flow. Conversion of semi-solid kerogen to liquid oil and/or gas is accompanied by a volume increase of 10 – 20 % (Goff, 1983) which is most effective when generating gas (Yükler and Dow, 1990; Hunt et al., 1994) and provides internal pressure to the source rock. This effect is counteracted by a porosity increase of kerogen which can be as high as 31 % (Eseme et al., 2012). But compaction and swelling of the kerogen by soaking with its generation products (Ertas et al., 2006; Kelemen et al., 2006b) reduce porosity. Overcoming the lithostatic pressure at depth of generation, expulsion fractures may be formed (Mann, 1994) significantly enhancing migration efficiency. Similar to the concept of

1: INTRODUCTION hydraulic fracturing in shale gas plays, a high excess pressure is necessary for fracturing depending on ductility and brittleness by clay minerals, organic matter and carbonate content. The healing of these fractures complicates confirmation of this concept.

Diffusion is a molecular motion to balance a difference in chemical potential (concentration, pressure, gravity, temperature) in a non-solid medium and is by orders of magnitude shorter and slower than bulk fluid flow (Leythaeuser et al., 1987; Sandvik and Mercer, 1990; Thomas and Clouse, 1990b). It can occur in aqueous phase, but due to higher solubility of petroleum in kerogen, potential differences and thus fluxes are higher in OM network (Sandvik and Mercer, 1990; Thomas and Clouse, 1990b, c). The tortuosity of this 3D network is a limiting factor with more efficient migration in OM-rich shales possessing low tortuosity (Stainforth and Reinders, 1990). A TOC cut-off is found at 2.5 % TOC for non- laminated shales (Thomas and Clouse, 1990c). Diffusion is most important where the critical saturation level for inducing bulk flow has not been reached, either during early generation stages, in lean source rocks, or in those with high porosity (Thomas and Clouse, 1990a), transporting bitumen to fractures or bitumen-filled pathways.

Solution in water is limited to the most soluble, light hydrocarbons such as methane, ethane, propane, benzene, toluene and other low boiling cyclic compounds (McAuliffe, 1966; Price, 1976; McAuliffe, 1980). Paraffins and the attachment of alkyl groups significantly lower water solubility, and thus this process may alter the initial petroleum composition.

Solution in gas may contribute when bitumen saturation is too low for oil phase flow (Sokolov et al., 1964). Especially in oxygen-rich Type III OM, the

solution in methane (Price, 1989; Leythaeuser and Poelchau, 1991), CO2 and other hydrocarbon gases (Bray and Foster, 1980) mobilizes liquid hydrocarbons.

During proceeding petroleum generation upon maturation of organic matter, the pores and migration conduits of a source rock have to reach a certain level of petroleum saturation until (pressure-driven) bulk flow is initiated (McAuliffe, 1980; England et al., 1987; Leythaeuser et al., 1987; Sandvik and Mercer, 1990; Forbes et al., 1991; Pepper, 1991; Düppenbecker and Welte, 1992). Welte (1987) and Mann

(1994) reported a minimum oil saturation (SO) of 15 – 25 % and Okui and Waples

1: INTRODUCTION

(1993) a maximum of 25 %. However, SO of 40 – 60 % can be reached during peak oil generation with porosity acting a buffer and may retard expulsion. During expulsion from the source rock, the frame conditions for hydrocarbon migration significantly change from OM-rich, less porous and permeable sequences that are enriched in petroleum fluid to OM-lean horizons with increased water-filled porosity and/or permeable conduits present, as interconnected matrix porosity or fracture systems. A thin source rock interval might expel its generated products very efficiently. A thick source rock generating throughout its entire succession is only depleted in generated yields at the edges to carrier beds (Mackenzie et al., 1988) while primary migration mechanisms are the rate-limiting factor. Here, the major geochemical fractionation occurs as a function of the distance between place of generation and carrier bed (Leythaeuser et al., 1984b; Leythaeuser and Schaefer, 1984; Mackenzie et al., 1987; Leythaeuser et al., 1988a; Leythaeuser et al., 1988b). Such effects were illustrated to be only significant in the initial phase of petroleum generation and/or in OM-lean source rocks (Leythaeuser et al., 1987; Mackenzie et al., 1988).

This chemical fractionation during primary migration and expulsion has an important impact on petroleum behaviour during migration and ultimately on its quality in the reservoir. Rate-limiting processes have been ascribed to either sedimentological and petrophysical development upon burial (e.g. Eseme et al., 2012) or organofacies characteristics influencing the formation of oil from kerogen breakdown. Petrophysical features attribute sedimentary facies of the source rock (pore size distribution, pore throat diameters and relative permeabilities), together with the extent of fine silt-sand interbedding (Raji et al., 2015a). Chemical fractionation can be caused by the interaction of the petroleum fluid with mineral matrix. Examples of such interactions include size exclusion in pelitic rocks (Krooss et al., 1991), fractures due to overpressure from compaction (Meissner, 1978; Hunt, 1979; du Rouchet, 1981) or hydrocarbon generation (Mann et al., 1997). In addition, sorption on mineral surfaces (Carlson and Chamberlain, 1986; Barrer, 1989; Brother et al., 1991; Taylor et al., 1997), partitioning with gas phase (gas washing) (Losh and Cathles III, 2010; Bourdet et al., 2014), or partitioning with formation water (Zhang et al., 2005) have been suggested as influencing compositions at the molecular level. In recent years, it has come to light that the indigenous organic matter in source rocks plays a direct role in retention and expulsion. Adsorption on

1: INTRODUCTION the kerogen surface (Lamberson and Bustin, 1993), absorption in the kerogen structure (Young and McIver, 1977; Sandvik et al., 1992) or diffusion through organic matter (Thomas and Clouse, 1990c) have been reported as causing selective fractionation. The efficiency is primarily controlled by TOC and kerogen composition (Sandvik et al., 1992). Based on relative proportions of free, generated petroleum and residual kerogen in immature and mature source rock sections within a sedimentary basin, the calculation of expulsion efficiencies of a particular source rock package or sample at particular maturity stages allows estimates on amounts of potentially accumulated petroleum in a petroleum system (Jones, 1980; Leythaeuser et al., 1984a; Cooles et al., 1986; Mackenzie et al., 1987; Düppenbecker et al., 1991; Braun and Burnham, 1992; Pepper and Corvi, 1995). Implementing the composition of petroleum for generated fractions (Forbes et al., 1991; Losh and Cathles III, 2010) or on a molecular level (Horsfield et al., 2001) enables the prediction of generated, retained and expelled (accumulated) portions and may reduce exploration risks in a particular hydrocarbon play or petroleum system. However, slight variations in kerogen type and preservation, thus in the source rock quality, complicate such efforts (Ziegs et al., 2015). Different migration mechanisms at particular generation stages controlling fluxes and expelled compositions represent a further obstacle in the correct compositional prediction (Kuske, S., pers. comm., 2017).

1.4.2 Migration fractionation

A compositional difference between the dispersed bitumen of shales and reservoir crudes, attributed to primary migration, was first described by Brenneman and Smith Jr (1958) who observed an enrichment of saturated hydrocarbons in reservoired petroleum and a prominence of asphaltenes and resins in source rock bitumens (e.g. Tissot and Pelet, 1971; Tissot and Welte, 1984). The preferential expulsion of aliphatic over aromatic hydrocarbons was noted by Baker (1962). For shale-sandstone contacts, Leythaeuser et al. (1984a; 1988a; 1988b) showed that preferential retention occurs in the sequence asphaltenes > NSO compounds > aromatic hydrocarbons > saturated hydrocarbons, and iso- paraffins over n-paraffins, and vice versa for the compositions of expelled petroleum. The simplest explanation for compositional fractionation is that small molecules migrate more easily than larger, more awkwardly shaped ones by

1: INTRODUCTION diffusion resulting in preferential primary migration of gas molecules (Stainforth and Reinders, 1990). This selective sequence has been reproduced in laboratory experiments on the solubility of organic compounds in kerogen (Lafargue et al., 1990; Sandvik et al., 1992; Ritter, 2003b, a).

Relative solubility of different organic components and mutual mixtures in water (McAuliffe, 1966; Price, 1976, 1981; Price et al., 1983; Yaws et al., 1991) plays a significant role in organic-lean source rocks whereas solubility of petroleum compounds in kerogen (Ritter, 2003b, a) is important for TOC-rich source rocks with high Hydrogen Indices (Han et al., 2015). Retaining its generated products by swelling, the kerogen network attracts compounds that are chemically similar, thus the affinity of polar NSO’s, especially of C14+ NSO’s, to absorb is higher than that of aromatics and n-paraffins (Kelemen et al., 2006a) reproducing the sequences and confirming mechanisms reported earlier (Leythaeuser et al., 1984b; Leythaeuser and Schaefer, 1984; Ritter, 2003b, a).

1.5 Secondary migration from source to reservoir

1.5.1 Mechanisms of secondary migration

Secondary migration is the movement of petroleum through OM-depleted, water-filled pores of a permeable carrier bed into reservoir horizons. This movement is controlled by three parameters determined by the differences of physical and chemical properties of organic petroleum compounds and the exterior of inorganic mineral and water phases (Tissot and Welte, 1984):

(1) The buoyant rise in water-saturated pores is the driving force due to oil and especially gas being the lowest density phases in the subsurface (0.7 – 1.0 g/cm³ and <0.001 g/cm³, respectively). (2) Capillary pressure, the contrary force to buoyancy, is the net force acting on contact surfaces of different phases, the surface or interfacial tension, to produce a pressure difference across the interface. It is defined as =

�� ( ) with γ is interfacial tension, r = radius of droplet, β is angle� of 2 ∙ γ �petroleum∙ cos � water interface and cos β defines the wettability of a rock.

1: INTRODUCTION

(3) Hydrodynamic fluid flow, following the laws of Darcy flow, can be directed towards or against the direction of hydrocarbon flow and thus modifies its efficiency determined by buoyant rise (Hubbert, 1953, 1954). Multiple immiscible phases in pore spaces reduce a capability of individual fluids to flow relative to their capacity, defining the concept of relative permeability.

Subsurface hydrocarbon flow is generally controlled by these modes, but due to frame conditions distinguished by reduced porosity, permeability and water availability in source rock horizons, they are only prevailing during secondary migration. In porous space, hydrocarbons follow the path of least resistance, i.e. through interconnected pore throats (e.g. Hobson, 1954; Berg, 1975). England et al. (1987) have calculated that a petroleum saturation of up to 50 % of a net pore volume or 1 – 10 % of the gross volume of a carrier rock is necessary to establish bulk flow. To initiate bulk flow, petroleum need to replace the free water in pores, reducing the interfacial tension to water until overcoming capillary pressure, which is determined by the pore throat diameter and the size of the oil globule. Interestingly, larger globules move faster than tiny ones due to lowered interfacial tension per volume of the droplet (Athy, 1930). If no pressure gradient is present for a particular incremental petroleum charge, this equilibrium will only be outbalanced by tectonic events or by additional generation. Reaching horizons with too small pore throats, capillary pressure exceeds buoyancy forces and migration is terminated within a reservoir rock. Vertical migration distances in matrix porosity may be as long as the thickness of a reservoir and if exceeding, it occurs only via fault or fracture systems or other preferred avenues as dikes, thrust plains and mud volcanoes (Tissot and Welte, 1984).

1.5.2 Processes of chemical fractionation

Mechanisms known from primary migration, such as diffusion, solution in gas and water may dominate the initial phase of secondary migration or the initial petroleum charge at the migration front composed of smaller, more dispersed oil droplets (Tissot and Welte, 1984). Silverman (1965) attributed an increase of nonpolar hydrocarbons and decrease of asphaltenes, resins, porphyrins and non- hydrocarbons in crude oils to adsorption on solid surfaces, depending on wettability of the carrier rock. Due to the steady contact of oil with the surrounding pore water,

1: INTRODUCTION interaction with water and mineral surfaces is pronounced for low-boiling, more water-soluble and/or polar compounds of the migrating petroleum (e.g. McAuliffe, 1966) and these molecules are particularly concentrated at oil-pore water interface. Under hydrodynamic conditions, also called water washing, such compounds may be partially removed from the migrating phase. Aromatic hydrocarbons are more water-soluble than cyclic and normal paraffins, thus the preferential loss of aromatics further results in a slight decrease of 13C/12C ratios (Silverman, 1965). A strong effect, also on aromatic moieties, may be due to increasing aliphatic chain lengths to core structures reducing water solubility, e.g. for phenol and cresol, but this was not observed for alkylated napthtalenes (Leythaeuser et al., 1988a).

Phenols and its alkylated homologues have been identified as strong agents in crude oils altering wetting properties of carrier and reservoir rocks or interacting with non-hydrocarbon phases (Larter and Aplin, 1995; Bennett and Larter, 1997; Larter et al., 1997; Bennett et al., 2004; Bennett et al., 2007b). Oil-water partitioning and solid-phase adsorption are controlled by the extent and position of alkylation to the functional group (e.g. Taylor, 1994). Phenols and cresols destabilize the water film around minerals by partitioning due to their elevated water solubility facilitating apolar sorption on polar surfaces (Bennett et al., 2004) and thus removes these compounds from the migrating oil-phase. Taylor et al.

(1997) observed a systematic decrease in total C0-3 phenols in four North Sea oils with migration distance. Partitioning coefficients of alkylphenols are negatively correlated to temperature, water salinity and the bulk NSO content of crude oils, whereas pressure has no influence (Bennett and Larter, 1997).

Similar fractionation effects related to migration distance have been observed for carbazole homologues (Li et al., 1994; Stoddart et al., 1995; Larter et al., 1996; Li et al., 1997; Bennett et al., 2002). Stoddart et al. (1995) and Li et al. (1995) showed that alkylation, molecular size and the activity of the polar atom or functional group influenced by shielding effects are distinguishing factors for molecular fractionation due to sorption onto clay minerals (e.g. Stoddart et al., 1995; Larter et al., 1996) or oil-water partitioning (van Duin and Larter, 2001). Longer migrated oils in the Eldfisk field, Central Graben are enriched in alkylcarbazoles relative to alkylbenzocarbazoles, higher methylated C3 homologues are less hindered in migration than C1 and C2 homologues, and the rod-shaped [a] configuration is depleted over the subspherical shaped [c] isomer of

1: INTRODUCTION benzocarbazoles. However, such distributions can be affected by source facies (Bakr and Wilkes, 2002) and/or maturity (Li et al., 1997; Clegg et al., 1998a; Clegg et al., 1998b; Horsfield et al., 1998). Li et al. (1994; 1995) suggested that migration fractionation has a higher impact on pyrrolic nitrogen abundances and distributions than the source environment, but varying lithology and geometry along the migration pathway also affect these distributions (Li et al., 1998).

1.5.3 Phase behaviour during up-rise

During generation and expulsion in depths exceeding 3500 m (Cornford, 1994), petroleum is expelled as a single saturated liquid phase containing dissolved gas due to pressure and temperature (pT)-conditions as high as ~150 °C and 400 – 500 bar (Larter and Mills, 1991). Following buoyancy and physical gradients during migration, pressure and temperature drop until separation into an exsolved, free gas and a remaining, shrinked liquid phase with higher viscosity (Neumann et al.,

1981). Physical properties which are typically described by saturation pressure Psat, shrinkage factor BO, gas-to-oil ratio (GOR) or within a pT-phase diagram are directly controlled by petroleum composition as a function of source, maturation and primary migration and expulsion (di Primio and Horsfield, 2006).

Phase behaviour-related effects occurring during migration or production are the dominant mechanism forming condensates (Larter and Mills, 1991) and have a severe impact on petroleum composition, thus physical and chemical properties. Such effects include phase separation (development of a two-phase-system) and concurring differential migration (physical separation of the two phases due to their significantly different physical properties) or phase fractionation (e.g. gas washing). Gas condensates are mostly of evaporative origin rather than a result of thermal cracking (Thompson, 1987).

Generally, compositional variations due to phase separation are controlled by molecular size, compound structure and partitioning between different petroleum compounds and fractions (Larter and Mills, 1991; van Graas et al., 2000). Strongest compositional influences on phase behaviour is carried out by the relative proportions of low-density, low-viscosity constituents in the C1–C4 range, on the one hand, and high-density, high-viscosity constituents, especially colloidal components, such as asphaltenes, on the other (Hunt, 1979; Connan and Coustau,

1: INTRODUCTION

1989) and thus the GOR. Multiple-stage separation upon proceeding vertical migration at various pT-conditions may intensify observed alterations (van Graas et al., 2000). The vapour phase is enriched in smaller molecules, preferentially in paraffins over naphthenes (Larter and Mills, 1991) in the low-molecular range, but contain more n-paraffins and elevated contributions of light aromatic and naphthenic hydrocarbons in their C6+ fraction (Thompson, 1987, 1988). This is due to the gas phase having a good solvent power for larger compounds including C20+ at higher pressures (>300 bar, van Graas et al., 2000), and because aromatic hydrocarbons retain preferentially in that liquid phase (Larter and Mills, 1991). Furthermore, the associated gas is wetter than the free, exsolved gas due to progressive loss of light ends in condensates and gas-saturated oils (Thompson,

1987). Consequently, the residual oil contains increased heavy C25+ fraction with decreased saturate/aromatic ratio, but most significant alterations to the evaporative phase can be determined in the lower molecular fraction (C<9) influencing the Thompson parameters: an increase in aromaticity and normality (unbranched and naphthenic over branched isomers) can be observed while paraffinicity is decreasing. The retrogressive changes in maturity indicators leads to spurious immaturity in residual oils and derived condensates (e.g. lower heptane value: nC7/sum of C6-7), while the gas/condensates derived from a first phase separation appear more mature (Thompson, 1987, 1988). The distribution of carbon isotopes remains untouched but varies due to compound specific fractionation within different boiling fractions (Larter and Mills, 1991; van Graas et al., 2000).

Molecular parameters in the C9+ fraction are varyingly affected depending on differences in molecular size of compounds but the partitioning effect distributes higher-molecular markers between vapour and liquid phases. Predicting geochemical parameters is difficult and depends on the pT-conditions of phase separation, thus migration history is of crucial importance (Larter and Mills, 1991). While Pr/Ph are increasing for the evaporative fraction and remaining stable in the residual oil, the size differences between Pr/n-C17 and Ph/n-C18 are lower. Thus, both parameters are not affected (van Graas et al., 2000). C27 to C29 sterane distributions and the methylphenantrene index (MPI-1) are not affected by single- stage fractionation but vary for lower pressures with lower homologues partitioning into the vapour phase (Larter and Mills, 1991; van Graas et al., 2000).

1: INTRODUCTION

1.6 The scope of this thesis

The correct assessment of the prospectivity of a given petroleum system depends on the understanding of chemical and physical processes taking place during its geological evolution. Among these processes petroleum generation, expulsion and migration are the most important. An initial screening of the principal source rock in the Central Graben, the Mandal Formation, using TOC and Rock Eval data has revealed significant deviations from the commonly applied sorption thresholds of e.g. Pepper and Corvi (1995). When assessing bulk volumes generated and expelled from a source rock, such approaches omit a range of chemical and physical details of the processes and effects involved, which may be critical for predicting the amount and type of retained and expelled fluids, and variations of these amounts and compositions as a function of maturity. Hydrocarbon accumulations predicted by 3D basin modelling to be filled have raised the question as to whether charging is an issue related to high thermal stability of the source organic matter or delayed expulsion from source to the carrier-reservoir system

The Central Graben represents a good natural laboratory containing numerous oil, gas and condensate accumulations that has been explored since 1969 when the giant Ekofisk oil field, still the 5th largest field at the Norwegian continental shelf (NCS) in 2014, was discovered. Crucial well bore information on geological, geochemical and petroleum engineering data of rock and fluid samples is easily assessable online supporting interpretation. This failed-rift basin contains a more or less homogenous source rock, the clay- and TOC-rich Mandal Formation, which was deposited in anoxic milieu during the Upper Jurassic main rifting phase and buried by continuous, differential subsidence to depths ranging from early to late mature stages and has been tectonically more or less unaffected during overburden deposition (Cornford, 1994, 1998). An overview about the geological history of the Central Graben can be found in the following chapters. Migration in the axial regions is seen to be principally vertical through fractures and faults with minor lateral components and in marginal regions reservoir and source rocks are in direct contact.

The current PhD study was initiated to re-assess the understanding of generated, retained and expelled fluid compositions as a function of maturity and

1: INTRODUCTION kerogen type, and unravel processes controlling compositional fractionation during expulsion occurring in different evaporative fractions generated.

Having evaluated different source rock intervals at the Norwegian parts of the Viking and Central Grabens, mass balance calculations have revealed that the Mandal Formation contains significantly higher amounts of is generated organic matter while other Type II OM-containing horizons, namely the Draupne, Heather and Drake Formations in the Viking Graben are extremely effective in expelling their generation products. Thus, the overall objective of this project is:

• Why is the Mandal Formation such a bad expeller? / Why does the Mandal possess higher sorption thresholds than e.g. the Draupne Formation in the Viking Graben although obviously of similar origin?

Following recent scientific developments in factors controlling expulsion (e.g. Sandvik et al., 1992; Pepper and Corvi, 1995; Kelemen et al., 2006a; Mahlstedt and Horsfield, 2013), we exclusively examined the organic matter portion of the source rock and evaluated generative and generated, volatile and high-molecular weight boiling fractions. The application of conventional GC-based and novel mass spectrometric methods for evaluating both, the hydrocarbon and high-molecular weight fractions was necessary to answer the following research questions. Descriptions to those methods are given in the respective chapters.

• How does the vertical and lateral variation of generative organic matter in the Mandal Formation contribute to this issue by its composition and generation characteristics? • What are the individual controls on petroleum expulsion and retention in the Central Graben? • What is the role of heavier products with higher boiling range?

Consecutively, expelled portions represented by accumulated oils in the Central Graben were examined focussing on their polar, high-molecular weight inventory. The principal question was:

• Are particular oils at the Northern Mandal High different to other crude oils found in the Central Graben?

1: INTRODUCTION

1.6.1 Sample selection

Reaching this goal, a database was provided by the project partner Aker BP ASA and the Norwegian Petroleum Directorate containing bulk Rock-Eval data sets and well bore information of 23,300 source rock samples from different regions of the Norwegian continental shelf (NCS). These data served as input for mass balance calculations of the expulsion efficiency and for sample selection. Principal source rocks in the Norwegian North Sea are Draupne (2778 samples), Heather (3545 samples), Drake Formations (330 samples) in the Viking Graben as well as Mandal (694 samples) and Farsund Formations (1414 samples) in the Central Graben. The Farsund Formation was excluded from the current study because at nearly all basin locations, the Mandal acts as seal to the Farsund Formation and therefore, the Mandal Formation was considered to be the principal source rock in the basin (Cornford, 1994).

Samples available for calculation of expulsion efficiency and with interest towards re-measurement at GFZ contain > 1 % TOC, do not exceed a maximum natural saturation of 200 mg HC/g TOC and do not contain > 10 % generation products at immature stage. Thus, around two third of the source rock sample sets were excluded and the Mandal sample set was confined from 694 to 222 samples. Based on this revised sample set, two natural maturity series’ of 28 Mandal source rock samples from different locations of the Central Graben and with different initial generative potentials (assumed kerogen sub-types) were selected for detailed geochemical analyses following widely accepted theories of decreasing TOC and HI (Tissot and Welte, 1984). Based on 68 available fields or discoveries in the Central Graben, 24 crude oils and condensates were selected from different regions of the Norwegian Central Graben. The sample choice was solely based on API gravity, fluid type, sample position to the source rock and some sparse additional information of maturity data. Unfortunately, when rock and fluid samples were available in the same or in closely neighbouring wells, mostly at marginal basin positions, fluids were discovered in deeper horizons than the Mandal Formation occurred leading to the conclusion that it must have been sourced from a deeper buried Mandal section. Thus, a direct comparison of source rock and fluid compositions was not possible in the course of this study as potential lateral variations could not be excluded.

1: INTRODUCTION

1.6.2 The structure of this dissertation

Subsequent to this introduction of primary and secondary factors on petroleum compositions in explored accumulations, the petroleum generation and bulk retention characteristics of the Mandal Formation as principal source rock in the Central Graben are presented in chapter 2, and molecular properties of different fractions of expelled crude oils in the Central Graben are described in chapter 3. Chapter 4 focussed then on the chemical structure of polar, high- molecular weight compound classes retained in the Mandal Formation source rock which have been shown to play an important role in retention of volatile products.

Chapter 2 reveals the influence of kerogen on retention of gas and oil in the source rock combining a NCS-wide data base of bulk Rock-Eval data and selected whole rock and extracted samples, applying mass balance calculations, bulk and molecular characteristics of the petroleum generative as well as of generated fractions.

Chapter 3 identifies molecular alterations during secondary migration in different carrier-reservoir rock systems over varying distances, and discusses individual processes controlling petroleum compositions. A combination of conventional GC methods and the novel FT-ICR-MS approach allows characterization of the hydrocarbon as well as heavy polar fractions.

Chapter 4 applies the same approach as in chapter 3 focussing on the retained portion within the source rock and shows which processes and physicochemical properties of the generated heavy, polar products are indicative for the Mandal Formation in comparison to an excellent expeller such as the German Posidonia Shale.

Chapter 5 summarizes this research project and gives an outlook on future topics.

1: INTRODUCTION

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2 Petroleum retention in the Mandal Formation, Central Graben, Norway

2.1 Abstract

Upper Jurassic organic matter-rich, marine shales of the Mandal Formation have charged major petroleum accumulations in the North Sea Central Graben including the giant Ekofisk field which straddles the graben axis. Recent exploration of marginal basin positions such as the Mandal High area or the Søgne Basin has been less successful, raising the question as to whether charging is an issue, possibly related to high thermal stability of the source organic matter or delayed expulsion from source to carrier.

The Mandal Formation is in part a very prolific source rock containing mainly Type II organic matter with <12 wt.- % TOC and HI <645 mg HC/g TOC but Type III-influenced organofacies are also present. The formation is therefore to varying degrees heterogeneous. Here we show, using geochemical mass balance modelling, that the petroleum expulsion efficiency of the Mandal Formation is relatively low as compared to the Upper Jurassic Draupne Formation, the major source rock in the Viking Graben system. Using maturity series of different initial source quality from structurally distinct regions and encompassing depositional environments from proximal to distal facies, we have examined the relationship between free hydrocarbon retention and organic matter structure. The aromaticity of the original and matured petroleum precursors in the Mandal source rock plays a major role in its gas retention capacity as cross-linked monoaromatic rings act on the outer surface of kerogen as sorptive sites. However, oil retention is a function of both kerogen and involatile bitumen compositions. Slight variations in total petroleum retention capacities within the same kerogen yields suggest that texture of organic matter (e.g. organic porosity) could play a role as well.

This chapter (post-print) is published as: Ziegs, V., Horsfield, B., Skeie, J.E., Rinna, J. (2017). Petroleum retention in the Mandal Formation, Central Graben, Norway. Marine and Petroleum Geology 83, 195-214. DOI: 10.1016/j.marpetgeo.2017.03.005

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2.2 Introduction

The North Sea is one of the most prolific hydrocarbon provinces in the world (Bowen, 1992; Brennand et al., 1998) that are mainly sourced by Upper Jurassic marine shales and containing numerous giant oil and gas fields (Brent, Statfjord, Troll, Oseberg, Gulfaks, Ekofisk). Known variably across the Northwest European Shelf as the Kimmeridge Clay, Draupne, Spekk, Mandal and Farsund Formations, these shales have been studied in great detail over the past 50 years, encompassing their depositional environment, petroleum generation potential, generated petroleum composition and phase behaviour (Cooper and Barnard, 1984; Demaison et al., 1984; Huc et al., 1985; Thomas, 1985; Baird, 1986; Karlsson, 1986; Cooper et al., 1995; Cornford, 1998; Isaksen and Ledje, 2001; di Primio and Skeie, 2004; di Primio and Horsfield, 2006). Thanks to the shales favourable organic richness, organic matter quality and maturation characteristics, the source has been said to be a prime example of a prolific petroleum expeller with Petroleum Expulsion Efficiency estimates as high as 95 % as calculated using mass balance models (Goff, 1983; Cooles et al., 1986; Wilhelms et al., 1990).

The Norwegian and Danish Central Graben fall in what might be termed the “excellent petroleum expeller” category. Upper Cretaceous chalk sequences representing the major reservoir in the Mandal-Ekofisk petroleum system (Ziegler, 1990; Cornford, 1994) contain 3061 mill. Sm³ in-place petroleum (87 % of total regional resources). The Ekofisk Field was still the 5th largest producing oil field in the Norwegian sector in 2014 (www.npd.no) with a total volume of 1134 mill. Sm³ oil-in-place (OIP) and 301 bill. Sm³ gas-in-place (GIP). Research on source rock characteristics in the Norwegian and Danish Central Graben has been undertaken by Cornford (1994), Petersen and Brekke (2001), Ineson et al. (2003), Petersen et al. (2011), Petersen et al. (2012b) and Petersen et al. (2013). It has been shown that the Mandal Formation source rock was deposited in sea-floor depressions controlled by reactivated Zechstein salt movements during the Upper Jurassic, and is enriched in marine organic matter residues (Cornford, 1994; Rossland et al., 2013). Surrounding landmasses were probably rich in higher plant vegetation and contributed lignocellulosic components to the kerogen. This is particularly the case along the graben flanks of e.g. the Mandal High where debris flows and turbidites transported coarser material deeper into the basin (Rossland et al., 2013). Cenozoic

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY siliciclastic deposition accounts for the majority of the up to 4000 m thick post-rift sedimentation and it is this subsidence phase that brought about the thermal maturation of the source rocks. The principal oil generation zone spans between 3650 and 4200 m burial depth in the central parts of the Central Graben (Stoddart et al., 1995) and the Mandal Formation is thermally mature to highly mature (Neumann, 2007). At basin margins Ryazanian successions have not yet reached the principal oil generation zone, but kinetic studies show a conversion of 15 % has been attained (di Primio et al., 2000).

central marginal

Fig. 2.1. Structural map of the investigation area. The general stratigraphy of the Central Graben area in the Norwegian sector is shown.

In the Søgne Basin of the Norwegian Central Graben (Fig. 2.1) the situation is more complex due to decreasing source rock quality from northeast to southwest and inverse maturity trends with highest maturity in the southwest (Petersen et al., 2011). But, as oil generation takes place at depths between 3250 and 3900 m, only in southwestern parts of this sub-basin the Mandal Formation is sufficiently thermally mature to generate oil. There has been only one discovery to date, located in the southernmost Trym area, and this was not sourced by the Mandal, but by the

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Middle Jurassic, coaly Bryne Formation (Petersen and Brekke, 2001; Petersen et al., 2011) which extends to form another petroleum system in the Danish part of the Central Graben (Andsbjerg et al., 2001). Mandal/Farsund-sourced production is in reality absent, and recent exploration has raised important questions about the generation and/or expulsion capabilities of the Mandal Formation in that area. Structural traps predicted by 3D modelling as containing economic amounts of Mandal-generated petroleum were in fact dry, despite trap, seal and reservoir risks being favourable. It was with this in mind that a study of primary migration from the Mandal Formation was initiated, and is the subject of the current communication.

Primary migration mechanisms have been comprehensively reviewed earlier by Pepper and Corvi (1995) and Mann et al. (1997). Diffusion through the kerogen network (Thomas and Clouse, 1990a, b, c) and discrete bulk petroleum phase (Vandenbroucke, 1993; Hunt, 1996) can occur, depending on the ratio of bitumen to kerogen. Molecular diffusion and solution are limited to the smallest molecules (Jones, 1980; Leythaeuser et al., 1982; Hunt, 1996). Diffusion is an important but slow mechanism at the immature to early mature stage of organic-rich source rocks or during the entire generation in organic leaner, moderate quality source rocks (Thomas and Clouse, 1990a). Petroleum migrates through the kerogen network towards migration avenues (interconnected pore spaces or fractures) where bulk flow can take over when the critical oil saturation has been reached. The critical oil saturation ranges between 15 and 25 % before bulk-flow expulsion (Welte, 1987; Okui and Waples, 1993; Mann, 1994) and between 40 and 60 % at peak oil generation (Mann et al., 1997). Bulk flow as a discrete petroleum phase is the principal and most effective primary migration mechanism in the subsurface, and is driven by pressure and buoyancy gradients within macropores and fractures. These gradients are mainly caused by compaction due to burial, augmented by volume expansion due to transformation from a solid kerogen phase to a liquid petroleum phase (Momper, 1978; Hunt et al., 1998; Modica and Lapierre, 2012), the so called “generation pressure” (Mann et al., 1997). A compositional difference between the dispersed bitumen of shales and reservoir crudes, attributed to primary migration, was first described by Brenneman and Smith Jr (1958) who observed an enrichment of saturated hydrocarbons in reservoired petroleum and a prominence of asphaltenes and resins

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY in source rock bitumens (e.g. Tissot and Pelet, 1971; Tissot and Welte, 1984). Models for that chemical fractionation include (1) the preferential movement of hydrocarbons away from kerogen within shales, and (2) the drainage of hydrocarbons out of shales. Rate-limiting processes have been ascribed to either sedimentological and petrophysical development or organofacies characteristics influencing the formation of oil from kerogen breakdown. Petrophysical features attribute sedimentary facies of the source rock (pore size distribution, pore throat diameters and relative permeabilities), together with the extent of fine silt-sand interbedding (Raji et al., 2015b). Chemical fractionation can be caused by the interaction of the petroleum fluid with mineral matrix. Examples of such interactions include size exclusion in pelitic rocks (Krooss et al., 1991), fractures due to overpressure from compaction (Meissner, 1978; Hunt, 1979; du Rouchet, 1981) or hydrocarbon generation (Mann et al., 1997). In addition, sorption on mineral surfaces (Carlson and Chamberlain, 1986; Barrer, 1989; Brother et al., 1991; Taylor et al., 1997), partitioning with gas phase (gas washing) (Losh and Cathles III, 2010; Bourdet et al., 2014), or partitioning with formation water (Zhang et al., 2005) have been suggested as influencing compositions at the molecular level.

In recent years it has come to light that the indigenous organic matter in source rocks plays a direct role in retention and expulsion. Adsorption on the kerogen surface (Lamberson and Bustin, 1993), absorption in the kerogen structure (Young and McIver, 1977; Sandvik et al., 1992) or diffusion through organic matter (Thomas and Clouse, 1990c) have been reported as causing fractionation. The efficiency is primarily controlled by TOC and kerogen composition (Sandvik et al., 1992). For example, Leythaeuser et al. (1984b) and Leythaeuser et al. (1988b) showed that preferential retention occurs in the sequence asphaltenes > NSO compounds > aromatic hydrocarbons > saturated hydrocarbons, and iso-paraffins over n-paraffins. The simplest explanation for compositional fractionation is that small molecules migrate more easily than larger ones by diffusion (preferential primary migration of gas molecules) (Stainforth and Reinders, 1990). This selective sequence has been reproduced in laboratory experiments on the solubility of organic compounds in kerogen (Sandvik et al., 1992; Ritter, 2003b, a). Relative solubility of different organic components and mutual mixtures in water (McAuliffe, 1966; Price, 1976, 1981; Price et al., 1983; Yaws et al., 1991) play a

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY significant role in organic-lean source rocks whereas solubility of petroleum compounds in kerogen (Ritter, 2003b, a) is important for organic-rich (high TOC) source rocks with high Hydrogen Indices (Han et al., 2015).

This study documents the source richness, source quality, thermal lability, and expulsion characteristics of the Mandal Formation. A large geochemical database (for benefits of statistically relevant datasets see Cornford et al., 1998) is utilised to assess regional variability using crude screening parameters, and provides the input to mass balance modelling of petroleum expulsion efficiency. The link between kerogen composition/abundance and adsorptive capacity for gas and oil is then explored using thermal analytical methods.

2.3 Databases, samples and methods

2.3.1 Regional geochemistry database

The foundation of this study is a comprehensive data set provided by AkerBP ASA containing Rock-Eval data and well information for 23,300 samples from various potential source rock formations across the Norwegian Continental Shelf. Samples with obvious mud-staining or migrated oil contents comprising S1/TOC > 200 mg/g TOC at early mature levels as well as those containing less than 1 wt.- % TOC content were excluded in order to minimize the impact of mineral matrix effects (Espitalie et al., 1980; Horsfield et al., 1983; Espitalié et al., 1984; Dembicki Jr., 1992). Based on the filtered dataset, 40 wells containing the Mandal Formation (ca. 222 samples) were analysed ranging from the Coffee Soil Fault complex in the east and northeast (Jæren High and Sørvestlandet High) to the Mid North Sea High marked by the political border of U.K.-Norway in the west, and the southern transition to the Danish sector with the Tail End and Gertrud Grabens. Regional data trends of the Mandal Formation, related to organic richness, organic matter quality and maturation characteristics were extracted from the database. Parameters related to expulsion are presented. Mass balance calculations (cf. Cooles et al., 1986) were made to calculate maximum petroleum expulsion efficiency taking only Rock-Eval S1 into account as retained petroleum (Fig. 2.2, upper part).

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2.3.2 Choosing samples for compositional analysis

25 cutting and 3 core samples were chosen for compositional analysis based on their Hydrogen Index and TOC values relative to the Mandal Formation dataset as a whole, and taking compositional evolution throughout maturation fully into account (Tissot et al., 1987; Horsfield and Rullkötter, 1994). Samples were selected from different wells and depth intervals of the northern and southern part of the Central Graben as well as from the Søgne Basin. The sample set was subdivided in maturity levels based on Rock-Eval Tmax from 400 – 431 °C, 432 – 438 °C, 439 – 445 °C and >445 °C, each representing a range of HI. According to the frequency distribution of HI values at the given maturity levels samples have been selected representing mean HI, together with the upper and lower 10 % probability of the normal distribution.

2.3.3 Sample preparation and screening

After collecting 28 source Mass Database: whole rock samples balance rock samples at the core storage calculation Sample collection at NPD facility of the Norwegian Solvent whole rock extracted rock Petroleum Directorate (NPD),

• Rock-Eval: • Rock-Eval: S1ex, samples were washed, picked S2 , T , TOC S1wr, S2wr, ex max,ex ex • for cavings and powdered Tmax,wr, TOCwr Open pyrolysis • Bulk kinetics following standard NIGOGA

procedures (Weiss et al., 2000). • Conventional parameters: TOC content and Rock-Eval S1wr: Volatile oil o pyrolysis parameters were o S2extracted = S2kerogen

• Unconventional, novel parameters: measured by APT AS, Norway o S2bitumen = S2wr – S2ex: involatile oil eqn. (1) using a Rock-Eval 6 instrument. Total oil = S1wr + S2bitumen eqn. (2) o Jet Rock-1 was run as a o Petroleum Quality = S1wr / Total Oil eqn. (3)

• Novel Mass balance calculation standard and checked against the acceptable range given in Fig. 2.2. Analytical workflow and overview about conventional and novel parameters that will NIGOGA. Aliquots of both the subsequently be used in the text. Blue boxes represent original and solvent extracted sample types; the bullet points indicate analyses and parameters conducted on and obtained from the distinct rock (24 h Soxhlet, using 99 % samples. dichloromethane+1 % methanol)

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY were analysed in order to distinguish how S1, S2, Tmax and TOC are relatable to kerogen and bitumen contributions. The parameters are given in Fig. 2.2.

2.3.4 Thermovaporisation-gas chromatography

Thermovaporisation-gas chromatography was used to measure the amount and composition of free hydrocarbons present in the source rock. A glass tube of approximately 40 µl volume, with a bore of 1.0 mm a flexure of 120° and sealed at one end is loaded to the elbow with pre-cleaned quartz sand, followed by approximately 10 mg of the crushed sample material, more glass beads and finally a glass wool plug. The tube was closed by a hydrogen flame. The analytical system was comprised of a purpose-built sample holder, a programmable pyrolysis furnace, a heated on-off split and a cryogenic trap, interfaced via a heated transfer line (all comprising the Quantum MSSV-2 Thermal Analysis System) to an Agilent gas chromatograph equipped with flame ionization detection. After loading the glass tube into the liner and closing the system, the outer surface of the tube was thermally cleaned and the sample mobilized for 5 min at 300 °C. Starting the temperature program and cracking the glass tube releases free hydrocarbons within the sample to the cryogenic trap that is pre-cooled to -178 °C. Within this closed, oxygen-free system the sample material is transported by a continuous helium flow (30 mL/min). After removing the cooling system of the trap after 10 min, the trap is ballistically heated to 300 °C and the sample material was released to the capillary column (50 m x 0.32 mm capillary column J&W Scientific HP-Ultra 1 [Dimethylpolysiloxan-phase], 0.52 µm film thickness). Products were detected using a flame ionization detector (FID) and displayed as an electric current response in pA. n-Butane was used as an external standard. Response factors for all compounds were assumed the same, except for methane whose response factor was 1.1.

2.3.5 Pyrolysis-gas chromatography

Non-isothermal open-system pyrolysis gas chromatography was utilised for characterizing the macromolecular structure of the kerogen using the parameters of Horsfield (1989), Larter (1984) and Eglinton et al. (1990). The Quantum MSSV-2 Thermal Analysis System® was utilised, analysing the products released over the

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY temperature range 300 – 600 °C (40 K/min) using the same gas chromatographic conditions as described above for thermovaporisation.

2.3.6 Bulk kinetics

Bulk pyrolysis was performed on 2 immature samples located in a marginal and a more central basin depositional setting of the Central Graben at heating rates between 0.7 and 15 K/min (0.7, 2, 5, 15 K/min) using a modified Source Rock Analyzer© (Humble). Products were transported to the FID in a constant helium flow of 50 ml/min. The discrete activation energy (EA) distribution optimisation with a single, variable frequency factor (A) was performed using the KINETICS 2000 and KMOD® programs (Burnham et al., 1987). For reliable timing predictions slow heating rates (≤5 K/min) were chosen to avoid heat transfer problems (Schenk and Dieckmann, 2004), and the chosen range in heating rates is broad enough to ensure correct iteration of the mathematical model and calculation of frequency factors.

2.4 Bulk source rock characteristics

The interpretation of geochemical screening data for the 222 samples selected from the Mandal Formation was undertaken with reference to the guidelines published by Peters and Cassa (1994) for Rock-Eval pyrolysis, and using the maturity cut-offs for the North Sea (Table 2.1) of Spencer et al. (1986b) and Cornford (1994).

Table 2.1. Levels of thermal maturity in the Central Graben The Mandal (Spencer et al., 1986b; Cornford, 1994). Formation is currently buried to depths ranging levels of thermal maturity RO depth

Spencer et al. (1986) Cornford (1994) % m between 2500 and 3500 m oil generation threshold immature 0.55 3350 in the Søgne Basin, and oil expulsion threshold early mature 0.60 3550 2900 m to 4870 m in the peak oil generation peak oil 0.80 4000 Central Graben (Fig. 2.3a). cracking of retained oil late mature 0.90 4200 Organic petrography has gas generation early (wet) gas 1.10 4500 main (dry) gas 1.30 ? previously shown that the onset of maturity starts

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY within the depth range of 3350 and 3550 m, and that of peak generation at ~4000 m (Spencer et al., 1986b; Cornford, 1994; Petersen et al., 2011). The deepest buried sections can be found in the Feda Graben and the adjacent southern parts of the Hidra High (Blocks 1/6 and 2/4 to 2/12) where burial of the Mandal Formation is consistently deeper than 4500 m. Other parts of the Feda Graben are less deeply buried (3500 – 4500 m). Cooper and Barnard (1984) have shown that those Central Graben areas are within the principal oil zone to gas and condensate zone (Spore Colouration Index = 5.5 to >8). But within the Søgne Basin no active petroleum system sourced by the Upper Jurassic Mandal Formation has been identified to date (Petersen et al., 1998; Petersen et al., 2011; Petersen et al., 2013).

Previous workers have identified average TOC contents of 5 – 6 % (Leonard and Munns, 1987), or 8 % with a maximum of 15 % TOC (Cayley, 1987) which is in accordance with values above 5 % TOC in the Danish Tail End Graben (e.g. Damtoft et al., 1987). Maps of present day source richness were provided by Demaison et al. (1984) and Cooper and Barnard (1984) without considering maturity effects. Immature sections of this source rock at depths shallower than 2 km have average HI values of 500 – 600 mg HC/g TOC and possess a gross hydrocarbon potential of 30.25 mg HC/g rock. In contrast, HI values can drop below 100 mg HC/g TOC at depths of ~5 km at maximum burial (Cornford, 1994). Rock- Eval and TOC data from the provided comprehensive database fit with descriptions of previous authors, but it has to be noted that source rock characteristics are far more heterogeneous than previously reported.

In general, the organic-rich mudstones of the Mandal Formation possess a very high potential for generating oil and gas, having TOC contents of up to 12 % and HI values as high as 647 mg HC/g TOC (Fig. 2.3a and b), both decreasing as a function of depth (Fig. 2.3a). TOC values are right-skewed with one mode at 1 – 3 % and a second mode at 5 – 6 % (Fig. 2.3c). It is noteworthy that TOC values above 6 % do not occur in the Søgne Basin. Within any given depth or maturity interval HI or S2 values occur across a broad range, though enhanced Hydrogen Indices are tied to elevated TOC contents. Importantly, this is not tied to specific geographic areas or sub-basins, but rather to the stratigraphic position within a well. Specifically, for most of the wells, HI continuously increases towards the top of the Mandal. This phenomenon occurs for all maturity levels and could be related to increasingly anoxic conditions in the water column caused by Ryazanian

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY transgression (Hamar et al., 1983; Rossland et al., 2013). Deposition and preservation of different types of organic matter are governed by sea level changes directly controlling sedimentation, nutrient supply, primary production and microbial metabolism (Arthur and Sageman, 2005). Regression of the shoreline shrinks the basin, enhances anoxic conditions, lowers dilution by terrigenous, clastic materials, and focusses organic matter transport to the basin centre. Lower horizons are characterized by maximum HI values of 367 mg HC/g TOC in the Central Graben and 315 mg HC/g TOC in the Søgne Basin. Upper horizons show

Hydrogen (mg HC /gIndex TOC) (a) HydrogenHydrogen Index Index (mg HC/g TOC) (b) HI 0 100 200 300 400 500 600 700 100 150 200 250 300 2500 4015 Tmax (°C) Kerogen Type 400 - 432 II well 2/1-9 432 - 439 II / III 439 - 445 445 - 450 3000 450 - 470 4020 470 - 600

II/III II ) m (

3500 4025

) m Depth (m) (

Depth Depth (m) 4000 4030 Depth

4500 4035

900 (e) 5000 0.5% Ro 800 T ype I

40 700 700 Tmax (°C) (c) Poor source 400 - 432 32 Fair source 432 - 439 Good source 439 - 445

Very good source 445 - 450 600 600 24 450 - 470 470 - 600

500 500

Frequency 16

8 Type II

400 400 (mg HC/g TOC) 0 Ula Trend 0 1 2 3 4 5 6 7 8 9 10 11 12 HI

TOC (%) TOC (%) 300 300 40 Tmax (°C) (d) 400 - 432 32 432 - 439

200 200 439 - 445 Region 24 445 - 450 SW' extent 450 - 470 Søgne Basin 1.3% Ro 16 470 - 600 Frequency

100 100 8 Type III

0 0 0 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 3 2 1 0 410 420 430 440 450 460 470 HydrogenHydrogen Index Index (mg (mg S2/gHC TOC)/g TOC) Tmax ( °C) Fig. 2.3. Bulk chemical characterisation of the Mandal Formation in the Central Graben using Rock-Eval pyrolysis obtained from a comprehensive geochemical data base of the Norwegian Continental Shelf including well info, stratigraphy and Rock-Eval parameter. Red circles indicate samples from the Søgne Basin, the blue line discriminates between kerogen types. The orange line is HI depletion trend of Northern areas, purple line represents same for southern area. The attached HI histogram to the HI-Tmax plot represents the distribution of immature source rock strata of the Mandal Formation in different structural regions.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY potentials of 513 and 573 mg HC/g TOC for the respective sub-basins. It is otherwise only at some marginal locations (wells 2/6-3, 7/8-3 and 2/2-1) where higher values occur in the lower strata. This is possibly an effect of locally increased terrigenous input before oxygenated, open-marine conditions had been established in the Lower Cretaceous (Rawson and Riley, 1982). Interestingly, an overall increase in HI values accompanied by source quality variation occurs towards the top of the Mandal in wells 2/11-1, 2/1-9, 2/1-6, 1/3-3 and 7/11-7 (Fig. 2.3b). This could indicate mass transportation of terrigenous material from basin margins into the basin centre, probably tectonically controlled (Rossland et al., 2013). In the northern sector no immature samples are present and the initial potential in the southwest can only be assumed or deduced. In that regard, it should be noted that maximum HI values in the northern Central Graben (Ula Trend, Oselvar, Cod and Brynhild fields) are lower by ~120 mg HC/g TOC than in the southwestern Feda Graben although ranging in the same maturity level (Fig. 2.3a).

Immature sections of the Mandal Formation (Tmax < 431 °C, e.g. Cornford, 1994) possess a very wide range in source richness of 126 – 624 mg HC/g TOC, and indeed this has been documented here in the less deeply buried Søgne Basin (wells 3/7-3, 3/5-1, 3/5-2, 2/3-3, 2/2-1), near the Mandal High (well 2/6-3), and along the inverted Lindesnes Ridge (well 2/11-1) in the Feda Graben. HI values show a very heterogeneous and bimodal distribution whose pattern is preserved in the mature Mandal sections (Fig. 2.3c and d). An interval displaying both diminished source quality (average HI = 273 mg HC/g TOC) and richness (average TOC = 2.7 %) is found in the lower Mandal of wells in the Søgne Basin (Blocks 3/5, 3/7 and 2/2-1). In contrast, a richer (average TOC = 7.1 %) and higher quality interval (average HI = 516 mg HC/g TOC) occurs in the upper horizons of the Mandal Formation in not only the Søgne Basin but also west of the Mandal High. The high quality interval is probably derived from algal sapropels (Cooper and Barnard, 1984) whereas the lower quality interval contains a mixed kerogen assemblage. Low TOC values and elevated Oxygen Indices imply that either a higher input of terrestrial organic matter or a more oxygenated water column could be responsible for this phenomenon. It must be noted that the Mandal Formation is early to peak mature over large swathes of its lateral extent, and therefore the interpretations derived from 20 immature out of 222 samples must be treated with caution. This is equally

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY true for extrapolated dead carbon contents that are 1.25 % for the Søgne Basin and 2.15 % for the Central Graben areas, reflecting an increased portion of inertinite which is typical for organic-lean shales. The facies map of Cooper and Barnard (1984) indicates waxy sapropel-containing (sub-group Type IIa) kerogen to be originally deposited in the presently less deeply buried basin margins (proximal areas of the Mandal High and northern parts of the Norwegian Central Graben) and in the Søgne Basin. General maturity trends can be revealed from the HI vs.

Tmax plot (Fig. 2.3e) indicating that the original potential of organic matter-rich sediments deposited in the Feda Graben might have been 500 – 700 and in the Søgne Basin 550 – 600 mg HC/g TOC. Higher initial source rock richness was assigned to the Feda Graben by Cooper and Barnard (1984) with algal Type Ia sapropels.

Early mature source rock intervals at depths down to 3400 m contain up to 12 % TOC and HI values up to 588 mg HC/g TOC. Importantly, average values of 128 (lower horizons) and 304 mg HC/g TOC (upper horizons) for early mature intervals are far below the optimum levels of contiguous source rock quality, and are consistent with a rather heterogeneous source formation. TOC contents are right-skewed and broadly distributed with mean values of 4.3 % for early mature sections. It is in the depth range 3400 – 3550 m that peak petroleum generation begins. HI values decrease continuously and nearly linearly at a rate of around 35 – 40 mg HC/g TOC per 100 m in the southwest and ~30 mg HC/g TOC per 100 m in the northern area. The average HI is 176 mg HC/g TOC, and TOC averages 2.7 %.

Interestingly, some peak mature source rock packages (445 – 453 °C Tmax) at marginal areas as the Gyda field or well 2/6-1 (3800 – 4200 m) still possess an elevated residual potential of 167 – 267 mg HC/g TOC and contain TOC contents ranging between 1 and 6 % of which 1/3 (12 samples) is above 4 %. This is a compositionally distinct and geographically confined source subset whose kerogen has a high paraffinicity and thermal stability (see later). Finally, only eight samples from the deepest well 1/6-6 possess TOC contents of 5 – 6 % and very high Production Indices (PI = S1/[S1 + S2]) of 0.46 – 0.64, and have HI values below 100 mg HC/g TOC, all of which points to late oil window phasing to gas-condensate potential.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

900 50 (a) 0.5% Ro (b) 800 HI (mg/gTOC) T ype I 900 700 700 40 500 300 100 600

(mg/g TOC) (mg/g 30 500

Type II (mg/g) 400 S2 (mg/g) S2

S2 20 Kerogen Type Kerogen Type II 300 II II/III II/III Tmax (°C) 200 S2 (mg/g) 400 - 432 0.0 - 2.5 10 432 - 439 Hydrogen Index Index Hydrogen 2.5 - 5.0 1.3% Ro 439 - 445 100 5.0 - 10.0 445 - 450 10.0 - 20.0 Type III 450 - 470 20.0 - 100.0 470 - 600 0 0 410 420 430 440 450 460 470 0 2 4 6 8 10 12 Tmax (°C) TOCTOC (%)(%) Bulk chemical characterisation of selected samples of the Mandal Formation using Fig. 2 .4. Rock-Eval pyrolysis of solvent-extracted samples representing the true nature of organic matter. Although samples represent a medium and lower quality interval of the Mandal Formation, the natural maturity series of the upper quality interval follows the kerogen degradation of a Type II kerogen (after Cornford et al., 1998).

Collected samples (see chapter 2.3.2) fit within the general trend of the Mandal Formation although representing lower HI values throughout the entire maturity trend (Fig. 2.4). Solvent extraction of whole rock samples reveals the true nature of organic matter uninfluenced by involatile bitumen, thus showing slightly lower generation potentials and TOC contents.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2.5 Petroleum-type organofacies

0 Solvent extracted Mandal (a) samples (Tmax range = 429 – 449 °C) were analysed by pyrolysis-gas chromatography in order to deduce the types of petroleums they are likely to generate in nature. Based on the chain length distribution diagram of Horsfield (1989) which

20 correlates pyrolysate compositions

0 with petroleum types, all samples of (b) the Mandal, irrespective of whether they are Type II or Type II/III, fall in the Paraffinic-Naphthenic-Aromatic (PNA) Low Wax Oil field (Horsfield, 1997). Thus, the samples with lower generative potential (Type II/III)

0 exhibit the same source quality as those with higher generative Fig. 2.5. Bulk compositional characterisation of potential (Type II). There is some selected, extracted Mandal Formation samples obtained by open pyrolysis-GC-FID and using variability in gas contribution (Fig. ternary plots of (a) the chain length distribution 2.5a), but the key point is that (Horsfield 1989, 1997) and (b) kerogen typing (Eglinton et al., 1990). petroleum type remains the same. This finding is consistent with the samples having similar organic input, but with variability in redox conditions, rather than a difference in the input of aquatic versus terrigenous having occurred. The Mandal Type II / Low wax P-N-A signature is the same as reported for the U.K. Kimmeridge Clay Formation (Horsfield, 1989; di Primio and Horsfield, 2006; Horsfield and di Primio, 2014). The position within the P-N-A field, and within the diagram of Eglinton et al. (1990) is to some degree governed by maturity. Early to peak mature organic matter (439 – 445 °C Tmax) is characterized by lower aromaticity and a slightly higher contribution of the mid-chain length range

(n-C6-14) than earliest mature sections. Oil-window mature organic matter of both kerogen types always contains shorter n-alkyl chains but interestingly the same

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY aromaticity as immature or early mature intervals (Fig. 2.5b) although a depleted kerogen structure with elevated aromaticity could be expected for Type II kerogens (Behar and Vandenbroucke, 1987; Rullkötter and Michaelis, 1990).

2.6 Generation characteristics

The calculated activation energy (EA) distribution and frequency factors for two immature samples from the marginal and central basin locations of the Central Graben are illustrated in Fig. 2.6a; data are shown in Table 2.2. The activation energy distributions of both of the samples are narrow, starting with the modal EA, with the frequency distribution dropping towards higher activation energies, and possibly representing the upper end of a more or less bell-shaped, Gaussian-like distribution documented for many marine Type II kerogens (Burnham et al., 1987; Espitalié et al., 1988; Moretti and Deacon, 1995; Reynolds and Burnham, 1995; Wei et al., 1995; Dessort et al., 1997). The distribution might signal the kerogen to be of a quite homogeneous composition but is depleted in the most labile components.

Immature Mandal Formation samples are characterized by main activation energies of 53 and 54 kcal/mol and frequency factors A = 8.38 x 1013 and

14 -1 2.23 x 10 s . The marginal sample (well 2/6-3) possesses lower Emean by 1.3 kcal/mol and frequency factor as compared to the basin-central sample from well 2/11-7 (Table 2.2) resulting in being insignificantly less stable by 3 °C geological Tmax at an average geological heating rate of the Central Graben

(1 °C/Ma, Fig. 2.7). A broad EA distribution and low relative fraction of the main EA generally indicate a more heterogeneous organic matter composition found in the central basin locations of the Mandal Formation, but this is not necessarily tied to higher stability of organic matter as can be observed from kinetic variability of higher mature samples (Tmax > 432 °C).

Dominant activation energies of mature Mandal samples range between 51 and 55 kcal/mol (Fig. 2.6b). Two EA distributions can be identified in the sample set: either a broad distribution as described for immature samples resembling a heterogeneous kerogen or a quite narrow one resembling a more homogeneous kerogen (Braun et al., 1991). Sharper distributions are found in marginal basin positions whereas broader distributions with higher deviation from the main EA are deposited in basin-central areas. Type II/III organic matter shows slightly higher

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY activation energies within the same frequency factor range (Fig. 2.7 inset) resulting in higher kinetic stability. It is noteworthy that organic matter in the southern Søgne Basin (well 3/7-3) is comparably stable, possessing a peak generation temperature (geological Tmax) of 146.4 °C and 146.2 °C for upper and lower Mandal horizons. Due to steep transformation ratio curves that are typical of homogeneous organic matter, a difference of 6 °C to the average geological Tmax of comparable

(a)

(b)

well 2/11-7 well 7/7-2 well 2/10-2 A = 1.63x1013 s-1 A = 1.25x1013 s-1 A = 1.18x1013 s-1

Tmax = 436 °C Tmax = 438 °C Tmax = 441 °C Type II Type II HIreal = 246 HIreal = 323 HIreal = 149 central marginal central

Fraction of Reaction

well 3/7-3 well 7/11-5 well 1/3-3 14 -1 13 -1 13 -1 III A = 6.47x10 s A = 2.74x10 s A = 2.45x10 s /

Tmax = 436 °C Tmax = 439 °C Tmax = 442 °C

HIreal = 140 HIreal = 112 HIreal = 113 Type II Søgne Basin central central Fraction of Reaction

Activation Energy Ea (kcal/mol) Activation Energy Ea (kcal/mol) Activation Energy Ea (kcal/mol)

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY source rock intervals in the Central Graben results in delayed onset of petroleum generation in the Søgne Basin at the same geological temperatures. This feature might have contributed to the low exploration success in that basin.

) 1 - (s

Frequency factor A factor Frequency

mean EA (kcal/mol)

Table 2.2 lists previously published kinetic parameters for immature Kimmeridge Clay Formation (KCF) in the UK Central Graben and its stratigraphic equivalents to the Mandal Formation (Burnham et al., 1987; Espitalié et al., 1988; Moretti and Deacon, 1995; Reynolds and Burnham, 1995; Wei et al., 1995; Dessort et al., 1997; Erdmann, 1999; Keym et al., 2006; Neumann, 2007). The mean activation energies of the reported Upper Jurassic source rocks range between 48 and 60 kcal/mol with frequency factors ranging between 3.7 x 1012 and 2.7 x 1016 s-1.

EA distributions are usually bell-shaped but with variable breadth (Fig. 2.6b) following the Gaussian distribution representing different varieties of chemical bonds as typical for marine source rocks.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Table 2.2. Bulk kinetic parameters of most immature Mandal sections (marginal and central basin position) and comparison to literature data from Viking Graben and U.K. Central Graben.

Sample 2/6-3 2/11-7 KCF KIMR UK Mandal

Specifics slow rates slow rates Tegelaar & Noble, 1994 Braun et al., 1991 Neumann, 2007

A (sec-1) 8.38x1013 2.24x1014 1.56x1014 2.90x1014 1.37x1014 E (kcal/mol) % % % % % 45 0.02 0.10 46 0.30 47 0.15 0.10 1.45 48 0.60 3.98 49 1.93 50 1.20 2.41 51 3.00 4.10 52 5.79 53 69.60 36.44 19.90 8.32 54 7.79 46.70 29.34 27.50 17.19 55 18.70 28.40 22.40 23.80 22.64 56 17.60 5.94 12.50 15.16 57 3.57 1.35 3.29 4.30 7.46 58 5.07 1.24 5.00 3.85 59 0.28 0.04 1.73 60 0.94 61 1.30 1.70 0.71 62 0.55 63 0.39 64 0.31 65 0.29 0.42 0.24 66 0.24 67 0.16 68 0.16 69 0.16 70 0.08 71 0.08

mean EA 53.61 54.91 54.21 54.52 54.52 HI 423 383 525 - - TOC 3.91 8.05 8.00 - -

Tmax 432 430 423 - -

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2.7 Petroleum expulsion

2.7.1 Hydrocarbon retention profiles

S1/TOC and S1/S2 data reveal that the Mandal Formation retains higher amounts of in-situ generated petroleum than do other Jurassic marine source rocks along the Norwegian Continental Shelf (Fig. 2.8). Only the directly underlying Farsund Formation retains more petroleum (Fig. 2.8a), arguably due to sealing by the Mandal Formation. Changes in maximum S1 yields correlate well with depth intervals of decreasing Hydrogen Indices in the different sub-basins and can be observed in depth intervals assigned by previous studies with beginning oil generation and cracking (Table 2.1). In the depth range between 3560 and 4200 m, where expulsion is expected (Table 2.1), maximum retained hydrocarbon yields remain stable indicating that the source rock is still saturated with generated products until late maturity commences (below 4500 m). In the Feda Graben major reservoirs have been charged vertically (Demaison and Huizinga, 1991; Cornford, 1994), thus initially generated yields must have been substantially higher. The slight decrease in maximum S1 yields (from 4200 to 4500 m) at this stage indicates only low expulsion efficiency. Noticeable is the southern Ula Trend (Fig. 2.1) containing significant petroleum volumes in thick, highly mature source rocks with high residual generation potential (3900 – 4200 m depth, black circles in Fig. 2.3e and Fig. 2.8b and c). In the shallow buried Søgne Basin (<3500 m) retained product yields are generally very low, not exceeding 2 mg/g rock.

Pepper and Corvi (1995) illustrated a strong dependency of volatile S1 yields and TOC (Fig. 2.8b and Fig. 2.9). The gradients have been referred to as either Bitumen Index BI (Killops et al., 1998; Sykes and Snowdon, 2002) or Oil Saturation Index (OSI, Jarvie, 2012), and are not higher than 106 mg S1/g TOC (UK KCF) in pre-expulsive sections (Pepper and Corvi, 1995). For the Mandal Formation, this North Sea-wide retention threshold is locally exceeded in the Feda Graben. While there is a positive correlation with maturity, significant overlap of OSI values occurs for any given maturity stage, independent of the samples lateral origin in different sub-basins. All samples show a very broad range in oil saturation OSI with early and peak mature source rocks scattering between 0 and 200 mg HC/g TOC and increasing as a function of depth (Fig. 2.8c), which shows

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY that up to 20 % of the organic matter of the source rock is saturated with petroleum

(Fig. 2.8b). At higher maturity stages (Tmax >448 °C; depth >4500 m) the source expels generated petroleum more efficiently, accompanied by a decrease in OSI. This is manifested by a switch to lower gradients on the S1/TOC plot. Interestingly, lower gradients at peak oil window maturity are neither observed in whole rock nor in extracted rock samples (Fig. 2.9a and b). In these depth ranges, mainly gaseous hydrocarbons, which are more easily expelled than oil (Stainforth and Reinders, 1990), have been formed within the source rock by either cracking from retained petroleum (>4200 m burial depth) or as a primary product from kerogen breakdown (>4500 m). Similar thermal stabilities of Mandal Type II and II/III organic

(a) S1whole rock (mg/g) 12 0 5 10 15 30% 2000 (b) 20% Mandal Fm Kerogen Type Farsund Fm II Oil crossover 10% Draupne Fm II / III 10 2500 Heather Fm Tmax (°C) Drake Fm

400 - 432 432 - 439 439 - 445 8 3000 445 - 450

450 - 470

470 - 600 (mg/g) 0.68 %RO (m)

3500 6

0.73 %RO whole rock whole

Depth Depth 4000 4 0.88 S1 Ula Trend 4500 2 0.53 %RO

5000 1.45 %RO Søgne Basin 0 0 2 4 6 8 10 12 5500 TOCwhole rock (%) (c) OSI (mg/g TOC) 0 50 100 150 200 12 2000 (d)

2500 10

3000 8

(mg/g)

(m) 3500 0.68 %RO

4000 whole rock whole Depth Depth 0.73 %RO 4

Ula Trend S1 4500 0.53 %RO 2 0.88 5000 1.45

0 Søgne Basin 5500 0 10 20 30 40 50 60 70 S2whole rock (mg/g) Fig. 2.8. Retention characteristics of Mandal Formation for bulk petroleum using Rock-Eval parameter of whole rock samples provided by the geochemical database. Coloured lines in depth plots represent maximum retained yields in marine source rocks of the Norwegian North Sea, excluding outliers in single horizons (<5 % of total number of samples, similar to as shown for Mandal Formation). Transparent circles represent generation and expulsion characteristics of Posidonia Shale (Germany) at different maturity stages. Dotted circles allow correlation of interesting samples between plots. In S1/TOC plot, Jarvie (2012)’s “oil cross over effect” with OSI values >10 % (>100 mg/g TOC) is indicated.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY matter and KCF sections from various sub-basins in the North Sea (Fig. 2.6a) suggest that depth ranges given in Table 2.1 can be applied to the Central Graben.

A strong correlation between retained hydrocarbon yields (S1) and the amount of labile kerogen (S2) has previously been shown to occur for the Barnett and Posidonia Shales, and to be better than the correlation between S1 and TOC (Mahlstedt and Horsfield, 2013; Han et al., 2015; Mahlstedt and Horsfield, submitted). In a paper on the Posidonia Shale by Mahlstedt and Horsfield

(submitted) it was revealed that S1/S2 ratios were initially low (beginning RO = ca.

0.5 %), but that at ca. RO = 0.7 % a new steeper S1/S2 gradient was observable, that remains constant but shifts towards lower S2 potentials during progressive maturation until at ca. RO = 1.45 % only low S1 and S2 values are observed. These

Whole rock Extracted rock

10 20 % 10 20 % (a) 10 % (b) 10 %

8 8

0.68 %RO

6 6 (mg/g) (mg/g)

S1 (mg/g) S1 0.73 %RO (mg/g) S1 4 4 whole rock whole Kerogen Type rock whole Kerogen Type 0.88 %RO II II S1 II/III S1 II/III 0.53 %RO Tmax (°C) Tmax (°C) 2 400 - 432 2 400 - 432 432 - 439 432 - 439 439 - 445 439 - 445 1.45 %RO 445 - 450 445 - 450 450 - 470 450 - 470 470 - 600 470 - 600 0 0 0 2 4 6 8 10 12 0 2 4 6 8 10 12 TOC (%) TOCwhole rock (%) TOCTOCextracted (%) (%)

10 10 (c) (d)

8 8

0.68 %R 6 O 6 (mg/g) (mg/g)

S1 (mg/g) S1 S1 (mg/g) S1 4 4 0.73 %RO whole rock whole whole rock whole Kerogen Type Kerogen Type 0.88 %RO II II S1 S1 II/III II/III Tmax (°C) Tmax (°C) 2 400 - 432 2 400 - 432 432 - 439 432 - 439 439 - 445 439 - 445 1.45 %RO 0.53 %RO 445 - 450 445 - 450 450 - 470 450 - 470 470 - 600 470 - 600 0 0 0 10 20 30 40 50 0 10 20 30 40 50 S2 (mg/g) S2whole rock (mg/g) S2extractedS2 (mg/g) (mg/g) Fig. 2.9. Bulk petroleum retention characteristics of selected samples of the Mandal Formation using Rock-Eval parameter of whole rock and extracted rock samples. For comparison, the retention series of the homogeneous and excellently expelling Posidonia Shale is plotted in shaded areas.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY changes are shown as shaded areas in Fig. 2.8d and Fig. 2.9c. Furthermore, the increasing contribution of aromatic components to S2 enhanced the retention of gases in the Posidonia Shale (Mahlstedt and Horsfield, submitted).

For the Mandal Formation absolute retained yields are significantly higher (see S1 values in Fig. 2.8a) and S1/S2 trends show the same development as a function of maturity but with a strong overlap between maturity levels (Fig. 2.8d and Fig. 2.9d). This overlap might be attributable to organofacies (Type II and II/III) variations. Intervals with lower source quality (Type II/III) possess steeper gradients at an immature stage (triangles in Fig. 2.9d), but after the beginning of oil generation average slopes are the same for high and low quality source intervals (light green and turquoise symbols). Slopes of Type II/III kerogen containing intervals are slightly shifted to lower S2 potentials indicating higher retention capacity of reactive carbon.

Here a comparison with Posidonia Shale is presented because the data necessary to make a direct comparison are available. The collected sample set contains significantly steeper gradients in S1 abundance than do the Posidonia Shale maturity series, and thus higher retention capacity of its organic matter is inferred. While unextracted whole rock samples do not show a well-developed maturity trend, the correlation is improved using extracted S2 yields (Fig. 2.9b and d). Both S1/TOC and S1/S2 cross plots show well-established maturity trends. The similarity of these plots showing extracted data can be attributed to the proportions of TOC being kerogen S2, a small portion of dead carbon and ultimately retained solvents from extraction (S1ex). Unlike the Posidonia Shale, Mandal-gradients on the S1/S2 and S1/TOC plots are increasing as function of maturity which is independent of kerogen type (see Type II vs. Type II/III symbols). Immature samples (Tmax = 400 – 432 °C) of the basin-central well 2/11-7 contain the overall highest absolute S1 yields and gradients similar to those of peak maturity (439 –

450 °C Tmax). Increasing retention capacity cannot be explained by the kerogen structure as the Mandal Formation’s reactive organic matter does not change significantly throughout catagenesis (Fig. 2.5) but rather with initial depositional environment. The appreciable changes between whole rock and extracted rock samples could indicate that the two fractions of S2, namely S2kerogen and S2bitumen, adsorb to different degrees.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2.7.2 Mass balance and total liquid petroleum profiles

Here we consider expulsion efficiency using the Cooles et al. (1986) model for a regional dataset built on whole rock analyses. Thereafter, a contrast is drawn between predictions made using selected whole rock versus solvent extracted samples, thus differentiating between bitumen versus kerogen contributions to the

S2 peak, and the volatile (S1) versus total oil (S1+[S2wr-S2ex]) petroleum in the models.

Cooles et al. (1986) defined two parameters, Petroleum Expulsion Efficiency PEE and Petroleum Generation Index PGI, as follows in which superscripts 0 and m define immature and mature samples:

( ) = = 0( 0 )� � eqn. (4) �� 1 +�2 �− �1 +�2 0 � 0 �� +��𝑐 �� 2 −�2 +�1 = = (0 � ) 0 eqn. (5) �� +��𝑐 �� 2 −�2 �+�1 0 0 �� 𝑝 �� 𝑖 �1 +�2

1.0 Source rocks which expel

efficiently are characterized by

0.8 rapidly increasing PEE as a function of PGI. This is illustrated for the

0.6 Northern German Posidonia Shale and Upper Jurassic Draupne 0.4 Kerogen Type II Formation in the Norwegian Viking

Petroleum Expulsion Efficiency Petroleum Expulsion II / III Tmax (°C) 0.2 400 - 432 Graben in Fig. 2.10. The Draupne 432 - 439 439 - 445 Petroleum Expulsion Efficiency Expulsion Petroleum 445 - 450 Formation is characterized by a 450 - 470 470 - 600 0.0 0.0 0.2 0.4 0.6 0.8 1.0 broad range of expulsion thresholds Petroleum Generation Index Petroleum Generation Index following the same evolutionary Fig. 2.10. Petroleum Expulsion Efficiency of the trends, of which the least efficient Upper Jurassic Mandal Formation as obtained from mass balance modelling using a modified expeller has a lower initial Hydrogen approach of Cooles et al. (1986). By means of Index of <400 mg HC/g TOC (Keym comparison, an area representing the trend of excellent expellers (Upper Jurassic Draupne and Dieckmann, 2006). Formation in the Norwegian North Sea and the Northern German Posidonia Shale) is plotted in grey.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Because the kerogen of the Mandal Formation is of variable starting composition, its assumed initial source quality in the model has been split into HI increments of 50 mg HC/g TOC, and kerogen degradation trends have been assigned using the bulk kinetic descriptions for two low maturity samples with higher (central position) and lower (marginal position) organic matter quality. While low expulsion thresholds and an exponential curve representing very good expulsion characteristics similar to the Draupne Formation were calculated for early mature levels, this is more likely to reflect the uncertainty in choice of initial composition than anything else. For higher maturity levels, a band of variable but overall low expulsion efficiency can be traced through the sample set. Higher expulsion efficiencies are linked to lower OSI values (=S1/TOC) (Fig. 2.8b, and shaded area in Fig. 2.10) and can be mainly be found for the Mandal on northern to eastern marginal terraces (light grey in Fig. 2.1).

In contrast to the models of Larter (1985) and Cooles et al. (1986), involatile as well as volatile oil needs to be taken into account when making the mass balance. The involatile portion, defined as S2bitumen, accounts for 24 – 63 % of the total S2 inventory as represented by S2ker/S2wr = 0.37 – 0.76 in Fig. 10b. Being composed of polar, high molecular weight components (Horsfield et al., 1991;

Jarvie, 2012) rich in heteroatoms, S2bitumen has a low mobility within the source rock. The involatile bitumen contribution is a function of kerogen type and maturity. At 439 – 445 °C Tmax both organic matter types contain the highest portions of bitumen that subsequently decrease towards peak oil window maturity (Fig. 2.11a). Throughout catagenesis the bitumen portion of S2 is subsequently increasing at the cost of reactive kerogen (Fig. 2.11b). Linked with a global data set of shale oil and gas plays (background dataset in Fig. 2.11), it can be deduced that the bitumen development of the Mandal Formation is similar to shale oil plays like the Niobrara, Eagle Ford, Wolfcamp (U.S.) and Vaca Muerta shales (Argentina). However, the excellently expelling Posidonia Shale possesses significantly higher initial kerogen proportions and S2bitumen does not exceed 30 % of S2whole rock throughout catagenesis.

Source rocks in a central basin position contain higher amounts of initial and newly generated involatile bitumen, whereas marginal sections of the northern and southern sector and the Søgne Basin intervals are dominated by the kerogen portion of the S2wr (Fig. 2.11). Total retained hydrocarbons (Total Oil = S1 +

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

20 1.0 Kerogen Type

II II/III (a) (b) Tmax (°C) 400 - 432 0.8

(mg/g) 432 - 439 15 439 - 445 445 - 450 450 - 470 470 - 600 0.6 whole rock whole

10 / S2

0.4 S1_bitumen (NA)S1_bitumen S2ex / S2wr (NA) itumen content content itumen Kerogen Type kerogen II

5 S2 II/III 0.2 S2 (mg/g) 0.0 - 2.5 2.5 - 5.0 5.0 - 10.0 Absolute b Absolute 10.0 - 20.0 20.0 - 100.0 0 0.0 0 10 20 30 40 50 425 435 445 455 465 475 S2kerogenS2 (mg/g) (mg/g) extracted Tmax (°C) extracted Tmax ( °C)

30 1.0 Kerogen Type

S2 (mg/g) 25 0.0 - 2.5 2.5 - 5.0 0.8 5.0 - 10.0 10.0 - 20.0 (mg/g) 0.7 20 20.0 - 100.0 bit 0.6

Oil) Total / S1 (= 15 0.5

= S1 + S2

Total OilTotal (mg/g) 0.4

Quality 10 Kerogen Type

S1 / (S1 + S2 - S2ex) (mg/g) - S2ex) + S2 (S1 / S1 0.3 II II/III 0.2 S2 (mg/g) 0.0 - 2.5

Total Oil 5 2.5 - 5.0 0.1 5.0 - 10.0

Petroleum Petroleum 10.0 - 20.0 20.0 - 100.0 0 0.0 425 435 445 455 465 475 425 435 445 455 465 475 extractedTmax_extracted Tmax (°C) ( °C) extractedTmax_extracted Tmax (°C) ( °C)

Fig. 2.11. Characterisation of bitumen portions generated by the Mandal Formation. Upper diagrams show absolute bitumen content denoted against extracted S2ker (a), and the proportion of reactive organic matter as part of the measured, unextracted S2wr yield (b). Lower diagrams illustrate properties of generated petroleum: (c) Total Oil yields = volatile hydrocarbons (S1) + involatile, heavy bitumen, and (c) determination of retained petroleum quality (= S1/Total Oil). Green circles indicate kerogen type II source rocks, red circles represent type II/III intervals. Gradients in (a) are colour-coded the same as maturity Tmax. The background data set comprises the Northern German Posidonia Shale, a typical marine source rock containing Type II organic matter, and unconventional U.S. oil and gas shale formations characterized as hybrid shale plays.

[S2whole rock – S2extracted rock]; Jarvie (2012)) describe the same trend as volatile products (S1; see Fig. 2.11c) as both fractions are miscible in each other, as indicated by an inherent mutual correlation with R² = 0.97 (see e.g. Han et al., 2015). The retained products show converging yields with increasing maturity, and plateau out at peak oil maturity (440 – 450 °C Tmax, Fig. 2.11c). Hereby, average retained Total Oil yields in Type II/III intervals are slightly lower (by ~5 mg/g rock, Fig. 2.11c) which can be attributed to lower S2 yields. With increasing maturity the

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY quality of the total retained petroleum (=S1/[S1 + S2bitumen]) improves in the lower

Tmax range, then appears to plateau out at 439 – 449 °C Tmax (Fig. 2.11d). Generally, the proportion of non-volatile retained products is decreasing with maturity and is relatively high compared to the excellently expelling Posidonia Shale but similar to worldwide oil shale plays.

As far as mass balance models are concerned, the involatile bitumen component is usually treated as “reactive carbon” (Cooles et al., 1986) but here we consider it as a reaction product. Fig. 2.12b integrates the S2bitumen portions into generated petroleum yields rather than considering it as reactive organic matter (Fig. 2.12a). Thus, initial expulsion thresholds are significantly increased and PEE is lowered at elevated maturity leading to later expulsion of lower total amounts of petroleum. The in-source cracking of retained products could of course increase PEE at very late maturity stages. Due to involatile bitumen acting as both a source and product, the true expulsion characteristics of the Mandal should lie in between those curves.

(a) Conventional (b) Modified using Total Oil

thresholds thresholds

Fig. 2 .12. Comparison of mass balance calculations conducted using (a) the conventional scheme modified from Cooles et al. (1986), and (b) a modification applying Total Oil (Han et al., 2015) and extracted S2 potential. Shift to lower PEE’s and higher PGI’s depends on S2bitumen yields. Saturation threshold until start of expulsion is significantly increased to 35 – 50 % conversion (PGI) for kerogen type II and II/III, respectively.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2.7.3 Linking retention to organic matter characteristics

2.7.3.1 Gas retention

Gas retention in hydrocarbon source rocks has long been attributed to adsorption on the organic and inorganic components in approximately equal portions (Krooss et al., 2002; Cheng and Huang, 2004; Ross and Bustin, 2009; Gasparik et al., 2012; Kuila et al., 2014; Rexer et al., 2014; Merkel et al., 2015). Micro- and meso-pores are the main sorptive site of HC gases (Gasparik et al., 2012; Rexer et al., 2014; Wang et al., 2016) wherein wet gases are preferentially sorbed (Ritter, 2003a; Cheng and Huang, 2004). In a study of Jurassic shales from western Canada, Ross and Bustin (2009) have suggested that the swelling of kerogen macerals could be important in retaining generated products (Sandvik et al., 1992; Ritter, 2003b; Kelemen et al., 2006b). Han et al. (2017)’s analysis of the Barnett and Posidonia Shales came to the same conclusion.

Han et al. (2015) demonstrated not only that petroleum retention in the Barnett Shale is primarily controlled by organic matter richness (TOC), but also that the “live” or “labile” component, rather than “dead” or “inert” carbon constitutes the most active adsorptive sites. Mahlstedt and Horsfield (2013) had earlier found a strong positive correlation between retention capacity (manometric method; Krooss et al. (2002)) and the yield of C1-C5 thermovaporization products normalised to S2 for the Posidonia Shale, whereas the TOC-normalised yield produced a weaker correlation. Here we use pyrolysis-gas chromatography to characterise the labile component in terms of aromatic versus aliphatic nature (aromaticity) and the progressive annelation of aromatics, (Giraud, 1970; Horsfield et al., 1983; Larter, 1984; Behar and Vandenbroucke, 1987; Horsfield, 1989) in order to draw correlations with gas concentrations measured in the samples.

In the Mandal Formation sorptive or retention capacity does change as a function of kerogen type (Fig. 2.13a). Gas sorption capacity increases as a function of the aromaticity of the labile portion of the macromolecular organic matter (Fig. 2.13b). A mixed Type II/III kerogen is able to store higher amounts of gaseous hydrocarbons at the same maturity stage than Type II kerogen, supposedly due to the higher proportion of aromatic structures within its labile component

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

(Vandenbroucke and Largeau, 2007). The correlation with genetic potential is significantly weaker (grey R² in Fig. 2.13b).

) 2.5 ) 2.5 Kerogen Type Kerogen Type ex ex (a) II II (b) II/III II/III Tmax (°C) Tmax (°C) 400 - 432 400 - 432 / g S2 2.0 R² = 0.82 / g S2

2.0 432 - 439 432 - 439 5 - 5

- 439 - 445 439 - 445 1

1 445 - 450 445 - 450 450 - 470 450 - 470 470 - 600 470 - 600 R² = 0.06 1.5 1.5

(mg Tvap C Tvap (mg (mg Tvap C Tvap (mg 1.0

1.0 Gas retention (mg/g S2er) Tvap Gas (mg/g S2ex) R² = 0.36

0.5 0.5 Gas retention Gas

Gas retention Gas 0.0 0.0 0 100 200 300 400 500 10 12 14 16 18 20 HydrogenHydrogen Index Index (mg/gTOC)(mg HC/g TOC) Total AromaticityAromaticity (%)

Cyclization and condensation reactions occur in both the bitumen and kerogen fractions during catagenesis (Larter and Douglas, 1980; Horsfield and Rullkötter, 1994; Horsfield, 1997; Poetz et al., 2014). Fig. 2.14 illustrates how the proportion of alkylbenzenes, alkylnaphthalenes and alkylphenanthrenes (left

column, Fig. 2.14a–c) and the C1-, C2- and C3-alkylbenzenes (right column, Fig. 2.14d–f), normalised to the GC-resolved fluid portion of the pyrolysate (mg/g

C6+ res), correlate to gas retention. The alkylbenzenes show the highest correlation with S2-normalised gas yields, pointing to aromatic structural elements in kerogen, releasable upon pyrolysis, as being most important in gas retention. The relative

concentration of alkylbenzenes shows the best correlation with relative Tvap C1-5 yields with an inherent structural relationship of R² = 0.81 (Fig. 2.14a). The significance of the correlation decreases as the number of fused rings increases. At this point it has to be mentioned that the method is less suitable for the alkylphenanthrenes because of lower signal to noise ratios and co-elution problems during column aging. The effect of the degree of alkylation on gas retention was investigated for the most significant compound class, the alkylbenzenes. The correlation between retained gas and the sum of styrene, o-, m- and p-xylenes is better than that with either toluene or the sum of methyl-, ethyl and propyl homologues; the correlation factor R² increases from medium to significant (Fig.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2.14d–f), before decreasing. M-, p- and o-xylenes together with styrene (C2- alkylbenzenes) give a slightly stronger correlation (R² = 0.85). These components represent mono-aromatic rings cross-linked by aliphatic chains of different length within the kerogen structure (e.g. Larter and Horsfield, 1993). β-Cleavage of extended alkyl chains attached to aromatic rings produces n-alkanes and alkylbenzenes on catagenesis or pyrolysis (Larter and Horsfield, 1993). The dimethyl moieties are one step towards a higher degree of aromatization or cyclization of the kerogen structure. As a consequence that the sum of C6+- normalised GC-resolved n-alkyl compounds remains more or less stable with increasing gas yields, it does not seem that aliphaticity of the kerogen structure controls gas sorption properties.

As shown in Fig. 2.13b, the general correlation of aromatic compounds and sorption capacity is linearly positive with different significance for individual compound groups of two kerogen types. However, when examining this correlation for individual kerogen types the significance coefficients R² drop significantly and no relationship between the observed parameter can be drawn. When evaluating the monoaromatics (Fig. 2.14) and C2-alkylbenzenes (Fig. 2.15) highest correlation coefficients reveal medium to low relationship for kerogen Types II and II/III, respectively, ranging between R² = 0.01 and 0.36. Having discriminated between kerogen types, the major alteration mechanism within individual kerogen types is thermal maturity generally changing the kerogen structure to more aromatic moieties. But as revealed by overall kerogen characteristics (Fig. 2.5 and Fig. 2.13b), the kerogen structure of the Mandal Formation does not change significantly within the early to peak mature oil window, and is expressed in the scatter of kerogen type.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY )

ex (a) (d) /g S2 5 - 1

R² = 0.81 R² = 0.51 (mg Tvap C Tvap (mg Gas retention Gas

mono-aromatics (mg/g C6+ res) monoaromatics with 1 side chain )

ex (b) (e) /g S2 5 - 1

R² = 0.43 R² = 0.76 (mg Tvap C Tvap (mg retention Gas Gas

di-aromatics (mg/g C6+ res) monoaromatics with 2 side chains

) (c) (f) ex /g S2

5 - 1

R² = 0.08 R² = 0.39 (mg Tvap C Tvap (mg

Gas retention Gas

tri-aromatics (mg/g C6+ res) monoaromatics with 3 side chains

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Interestingly, the sorption

)

ex capacity of kerogen Type II/III /g S2

decreases when some constituents, 5 - 1 such as polyaromatics, occur in a R² = 0.85 higher proportion in the kerogen (mg Tvap C Tvap (mg structure. This trend is the most pronounced for the relative presence of di- and triaromatic compounds

Gas retention Gas (Fig. 2.14b+c). With increasing degree of alkylation the negative C2 alkylbenzenes (mg/g C6+ res) slope with respect to absorptive Fig. 2.15. Cross plot of gas sorption capacity of capacity as well as the correlation reactive carbon versus aromatic compound group showing the best correlation. The correlation factor factor of this phenomena increases R² for both kerogen types (dark blue) is (R² > 0.39 for trimethyl- significantly higher than for individual kerogen types (grey). naphthalenes, not shown here). 3 to 5 samples containing kerogen type II/III fall out of the general trend between sorption capacity and aromaticity for total aromaticity but also individual aromatic compound groups and compounds (Fig. 2.13 and Fig. 2.14). These samples possess highest sorption capacity but have the same aromaticity level as other Type II/III samples and are the reason for partly negative linear correlations for Type II/III kerogens between both parameters. The three samples with highest sorption levels are the uppermost samples in the stratigraphic columns in different basin locations and are characterized by very high GR readings (140 – 200 gAPI). They might mark the beginning of the transition from restricted, anoxic to a fully marine depositional setting. Interestingly, a sample from the middle Mandal Formation in well 2/6-3 shows lower GR readings and lower electrical resistivity than its underlying three samples. This well is located very proximal to the paleo shore face of the Mandal High and has been influenced by the fluctuating shore face intercalating with inner shelf deposition (Rossland et al., 2013) introducing either terrestrial organic material to the marine environment or diluting the condensed shale deposition with siliciclastic input. Nevertheless, the vertical change in organic matter properties within this well is manifested in petrophysical property changes.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Such complex observations could be the result of different mechanisms controlling gas sorption in shales. Gas retention in TOC-rich shales is dominated by the organic matter component of the source rock (e.g. Cheng and Huang, 2004; Ross and Bustin, 2009; Kuila et al., 2014). But the above-described samples are characterized by comparably low labile kerogen portion (<2.5 mg/g). Gasparik et al. (2012) have shown that low-TOC, clay-rich shales can show excess gas retention. Due to the lack of organic matter, logically, this points to the gas being then dominantly adsorbed on the pore surface of the inorganic component (micro- and meso-porosity within/in between clay minerals, Cheng and Huang, 2004; Rexer et al., 2014; Merkel et al., 2015). Appreciable methane sorption capacities similar to those of TOC-rich shales have been observed on pure illite and montmorillonite (e.g. Ross and Bustin, 2009). However, the influence of mineralogy on retention capacities has not been tested in the course of this study.

2.7.3.2 Total oil retention

As introduced earlier, Han et al. (2015) showed that the amount of retained volatile oil is mainly a function of TOC and S2 in particular, for a single maturity stage (Fig. 2.8b+d and Fig. 2.9b+d). Our results from analyses conducted on extracted and unextracted Mandal Formation source rocks also show a very good correlation with both the S2extracted (kerogen portion) as well as S2whole rock (kerogen + bitumen). This indicates that the volatile portion of generated products is retained in both the kerogen and involatile bitumen. Together with an inherent correlation of Total Oil and S1 yields, R² = 0.94 (Han et al., 2015), this finding implies that volatile fractions may be partially dissolved in heavy bitumen which likely occurs in a viscous state in the rock, closely associated with kerogen (Wilhelms et al., 1990).

The portion of the conventional S2 which has been removed by solvent extraction, the S2bitumen, plays a dual role as generation product from kerogen (Fig. 2.11a) and as a retention site of volatile HC (Fig. 2.11c+d). The generation of volatiles as well as involatile bitumen from kerogen is evident from the development of Total Oil as a function of S2 showing the same gradients at different maturity levels as volatiles (Fig. 2.9d and Fig. 2.11a). The correlation coefficient is slightly better with S2whole rock (R² = 0.82) than for S2kerogen (R² = 0.66).

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

500 The bitumen portion of the

) conventional S2 is a product

400 kerogen dominated by high molecular weight compounds resulting in its highly 300 viscous nature. Thus, migration R² = 0.94 200 distances of the involatile bitumen Total Oil/S2exTotal (NA) portion away from its place of 100 generation are most likely not very (mg/g S2 (mg/g retention Total Oil long. Assuming that S2bitumen is not a 0 0.0 0.2 0.4 0.6 0.8 1.0 result of degradation of early Bitumen portionbitumen/S2wr of S2 (NA)whole rock (%) migrated oil into the source rock Fig. 2.16. Retention of generated petroleum fractions in kerogen as a function of bitumen (Sanei et al., 2015) since maturity is proportion of S2 indicating that bitumen is the too low for extensive cracking, then major retention site of products from the Mandal Formation. Worldwide dataset (see Fig. 2.11) we can further assume that Total Oil shows higher yields of retained products for low-S2 remains within or nearby kerogen. intervals of peak to late oil window which are thus dominated by the heavy bitumen portion. This can as well be taken from the inherent correlations of the generated fractions S1, S2bitumen and its sum Total Oil showing R² of 0.84 – 0.97; being part of the kerogen swelling theory (Kelemen et al., 2006b). Total Oil retention in kerogen shows a very good correlation with the bitumen portion of the conventional S2whole rock (R² = 0.94, Fig. 2.16) indicating that the bitumen is the major retention site of generated products. This correlation is more convincing than that with oil retention in S2whole rock (R² = 0.87) although it is relating two parameters to each other that both contain S2bitumen as a significant component. Interestingly, the quality of petroleum, being the proportion of volatiles in total generated petroleum, cannot be related to any amounts or proportions of the organic matter fractions determined from bulk Rock-Eval pyrolysis.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Although being a function of active carbon yields, neither the volatile nor the total petroleum retention in S2kerogen can be related to compositional parameters of the pyrolysates, aromaticity and aliphaticity (R² = 0.00 and 0.17, respectively, Fig. 2.17). However and unexpectedly, the correlation of volatiles as well as total oil retention to aliphaticity is less weak when excluding two outliers (from wells 1/3-3 and 2/1-9) as strong as R² = 0.57 for volatile and R² = 0.51 for Total Oil retention. In addition, petroleum quality seems to be a function of the source rock pyrolysate’s aliphaticity with different correlation factors for kerogen Types II and II/III (R² = 0.50 and 0.87, respectively; not shown here).

300 300

Kerogen Type Kerogen Type ) II II II/III II/III 250 S2 (mg/g) 250 S2 (mg/g) kerogen 0.0 - 2.5 0.0 - 2.5 2.5 - 5.0 2.5 - 5.0 5.0 - 10.0 5.0 - 10.0 200 10.0 - 20.0 200 10.0 - 20.0 20.0 - 100.0 20.0 - 100.0

(mg/g S2 (mg/g R² = 0.00 R² = 0.17 150 150

100 100 Total OilTotal retention (mg/g S2)

50 50 Total Oil retention retention Oil Total

0 0 10 11 12 13 14 15 16 17 18 19 20 18 20 22 24 26 28 30 32 34 36 AromaticityAromaticity (%) (%) AliphaticityAliphaticity (%) (%) Fig. 2.17. Correlation of Total Oil retention on reactive organic matter with compositional parameter of the kerogen from extracted Mandal Formation samples showing no or weak correlation but a distinct influence of S2 yields (see Fig. 2.11). When excluding the two outliers a “pseudo-correlation” with the kerogen’s aliphaticity (R² = 0.51) can be observed.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

2.8 Conclusions

The Mandal Formation is a very prolific source rock with up to 12 wt.- % TOC and remaining generation potentials between 647 and 36 mg HC/g TOC as a function of maturity and kerogen type. Original HI values are calculated to have ranged between 500 and 700 indicating hydrogen-rich Type II kerogen. Contributions of Type III can be found in lower stratigraphic horizons and more marginal basin locations. Fingerprints from source rock pyrolysis-gas chromatography place the kerogen within the PNA Low Wax oil generating facies with a trend from longer aliphatic chains towards short-chained aliphatics at each maturity stage. The kerogen structure is intermediate for Type II kerogens and more aromatic for horizons with Type II/III organic matter.

Retention characteristics calculated from detailed inverse mass balance modelling are highly variable throughout the basin with expulsion starting locally at 40 % generation. Higher expulsion efficiencies can be found at marginal areas of the Central Graben. Having shown that generated volatiles (S1) are retained within the Rock-Eval S2 portion of the source rock, solvent extraction reveals that

S2 consists of reactive kerogen (the true, extracted S2kerogen) and high molecular weight bitumen (S2bitumen). The latter is generated from kerogen during catagenesis and acts itself as source of volatile hydrocarbons as well as retention site for oil. Retention of the gaseous fraction of generated products is a function of aromaticity of the kerogen and reveals higher sorption capacities for kerogen Type II/III than for Type II. Mono-aromatic constituents, and especially alkylbenzenes with two attached aliphatic chains, seem to play an important role in the kerogen structure when considering its sorption characteristics. Although existing in a mono-phase system, different controls on retention infer separate migration and sorption mechanisms for gaseous and liquid components. Thermal stability of both organic matter types is similar to KCF in other North Sea sub-basins.

If the dual origin of S2bitumen is taken into account for mass balancing, the Mandal Formation’s Petroleum Generation Index is underestimated. The extent is a question of its proportions in the conventional S2, together with the generation potential and kinetic parameters of the involatile bitumen.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

In the Søgne Basin, a half-graben system with high lateral variability in thickness and source rock quality of marine shales, the Mandal Formation in its southern location occurs with slightly different kinetic properties as compared to average Central Graben characteristics: the Type II/III marine mudstones contain organic matter with HI values below 200 mg S2extracted/g TOC, mean activation energies of 55.5 kcal/mol and elevated frequency factors A of ~2.4 x 1014 s-1. Calculated geological maximum generation temperatures of 146.3 °C are higher by 6 °C to comparable samples in the Central Graben. This feature might have contributed to the low exploration success in the Søgne Basin.

2.9 Acknowledgement

This study is part of the Ph.D. thesis of Volker Ziegs at GFZ Potsdam as an Industry Partnership with Aker BP ASA. We are grateful for financial support and the permission to publish. Special thanks goes to Rolando di Primio (Lundin Norway, formerly GFZ) for initiation of the Ph.D. study as well as Ferdinand Perssen, Cornelia Karger and Anke Kaminsky (all GFZ Potsdam) for assistance with the lab work. The manuscript benefited substantially from excellent critical reviews by Alexander Hartwig (Aker BP ASA), Chris Cornford, Ray McBride (both IGI Ltd) and two anonymous reviewers.

2: PETROLEUM RETENTION IN THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

70 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

3 Unravelling maturity- and migration-related carbazole and phenol distributions in Central Graben crude oils

3.1 Abstract

In this contribution we present the results of an integrated investigation of selected nitrogen- and oxygen-bearing compounds in Norwegian Central Graben crude oils. We first provide an interpretation framework built on hydrocarbon biomarkers, and then use this framework to relate polar compound geochemistry to the influences of source facies (Farsund versus Mandal Formations), maturity, migration and reservoir lithology.

Oil maturity could be assessed using established changes in carbazole annelation (N1 DBE 9 vs. 12 vs. 15 classes), as well as hydrocarbon biomarkers. 29Ts/(29Ts+NH) correlated best with the polar compound maturity data. Secondary migration fractionation appears nevertheless to have played a role, as seen by increased DBE 9 and 12 carbazole and benzocarbazole proportions and a loss of C2-3 DBE 12 homologues within carbonate reservoirs as compared to intraformational Upper Jurassic siliciclastic reservoirs. Thus, migration distances, pathways and wettability of carrier systems ostensibly play a significant role in carbazole distributions of the Central Graben oils, manifesting itself as apparent maturity retardation. In an attempt to eliminate the migration component from maturity assessment, we here present a novel ternary diagram including dibenzocarbazoles (N1 DBE 15) and phenolic species (O1 DBE 4 and 5) based on a single measurement using the FT-ICR-MS. However, the integration of such results into 3D-modelling software must be conducted to clarify source kitchen, migration pathways and distances.

This chapter (post-print) is published as: Ziegs, V., Horsfield, B., Noah, M., Poetz, S., Hartwig, A., Rinna, J., Skeie, J.E. (2018). Unravelling maturity- and migration-related carbazole and phenol distributions in Central Graben crude oils. Marine and Petroleum Geology 94, 114- 130. DOI: 10.1016/j.marpetgeo.2018.03.039

71 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

3.2 Introduction

Petroleum consists of an exceedingly complex mixture of hydrocarbons and non-hydrocarbons, extending from methane to macromolecular aggregates such as asphaltenes. The relative proportions of these components are quite variable and depend initially on the type of organic matter in the parent source rock, its level of maturity at the time of expulsion, and subsequently upon secondary effects including phase behaviour (Silverman, 1965; England et al., 1987; England and Mackenzie, 1989; Larter and Mills, 1991; di Primio and Skeie, 2004), fractionation during expulsion (Han et al., 2015; Ziegs et al., 2017), biodegradation and water washing (Connan, 1984; e.g. Ahsan et al., 1998) and in-reservoir maturation (Gabrielsen et al., 1985; Horsfield et al., 1992). The mixing of charges from co- sources (e.g. Ohm et al., 2012) and the acquisition of organics along migration routes (Curiale and Bromley, 1996) also occur. Deciphering these various influences remains a key challenge in petroleum geochemistry. Recognising and predicting the occurrence of migration fractionation is an important element in petroleum system assessment, because the processes, strongly influenced by water-oil and rock-oil interactions (Li et al., 1994; Larter et al., 1996; Mann et al., 1997), can alter the physicochemical properties of crude oils appreciably. Being of low polarity, hydrocarbon biomarkers reflect source and maturity and are not suitable for tracing fractionation (Peters et al., 2005). Polar compounds on the other hand lend themselves well to this application.

Early attempts to integrate the alkylcarbazoles and alkylphenols into petroleum geochemistry protocols were made by the Newcastle Research Group (Li et al., 1994; Stoddart et al., 1995; Larter et al., 1996; Li et al., 1997) using gas- chromatography mass-spectrometry (GC-MS) data. The main concept was that migration-related fractionation is controlled by the surface activity of polar compound isomers, as determined by the local chemical environment of the functional group or polar atom. Shielding effects of fused/annelated rings or alkyl side chains and therefore, the molecular size and shape of individual molecules, influence polarity and thus have been proposed as maturity-independent measures of migration distance. They are affected by sorption onto clay minerals or into solid organic matter (Stoddart et al., 1995; e.g. Larter et al., 1996) or by oil-water- partitioning (van Duin and Larter, 2001). Li et al. (1994) and (1995) suggested that

72 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS migration fractionation has a more remarkable impact on pyrrolic nitrogen abundances and distributions than source environments. Varying lithology and geometry along the migration pathway affect distributions (Li et al., 1998). Nevertheless, the preferential removal of the rod-shaped [a] isomer relative to the sub-spherical [c] configuration, originally considered to be migration-driven (Larter et al., 1996), can be influenced by maturity (Li et al., 1997; Clegg et al., 1998a; Clegg et al., 1998b; Horsfield et al., 1998), and facies (Bakr and Wilkes, 2002) in vertically drained petroleum systems. This is also the case for the predominant enrichment of alkylcarbazoles over alkylbenzocarbazoles and generally an enrichment of more methylated over less methylated homologues. Clegg et al. (1997) compared relative abundances of alkylated carbazoles in the carbonate-rich

Keg River Formation and found that C4-5 carbazoles prevailed in the lower transgressive facies and C0-1 carbazoles dominate the upper regressive facies. Source and maturity variations can also have a strong impact on the hydrocarbon fraction, but not on the pyrrolic N fraction of crude oils in the Rainbow-Shekilie- Zama sub-basins, NW Alberta, Canada (Li et al., 1999). Due to short migration distances, strong correlations between oil maturity and carbazole distributions were shown for the marine carbonate Tithonian source rocks of the Sonda de Campeche (Clegg et al., 1998b; Horsfield et al., 1998).

Fourier Transform-Ion Cyclotron Resonance-Mass Spectrometry (FT-ICR-MS) has offered new insights into polar compound geochemistry (e.g. Hughey et al., 2002; Marshall and Rodgers, 2008). While it cannot identify stereochemical isomers, FT-ICR-MS offers the big advantage of extending the molecular weight range up to 1000 Da. In addition, it differentiates between the multitudes of N-, S- and O-compounds in complex mixtures according to aromaticity, annulation and aliphaticity. The ESI negative mode is able to ionize acidic ring structures with one or more nitrogen (N1 to N2), oxygen (O1 to Ox) and sulphur atoms (S1 to Sz) or combinations of all three (NyOx, NySz or SzOx classes). Double Bond Equivalents define the structure of these compounds by the number of aromatic and saturated rings plus unsaturated carbon bonds incorporated into the core structure and attached side chains (Oldenburg et al., 2014; Poetz et al., 2014), thus defining their aromaticity (Hughey et al., 2002; Hughey et al., 2004). The relative abundance of carbazoles, benzo- and dibenzocarbazoles in a North Sea Viking Graben oil series, as determined using Fourier Transform-Ion Cyclotron Resonance-Mass

73 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

Spectrometry, showed a strong dependence on maturity (FT-ICR-MS, Oldenburg et al., 2014). The term aliphaticity has been used to define the chain length distribution (short, intermediate, long) of the side chains attached to a given core structure (Mahlstedt et al., 2016), this being relatable to similar structural moieties in kerogens (cf. Muscio et al., 1991).

Here, we present the results of an integrated investigation of selected polar compounds of Central Graben crude oils, utilising FT-ICR-MS data. Importantly, we first provide an interpretational context for the petroleum system using hydrocarbon biomarkers, and then use this framework to relate polar compound geochemistry to the influences of source facies, maturity, migration fractionation and reservoir lithology. An assessment of expulsion efficiency and factors controlling adsorption has already been presented (Ziegs et al., 2017).

3.3 Geological setting & hydrocarbon habitat

The Central Graben is an intracratonic basin forming the southern branch of the North Sea triple rift complex (Fig. 3.1, inset) whose Norwegian part is bounded by the Mid-North Sea High to the southwest and the Coffee Soil Fault to the northeast (e.g. Gowers et al., 1993; Japsen et al., 2003; Møller and Rasmussen, 2003). The Mandal High in the southernmost part of the Norwegian sector split the main graben, having evolved as an extensional basin, from the adjacent Søgne Basin which developed as a half-graben (Rossland et al., 2013). Its present-day configuration results from failed rift development starting during Permian times and culminating during the Upper Jurassic (Ziegler, 1988; e.g. Glennie, 1990, 1990; Gowers et al., 1993), followed by a thermal sag phase from the Cretaceous up to the present day. The deposition of thick Zechstein salt successions during prolonged extension in Permian-Triassic times is followed by a Triassic sequence of continental red-bed mudstones, siltstones and minor sandstones. Lower Jurassic successions are largely missing due to pre-rift doming which represents the earliest stage of the main extensional phase (Ziegler, 1990; Underhill and Partington, 1993; Andsbjerg et al., 2001; Andsbjerg and Dybkjaer, 2003; Nielsen, 2003). Middle Jurassic shallow-marine shaly-coaly sandstones of the Lulu and Bryne Formations constitute the first horizons of the Jurassic extensional phase, followed by the deposition of thick, OM-rich, marine black shales in the Upper Jurassic/basal

74 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

Cretaceous. The opening of the Atlantic Ocean during the Early Cretaceous completely changed the regional stress regime causing the cessation of rifting and flooding of the basin with fresh water, marking the initiation of the post-rift thermal subsidence. The sequential deposition of Lower Cretaceous organic-lean, marine shales, Upper Cretaceous chalks and limestones, and thick Cenozoic shales successions, resulted in the progressive burial and thermal maturation of Upper Jurassic source rocks in the Central Graben. Present-day burial depths along the graben axis are ~4700 m and up to 3300 m along the flanks. Chalk sequences represent the major reservoir type in the Central Graben axis (Ziegler, 1990; Cornford, 1994), e.g. the Ekofisk, Eldfisk, and Valhall fields, which account for 72 % of the Central Graben’s initial petroleum in-place reserves (npd.no, Dec 2015). At marginal basin locations (light grey in Fig. 3.1), clean, shallow-marine sandstones of the Ula Formation were deposited during Middle and Upper Jurassic times constituting important reservoirs (dark grey in Fig. 3.1).

distal proximal

75 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

Upper Jurassic black shales are subdivided into the Haugesund, Farsund and Mandal Formations in the Norwegian part of the Central Graben. The Mandal Formation constitutes the principal source rock in the graben centre containing Type II algal marine OM. The Farsund Formation, a less prolific source rock which contains higher terrigenous input, has charged individual petroleum accumulations (e.g. 2/2-1, Gabrielsen et al., 1985), generally at marginal locations where the Mandal is thin or eroded. Although being a TOC-rich source rock, the Mandal Formation possesses a low efficiency in expelling its generated products which was attributed to enhanced generation of heavy, involatile and thus less movable petroleum fractions in which volatile products are partially retained (Ziegs et al., 2017). Different mechanisms for retention of oil and gas may as well affect the gas- oil-ratio of expelled products.

3.4 Samples & methods

3.4.1 Sample set

Crude oils were selected based on publically available screening data in the online database of the Norwegian Petroleum Directorate (NPD), including fluid test type, reservoir depth and stratigraphy, API gravity and GOR. A screening database provided by Aker BP ASA contained supplementary bulk compositional data. The oils obtained from drill stem tests (DST) were provided by the NPD.

Twenty-four DST oils from 20 well locations were selected as representing the API and likely maturity range typical for the Central Graben. Cornford (1998) identified the main mode of North Sea oils representing a full maturity spectrum at 38 °API. The sample choice included oils potentially sourced from the Mandal Formation that is the principal source rock in the Mandal-Ekofisk petroleum system (Cornford, 1994). Three oils which have been generated by source rocks with different organic matter types were selected as a compositional comparison to intra-Mandal facies variations. The 2/2-5 oil was generated from a source rock formed under locally restricted, hypersaline and highly anoxic depositional environment (Pedersen et al., 2006). The Farsund Formation, a marine shale with a higher input of terrigenous organic matter, has charged the 2/2-1 discovery (Gabrielsen et al., 1985). The Trym condensate play (well 3/7-4) was sourced by the

76 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS coaly Middle Jurassic Lulu Formation from the Danish part of the Søgne Basin and represents the only active petroleum system there (Petersen and Brekke, 2001). The sampled crude oils were collected from either Upper Cretaceous to Lower Palaeocene carbonate reservoirs or Devonian to Upper Jurassic (silici)clastic reservoirs.

Five petroleum provinces have been distinguished in the Norwegian Central Graben (NCG) based on literature (Cooper and Barnard, 1984; e.g. Gabrielsen et al., 1985; Hughes et al., 1985; Petersen and Brekke, 2001; Ohm et al., 2012): Chalk and limestone reservoir occur in (1) deep axial regions developed in the southwest of the NCG (Greater Ekofisk area) and as well (2) at the Steinbit Terrace. (3) Shallow-marine sandstone reservoirs are restricted to shallower, marginal areas in the north/northeast, called here the “Extended Ula Trend”, (4) In the Søgne Basin no active Upper Jurassic sourced petroleum system has been identified, despite sufficient source rock presence and maturity (Petersen et al., 2013; Ziegs et al., 2017). The Trym condensate (well 3/7-4), on the other hand, is charged from the marine-influenced, coaly Lulu Formation. The 2/6-5 discovery is located within the structural borders of the Søgne Basin but oil-source correlation is to-date unclear. (5) Although structurally located in the basin axis, the 2/7 oils of the siliciclastic Embla field (SW region) are distinguishable due to a mixed origin with minor input of Palaeozoic oils that were biodegraded before receiving fresh, undegraded petroleum from the Upper Jurassic Farsund Formation (Ohm et al., 2012). Furthermore but not linked to a source kitchen, the Mjølner oil field close to the Danish border has been produced from shoreline sandstones of the Eldfisk Formation which was deposited synchronously to the Haugesund Formation (Söderström et al., 1991).

3.4.2 Analytical methods

Geochemical data of conventional geochemical analyses on crude oils are provided by the project partner Aker BP ASA and measured by APT Norway AS. FT-ICR-MS measurements are conducted at GFZ.

For GC-FID on whole oils, an Agilent 7890 A instrument is used. The column is a HP PONA, length 50 m, i.d. 0.2 mm, film thickness 0.5 µm. 2,2,4-tri-methyl- pentane is used as an internal standard. The temperature programme is as follows:

77 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

30 °C (10 min. isothermal) - 2 °C/min. - 60 °C (10 min.) - 2 °C/min - 130 °C (0 min.) - 4 °C/min. - 320 °C (25 min.). The oils were fractionated into aliphatic, aromatic and polar fractions using an automated medium pressure liquid chromatography (MPLC) procedure (Radke et al., 1980). Sample weights of the individual fractions were used for the chemical gross composition. GC-MS was conducted on the saturate and aromatic faction, but only the former fraction was useful for interpretation. For GC-MS analysis of the saturate fraction, a Micromass ProSpec high resolution instrument is used. The instrument is tuned to a resolution of 3000 and data is acquired in Selected Ion Recording (SIR) mode. The column used is a 60 m CP-Sil-5 CB-MS with an i.d. of 0.25 mm and a film thickness 0.25 µm. δ4-27ααR is used as internal standard for quantification. The temperature programme is the following: 50 °C (1 min.) - 20 °C/min. - 120 °C - 2 °C/min - 320 °C (20 min.).

Mass analyses on the polar compounds were performed with a 12 T FT-ICR mass spectrometer in negative ion ESI mode equipped with an Apollo II ESI source both from Bruker Daltonik GmbH (Bremen, Germany). Details on the analytical procedure, mass calibration and subsequent data analysis were described in Mahlstedt et al. (2016). Consequently developing and adjusting the method to the needs of the samples, we have used stock solutions with a concentration of

1000 µg/ml MeOH:Tol [1:1, v:v] + 10 µl aqueous NH3 solution to facilitate deprotonation. The mass spectra, having been accumulated in a mass range from m/z 147 to 750 Da during 200 scans, were internally recalibrated using known homologous series of saturated fatty acids and carbazoles with one, two, and three additional fused aromatic rings. A quadratic calibration mode was chosen for all samples. Elemental formulas were assigned to the recalibrated m/z values with a maximal error of 0.5 ppm using S/N ratio of 12 and allowing 0−∞ C, 0 −∞ H, 0−8 O, 0−2 N, 0−2 S, and 0−1 Na atoms as well as 0−2 13C isotopes. The number of assigned signals, mostly above 80 % of all detected peaks, the mean molecular weight (normalized to # of signals, Mn, and molecular weight, Mw), and the relative monoisotopic ion abundances of major compound classes are listed in Table 3.2, whereby the subscripts denote the number of heteroatoms incorporated into individual chemical structures. Elemental classes, compound classes and individual compounds are given as Total Monoisotopic Ion Abundance (% TMIA) representing a relative abundance of individual compounds based on all detected compounds, excluding the 13C species.

78 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

3.5 Results & discussion

3.5.1 Oil and condensate properties using conventional parameters

3.5.1.1 Variations of API with depth

The sample set covers a wide range of API gravities from different regions of the Norwegian Central Graben encompassing 29 °API black oils (Krabbe discovery, 7/8-3) to 55 °API condensates of the Tommeliten Alpha field (1/9-1). Present-day bottom hole temperatures (BHT) of selected DSTs do not fall below 80 °C (Table 3.1), drastically reducing the likelihood of biodegradation as a compositional alteration factor.

Crude oils in sandstone reservoirs of the Ula, Bryne, Sandnes Formations or Rotliegend Group show API gravities that are increasing as a function of depth with a gradient of 2 °API/100 m. These results are similar to those of Justwan et al. (2006) and Barnard and Bastow (1991) for the southern Viking Graben (N, GB) showing a gradient of 3 ° and 6 °API/100 m, respectively. However, oil qualities in Upper Cretaceous to Lower Palaeocene chalk and limestone reservoirs (Ekofisk and Tor Formations) show a wide API gravity range independent of depth (Fig. 3.2) and reservoir temperature. Based on the general relationship between API gravity and maturity (e.g. Peters et al., 2005), and thus depth of generation and expulsion, crude oils in carbonate reservoirs must have experienced vertical and/or longer migration distances than those in clastic reservoirs which are, in most cases, in close juxtaposition to the source rocks.

The trends are likely related to facies and maturity of the source kitchens, though phase behaviour cannot be ruled out. Cretaceous and Tertiary horizons are lean in organic matter, and thus contamination during secondary migration (Curiale and Bromley, 1996) can be ruled out. Measured BHTs of sampled DSTs range between 80 and 165 °C and thus only oils in wells 1/3-3 and 2/7-22 are near the temperature range needed for in-reservoir thermal alteration (165 − 174 °C; Horsfield et al., 1992; Pepper and Dodd, 1995; Isaksen, 2004). However, phase separation might be a factor influencing physical behaviour and chemical composition. In well 1/9-1, three DST oils from different carbonate reservoir

79 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

Oil Density (° API) intervals have been sampled with 20 25 30 35 40 45 50 55 60 2500 API densities (35 and 55 °API) and GOR’s increasing up-well. In

2/6-5 3000 between oil-filled intervals, dry 1/9-1 1/9-1 1/9-6 S 1/9-1 1/6-1 1/9-4 horizons are present, thus intra- 2/5-7 1/9-4 7/7-2 2/5-11 reservoir equilibration or evaporative 3500 7/12-6 3/7-4 7/12-6 fractionation can be ruled out, but 2/2-5 2/2-1 7/8-3 lateral hydraulic connection could be 4000 possible.

True Vertical Depth (m) 2/1-9 Res Lith 1/3-3 carbonate 2/7-20 siliciclastic 2/7-21 S Stratigraphy 2/7-27 S 4500 Ekofisk Fm 2/7-22 Tor Fm 2/12-1 Ula Fm Bryne Fm Skagerrak Fm pre-Jurassic Rotliegend 5000

80 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

- - - - 80 90 148 165 160 143 149 154 117 131 160 136 116 118 120 163 141 131 131 157 (°C) BHT BHT

- - - regional 66 36 20 32 70 44 48

606 131 196 548 285 309 375 1491 2740 1201 2850 1330 1181 GOR 14000 365 - 989 - 890 - 2330 - 890 -

43 30 29 55 46 46 35 46 39.4 39.2 36.6 43.2 38.8 44.5 45.4 29.7 40.5 47.6 45.4 32.5 38.7 50.5 42.5 40.7 (°API) Density

Type light oil light oil black oil light oil black oil light oil black oil black oil light oil light oil black oil light oil light oil light oil light oil black oil light oil light oil black oil light oil light oil light oil light oil black oil

Gamma gas field

Reservoir name Mjølner oil field oil Mjølner Tambar field oil SørGyda field oil field Ula field Ula field oil Brynhild Krabbe discovery - field gas/condensate Trym Embla oil field Embla oil Møyfrid discovery TommelitenAlpha gas field Albuskjell field gas Bumblebee discovery TommelitenAlpha gas field TommelitenAlpha gas field Embla oil field Embla oil field Embla oil Tommeliten Tjatse discovery Tommeliten Gamma gas field Tommeliten Gamma gas field - Embla oil field Embla oil

Region SW' SW' SW' SW' N' margin N' margin N' margin N' margin N' margin N' margin Søgne Basin Søgne Basin SW' SW' SW' SW' SW' SW' SW' SW' SW' SW' SW' SW'

Lithology clastic clastic carbonate carbonate clastic clastic clastic clastic clastic clastic carbonate clastic clastic carbonate carbonate clastic clastic carbonate carbonate carbonate carbonate carbonate clastic clastic

Formation Ula Fm Ula Fm Ekofisk Fm Ekofisk Fm Ula Fm Ula Fm Ula Fm Skagerrak Fm Ula Fm Ula Fm Tor Fm Bryne Fm Ula Fm Tor Fm Tor Fm Rotliegend Ekofisk Fm Palaeozoic Tor Fm Ekofisk Fm Tor Fm Ekofisk Fm Palaeozoic Devonian

Epoch Malm Malm Malm Malm Middle Triassic Malm Malm CretaceousUpper Dogger Malm Paleocene Paleocene Malm Upper CretaceousUpper Upper CretaceousUpper Rotliegends Paleocene

Upper CretaceousUpper Paleocene Upper CretaceousUpper Paleocene

Period Jurassic Jurassic Jurassic Jurassic Triassic Jurassic Jurassic Cretaceous Jurassic Jurassic Paleogene Paleogene Permian Jurassic Cretaceous Cretaceous Paleogene

Cretaceous Paleogene Cretaceous Paleogene Devonian

Base 4610 4214 4108 3511 3612 3342 3767 2953 3537 3730 3042 3159 4496 3670 3083 3131 3127 4435 3381 3123 3296 3286 4258 4347 (m)

TVD Top 4598 4211 4078 3434 3543 3333 3762 2927 3470 3715 3029 3153 4489 3666 3079 3122 3137 4178 3363 3115 3292 3262 4099 4314

Specifications of analysed oil and condensate reservoir samplesoriginating from the Central Graben and ordered according to

DST 1 DST DST 3.3 DST 1 DST 2 DST 1C DST 2 DST 1 DST 1 DST 1A DST 1 DST 8 DST 4 DST 1 DST 1 DST 6 DST 4 DST 4 DST 1 DST 1 DST 4 DST 1 DST 2 DST 1 DST 1 . 1 .

1 6 6

- - 3 9 2 3 5 4 1 1 1 22 5 1 1 4 27 S 11 4 7 6 S 20 21 S ------2/12 - 2/2 - 1/9 - Well 1/3 2/1 7/12 7/12 7/7 7/8 1/6 - 2/6 3/7 2/7 - 2/2 - 1/9 - 1/9 - 1/9 - 2/7 - 2/5 - 1/9 - 2/5 - 1/9 - 2/7 - 2/7 - occurrence and structural affiliation within the basin. Table

81 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

3.5.1.2 Gross chemical composition

(wt.-%) Fig. 3.3 shows that Lithology Aromatics carbonate 100 0 siliciclastic increasing API gravity is API gravity 29 - 31 accompanied by an 31 - 33 80 20 33 - 35 increase in the relative 35 - 37 37 - 39 39 - 41 60 40 abundance of saturated 41 - 43 43 - 45 45 - 47 hydrocarbons, describing a 47 - 49 40 60 49 - 51 single trend for oils 53 - 56 2/2-1 produced from both clastic 3/7-4 20 80 2/2-5 and carbonate reservoirs. 1/9-1 Saturates NSO (wt.-%) 0 100 (wt.-%) 100 80 60 40 20 0 Oils in clastic reservoirs generally follow Fig. 3.3. Gross chemical composition of collected Central Graben crude oils showing common trend towards more a maturity trend becoming saturate composition with increasing API gravity, independent increasingly lighter and of reservoir lithology. No regional features can be observed; the green circle identifies oils accumulated in the eastern more enriched in saturate carbonate reservoirs and the blue circles represent oils from compounds as a function of the western carbonate reservoirs (Fig. 3.1). maturity. On the other hand, oils from carbonate reservoirs are characterized by regional differences. While compositionally similar within the regional sub-sample sets, oils at the eastern margins contain less saturates than those from the western basin centre (green and blue circles, respectively).

Although describing a general maturity trend by decreasing density, secondary alteration processes can affect the composition. The increased NSO content of the lowermost reservoir interval in well 1/9-1 explains the reduced API gravity (see chapter 3.5.1.1) and can affect phase behaviour (Bailey et al., 1973; Bennett et al., 2007a). If these compartments are hydraulically connected, this feature could point to phase separation during secondary migration and charging of the reservoir horizons.

Other samples falling apart from the general trend are characterized by facies differences. While oils charged from more terrigenous-influenced source rocks (2/2-1, 3/7-4) contain higher aromatic proportions, the hypersaline 2/2-5 oil is richer in the NSO fraction.

82 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

3.5.1.3 Molecular geochemistry of hydrocarbons

1.5 0.20 (a) (b) highly anoxic anoxic / dysoxic 7/7-2

0.15 Ext.

1.0 Ula 7/8-3

7/12-6 2/1- 2/1-92/7-20 1/3-31/3-3 7/12-6 2/6-5 1/9-4 7/7-2 Reservoir Lithology 2/2-5 1/6-12/5-72/5-112/6-5 0.10 carbonate 1/9-1 1/9-1 homohopanes DST 8 7/8-3 siliciclastic 1/9-11/9-4 4 1/9-6 S 1/9-1 Sterane Index Sterane 3 2/12-1 3/7-4 30

C29 Ts/(Ts+Tm) /C organic matter organic % St 30/(27-30)

3/7-4 5 C35 / C34 hopanes C 3 2/2-1 0.19 - 0.25 0.5 2/7-21 S C 2/7 2/2-1 0.25 - 0.32 Reservoir Lithology 0.05 0.38 - 0.44 0.44 - 0.51 carbonate siliciclastic 0.51 - 0.57 oxic 0.57 - 0.63 Location ext. Ula Trend

terrestrial 0.69 - 0.76 environment SW' extent lacustrine 0.76 - 0.82 Søgne Basin 0.00 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

CC3131 22 22R R / /H H PrPr/Ph / Ph

The primary factors influencing crude oil compositions are the organofacies and maturity of the source rock. The secondary influence of biodegradation can be eliminated due to a lack of 25-NH and no alteration of n-alkanes. The principal source rock in the Norwegian Central Graben is the Upper Jurassic marine Hot Shale of the Mandal Formation containing 5 – 6 % TOC and a generation potential of ~450 mg HC/g TOC in its immature to early mature sections (e.g. Cornford, 1994; Ziegs et al., 2017). Though having a high overall source potential containing Type II kerogen, the Mandal Formation consists of alternating sequences with high and low source potential due to fluctuations in redox conditions during deposition, while the quality of marine zoo- and phytoplanktonic input remained consistent (Herbin et al., 1993). Fig. 3.4a illustrates this biological uniformity, while Fig. 3.4b shows the fluctuations in chemical conditions of the source intervals having principally charged oil accumulations. Oils along the northern margins of the Central Graben and the 2/2-5 discovery have been charged by highly anoxic to mixed anoxic/dysoxic source intervals with anoxia increasing further northwards. Petroleum accumulations in the SW’ basin centre and margins were charged from compositionally variable Type II facies. Interestingly, the eastern carbonate- reservoired oils (2/5-7, 2/5-11 and 2/6-5) show identical properties of marine anoxic facies (Fig. 3.4b, green circle), this being a hint that the northern Søgne Basin well

83 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

2/6-5 has been sourced from the Central Graben or an equivalent facies. The 2/2-1 discovery charged by the Farsund Formation shows higher terrigenous organic matter input in a more dysoxic marine environment. The Trym condensate (3/7-4) is indicated to be charged from a source rock with increased marine organic matter input in a more oxic marine environment. The T-seams of the coaly Lulu Formation in the Danish part of the Søgne Basin were accumulated during a high-standing water table resulting in stronger marine influence to the peat mires (Petersen and Brekke, 2001).

When assessing both depositional environment and maturity, it has to be kept in mind that all crude oils just represent cumulative mixtures of generated petroleum portions having been expelled from a potentially inhomogeneous source rock at different maturity levels (Wilhelms and Larter, 2004).

1.0 1.0 Res Lith carbonate siliciclastic (a) (b) Location Early maturemature ext. Ula Trend 0.8 0.8 Mid mature 2/7-21 S

SW' extent 2/7-20 Late mature Søgne Basin Late mature Bacterial degradation 1/3-3

0.6 High 0.6 2/1-9 maturity Most oil 2/12-1 1/9-6 S 2/2-5 oils 7/12-67/12-62/5-72/6-51/9-41/6-1 2/5-11 1/9-11/9-1 0.4 0.4 1/9-11/9-4 St29 20S/(S+R ) 20S/(S+R St29 Low maturity oil Migrational fractionation 7/8-3 7/7-2

29Ts / (29Ts + 30NH) + (29Ts 29Ts / steranes: 20S / (20S + 20R) + (20S / 20S steranes:

3/7-4

29 Res Lith

C 2/2-1 carbonate 2/2-5 0.2 0.2 siliciclastic Location ext. Ula Trend Immature source rock SW' extent Søgne Basin 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

C29 steranes:St29 bb/(aa+bb) ββ / (αα + ββ) Ts / (Ts + Tm)

Fig. 3.5. Thermal maturity of the analysed oil and condensate samples based on (a) sterane

A cross plot of the %20S and %ββ sterane maturity biomarker (Fig. 3.5a) shows that most of the Central Graben crude oils have reached the stage where chemical equilibrium of the αα sterane isomerization (%20S > 0.55) has taken place, but %ββ which is more effective at slightly higher maturities (Peters et al., 2005) ranges between 0.57 and 0.65 indicating the beginning of peak oil generation, according to correlations in Waples and Machihara (1990). The samples plot in a

84 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS very confined area and do not show any specific trend related to spatial or stratigraphic occurrence. The oil from well 2/2-5 is the only oil falling away from the general trend showing values of 0.52 (%20S) and 0.47 (%ββ) indicating expulsion at an early maturity level.

According to a cross-plot of Ts/(Ts+Tm) and 29Ts/(29Ts+NH) (Fig. 3.5b) the oils can be separated into peak- to late oil window maturity, high mature and supermature ranges. Crude oils in siliciclastic Jurassic reservoirs show a very broad maturity range whereas Cretaceous carbonate reservoir oils lie in a confined, high maturity range between 0.40 and 0.51 in their 29Ts/(29Ts+NH) ratio. While the 1/9-1 oils (Tommeliten Alpha field) are principally of the same maturity, the Tommeliten Gamma crudes are very different. 1/9-4 DST 4 (50.5 °API) is indicated as lower mature than the deeper DST 1 (46 °API) and 1/9-6 S (30 °API), a side track of the 1/9-4 well, which are both similar to the Albuskjellet condensate, the highest mature crude in the Greater Ekofisk area trend (Hughes et al., 1985). The eastern carbonate reservoirs 2/6-5 and 2/5-7 contain oils of the similarly high maturity, while the 2/5-11 oil is of intermediate maturity within the carbonate-reservoired oils.

Crude oils from the Steinbit Terrace, north of the Mandal High, block 2/2, are least mature, according to sterane epi-and isomerization indicators (Fig. 3.5a). Still, the maturity of the 2/2-5 oil might be overestimated using C29 steranes and Ts/(Ts+Tm) (Fig. 3.5b) due to its hypersaline source environment (Peters et al., 2005; Pedersen et al., 2006). Oils from the Jæren High (wells 7/7-2 and 7/8-3) were expelled at a late maturity stage (Fig. 3.5b). Also, the Trym condensate, sourced by the coaly Jurassic Lulu Formation from the south, seems to be charged from source intervals with intermediate maturity of the hopanes (Fig. 3.5b). Oils in siliciclastic reservoirs from the basin-marginal Ula Trend, comprising wells 7/12-6, 1/3-3 and 2/1-9, and the Mjølner field seem to be charged from a highly mature source rock. Here, well 1/3-3 crude oil is at the supermature stage similar to the Middle Jurassic and Rotliegend reservoirs in block 2/7, wells -20, -21 S and -27 S, at the western margin of the Central Graben.

Oil-water interactions typically accompany biodegradation of petroleum and results in the selective loss of light hydrocarbons, especially benzene, toluene, phenols and other aromatics of the light hydrocarbon fraction (Bailey et al., 1973;

85 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

1.5 Reservoir rock Palmer, 1984; Palmer, 1993). Acting carbonate Heavy biodegradation siliciclastic alone at temperatures >80 °C, these Res Temp (°C) 1.2 80 - 90 90 - 100 processes have only a minor effect 110 - 120 120 - 130 1/9-4 130 - 140 1/9-1 1/9-4 on the chemical and physical 0.9 140 - 150 1/9-1 150 - 160 1/9-61/6-1 S 160 - 170 2/7-22 unknown Pristine oils properties of oil (Peters et al., 2005), 3/7-4 2/1-9 0.6 1/9-1 and if related to secondary 2/7-20 2/7-21 S

Toluene/Methyl Cyclo-Hexane 2/2-1 migration they have an insignificant 2/12-12/2-57/7-2 7/12-6 0.3 Water washing1/3-3 2/6-5 2/5-11 impact (Lafargue and Barker, 1988). 7/8-3 2/5-7 However, it must be noted in this 0.0 0.0 0.2 0.4 0.6 0.8 1.0 Benzene/Cyclo-Hexane context that all oils from the Fig. 3.6. Cross plot of benzene and toluene ratios northern and eastern margins of the against their water-insoluble structural equivalents. Benzene is more hydrophilic than is toluene. Central Graben, except the 2/1-9 oil, are depleted in toluene and benzene, independent of reservoir lithology (Fig. 3.6). Oils from the graben axis (blocks 1/9, 1/6) and western margins (block 2/7) are not depleted in water-soluble aromatic moieties. However, benzene ratios are comparably elevated in Embla oils and the adjacent 2/7-22 field while the slightly less soluble toluene shows normal ratios to its cycloalkane equivalent. This could represent an effect of mixing with biodegraded Palaeozoic oils (Ohm et al., 2012) that is only observable in the light fraction as benzene and toluene are abundant compounds in crude oils.

3.5.2 The polar NSO fraction of petroleum

For FT-ICR-MS evaluation the oil sample set was condensed from 24 to 16 oil samples covering the whole maturity and spatial ranges of oils originating from different organofacies within the Central Graben. Putting the focus on samples containing higher rather than lower NSO fraction weights ensures reproducible FT-ICR-MS measurements with overall high signal numbers, a natural Gaussian distribution of signals, and low influence of potential contaminations from drill mud additives. The latter cannot be completely excluded as surface active and bioactive agents used during drilling are rich in cyclic and acyclic sulphur and oxygen constituents that could be ionized using ESI negative mode.

86 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

3 O 1 0.86 0.23 0.05 0.11 3.85 1.56 7.54 1.64 0.39 0.13 0.61 0.10 0.16 0.86 0.13 S

------2

O 1 0.01 0.13 0.06 0.02 0.74 0.02 0.08 0.02 N

1 1.41 4.51 4.85 3.13 4.40 3.65 1.13 3.31 5.07 4.45 2.20 3.85 2.83 2.23 0.08 N

- 1 S (% TMIA) 1

0.31 0.49 3.96 1.03 0.19 0.18 0.09 0.30 0.16 0.16 0.25 0.12 0.47 0.22 N

------2 N

0.09 1.02 0.61 0.07 0.18 0.88 0.12 0.48

1 N

68.14 62.28 64.12 61.21 53.25 61.03 58.59 60.98 45.32 41.35 45.72 54.40 64.97 53.87 10.15

- 3 O 0.04 0.09 0.02 0.01 0.65 0.05 0.12 0.06 0.33 1.42 0.04 0.70 0.13 0.37

Ion Abundance Relative

2 O 2.13 3.55 2.84 1.98 7.32 5.05 4.45 6.28 8.76 6.61 5.40 9.74

13.80 10.32 16.48

1 O

26.07 28.03 20.81 31.11 19.00 24.79 23.73 28.55 41.82 42.44 34.33 33.81 23.79 30.84 69.16 MS forMS 16 oil samples. -

ICR

- w 385 414 438 428 362 376 378 401 393 484 409 402 394 400 387 M total ) FT -

n M 362 388 411 402 342 356 357 378 372 451 384 380 371 375 368 total

# of 2000 3237 4301 3262 1292 1370 1215 2564 2833 3449 2813 1818 2188 2177 1525 signals

%) - (wt Polars 4.73 6.14 14.81 11.28 2.00 3.74 3.07 6.84 5.18 19.05 6.83 6.69 2.87 5.71 1.80

°API 43 43 30 29 55 46 46 39 41 33 39 41 45 39 45

Region N' margin N' margin N' margin N' margin SW' SW' SW' SW' SW' SW' SW' SW' SW' Søgne Basin Søgne Basin

TVD (m) 4108 3612 3342 3767 3042 3083 3296 4610 3730 3670 3381 3286 4347 2953 3537 abundances of compound classes obtained from ( ESI

DST 1 DST 1C DST 2 DST 1 DST 8 DST 6 DST 1 DST 1 DST 1 DST 1 DST 1 DST 2 DST 1 DST 1 DST 1A DST Relative

. 2

. 3

6 1 - - 9 2 3 1 1 4 1 5 11 7 21 S 5 4 ------Well 2/1 7/12 7/7 7/8 1/9 1/9 1/9 2/12 2/2 2/2 2/5 2/5 2/7 2/6 3/7

Table

87 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

3.5.2.1 Gross composition

The number of assigned peaks in the mass spectrum of the acidic polar compounds ranges between 489 and 4300 but for the majority of samples it lies between 1200 and 2800 (Table 3.2). While the number of assigned signals is not necessarily indicative of the absolute concentration of acidic NSO compounds in the oils, a rough trend of decreasing number of signals with increasing API gravity is discernible. Variability is broadest in the range of light oils (32 – 45 °API). The number average and the weight average molecular weights (Mn: 320 – 451 Da and

Mw: 331 – 484 Da) decrease as a function of petroleum density and is determined by the relative concentration of individual compounds in the whole oil. Heavier polar compounds become less abundant with increasing API gravity.

3.5.2.2 Compound class distribution

100 0 Location Table 3.2 and Fig. 3.7 show the O1 Northern 60 0 SW' extent C29 Ts/(Ts+Tm) relative abundances of the major 3/7-4 50 50 0.19 - 0.25 0.25 - 0.32 0.38 - 0.44 compound classes of Central Graben 50 10 0.44 - 0.51 0.51 - 0.57 2/2-5 0 100 2/2-1 0.57 - 0.63 crude oils being dominated by N1, O1 100 50 0 0.69 - 0.76 0.76 - 0.82 40 20 2/5-11 and O2, thus resembling the 2/5-7 7/8-3 2/6-5 elemental class distributions of 30 7/12-62/12-1 30

2/1-9 1/9-41/9-1 2/7-21 S undegraded crude oils worldwide 7/7-2 1/9-1 20 40 (e.g. Hughey et al., 2002; Hughey et al., 2004; Li et al., 2010; Oldenburg 10 50 N1 90 80 70 60 50 40 O2 et al., 2014; Poetz et al., 2014; Liu et Fig. 3.7. Rudimentary oil family discrimination al., 2015). In all oils sourced by using elemental data from FT-ICR-MS. Interferences with maturity can be expected. marine shales N1-compounds are the Combined with maturity assessment, compositional major components with 41 to variations within the same maturity stage can be used for classification. Circles represent regions as 68 % TMIA, and that abundance in previous illustrations. generally increases with progressive maturity of the oils (see as well Poetz et al., 2014). The O1 class most likely comprising phenols, sterols, alcohols or indols (Wang et al., 2011; Pan et al., 2013) contributes with 19 to 43 % TMIA (1/9-1 DST 8 and 2/2-5) to the total signal, and

O2 compounds comprised of carboxylic acids, range between 2 and 14 % TMIA (7/8-3 and 1/9-1). Thus, components with one heteroatom account for 72 –

88 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

94 % TMIA of the acidic polar fraction, as determined using the ESI (-) ionization mode. Lower mature samples, located in the 2/2 block, contain lower N1 proportions than higher mature ones, but an influence of the source environment within marine oils could not be ruled out: the 2/2-5 oil was charged from a hypersaline, marine source rock, and the source of the 2/2-1 oil contains a relatively high terrigenous contribution (Pedersen et al., 2006). Significant compositional differences of the

Trym condensate (well 3/7-4) containing 86 % TMIA of O1 and O2 components can be addressed to its origin being coal-sourced from the shallow-marine Lulu Formation.

The maturity effect on compound class distributions has been shown for source rock extracts with the N1 class increasing in relative abundance as a function of thermal stress (Poetz et al., 2014). Differently, the relative abundances of the N1 compound classes in crude oils does not show a consistent behaviour related to maturity (e.g. Oldenburg et al., 2014 for a Viking Graben oil suite; or this communication), and must rather be a result of migration-related effects when source variations can be excluded.

Regional variations in the N1 and O1 classes have been documented (Fig. 3.7).

At the outer Steinbit Terrace (block 2/2), oil samples are N1 dominated but slightly enriched in O1 compounds. Oils at the Cod Terrace and Jæren High, northwest of the Ula Trend (7/12-6, 7/8-3 and 7/7-2), are strongly dominated by the N1 compound class and depleted in the O2 class, ranging between 61 – 64 % TMIA N1, 21 –

31 % TMIA O1 and only 2.0 – 3.6 % TMIA O2 compounds. A clear regional differentiation can be observed between carbonate-reservoired oils from the eastern margin and western basin centre which are most likely not related to maturity (see Fig. 3.5b). While oils at the Steinbit Terrace (2/5-7, 2/5-11) and the northern Søgne

Basin (2/6-5, Fig. 3.1) contain 46 – 54 % TMIA N1 compounds, 31 – 34 % TMIA O1 and 7 – 10 % TMIA O2 compounds, those in the West in block 1/9 contain higher N1 portion with 53 – 61 % TMIA and lower O1 with 19 – 25 % TMIA.

3.5.2.3 N1 compounds & maturity assessment

The DBE class distribution is clearly dominated by DBE 9, 12 and 15 classes (Fig. 3.8a). Previous studies have shown that these DBE classes belong to the carbazole-family (Hughey et al., 2002; Purcell et al., 2007; Oldenburg et al., 2014;

89 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

Poetz et al., 2014) consisting of a carbazole-core unit (DBE 9) with one or more ortho-fused benzene rings adding 3 DBE to the core unit: DBE 12 = benzocarbazole, DBE 15 = dibenzocarbazole. Higher fused homologues of the DBE 18+ classes can be annelated in ortho- and peri-positions (Poetz et al., 2014) and thus the DBE class distribution may vary by ±2.

The N1 class further

contains elevated contributions TMIA) of DBE 10, 13 and 16 (Fig. 3.8) (%

compounds representing phenylindoles and showing a similar thermal behaviour like

relative abundance carbazoles with increasing abundances from DBE’s 9 to 12 DBE class / 10 to 13 while DBE’s 15 and Fig. 3.8. DBE class distribution of the N1 compound 16 show lower abundances. class. Dotted and dashed lines indicate oils from eastern and western carbonate reservoirs, respectively. Although of different organofacies origin and containing significantly lower N1 proportions, the 3/7-4 condensate shows a predominance of DBE classes representing carbazole and phenylindol homologues.

Fig. 3.9 shows a decrease in the relative intensities of carbazoles (DBE 9) with increasing maturity and an increase in relative intensities of benzo- (DBE 12) and dibenzocarbazoles (DBE 15), reflecting annulation and aromatization of core structures as a function of maturity. Oldenburg et al. (2014) have shown that the linear decrease of the DBE 9 class correlates well with vitrinite reflectance data

(VR) of crude oils in the Viking Graben (Fig. 3.9, inset bar) ranging between 0.68 and 1.1 %Rc within their sample set. Further testing with worldwide oil samples of different origin (Mahlstedt et al., 2016) supports this VR scale. Thus, we can propose that the investigated oils in the Central Graben most likely range between

0.75 %Rc for the earliest mature oil (2/2-5) and 1.1 %Rc for the most mature oil (2/7-21 S). According to vitrinite reflectances assigned by Oldenburg et al. (2014), the majority of selected Central Graben oils range between 0.8 to 1.0 %Rc and are thus in the peak oil window.

90 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

Benzocarbazoles (BCs)

Lithology carbonate siliciclastic 29Ts/29Tm 100 0 0.19 - 0.25 0.25 - 0.32 0.38 - 0.44

0.44 - 0.51 80 20 0.51 - 0.57 0.57 - 0.63 0.69 - 0.76 0.76 - 0.82

60 40 1/9-1 DST 6 1/9-1 DST 8 2/2-1 3/7-4

40 1.1 60

0.84- 0.68 0.74- 0.85 0.78

20 80

0 100 100 80 60 40 20 0

N1 DBE 9 (Cs) Carbazoles (Cs) Dibenzocarbazoles (DBCs)

Fig. 3.9. Maturity assessment based on the distribution of carbazole and its higher fused homologues (Oldenburg et al., 2014). Maturities of the polar and hydrocarbon components of petroleum fit quite well, but show a higher contribution of N1 DBE 12 benzocarbazoles for carbonate oils and thus an apparent maturity retardation of those oils.

3.5.2.4 Correlating GC-MS maturity and degree of N1 annulation

Aliphatic and aromatic maturity biomarkers with specifity for the entire maturity range from early to high mature are here correlated with the N1 annulation of FT-ICR-MS data. As can be observed from relative distributions in Fig. 3.9, DBE 9 relatively decreases while DBE 15 relatively increases as function of maturity. Fig. 3.10 shows abundances of DBE classes with 29Ts/(29Ts+NH) maturity ratio correlating the best. Although showing similar geochemical behaviour like Ts/(Ts+Tm) due to their same bacterial origin (Moldowan et al., 1991) and maturity effect (Kolaczkowska et al., 1990), 29Ts/(29Ts+NH) might be less influenced by the source rock facies (Peters et al., 2005) but carbazoles might show similar effects (Bakr and Wilkes, 2002). However, elevated 30-NH (nor- hopane) in the 2/2-1 oil could result in an apparently lower maturity level obtained from 29Ts/(29Ts+NH).

91 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

12 2/1-9 10

2/7-21 S 7/12-6 2/12-1 7/8-3 8 1/9-4 7/7-2 DBE 9, 12 and 15 1/9-1

2/6-5 1 2/5-7

2/2-1 6 1/9-1

2 2/5-11 N1 DBE 15 (% TMIA) 4 2/2-5 tive TMIA’s in a ternary plot. R² R² = 0.5 2

)

c ( 3/7-4 ( 0 1.0 0.8 0.6 0.4 0.2 0.0

14 1/9-1 1/9-4 12 2/1-9 1/9-1 7/7-2

2/5-7

2/7-21 S 10 2/6-5 7/12-6 2/12-1 7/8-3 09 2/2-1 8 2/5-11 2/2-5 R² = 0. 6

N1 DBE 12 (% TMIA)

)

2 ( 3/7-4

Location ext. Ula Trend SW' extent BasinSøgne carbonate siliciclastic Res Lith (b)

1.0 0.8 0.6 0.4 0.2 0.0

12 4 because of too strong facies variations. 10

34 8 1/9-1 1/9-1

R² = 0. 2/2-5 7/7-2 6 1/9-4 2/5-7 N1 DBE 9 (% TMIA) (% 9 DBE N1

2/6-5 7/8-3 4 2/1-9 2/2-1 7/12-6 2/5-11 2/12-1

)

2 Correlation of maturity indicator 29Ts/(29Ts+NH) to the total monoisotopic abundances of N 2/7-21 S ( 3/7-4

. ) a ( 10 . 0 3

1.0 0.8 0.6 0.4 0.2 0.0 29Ts / (29Ts + NH) + (29Ts / 29Ts Fig. compounds illustrating that individual abundances (TMIA) react in the samemanner as their rela has been calculated, excluding 3/7 -

92 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

Comparing Fig. 3.9 and Fig. 3.10, oils charged by the terrigenous-influenced Farsund Formation (2/2-1, Gabrielsen et al., 1985) and the coaly Lulu Formation (Trym, 3/7-4) do not follow the trend of oils described by Oldenburg et al. (2014) and others (e.g. Hughey et al., 2004) that are charged from source rocks containing pure marine organic matter. The Trym condensate is significantly depleted in N1 compounds (5 % TMIA) and cannot be used for maturity assessment. However, the

2/2-1 shows similar N1 DBE distributions as purely marine oils, thus falling within Oldenburg’s trend, but is slightly depleted in DBE 9 compounds (Fig. 3.10a), when compared to its maturity level from conventional biomarkers (Fig. 3.5).

29Ts/(29Ts+NH) suggest expulsion at early mature levels, but the N1 DBE distribution assigns a maturity of 0.95 %Ro in the peak oil window (Fig. 3.9). These outliers could indicate that the proposed maturity-related trend (Oldenburg et al., 2014) is only valid for oils sourced from purely marine source rocks.

Interestingly, lowest mature oils from eastern (2/5-11) and western (1/9-1) carbonate reservoirs show very different DBE distributions with maturity (Fig. 3.10). While 1/9-1 condensates are enriched in smaller DBE 9 and 12 compounds, the 2/5-11 oil is depleted in bigger and potentially more rod-shaped DBE 12 and 15 compounds for samples of comparable maturity level, with highest variations occurring within the DBE 12 class. Both, a relative enrichment in DBE 9 and relative depletion of DBE 15 classes result in apparently higher relative DBE 12 contribution causing an apparent maturity retardation when using the polar maturity assessment for carbonate oils. Thus, parameters based on the relative stability of hydrocarbon isomers record differently to that based on annulation and aromatisation of the carbazoles, the latter of which is related to aliphatic chain shortening, as presented below (Fig. 3.11).

3.5.2.5 Detailed variations due to reservoir lithology

As alluded to above, the Central Graben crude oils stored in carbonate reservoirs plot above the maturity trend, containing a higher DBE 12 portion by up to 5 % at the expense of the DBE 9 and 15 compounds (Fig. 3.9). Here, carbonate reservoired oils from well 1/9-1 at the western rim of the Central Graben (condensates with API = 46 – 55 °) show a higher deviation from clastic oils than those from the eastern rims near the Mandal High (Fig. 3.1). Well 1/9-4 from the

93 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS axial part of the Central Graben is similar to the eastern wells in carbazole abundance distribution. For the eastern wells, 2/5-11 and 2/5-7 at the Steinbit Terrace are most detached from the maturity trend whereas 2/6-5 in the northern Søgne Basin is most similar to oils in clastic reservoirs. The levels of detachment reflect maturity developments within each regional occurrence (Fig. 3.5, inset) with less mature samples being most detached (1/9-1 and 2/5-11).

ng on

DBE 12). Earliest

1 earliest mature clastic oils

E‘

components with significantly longer chained substituents

1 W‘

highest mature clastic oils

(b)

normed rel. Intensity rel. normed TMIA) (%

15+ C

3.1 ) 100 0 80 Fig. 60 20

40 20 40

14 - 0 6

100 C 60 80 60 (a) 80 40 Chain length distribution of aliphatic carbon attached to benzocarbazole core structures (N

. 20 11 . 3 100

0 5 - C29 Ts/(Ts+Tm) 0.19 0.25 - 0.25 0.32 - 0.38 0.44 - 0.44 0.51 - 0.51 0.57 - 0.57 0.63 - 0.69 0.76 - 0.76 0.82 - 1 carbonate siliciclastic Reservoir LithologyReservoir C Fig. mature oils contain longest side chains (a). Peak oil window mature oils stored Upper in Cretaceous carbonate reservoirs (dashed) contain aliphatic chains with higher contribution of intermediate carbon numbers than clastic oils, dependi regional occurrence (b). Crude oils at the eastern rim contain N thethan at westernmargin (compare

94 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

Using ternary diagrams to depict chain length distributions of the major N1 DBE classes related to the carbazole family (12, 15, 18), oils from carbonate and clastic reservoirs fall along the same evolutionary trend, beginning at early mature levels with high C15+ contribution, followed by intermediate aliphatic carbon numbers dominating peak mature oil compositions and finally short side chains prevailing in late oil window oils. This chain length shortening is exemplified by the benzocarbazoles (DBE 12) in Fig. 3.11a. Significant compositional differences exist between the oils in the two reservoir lithologies when the individual chain lengths are considered. Fig. 3.11b shows the carbon number (CN) distribution of DBE class 12 normalized to the individually highest peak for each sample and outlines the differences between regional oil compositions. While oils in clastic reservoirs (solid lines) get more enriched in C2-4 aliphatic side chains and depleted in higher CN’s with increasing maturity, which is most evident for highest mature oil, 2/7-21 S, the oils from eastern carbonate reservoirs are different. Although most deviated from the general trend in Fig. 3.9, chain length distributions of western 1/9 oils are similar to clastic oils of similar maturity (compare inset of Fig. 3.5). However, the eastern carbonate reservoirs (2/5-11, 2/5-7, 2/6-5) contain oils with significantly increased C6+ side chains apparently resulting from a depletion in the highest abundant, short-chained (C2-3) aliphatic carbon numbers. This relative increase of longer chained, heavier compounds in the major compound class of the polar fraction might affect API gravities with higher densities of eastern oils (Fig. 3.2).

The intensity of increase in intermediate to long chained homologues (or the loss of short chained compounds) of the oils in the eastern carbonate reservoirs coincides with their ratio of toluene and methylcyclohexane, which is an indicator of the severity of oil-water interactions. Fig. 3.6 shows that oils in the western carbonate reservoirs are not affected by oil-water interactions. Normally accompanied by biodegradation, the effect of water washing is covered by those bacterial processes, and may occur alone during migration of petroleum through the water-saturated pores of both clastic and carbonate carrier beds and intensifies with longer migration distances (Lafargue and Barker, 1988). Although N1- containing carbazoles are less polar and thus less water-soluble than oxygen- bearing heterocompounds, water solubility mostly increases with temperature and can be important at higher temperatures prevailing in burial depth of early or main

95 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS petroleum generation. A more detailed discussion on physicochemical interactions of the particular heterocompounds with polar sites in the carrier-reservoir system is carried out in the following paragraphs after the description of individual distributions.

3.5.2.6 Other influences on N1 chain length distributions

Maturity is the major control on crude oil compositions in the Central Graben (Hughes et al., 1985; Cornford, 1994) affecting distributions of carbazole compounds (DBE classes), but not necessarily their chain length distributions. Therefore, we calculated the PEARSON correlation factor R of the relative compound abundances of individual chain lengths in ESI negative FT-ICR-MS mass spectra (Fig. 3.11b) and the oil’s maturity biomarker ratio of 29Ts/(29Ts+NH) (Fig. 3.9). Such a correlation is exemplified in the appendix (Fig. 3.A.1). A cross plot with the carbon number of aliphatic side chains of each DBE class shows the development with increasing chain length for Central Graben crude oils stored in clastic (Fig. 3.12a) and in carbonate reservoirs (Fig. 3.12b). Although charged from a marine source environment (Fig. 3.4a), 3/7-4, 2/2-1 and 2/2-5 oils have been excluded from correlations due to high facies variations.

inherent (a) clastic oils (b) carbonate oils very strong

moderate

low

very low

low

moderate

very strong inherent

96 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

In general, chain lengths shortening of aliphatic carbon should result in negative correlations of long chains with maturity, while the relative increase in short aliphatic moieties results in positive correlation. Oils in clastic reservoirs with apparently very short migration distances follow the expected compositional development from cracking of long to shorter chains in the source rock, effective expulsion of all generated and cracked moieties, and show no significant compositional fractionation during secondary migration. DBE classes 9 and 12, carbazoles and benzocarbazoles, show very similar correlation trends (Fig. 3.12a) with strongly positive R of short aliphatic chains (C0-3), subsequently changing to strongly negative R for intermediate to long chains (C11+). Larger core molecules of dibenzo- and naphtobenzocarbazoles (DBE 15 and 18) show strongly positive trends for C0-8 and C0-12 short to intermediate number of aliphatic carbon atoms. The strongly positive correlation coefficients of DBE 15 and 18 carbazoles for short to intermediate aliphatic chains can be interpreted as a result of maturity-related aromatization of even longer chained, lower-fused carbazole homologues following reaction mechanisms suggested by Poetz et al. (2014). It is to be expected that carbonate reservoired oils should fall within the same linear trend as clastic oils (Fig. 3.12b). However, due to their confined range in thermal maturity (Fig. 3.5), the correlation of compound abundances with 29Ts/(29Ts+NH) could indicate different mechanisms than maturity. The maturity-related decrease in DBE 9 carbazoles can be observed, but the relative increase of short alkyl chains as a result of chain shortening is not well developed. Otherwise the formation of higher fused DBE 15 and 18 carbazole homologues by ring annulation as a function of maturity is expressed by a strongly positive R. The inverse correlation trend with increasing carbon number within the DBE 12 class supports the conclusion from chain length distributions that this class is significantly depleted in short chained compounds (chapter 3.5.2.5). Assuming a homogeneous source and comparing oils of similar maturity, the removal of short-chained benzocarbazoles is a feature of migration fractionation within the carrier-reservoir system.

In that regard, Stoddart et al. (1995) showed that longer migrated oils in the chalk-hosted Eldfisk field, Central Graben are enriched in alkylated carbazoles

(equivalent to N1 DBE 9) relative to benzocarbazoles (equivalent to DBE 12) and that higher methylated C3 homologues are less hindered in migration than C1 and

C2 homologues. These migration distance trends based on GC-MS results were

97 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS confirmed for relative abundances using FT-ICR-MS on the well investigated Duvernay petroleum system, Canada (Liu et al., 2015) and can be related to a higher surface activity of DBE 12 benzocarbazoles (Bennett et al., 2004). Benzocarbazoles were believed to sorb on mineral surfaces (Li et al., 1995; Larter et al., 1996; Larter et al., 1997; Li et al., 1997). Thus, purely water-wet siliciclastic systems might retain/adsorb less surface-active compounds than oil-/mixed-wet carbonate carrier systems. However, carbonate reservoirs might be water-wet prior to oil migration (Abdallah et al., 2007). Even minuscule amounts of polar high- molecular weight (HMW) compounds can sorb to mineral surfaces causing a change to oil-wetting conditions (Akbarzadeh et al., 2007) and enabling a continuous oil flow without major compositional alterations. Bennett et al. (2004) have shown that low-molecular weight (LMW) C0-3 alkylphenols are able to pre-condition water-wet surfaces to make lipophilic hydrocarbons and lower-polar compounds sorb within the carrier system until saturated. Furthermore, van Duin and Larter (2001) have proved that benzocarbazole distributions are rather altered by oil-water partitioning than by oil-rock partitioning, thus excluding an influence of primary migration within oil-saturated source rocks while promoting migration-related variations within water-filled pore spaces. Following ratios on water-oil partitioning and solubility in Fig. 3.6, this would suggest that the 7/8-3 oil in a clastic system might have migrated longer than other marginal oils, but this cannot be observed in the DBE 12 CN distribution, due to its lower maturity.

With regard to the solubility context and the negative “maturity” trend in carbonate-reservoired oils (Fig. 3.12b), it can be said that those are likely related to secondary effects during migration rather than to maturity effects. Such secondary effects are caused by oil-water interactions at the front of the initial oil charge flowing through water-filled pores, fractures and faults. If longer migrated oils contain dominantly alkylated carbazoles over alkylated benzocarbazole species (Li et al., 1995; Stoddart et al., 1995), then oils in carbonate reservoirs at the eastern margin and the 1/9-4 condensate would have migrated shorter distances than the 1/9-1 condensates from the graben axis due to their lower deviation from the general clastic maturity trend and thus higher DBE 12/9 ratios. However, the higher DBE 9 portion of 1/9-1 oils is to be explained with slightly lower maturity (Fig. 3.5b, inset). While migrating into water-filled pore or fracture systems of carrier rocks, benzocarbazoles with shorter chains (2 – 3 aliphatic carbon atoms)

98 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS get removed from the oil phase. This mechanism would explain higher relative DBE 12 compound abundances of intermediate and long aliphatic chains in the eastern oils pointing to longer migration within chalk sequences, when assuming only one major expulsion charge from a badly expelling source rock (Ziegs et al., 2017). An excellent expeller would rather instantaneously expel multiple smaller increments of oil charges, of which the very first one would be significantly altered by pre-conditioning of migration pathways. Subsequent oil charges of higher maturity are less or not compositionally altered during migration but make up a higher portion of an oil accumulation, thus overprinting migration-related information with the maturity-related chain length distributions. Based on this, the loss of short chained benzocarbazoles and other surface-active compounds in the Central Graben petroleum system, containing a source rock with low expulsion efficiency, might be an indicator for migration distances within different lithologies, thus resulting in different controlling mechanisms.

Following on the above observations, it further needs to be mentioned that for oils in carbonate reservoirs an assignment of biomarker equivalent vitrinite reflectances seems to be difficult, because the maturity assessments based on the hydrocarbon and polar NSO fractions give different results (compare Fig. 3.5 and Fig. 3.9). Due to secondary controls other than maturity on DBE 12 class contribution and distribution within the DBE 9/12/15 ternary diagram, Oldenburg’s maturity assignment may not be applicable to carbonate oils or an adjustment for migration distances would be necessary.

3.5.2.7 Factors controlling O1 compounds

O1 compounds, being the second most abundant compound class in the Central Graben crude oils, might be even more influenced by secondary processes than N1 compounds due to higher electron negativity and a stronger dipole moment depending on bonding type. They are dominated by compounds with DBE 4, 5, 6, and 7. O1 components with much higher aromaticity (up to DBE 22) are also found, but their relative abundances continuously decrease with increasing DBE number (Fig. 3.13). Possible core structures for the DBE 4 class could be phenol and its alkylated homologues (Kim et al., 2005; Shi et al., 2010; Liao et al., 2012; Pan et al., 2013), while those with DBE 5 and 6 could refer to indanols and indenols (Liao et

99 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS al., 2012; Pan et al., 2013). Geng et al. (2012) suggested that DBE classes 7, 9 and 11 correspond to naphthols, fluorenols and phenanthrenols, respectively. DBE 4, 5 and 7 class might also refer to nonaromatic, monoaromatic or triaromatic sterols.

The relative abundances

of O1 DBE classes are

TMIA)

TMIA principally influenced by facies

and thermal maturity and range between 1.3 and

DBE class 22.1 % TMIA for the

relative abundance alkylphenols (DBE 4), 1.5 and 11.9 % TMIA for indanols (DBE DBE class 5), 2.6 and 8.6 % TMIA for DBE

Fig. 3.13. DBE class distribution of the O1 compound 6 and 1.9 and 7.2 % TMIA for class. Dashed lines indicate oils from carbonate reservoirs. Inset shows oils with lower O1 contribution (cf. DBE 7. Early mature oils from Fig. 3.7). the Eastern rim of the Central Graben contain mainly DBE 4 compounds, and only insignificant amounts of species with higher aromaticity. The same feature can be observed for the crude oil 3/7-4 derived from the coaly Lulu

Formation and containing the overall highest proportions of O1 compounds related to a source dominated by terrigenous organic matter. This correlates to high phenol contents in lignocellulosic kerogen pyrolysates (Tissot and Welte, 1984; Horsfield, 1989; Larter, 1990; Rullkötter and Michaelis, 1990; Vandenbroucke and Largeau,

2007). With increasing maturity, the aromaticity and annulation of O1 compounds increases with increasing abundances of higher DBE classes relative to DBE 4 and 5 classes. However, peak mature oils contain DBE classes 4 – 7 representing a broad chemical inventory. The high-GOR condensates in the graben axis (well 1/9-1) which are peak to late mature are dominated by DBE 4 and show a subsequent decrease for higher DBE classes. Interestingly, other carbonate- reservoired oils (1/9-4 and at the Steinbit Terrace) show several secondary modes at DBE 6, 7 and 11 which is rather a feature of peak mature oils and represents a broad chemical variability. The 2/5-11 oil is slightly depleted in DBE 4 and 5 and dominated by DBE 6 and 7. This is a similar feature as observed in the Mjølner field (2/12-1). Two peak mature oils from the Ula Trend at NE’ terraces are dominated by DBE 4 compounds and show a second mode at DBE 6 whereas the

100 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS northernmost oils from wells 7/7-2 and 7/8-3 show highest abundances of 4 and 5 DBE classes which could be attributed to their slightly lower maturity (see Fig. 3.5b). The highest maturity oil 2/7-21 S is strongly dominated by higher DBE classes of 6 – 8, whereas DBE classes 4 and 5 (phenols and indanols) are largely absent. A ring annulation as developing within the N1 class cannot be observed in

O1 DBE distributions. Similar to N1 DBE 9, 12, 15 and 18 classes, a correlation with 29Ts/(29Ts+NH) was conducted for O1 DBE 4 and 5 classes showing the same evolutions with increasing number of carbon atoms for oils in different reservoir lithologies, and supporting the surface-active and interphase mechanisms influencing carbazole distributions. An illustration, explanation of the correlation and brief discussion can be found in the appendix (Fig. 3.A.2).

TMIA) TMIA

(a) DBE 4 C# relative abundance

TMIA

TMIA) %

(b) DBE 5 C# relative abundance

carbon number: C#

101 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

In contrast to the abundance distributions of carbazoles, O1 DBE 4 and 5 compounds are depleted in their low molecular weight species until C12 (Fig. 3.14) potentially containing only few aliphatic carbon atoms. This is related to more insensitive compound detection in the ESI negative FT-ICR-MS at m/z <200 Da corresponding to C12-14 O1 DBE 4 and 5 compounds. The C# distributions are generally Gaussian distributed with maximum abundances in the range C20 – C30 and outstanding abundances of C23 and C27 compounds. Facies is the dominating control on abundances but not on C# distributions. However, less mature oils are generally more abundant in O1 DBE 4 and 5 compounds than more mature marine oils in clastic reservoirs. Interestingly, oils in carbonate reservoirs contain a slightly higher portion of shorter chained aliphatic carbon. For the condensates at the western rim of the Central Graben, this phenomenon is more pronounced than for the eastern rim carbonate oils although the latter are more enriched in O1 compounds. The 2/5-11 oil is more strongly depleted in shorter chained DBE 4 and 5 moieties than 2/5-7 and 2/6-5. The similarity of DBE 4 and 5 compounds (Fig. 3.14a and b) could hint to similar migration and expulsion mechanisms or a structural relation of both compound classes, either caused by common natural origin in lignins (Freudenberg, 1966) or due to cyclisation reactions of longer chained alkylphenols (McClennen et

N1 DBE 15 Reservoir Lithology 100 0 carbonate al., 1983) similar to the proposed siliciclastic C29 Ts/(Ts+Tm) 0.19 - 0.25 mechanism in Poetz et al. (2014). 80 20 0.25 - 0.32 0.38 - 0.44 0.44 - 0.51 Therefore, we anticipate that 0.51 - 0.57 0.57 - 0.63 60 40 0.69 - 0.76 phenolic and indanolic compounds 0.76 - 0.82 with long aliphatic carbon numbers 40 60 (C25+) are less influenced by 20 80 migration-related processes and

0 reflect “true” maturity developments O1 10O1 100 80 60 40 20 0 DBE 4 DBE 5 (see as well Fig. 3.A.2 in the Fig. 3.15. Ternary plot of three DBE classes appendix). Thus, a ternary plot of correlating linearly with 29Ts/(29Ts+NH) maturity indicator. Using different compound classes both O1 DBE 4 and 5 with N1 DBE excludes the influence of rock-fluid interactions. 15 might be well applicable for maturity assessment (Fig. 3.15).

Besides carbazoles, phenols and its alkylated homologues have been identified as strong agents in crude oils, altering wetting properties of carrier and reservoir

102 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS rocks or interacting with non-hydrocarbon phases (Larter and Aplin, 1995; Bennett and Larter, 1997; Larter et al., 1997; Bennett et al., 2004; Bennett et al., 2007b).

Similarly to the N1 compounds, solubility, oil-water partitioning and solid-phase adsorption are controlled by the extent and position of alkylation to the functional group (e.g. Taylor, 1994). Partitioning coefficients of alkylphenols are negatively correlated to temperature, water salinity and the bulk NSO content of crude oils, whereas pressure has no influence (Bennett and Larter, 1997). Oil-water partitioning of phenols and cresols destabilizes the water film around minerals, facilitating apolar sorption on polar surfaces (Bennett et al., 2004) and thus removing these compounds from the migrating oil-phase. Thus, C0-1 phenols can rather occur in the water phase, whereas C2-3 phenols (Larter et al., 1997) and higher alkylated homologues partition within the oil phase. Due to higher subsurface temperatures during expulsion (T > 130 °C, Ziegs et al., 2017), higher alkylated homologues might still interact with the water phase even though water solubility decreases with increasing alkyl chain lengths. Although no change in the relative distributions of phenols was reported, Taylor et al. (1997) observed a systematic decrease in total C0-3 phenols in four North Sea oils with migration distance. This process occurs only until saturation in the carrier rocks is reached, because crude oil composition is no longer altered for later charges when using the same migration pathway. However, these conclusions were drawn from GC-based experiments which are sensitive for lower molecular weights (LMW), but might be extended to high-molecular weight (HMW) homologues measured with FT-ICR-MS in ESI negative mode. Within a DBE class, the chemical behaviour of LMW compounds is rather influenced by its polar, hydrophilic atom than the hydrophobic nature of aliphatic side chains. Consequently, if short chained homologues are strongly removed during migration in water-wet clastic carrier systems or in water- saturated, mixed-wet carbonate systems, long chained phenols relatively increase in abundance with increasing migration distance. This effect might be intensified for a water-saturated carrier system that is water-wet, rather than for an oil- /mixed-wet, water-filled carbonate rock. Observations from carbazole C# distributions are then confirmed: 1/9-1 carbonate oils have migrated shorter, whereas eastern carbonate oils migrated longer distances.

Carbon number distributions (Fig. 3.14) and low maturity correlations until

C27 (Fig. 3.A.2) likewise allow a different interpretation of compounds represented

103 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS by the O1 DBE 4 class. C23 and C27 compounds stand out from the natural CN distribution ranging between C14 and C45, and thus could represent other compounds than n-alkyl phenol homologues. Zhang et al. (2011) proposed isoprenoidal phenols and Shi et al. (2010) suggested cholesterol and

5,6-dihydrocholesterol as representing C27 and C28 compounds in Chinese crude oils from Cenozoic source rocks. Cholesterols are precursor structures of diasteranes which are ultimate products of clay-catalysed structural rearrangement via diasterenes (Peters et al., 2005) and are very abundant in the Mandal source rock and Central Graben crude oils. Diasteranes increase as a function of thermal maturity, thus its precursor structures must decrease in abundance which can be observed in Fig. 3.A.2 and Fig. 3.14 when excluding facies variations from the source rock (inset of Fig. 3.14). C23 compound do as well stand out from the natural distribution. Although no relationship to a distinct steroidal precursor structure has yet been identified, such a structure could be related to most saturated biomarkers containing 21 – 29 carbon atoms. However, we believe that the gradual, extremely smooth evolution of correlation coefficients R from LMW to HMW DBE 4 and 5 compounds generally represents a homologous series. However, variations for

C23, C25 and C27 compounds in maturity correlation of carbonate and clastic oils suggest other compounds with slightly different thermal stability than n-alkylphenols to dominate or influence constituents represented by these O1 DBE 4 carbon numbers.

3.6 Conclusion

The present study has investigated 24 crude oil and condensate samples from 16 petroleum accumulations in the Central Graben covering a range in oil qualities of 29 to 55 °API and a regional variety from central to marginal basin positions. Major reservoirs in the Central Graben are either Upper Cretaceous to Lower Palaeocene carbonate reservoirs having most likely been vertically charged, or clastic reservoirs of Devonian to Upper Jurassic age. The clastic reservoirs are in direct contact to Upper Jurassic source rocks.

API gravity depth profiles for oils in carbonate reservoirs are different to those in siliciclastic reservoirs. While API gravities of clastic oils increase with depth following depth trends described in the Viking Graben, carbonate oils

104 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS representing a broad API range are scattered across shallow depth intervals. Whole oils from both reservoir types follow the same compositional trend towards more saturate compositions as a function of maturity. The evaluation of hydrocarbon biomarkers confirmed that the oils are sourced from a homogeneous source rock containing dominantly marine organic matter, sometimes with locally higher oxygenation level and/or terrigenous contribution. They represent a maturity range from peak mature to overmature. Oils are not biodegraded but most marginal oils in clastic reservoirs show indications of oil-water interactions. Importantly, the low petroleum expulsion efficiency of the Mandal Formation (Ziegs et al., 2017) most likely brings about expulsion in one major charge into carrier and reservoir rocks. This is in contrast to source rocks which are efficient expellers, where multiple oil charges leave the source rock. In this case it can be anticipated that the composition of the earliest oil charge that is altered upon secondary migration is overprinted by the composition of later, more mature and less altered charges.

The polar fraction of crude oils is mainly composed of N1 and O1 compounds and although influenced by thermal maturity, basic oil families could be recognised on a regional basis. The thermal maturity assessment based on the carbazole distribution of crude oils in clastic reservoirs of the Viking Graben (Oldenburg et al., 2014) has been compared to several hydrocarbon maturity markers and correlates best with 29Ts/(29Ts+NH). When compared to hydrocarbon maturity biomarkers, its application is limited to clastic reservoir oils charged from a purely marine source rock. An apparent maturity retardation of oils in carbonate reservoirs is caused by a relative increase of N1 DBE 9 carbazoles and DBE 12 benzocarbazoles accompanied by a partial loss of short aliphatic chained substituents that are attached to the core structure. While the first observation can be related to subtle maturity variations, the latter is here interpreted as a migration effect. Compared to phenol-related, surface-active O1 DBE 4 and 5 compounds, benzocarbazoles in carbonate reservoirs show a similar relative depletion of their short-chained homologues as compared to oils in clastic reservoirs, as expressed in altered chain length distributions and maturity trends. This is likely a result of physicochemical interactions of petroleum, water and mineral surfaces in the carrier and reservoir systems and can be used to detect differences in carrier lithologies, migration routes and distances.

105 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS

3.7 Acknowledgements

This study is part of the Ph.D. thesis of Volker Ziegs at GFZ Potsdam as an Industry Partnership with Aker BP ASA. We are grateful for financial support and the permission to publish. Special thanks to Cornelia Karger and Anke Kaminsky (both GFZ Potsdam) for assistance during the lab work. We want to thank Drs. Maowen Li, Meijun Li and Hui Tian for their critical review having helped to refine the manuscript into its present shape.

3.8 Appendix

3.8.1 Maturity correlations: background information

The sample set of 16 crude oils measured using FT-ICR-MS contains

R² = 0.08 oils from different source input and R = -0.29 environments, but only oils originating from the principal anoxic R² = 0.61 R = -0.78 facies composed of marine algal

N1 DBE 12 C# 12 organic matter has been used for maturity correlations. As shown in Fig. 3.9, the compositional evolutions Fig. 3.A.1. Cross plot of compound abundance and biomarker ratio indicating thermal maturity of DBE 9, 12 and 15 compounds are serving for calculation of maturity correlations of N1 maturity dependent. Hereby, a DBE and O1 compounds; here exemplified for C12- alkylated benzocarbazoles from clastic and class combines compounds with carbonate rocks. The strengths of correlations are given by the determination factor R² and PEARSON short, intermediate and longer coefficient R. saturated chains attached to the core structure (Fig. 3.11). If a maturity influence of individual compounds cannot be observed, different mechanisms might act upon the distribution of relative abundances of these compounds.

Fig. 3.A.1 illustrates a cross plot of compound abundances (% TMIA) and 29Ts/(29Ts+NH) which is used as indicator of thermal maturity of North Sea crude oils and source rocks. Different developments as function of maturity for crude oils of different reservoir types are represented by correlation coefficients R. R indicates

106 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS both, strength and direction of a linear correlation (positive or negative) and has been applied to the y-axis of Fig. 3.12 and Fig. 3.A.2. The figure further illustrates that evolution must not be linearly as assumed for correlation studies, but may develop differently at different maturity levels. It further shows a relative enrichment (% TMIA) of C12-alkylated benzocarbazoles relative to the general inventory of acidic polar NSO compounds for oils in the eastern carbonate reservoirs as compared to clastic oils of the same maturity level. Contrary, the 29Ts/(29Ts+NH) ratio might overestimate their maturity but Ts/(Ts+Tm) ratios are in a similar range (Fig. 3.5).

3.8.2 Maturity correlations of O1 DBE 4 and 5 compounds

inherent inherent (a) clastic reservoirs very strong (b) carbonate reservoirs very strong

moderate moderate low low

very low very low

low low

moderate moderate very strong very strong inherent inherent

Fig. 3.A.2. Maturity correlation of relative abundances of individual O1 DBE 4 and 5 compound homologues. Correlation (a) represents these of oil suite produced from clastic intervals interbedded with Upper Jurassic OM-rich, marine black shales, and (b) represents oils from carbonate reservoirs that lie stratigraphically higher than the principal source rocks. 29Ts / (29Ts + NH) is used as maturity indicator.

Compound class distributions (Fig. 3.7) reveal that O1 compounds are generally decreasing with increasing maturity relative to N1 compounds. Having screened the maturity correlations for all dominant O1 DBE classes for the sub- sample sets, only compounds belonging to the DBE 4 and DBE 5 classes show clear trends with increasing carbon number which can be seen as an indicator for continuous homologous series represented by these DBE classes.

O1 compounds in oils produced from clastic and carbonate reservoirs behave similar to maturity correlations found in N1 DBE 12 compounds. While oils from clastic reservoirs show a strongly negative correlation to 29Ts/(29Ts+NH) for C27+ compounds, the carbonate-reservoired oils indicate a significant loss of LMW

107 3: UNRAVELLING MATURITY- AND MIGRATION-RELATED CARBAZOLE AND PHENOL DISTRIBUTIONS IN CENTRAL GRABEN CRUDE OILS compounds and thus a relative increase in HMW DBE 4 and 5 compounds (Fig. 3.A.2).

Assuming either phenolic or sterolic structures dominantly representing these DBE classes or single C#, the results can be explained coherently in both cases. The strong affinity of LMW phenolic compounds to interact with water-wet surfaces or water phases in migration pathways removes these compounds depending on the distance of migration. Even for short migration distances, clastic pathways need to be wettened making them available for migration of lipophilic compounds. Longer migration distances to eastern carbonate reservoirs containing slightly higher mature oils might result in negative correlations. Similar to steroidal acids representing the O2 DBE 5 class (Pan et al., 2013), sterolic homologues might represent particular compounds in the O1 DBE 4 class. Steroidal structures in crude oils ranging between C21 and C28 are not necessarily maturity dependent, but a chain length shortening might be possible explaining moderately negative correlations in this C# range. The similar reaction to thermal stress of DBE 4 and 5 classes could indicate similar or genetically related core structures and the continuous development with increasing C# to a homologous series of aliphatic chains attached to the same core structures. This might point to phenols and indanols representing DBE 4 and 5 classes.

108 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

4 Deeper insights into oxygen-containing compounds of the Mandal Formation, Central Graben, Norway

4.1 Abstract

Beside the kerogen composition, the amounts of generated bitumen play a major role when assessing the petroleum retention and expulsion behaviour of a source rock. High-molecular weight (HMW) products dominate the source rock extracts during early stages of generation in the TOC-rich, inefficiently expelling Mandal Formation. Such GC-unresolvable, bituminous compounds have not yet been structurally described. Based on 20 immature to peak-oil mature whole rock samples from different locations of the Central Graben, a compositional comparison of seven samples of different maturity stages is drawn to the excellently expelling Posidonia Shale, Germany of similar maturity. ESI negative FT-ICR-MS allows to investigate the acidic heteroelemental interior of the in-source retained petroleum.

Rather than the quantities of bitumen, its composition seems to be important for petroleum migration efficiency and fractionation. While Posidonia Shale extracts contain slightly higher proportions of NSO constituents than extracts of the Mandal Formation, they are dominated by lower polar nitrogen-compounds. Contrary, Mandal Formation extracts are strongly enriched in highly polar oxygen- containing (Ox)-compounds (O2 to O6) which are more aromatic but contain longer aliphatic chains than the Posidonia Shale samples, thus increasing their molecular size and the number of polar sites. In particular, it is the C16 and C18 aliphatic and C20 aromatic homologues occurring in all Ox classes which most likely represent fatty and aromatic acids linked with additional oxygen-functional groups. We suggest that these features might be specific to the Mandal Formation of the Central Graben and are related to complex interactions of geological/ palaeogeographic evolution, climate fluctuations and biological input in Upper Jurassic times. Consequently, the compositional features of petroleum generated from the Mandal Formation – highly polar, very aromatic but also strongly aliphatic – control the its physical properties and cause interaction with polar phases, such as the residual kerogen or clay minerals, and nonpolar phases in the source rock.

This chapter is reproduced as pre-print with permission from Energy & Fuels, submitted for publication. Unpublished work 2018 American Chemical Society; authored by Ziegs, V., Poetz, S., Horsfield, B., Hartwig, A., Rinna, J., Skeie, J.E.

109 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

4.2 Introduction

Amounts of petroleum available within a source rock-reservoir system are determined by quantities of generative organic matter in the source rock, its burial to depths of the principal oil window, the efficiency of expulsion and migration pathways as well as the presence and timing of geological processes causing subsidence and trap formation. The generation of low- to medium-molecular weight petroleum and higher-molecular weight bitumen from the macromolecular kerogen structure need to reach a certain threshold until generation pressure causes products to migrate within and finally from the source rock (Tissot and Welte, 1984). Constraints on expulsion from organic matter-rich, variably porous but low- to impermeable source rock into more permeable carrier beds are differences in pressure, temperature and/or relative concentrations of water and petroleum (Tissot and Welte, 1984; Mann et al., 1997). The quantitative retardation of expulsion after onset of generation has been attributed to the saturation of pores and generally amounts to 15 – 25 % of the net pore volume of the source rock (Welte, 1987; Okui and Waples, 1993; Mann, 1994). Pores in a source rock are either inter- and intra-matrix porosity (e.g. Han et al., 2017) or organic pores (e.g. Curtis et al., 2012; Loucks and Reed, 2014; Löhr et al., 2015).

During investigation of source rock extracts and expelled oils (Brenneman and Smith Jr, 1958), qualitative changes of petroleum compositions have been detected in a sequence of preferential retention: Asphaltenes > NSO compounds > aromatic hydrocarbons > aliphatic hydrocarbons, as well as cyclic > acyclic and iso- over n-paraffins (Baker, 1962; Lafargue et al., 1990; Sandvik et al., 1992). Having been observed in directly adjacent shale-sandstone sequences (Mackenzie et al., 1987; Leythaeuser et al., 1988b), this chemical fractionation has been attributed to primary migration and expulsion, rather than secondary migration effects. Although known for a long time, the precise mechanisms were poorly understood (Tissot and Pelet, 1971) and have been attributed to different solid and liquid phases present in a source rock: e.g. size exclusion in pelitic rocks (Krooss et al., 1991), sorption on mineral surfaces (Carlson and Chamberlain, 1986; Barrer, 1989; Brother et al., 1991; Taylor et al., 1997), partitioning with gas phase (gas washing) (Meulbroek et al., 1998; Losh and Cathles III, 2010; Bourdet et al., 2014), or partitioning with formation water (Lafargue and Barker, 1988; Zhang et al., 2005).

110 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

As a result of laboratory experiments reproducing the preferential expulsion sequence (Ritter, 2003a; Ertas et al., 2006; Kelemen et al., 2006a; Kelemen et al., 2006b) the place of petroleum formation and retention have been considered to be the same. Adsorption on the kerogen surface (Lamberson and Bustin, 1993), absorption in the kerogen structure (Young and McIver, 1977; Sandvik et al., 1992) or diffusion through the organic matter network (Thomas and Clouse, 1990c) were reported as causing selective fractionation. The efficiency is primarily controlled by TOC and kerogen composition (Sandvik et al., 1992). Fractionation processes are particularly important during early stages of petroleum generation when the expulsion pressure is not the main driving force of petroleum expulsion in a bulk fluid phase (Mann, 1994; Mann et al., 1997). Hereby, partitioning of kerogen and fluid phases is more important than diffusivity (Thomas and Clouse, 1990c). Generally, smaller particles move faster (Stainforth and Reinders, 1990) and more polar compounds, e.g. those with keto- and ether-functionalities, interact more intensively with surfaces and phases of similar physicochemical properties, such as carbonates, clay minerals or water (e.g. Thomas and Clouse, 1990b; Oldenburg et al., 2002).

The close relationship between generative organic matter (OM) and generated petroleum was quantitatively validated for the Mandal Formation (Ziegs et al., 2017). Applying Rock-Eval pyrolysis to unextracted and extracted source rocks, the solvent-extractable organic matter generated from kerogen (S2 peak of extracted samples) can be separated into the free, movable portion (S1 peak) that is enriched in aliphatic and aromatic hydrocarbons and a phase that is carried-over into the S2 peak that represents an immovable, medium- to high-molecular weight bituminous portion (S2bitumen) which is enriched in polar compounds (Jarvie, 2012, 2014). Being composed to >50 % of such bituminous constituents, the generated products of the Mandal Formation might exist in a single, mutually solved phase (Ziegs et al., 2017) whose movability is strongly restricted and thus retained in the source rock (Jarvie, 2014). These polar constituents cannot be compositionally resolved using GC-methods.

The present study aims to unravel the structure of retained heavy portions in the Mandal Formation using FT-ICR-MS in ESI negative mode that is able to characterize the medium- to high-molecular weight, acidic polar fraction of extractable organic matter. Contrasting the composition of the heterocompound

111 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY fraction with that of the efficiently expelling Posidonia Shale allows linking compound class-specific physical properties with overall retention characteristics of a good and bad expeller. The Lower Toarcian Posidonia Shale is well characterized (Littke et al., 1988; Mann and Müller, 1988; Rullkötter and Marzi, 1988; Littke et al., 1991) containing OM composition of marine origin and represents a natural maturity series encompassing immature, mature, and overmature zones with vitrinite reflectance (RO) values increasing from 0.48 to 1.45 %. Its acidic polar inventory was characterized in the context of maturity (Poetz et al., 2014) and retention characteristics (Mahlstedt et al., 2016).

4.3 Geological evolution of the relevant petroleum system elements

The Upper Jurassic Mandal Formation is considered to be the major source rock of petroleum accumulations in the Norwegian part of the Central Graben (Cornford, 1994). The NW-SE trending Central Graben as part of the North Sea rift system is constricted by the Coffee Soil Fault Complex in the NE and by the Mid

central marginal

112 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

North Sea High in the SW. This basin has developed as an extensional basin between Permo-Triassic and Upper Jurassic times followed by subsidence since the Lower Cretaceous. Detailed descriptions of its relatively simplistic geological evolution are published in the context of the entire North Sea rift system (Ziegler, 1990; Cornford, 1998), as well as within their differential evolution in the Norwegian (Gowers and Sæbøe, 1985; Gowers et al., 1993; Cornford, 1994; Faleide et al., 2010; Rossland et al., 2013) and Danish sections (Andsbjerg et al., 2001; Ineson et al., 2003; Møller and Rasmussen, 2003; Petersen et al., 2011; Petersen et al., 2013). Abbink et al. (2001) present some insights into the palaeogeography, dominated by local palaeo-highs, and impacts of climate on organic matter (OM) deposition in the Central Graben.

The potential source rocks were deposited during the highest subsidence of the main extensional rifting phase between Middle to uppermost Jurassic times. These clay-rich shale formations, the Haugesund, Farsund and Mandal Formations (Hamar et al., 1983), were deposited with up to 1200 m total thickness in the basin centre and <500 m at the basin margins (Cornford, 1994). In the Norwegian sector, only the uppermost Ryazanian Mandal Formation is referred to as ‘Hot Shale’ due to the elevated gamma-ray log signatures, a result of higher radiogenic element concentrations (Uranium, Thorium and Potassium) contained in the organic matter, and due to its unusually high portion of organic carbon composed of bacterially degraded algal debris forming a Type II kerogen (Doré et al., 1985; Cornford, 1994; Dybkjaer, 1998; Andsbjerg et al., 2001). Containing in average 5.0 to 5.5 % TOC, but also up to 12 % in some areas, and Hydrogen Indices up to 650 mg HC/g TOC, source rock richness and generation potentials of the Mandal Formation are uniformly broadly distributed generating Paraffinic-Naphthenic- Aromatic oil at all maturity stages (own database, Cornford, 1998 and references therein; Ziegs et al., 2017). This variance is mainly an effect of maturity but may be due to variations in the depositional environment. Tyson et al. (1979) and Herbin et al. (1993) explained such variations in Kimmeridge Clay Formation outcrops in southern and northern England with different levels of oxygen deficiency during deposition of the same planktonic organisms that yield a Type II kerogen. However, at the flanks of the Central Graben near the paleo-shore line or near intrabasinal highs (e.g. the Mandal High, Fig. 4.1) reduced oxygen levels may be a result of increased land plant input and its partial oxidation (Cornford et al., 1980)

113 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY contributing a vitrinitic kerogen component. In contrast, the Farsund Formation contains mainly mixed Type II/III kerogen generating gas and oil and is sealed by the extremely clay-rich, impermeable Mandal Formation at most locations of the Norwegian Central Graben. Only at basin margins where the Mandal Formation is condensed or absent, the Farsund Formation has charged some petroleum accumulations, e.g. the 2/2-1 discovery (Gabrielsen et al., 1985; Pedersen et al., 2006).

The Mandal Formation is buried as deep as 4000 m in the axis and ~2500 m at the flanks of the Central Graben, and thus is within the early to peak oil window. Thermal maturity is determined by the thickness of the overburden rock. After connecting the North Sea with fresh water influx from the North at the Jurassic/Cretaceous transition, impermeable OM-poor (<1 – 2 % TOC) calcareous clays, marls, limestones and chalks were deposited during Cretaceous to Lower Palaeocene, and thick, clastic sediments in Cenozoic times. While the Upper Cretaceous chalks constitute the major reservoirs in the Central Graben (Cornford, 1994), the Cenozoic overburden has been the principal driver of thermal maturation, thus source rock intervals have experienced a fast subsidence to oil window temperatures (Nielsen et al., 1986). Due to low permeability of Lower Cretaceous shales, secondary migration pathways in the Central Graben are mainly fault-related migrating up-flank rotated fault blocks or in marginal strike- slip faults, and also along micro-fractures related to Zechstein salt movement (Cornford, 1994, 1998). Intercalation or juxtaposition of source and sandstone reservoir intervals may occur at the graben flanks (Curtin and Ballestad, 1986).

114 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

4.4 Samples & methods

4.4.1 The sample set

The basis of this study is a sample set representing natural maturity series’ of the Mandal Formation with different starting generation potentials and with samples originating from marginal and central basin positions within the Central Graben and different stratigraphic positions within the Mandal source rock. The maturity series’ are composed of 21 cutting and 3 core samples and were selected from a comprehensive data base containing 222 bulk chemical data of the Mandal Formation which was provided by the project partner Aker BP ASA. Organic richness, organic matter quality and maturation characteristics were the principle criteria for sample selection. The same samples were investigated in an earlier study addressing the controls of organic matter on bulk retention capacity of the Mandal Formation applying Rock-Eval pyrolysis and pyrolysis-GC-FID on unextracted and extracted source rock samples (Ziegs et al., 2017).

The distinction between kerogen types “II” and “II/III” in this and the previous study was made based on bulk and pyrolytic data, such as Hydrogen Indices, aliphatic chain length distributions and aromaticity of pyrolysates (Larter, 1984; Horsfield, 1989; Eglinton et al., 1990; Horsfield, 1997). Although the OM quality can be described as heterogeneous, Herbin et al. (1993) pointed out that its origin is homogenously algal-planktonic throughout the basin. Thus, heterogeneities might be attributed to variations in the Eh and pH conditions during deposition. Minuscule variations in aliphatic chain length distributions at particular maturity levels and possibly in aromaticity (Ziegs et al., 2017) may confirm this hypothesis.

Having conducted bulk and compositional pyrolysis measurements on 24 samples in the sample set, compositional investigations using both, GC-MS and FT-ICR-MS were conducted concurrently only on 20 out of 24 samples covering all boreholes represented by the sample set described in Ziegs et al. (2017). They represent different basin positions and generation potentials of the Mandal Formation at varying maturity levels. An overview of the respective samples is presented in Table 4.1.

115 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

A maturity series of the Toarcian Posidonia Shale, Northern German Basin, with an initial generation potential of 663 mg HC/g TOC that ranges from immature to overmature (423 – 457 °C Tmax) serves as an example of an efficiently expelling source rock (Horsfield et al., 2010). The geochemical source rock characteristics of immature to early mature sections of Posidonia Shale and Mandal Formation are comparable. Both source rocks contain marine OM characterized as Type II kerogen with variably low aromaticity (Larter, 1984; Eglinton et al., 1990) at the same maturity levels, and generate Paraffinic- Naphthenic-Aliphatic (PNA) Low Wax oil (Horsfield, 1989, 1997) until reaching peak oil window. Detailed characteristics of the particular Posidonia Shale samples are described in literature (Wilkes et al., 1998; Bernard et al., 2012a; Poetz et al., 2014; Mahlstedt et al., 2016) having characterized the kerogen composition as well as the hydrocarbon and polar fractions of extractable OM.

4.4.2 Analytical methods

Generated and retained petroleum in the Mandal Formation was quantified by applying solvent extraction for 24 hours on whole rock samples using dichloromethane and 1 vol.-% methanol. The acidic polar inventory of the extractable organic matter (EOM) was measured on whole extracts using FT-ICR- MS in ESI negative mode. The method is described in Mahlstedt et al. (2016) and Ziegs et al. (2018), chapter 3. Prior to GC-analyses, whole oils and extracts were separated into aliphatic, aromatic and polar fractions using an automated medium pressure liquid chromatography (MPLC) procedure (Radke et al., 1980).

The aromatic hydrocarbon fractions from MPLC separation were analysed using GC-MS to evaluate distributions of alkylated phenanthrenes and dibenzothiophenes. Analysis was performed using a DSQ spectrometer coupled to a Trace GC Ultra chromatograph. The latter was equipped with a BPX5 fused silica column (50 m · 0.22 mm i.d; film thickness, 0.25 µm). The injector temperature was set to 50 °C and increased to 300 °C at 600 °C/min (held 10 min). The oven temperature was increased from 50 °C (1 min isothermal) to 310 °C at 3 °C/ min (held 30 min). Helium was the carrier gas at a constant flow of 1 ml/min. The mass spectrometer was operated in the electron ionization (EI) mode at 70 eV and 230 °C source temperature. Full scan mass spectra were recorded over a range of m/z 100

116 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

– 330 at 1 s per decade, an inter-scan time of 0.2 s, resulting in a scan cycle time of 0.719 s. Quantification was performed using m/z 176 and m/z 178 for phenanthrenes, m/z 184 for dibenzothiophene.

GC-MS-MS was used to evaluate the saturated biomarkers. Analysis was performed using a Finnigan MAT 95XL spectrometer coupled to a HP 6890A chromatograph which is equipped with the described BPX5 column. The PVT splitless injector was set to 52 °C (1 min isothermal) and increased to 300 °C at 720 °C/min (10 min isothermal). The oven temperature program is the same as described above. Biomarkers were recorded in the metastable reaction monitoring (MRM) mode with a dwell time of 21 ms and an inter-dwell time of 20 ms per metastable transition, resulting in a scan cycle time of 0.984 s for the detection of 24 metastable transitions. Hopanes and steranes were evaluated using the metastable transitions m/z 370, 398, 412, 426 → m/z 191 and m/z 372, 386, 400, 414 → m/z 217, respectively.

117 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

- - x x x x x x x x x x x x x x x x x x GC

- - - - - x x x x x x x x x x x x x x x

- ex haracteristics 48.12 44.03 52.51 58.09 80.05 78.45 47.47 73.76 93.85 42.68 51.33 97.24 150.09 202.20 128.90 116.71 260.09 133.67 140.04 mg/g S2 mg/g Total Oil Retention

ex Gas 0.35 2.28 1.64 0.86 0.24 0.80 1.96 0.37 0.69 0.67 0.21 0.20 1.46 1.22 0.15 2.34 1.89 0.67 0.14 2.08 mg/g S2 mg/g

Retention

= mg/g

S2ex 1.32 7.49 7.93 6.18 3.66 1.11 6.96 4.79 1.35 0.40 9.51 1.77 4.14 4.08 8.87 2.75 - - 11.04 10.26 15.89 15.11 bitumen S2wr S2

- - % - Retained Petroleum wt. NSO 37.22 37.66 48.94 39.70 28.29 47.42 25.90 25.03 51.39 48.26 29.43 49.18 42.92 32.61 27.17 34.57 19.94 43.43

content

TOC mg/g mg/g OSI 78.49 19.42 29.12 32.48 41.64 65.88 49.73 88.41 89.33 13.81 34.33 28.18 72.35 84.42 79.73 18.73 12.14 63.18 49.69 34.90

PI 0.36 0.11 0.10 0.09 0.22 0.43 0.25 0.37 0.48 0.08 0.09 0.11 0.33 0.22 0.22 0.12 0.05 0.21 0.31 0.37 a/(a+b)

II II II II II II II II II II II II OM II/III II/III II/III II/III II/III II/III II/III II/III Type

6 ex 17 53 49 24 48 25 26 42 55 19 38 16 19 22 48 52 14 25 16 TOC mg/g mg/g OI

ex 85 72 TOC mg/g mg/g 179 188 294 402 189 114 214 182 127 172 423 288 193 383 356 150 191 253 HI

°C max 440 437 432 438 437 439 441 449 449 437 432 432 441 430 430 436 436 441 439 445 T

ex mg/g S3 1.13 1.04 2.13 1.28 1.02 1.30 1.01 0.37 0.72 1.02 0.74 1.36 1.00 1.51 1.85 1.27 0.91 1.11 1.07 1.14

ex Kerogen Characterization mg/g S2 3.72 5.92 8.27 2.18 3.18 3.96 3.34 3.58 5.02 3.98 11.66 10.59 16.53 10.38 12.06 30.82 30.54 20.50 12.91 21.49

wr mg/g 6.46 0.47 4.48 2.73 6.18 2.01 0.29 1.72 1.24 5.86 8.78 8.85 0.53 0.17 5.61 1.59 2.89 1.43 2.16 1.12 S1

% 6.51 1.98 5.19 3.86 5.83 1.72 1.85 3.91 3.60 6.26 8.05 8.59 2.64 1.75 8.11 4.23 6.95 4.39 5.35 2.11 TOC

Cretaceous High

Steinbit Terrace Steinbit Terrace Steinbit Terrace Steinbit Terrace HighMandal HighMandal HighMandal Cretaceous High Cretaceous High Cretaceous High Søgne Basin Hidra High Hidra High Hidra High Hidra High Location HighMandal Søgne Basin Hidra High Cretaceous High

Selected sample set of Formation the Mandal in the Central Graben containing information on sampled wells, bulk sourcerock c

. 3887 m 4070 3795 4020 4028 3410 3415 3420 3884 3775 3780 3500 3245 3658 4025 4420 Depth 3425 3438 3242 3593 1 . Well information Well Eval) and retained petroleum. 4

- 2 2 7 7 5 7

- - - - - 3 3 9 9 3 3 3 3 2 3 3 3 2 15 ------

Well 1/3 2/1 2/1 2/1 2/6 2/6 2/6 2/6 - 2/10 2/10 - 2/11 2/11 3/7 - 3/7 7/7 - 7/7 7/8 7/11 7/11 2/7 -

Table (Rock

118 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

4.5 Results

4.5.1 Environmental information from the hydrocarbon fraction

Table 4.2. GC-FID and GC-MS data of Mandal Formation extracts characterizing the organic matter input and depositional setting.

well 2/11-7 2/1-3 2/1-9 2/1-9 2/6-3 7/11-5 7/11-7 7/7-2 7/7-2 depth (m) 3780 3795 4020 4028 3425 4025 4420 3242 3445

Tmax (°C) 430 441 449 449 432 439 445 438 436

Pr/n-C17 0.75 0.72 0.51 0.50 2.33 0.42 0.29 1.35 0.70

Ph/n-C18 0.68 0.58 0.33 0.32 3.43 0.34 0.26 0.81 0.37 Pr/Ph 1.35 1.76 1.97 1.79 1.18 1.71 1.44 1.89 3.36 CPI-1 1.02 1.06 1.03 1.05 1.05 1.06 0.91 0.94 - TAR 0.22 0.03 0.05 0.06 0.08 0.07 0.00 0.06 0.00 DBT/PHEN 0.19 0.11 - 0.01 0.15 0.02 - 0.10 -

T23/(T23+H) 0.05 0.00 - 0.00 0.05 - - 0.06 - %BNH 4.70 6.31 - - 4.50 11.03 - 21.50 - NH/H 0.51 0.26 0.24 0.24 0.39 0.48 - 0.31 - DH/29Ts 0.27 1.27 1.36 2.03 0.33 2.36 3.87 0.64 -

C31 22R/H 0.38 0.42 0.66 0.61 0.42 0.55 - 0.43 - Gam/(Gam+H) 0.05 0.05 0.00 0.00 0.13 - 0.21 0.04 0.00

C35/(C34+C35) 0.37 0.49 0.50 0.43 0.54 - - 0.52 - DSt/(DSt+rSt) 0.23 0.73 0.85 0.86 0.30 0.82 - 0.67 -

%St30 0.11 0.12 0.13 0.13 0.09 0.11 - 0.11 -

%DSt30 0.06 0.09 0.09 0.11 0.07 0.10 0.10 0.08 -

Pr GC-FID Pristane Ph GC-FID Phytane

CPI-3 GC-FID (n-C27+6∙n-C29+n-C31)/ (n-C +n-C ) 28 30 TAR GC-FID (n-C27+n-C29+n-C31)/ (n-C +n-C +n-C ) 15 17 19 DBT m/z 184 Dibenzothiophene Phen m/z 178 Phenantrene T23 m/z 191 C23 tricyclic terpane BNH m/z 191 28,30-Bisnorhopane H m/z 191 C30 hopane %BNH m/z 191 100*BNH/(BNH+H) NH m/z 191 nor-hopane

DH m/z 191 diahopane, C30* 29Ts m/z 191 18α-nor-Neohopane Gam m/z 191 Gammacerane

C34 / C35 m/z 191 C34/C35 (S+R) homohopanes

DSt m/z 217 C27-29 βα-steranes

rSt m/z 217 C27-29 αβ-steranes

%St30 m/z 217 C30/C27-30 αβ-steranes

%DSt30 m/z 217 C30/C27-30 βα-steranes

119 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Biomarker data shown in in Table 4.2 suggest that the extractable organic matter in the Mandal Formation originates from a purely marine environment

(C31 22R) in which marine algal organisms have been deposited as indicated by elevated C30 steranes. Elevated diasterane and diahopane ratios at immature and early mature levels confirm the petroleum origin from shaly source rocks. Low to absent tricyclic terpanes indicate more or less marine fresh water environment, but varying Pr/Ph values are a sign of varying environmental conditions. These can be attributed to variations in anoxic conditions (C35/C34 homohopanes) which are not due to salinity-induced stratification (low gammaceranes) but rather to varying terrestrial OM input (preference of odd carbon numbers and elevated C21-35 alkanes in immature samples) locally creating anoxic zones due to increased bioproductivity. However, C27-29 sterane and diasterane distributions as well as

Pr/n-C17 vs. Ph/n-C18 values do not yield considerable variations and suggest mixed marine and terrestrial OM input.

Elevated C35 homohopanes (C35/C34 > 1.0) in basin-marginal wells 2/6-3, 7/7-2 and 7/8-3 do not indicate locally carbonatic source rocks. Low NH/H, high diasteranes, low gammaceranes as well as high Pr/Ph ratios point to elevated anoxic conditions, thus following Peters et al. (2005)’s suggestion for interpretation of homohopane distributions.

4.5.2 Gross composition of source rock extracts

The gross composition of extractable OM (EOM) retained in the Mandal Formation and Posidonia Shale is shown in Fig. 4.2. Immature and earliest mature sections are dominated by NSO’s, and is similar to those of Type II Kimmeridge Clay and Type I Green River Shale pyrolysates (cf. Horsfield, 1997), but as well to immature and early mature Posidonia samples. Upon further maturation, the EOM compositions of Mandal Formation samples become depleted in NSO’s and develop to higher aromatic portions as a function of thermal maturity, the peak and overmature Posidonia Shale samples, on the other hand, exhibit a more aliphatic composition. Interestingly, the saturate content in the Mandal Formation remains relatively stable with increasing maturity, although generated petroleum becomes enriched in aliphatic hydrocarbons at higher maturities (Tissot and Welte, 1984) as can be observed in the Posidonia maturity series. Regional variations within the

120 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Kerogen Type Kerogen Type Mandal Formation sample set II 100 0 II-A II/III %Ro Tmax (°C) 0.00 - 0.55 cannot be observed. Type II/III 400 - 432 0.55 - 0.70 432 - 439 80 20 0.70 - 0.80 439 - 445 0.80 - 0.90 samples contain more aromatic 445 - 450 NSO (wt.-%) 0.90 - 1.50 portions and a lower NSO content 60 40 than Type II samples of comparable

40 60 maturity. The highest mature Aromatics (wt.-%) Posidonia Shale sample (1.45 %RO) 20 80 exemplifies the compositional trend 0 100 expected for late to overmature 100 80 60 40 20 0 Saturates (wt.-%) Mandal Formation samples although

Fig. 4.2. Chemical gross composition of the compositional evolution in both maltene fraction of extracts from the Mandal Formation (squares for Type II and triangles for formations is different. Type II/III) and Posidonia Shale (circles) illustrated as a function of maturity, Rock-Eval Tmax and %RO.

4.5.3 Elemental class distribution

Kerogen Type 100 0 As pointed out in great detail II II/III II-A by Poetz et al. (2014) and Mahlstedt Tmax (°C) 400 - 432 80 20 432 - 439 et al. (2016), the acidic polar 439 - 445 445 - 450 450 - 470 inventory of Posidonia Shale extracts 470 - 600 60 40 is dominated by constituents

containing 1 nitrogen atom in their 40 60 heterocyclic structure (N1

20 80 compounds) increasing as a function of maturity (Fig. 4.3), while oxygen-

0 100 containing compounds with 1 to 4 O 100 80 60 40 20 0

N1Ox (% TMIA) atoms (O1-4) and N1Ox compounds are Fig. 4.3. Distribution of major compound groups of less abundant in matured Posidonia the Mandal Formation (Type II and II/III) and Posidonia Shale (Type II-A). Shale samples. Representing relative abundances (in % Total Mono- Isotopic Abundances) within a maturity series, only the most immature Posidonia

Shale sample contains elevated (relative) amounts of O2 components while N1, O1 and N1O1 are similar in abundance (Fig. 4.4c). The intensities of N1 and N1O1

121 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY compounds relatively increase while O1-4 intensities generally decrease as a function of maturity. With increasing number of heteroatoms abundances of individual compounds decrease significantly, particularly O3+ and N1O2+.

w 383 409 402 388 388 395 379 376 387 379 378 405 401 379 377 398 368 408 387 412 420 402 392 M total

365 M 385 380 368 368 376 359 355 365 359 359 382 379 361 357 377 350 383 367 386 393 381 374 total

MS.

- 4 O ICR 1 - 0.87 1.25 1.53 1.72 1.70 0.16 0.03 0.19 0.93 1.59 1.72 1.68 1.66 0.51 0.34 0.05 3.61 2.70 0.15 0.30 0.06 0.07 N

3 ) FT - O 1 3.46 4.52 4.45 5.04 1.65 5.64 2.27 2.03 2.56 3.10 4.63 4.49 4.17 4.76 2.14 2.55 0.85 5.98 6.15 2.36 3.29 1.70 0.42 N

2 O 1 7.91 8.34 7.99 6.06 7.86 6.87 7.42 6.56 7.82 7.56 8.23 5.39 7.37 2.83 8.83 6.86 8.86 6.47 1.86 N 10.06 11.61 10.01 11.43

1 O 1 8.94 7.05 8.76 6.67 7.61 7.16 9.41 7.22 4.34 8.78 9.78 N 12.37 11.58 10.18 16.24 14.64 17.96 13.39 11.22 15.52 13.27 16.08 23.07

S 1 0.21 0.06 0.07 0.06 2.45 0.14 0.19 0.17 0.10 0.01 0.02 0.02 0.02 0.01 0.25 0.22 1.61 0.03 0.20 3.94 3.28 5.40 5.27 N

2 (% TMIA) N

1.61 0.42 2.15 1.72 2.23 3.21 0.46 0.23 0.07 1.39 0.89 0.97 1.34 1.69 1.78 0.94 1.86 0.04 1.15 1.31 1.18 1.71 2.63

1 N 8.64 8.06 6.36 6.80 4.52 4.20 4.99 4.99 7.90 1.87 9.23 10.54 12.57 15.12 10.33 13.93 10.21 11.63 16.87 36.14 30.69 46.57 62.02

------

6 O 0.51 0.31 1.02 0.53 0.81 0.42 0.82 0.70 2.77 0.71 2.20 0.10 3.55 0.28 0.01 0.01 1.10

5 O 2.63 2.39 3.80 1.51 3.07 0.77 0.72 1.01 2.04 3.53 3.21 5.30 3.44 5.80 0.79 5.98 1.95 0.35 0.08 0.03 0.01 0.10 3.66 relative ion abundance ion relative

4 O 7.55 7.67 9.44 3.66 8.01 5.35 6.41 7.25 9.61 9.05 8.97 4.94 7.38 2.15 1.07 1.36 0.16 0.32 9.25 11.41 10.73 13.45 10.13

3 O 6.01 6.52 1.55 1.38 16.23 16.97 16.44 10.54 14.49 15.42 17.80 19.60 19.11 17.81 18.91 19.38 17.90 19.18 16.75 17.20 15.85 10.56 16.73

2 O 5.07 8.25 20.84 22.96 20.10 19.41 15.00 22.65 28.39 28.94 23.65 24.86 24.99 23.51 24.29 15.55 22.42 19.86 20.77 17.82 15.67 16.26 18.05

1 O 9.93 8.23 7.74 5.41 5.42 8.15 9.52 6.54 6.44 6.84 4.94 7.32 3.81 7.06 3.75 3.49 6.87 6.09 5.64 1.38 1.67 6.94 10.33

(°C)

%R 430 432 438 441 441 449 448 449 436 436 437 437 437 439 439 445 441 0.48 0.53 0.68 0.73 0.88 1.45 / max

T Relative abundances ofmajor NSO compound classes all samples investigated by ESI (

7 2 5 7

- - - - 3 2 3 9 9 9 3 2 3 15 3 3 3

------

well 4

2/11 2/6 7/7 2/10 2/1 7/8 - 2/1 2/1 2/1 3/7 7/7 3/7 2/7 2/6 7/11 1/3 7/11 WE WI DIEL DO HAR HAD

II Type Mandal II/III Type Mandal Shale Posidonia

Table

122 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

(a) Mandal Type II (b) Mandal (c) Posidonia Type II/III (% TMIA)

rel. Intensity rel.

In contrast, EOM composition of the Mandal Formation is strongly dominated by oxygen-containing compounds throughout all maturity stages. Exact data of 7 Mandal Formation extracts is given in Fig. 4.4 and Table 4.3, while molecular parameters of Posidonia Shale extracts can be obtained from Poetz et al. (2014).

Interestingly, the clearly positive evolution of N1 and subsequently negative evolution of Ox compound abundances as observed in the Posidonia Shale extracts cannot be identified within the Mandal Formation (Fig. 4.3). While the Posidonia

Shale becomes depleted in other elemental classes than N1, the Mandal Formation is compositionally very similar at all maturity stages resembling the composition of immature Posidonia extracts. This confirms observations from pyrolysate compositions that show a bulk compositional uniformity until peak oil window (Ziegs et al., 2017). Observing major compound classes,

ternary plots of O2 vs. O3 vs. N1 or

N1O1 compound classes show the very same maturity evolution as elemental classes.

Looking into the detailed compound classes of 7 Mandal Fig. 4.5. Relative distributions of O1 to O5 Formation samples at similar compounds within the Ox class of extracts from the Mandal Formation and Posidonia Shale that maturity level as the Posidonia Shale illustrate the dominance of O3+ compounds in the Mandal Formation. maturity series, N1 compound

123 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY abundances slightly decrease from immature to early mature Mandal Formation samples (Fig. 4.4a) but increase until entering the peak oil window (Fig. 4.4a and b). A contrary negative evolution can be observed for O2 compounds (Fig. 4.4). O2 and O3 compounds clearly dominate the polar inventory of the Mandal Formation extracts, but O4 and O5 compounds, representing a mixture of multiple functional groups such as carboxylic acids, aldehyds, alcohols, etc., are clearly more abundant than in Posidonia Shale extracts (5 – 14 % and 4 – 6 % TMIA, respectively). A star plot of all Ox compound classes (Fig. 4.5) illustrates the significantly higher relative contributions of O3, O4 and O5 classes in the Mandal Formation extracts, while O6 compounds are practically not present in the Posidonia Shale extracts. Relative abundances of N1Ox compounds and their evolution as function of maturity are very similar for Mandal Formation and Posidonia Shale extracts, ranging up to

17 % TMIA with a general increase for N1O1 and decrease for N1O2 and N1O3 compounds.

4.5.4 Characterization of selected compound classes

In the following, O2 to O4 compound classes are described in more detail since these classes were found to account for the major compositional differences between the efficiently expelling Posidonia Shale and the locally bad expelling Mandal Formation. Colour coding follows kerogen type, differentiating between Posidonia Type II-A (according to Delvaux et al., 1990) in black, Mandal Type II (greenish) and Mandal Type II/III (reddish). The number of rings plus double bonds is usually expressed by a double bond equivalent value, DBE (Hughey et al., 2007). DBE indicates hydrogen deficiency of a given molecular formula and is commonly used to investigate the high-resolution mass spectra of crude oil (Bae et al., 2010).

Poetz et al. (2014) and Mahlstedt et al. (2016) have illustrated a strong control of thermal maturity on N1 and N1O1 compound classes and their DBE distributions in source extracts and pyrolysate extracts of the Posidonia Shale representing ring addition to initial core structures. DBE distributions of these compound classes extracts of the Mandal Formation fit with data presented by the above authors for similarly mature Posidonia Shale samples (not presented here). Consequently, these compound classes are not available for evaluating different retention behaviour. Furthermore, the DBE distribution of the O1 class was also

124 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY shown to be strongly maturity-dependent in crude oils (Ziegs et al., 2018), thus they are not described in the context of variations in retention capacities.

With increasing number of heteroatoms as well as DBE classes, possible combinations of cyclic structures and double bonds in rings and side chains as well as the position of the heteroatom increase exponentially. Consequently, the assignment of distinct chemical compounds or functional groups based on sum formulas becomes rather speculative than precise. Therefore, descriptions in the following paragraphs are tied to DBE classes and carbon numbers (C#’s), while the functionality and thus their distinct properties of particular compounds are addressed in the discussion part and extrapolated to general trends within an elemental or compound class.

O2 compounds constitute the

(a) O2: immature major compound group in Mandal (% TMIA) Formation source extracts and are the most abundant oxygen compound class

rel. Intensity in Posidonia extracts (Fig. 4.4). In both

formations, the O2 class is mainly

(b) O : early mature 2 composed of the DBE 1 class and, as a

(% TMIA) second mode, of DBE 5 to 10 compounds

phasing out until DBE 20 (Fig. 4.6). DBE 1 most likely represents saturated rel. Intensity n-fatty acids and could be influenced by

sample storage (Peters et al., 2005, and references therein, p. 701). However, rel. Intensity naphthenic acids, represented by DBE classes 2 – 4 and indicating

Fig. 4.6. DBE class distribution of the O2 compound class comparing source rock biodegradation, are not abundant in the extracts of the Mandal Formation and Posidonia samples. DBE 5+ classes represent Shale at three maturity stages. unsaturated aliphatic as well as aromatic acids (Peters et al., 2005) and are more abundant at comparable maturity stages in the Mandal Formation samples than in the Posidonia Shale samples (compare Fig. 4.6a, b and c).

125 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

(a) O2 DBE 1: immature (d) O2 DBE 7: immature (% TMIA)

rel. Intensity

(b) O2 DBE 1: (e) O2 DBE 7: early mature early mature (% TMIA)

rel. Intensity

Fig. 4.7. Carbon number distributions of selected O2 DBE classes addressing detailed compositional differences of the Mandal Formation and Posidonia Shale extracts at three maturity stages.

The Mandal Formation is extremely variable in the DBE 1 contribution and to a lower degree within the DBE 5+ fraction. Generally, Mandal Type II samples contain lower abundances of fatty acids (DBE 1) similarly to Posidonia samples, while Type II/III samples as well as the marginal well 2/6-3 show a relative enrichment. A relative depletion in n-fatty acids results in relative enrichment, thus higher TMIA’s, of higher DBE classes, the aromatic acids, which is nicely exemplified in mature sample 2/1-9. Within the aromatic acids of the Mandal Formation sample set, interesting compositional variances occur. While Type II samples are generally more enriched in aromatic acids (e.g. wells 2/11-7, 7/7-2 and 2/1-3), the DBE 5+ fraction in Type II/III samples (7/11-5, 7/11-7) and the marginal well 2/6-3 is dominated by the DBE 7 class. In Posidonia Shale extracts, the DBE 5+ fraction is less abundant and does not show a dominance of a particular DBE class. Thus, the DBE 7 enrichment might be a feature of OM type and/or regional variance.

126 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

A carbon number (C#) distribution of the O2 DBE 1 class (Fig. 4.7a–c) reveals that the fatty acids are dominated by C16 and C18 fatty acids pointing to microbial activity or contamination as mentioned above. Higher alkylated homologues are more abundant in the Mandal Formation than in Posidonia Shale. An even-to-odd predominance for the C20-30 fatty acids can be observed in Mandal Formation samples which are stronger in Type II/III and the marginal 2/6-3 well than in Type II samples but does not occur in Posidonia Shale samples. Such compounds may be precursor structures of even-numbered n-alkanes (Peters et al., 2005) stemming from terrigenous OM (i.e. plant cuticular waxes, Tegelaar et al., 1989b). A lack of C# predominance is an indicator of predominantly marine OM input (Peters et al., 2005) and/or higher maturity such as in mature well 2/1-9. An interesting feature is the occurrence of C12 and C14 fatty acids in the early mature Type II/III sample (7/11-5).

The DBE 7 compounds are more abundant in the Mandal Fm samples than in the Posidonia samples and C# distributions within the O2 DBE 7 class (Fig. 4.7d–f) illustrate broad Gaussian distributions. In Mandal Fm extracts, apices of these C# distributions are at higher C#’s than in Posidonia Shale samples, thus representing longer aliphatic chains attached to potential core structures. These variations in DBE 7 abundances between both formations increase as function of maturity. A strong dominance of C20 compounds in the Mandal Formation samples can be observed, particularly for the marginal Type II sample 2/6-3 (0.8 % TMIA), and Type II/III OM containing samples of wells 7/11-5 and 7/11-7 with up to 4.5 and 0.85 % TMIA, respectively (Fig. 4.7e–f). This compositional feature does as well occur to a lower degree in Posidonia samples and thus is most likely not a result of contamination.

O3 compounds are the second most abundant compound class in the Mandal

Formation extracts and abundant in immature Posidonia extracts (Fig. 4.4). O3 compounds in the Mandal Formation show variably elevated contributions of DBE classes 1, 2 and 6 – 9 phasing out until DBE 20/21 (Fig. 4.8). While the Posidonia Shale shows interesting variances at DBE classes 6, 7 and 10, the Mandal Formation extracts are extremely variable in their contributions of DBE classes 2 and 8. The increase of 1 DBE in comparison to distribution patterns of the O2 class (cf. Fig. 4.6) suggests a combination of carboxylic acids and an ether-bound oxygen atom. However and similar to the O2 compound class, the Mandal Formation

127 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

samples contain significantly higher (a) O3: immature contributions in all DBE classes, and (% TMIA) the dominance in the mentioned DBE classes 2, 5 – 9 and especially DBE 8, is

rel. Intensity very pronounced.

(b) O3: early mature The O3 DBE 2 compounds are less abundant in Type II samples, but (% TMIA)

abundant or very abundant in Type II/III samples and the well 2/6-3 which rel. Intensity is at the margin to the palaeo-high of

the Mandal High (Fig. 4.1). Only the (c) O3: mature immature Wickensen well shows a (% TMIA) similar contribution of DBE 2 compounds as Mandal Formation

rel. Intensity samples of all maturity ranges, while more mature Posidonia Shale samples

Fig. 4.8. DBE class distribution of the O3 are completely depleted in that DBE compound class comparing source rock extracts of the Mandal Formation and Posidonia class. The O3 DBE 8 class is more Shale at three maturity stages. enriched than other abundant aromatic

DBE classes (DBE 6 – 9). The dominance of the O3 DBE 8 is a regional feature which might be linked to OM contribution but not necessarily to generation potential which was the characteristic to “kerogen typing”: extracts from wells 7/11- 5 and 7/11-7 containing Type II/III OM and the 2/6-3 well are strongly enriched in DBE class 8. Although quite close to each other, the 2/1-3 and 2/1-9 wells (Fig. 4.8a and c) are again very different which might not necessarily attributed to maturity as Ox compounds do not show maturity trends in the Mandal Formation. Interestingly and different to the DBE 2 class, distribution patterns of the DBE 6+ fraction do not change within the same well (cf. Type II and II/III samples in well 7/7-2).

A more detailed look into O3 DBE class 2 (Fig. 4.9a–c) reveals a broad natural distribution of compound homologues until C43 in Mandal Formation extracts and

C33 for Posidonia Shale. The high relative abundance of this DBE class is determined by the dominance of C16 and C18 homologues, similar to O2 DBE 1. While very abundant in Type II/III samples, these homologues do not prevail in

128 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Type II extracts. In Mandal Formation samples of both OM types and at comparable maturity stages, the apex of C# distributions is at higher carbon numbers than in Posidonia Shale samples (~C24 vs. ~C18).

(a) O3 DBE 2: immature (d) O3 DBE 8: immature (% TMIA)

rel. Intensity

(b) O3 DBE 2: (e) O3 DBE 8: early mature early mature (% TMIA)

rel. Intensity

Fig. 4.9. Carbon number distributions of selected O3 DBE classes addressing detailed compositional differences of the Mandal Formation and Posidonia Shale extracts at three maturity stages.

DBE 8 compounds are more abundant in the Mandal Formation (Fig. 4.8) and are strongly dominated by C20 compounds in the Mandal Formation samples (Fig. 4.9d–f), while Posidonia Shale samples are characterized by a broad natural C# homologue distribution until C33, again similar to O2 class (DBE 7, cf. Fig. 4.7). In the range C22-30, an even-over-odd-predominance of variable intensity can be observed in the Mandal Formation extracts, while not in Posidonia Shale samples.

Even-odd-predominance of C20-30 n-alkanes, which might be daughter structures of such saturated acidic chains, can indicate hypersaline or carbonate environments

(Peters et al., 2005). However, and abundant diasteranes and C30 steranes point to

129 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

marine, clay-rich environment, and

(a) O4: immature elevated CPI values of n-alkanes reflect (% TMIA) a certain land plant input into the system (Table 4.2). rel. Intensity

O4 compounds (Fig. 4.10) are

even more dominated by higher DBE (b) O4: early mature classes, meaning that this compound (% TMIA)

class contains more aromatic moieties

than O2 and O3 compounds (Fig. 4.6 and rel. Intensity Fig. 4.8). While DBE 1 – 4 classes are variably abundant as a function of

(c) O4: mature kerogen type and as a local feature with (% TMIA) enrichment in Type II/III samples 7/11-5 and 7/11-7 (Fig. 4.10b and c), a

rel. Intensity range of DBE 6 – 19 compounds with apices at DBE 7 – 10 forms the majority

Fig. 4.10. DBE class distribution of the O4 of core structures in the O4 compound compound class comparing source rock extracts of the Mandal Formation and class. The dominance of DBE class 8 is Posidonia Shale at three maturity stages. pronounced in wells 2/6-3 and 7/11-7

(Fig. 4.10a and c), just similar to O3 compounds, and to a lower degree in well 7/11-5 (Fig. 4.10b). Other Type II samples are not much dominated by DBE class 8 and rather show a natural distribution of

DBE classes representing aromatic core structures. Just similar as in the O3 compound class, most mature 2/1-9 extract shows a lower and more uniform contribution of aromatic DBE classes, and both extracts of well 7/7-2 containing high and low generation potential (“OM Types” II and II/III) are quite similar which might point to an intra-source contamination of the Type II/III horizon from the Type II horizon. O4 compounds in Posidonia Shale extracts are not abundant and contain rather less aromatic DBE classes 2 – 4 in immature stage and DBE 5 – 9 in early mature samples, thus are less aromatic than Mandal Formation extracts.

While DBE classes 2 and 3 as well as 8 are varying in abundance as a regional feature for samples of different kerogen type, a more detailed look into the chemical composition reveals that these classes are dominated by particular compounds. DBE class 2 is here described in more detail (Fig. 4.11a–c), but both, 2

130 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY and 3 classes are compositionally very similar suggesting a genetic relationship. All Mandal Formation samples, irrespective of “kerogen type”, are dominated by the

C18 homologue and show a lower contribution of the C16 homologue. This distribution is similar to O2 DBE 1 and further suggests 2 acid groups to be incorporated in the chemical structure of O4 DBE 2 and 3 compounds. An even- predominance in the range C22-28 homologues points to similar biological precursors as for n-fatty acids. However, these features are not observed in the Posidonia Shale extracts. Striking is the composition of well 7/11-5 being composed of even- numbered C12-20 homologues (Fig. 4.11b) and thus aligns with the broader composition of lower C# homologues as found in O2 and O3 classes (Fig. 4.7b and Fig. 4.9b).

(a) O4 DBE 2: immature (d) O4 DBE 8: immature (% TMIA)

rel. Intensity

(b) O4 DBE 2: (e) O4 DBE 8: early mature early mature (% TMIA)

rel. Intensity

Fig. 4.11. Carbon number distributions of selected O4 DBE classes addressing detailed compositional differences of the Mandal Formation and Posidonia Shale extracts at three maturity stages.

131 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

DBE class 8 is strongly dominated by the C20 homologue (Fig. 4.11d–f) with extreme dominance in wells 2/6-3, 7/11-5 and 7/11-7 while other extracts, irrespective of OM type, are not dominated by this compound. it is interesting to note that the odd-C# predominance in the C23-30 range shows an enrichment in C27 to C29 homologues, similar to C# predominances in steranes. Actually, a possible precursor of monoaromatic steroids (C25+ homologues composed of 3 rings and 3 double bonds) containing 2 acidic functional groups could represent the long- chained homologues of this O4 DBE class. While Posidonia samples do not contain the C20 homologue, their contribution to this DBE class is solely a matter of the

C23-30 homologues showing odd-predominance as well.

In summary, high abundances of O2-4 compound classes are present in solvent extracts of the Mandal Formation. Notably, the C16 and C18 homologues in the O2

DBE 1, O3 DBE 2 and O4 DBE 2 classes are particularly enriched, as well as the C20 homologues of the O2 DBE 6 and 7, O3 DBE 8 and O4 DBE 8. While some compounds exclusively occur in immature and early mature samples (e.g. C27 homologues in O2 DBE 6), abundances of other compounds (e.g. the C16 and C18 homologues in O2-4 classes) remain at a high level or are seen to become enriched upon maturation (e.g. the C20 homologues in O2 DBE 7, O3 and O4 DBE 8 classes).

4.6 Discussion

4.6.1 On the origin of elevated oxygen content in Mandal extracts

The overall high oxygen content of the polar fraction as well as the abundance of C16 and C18 homologues in aliphatic acidic compounds and C20 components in aromatic oxygen compounds likely reflects the provenance of organic matter contained in the source rock.

Carbon dioxide originating from carboxyl functions in the thermally labile organic matter, as measured using Rock-Eval pyrolysis (Oxygen Index, OI) correlates quite well with the relative abundances of Ox compounds in the polar fraction (Fig. 4.12), while the phenol content in source rock pyrolysates does not (compare Ziegs et al., 2017).

132 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

60 C16 and C18 saturated and Kerogen Type II II/III unsaturated fatty acids are ubiquitous 50 Tmax (°C) 400 - 432 in biological systems. While C16:0 and 432 - 439 40 439 - 445 445 - 450 C18:0 (O2 DBE 1) are abundant faunal

30 and floral lipogenetic products,

18:1 2

OI (mg/gTOC) unsaturated oleic acid (C , O DBE 2) 20 would chiefly be of phytoplanktonic

10 origin. The C16:1 acid is considered as a

0 bacterial marker (Barouxis et al., 1988). 0 10 20 30 40 50 60 70 O x (% T M IA) Longer chained homologues are

Fig. 4.12. Oxygen contents of the unspecific and can be of terrestrial, macromolecular organic matter obtained using Rock-Eval pyrolysis and FT-ICR-MS correlate algal or bacterial origin (Volkman et al., well with each other suggest genetic 1980; Barouxis et al., 1988; Gong and relationship between kerogen and extracts of the source rock. Hollander, 1997) depending on carbon number distributions and odd-even preferences. C16 and C18 fatty acids are abundant in cutins and suberins which contain mid-chain functionalities such as hydroxyl and epoxy groups (de Leeuw and Largeau, 1993), represented by higher

Ox numbers of those homologues (O3, O4, etc.). Bacteria living contemporaneously with algae have been observed in younger geologic systems, such as Cretaceous black shales or near- surface New Zealand coals (Cole, 1982; Pacton et al., 2011;

Glombitza et al., 2016). They yield significant C16 and C18 fatty acids and are competing for nutrients with algae, thus having an impact on preserved OM quality.

The C20 homologues of aromatic O2 DBE 7 acids, abundant in Type II/III (but not Type II) samples of the Mandal Formation, could represent diterpenoid structures, as previously documented in Athabasca oil sands and New Zealand coals using ESI Fig. 4.13. Carbon number distributions of the O2 DBE 7 class of immature New Zealand coal negative FT-ICR-MS (Noah et al., extracts are clearly dominated by the C20 homologue representing a diterpenoid structure 2015, Pötz, pers. comm., 2018, Fig. typical for resinous material. 4.13). Interestingly, the C20 homologues are also abundant in

133 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

DBE classes 5 – 8 of O2 to O4 compound classes of the Type II/III Mandal Formation samples. Noah et al. (2015) suggested abietic and/or palustric acid as representing the O2 DBE 6 homologue, stemming from various pines, larches, firs and their resins and roots (Breitmaier, 2005). Although the Mandal Formation is of marine origin, the structural evolution of the Central Graben does not preclude an input of allochthonous contemporaneous or reworked Type III OM. Gowers and Sæbøe (1985) stated that uplift of marginal areas occurred during the Upper Jurassic extension and subsidence of the Central Graben, especially the Mid North Sea High in the west of the graben. There is evidence of erosion and transport from the eastern Ringkøbing-Fyn High, too. Humphreys et al. (1991) suggested a dilution of autochthonous deposits by the reworking of mature Triassic, Permian and Carboniferous rocks from the Mid North Sea High. Bjørlykke et al. (1975) has earlier found such indication in the basin axial well 2/11-1 of the Central Graben. Additionally, at that time immature Middle and Lower Jurassic sediments containing abundant mires represented by the Sandnes, Bryne and Lulu Formations (Gowers and Sæbøe, 1985; Mellere et al., 2017) were partially eroded off local highs, and deposited in the basin centre as indicated by «scattered tiny particles of reworked vitrinite» (Petersen et al., 2012a). Coals of the Bryne Formation are rich in vitrinitic (telo- and detrovitrinite) and inertinitic (inertodetrinite) macerals representing back-barrier deposits in a strongly aggradational coastal plain environment (Petersen and Andsbjerg, 1996). The above examples represent local events and this might be also restricting the observations in the Mandal Formation assuming resedimentation of eroded material by mass transports, debris flows or slumping. A WSW tilt of the palaeosurface causing the erosion of eastern areas can be inferred from a general movement of Upper Jurassic and Lower Cretaceous depocentres into this direction (Per Varhaugh, Aker BP, pers. comm. 15.3.2018), as well as an erosion of eastern marginal locations of the Mandal Formation (Per Erik Øverli, Aker BP, pers. comm. 15.3.2018).

A general enrichment in oxygen-containing compounds in the polar fraction of the Mandal Formation may be caused by erosion and reworking of slightly older Upper Jurassic sediments of the Farsund Formation by slumping, sliding and debris flows as described above for the Greater Mandal High area has also been proven in successions of the Danish Hot Shales at the Gert Ridge and Jeppe-1 well

134 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

(Ineson et al., 2003). Biostratigraphic analysis in the Jeppe-1 well revealed reworking of older strata into the Bo Member of the Farsund Formation, a stratigraphic horizon equivalent to the lower Mandal Fm.

Inherited biological signatures might also partly account for the enhanced oxygen content of the Mandal polar fraction. Algae tend to incorporate more oxygen into their resistant biomacromolecules when experiencing environmental stress. Microorganisms such as Gloeocapsomorpha Prisca or Botryococcus braunii, respond to higher oxygen supply, fluctuating salinity or lower nutrient supply with thickening of their cell walls, leading to higher aliphaticity of resistant biopolymers, and with an incorporation of oxygen into their outer cell wall structure acting as antioxidants to the organism (e.g. Cole, 1982; Derenne et al., 1992; Stasiuk et al., 1993; Cassab, 1998; Gašparović et al., 2013; Gašparović et al., 2014). Gašparović et al. (2013; 2014) have shown that marine planktonic algae in the N’ Adriatic Sea and East Atlantic Ocean respond to oligotrophic summer conditions and water temperatures >19 °C with cell wall thickening and enhanced incorporation of glycolipids to prevent photooxidation and to achieve thermal stability in warm waters. Having observed increased absorbance in the IR spectra of aromatic bands (Stasiuk et al., 1993), these ring structures might be diagenetically restructured upon burial yielding aromatic oxygen compounds in FT- ICR-MS spectra. Depending on the exact species of the organisms, long aliphatic chains are reformed to cyclic bitumen-like macerals or form free hydrocarbons

(Stasiuk et al., 1993). Based on high abundances of C16 and C18 fatty acids and cross-linked moieties – most likely by ether bonds (de Leeuw and Largeau, 1993; Riboulleau et al., 2000) yielding the explanation for an increase by 1 DBE class from O2 to O3 and O4 compound – a significant input of bacterial OM can be deduced. Cole (1982) has stressed the competitive nature of nutrient uptake between algae and bacteria. Fast nutrient uptake of bacteria might set competitive food-related stress on algae due to lower metabolizable nitrogen available, thus reacting with increased incorporation of oxygen as a barrier for the O-sensitive nitrogenase, a process essential for all metabolic reactions (Cassab, 1998). Furthermore, Weissert and Mohr (1996) have pointed out the impact of climatic changes on nutrient supply. The stepwise warming and aridization until Ryazanian times due to income of cooler, high latitude waters (Abbink et al., 2001) ultimately resulted in a regional collapse of the monsoonal climate causing lower nutrient

135 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY supply and a deceleration of the water cycling (Weissert and Mohr, 1996) which fits with generally low gammacerane ratios identified in the Mandal Formation (Table 4.2). Anoxic to dysoxic depositional environments with low nutrient supply can form TOC-rich, oil-prone source rocks such as the Mandal Formation (Cornford, 1994; Ziegs et al., 2017) under the premise of low sedimentation rates (Tyson, 2005). Anoxia and slow sedimentation rates may cause a feedback loop resulting in the regeneration of nutrients and elevated productivity (Katz, 2005). Ineson et al. (2003) and Michelsen et al. (2003) proposed lower subsidence rates and sediment influx to the Danish Central Graben in Late Volangian to Early Ryazanian times when the Mandal Formation and the Bo Member of the Farsund Formation were deposited. In this context, it need to be mentioned that the Ryazanian ‘Hot Shales’ in the Central Graben represent the latest phase of source rock deposition in the North Sea with the Draupne Formation in the Viking Graben and UK Kimmeridge Clay Formation having been deposited stratigraphically lower until Volgian times only (Cornford, 1994). In general, climatic fluctuations, sea-level changes and paleogeographic reorganizations form a complex interplay, such as in the Miocene Paratethys ocean (Chamley et al., 1986; Kováč et al., 2017), that are likely to have influenced or controlled biotic distributions and evolutions forming a specification of the Ryazanian Mandal Fm in the Central Graben.

4.6.2 Broad inferences regarding petroleum expulsion efficiency

The bulk composition of organic matter (e.g. Sandvik et al., 1992; Pepper and Corvi, 1995; Kelemen et al., 2006a) and to a certain degree the types of clay mineral surfaces (e.g. Huc, 2003; Kahle et al., 2004; Wu et al., 2012; Lewan et al., 2014) certainly play key roles in the retention and expulsion of petroleum. Petroleum expulsion and migration are complex processes and the degree of physicochemical interactions between solid and fluid phases in a source rock depends on various factors that are pressure-temperature sensitive, such as OM type and distribution, its level of maturity and thus the detailed petroleum composition (Mackenzie et al., 1983; Stainforth and Reinders, 1990; Sandvik et al., 1992), but as well on lithology, sedimentological features determining and controlling fluid flow (Okui and Waples, 1993; Mann, 1994; Khavari-Khorasani et al., 1998). HC migratability is determined by its physical properties that are determined by gas content, low-molecular weight and high-molecular weight phase quantities and composition (e.g. Mackenzie et al.,

136 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

1983; Ritter, 2003b, a; Kelemen et al., 2006a; Ziegs et al., 2017). Here, we explore how the bulk chemical composition of the acidic polar fraction, as measured using FT-ICR-MS, might influence retention behaviour.

Extracts of the Mandal Formation are considerably enriched in highly polar oxygen-containing compounds and notably those containing several oxygen atoms

(O4+ classes). Thus, generated and retained petroleum molecules in the Mandal Formation contain more polar functionalities that can interact with polar phases in the source rock, such as water, hydrous clay minerals or residual organic matter. In the context of preferential retention series of asphaltenes > polar NSO’s > aromatic hydrocarbons > aliphatic hydrocarbons (Leythaeuser et al., 1984b; Mackenzie et al., 1987; Sandvik et al., 1992; Pepper and Corvi, 1995), it can be inferred that polarity variations within the NSO fraction play a role in mutual solubility of these fractions (Ritter, 2003b; Ertas et al., 2006; Kelemen et al., 2006b). By way of contrast, the Posidonia Shale representing an excellently expelling marine source rock (e.g. Rullkötter et al., 1988) is a nitrogen-rich system characterized by low polarity. While the N1 fraction increases as function of maturity, less abundant and highly polar oxygen moieties are further reduced upon maturation (Poetz et al., 2014; Mahlstedt et al., 2016).

While an empirical correlation appears to exist between expulsion efficiency and the relative abundance of oxygen- versus pyrrolic nitrogen moieties, a distinct correlation of specific compound classes, DBE classes or C# homologues with retention capacity indices of volatile and bituminous oil (= Total Oil; as presented in Ziegs et al., 2017) has not been observed. The number of possible chemical structures underlying a particular sum formula in the mass spectrum increases exponentially with molecular mass (Koch et al., 2007). Thus, possible combinations of distinct chemical characteristics which determine the exact physicochemical properties of a particular compound increase as well. Each individual chemical structure results in different physical interactions with organic matter and/or the mineral interior.

137 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

4.7 Conclusion

The expulsion efficiency of a source rock is crucial for petroleum yields available in petroleum systems. While the Toarcian Posidonia Shale has expelled its generated petroleum into the Northern German Basin, the Upper Jurassic Mandal Formation in the Central Graben locally retains significant petroleum portions. Retention is attributed to the organic matter inventory of a source rock conventionally approached as S2 using Rock-Eval pyrolysis. But neither amounts nor structural features of the macromolecular kerogen gives plausible hints to enhanced retention capacities of the principle source rock in the North Sea Central Graben. The S2 of whole rock samples containing as well the polar, high-molecular weight portion of generated petroleum cannot be structurally resolved using GC- based analyses, but its acidic polar inventory is unlocked using FT-ICR-MS in ESI negative mode.

Our results indicate that it is not the amount of the NSO fraction but rather its composition that results in high petroleum retention in the source rock. While the Mandal Formation contains less NSO’s at comparable maturity stages than the Posidonia Shale, it possesses the following structural features that we believe to influence retention capacities and fractionation:

(1) The Mandal Formation is enriched in highly polar Ox compound classes containing 3 to 6 oxygen atoms. Each oxygen hetero-bond is affine for surface interaction with other polar sites in the source rock, e.g. kerogen or minerals. (2) These more polar compound classes contain bigger and more aromatic molecules than the Posidonia Shale which is illustrated by higher DBE classes. A DBE class represents the number of rings and double bonds in a chemical structure. (3) The particular DBE classes in Mandal Formation extracts are enriched in longer aliphatic chains than Posidonia Shale extracts, therefore reducing their aromatic character.

(4) Furthermore, the O2-4 classes in Mandal Formation extracts are dominated

by particular compounds, C16 and C18 homologues of aliphatic structures and

C20 homologues of aromatic core structures that are indicative for terrigenous OM input.

138 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

Based on the special characteristics of this marine source rock which are, in part, similar to Type III OM, we suggest various sources for the origin of the generally elevated oxygen content and the particularly high concentrations of C16,

C18 and C20 homologues of O2 to O4 compounds in different DBE classes: C16 and C18 fatty acids are broadly distributed in the floral and faunal universe but might originate from cuticular or bacterial origin. Oxygen-rich C20 compounds most likely representing diterpenoids are indicative for terrigenous OM stemming from erosion of low mature coals from palaeo-highs. The generally high oxygen content could be explained by basin-marginal reworking of freshly deposited marine shales or could be a result of environmental stress on algae, which tend to incorporate oxygen into their resistant cell walls when facing changing conditions. Such variations in temperature, salinity or nutrient supply, as a consequence of changing climate in uppermost Jurassic times, are reported in literature.

Expulsion efficiency and migration behaviour can be evaluated based on relative abundances of heteroatomic N-S-O functionalities but not broken down to the molecular level. Using the FT-ICR-MS, specific identification of functional groups is not readily possible, thus individual physical properties cannot be inferred. Nevertheless, oxygen-rich compounds with a high number of polar functionalities interact stronger with the polar source rock environment.

4.8 Acknowledgements

This study is part of the Ph.D. thesis of Volker Ziegs at GFZ Potsdam as an Industry Partnership with Aker BP ASA. We are grateful for financial support and the permission to publish. Special thanks to Cornelia Karger and Anke Kaminsky (both GFZ Potsdam) for assistance during the lab work, and to Mareike Noah for invaluable help during the sample evaluation using the ‘Dicke Berta’.

139 4: DEEPER INSIGHTS INTO OXYGEN-CONTAINING COMPOUNDS OF THE MANDAL FORMATION, CENTRAL GRABEN, NORWAY

140 5: SUMMARY & PERSPECTIVES

5 Summary & perspectives

5.1 Summary

Although being a very prolific source rock in the Central Graben, the Mandal Formation has been identified by geochemical mass balance modelling as a locally very ineffective expeller of its generated products as compared to the Upper Jurassic Draupne Formation, the major source rock in the Viking Graben system. The first part of the study tied organic matter characteristics to the expulsion efficiency of generated bulk products as well as properties of volatile, free and heavy petroleum fractions. In the subsequent sections, structural characteristics of retained and expelled petroleum portions, using conventional approaches for their hydrocarbon fraction and high-resolution mass spectrometry for their polar inventory, are characterized and correlated to inferred physical properties of retained products.

5.1.1 The Mandal source rock

Source rock characteristics: The Mandal Formation is in part a very prolific source rock having been deposited in an anoxic marine environment during Upper Jurassic sea floor extension and containing mainly Type II organic matter (OM) but Type III-influenced organofacies are also present. While TOC contents with <12 wt.-% show two modes at 1 – 3 % and 5 – 6 %, generation potentials are well distributed ranging up to HI = 647 mg HC/g TOC. The formation is therefore to varying degrees heterogeneous in its generation potential, and according to hydrocarbon as well as polar fractions, evaluated using GC-MS and ESI negative FT-ICR-MS data, these variations can be attributed to varying input of Type III OM into the marine depositional system. Pyrolysis-GC has shown that both kerogen types of differently mature sections of the Mandal Formation have generated Paraffinic-Naphthenic-Aromatic (PNA) Low Wax oil according to aliphatic chain length distributions (Horsfield, 1997), and vary in aromaticity with Type III-influenced organic matter being more aromatic than Type II intervals. However, the kerogen structure does not change significantly from immature to peak oil mature intervals. Bulk kinetic studies confirmed general expectations of higher activation energies (EA) for Type III-influenced intervals and added the

141 5: SUMMARY & PERSPECTIVES

information that broader EA distributions prevail in the basin centre while narrower EA distributions are tied to some marginal locations indicating different preservation modes or more confined organic matter inventory, and thus could result in a slightly later onset of generation.

Retention of volatile hydrocarbons: Previous studies have found that generated petroleum is mainly adhesive to active organic matter and during primary migration (Sandvik et al., 1992; Mahlstedt and Horsfield, 2013). Differently to an effective expeller like the Northern German Posidonia Shale, generated product yields (S1) do not reach their peak at early generation being expelled upon thermal degradation of organic matter, represented by Rock-Eval S2 values. Product yields remain at high levels or slightly increase in abundance. Thus, the bulk retention capacity of OM is increasing throughout maturation of the Mandal Formation source rock until reaching peak oil maturity. These trends become more distinct using the S2 of extracted source rock samples reflecting their true kerogen portion, while the difference between whole rock and extracted S2 values can be ascribed as heavy bitumen (S2bitumen) which is either a product of selective preservation of kerogen precursor structures or a metastable, NSO-rich product of kerogen conversion into petroleum. And it is this S2bitumen portion that correlates linearly with the volatile petroleum portion, S1, as well as S2 of extracted rocks drawing the deduction that generated petroleum is retained within the kerogen or on its surface, and that volatile product amounts are closely linked to this heavy portion. Interestingly, the Mandal Formation has generated elevated amounts of the heavy, immobile products (24 – 63 % of the conventional, unextracted S2) as compared to a worldwide set of marine shales, resulting in higher thresholds until expulsion commences which has been shown by mass balance calculations (after Cooles et al., 1986). While oil retention does not seem to be linked to structural properties of the kerogen as identified from open pyrolysis- GC, aromaticity of the original and matured kerogen plays a major role in its gas retention capacity as cross-linked monoaromatic rings on the outer surface of kerogen act as sorptive sites.

Characteristics of heavy products: The comparison of relative NSO yields within the SARA gross composition of the Mandal Formation and an excellently expelling marine shale, the Northern German Posidonia Shale, illustrates that it is not only the abundance but rather the composition of the NSO fraction that

142 5: SUMMARY & PERSPECTIVES influences yields of retained generation products. Being not able to properly characterize the heavy, polar generation products above m/z ≈ 280 Da using gas- chromatographic methods, Fourier Transform-Ion Cyclotron Resonance-mass spectrometry in negative Electrospray Ionizing mode (ESI negative FT-ICR-MS) is a powerful tool to identify the acidic polar inventory on a high-resolution level containing medium- to high-molecular weight compounds up to 1000 Da. The polar generation products retained in the Mandal Formation are mainly composed of straight-chained and cyclic compounds containing 2 and 3 highly polar oxygen atoms (O2, O3) as part of their cyclic structure or in a functional group. Less polar

N1 and N1O1 compounds and as well constituents with 4 to 6 oxygen atoms (O4-6) occur in lower abundances. Highly polar oxygen atoms are part of fatty acids represented by lower DBE classes or can be incorporated in aromatic structures represented by DBE classes higher than 4 in the O2 compound class. Double bond equivalents (DBE) of a molecule reveals structural information about the degree of unsaturation representing their number of rings plus the number of double bonds involved in the molecule’s core structure or in attached side chains (Hughey et al., 2004). With increasing number of oxygen atoms, the relative aromaticity of compound classes does increase as well, represented by DBE class distributions shifting to higher DBEs resulting in larger molecular core structures. Contrary, retained petroleum in the excellently expelling Posidonia Shale is dominated by less polar N-containing compounds and contains oxygen (Ox) compounds with less heteroatoms and with lower aromaticity, thus molecules are smaller and contain less polar sites for interaction with other polar phases. While the Ox classes in

Mandal extracts are enriched in varying abundance of C20 aromatic and C16 and C18 saturated acids, Posidonia extracts are not dominated by particular compounds, carbon number distributions are rather homogenous and contain less longer- chained moieties. The retained polar inventory of the Mandal source rock is dominated by highly polar oxygen-containing compounds that are quite aromatic but contain long aliphatic chains as well, reducing its aromatic character and possibly its polarity. The activity of the polar atom or functional group is a crucial factor in controlling phase interaction between hydrophobic surfaces and is determined by isomeric positions of side chains. Isomeric shielding effects cannot be identified using FT-ICR-MS. On the other hand, the length of aliphatic side chains influences the in-source interaction with hydrophilic phases, i.e. low- and high- molecular weight OM as well as inorganic materials. In general, those polar

143 5: SUMMARY & PERSPECTIVES molecules may interact strongly with the polar environment provided by a source rock, e.g. hydrous clay minerals or residual organic matter, and solely due to their size and awkward shape they do not migrate easily. However, a link between retention capacity and abundances of specific compounds or DBE classes has not been observed which can be attributed to the number of possible chemical structures represented by a sum formula assigned from FT-ICR-MS data and increasing as a function of molecular mass. Having illustrated the consequences of high polarity and aromaticity of retained organic matter, the origin of high oxygen contents in the petroleum’s heavy portion has been discussed using worldwide equivalents and local examples of the geologic evolution specific to the Central

Graben in Upper Jurassic times. C20 homologues are typical for diterpenoid structures occurring in terrigenous OM having diluted the autochthonous marine deposits by either contemporaneous input of Type III OM or reworking of older, allochthonous, terrigenous material. C16 and C18 saturated acids are unspecific but might be of bacterial origin accompanying algal communities. The generally high oxygen content in bitumens from the Mandal Formation is discussed based on local reporting and can be partly cross-linked. Reworking of freshly deposited shale sequences in marginal sections of the Central Graben basin can be a local factor occurring not only in the Danish sector but also in Norwegian areas, as indicated by confidential data provided by our project partner Aker BP. Furthermore, microorganisms tend to cell wall thickening and incorporation of oxygen into these resistant molecular structures when facing environmental stress. Variations in salinity and/or nutrient supply, e.g. as a consequence of competition for nutrient uptake or climate variations due to palaeogeographic features, have been reported to increase environmental stress.

5.1.2 Central Graben crude oils

Based on the data set of 24 available crude oils from different regions and source rocks in the Central Graben covering an API range of 29 – 55 °API, the present thesis successfully unravelled the effects of source variations, thermal maturation and migration on bulk properties and molecular compositions of expelled petroleum fractions. Having collected crude oils from siliciclastic reservoirs from the entire basin which are in direct contact with potential source rock intervals and stratigraphically higher carbonate reservoirs at the eastern rim and

144 5: SUMMARY & PERSPECTIVES at western basin locations, oils show different depth gradients of their API gravities depending on reservoir lithology. While API gravities of oils in clastic reservoirs are increasing with depth as expected for a maturity suite, oils in carbonate reservoirs characterized by a broad API range and confined thermal maturity stages are accumulated in similar depth levels indicating different migration distances. Otherwise, the gross composition of all oils follows a maturity evolution becoming enriched in saturate hydrocarbons while depleted in aromatic hydrocarbons and the NSO fractions. However, source influences were identified. Interestingly, biomarker information obtained from the saturated fraction suggests a marine source rock for the majority of oils that is homogeneous in OM type and depositional environment with algal organic matter having experienced different levels of oxygenation during deposition (Herbin et al., 1993). The other oils have been charged by the terrestrial influenced, marine Farsund Formation (2/2-1, Gabrielsen et al., 1985), from a highly anoxic, hypersaline marine source rock (2/2- 5, Pedersen et al., 2006) and from the coal-bearing Middle Jurassic Lulu Formation having migrated from the Danish part of the Central Graben (3/7-4, Petersen and Brekke, 2001).

The polar inventory of crude oils is, different to in-source retained petroleum, dominated by compounds with one heteroatom, N1 and O1, and contains minor contributions of O2 and N1O1 classes. The major compound classes, N1, O1 and O2, can be used for oil family classification based on subtle changes of source contribution which is a fast and powerful approach when comparing oils of the same maturity range. The maturity assessment is based on carbazole homologue distributions using N1 DBE classes 9, 12 and 15 as introduced by Oldenburg et al. (2014) and Poetz et al. (2014) for crude oils and source extracts with ring annelation to the carbazole core structure on the expense of aliphatic side chain lengths. Decreasing N1 DBE 9 and increasing N1 DBE 12 and 15 abundances of the polar fraction can be correlated with hydrocarbon maturity ratios. The best fit was obtained for 29Ts/(29Ts+NH) for marine oils stored in clastic reservoirs, while oils from terrestrial-influenced marine source rocks are indicated as higher mature using the polar maturity assessment. Subtle variations have been observed for oils in carbonate reservoirs with higher relative contribution and abundances of the N1 DBE 9 and 12 classes falling out of the range defined by Oldenburg et al. (2014) and can be related to maturity differences. While oils in the western carbonate

145 5: SUMMARY & PERSPECTIVES reservoirs show the highest deviation from the maturity range of oils in clastic reservoirs, oils from the eastern carbonate reservoirs are closer to this range. The carbon number distribution within the DBE 12 class shows an inverse trend with a significant loss of short (C2-3) aliphatic side chains for the eastern oils. The intensity of this loss correlates with the removal of water-soluble, light hydrocarbon compounds and could be attributed to different migration distances and pathways of these oils as even more water-soluble phenolic homologues of the O1 DBE 4 and 5 classes show the same behaviour. Having attributed organic matter as major retention site within TOC-rich source rocks, rock-fluid and water-petroleum interactions in the carrier-reservoir system might have altered N1 and O1 compound distributions of the polar, high-molecular weight fraction. Hereby, structural similarities of all migrating petroleum phases control the adsorption on water- or mixed water/oil-wet mineral surfaces as well as the interaction with water in saturated migration conduits, matrix porosity or fracture/fault systems. In particular, it is an interplay of polarity, aromaticity and aliphaticity determining the physicochemical properties of the migrating petroleum. Polarity is defined as the strength of the dipole moment of the heteroatom and thus influenced by bonding type and shielding effects caused by the molecular shape and in detail by the position of fused rings or annelated side chains. Although isomeric effects cannot be investigated using ESI negative FT-ICR-MS, it might be water-solubility of particular polar petroleum compounds defined by their aliphaticity playing a role at the migration front before bulk oil flow is initiated controlling individual N1 and

O1 compound distributions in crude oils.

5.1.3 Migratability of high-molecular weight generated petroleum

The present study has shown that different fractions generated during thermal maturation of sedimentary organic matter upon burial have varying influence on migration behaviour of the bulk petroleum according to their physicochemical properties. The high-molecular weight portion of petroleum that has yet been less understood due to missing experimental procedures was investigated using FT-ICR-MS in ESI negative mode focussing on polar NSO- containing compounds with an acidic hydrogen atom attached to the heteroatom. This fraction was subject of investigation at different steps of the generation- migration process in a sedimentary basin, namely as part of the residual petroleum

146 5: SUMMARY & PERSPECTIVES in the source rock after expulsion of its major amounts and within the expelled portion of accumulated crude oils.

As pointed out in chapter 2, producible bulk yields (S1) are retained within the S2 portion of a source rock which is thermally not extractable at 300 °C measured during standard Rock-Eval procedure (Behar et al., 2001). Measured volatile yields represent ultimately retained amounts (S1) because rock samples are washed, air-dried and pulverized removing light hydrocarbons that have not been lost during production of the samples. Even an underestimation of ultimately retained portions is possible as subsurface pT-, pH- and eH-conditions may promote additional phase interactions, e.g. solving effects in gas or higher reactivity of included fluid phases, cannot be reproduced having used the available standard procedures.

The S2 portion composed of kerogen and solvent extractable, meta-stable products are both highly complex to macromolecular, not resolvable during GC-FID procedures and mainly composed of NSO constituents with minor contributions of aliphatic and aromatic hydrocarbons (Han et al., 2015 and own, unpublished investigations). Due to different retention modes for gas and oil – gas is strongly retained at the aromatic moieties of the kerogen structure and oil within the overall

S2 portion (S2whole rock) – it was important to investigate this fraction in more structural detail.

The study and structural comparison of source rock extracts and crude oils revealed, and supported earlier findings (Ritter, 2003b, a; Kelemen et al., 2006a), that polarity and structural similarity (aromaticity and aliphaticity) of highly molecular fractions linked with similar physicochemical properties as solubility and phase (semi)-solid-fluid interactions have major effects on compositional variations during expulsion and migration. While highly polar compounds, such as O2 to O6 and N1O2 to N1O5 compounds are entirely or partially retained within the source rock, expelled lower polar compounds (N1, N1O1) strongly interact with rock and water phases upon their way into the reservoir. With increasing number of heteroatoms incorporated in a molecule, the complexity of molecular size and shape and the number of possible structural formulas increases exponentially. Thus, big and awkwardly shaped molecules are hindered in being expelled from the low- permeable source rock or in migrating within a pore network. Nevertheless, larger

147 5: SUMMARY & PERSPECTIVES molecules may be flushed within the bulk flow of expelled petroleum through wider pore throats or faults and fractures.

Although a smaller portion of heavy, polar, less moveable products is expelled, the compositional changes of these compounds are measurable and might influence migration behaviour. It is particularly the low-molecular weight component that clearly affects relative portions of migrating fluids, thus altering the GOR, petroleum movability and phase behaviour, leading to structural changes on migration and alteration of possible migration pathways induced by plugging of pore throats and other migration conduits such as small fractures. In the consequence, existing migration pathways may be less effective or might not be easily assessable at all. In this case, alternative pathways need to be taken into account for basin-wide 3D migration studies. A delay of charging of formerly identified trap structures is possible in such scenarios.

5.2 Perspectives

5.2.1 Retention in and expulsion from the source rock

During this thesis, the primary focus was on the characteristics of the TOC- rich Mandal Formation and compositional evolution of its organic matter upon thermal maturation based on the hypothesis that the organic matter portion is the major retention site of generated products. A compositional comparison of heavy products and their relation to retention characteristics may be applied to stratigraphic equivalents in the North Sea, the oil-prone Bo Member of the (upper) Farsund Formation in the Danish Central Graben, the excellently expelling Draupne Formation in the Viking Graben and the Kimmeridge Clay Formation in the UK sector, or to stratigraphically lower and mixed oil-/gas-prone Heather and Farsund Formations in the Viking and Central grabens, respectively. Having partially contributed to petroleum accumulations where the principal source rock was absent or undermature, these secondary source rock intervals are characterised by subtle variations within the marine depositional environment which thus can be identified quantitatively and qualitatively serving for correlation with crude oils.

148 5: SUMMARY & PERSPECTIVES

A comparison of the composition of the heavy, polar retention products in a suite of worldwide marine shales has revealed that significant compositional variations within compound groups vary between individual marine shales, U.S. Barnett shale, German Posidonia shale, and Argentinian Vaca Muerta shale (not published in course of this thesis). Addressing such variations in the meta-stable NSO products to sedimentological-petrophysical evolution during maturation of subtly different source horizons may give further insight to the expulsion of volatile products in worldwide shale plays. Unmingling generation from kerogen and from these meta-stable NSO’s (cf. Behar et al., 2008) may be of importance at temperatures exceeding the early oil window to properly evaluate the generation potential of oil, gas and late gas of a source rock. FT-ICR-MS analysis of extractable or non-extractable fractions of differently matured MSSV-pyrolysis-GC products, the kerogen, heavy bitumen and free HC portions, might shed light on their individual characteristics. Moreover and after having evaluated the whole extract composition pointing to polarity variations to play a major role in retention characteristics, differently polar portions of generated and retained petroleum should to be analysed individually after separation using Hetereo-MPLC (Willsch et al., 1997). A combination with GC-MS measurements to evaluate the respective low-molecular weight compounds can serve for transferring compositional information into medium- and high-molecular weight portions being resolved using FT-ICR-MS.

However, subtle variations in retention capacity may not have been covered by investigating the organic matter composition alone. Petroleum source rocks can be further characterized by their inorganic inventory, mineral distributions, clay content or porosity evolution (e.g. Eseme, 2006) and even within apparently massive clay intervals, small-scaled structural and textural variations may have a quantitative and qualitative impact on expulsion of bulk volumes. However, the evolution of organic pores as a result of OM swelling and shrinkage during kerogen or bitumen cracking has recently been attributed as significant factor on enhanced petroleum storage with increasing capacities until major kerogen depletion in the principal oil window (Ertas et al., 2006; Kelemen et al., 2006a; Kelemen et al., 2006b; Loucks et al., 2009; Bernard et al., 2012b; Löhr et al., 2015; Han et al., 2017). Due to compaction acting on kerogen as promoting force to expulsion (Tissot and Welte, 1984), an integration of organic geochemistry, sedimentology /

149 5: SUMMARY & PERSPECTIVES microscopy and petrophysical properties such as porosity and permeability evolution with detailed basin development is necessary for evaluation in an appropriate pT-regime.

5.2.2 Crude oils

A major insecurity in the present evaluation of crude oils of the Central Graben was the effect of phase behaviour on observed compositions. Molecular alterations in the NSO fraction, a minor proportion of crude oils, related to increasing migration distances in various carrier lithologies has insignificant effects on bulk or fractional properties by interaction of their hydrogen and ionic bonds with each other and with mineral surfaces (Larter and Aplin, 1995). The low- polarity fraction, represented by carbazoles and phenols and acting as a “molecular velcro” for higher molecular weight compounds, shows a systematic behaviour as a function of increasing viscosity (Oldenburg et al., 2010), thus its removal severely affects phase behaviour, e.g. by asphaltene precipitation during migration. An increase in molecular weight of the C7+ fraction results in an increase of pressure and temperature of phase separation conditions while its density only affects pressure (Kuske and Horsfield, unpubl., pers. comm., Sascha Kuske, 2017). Thus, incorporation of the NSO fraction into phase prediction approaches may be substantial for proper prediction of accumulated compositions.

The observed variations within the selected sample set have been attributed to lithologies of carrier beds, different migration pathways and distances. Due to heterogeneities in the sample set – oils originate from different fields that are heterogeneous themselves and reliable information on source kitchens and migration pathways are missing – the findings are rather theoretical but not unconvincing. In the Tampen Spur area, northern Viking Graben, North Sea, such a confined natural laboratory does exist with bore holes drilled into a well-explored migration pathway having sourced the Snorre oil field (di Primio, 2002; di Primio and Skeie, 2004). Adapting empiric theories to actual conditions in migration pathways, such as mineralogy, water saturation, porosity and permeability measurements, may shed light on factors controlling the compositional evolution of polar, high molecular weight compounds on their movement into reservoir structures.

150 6: REFERENCES

6 References

Abbink, O., Targarona, J., Brinkhuis, H., Visscher, H., 2001. Late Jurassic to earliest Cretaceous palaeoclimatic evolution of the southern North Sea. Global and Planetary Change 30, 231- 256. Abdallah, W., Buckley, J.S., Carnegie, A., Herold, J.E.B., Fordham, E., Graue, A., Signer, T., Hussain, H., Montaron, B., Ziauddin, M., 2007. Fundamentals of Wettability. Oilfield Review, Schlumberger 19, 44-61. Ahsan, S.A., Karlsen, D.A., Mitchell, A.W., Rothwell, N., 1998. Inter and intrafield hydrocarbon compositional variations in the Ula and the Gyda fields (Central Graben-North Sea)- implication for understanding the controls on hydrocarbon distribution within and between these fields. Organic Geochemistry 29, 429-448. Akbarzadeh, K., Hammami, A., Kharrat, A., Zhang, D., Allenson, S., Creek, J., Kabir, S., Jamaluddin, A., Marshall, A.G., Rodgers, R.P., Mullins, O.C., Solbakken, T., 2007. Asphaltenes—Problematic but rich in potential. Oilfield Review, Schlumberger 19, 22-43. Andsbjerg, J., Dybkjaer, K., 2003. Sequence stratigraphy of the Jurassic of the Danish Central Graben. Geological Survey of Denmark and Greenland Bulletin 1, 265-300. Andsbjerg, J., Nielsen, L.H., Johannessen, P.N., Dybkjær, K., 2001. Divergent development of two neighbouring basins following the Jurassic North Sea doming event: the Danish Central Graben and the Norwegian-Danish Basin, in: Ole, J.M., Tom, D. (Eds.), Sedimentary environments offshore Norway - Palaeozoic to recent, 1 ed. Elsevier, Amsterdam, pp. 175- 197. Arthur, M.A., Sageman, B.B., 2005. Sea-level control on source-rock development: Perspectives from the Holocene Black Sea, the Mid-Cretaceous Western Interior Basin of North America, and the Late Devonian Appalachian Basin, in: Harris, N.B. (Ed.), The Deposition of Organic- Carbon-Rich Sediments: Models, Mechanisms, and Consequences. SEPM (Society for Sedimentary Geology), Tulsa, OK, pp. 35-59. Athy, L., 1930. Compaction and oil migration. AAPG Bulletin 14, 25-35. Bae, E., Na, J.-G., Chung, S.H., Kim, H.S., Kim, S., 2010. Identification of about 30 000 chemical components in shale oils by electrospray ionization (ESI) and atmospheric pressure photoionization (APPI) coupled with 15 T Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS) and a comparison to conventional oil. Energy & Fuels 24, 2563- 2569. Bailey, N.J.L., Jobson, A.M., Rogers, M.A., 1973. Bacterial degradation of crude oil: Comparison of field and experimental data. Chemical Geology 11, 203-221. Baird, R.A., 1986. Maturation and source rock evaluation of Kimmeridge clay, Norwegian North Sea. AAPG Bulletin 70, 1-11. Baker, D.R., 1962. Organic geochemistry of Cherokee Group in southeastern Kansas and northeastern Oklahoma. AAPG Bulletin 46, 1621-1642. Baker, E.G., 1959. Origin and Migration of Oil. Oil may consist of hydrocarbons that collected from waters containing natural solubilizers 129, 871-874. Bakr, M.M.Y., Wilkes, H., 2002. The influence of facies and depositional environment on the occurrence and distribution of carbazoles and benzocarbazoles in crude oils: a case study from the Gulf of Suez, Egypt. Organic Geochemistry 33, 561-580. Barker, C., 1972. Aquathermal pressuring--role of temperature in development of abnormal- pressure zones. AAPG Bulletin 56, 2068-2071. Barnard, P.C., Bastow, M.A., 1991. Hydrocarbon generation, migration, alteration, entrapment and mixing in the Central and Northern North Sea. Geological Society, London, Special Publications 59, 167-190.

151 6: REFERENCES

Barouxis, A., Scribe, P., Dagaut, J., Saliot, A., 1988. Free and bound lipids from equatorial surficial sediments separated as a function of particle size. Organic Geochemistry 13, 773- 783. Barrer, R.M., 1989. Shape-selective sorbents based on clay minerals: a review. Clays & Clay Minerals 37, 385-395. Behar, F., Beaumont, V., De B. Penteado, H.L., 2001. Technologie Rock-Eval 6 : performances et développements. Oil & Gas Science and Technology 56, 111-134. Behar, F., Lorant, F., Lewan, M., 2008. Role of NSO compounds during primary cracking of a Type II kerogen and a Type III lignite. Organic Geochemistry 39, 1-22. Behar, F., Vandenbroucke, M., 1987. Chemical modelling of kerogens. Organic Geochemistry 11, 15-24. Behar, F., Vandenbroucke, M., Tang, Y., Marquis, F., Espitalie, J., 1997. Thermal cracking of kerogen in open and closed systems: determination of kinetic parameters and stoichiometric coefficients for oil and gas generation. Organic Geochemistry 26, 321-339. Bennett, B., Buckman, J.O., Bowler, B.F.J., Larter, S.R., 2004. Wettability alteration in petroleum systems: the role of polar non-hydrocarbons. Petroleum Geoscience 10, 271-277. Bennett, B., Chen, M., Brincat, D., Gelin, F.J.P., Larter, S.R., 2002. Fractionation of benzocarbazoles between source rocks and petroleums. Organic Geochemistry 33, 545-559. Bennett, B., Lager, A., Larter, S.R., 2007a. Polar non-hydrocarbons in crude oils and rock extracts: Recovery and impact of sample storage protocols. Journal of Petroleum Science and Engineering 58, 367-374. Bennett, B., Lager, A., Potter, D.K., Buckman, J.O., Larter, S.R., 2007b. Petroleum geochemical proxies for reservoir engineering parameters. Journal of Petroleum Science and Engineering 58, 355-366. Bennett, B., Larter, S.R., 1997. Partition behaviour of alkylphenols in crude oil/brine systems under subsurface conditions. Geochimica et Cosmochimica Acta 61, 4393-4402. Berg, R.R., 1975. Capillary pressures in stratigraphic traps. AAPG Bulletin 59, 939-956. Bernard, S., Horsfield, B., Schulz, H.-M., Wirth, R., Schreiber, A., Sherwood, N., 2012a. Geochemical evolution of organic-rich shales with increasing maturity: A STXM and TEM study of the Posidonia Shale (Lower Toarcian, northern Germany). Marine and Petroleum Geology 31, 70-89. Bernard, S., Wirth, R., Schreiber, A., Schulz, H.-M., Horsfield, B., 2012b. Formation of nanoporous pyrobitumen residues during maturation of the Barnett Shale (Fort Worth Basin). International Journal of Coal Geology 103, 3-11. Berthelot, M., 1866. Sur l’origine des carbures et des combustibles minéraux. Ann. de Chimie et des Physiques 4. ser 9, 481-483. Bjørlykke, K., 1998. Clay mineral diagenesis in sedimentary basins - A key to the prediction of rock properties. Examples from the North Sea Basin. Clay Minerals 33, 15-34. Bjørlykke, K., Dypvik, H., Finstad, K., 1975. The Kimmeridge Shale, its composition and radioactivity, Proceedings of the Jurassic Northern North Sea Symposium, Stavanger, p. 12. Bourdet, J., Burruss, R.C., Chou, I.M., Kempton, R., Liu, K., Hung, N.V., 2014. Evidence for a palaeo-oil column and alteration of residual oil in a gas-condensate field: Integrated oil inclusion and experimental results. Geochimica et Cosmochimica Acta 142, 362-385. Bowen, J.M., 1992. Exploration of the Brent Province. Geological Society, London, Special Publications 61, 3-14. Braun, R.L., Burnham, A.K., 1992. PMOD: a flexible model of oil and gas generation, cracking, and expulsion. Organic Geochemistry 19, 161-172. Braun, R.L., Burnham, A.K., Reynolds, J.G., Clarkson, J.E., 1991. Pyrolysis kinetics for lacustrine and marine source rocks by programmed micropyrolysis. Energy & Fuels 5, 192-204. Bray, E.E., Evans, E.D., 1961. Distribution of n-paraffins as a clue to recognition of source beds. Geochimica et Cosmochimica Acta 22, 2-15.

152 6: REFERENCES

Bray, E.E., Foster, W.R., 1980. A process for primary migration of petroleum. AAPG Bulletin 64, 107-114. Breitmaier, E., 2005. Terpene: Aromen, Düfte, Pharmaka, Pheromone. John Wiley & Sons, Limited,. Brennand, T.P., van Hoorn, B., James, K.H., Glennie, K.W., 1998. Historical Review of North Sea Exploration, Petroleum Geology of the North Sea. Blackwell Science Ltd, pp. 1-41. Brenneman, M.C., Smith Jr, P., 1958. The chemical relationships between crude oils and their source rocks. Habitat of oil: Am. Assoc. Petroleum Geologists, 818-849. Brother, L., Engel, M.H., Krooss, B.M., 1991. The effects of fluid flow through porous media on the distribution of organic compounds in a synthetic crude oil. Organic Geochemistry 17, 11-24. Burnham, A.K., Braun, R.L., Gregg, H.R., Samoun, A.M., 1987. Comparison of methods for measuring kerogen pyrolysis rates and fitting kinetic parameters. Energy & Fuels 1, 452-458. Burst, J.F., 1969. Diagenesis of Gulf Coast clayey sediments and its possible relation to petroleum migration. AAPG bulletin 53, 73-93. Carlson, R.M.K., Chamberlain, D.E., 1986. Steroid biomarker-clay mineral adsorption free energies: Implications to petroleum migration indices. Organic Geochemistry 10, 163-180. Cassab, G.I., 1998. PLANT CELL WALL PROTEINS. Annual Review of Plant Physiology and Plant Molecular Biology 49, 281-309. Cayley, G.T., 1987. Hydrocarbon migration in the central North Sea, in: Brooks, J., Glennie, K.W. (Eds.), Petroleum geology of North West Europe. Graham & Trotman, London, London, pp. 549-556. Chamley, H., Meulenkamp, J., Zachariasse, W., Vanderzwaan, G., 1986. Middle to late miocene marine ecostratigraphy-clay-minerals, planktonic-foraminifera and stable isotopes from sicily. Oceanologica Acta 9, 227-238. Cheng, A.-L., Huang, W.-L., 2004. Selective adsorption of hydrocarbon gases on clays and organic matter. Organic Geochemistry 35, 413-423. Clegg, H., Horsfield, B., Wilkes, H., Sinninghe Damsté, J., Koopmans, M.P., 1998a. Effect of artificial maturation on carbazole distributions, as revealed by the hydrous pyrolysis of an organic-sulphur-rich source rock (Ghareb Formation, Jordan). Organic Geochemistry 29, 1953-1960. Clegg, H., Wilkes, H., Horsfield, B., 1997. Carbazole distributions in carbonate and clastic source rocks. Geochimica et Cosmochimica Acta 61, 5335-5345. Clegg, H., Wilkes, H., Oldenburg, T., Santamaría-orozco, D., Horsfield, B., 1998b. Influence of maturity on carbazole and benzocarbazole distributions in crude oils and source rocks from the Sonda de Campeche, Gulf of Mexico. Organic Geochemistry 29, 183-194. Cole, J.J., 1982. Interactions between bacteria and algae in aquatic ecosystems. Annual Review of Ecology and Systematics 13, 291-314. Connan, J., 1984. Biodegradation of crude oils in reservoirs, in: Brooks, J., Welte, D.H. (Eds.), Advances in Petroleum Geochemistry. Academic Press, London, pp. 299-335. Connan, J., Coustau, H., 1989. Influence of Geological and Geochemical Characteristics of Heavy Oils On Their Recovery, in: Meyer, R.F. (Ed.), Exploration for heavy crude oil and natural bitumen, pp. 261-279. Cooles, G.P., Mackenzie, A.S., Quigley, T.M., 1986. Calculation of petroleum masses generated and expelled from source rocks. Organic Geochemistry 10, 235-245. Cooper, B.S., Barnard, P.C., 1984. Source rock and oils of the central and northern North Sea. AAPG Memoir 35, 303-314. Cooper, B.S., Barnard, P.C., Telnaes, N., 1995. The Kimmeridge Clay Formation of The North Sea, in: Katz, B.J. (Ed.), Petroleum Source Rocks. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 89-110.

153 6: REFERENCES

Cornford, C., 1994. Mandal-Ekofisk(!) petroleum system in the Central Graben of the North Sea, in: Magoon, L.B., Dow, W.G. (Eds.), The petroleum system - From source to trap. AAPG Tulsa, Oklahoma, pp. 537-571. Cornford, C., 1998. Source Rocks and Hydrocarbons of the North Sea, in: Glennie, K.W. (Ed.), Petroleum Geology of the North Sea: Basic Concepts and Recent Advances, 4th ed. Blackwell Science Ltd, Oxford, UK, pp. 376-462. Cornford, C., Gardner, P., Burgess, C., 1998. Geochemical truths in large data sets. I: Geochemical screening data. Organic Geochemistry 29, 519-530. Cornford, C., Rullkötter, J., Welte, D., 1980. A synthesis of organic petrographic and geochemical results from DSDP sites in the eastern central North Atlantic. Physics and Chemistry of the Earth 12, 445-453. Curiale, J.A., Bromley, B.W., 1996. Migration induced compositional changes in oils and condensates of a single field. Organic Geochemistry 24, 1097-1113. Curtin, D., Ballestad, S., 1986. South Viking Graben: habitat of Mesozoic hydrocarbons, in: Spencer, A.M., Holter, E., Campbell, C.J., Hanslien, S.H., Nelson, P.H.H., Nysaether, E., Ormaasen, E.G. (Eds.), Habitat of hydrocarbons on the Norwegian Continental Shelf. Norwegian Petroleum Society, Graham and Trotman, London, pp. 153-157. Curtis, J.B., 2002. Fractured shale-gas systems. AAPG Bulletin 86, 1921-1938. Curtis, M.E., Cardott, B.J., Sondergeld, C.H., Rai, C.S., 2012. Development of organic porosity in the Woodford Shale with increasing thermal maturity. International Journal of Coal Geology 103, 26-31. Damtoft, K., Andersen, C., Thomsen, E., 1987. Prospectivity and hydrocarbon plays of the Danish Central Trough, in: Brooks, J., Glennie, K.W. (Eds.), Petroleum geology of North West Europe. Graham & Trotman, London, London, pp. 403-417. de Leeuw, J.W., Largeau, C., 1993. A review of macromolecular organic compounds that comprise living organisms and their role in kerogen, coal, and petroleum formation. Organic Geochemistry: Principles and Applications 11, 23-72. Delvaux, D., Martin, H., Leplat, P., Paulet, J., 1990. Geochemical characterization of sedimentary organic matter by means of pyrolysis kinetic parameters. Organic Geochemistry 16, 175-187. Demaison, G., Holck, J.J., Jones, R.W., Moore, G.T., 1984. Predictive Source Bed Stratigraphy: A Guide to Regional Petroleum Occurrence, 11th World Petroleum Congress. World Petroleum Congress, London, p. 17. Demaison, G., Huizinga, B.J., 1991. Genetic Classification of Petroleum Systems. AAPG Bulletin 75, 1626-1643. Dembicki Jr., H., 1992. The effects of the mineral matrix on the determination of kinetic parameters using modified Rock Eval pyrolysis. Organic Geochemistry 18, 531-539. Dembicki Jr., H., Horsfield, B., Ho, T.T.Y., 1983. Source Rock Evaluation by Pyrolysis-Gas Chromatography. AAPG Bulletin 67, 1094-1103. Derenne, S., Metzger, P., Largeau, C., Van Bergen, P.F., Gatellier, J.P., Sinninghe Damsté, J.S., de Leeuw, J.W., Berkaloff, C., 1992. Similar morphological and chemical variations of Gloeocapsomorpha prisca in Ordovician sediments and cultured Botryococcus braunii as a response to changes in salinity. Organic Geochemistry 19, 299-313. Dessort, D., Connan, J., Derenne, S., Largeau, C., 1997. Comparative studies of the kinetic parameters of various algaenans and kerogens via open-system pyrolyses. Organic Geochemistry 26, 705-720. di Primio, R., 2002. Unraveling secondary migration effects through the regional evaluation of PVT data: a case study from Quadrant 25, NOCS. Organic Geochemistry 33, 643-653. di Primio, R., Horsfield, B., 1996. Predicting the generation of heavy oils in carbonate/evaporitic environments using pyrolysis methods. Organic Geochemistry 24, 999-1016.

154 6: REFERENCES di Primio, R., Horsfield, B., 2006. From petroleum-type organofacies to hydrocarbon phase prediction. AAPG Bulletin 90, 1031-1058. di Primio, R., Horsfield, B., Guzman-Vega, M.A., 2000. Determining the temperature of petroleum formation from the kinetic properties of petroleum asphaltenes. Nature 406, 173-176. di Primio, R., Skeie, J.E., 2004. Development of a compositional kinetic model for hydrocarbon generation and phase equilibria modelling: A case study from Snorre Field, Norwegian North Sea, in: Cubitt, J.M., England, W.A., Larter, S.R. (Eds.), Understanding Petroleum Reservoirs: towards an Integrated Reservoir Engineering and Geochemical Approach. Geological Society, London, Special Publications, pp. 157-174. Dickey, P.A., 1975. Possible Primary Migration of Oil from Source Rock in Oil Phase. AAPG Bulletin 59, 337-345. Dieckmann, V., Horsfield, B., Schenk, H.J., 2000. Heating rate dependency of petroleum- forming reactions: Implications for compositional kinetic predictions. Organic Geochemistry 31, 1333-1348. Dieckmann, V., Schenk, H.J., Horsfield, B., Welte, D.H., 1998. Kinetics of petroleum generation and cracking by programmed-temperature closed-system pyrolysis of Toarcian Shales. Fuel 77, 23-31. Doré, A.G., Vollset, J., Hamar, G.P., 1985. Correlation of the offshore sequences referred to the Kimmeridge Clay Formation - relevance to the Norwegian sector, in: Thomas, B.M., Doré, A.G., Eggen, S., Home, P.C., Larsen, R.M. (Eds.), Petroleum geochemistry in exploration of the Norwegian Shelf. Graham & Trotman, for Norwegian Petroleum Society, London, pp. 27- 37. du Rouchet, J., 1981. Stress fields, a key to oil migration. AAPG bulletin 65, 74-85. Düppenbecker, S.J., Dohmen, L., Welte, D.H., 1991. Numerical modelling of petroleum expulsion in two areas of the Lower Saxony Basin, Northern Germany, in: England, W.A., Fleet, A.J. (Eds.), Petroleum Migration. Geological Society Special Publications, London, pp. 47-64. Düppenbecker, S.J., Horsfield, B., 1990. Compositional information for kinetic modelling and petroleum type prediction. Organic Geochemistry 16, 259-266. Düppenbecker, S.J., Welte, D.H., 1992. Petroleum expulsion from source rocks - Insights from geology, geochemistry and computerized numerical modelling, 13th World Petroleum Congress. 13th World Petroleum Congress, Buenos Aires, Argentina, pp. 165-177. Durand-Souron, C., 1980. Thermogravimetric analysis and associated techniques applied to kerogens, in: Durand, B. (Ed.), Kerogen: Insoluble organic matter from sedimentary rocks. Editions technip, Paris, pp. 143-162. Durand, B., 1980. Kerogen: Insoluble organic matter from sedimentary rocks. Editions technip, Paris. Durand, B., 1988. Understanding of HC migration in sedimentary basins (present state of knowledge). Organic Geochemistry 13, 445-459. Durand, B., 2003. A history of organic geochemistry. Oil and Gas Science and Technology 58, 203-231. Durand, B., Espitalié, J., 1973. Evolution de la matie ‘re organique au cours de l’enfouissement des sédiments. Compte rendus de l’Académie des Sciences (Paris) 276, 2253-2256. Durand, B., Espitalié, J., 1975. Méthode et appareil permettant l’évaluation rapide du potentiel de production pétrolière de sédiments géologiques sur la base de faibles prélèvements. Brevet national 74 14347. Durand, B., Monin, J.C., 1980. Elemental analysis of kerogens (C, H, O, N, S, Fe), in: Durand, B. (Ed.), Kerogen: Insoluble organic matter from sedimentary rocks. Editions technip, Paris, pp. 113-142.

155 6: REFERENCES

Dybkjaer, K., 1998. Palynological dating of the Mandal Formation (uppermost Jurassic- lowermost Cretaceous, Norwegian Central Graben) and correlation to organic-rich shales in the Danish sector. Marine and Petroleum Geology 15, 495-503. Eglinton, T.I., Philp, R.P., Rowland, S.J., 1988. Flash pyrolysis of artificially matured kerogens from the Kimmeridge Clay, U.K. Organic Geochemistry 12, 33-41. Eglinton, T.I., Sinninghe Damsté, J.S., Kohnen, M.E.L., de Leeuw, J.W., 1990. Rapid estimation of the organic sulphur content of kerogens, coals and asphaltenes by pyrolysis-gas chromatography. Fuel 69, 1394-1404. Eglinton, T.I., Sinninghe Damsté, J.S., Pool, W., de Leeuw, J.W., Eijk, G., Boon, J.J., 1992. Organic sulphur in macromolecular sedimentary organic matter. II. Analysis of distributions of sulphur-containing pyrolysis products using multivariate techniques. Geochimica et Cosmochimica Acta 56, 1545-1560. England, W.A., Mackenzie, A.S., 1989. Some aspects of the organic geochemistry of petroleum fluids. Geol Rundsch 78, 291-303. England, W.A., Mackenzie, A.S., Mann, D.M., Quigley, T.M., 1987. The movement and entrapment of petroleum fluids in the subsurface. Journal of the Geological Society 144, 327-347. Engler, C., 1888. Zur Bildung des Erdöles. Berichte der deutschen chemischen Gesellschaft 21, 1816-1827. Erdmann, M., 1999. Gas generation from overmature Upper Jurassic source rocks, northern Viking Graben. Forschungszentrum, Zentralbibliothek, p. 190. Ertas, D., Kelemen, S.R., Halsey, T.C., 2006. Petroleum Expulsion Part 1. Theory of Kerogen Swelling in Multicomponent Solvents. Energy & Fuels 20, 295-300. Eseme, E., 2006. Oil shales : Compaction, Petroleum Generation and Expulsion, RWTH Aachen, Fakultät für Georessourcen und Materialtechnik RWTH Aachen, p. 143. Eseme, E., Krooss, B.M., Littke, R., 2012. Evolution of petrophysical properties of oil shales during high-temperature compaction tests: Implications for petroleum expulsion. Marine and Petroleum Geology 31, 110-124. Espitalié, J., 1986. Use of Tmax as a maturation index for different types of organic matter. Comparison with vitrinite reflectance, in: Burrus, J. (Ed.), Thermal modelling in sedimentary basins. IFP, Paris, France, pp. 475-496. Espitalie, J., Deroo, G., Marquis, F., 1985. La pyrolyse Rock-Eval et ses applications. Première partie. Rev. Inst. Fr. Pét. 40, 563-579. Espitalié, J., Laporte, J., Madec, M., Marquis, F., Leplat, P., Paulet, J., Boutefeu, A., 1977. Méthode rapide de caractérisation des roches mètres, de leur potentiel pétrolier et de leur degré d'évolution. Oil & Gas Science and Technology 32, 23-42. Espitalie, J., Madec, M., Tissot, B., 1980. Role of mineral matrix in kerogen pyrolysis; influence on petroleum generation and migration. AAPG Bulletin 64, 59-66. Espitalié, J., Senga Makadi, K., Trichet, J., 1984. Role of the mineral matrix during kerogen pyrolysis. Organic Geochemistry 6, 365-382. Espitalié, J., Ungerer, P., Irwin, I., Marquis, F., 1988. Primary cracking of kerogens. Experimenting and modelling C1, C2–C5, C6–C15 and C15+ classes of hydrocarbons formed. Organic Geochemistry 13, 893-899. Faleide, J.I., Bjørlykke, K., Gabrielsen, R.H., 2010. Geology of the Norwegian continental shelf, in: Bjørlykke, K., Avseth, P. (Eds.), Petroleum geoscience: From sedimentary environments to rock physics. Springer, pp. 467-499. Forbes, P.L., Ungerer, P.M., Kuhfuss, A.B., Riis, F., Eggen, S., 1991. Compositional modeling of petroleum generation and expulsion: Trial application to a local mass balance in the Smørbukk Sør Field, Haltenbanken area, Norway. AAPG Bulletin 75, 873-893. Forsman, J.P., Hunt, J.M., 1958. Insoluble organic matter (kerogen) in sedimentary rocks. Geochimica et Cosmochimica Acta 15, 170-182.

156 6: REFERENCES

Freudenberg, K., 1966. Analytical and Biochemical Background of a Constitutional Scheme of Lignin, in: Marton, J. (Ed.), Lignin Structure and Reactions. AMERICAN CHEMICAL SOCIETY, pp. 1-21. Gabrielsen, R.H., Ulvøen, S., Elvsborg, A., Ekern, O.F., 1985. The geological history and geochemical evaluation of Block 2/2, Offshore Norway, in: Thomas, B.M. (Ed.), Petroleum Geochemistry in Exploration of the Norwegian Shelf: Proceedings of a Norwegian Petroleum Society (NPF) conference Organic Geochemistry in Exploration of the Norwegian Shelf held in Stavanger, 22–24 October 1984. Springer Netherlands, Dordrecht, pp. 165-178. Gasparik, M., Ghanizadeh, A., Bertier, P., Gensterblum, Y., Bouw, S., Krooss, B.M., 2012. High- Pressure Methane Sorption Isotherms of Black Shales from The Netherlands. Energy & Fuels 26, 4995-5004. Gašparović, B., Frka, S., Koch, B.P., Zhu, Z.Y., Bracher, A., Lechtenfeld, O.J., Neogi, S.B., Lara, R.J., Kattner, G., 2014. Factors influencing particulate lipid production in the East Atlantic Ocean. Deep Sea Research Part I: Oceanographic Research Papers 89, 56-67. Gašparović, B., Godrijan, J., Frka, S., Tomažić, I., Penezić, A., Marić, D., Djakovac, T., Ivančić, I., Paliaga, P., Lyons, D., Precali, R., Tepić, N., 2013. Adaptation of marine plankton to environmental stress by glycolipid accumulation. Marine Environmental Research 92, 120- 132. Geng, C., Li, S., Ma, Y., Yue, C., He, J., Shang, W., 2012. Analysis and Identification of Oxygen Compounds in Longkou Shale Oil and Shenmu Coal Tar. Oil Shale 29, 322-333. Giraud, A., 1970. Application of pyrolysis and gas chromatography to geochemical characterization of kerogen in sedimentary rock. AAPG Bulletin 54, 439-455. Glasby, G.P., 2006. Abiogenic Origin of Hydrocarbons: An Historical Overview. Resource Geology 56, 83-96. Glennie, K.W., 1990. Lower Permian – Rotliegend, in: Glennie, K.W. (Ed.), Introduction to the petroleum geology of the North Sea, 3rd ed. Blackwell Scientific Publications, Oxford, pp. 120-152. Glombitza, C., Mangelsdorf, K., Horsfield, B., 2016. Differences in bitumen and kerogen-bound fatty acid fractions during diagenesis and early catagenesis in a maturity series of New Zealand coals. International Journal of Coal Geology 153, 28-36. Goff, J.C., 1983. Hydrocarbon generation and migration from Jurassic source rocks in the E Shetland Basin and Viking Graben of the northern North Sea. Journal of the Geological Society 140, 445-474. Gold, T., 1985. The origin of natural gas and petroleum, and the prognosis for future supplies. Annual Review of Energy 10, 53-77. Gong, C., Hollander, D.J., 1997. Differential contribution of bacteria to sedimentary organic matter in oxic and anoxic environments, Santa Monica Basin, California. Organic Geochemistry 26, 545-563. Gowers, M.B., Holtar, E., Swensson, E., 1993. The structure of the Norwegian Central Trough (Central Graben area). Petroleum Geology of Northwest Europe: Proceedings of the ... Conference 4, 1245-1254. Gowers, M.B., Sæbøe, A., 1985. On the structural evolution of the Central Trough in the Norwegian and Danish sectors of the North Sea. Marine and Petroleum Geology 2, 298-318. Hamar, G., Fjaeran, T., Hesjedal, A., 1983. Jurassic stratigraphy and tectonics of the south- southeastern Norwegian offshore, in: Kaasschieter, J., Reijer, T. (Eds.), Petroleum geology of the southeastern North Sea and the adjacent onshore areas. Springer, Amsterdam, pp. 103- 114. Han, Y., Horsfield, B., Wirth, R., Mahlstedt, N., Bernard, S., 2017. Oil retention and porosity evolution in organic rich shales. AAPG Bulletin.

157 6: REFERENCES

Han, Y., Mahlstedt, N., Horsfield, B., 2015. The Barnett Shale: Compositional fractionation associated with intraformational petroleum migration, retention and expulsion. AAPG Bulletin 99, 2173-2202. Harwood, R.J., 1977. Oil and gas generation by laboratory pyrolysis of kerogen. AAPG Bulletin 61, 2082-2102. Herbin, J.P., Müller, C., Geyssant, J.R., Mélières, F., Penn, I.E., 1993. Variation of the distribution of organic matter within a transgressive system tract: Kimmeridge Clay (Jurassic), England, in: Katz, B.J., Pratt, L.M. (Eds.), Source Rocks in a Sequence Stratigraphic Framework. AAPG Special Volume, Tulsa, OK, pp. 67-100. Hobson, G.D., 1954. Some fundamentals of petroleum geology. Oxford University Press, London-NY-Toronto. Horsfield, B., 1984. Pyrolysis Studies and Petroleum Exploration. Academic Press. Horsfield, B., 1989. Practical criteria for classifying kerogens: Some observations from pyrolysis- gas chromatography. Geochimica et Cosmochimica Acta 53, 891-901. Horsfield, B., 1997. The Bulk Petroleum Composition of First-Formed Petroleum in Source Rocks, in: Welte, D.H., Horsfield, B., Baker, D.R. (Eds.), Petroleum and Basin Evolution: Insights from Petroleum Geochemistry, Geology and Basin Modeling. Springer, Berlin, p. 68. Horsfield, B., Clegg, H., Wilkes, H., Santamaría-Orozco, D., 1998. Effect of maturity on carbazole distributions in petroleum systems: New insights from the Sonda de Campeche, Mexico, and Hils Syncline, Germany. Naturwissenschaften 85, 233-237. Horsfield, B., Dembicki Jr, H., Ho, T.T.Y., 1983. Some potential applications of pyrolysis to basin studies. Journal of the Geological Society 140, 431-443. Horsfield, B., di Primio, R., 2014. Fluid Compositional Prediction in Conventional and Unconventional Petroleum Systems, SPE Unconventional Resources Conference. Society of Petroleum Engineers, The Woodlands, Texas, USA Horsfield, B., di Primio, R., Mahlstedt, N., Vu, T.A.T., 2011. Compositional Kinetic Models of Petroleum Generation and Destruction for a Devonian Source Rock. GeoS4 GmbH. Horsfield, B., Dieckmann, V., Santamarıa-Orozco, D., Curry, D., Schenk, H., 2001. Compositional mass balance of petroleum formation, 20th International Meeting of Organic Geochemistry (EGOA). P-THU3-14, Nancy, France. Horsfield, B., Disko, U., Leistner, F., 1989. The micro-scale simulation of maturation: outline of a new technique and its potential applications. Geol Rundsch 78, 361-373. Horsfield, B., Düppenbecker, S.J., 1991. The decomposition of posidonia shale and green river shale kerogens using microscale sealed vessel (MSSV) pyrolysis. Journal of Analytical and Applied Pyrolysis 20, 107-123. Horsfield, B., Heckers, J., Leythaeuser, D., Littke, R., Mann, U., 1991. A study of the Holzener Asphaltkalk, northern Germany: observations regarding the distribution, composition and origin of organic matter in an exhumed petroleum reservoir. Marine and Petroleum Geology 8, 198-211. Horsfield, B., Littke, R., Mann, U., Bernard, S., Vu, T., di Primio, R., Schulz, H., 2010. Shale gas in the Posidonia Shale, Hils area, Germany, AAPG Annual Convention, New Orleans, LA,. Horsfield, B., Rullkötter, J., 1994. Diagenesis, catagenesis, and metagenesis of organic matter. The petroleum system - from source to trap, 189-199. Horsfield, B., Schenk, H.J., Mills, N., Welte, D.H., 1992. An investigation of the in-reservoir conversion of oil to gas: compositional and kinetic findings from closed-system programmed-temperature pyrolysis. Organic Geochemistry 19, 191-204. Hubbert, M.K., 1953. Entrapment of petroleum under hydrodynamic conditions. AAPG Bulletin 37, 1954-2026. Hubbert, M.K., 1954. Entrapment of Petroleum Under Hydrodynamic Conditions: Discussion. AAPG Bulletin 38, 341-342.

158 6: REFERENCES

Huc, A.-Y., Irwin, H., Schoell, M., 1985. Organic matter quality changes in an Upper Jurassic shale sequence from the Viking Graben, in: Thomas, B.M., Doré, A.G., Eggen, S., Home, P.C., Larsen, R.M. (Eds.), Petroleum geochemistry in exploration of the Norwegian Shelf. Graham & Trotman, for Norwegian Petroleum Society, London, pp. 179-183. Huc, A., Durand, B., 1973. Etude des acides humiques et de l’humine de sédiments récents considérés comme précurseurs des kérogènes, in: Tissot, B., Bienner, F. (Eds.), Actes du 6e congrès de géochimie organique. Éditions Technip, Paris, pp. 53-72. Huc, A.Y., 2003. Petroleum Geochemistry At the Dawn of the 21st Century. Oil & Gas Science and Technology - Rev. IFP 58, 233-241. Hughes, W.B., 1984. Use of thiophenic organosulfur compounds in characterizing crude oils derived from carbonate versus siliciclastic sources, in: Palacas, J.G. (Ed.), Petroleum Geochemistry and Source Rock Potential of Carbonate Rocks. American Association of Petroleum Geologists, Tulsa, OK, pp. 181-196. Hughes, W.B., Holba, A.G., Miller, D.E., Richardson, J.S., 1985. Geochemistry of the greater Ekofisk crude oils, in: Thomas, B.M., Doré, A.G., Eggen, S., Home, P.C., Larsen, R.M. (Eds.), Petroleum geochemistry in exploration of the Norwegian Shelf. Graham & Trotman, for Norwegian Petroleum Society, London, pp. 75-92. Hughey, C.A., Galasso, S.A., Zumberge, J.E., 2007. Detailed compositional comparison of acidic NSO compounds in biodegraded reservoir and surface crude oils by negative ion electrospray Fourier transform ion cyclotron resonance mass spectrometry. Fuel 86, 758- 768. Hughey, C.A., Rodgers, R.P., Marshall, A.G., Qian, K., Robbins, W.K., 2002. Identification of acidic NSO compounds in crude oils of different geochemical origins by negative ion electrospray Fourier transform ion cyclotron resonance mass spectrometry. Organic Geochemistry 33, 743-759. Hughey, C.A., Rodgers, R.P., Marshall, A.G., Walters, C.C., Qian, K., Mankiewicz, P., 2004. Acidic and neutral polar NSO compounds in Smackover oils of different thermal maturity revealed by electrospray high field Fourier transform ion cyclotron resonance mass spectrometry. Organic Geochemistry 35, 863-880. Humphreys, B., Morton, A.C., Hallsworth, C.R., Gatliff, R.W., Riding, J.B., 1991. An integrated approach to provenance studies; a case example from the Upper Jurassic of the Central Graben, North Sea. Geological Society Special Publications 57, 251-262. Hunt, J., 1979. Petroleum geochemistry and geology. Freeman, San Francisco. Hunt, J., 1996. Petroleum geochemistry and geology Freeman and Company, New York. Hunt, J.M., 1991. Generation of gas and oil from coal and other terrestrial organic matter. Organic Geochemistry 17, 673-680. Hunt, J.M., Whelan, J.K., Eglinton, L.B., Cathles III, L.M., 1998. Relation of shale porosities, gas generation, and compaction to deep overpressures in the U.S. Gulf Coast, in: Law, B.E., Ulmishek, G.F., Slavin, V.I. (Eds.), Abnormal pressures in hydrocarbon environments, pp. 87- 104. Hunt, J.M., Whelan, J.K., Eglinton, L.B., Cathles, L.M.I., 1994. Gas generation - A major cause of deep Gulf Coast overpressures. Oil and Gas Journal 92, 59-63. Hunt, T.S., 1863. Contributions to the chemical and geological history of bitumens and of pyroschists or bituminous shales. Am. J. Sci., 157-171. Ineson, J.I., Bojesen-Koefoed, J.A., Dybkjær, K., Nielsen, L.H., 2003. Volgian–Ryazanian ‘hot shales’ of the Bo Member (Farsund Formation) in the Danish Central Graben, North Sea: stratigraphy, facies and geochemistry, in: Ineson, J.I., Surlyk, F. (Eds.), The Jurassic of Denmark and Greenland. Geological Survey of Denmark and Greenland, Copenhagen, pp. 403-436. Isaksen, G.H., 2004. Central North Sea hydrocarbon systems: Generation, migration, entrapment, and thermal degradation of oil and gas. AAPG bulletin 88, 1545-1572.

159 6: REFERENCES

Isaksen, G.H., Ledje, K.H.I., 2001. Source rock quality and hydrocarbon migration pathways within the greater Utsira High area, Viking Graben, Norwegian North Sea. AAPG Bulletin 85, 861-884. Japsen, P., Britze, P., Andersen, C., 2003. Upper Jurassic-Lower Cretaceous of the Danish Central Graben; structural framework and nomenclature. Geological Survey of Denmark and Greenland Bulletin 1, 233-246. Jarvie, D.M., 2012. Shale resource systems for oil and gas: Part 2 - Shale-oil resource systems. Shale reservoirs - Giant resources for the 21st century: AAPG Memoir 97, 89-119. Jarvie, D.M., 2014. Components and processes affecting producibility and commerciality of shale resource systems. Geologica Acta 12, 307-325. Jones, R., 1980. Some mass balance and geological constraints on migration mechanisms, in: Roberts, W., Cordell, R. (Eds.), Problems of petroleum migration. AAPG, pp. 47-68. Jüntgen, H., van Heek, K.H., 1968. Gas release from coal as a function of the rate of heating. Fuel 48, 15. Justwan, H., Dahl, B., Isaksen, G.H., 2006. Geochemical characterisation and genetic origin of oils and condensates in the South Viking Graben, Norway. Marine and Petroleum Geology 23, 213-239. Kahle, M., Kleber, M., Jahn, R., 2004. Retention of dissolved organic matter by phyllosilicate and soil clay fractions in relation to mineral properties. Organic Geochemistry 35, 269-276. Karlsson, W., 1986. The Snorre, Statfjord and Gullfaks oilfields and the habitat of hydrocarbons on the Tampen Spur, offshore Norway ( North Sea). Habitat of hydrocarbons on the Norwegian continental shelf. Proc. conference, Stavanger, 1985, 181-197. Katz, B.J., 2005. Controlling factors on source rock development—a review of productivity, preservation, and sedimentation rate, The Deposition of Organic-Carbon-Rich Sediments: Models, Mechanisms, and Consequences (SEPM Special Publications). SEPM (Society for Sedimentary Geology), ISBN, pp. 7-16. Kelemen, S.R., Walters, C.C., Ertas, D., Freund, H., Curry, D.J., 2006a. Petroleum Expulsion Part 3. A Model of Chemically Driven Fractionation during Expulsion of Petroleum from Kerogen. Energy & Fuels 20, 309-319. Kelemen, S.R., Walters, C.C., Ertas, D., Kwiatek, L.M., Curry, D.J., 2006b. Petroleum Expulsion Part 2. Organic Matter Type and Maturity Effects on Kerogen Swelling by Solvents and Thermodynamic Parameters for Kerogen from Regular Solution Theory. Energy & Fuels 20, 301-308. Kenney, J.F., Kutcherov, V.A., Bendeliani, N.A., Alekseev, V.A., 2002. The evolution of multicomponent systems at high pressures: VI. The thermodynamic stability of the hydrogen–carbon system: The genesis of hydrocarbons and the origin of petroleum. Proceedings of the National Academy of Sciences 99, 10976-10981. Keym, M., Dieckmann, V., 2006. Predicting the timing and characteristics of petroleum formation using tar mats and petroleum asphaltenes: A case study from the northern North Sea. Journal of Petroleum Geology 29, 273-296. Keym, M., Dieckmann, V., Horsfield, B., Erdmann, M., Galimberti, R., Kua, L.C., Leith, L., Podlaha, O., 2006. Source rock heterogeneity of the Upper Jurassic Draupne Formation, North Viking Graben, and its relevance to petroleum generation studies. Organic Geochemistry 37, 220-243. Khavari-Khorasani, G., Dolson, J.C., Michelsen, J.K., 1998. The factors controlling the abundance and migration of heavy versus light oils, as constrained by data from the Gulf of Suez. Part I. The effect of expelled petroleum composition, PVT properties and petroleum system geometry. Organic Geochemistry 29, 255-282. Killops, S.D., Funnell, R.H., Suggate, R.P., Sykes, R., Peters, K.E., Walters, C., Woolhouse, A.D., Weston, R.J., Boudou, J.P., 1998. Predicting generation and expulsion of paraffinic oil from vitrinite-rich coals. Organic Geochemistry 29, 1-21.

160 6: REFERENCES

Kim, S., Stanford, L.A., Rodgers, R.P., Marshall, A.G., Walters, C.C., Qian, K., Wenger, L.M., Mankiewicz, P., 2005. Microbial alteration of the acidic and neutral polar NSO compounds revealed by Fourier transform ion cyclotron resonance mass spectrometry. Organic Geochemistry 36, 1117-1134. Koch, B.P., Dittmar, T., Witt, M., Kattner, G., 2007. Fundamentals of Molecular Formula Assignment to Ultrahigh Resolution Mass Data of Natural Organic Matter. Analytical Chemistry 79, 1758-1763. Kolaczkowska, E., Slougui, N.-E., Watt, D.S., Maruca, R.E., Michael Moldowan, J., 1990. Thermodynamic stability of various alkylated, dealkylated and rearranged 17α- and 17β- hopane isomers using molecular mechanics calculations. Organic Geochemistry 16, 1033- 1038. Kováč, M., Hudáčková, N., Halásová, E., Kováčová, M., Holcová, K., Oszczypko-Clowes, M., Báldi, K., Less, G., Nagymarosy, A., Ruman, A., 2017. The Central Paratethys palaeoceanography: a water circulation model based on microfossil proxies, climate, and changes of depositional environment. Acta Geologica Slovaca 9, 75-114. Kroos, B., Leythaeuser, D., Hanebeck, D., 1993. Volume balance of a maturity sequence of the Lower Jurassic Toarcian (Lias E) source rock reflects progress of petroleum generation/explusion, Geofluids' 93, Contributions to an International Conference on Fluid Evolution, Migration and Interaction in Rocks. British Gas. Krooss, B.M., Hanebeck, D., Leythaeuser, D., 1991. Experimental measurement of the molecular migration of light hydrocarbons in source rocks at elevated temperatures. Basin Modelling: Advances and Applications 3, 277-291. Krooss, B.M., van Bergen, F., Gensterblum, Y., Siemons, N., Pagnier, H.J.M., David, P., 2002. High-pressure methane and carbon dioxide adsorption on dry and moisture-equilibrated Pennsylvanian coals. International Journal of Coal Geology 51, 69-92. Kudryavtsev, N., 1951. Against the organic hypothesis of the origin of petroleum. Petroleum Economy [Neftianoye Khozyaistvo] 9, 17-29. Kuila, U., McCarty, D.K., Derkowski, A., Fischer, T.B., Topór, T., Prasad, M., 2014. Nano-scale texture and porosity of organic matter and clay minerals in organic-rich mudrocks. Fuel 135, 359-373. Kuske, S., Horsfield, B., unpubl. Geochemical factors controlling the phase behavior of Eagle Ford Shale petroleum fluids. Lafargue, E., Barker, C., 1988. Effect of Water Washing on Crude Oil Compositions. AAPG Bulletin 72, 263-276. Lafargue, E., Espitalie, J., Jacobsen, T., Eggen, S., 1990. Experimental simulation of hydrocarbon expulsion. Organic Geochemistry 16, 121-131. Lafargue, E., Marquis, F., Pillot, D., 1998. Rock-Eval 6 applications in hydrocarbon exploration, production, and soil contamination studies. Oil & Gas Science and Technology 53, 421-437. Lamberson, M.N., Bustin, R.M., 1993. Coalbed methane characteristics of Gates Formation coals, northeastern British Columbia: effect of maceral composition. AAPG Bulletin 77, 2062-2076. Larter, S., 1985. Integrated kerogen typing in the recognition and quantitative assessment of petroleum source rocks, in: Thomas, B.M., Doré, A.G., Eggen, S., Home, P.C., Larsen, R.M. (Eds.), Petroleum geochemistry in exploration of the Norwegian Shelf. Graham & Trotman, for Norwegian Petroleum Society, London, pp. 269-286. Larter, S., 1989. Chemical models of vitrinite reflectance evolution. Geol Rundsch 78, 349-359. Larter, S.R., 1984. Application of analytical pyrolysis techniques to kerogen characterization and fossil fuel exploration/exploitation, in: Voorhees, K. (Ed.), Analytical Pyrolysis - Methods and Applications. Based upon the invited lectures presented at the Fifth International Symposium on Analytical Pyrolysis. Butterworths, London, pp. 212-275.

161 6: REFERENCES

Larter, S.R., 1990. Molecular characterization of kerogen — Applications to primary and secondary migration studies and to maturation modelling. Review of Palaeobotany and Palynology 65, 379-391. Larter, S.R., Aplin, A.C., 1995. Reservoir geochemistry: methods, applications and opportunities. Geological Society, London, Special Publications 86, 5-32. Larter, S.R., Aplin, A.C., Corbett, P.W.M., Ementon, N., Chen, M., Taylor, P.N., 1997. Reservoir Geochemistry: A Link Between Reservoir Geology and Engineering? SPE Reservoir Engineering. Larter, S.R., Bowler, B.F.J., Li, M., Chen, M., Brincat, D., Bennett, B., Noke, K., Donohoe, P., Simmons, D., Kohnen, M., Allan, J., Telnaes, N., Horstad, I., 1996. Molecular indicators of secondary oil migration distances. Nature 383, 593-597. Larter, S.R., Douglas, A.G., 1980. A pyrolysis-gas chromatographic method for kerogen typing. Physics and Chemistry of the Earth 12, 579-583. Larter, S.R., Horsfield, B., 1993. Determination of Structural Components of Kerogens by the Use of Analytical Pyrolysis Methods, in: Engel, M., Macko, S. (Eds.), Organic Geochemistry. Springer US, pp. 271-287. Larter, S.R., Mills, N., 1991. Phase-controlled molecular fractionations in migrating petroleum charges, in: England, W.A., Fleet, A.J. (Eds.), Petroleum Migration. Geological Society Special Publications, London, pp. 137-147. Leonard, R.C., Munns, J.W., 1987. Valhall, in: Spencer, A.M. (Ed.), Habitat of hydrocarbons on the Norwegian Continental Shelf. Graham & Trotman, Norwell, MA. Lewan, M., Dolan, M., Curtis, J., 2014. Effects of smectite on oil-expulsion efficiency of the Kreyenhagen Shale based on hydrous-pyrolysis experiments. AAPG Bulletin 98, 1091-1109. Lewan, M., Winters, J., McDonald, J., 1979. Generation of oil-like pyrolyzates from organic-rich shales. Science 203, 897-899. Lewan, M.D., 1997. Experiments on the role of water in petroleum formation. Geochimica et Cosmochimica Acta 61, 3691-3723. Leythaeuser, D., Mackenzie, A., Schaefer, R.G., Bjoroy, M., 1984a. A novel approach for recognition and quantification of hydrocarbon migration effects in shale-sandstone sequences. AAPG Bulletin 68, 196-219. Leythaeuser, D., Poelchau, H.S., 1991. Expulsion of petroleum from type III kerogen source rocks in gaseous solution: modelling of solubility fractionation, in: England, W.A., Fleet, A.J. (Eds.), Petroleum Migration. Geological Society Special Publications, London, pp. 33-46. Leythaeuser, D., Radke, M., Schaefer, R.G., 1984b. Efficiency of petroleum expulsion from shale source rocks. Nature 311, 745-748. Leythaeuser, D., Radke, M., Willsch, H., 1988a. Geochemical effects of primary migration of petroleum in Kimmeridge source rocks from Brae field area, North Sea. II: Molecular composition of alkylated naphthalenes, phenanthrenes, benzo- and dibenzothiophenes. Geochimica et Cosmochimica Acta 52, 2879-2891. Leythaeuser, D., Schaefer, R.G., 1984. Effects of hydrocarbon expulsion from shale source rocks of high maturity in Upper Carboniferous strata of the Ruhr area, Federal Republic of Germany. Organic Geochemistry 6, 671-681. Leythaeuser, D., Schaefer, R.G., Radke, M., 1987. On the primary migration of petroleum, 12th World Petroleum Congress. World Petroleum Congress, pp. 227-236. Leythaeuser, D., Schaefer, R.G., Radke, M., 1988b. Geochemical effects of primary migration of petroleum in Kimmeridge source rocks from Brae field area, North Sea. I: Gross composition of C15+-soluble organic matter and molecular composition of C15+-saturated hydrocarbons. Geochimica et Cosmochimica Acta 52, 701-713. Leythaeuser, D., Schaefer, R.G., Yukler, A., 1982. Role of diffusion in primary migration of hydrocarbons. AAPG Bulletin 66, 408-429.

162 6: REFERENCES

Leythaeuser, D., Welte, D.H., 1969. Relation between Distribution of Heavy n-Paraffins and Coalification in Carboniferous Coals from the Saar District, Germany, Advances in Organic Geochemistry 1968. Pergamon, pp. 429-442. Li, M., Cheng, D., Pan, X., Dou, L., Hou, D., Shi, Q., Wen, Z., Tang, Y., Achal, S., Milovic, M., Tremblay, L., Hinrichs, K.-U., Michaelis, W., Rullkotter, J., 2010. Characterization of petroleum acids using combined FTIR, FTICR-MS and GC-MS; implications for the origin of high acidity oils in the Muglad Basin, Sudan. Organic Geochemistry 41, 959-965. Li, M., Fowler, M.G., Obermajer, M., Stasiuk, L.D., Snowdon, L.R., 1999. Geochemical characterisation of Middle Devonian oils in NW Alberta, Canada: possible source and maturity effect on pyrrolic nitrogen compounds. Organic Geochemistry 30, 1039-1057. Li, M., Larter, S.R., Frolov, Y.B., Bjoroy, M., 1994. Adsorptive interaction between nitrogen compounds and organic and/or mineral phases in subsurface rocks. Models for compositional fractionation of pyrrolic nitrogen compounds in petroleum during petroleum migration. Journal of High Resolution Chromatography 17, 230-236. Li, M., Larter, S.R., Stoddart, D., Bjorøy, M., 1995. Fractionation of pyrrolic nitrogen compounds in petroleum during migration: derivation of migration-related geochemical parameters. Geological Society, London, Special Publications 86, 103-123. Li, M., Yao, H., Fowler, M.G., Stasiuk, L.D., 1998. Geochemical constraints on models for secondary petroleum migration along the Upper Devonian Rimbey-Meadowbrook reef trend in central Alberta, Canada. Organic Geochemistry 29, 163-182. Li, M., Yao, H., Stasiuk, L.D., Fowler, M.G., Larter, S.R., 1997. Effect of maturity and petroleum expulsion on pyrrolic nitrogen compound yields and distributions in Duvernay Formation petroleum source rocks in central Alberta, Canada. Organic Geochemistry 26, 731-744. Liao, Y., Shi, Q., Hsu, C.S., Pan, Y., Zhang, Y., 2012. Distribution of acids and nitrogen-containing compounds in biodegraded oils of the Liaohe Basin by negative ion ESI FT-ICR-MS. Organic Geochemistry 47, 51-65. Lind, S., 1938. The Science of Petroleum. New York, Oxford University Press. Littke, R., Baker, D.R., Leythaeuser, D., 1988. Microscopic and sedimentologic evidence for the generation and migration of hydrocarbons in Toarcian source rocks of different maturities. Organic Geochemistry 13, 549-559. Littke, R., Horsfield, B., Leythaeuser, D., 1989. Hydrocarbon distribution in coals and in dispersed organic matter of different maceral compositions and maturities. Geol Rundsch 78, 391-410. Littke, R., Leythaeuser, D., Rullkötter, J., Baker, D.R., 1991. Keys to the depositional history of the Posidonia Shale (Toarcian) in the Hils Syncline, northern Germany. Geological Society, London, Special Publications 58, 311-333. Liu, P., Li, M., Jiang, Q., Cao, T., Sun, Y., 2015. Effect of secondary oil migration distance on composition of acidic NSO compounds in crude oils determined by negative-ion electrospray Fourier transform ion cyclotron resonance mass spectrometry. Organic Geochemistry 78, 23-31. Löhr, S.C., Baruch, E.T., Hall, P.A., Kennedy, M.J., 2015. Is organic pore development in gas shales influenced by the primary porosity and structure of thermally immature organic matter? Organic Geochemistry 87, 119-132. Losh, S., Cathles III, L., 2010. Phase fractionation and oil-condensate mass balance in the South Marsh Island Block 208–239 area, offshore Louisiana. Marine and Petroleum Geology 27, 467-475. Loucks, R.G., Reed, R.M., 2014. Scanning-Electron-Microscope Petrographic Evidence for Distinguishing Organic-Matter Pores Associated with Depositional Organic Matter versus Migrated Organic Matter in Mudrock.

163 6: REFERENCES

Loucks, R.G., Reed, R.M., Ruppel, S.C., Jarvie, D.M., 2009. Morphology, Genesis, and Distribution of Nanometer-Scale Pores in Siliceous Mudstones of the Mississippian Barnett Shale. Journal of Sedimentary Research 79, 848-861. Mackenzie, A., Price, I., Leythaeuser, D., Muller, P., Radke, M., Schaefer, R., 1987. The expulsion of petroleum from Kimmeridge Clay source-rocks in the area of the Brae Oilfield, UK continental shelf. Petroleum geology of North West Europe. Graham & Trotman, London, 865-877. Mackenzie, A.S., Leythaeuser, D., Muller, P., Quigley, T.M., Radke, M., 1988. The movement of hydrocarbons in shales. Nature 331, 63-65. Mackenzie, A.S., Leythaeuser, D., Schaefer, R.G., Bjoroy, M., 1983. Expulsion of petroleum hydrocarbons from shale source rocks. Nature 301, 506-509. Mahlstedt, N., 2012. Evaluating the late gas potential of source rocks stemming from different sedimentary environments, Institute of Applied Geosciences. Technical University of Berlin, Berlin, p. 369. Mahlstedt, N., Horsfield, B., 2013. A new screening tool for the rapid evaluation of gas sorption capacity in shales in: Jin, Z., Katz, B.J. (Eds.), AAPG Hedberg. AAPG Beijing, China. Mahlstedt, N., Horsfield, B., submitted. A new screening tool for the rapid evaluation of gas sorption capacity in shales. Mahlstedt, N., Horsfield, B., Wilkes, H., Poetz, S., 2016. Tracing the Impact of Fluid Retention on Bulk Petroleum Properties Using Nitrogen-Containing Compounds. Energy & Fuels 30, 6290- 6305. Mann, U., 1994. An integrated approach to the study of primary petroleum migration. Geological Society, London, Special Publications 78, 233-260. Mann, U., Hantschel, T., Schaefer, R.G., Krooss, B., Leythaeuser, D., Littke, R., Sachsenhofer, R.F., 1997. Petroleum migration: Mechanisms, Pathways, efficiencies and numerical simulations, in: Welte, D., Horsfield, B., Baker, D. (Eds.), Petroleum and Basin Evolution. Springer Berlin Heidelberg, pp. 403-520. Mann, U., Müller, P.J., 1988. Source rock evaluation by well log analysis (Lower Toarcian, Hils syncline). Organic Geochemistry 13, 109-119. Marshall, A.G., Hendrickson, C.L., 2002. Fourier transform ion cyclotron resonance detection: principles and experimental configurations. International Journal of Mass Spectrometry 215, 59-75. Marshall, A.G., Rodgers, R.P., 2004. Petroleomics: The Next Grand Challenge for Chemical Analysis. Accounts of Chemical Research 37, 53-59. Marshall, A.G., Rodgers, R.P., 2008. Petroleomics: Chemistry of the underworld. Proceedings of the National Academy of Sciences 105, 18090-18095. McAuliffe, C.D., 1966. Solubility in Water of Paraffin, Cycloparaffin, Olefin, Acetylene, Cycloolefin, and Aromatic Hydrocarbons. The Journal of Physical Chemistry 70, 1267-1275. McAuliffe, C.D., 1980. Oil and gas migration; chemical and physical constraints. AAPG Bulletin 63, 761-781. McClennen, W.H., Meuzelaar, H.L.C., Metcalf, G.S., Hill, G.R., 1983. Characterization of phenols and indanols in coal-derived liquids: Use of Curie-point vaporization gas chromatography/mass spectrometry. Fuel 62, 1422-1429. Meissner, F.F., 1978. Petroleum geology of the Bakken Formation, Williston Basin, North Dakota and Montana, Williston Basin Proceedings. Montana Geological Society, Billings, pp. 207-227. Mellere, D., Mannie, A., Longhitano, S., Mazur, M., Kulausa, H., Brough, S., Cotton, J., 2017. Tidally influenced shoal water delta and estuary in the Middle Jurassic of the Søgne Basin, Norwegian North Sea: sedimentary response to rift initiation and salt tectonics. Geological Society, London, Special Publications 444, 173-213. Mendeleiev, D.I., 1878. Revue Scientific XIII, 409.

164 6: REFERENCES

Merkel, A., Fink, R., Littke, R., 2015. The role of pre-adsorbed water on methane sorption capacity of Bossier and Haynesville shales. International Journal of Coal Geology 147–148, 1-8. Meulbroek, P., Cathles Iii, L., Whelan, J., 1998. Phase fractionation at South Eugene Island Block 330. Organic Geochemistry 29, 223-239. Michelsen, O., Nielsen, L.H., Johannessen, P.N., Andsbjerg, J., Surlyk, F., 2003. The Jurassic of Denmark and Greenland: Jurassic lithostratigraphy and stratigraphic development onshore and offshore Denmark. Geological Survey of Denmark and Greenland Bulletin 1, 147-216. Modica, C.J., Lapierre, S.G., 2012. Estimation of kerogen porosity in source rocks as a function of thermal transformation: Example from the Mowry Shale in the Powder River Basin of Wyoming. AAPG bulletin 96, 87-108. Moldowan, J.M., Fago, F.J., Carlson, R.M.K., Young, D.C., Vanduyne, G., Clardy, J., Schoell, M., Pillinger, C.T., Watt, D.S., 1991. Rearranged Hopanes in Sediments and Petroleum. Geochimica Et Cosmochimica Acta 55, 3333-3353. Møller, J.J., Rasmussen, E.S., 2003. Middle Jurassic–Early Cretaceous rifting of the Danish Central Graben. Geological Survey of Denmark and Greenland Bulletin 1, 247-264. Momper, J.A., 1978. Oil migration limitations suggested by geological and geochemical considerations, in: Roberts, W., Cordell, R. (Eds.), Physical and Chemical Constraints on Petroleum Migration. American Association of Petroleum Geologists. Education Department, Tulsa, OK. Moretti, I., Deacon, K., 1995. Subsidence, maturation and migration history of the Tampen Spur area. Marine and Petroleum Geology 12, 345-375. Neumann, H.-J., Paczynska-Lahme, B., Severin, D., 1981. Composition and properties of petroleum, Geology of petroleum. Wiley, Chichester, p. 137. Neumann, V., 2007. Numerical modeling and phase prediction in deep overpressured basinal settings of the Central Graben, North Sea, Department of Applied Geosciences. TU Berlin, Berlin, p. 161. Nielsen, L.H., 2003. Late Triassic–Jurassic development of the Danish Basin and the Fennoscandian Border Zone, southern Scandinavia, in: Ineson, J.I., Surlyk, F. (Eds.), The Jurassic of Denmark and Greenland. Geological Survey of Denmark and Greenland, Copenhagen, pp. 459-526. Nielsen, O.B., Sorensen, S., Thiede, J., Skarbo, O., 1986. Cenozoic differential subsidence of North Sea. AAPG Bulletin 70, 276-298. Noah, M., Poetz, S., Vieth-Hillebrand, A., Wilkes, H., 2015. Detection of Residual Oil-Sand- Derived Organic Material in Developing Soils of Reclamation Sites by Ultra-High-Resolution Mass Spectrometry. Environmental Science & Technology 49, 6466-6473. Ohm, S.E., Karlsen, D.A., Phan, N.T., Strand, T., Iversen, G., 2012. Present Jurassic petroleum charge facing Paleozoic biodegraded oil: Geochemical challenges and potential upsides, Embla field, North Sea. AAPG bulletin 96, 1523-1552. Okui, A., Waples, D.W., 1993. Relative permeabilities and hydrocarbon expulsion from source rocks, Basin Modelling: Advances and Applications. Elsevier Amsterdam, pp. 293-301. Oldenburg, T., Horsfield, B., Wilkes, H., Stoddart, D., Wilhelms, A., 1999. Benzo[b]carbazole as a continental-deltaic source indicator, offshore Norway, Book of Abstracts, 19th International Meeting on Organic Geochemistry. EAOG, Istanbul, pp. 541-542. Oldenburg, T.B.P., Brown, M., Bennett, B., Larter, S.R., 2014. The impact of thermal maturity level on the composition of crude oils, assessed using ultra-high resolution mass spectrometry. Organic Geochemistry 75, 151-168. Oldenburg, T.B.P., Wilkes, H., Horsfield, B., van Duin, A.C.T., Stoddart, D., Wilhelms, A., 2002. Xanthones—novel aromatic oxygen-containing compounds in crude oils. Organic Geochemistry 33, 595-609.

165 6: REFERENCES

Oldenburg, T.B.P., Yarranton, H.W., Larter, S., 2010. The Effect of Low Molecular Weight Multifunctional Additives on Heavy Oil Viscosity, Canadian Unconventional Resources and International Petroleum Conference. Society of Petroleum Engineers, Calgary, Alberta, Canada. Orr, W.L., 1986. Kerogen/asphaltene/sulfur relationships in sulfur-rich Monterey oils. Organic Geochemistry 10, 499-516. Pacton, M., Schmid, T., Gorin, G., Massault, M., Stadler, J., 2011. Cretaceous black shale: a window into microbial life adaptation. Terra Nova 23, 362-368. Palmer, S.E., 1984. Effect of water washing on C15+ hydrocarbon fraction of crude oils from northwest Palawan, Philippines. AAPG Bulletin 68, 137-149. Palmer, S.E., 1993. Effect of Biodegradation and Water Washing on Crude Oil Composition, in: Engel, M.H., Macko, S.A. (Eds.), Organic Geochemistry: Principles and Applications. Springer US, Boston, MA, pp. 511-533. Pan, Y., Liao, Y., Shi, Q., Hsu, C.S., 2013. Acidic and Neutral Polar NSO Compounds in Heavily Biodegraded Oils Characterized by Negative-Ion ESI FT-ICR-MS. Energy & Fuels 27, 2960- 2973. Pedersen, J.H., Karlsen, D.A., Backer-Owe, K., Lie, J.E., Brunstad, H., 2006. The geochemistry of two unusual oils from the Norwegian North Sea: implications for new source rock and play scenario. Petroleum Geoscience 12, 85-96. Pelet, R., Behar, F., Monin, J.C., 1986. Resins and asphaltenes in the generation and migration of petroleum. Organic Geochemistry 10, 481-498. Pepper, A.S., 1991. Estimating the petroleum expulsion behaviour of source rocks: A novel quantitative approach, in: England, W.A., Fleet, A.J. (Eds.), Petroleum migration. Geological Society Special Publications, London, pp. 9-31. Pepper, A.S., Corvi, P.J., 1995. Simple kinetic models of petroleum formation. Part III: Modelling an open system. Marine and Petroleum Geology 12, 417-452. Pepper, A.S., Dodd, T.A., 1995. Simple kinetic models of petroleum formation. Part II: oil-gas cracking. Marine and Petroleum Geology 12, 321-340. Perry Jr, E.A., Hower, J., 1972. Late-stage dehydration in deeply buried pelitic sediments. AAPG Bulletin 56, 2013-2021. Peters, K.E., Cassa, M.R., 1994. Applied source rock geochemistry, in: Magoon, L.B., Dow, W.G. (Eds.), The petroleum system - From source to trap. AAPG Memoirs, pp. 93-120. Peters, K.E., Walters, C.C., Moldowan, J.M., 2005. The biomarker guide; II, Biomarkers and isotopes in petroleum systems and Earth history. Cambridge University Press, Cambridge. Petersen, H.I., Andersen, C., Holme, A.C., Carr, A.D., Thomsen, E., 2012a. Vitrinite reflectance gradients of deep wells with thick chalk sections and high pressure; implications for source rock maturation, Danish-Norwegian Central Graben, North Sea. International Journal of Coal Geology 100, 65-81. Petersen, H.I., Andsbjerg, J., 1996. Organic facies development within Middle Jurassic coal seams, Danish Central Graben, and evidence for relative sea-level control on peat accumulation in a coastal plain environment. Sedimentary Geology 106, 259-277. Petersen, H.I., Andsbjerg, J., Bojesen-Koefoed, J.A., Nytoft, H.P., Rosenberg, P., 1998. Petroleum potential and depositional environments of Middle Jurassic coals and non- marine deposits, Danish Central Graben, with special reference to the Sogne Basin. Geology of Denmark Survey Bulletin 36, 78. Petersen, H.I., Brekke, T., 2001. Source rock analysis and petroleum geochemistry of the Trym discovery, Norwegian North Sea: a Middle Jurassic coal-sourced petroleum system. Marine and Petroleum Geology 18, 889-908. Petersen, H.I., Holme, A.C., Andersen, C., Whitaker, M.F., Nytoft, H.P., Thomsen, E., 2013. The source rock potential of the Upper Jurassic-lowermost Cretaceous in the Danish and southern Norwegian sectors of the Central Graben, North Sea. First Break 31, 43-53.

166 6: REFERENCES

Petersen, H.I., Holme, A.C., Thomsen, E., Whitaker, M.F., Brekke, T., Bojesen-Koefoed, J.A., Hansen, K.H., Larsen, B.T., 2011. Hydrocarbon potential of Middle Jurassic coaly and lacustrine and Upper Jurassic-lowermost Cretaceous marine source rocks in the Søgne Basin, North Sea. Journal of Petroleum Geology 34, 277-304. Petersen, H.I., Holme, A.C., Whitaker, M.F., Andersen, C., Nytoft, H.P., Thomsen, E., 2012b. The Upper Jurassic-lowermost Cretaceous source rocks in the Central Graben, North Sea; status and future challenges. Conference and Technical Exhibition - European Association of Geoscientists and Engineers 74, 0-Paper D009. Philippi, G.T., 1965. On the depth, time and mechanism of petroleum generation. Geochimica et Cosmochimica Acta 29, 1021-1049. Poetz, S., Horsfield, B., Wilkes, H., 2014. Maturity-Driven Generation and Transformation of Acidic Compounds in the Organic-Rich Posidonia Shale as Revealed by Electrospray Ionization Fourier Transform Ion Cyclotron Resonance Mass Spectrometry. Energy & Fuels 28, 4877-4888. Price, L.C., 1976. Aqueous solubility of petroleum as applied to its origin and primary migration. AAPG Bulletin 60, 213-244. Price, L.C., 1981. Primary petroleum migration by molecular solution: consideration of new data. Journal of Petroleum Geology 4, 89-101. Price, L.C., 1989. Primary petroleum migration from shales with oxygen-rich organic matter. Journal of Petroleum Geology 12, 289-324. Price, L.C., Wenger, L.M., Ging, T., Blount, C.W., 1983. Solubility of crude oil in methane as a function of pressure and temperature. Organic Geochemistry 4, 201-221. Purcell, J.M., Rodgers, R.P., Hendrickson, C.L., Marshall, A.G., 2007. Speciation of Nitrogen Containing Aromatics by Atmospheric Pressure Photoionization or Electrospray Ionization Fourier Transform Ion Cyclotron Resonance Mass Spectrometry. Journal of the American Society for Mass Spectrometry 18, 1265-1273. Quigley, T., Mackenzie, A., Gray, J., 1987. Kinetic theory of petroleum generation, in: Doligez, B. (Ed.), Migration of Hydrocarbons in Sedimentary Basins. Editions Technip, Paris, pp. 649- 665. Radke, M., Horsfield, B., Littke, R., Rullkötter, J., 1997. Maturation and Petroleum Generation, in: Welte, D.H., Horsfield, B., Baker, D.R. (Eds.), Petroleum and Basin Evolution: Insights from Petroleum Geochemistry, Geology and Basin Modeling. Springer, Berlin, p. 62. Radke, M., Willsch, H., Welte, D.H., 1980. Preparative hydrocarbon group type determination by automated medium pressure liquid chromatography. Analytical Chemistry 52, 406-411. Raji, M., Gröcke, D.R., Greenwell, C., Cornford, C., 2015a. Pyrolysis, Porosity and Productivity in Unconventional Mudstone Reservoirs:‘Free’and ‘Adsorbed’Oil. Unconventional Resources Technology Conference (URTEC). Raji, M., Gröcke, D.R., Greenwell, H.C., Gluyas, J.G., Cornford, C., 2015b. The effect of interbedding on shale reservoir properties. Marine and Petroleum Geology 67, 154-169. Rawson, P.F., Riley, L.A., 1982. Latest Jurassic - Early Cretaceous Events and the "Late Cimmerian Unconformity" in North Sea Area. AAPG Bulletin 66, 2628-2648. Raynaud, J., Robert, P., 1976. Les méthodes d'étude optique de la matière organique. Bulletin du Centre Recherche Pau, SNPA 10, 109-127. Rexer, T.F., Mathia, E.J., Aplin, A.C., Thomas, K.M., 2014. High-Pressure Methane Adsorption and Characterization of Pores in Posidonia Shales and Isolated Kerogens. Energy & Fuels 28, 2886-2901. Reynolds, J.G., Burnham, A.K., 1995. Comparison of kinetic analysis of source rocks and kerogen concentrates. Organic Geochemistry 23, 11-19. Riboulleau, A., Derenne, S., Sarret, G., Largeau, C., Baudin, F., Connan, J., 2000. Pyrolytic and spectroscopic study of a sulphur-rich kerogen from the “Kashpir oil shales” (Upper Jurassic, Russian platform). Organic Geochemistry 31, 1641-1661.

167 6: REFERENCES

Ritter, U., 2003a. Fractionation of petroleum during expulsion from kerogen. Journal of Geochemical Exploration 78–79, 417-420. Ritter, U., 2003b. Solubility of petroleum compounds in kerogen: implications for petroleum expulsion. Organic Geochemistry 34, 319-326. Ross, D.J.K., Bustin, M.R., 2009. The importance of shale composition and pore structure upon gas storage potential of shale gas reservoirs. Marine and Petroleum Geology 26, 916-927. Rossland, A., Escalona, A., Rolfsen, R., 2013. Permian–Holocene tectonostratigraphic evolution of the Mandal High, Central Graben, North Sea. AAPG Bulletin 97, 923-957. Rouxhet, P.G., Robin, P.L., Nicaise, G., 1980. Characterization of kerogens and of their evolution by infrared spectroscopy, in: Durand, B. (Ed.), Kerogen: Insoluble organic matter from sedimentary rocks. Editions technip, Paris, pp. 163-190. Rullkötter, J., Leythaeuser, D., Horsfield, B., Littke, R., Mann, U., Müller, P.J., Radke, M., Schaefer, R.G., Schenk, H.J., Schwochau, K., Witte, E.G., Welte, D.H., 1988. Organic matter maturation under the influence of a deep intrusive heat source: A natural experiment for quantitation of hydrocarbon generation and expulsion from a petroleum source rock (Toarcian shale, northern Germany). Organic Geochemistry 13, 847-856. Rullkötter, J., Marzi, R., 1988. Natural and artificial maturation of biological markers in a Toarcian shale from northern Germany. Organic Geochemistry 13, 639-645. Rullkötter, J., Michaelis, W., 1990. The structure of kerogen and related materials. A review of recent progress and future trends. Organic Geochemistry 16, 829-852. Sandvik, E.I., Mercer, J.N., 1990. Primary migration by bulk hydrocarbon flow. Organic Geochemistry 16, 83-89. Sandvik, E.I., Young, W.A., Curry, D.J., 1992. Expulsion from hydrocarbon sources: the role of organic absorption. Organic Geochemistry 19, 77-87. Sanei, H., Wood, J.M., Ardakani, O.H., Clarkson, C.R., Jiang, C., 2015. Characterization of organic matter fractions in an unconventional tight gas siltstone reservoir. International Journal of Coal Geology 150–151, 296-305. Schenk, H.J., di Primio, R., Horsfield, B., 1997. The conversion of oil into gas in petroleum reservoirs. Part 1: Comparative kinetic investigation of gas generation from crude oils of lacustrine, marine and fluviodeltaic origin by programmed-temperature closed-system pyrolysis. Organic Geochemistry 26, 467-481. Schenk, H.J., Dieckmann, V., 2004. Prediction of petroleum formation: the influence of laboratory heating rates on kinetic parameters and geological extrapolations. Marine and Petroleum Geology 21, 79-95. Schenk, H.J., Witte, E.G., Littke, R., Schwochau, K., 1990. Structural modifications of vitrinite and alginite concentrates during pyrolitic maturation at different heating rates. A combined infrared, 13C NMR and microscopical study. Organic Geochemistry 16, 943-950. Senftle, J.T., Larter, S.R., Bromley, B.W., Brown, J.H., 1986. Quantitative chemical characterization of vitrinite concentrates using pyrolysis-gas chromatography. Rank variation of pyrolysis products. Organic Geochemistry 9, 345-350. Shi, Q., Zhao, S., Xu, Z., Chung, K.H., Zhang, Y., Xu, C., 2010. Distribution of Acids and Neutral Nitrogen Compounds in a Chinese Crude Oil and Its Fractions: Characterized by Negative-Ion Electrospray Ionization Fourier Transform Ion Cyclotron Resonance Mass Spectrometry. Energy & Fuels 24, 4005-4011. Silverman, S.R., 1965. Migration and Segregation of Oil and Gas, Fluids in Subsurface Environments. AAPG, Tulsa, OK, USA, pp. 53-65. Sinninghe Damste, J.S., De Leeuw, J.W., 1990. Analysis, structure and geochemical significance of organically-bound sulphur in the geosphere: State of the art and future research. Organic Geochemistry 16, 1077-1101. Sinninghe Damsté, J.S., Eglinton, T.I., De Leeuw, J.W., Schenck, P., 1989. Organic sulphur in macromolecular sedimentary organic matter: I. Structure and origin of sulphur-containing

168 6: REFERENCES

moieties in kerogen, asphaltenes and coal as revealed by flash pyrolysis. Geochimica et Cosmochimica Acta 53, 873-889. Sinninghe Damsté, J.S., Irene, W., Rijpstra, C., de Leeuw, J.W., Schenck, P.A., 1988. Origin of organic sulphur compounds and sulphur-containing high molecular weight substances in sediments and immature crude oils. Organic Geochemistry 13, 593-606. Söderström, B., Forsberg, A., Holtar, E., Rasmussen, B.A., 1991. The Mjolner Field, a deep Upper Jurassic oil field in the central North Sea. First Break 9, 156-171. Sokolov, V., Zhause, T., Vassoyevich, N., Antonoc, P., Grigoriyev, G., Kozlov, V., 1964. Migration processes of gas and oil, their intensity and directionality, 6th World Petroleum Congress, Frankfurt, pp. 493-505. Spackman, W., 1958. Section of geology and mineralogy: the maceral concept and the study of modern environments as a means of understanding the nature of coal. Transactions of the New York Academy of Sciences 20, 411-423. Spencer, A.M., Home, P.C., Wiik, V., 1986b. Habitat of hydrocarbons in the Jurassic Ula trend, Central Graben, Norway, in: Spencer, A.M. (Ed.), Habitat of hydrocarbons on the Norwegian continental shelf. Graham & Trotman, London, p. 354. Stach, E., Mackowsky, M.T., Teichmüller, M., Taylor, G., Chandra, D., Teichmüller, R., 1982. Textbook of Coal Petrology. Gebrüder Borntraeger, Berlin, Stuttgart. Stainforth, J.G., 2009. Practical kinetic modeling of petroleum generation and expulsion. Marine and Petroleum Geology 26, 552-572. Stainforth, J.G., Reinders, J.E.A., 1990. Primary migration of hydrocarbons by diffusion through organic matter networks, and its effect on oil and gas generation. Organic Geochemistry 16, 61-74. Stasiuk, L.D., Kybett, B.D., Bend, S.L., 1993. Reflected light microscopy and micro-FTIR of Upper Ordovician Gloeocapsomorpha prisca alginite in relation to paleoenvironment and petroleum generation, Saskatchewan, Canada. Organic Geochemistry 20, 707-719. Steen, A.S., Wilhelms, A., Schulz, L., Oldenburg, T., Horsfield, B., 2000. Benzo[b]carbazole, an indicator on terrestrial organic matter? Results from offshore Mid-Norway, Book of Abstracts, 7th Norwegian Meeting in Organic Geochemistry, Stavanger. Stoddart, D.P., Hall, P.B., Larter, S.R., Brasher, J., Li, M., Bjorøy, M., 1995. The reservoir geochemistry of the Eldfisk Field, Norwegian North Sea, The Geochemistry of Reservoris. Geological Society, London, pp. 257-279. Stone, R.W., ZoBell, C.E., 1952. Bacterial aspects of the origin of petroleum. Industrial & Engineering Chemistry 44, 2564-2567. Stopes, M.C., 1935. On the petrology of banded bituminous coal. Fuel in science and practice. Sykes, R., Snowdon, L.R., 2002. Guidelines for assessing the petroleum potential of coaly source rocks using Rock-Eval pyrolysis. Organic Geochemistry 33, 1441-1455. Taylor, P., Larter, S., Jones, M., Dale, J., Horstad, I., 1997. The effect of oil-water-rock partitioning on the occurrence of alkylphenols in petroleum systems. Geochimica et Cosmochimica Acta 61, 1899-1910. Taylor, P.N., 1994. Controls on the occurrence of phenols in petroleums and waters. University Newcastle upon Tyne, Newcastle upon Tyne. Tegelaar, E.W., de Leeuw, J.W., Derenne, S., Largeau, C., 1989a. A reappraisal of kerogen formation. Geochimica et Cosmochimica Acta 53, 3103-3106. Tegelaar, E.W., Matthezing, R.M., Jansen, J.B.H., Horsfield, B., De Leeuw, J.W., 1989b. Possible origin of n-alkanes in high-wax crude oils. Nature 342, 529-531. Tegelaar, E.W., Noble, R.A., 1994. Kinetics of hydrocarbon generation as a function of the molecular structure of kerogen as revealed by pyrolysis-gas chromatography. Organic Geochemistry 22, 543-574. Teichmüller, M., 1971. Anwendung kohlenpetrographischer Methoden bei der Erdöl-und Erdgasprospektion. Erdöl und Kohle 24, 69-76.

169 6: REFERENCES

Thomas, B.M., 1985. Petroleum Geochemistry in Exploration of the Norwegian Shelf - Proceedings of a Norwegian Petroleum Society (NPF) conference Organic Geochemistry in Exploration of the Norwegian Shelf held in Stavanger, 22–24 October 1984. Graham & Trotman, for Norwegian Petroleum Society. Thomas, M.M., Clouse, J.A., 1990a. Primary migration by diffusion through kerogen: I. Model experiments with organic-coated rocks. Geochimica et Cosmochimica Acta 54, 2775-2779. Thomas, M.M., Clouse, J.A., 1990b. Primary migration by diffusion through kerogen: II. Hydrocarbon diffusivities in kerogen. Geochimica et Cosmochimica Acta 54, 2781-2792. Thomas, M.M., Clouse, J.A., 1990c. Primary migration by diffusion through kerogen: III. Calculation of geologic fluxes. Geochimica et Cosmochimica Acta 54, 2793-2797. Thompson, K.F.M., 1987. Fractionated aromatic petroleums and the generation of gas- condensates. Organic Geochemistry 11, 573-590. Thompson, K.F.M., 1988. Gas-condensate migration and oil fractionation in deltaic systems. Marine and Petroleum Geology 5, 237-246. Tissot, B., Espitalie, J., 1975. L'evolution thermique de la matière organique des sédiments: applications d'une simulation mathématique. Potentiel pétrolier des bassins sédimentaires de reconstitution de l'histoire thermique des sédiments. Oil & Gas Science and Technology 30, 743-778. Tissot, B., Pelet, R., 1971. Nouvelles données sur les mécanismes de genèse et de migration du pétrole: Simulation mathématique et application à la prospection. Proc. 8th World Petroleum Congress 2, 35-46. Tissot, B., Pelet, R., Ungerer, P., 1987. Thermal history of sedimentary basins, maturation indices, and kinetics of oil and gas generation. AAPG bulletin 71, 1445-1466. Tissot, B., Welte, D., 1984. Petroleum formation and occurrence: A new approach to oil and gas exploration. Springer, Berlin, Heidelberg. Trask, P.D., Hammar, H.E., Wu, Z., 1932. Origin and environment of source sediments of petroleum. Gulf Publishing Company. Trask, P.D., Patnode, H.W., 1942. Source beds of petroleum. JSTOR. Trask, P.D., Wu, C., 1930. Does petroleum form in sediments at time of deposition? AAPG Bulletin 14, 1451-1463. Treibs, A., 1934. Chlorophyll- und Häminderivate in bituminösen Gesteinen, Erdölen, Erdwachsen und Asphalten. Ein Beitrag zur Entstehung des Erdöls. Justus Liebigs Annalen der Chemie 510, 42-62. Tyson, R., 2005. The" productivity versus preservation" controversy: cause, flaws, and resolution. Special Publication-SEPM 82, 17. Tyson, R.V., Wilson, R.C.L., Downie, C., 1979. A stratified water column environmental model for the type Kimmeridge Clay. Nature 277, 377-380. Underhill, J.R., Partington, M.A., 1993. Jurassic thermal doming and deflation in the North Sea: implications of the sequence stratigraphic evidence. Geological Society, London, Petroleum Geology Conference series 4, 337-345. Ungerer, P., 1990. State of the art of research in kinetic modelling of oil formation and expulsion. Organic Geochemistry 16, 1-25. Ungerer, P., Behar, F., Villalba, M., Heum, O.R., Audibert, A., 1988. Kinetic modelling of oil cracking. Organic Geochemistry 13, 857-868. Ungerer, P., Pelet, R., 1987. Extrapolation of the kinetics of oil and gas formation from laboratory experiments to sedimentary basins. Nature 327, 52-54. van Duin, A.C.T., Larter, S.R., 2001. Molecular dynamics investigation into the adsorption of organic compounds on kaolinite surfaces. Organic Geochemistry 32, 143-150. van Graas, G.W., Elin Gilje, A., P. Isom, T., Aase Tau, L., 2000. The effects of phase fractionation on the composition of oils, condensates and gases. Organic Geochemistry 31, 1419-1439. van Heek, K.H., 1982. Druckpyrolyse von Steinkohlen. VDI Forschungsheft 612, 48.

170 6: REFERENCES

Van Krevelen, D., 1950. Graphical-statistical method for the study of structure and reaction processes of coal. Fuel 29, 269-284. van Krevelen, D.W., 1993. Coal: Typology, chemistry, physics, constitution, 3rd ed. Elsevier Amsterdam. Vandenbroucke, M., 1980. Structure of kerogens as seen by investigations on soluble extracts, in: Durand, B. (Ed.), Kerogen: Insoluble organic matter from sedimentary rocks. Editions technip, Paris, pp. 415-444. Vandenbroucke, M., 1993. Migration of hydrocarbons. Applied Petroleum Geochemistry, Éditions Technip, Paris, 123-148. Vandenbroucke, M., Behar, F., Rudkiewicz, J.L., 1999. Kinetic modelling of petroleum formation and cracking: implications from the high pressure/high temperature Elgin Field (UK, North Sea). Organic Geochemistry 30, 1105-1125. Vandenbroucke, M., Largeau, C., 2007. Kerogen origin, evolution and structure. Organic Geochemistry 38, 719-833. Vernadsky, V., 1934. Outlines of geochemistry ONTI. Gornogeolog. Neft. Izd. Vitorovic, D., 1980. Structure elucidation of kerogen by chemical methods, in: Durand, B. (Ed.), Kerogen: Insoluble organic matter from sedimentary rocks. Editions technip, Paris, pp. 301- 338. Volkman, J.K., Johns, R.B., Gillan, F.T., Perry, G.J., Bavor, H.J., 1980. Microbial lipids of an intertidal sediment—I. Fatty acids and hydrocarbons. Geochimica et Cosmochimica Acta 44, 1133-1143. Wang, J., Zhang, X., Li, G., 2011. Detailed characterization of polar compounds of residual oil in contaminated soil revealed by Fourier transform ion cyclotron resonance mass spectrometry. Chemosphere 85, 609-615. Wang, Y., Zhu, Y., Liu, S., Zhang, R., 2016. Pore characterization and its impact on methane adsorption capacity for organic-rich marine shales. Fuel 181, 227-237. Waples, D.W., Machihara, T., 1990. Application of sterane and triterpane biomarkers in petroleum exploration. Bulletin of Canadian Petroleum Geology 38, 357-380. Wei, H., Bjoroy, M., Roaldset, E., 1995. Kinetic study of source rocks and coals from the Norwegian North Sea; inorganic gas generation and hydrocarbon cracking. Conference and Technical Exhibition - European Association of Geoscientists and Engineers 57, abstr. P545. Weiss, H.M., Wilhelms, A., Mills, N., Scotchmer, J., Hall, P.B., Lind, K., Brekke, T., 2000. NIGOGA - The Norwegian Industry Guide to Organic Geochemical Analyses [online], 4.0 ed. Norsk Hydro, Statoil, Geolab Nor, SINTEF Petroleum Research and the Norwegian Petroleum Directorate. Weissert, H., Mohr, H., 1996. Late Jurassic climate and its impact on carbon cycling. Palaeogeography, Palaeoclimatology, Palaeoecology 122, 27-43. Welte, D., 1987. Migration of hydrocarbons. Facts and theory, in: Doligez, B. (Ed.), Migration of hydrocarbons in sedimentary basins. Technip, Paris, pp. 393-413. Whitehead, E.V., 1973. Molecular evidence for the biogenesis of petroleum and natural gas, in: Ingerson, E. (Ed.), Proceedings of Symposium on Hydrogeochemistry and Biogeochemistry. The Clarke Co., Tokyo, 1970, pp. 158-211. Wilhelms, A., Larter, S., 2004. Shaken but not always stirred. Impact of petroleum charge mixing on reservoir geochemistry. Geological Society, London, Special Publications 237, 27- 35. Wilhelms, A., Larter, S.R., Leythaeuser, D., Dypvik, H., 1990. Recognition and quantification of the effects of primary migration in a Jurassic clastic source-rock from the Norwegian continental shelf. Organic Geochemistry 16, 103-113. Wilkes, H., Clegg, H., Disko, U., Willsch, H., Horsfield, B., 1998. Fluoren-9-ones and carbazoles in the Posidonia Shale, Hils Syncline, northwest Germany. Fuel 77, 657-668.

171 6: REFERENCES

Willsch, H., Clegg, H., Horsfield, B., Radke, M., Wilkes, H., 1997. Liquid Chromatographic Separation of Sediment, Rock, and Coal Extracts and Crude Oil into Compound Classes. Analytical Chemistry 69, 4203-4209. Wilson, M., Pugmire, R., Karas, J., Alemany, L., Woolfenden, W., Grant, D., Given, P., 1984. Carbon distribution in coals and coal macerals by cross polarization magic angle spinning carbon-13 nuclear magnetic resonance spectrometry. Analytical Chemistry 56, 933-943. Wu, L.M., Zhou, C.H., Keeling, J., Tong, D.S., Yu, W.H., 2012. Towards an understanding of the role of clay minerals in crude oil formation, migration and accumulation. Earth-Science Reviews 115, 373-386. Yang, S., 2017. Roles of organic-inorganic interactions in the generation of petroleum as exemplified by Lower Palaeozoic petroleum systems, Europe. TU Berlin, Berlin, p. 151. Yaws, C., Yang, H.-C., Pan, X., 1991. Henry's Law of Constants for 362 Organic Compounds in Water. Chemical Engineering 98, 179. Yen, T.F., 1974. Structure of Petroleum Asphaltene and Its Significance. Energy Sources 1, 447- 463. Young, A., McIver, R.D., 1977. Distribution of hydrocarbons between oils and associated fine- grained sedimentary rocks - Physical chemistry applied to petroleum geochemistry, II. AAPG Bulletin 61, 1407-1436. Yükler, M., Dow, W., 1990. Temperature, pressure, and hydrocarbon generation histories in San Marcos Arch area, in: Schumacher, D., Perkins, B. (Eds.), 9th Annual Research Conference SEPM. Gulf Coast Society of SEPM, pp. 99-104. Zhang, Y., Person, M., Merino, E., 2005. Hydrologic and geochemical controls on soluble benzene migration in sedimentary basins. Geofluids 5, 83-105. Zhang, Y., Shi, Q., Li, A., Chung, K.H., Zhao, S., Xu, C., 2011. Partitioning of Crude Oil Acidic Compounds into Subfractions by Extrography and Identification of Isoprenoidyl Phenols and Tocopherols. Energy & Fuels 25, 5083-5089. Ziegler, P., 1988. Evolution of the Arctic-North Atlantic and the Western Tethys. AAPG, Tulsa, Oklahoma. Ziegler, P., 1990. Tectonic and paleogeographic development of the North Sea rift system, in: Blundell, D., Gibbs, A. (Eds.), Tectonic Evolution of the North Sea Rifts. Oxford Science Publications, Oxford, pp. 1-36. Ziegs, V., 2013. Development of a compositional kinetic model for primary and secondary petroleum generation from Lower Cretaceous Wealden Shales, Lower Saxony Basin, Northern Germany, Department of Geology. Freiberg University of Mining and Technology, Freiberg, p. 82. Ziegs, V., Horsfield, B., Di Primio, R., Rinna, J., Skeie, J.E., 2015. Petroleum expulsion from the Mandal Formation, Norwegian North Sea, 27th International Meeting on Organic Geochemistry, Sept, 13-18. EAOG, Prague, Czech Republic. Ziegs, V., Horsfield, B., Skeie, J.E., Rinna, J., 2017. Petroleum retention in the Mandal Formation, Central Graben, Norway. Marine and Petroleum Geology 83, 195-214. Ziegs, V., Noah, M., Poetz, S., Horsfield, B., Hartwig, A., Rinna, J., Skeie, J.E., 2018. Unravelling maturity- and migration-related carbazole and phenol distributions in Central Graben crude oils. Marine and Petroleum Geology 94, 114-130. ZoBell, C.E., 1947. Microbial transformation of molecular hydrogen in marine sediments, with particular reference to petroleum. AAPG Bulletin 31, 1709-1751.

172