Document [original]

Investigations on Hydrodesulfurization Reactions

using Slurry Catalysts and Supercritical Water

vorgelegt von

M. Sc.

Stephan Risse

an der Fakultät III – Prozesswissenschaften

der Technischen Universität Berlin

zur Erlangung des akademischen Grades

Doktor der Ingenieurwissenschaften

– Dr.-Ing. –

genehmigte Dissertation

Promotionsausschuss:

Vositzender: Prof. Dr. Jadran Vrabec

Gutachter: Prof. Dr. Frank Behrendt

Gutachter: Prof. Dr. Rainer Reimert

Tag der wissenschaftlichen Aussprache: 6. November 2020

Berlin 2021

Ich erkläre hiermit, dass ich die vorliegende Arbeit selbständig verfasst und keine

anderen als die angegebenen Quellen und Hilfsmittel verwendet habe.

Berlin, den 21. Dezember 2020

Acknowledgement

I would like to thank a number of people who strongly supported me during the

whole time of this thesis, especially, my former colleague Ivo Schneider, without

whom this thesis would not have been possible. Also my coordinators Prof. Frank

Behrendt and Alba Dieguez Alonso as well as the whole team at the RDH, who

created a very productive and fun atmosphere to work in. I would also like to

express my gratitude to Christian Hecka and Claus Strecker, who not only supported

the research financially, but also supported the progress of this work with intense

discussions. Last but not least, I would like to thank my friends and especially my

family, who kept supporting me through the whole time of my thesis.

Abstract

Depletion of easily accessed and light crude oil sources is shifting future crude oil

supply to less attractive heavier oil fractions. Vacuum residue, as the heaviest frac-

tion found in crude oil and the bottom product of the vacuum distillation column

in a refinery, has therefore been gaining more and more attention in the past years.

It is mostly used as heavy fuel oil in the shipping sector on open seas. Globaliza-

tion and rising demand of international transportation of goods has led to a strong

increase of heavy fuel oil consumption. Sulfur, present in vacuum residue to high

extent, forms environmentally hazardous sulfur oxides (SO2and SO3) when being

burnt in the ships engines. In order to counteract rising sulfur oxide emissions, the

International Maritime Organization has released a number of restrictions of which

the last reduces the maximum allowed sulfur content in fuel for ships on open seas

from 3.5% to 0.5%. These new sulfur regulations have come into force on January

1st 2020 and have led to increasing prices for fuels that comply with the sulfur reg-

ulations. On the other hand, high sulfur fuels exceeding the sulfur cap of 0.5% are

declining in price, also affecting refineries margins when selling high sulfur vacuum

residue. Developing a solution for removing sulfur directly from vacuum residue

therefore bears large economic potential. Existing hydrodesulfurization technologies

severely suffer from catalyst deactivation, coking and plugging as a result of high

asphaltenes, heavy metals, and Conradson Carbon present in the viscous vacuum

residue. Slurry phase hydrodesulfurization presents potentials for overcoming some

of these downsides by using unsupported, highly dispersed catalysts and additives.

In this work, process conditions, catalysts, and additives were tested and evaluated

with respect to their activity in enhancing desulfurization of high sulfur vacuum

residue. Experiments were conducted in a 2 l semi-batch slurry reactor that had

the possibility of collecting oil and residue fractions separately. At low temperature,

long residence time, and high H2partial pressure, undesired conversion reactions

could be minimized while desulfurization reactions could be maximized. Screening

of active substances that influence the reactions, revealed a catalyst and an additive

that showed very different effects on conversion and desulfurization reactions than

the majority of the tested catalysts. Supercritical water, as very cheap additive,

supported solely conversion reactions yielding large amounts of oil, while leaving a

high sulfur residue behind. Investigations on sulfur containing model compounds

underlined the absence of desulfurization reactions with supercritical water. The

novel catalyst showed a strong hydrogenation activity thus destabilizing the sulfur

bonds and enabling desulfurization of the high sulfur vacuum residue. Desulfuriza-

tion of above 90% was achieved while at 400 ◦Cand 310 bar undesired conversion

reactions could be kept below 30%. Both mechanistic pathways behind supercritical

water hydroconversion and hydrodesulfurization with the novel catalyst were inves-

tigated intensively and laid basis for an economic assessment of both hydrotreating

routes. On basis of a 50,000 barrels per day vacuum residue upgrading plant, the

two cases were calculated. Catalyst cost displayed a large penalty on the economic

performance in case of the hydrodesulfurization with the novel catalyst. High sulfur

content of the remaining residue in case of the supercritical water process presented

the downside of the investigated hydroconversion path. The potential of utilizing

the novel catalyst in direct hydrodesulfurization of vacuum residue for the produc-

tion of Very Low Sulfur Fuel Oil is given though at the current state of research,

catalyst cost make the process economically unfeasible. The long term price trends

for heavy fuel oils as well as future research on improvement of the novel catalyst

will show the potential for commercial applications.

Zusammenfassung

Die Erschöpfung leicht zugänglicher und leichter Rohölquellen verlagert die künftige

Rohölversorgung auf weniger attraktive schwerere Ölfraktionen. Die Aufbereitung

von Vakuumrückstand, als schwerste Fraktion des Rohöls und Sumpfprodukt der

Vakuumdestillationskolonne einer Raffinerie, hat daher in den vergangenen Jahren

besonders an Bedeutung gewonnen. Vakuumrückstand wird meist als Treibstoff

in der Schifffahrt auf offener See verwendet. Die Globalisierung und die steigende

Nachfrage am internationalen Transport von Gütern hat zu einem starken Anstieg

des Schwerölverbrauchs geführt. Schwefel, der in hohem Maße in Vakuumrückstän-

den enthalten ist, bildet bei der Verbrennung in Schiffsmotoren umweltschädliches

Schwefeldioxid. Um die steigenden Schwefeldioxidemissionen einzugrenzen, hat die

International Maritime Organization eine Reihe von Gesetzen erlassen, von denen

das letzte den maximal zulässigen Schwefelgehalt für auf internationalen Gewässern

verwendeten Schiffstreibstoffen von 3,5% auf 0,5% reduziert. Diese neuen Schwe-

felvorschriften sind am 1. Januar 2020 in Kraft getreten und haben zu steigenden

Preisen für die Kraftstoffe geführt, die den Schwefelvorschriften entsprechen. An-

dererseits ist bei Kraftstoffen mit hohem Schwefelgehalt, die den Schwefelgrenzwert

von 0,5% überschreiten, ein Preisrückgang zu verzeichnen, der sich auch auf die

Gewinnspannen der Raffinerien beim Verkauf von hochschwefelhaltigen Vakuum-

rückständen auswirkt. Die Entwicklung eines Verfahrens zur direkten Entschwe-

felung von Vakuumrückständen birgt daher ein großes wirtschaftliches Potenzial.

Bestehende Entschwefelungstechnologien würden bei der Verarbeitung von Vakuum-

rückstand stark unter der Deaktivierung des Katalysators durch hohen Asphalten,

Schwermetall und Conradson-Kohlenstoff-Gehalt leiden. Die Hydrodesulfurierung

in der Slurry-Phase bietet das Potential, einige dieser Nachteile durch die Verwen-

dung von ungeträgerten, hochdispergierten Katalysatoren und Additiven zu über-

winden. In dieser Arbeit wurden Prozessbedingungen, Katalysatoren und Additive

hinsichtlich ihrer Entschwefelungsaktivität von hochschwefelhaltigem Vakuumrück-

stand untersucht und bewertet. Die Experimente wurden in einem 2 l Semi-Batch-

Slurry-Reaktor durchgeführt, der die Möglichkeit hatte, Öl und Rückstandsfraktio-

nen getrennt zu sammeln. Bei niedriger Temperatur, langer Verweilzeit und ho-

hem H2-Partialdruck konnten unerwünschte Konversionsreaktionen minimiert und

Entschwefelungsreaktionen maximiert werden. Das Screening katalytisch aktiver

Substanzen ergab einen Katalysator und ein Additiv, die sehr unterschiedliche

Auswirkungen auf Umwandlungs- und Entschwefelungsreaktionen zeigten als die

restlichen getesteten Katalysatoren. Überkritisches Wasser, als sehr billiges Addi-

tiv, unterstützte ausschließlich Konversionsreaktionen und lieferte große Mengen an

leichter siedendem Öl, während ein Rückstand zurückblieb, der einen sehr hohen

Schwefelgehalt aufwies. Untersuchungen an schwefelhaltigen Modellverbindungen

untermauerten die Abwesenheit von Entschwefelungsaktivitäten des überkritischen

Wassers. Ein neuartiger Katalysator zeigte eine sehr starke Hydrierungsaktivität,

wodurch die Schwefelbindungen destabilisiert wurden und die Entschwefelung des

hochschwefelhaltigen Vakuumrückstandes ermöglicht wurde. Eine Entschwefelung

von über 90% wurde erreicht, während bei 400 ◦Cund 310 bar unerwünschte Kon-

versionsreaktionen unter 30% gehalten werden konnten. Die Reaktionsnetzwerke

hinter der Konversion mittels überkritischem Wasser und der Entschwefelung mit

dem neuartigen Katalysator wurden intensiv untersucht. Die Ergebnisse bildeten

die Grundlage für eine Wirtschaftlichkeitsbetrachtung beider Prozesse. Auf Ba-

sis einer 50 000 b/d Vakuumrückstands-Veredelungsanlage wurden die beiden Fälle

berechnet. Die Kosten des neuartigen Katalysators stellen einen großen Nachteil

für die Wirtschaftlichkeit der Entschwefelungsanlage dar. Der hohe Schwefelgehalt

des verbleibenden Rückstandes im Falle des Konversionsprozesses mit überkritis-

chem Wasser stellt die Kehrseite des untersuchten konversionsprozesses dar. Das

Potential der Verwendung des neuartigen Katalysators bei der direkten Entschwe-

felung von Vakuumrückständen zur Herstellung von sehr schwefelarmem Treibstoff

ist gegeben, obwohl beim derzeitigen Forschungsstand die Katalysatorkosten das

Verfahren wirtschaftlich nicht realsierbar machen. Die langfristigen Preistrends für

schwefelreiche und schwefelarme Schiffstreibstoffe sowie zukünftige Forschungen zur

Verbesserung des neuartigen Katalysators werden das Potential für kommerzielle

Anwendungen zeigen.

Contents

1 Introduction 1

2 Background 5

2.1 Crudeoil.................................. 5

2.1.1 Sulfurincrudeoil......................... 6

2.1.2 Vacuumresidue.......................... 7

2.1.3 Heavyfueloil........................... 13

2.2 Refinery.................................. 14

2.2.1 Distillation ............................ 14

2.2.2 Thermalcracking......................... 16

2.2.3 Catalytic cracking . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.2.4 Hydroprocessing ......................... 18

2.2.5 Catalytic reforming and isomerization . . . . . . . . . . . . . . 19

2.2.6 Alkylation and polymerization . . . . . . . . . . . . . . . . . . 21

2.3 Catalysts and additives . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2.3.1 Supercritical water . . . . . . . . . . . . . . . . . . . . . . . . 23

2.4 Chemistry and thermodynamics . . . . . . . . . . . . . . . . . . . . . 25

2.4.1 Thermalcracking......................... 25

2.4.2 Catalytic cracking . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.4.3 Hydrodesulfurization . . . . . . . . . . . . . . . . . . . . . . . 27

2.4.4 Bondenergy............................ 31

2.5 Reaction mechanisms for heavy feedstock hydroconversion and hy-

drodesulfurization............................. 32

II Contents

3 State of technology of hydrodesulfurization and hydroconversion

technologies 34

3.1 Desulfurization technologies for lighter fractions . . . . . . . . . . . . 35

3.1.1 Hydrodesulfurization . . . . . . . . . . . . . . . . . . . . . . . 35

3.2 Residueupgrading ............................ 38

3.2.1 VEBA OEL Combi Cracker (VCC) . . . . . . . . . . . . . . . 41

3.2.2 ENIEST ............................. 42

3.2.3 UOPUniflex ........................... 43

3.3 Summary of the state of technology . . . . . . . . . . . . . . . . . . . 45

4 State of research on hydroprocessing catalysts and additives 46

4.1 Supportedcatalysts............................ 48

4.2 Unsupported catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . 49

4.2.1 Iron based catalysts . . . . . . . . . . . . . . . . . . . . . . . . 49

4.2.2 Molybdenum based catalysts . . . . . . . . . . . . . . . . . . . 50

4.2.3 Supercritical water residue upgrading . . . . . . . . . . . . . . 52

4.3 Conclusion from recent research . . . . . . . . . . . . . . . . . . . . . 53

5 Material and methods 55

5.1 Hydrogenationset-up........................... 55

5.1.1 Liquidaddition .......................... 58

5.2 Safetyprecautions ............................ 58

5.3 Materials ................................. 60

5.3.1 Vacuumresidue.......................... 60

5.3.2 Model components . . . . . . . . . . . . . . . . . . . . . . . . 60

5.3.3 Catalysts and additives . . . . . . . . . . . . . . . . . . . . . . 61

5.4 Experimental procedure . . . . . . . . . . . . . . . . . . . . . . . . . 62

5.4.1 Operation of the hydrogenation setup . . . . . . . . . . . . . . 62

5.4.2 Productanalysis ......................... 66

5.5 Calculations................................ 70

5.5.1 Productyields........................... 70

5.5.2 Conversion............................. 72

5.5.3 Desulfurization .......................... 72

Contents III

6 Results and discussion 73

6.1 Investigations without catalyst . . . . . . . . . . . . . . . . . . . . . . 73

6.1.1 Influence of pressure, H2throughput, agitator rotation speed,

and vacuum residue amount . . . . . . . . . . . . . . . . . . . 74

6.1.2 Influence of residence time . . . . . . . . . . . . . . . . . . . . 79

6.1.3 Influence of temperature . . . . . . . . . . . . . . . . . . . . . 83

6.1.4 Determination of a reaction mechanism and reaction kinetics . 84

6.2 Catalystscreening ............................ 96

6.2.1 Evaluation of the catalysts screening . . . . . . . . . . . . . . 96

6.3 Experiments with supercritical water . . . . . . . . . . . . . . . . . . 102

6.3.1 Experimental procedure . . . . . . . . . . . . . . . . . . . . . 103

6.3.2 Effect of temperature . . . . . . . . . . . . . . . . . . . . . . . 105

6.3.3 Combination of catalyst and supercritical water (SCW) . . . . 105

6.4 Experiments with catalyst B14 . . . . . . . . . . . . . . . . . . . . . 114

6.4.1 Experimental results . . . . . . . . . . . . . . . . . . . . . . . 114

6.4.2 Interpretation of the observations . . . . . . . . . . . . . . . . 117

6.5 Dataanalysis ...............................122

6.5.1 Total sulfur removal and sulfur removal from the different

productfractions .........................122

6.5.2 H2consumption, conversion and sulfur removal . . . . . . . . 123

6.5.3 Asphaltene and maltene conversion . . . . . . . . . . . . . . . 125

6.5.4 Denitrification...........................126

6.5.5 Conclusion from the data analysis . . . . . . . . . . . . . . . . 127

7 Economic evaluation of findings 128

7.1 Boundaries and assumptions . . . . . . . . . . . . . . . . . . . . . . . 129

7.2 Cost estimation for the large scale process . . . . . . . . . . . . . . . 130

7.2.1 Capitalcost............................130

7.2.2 Variablecost ...........................134

7.3 Economic analysis of the desulfurization process with catalyst B14 . . 135

7.4 Economic analysis of SCW upgrading . . . . . . . . . . . . . . . . . . 136

7.5 Evaluation and comparison . . . . . . . . . . . . . . . . . . . . . . . . 137

8 Summary and outlook 141

IV Contents

A Appendix 146

A.1 Determination of kinetic parameters . . . . . . . . . . . . . . . . . . . 147

A.2 List of product specifications . . . . . . . . . . . . . . . . . . . . . . . 148

A.3 GC/MSofoil ...............................150

A.4 GC/MSofoil ...............................151

A.5 Datapreparation .............................152

A.6 MATLABscripts .............................152

A.7 ExperimentalSetup............................158

Bibliography 161

List of Figures

1.1 Implementation of sulfur limits by the IMO globally and in Emission

ControlledAreas(EMA)......................... 2

1.2 Hypothetical vacuum residue molecule with sulfur and the ideal re-

moval of sulfur via hydrodesulfurization (HDS) . . . . . . . . . . . . . 3

2.1 Sulfur containing structures in crude oil . . . . . . . . . . . . . . . . . 6

2.2 Distribution of heteroatoms in crude oil according to the boiling point

fraction .................................. 7

2.3 Solventfractionation ........................... 8

2.4 Atomic Force Microscopy of crude oil asphaltenes . . . . . . . . . . . 9

2.5 Hypothetical structure of different asphaltene molecules (the author

does not claim that these exact molecules exist in reality) . . . . . . . 10

2.6 Hypothetical structure of different maltene molecules with (a) resin,

(b) aromatic, (c) resin and (d) saturate . . . . . . . . . . . . . . . . 12

2.7 Basic flowsheet of a refinery . . . . . . . . . . . . . . . . . . . . . . . 14

2.8 Vacuum distillation column and the resulting product streams . . . . 16

2.9 Reactions taking place in the isomerization unit . . . . . . . . . . . . 20

2.10 Polymerization reaction . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.11 Alkylation reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.12 Phase diagram of water showing the critical point . . . . . . . . . . . 23

2.13 Critical curve of the binary mixture H2O - H2............. 26

2.14 Desulfurization reactions via hydrogenation and hydrolysis route for

(a) Benzothiophene (BT) and (b) Dibenzothiophene (DBT) based on

[1,2] .................................... 29

2.15 Stability of sulfur species . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.16 Influence of functional group on the bond dissociation energy . . . . . 31

2.17 Complex mechanistic model of heavy feedstock hydrochracking . . . . 32

VI List of Figures

2.18 Mechanistic model of heavy feedstock hydrochracking and hy-

drodesulfurization............................. 33

2.19 Reaction mechanism for desulfurization reactions . . . . . . . . . . . 33

3.1 Process flow scheme for a HDS unit . . . . . . . . . . . . . . . . . . . 36

3.2 Typical fixed-bed reactor for HDS . . . . . . . . . . . . . . . . . . . . 36

3.3 Residue upgrading technologies according to their operating pressure

andtemperature ............................. 39

3.4 Different reactor types - left: slurry reactor; middle: fixed bed reactor

(counter-current flow); right: ebullated bed . . . . . . . . . . . . . . . 40

3.5 Process scheme of the VCC technology for high conversion of residual

feedstock ................................. 41

3.6 Simplified process scheme of the EST technology . . . . . . . . . . . 43

3.7 Process scheme of the Uniflex technology for high conversion of resid-

ualfeedstock ............................... 44

5.1 PFD of the experimental setup . . . . . . . . . . . . . . . . . . . . . 56

5.2 Reactor with internals . . . . . . . . . . . . . . . . . . . . . . . . . . 58

5.3 Explosion limits for H2-O2mixtures at different pressure . . . . . . . 59

5.4 Model compounds used to represent sulfur occurring in vacuum residue 60

5.5 EDX signal and corresponding BSE image . . . . . . . . . . . . . . . 61

5.6 Process steps for each experiment . . . . . . . . . . . . . . . . . . . . 63

5.7 Typical temperature and pressure evolution throughout an experi-

mentwithoutSCW............................ 65

5.8 Laboratory vacuum distillation . . . . . . . . . . . . . . . . . . . . . 66

5.9 Nomograph for conversion of the boiling point . . . . . . . . . . . . . 67

5.10 Comparison of calculated residence time curve and actually measured

values ................................... 70

6.1 Variation of (a) vacuum residue amount while simultaneously adjust-

ing the H2to maintain a constant ratio, (b) H2- vacuum residue

ratio, (c) agitator rotation speed effect on desulfurization and conver-

sion, and (d) agitator rotation speed effect on temperature difference

between wall and inside . . . . . . . . . . . . . . . . . . . . . . . . . 75

6.2 Variation of pressure for 12 h experiments at 385 ◦C.......... 77

6.3 Coke deposition on the stirrer formed during an experiment at 415 ◦C

and 200 bar after 12 h residence time . . . . . . . . . . . . . . . . . . 77

6.4 Conversion to lighter boiling products plotted over residence time . . 79

List of Figures VII

6.5 Desulfurization of the product fractions oil and residue plotted over

theresidencetime............................. 80

6.6 Influence of residence time on desulfurization of (a) the Cold High

Pressure Separator (CHPS) oil fraction and (b) the distilled heavy oil

fraction .................................. 81

6.7 (a) H/C ratio plotted over the residence time and (b) increase of H2

consumption with increasing gas production . . . . . . . . . . . . . . 82

6.8 Variation of temperature with 3 h residence time . . . . . . . . . . . 83

6.9 (a) H/C ratio of the residue plotted over the reaction temperature

and (b) ratio of asphaltenes to maltenes in residue . . . . . . . . . . . 84

6.10 Reaction mechanism of the four lumps . . . . . . . . . . . . . . . . . 85

6.11 Simplified reaction scheme for the conversion of vacuum residue to

lowerboilingproducts .......................... 86

6.12 Reaction mechanism for desulfurization reactions . . . . . . . . . . . 86

6.13 Reaction mechanism for desulfurization reactions considering to par-

allelreactions ............................... 86

6.14 Determination of the reaction order for the conversion reaction at

385 ◦Cand 415 ◦C............................. 89

6.15 Determination of the reaction order for the desulfurization reaction

at 385 ◦Cand 415 ◦C........................... 89

6.16 Determination of the activation energy for both the conversion reac-

tions and the desulfurization by plotting ln(k) over 1/T . . . . . . . . 90

6.17 Determination of the accuracy of the reaction rate expressions for

conversion (top) and desulfurization (bottom) . . . . . . . . . . . . . 92

6.18 H2S, CH4and C2H6concentration measured in product gas for a 3 h

experiment at 310 bar and 415 ◦C.................... 93

6.19 Desulfurization over conversion for experiments without a catalyst . . 94

6.20 Effect of cracking of the feedstock on the desulfurization . . . . . . . 94

6.21 Conversion and residue desulfurization for the tested catalysts . . . . 97

6.22 Comparison of chromatograms obtained from oil fractions from ex-

periments with different catalysts . . . . . . . . . . . . . . . . . . . . 99

6.23 GC/MS chromatograms of the light (a) and heavy (b) oil fractions and

the identified sulfur species for the experiment with activated lignite

(Herdofenkoks) (HOK) and the process conditions from parameter set 1100

6.24 Paths for cooling and depressurization with SCW . . . . . . . . . . . 103

6.25 Foam produced from SCW and vacuum residue . . . . . . . . . . . . 104

VIII List of Figures

6.26 Influence of temperature on the (a) conversion reactions, (b) conver-

sion of asphaltenes and maltenes, (c) total organic sulfur content and

(d) sulfur content in the residue fraction for experiments with SCW

in comparison with no additive . . . . . . . . . . . . . . . . . . . . . 106

6.27 Results from experiments with model components . . . . . . . . . . . 107

6.28 Results from experiments with crude oil based vacuum residue; Top:

product fractions asphaltenes, maltenes, oil and gas; Bottom: the

sulfur distribution among the fractions . . . . . . . . . . . . . . . . . 108

6.29 gas chromatography - mass spectrometry (GC/MS) analysis of the oil

fraction obtained from an experiment with SCW . . . . . . . . . . . . 109

6.30 Possible mechanistic pathway for an asphaltene molecule reacting un-

derSCWconditions............................111

6.31 Gas composition throughout the residence time for a) H2S for model

compounds, b) H2S for vacuum residue, c) CH4for vacuum residue

and d) C2H6for vacuum residue . . . . . . . . . . . . . . . . . . . . . 112

6.32 SEM images of (a) virgin HOK (300 m2/g) and (b) HOK after being

exposed to the process conditions (60 m2/g) ..............113

6.33 Behavior of gas composition over time with use of 7.5% catalyst B14

(a) H2S, CH4and C2H6(b) total mass flow out of the system . . . . . 115

6.34 Repeatability of the detected gas composition with catalyst B14 . . . 116

6.35 H/C ratio of the residue fraction and asphaltene content . . . . . . . 118

6.36 Comparison of the H2S content in the product gas during experiments

with catalyst B14 and with no catalyst . . . . . . . . . . . . . . . . . 119

6.37 Reaction path proposed for the desulfurization using catalyst B14 . . 120

6.38 Desulfurization of oil and residue over total desulfurization . . . . . . 122

6.39 Clustering data for a) 3D-plot of H2consumption, conversion and

sulfur removal in residue fraction, b) 2D-plot of H2consumption

over conversion, c) 2D-plot of H2consumption over sulfur removal

in residue fraction and d) 2D-plot of sulfur removal over over conversion124

6.40 Influence of residue fractions on conversion . . . . . . . . . . . . . . . 126

6.41 a) Clusters in residue transformation b) Desulfurization of residue

over asphaltene content . . . . . . . . . . . . . . . . . . . . . . . . . . 126

6.42 Correlation of desulfurization and denitrification . . . . . . . . . . . . 127

7.1 Boundaries of the economic evaluation . . . . . . . . . . . . . . . . . 129

7.2 Basic flow diagram of the residue desulfurization process . . . . . . . 131

7.3 Economics of the vacuum residue upgrading using catalyst B14 . . . . 135

7.4 Sensitivity analysis for the recycle ratio of catalyst B14 . . . . . . . . 136

List of Figures IX

7.5 Economics of the vacuum residue upgrading using SCW . . . . . . . . 137

7.6 Price development for HFO and VLSFO . . . . . . . . . . . . . . . . 138

7.7 Long term price trend for bunker fuels . . . . . . . . . . . . . . . . . 139

A.1 Determination of the reaction order . . . . . . . . . . . . . . . . . . . 147

A.2 Determination of the activation energy . . . . . . . . . . . . . . . . . 148

A.3 Compressor ................................158

A.4 Reactor ..................................159

List of Tables

2.1 Fuel types and sulfur regulations for the use of heavy oils as fuel on

openseas[3,4]............................... 13

2.2 ASTM boiling range of crude oil distillation products . . . . . . . . . 15

2.3 Bond dissociation energies in diatomic molecules [5] . . . . . . . . . . 31

3.1 Residue upgrading technologies . . . . . . . . . . . . . . . . . . . . . 39

4.1 Overview on relevant research . . . . . . . . . . . . . . . . . . . . . . 47

5.1 Composition of vacuum residue . . . . . . . . . . . . . . . . . . . . . 60

5.2 Catalysts and additives . . . . . . . . . . . . . . . . . . . . . . . . . . 61

5.3 Inlet temperature for chosen hydrogen throughput . . . . . . . . . . . 64

6.1 Values for the reaction rate coefficient for the conversion kcon(T) and

the desulfurization reactions kdesulf (T) their logarithmic ln(kcon(T))

ln(kdesulf (T)) ............................... 90

6.2 Literatur values for the activation energy EA............. 91

6.3 k0and EAfor conversion and desulfurization . . . . . . . . . . . . . . 91

6.4 Choice of process parameters to achieve good desulfurization and high

conversion................................. 95

6.5 Results from catalyst screening . . . . . . . . . . . . . . . . . . . . . 96

6.6 Performed experiments for SCW combinations . . . . . . . . . . . . . 106

6.7 Desulfurization results with catalyst B14 . . . . . . . . . . . . . . . . 114

6.8 Variation of the catalyst composition for B14 . . . . . . . . . . . . . . 116

6.9 Improvements for catalyst B14 at 400 ◦C................117

6.10 Residue analysis from catalyst B14 with * as the molar ratio . . . . . 118

6.11 Centroids for the clusters conversion - sulfur reduction residue - hy-

drogenconsumption ...........................123

List of Tables XI

7.1 Key figures from the experiments the economic calculations are based

on .....................................130

7.2 Unit operations in the large scale plant together with the correspond-

ing capacity for each unit operation and the resulting cost . . . . . . 133

7.3 OPEXforbothcases...........................134

7.4 Produced products for both cases . . . . . . . . . . . . . . . . . . . . 134

7.5 Economic comparison of both cases . . . . . . . . . . . . . . . . . . . 138

8.1 Choice of process parameters to achieve good desulfurization and high

conversion.................................141

A.1 List of product specifications; not for every experiment all product

specifications were retrieved . . . . . . . . . . . . . . . . . . . . . . . 149

A.2 Table of identified compounds from the oil fraction from experiment

withHOK.................................150

A.3 Table of identified compounds from the oil fraction from experiments

withcatalystB14.............................151

Abbreviations

AFM atomic force microscope

AGO Atmospheric Gas Oil

ASTM American Society for Testing and Materials

BET Brunauer-Emmett-Teller

BP boiling point

BSE back-scattered electrons

BT Benzothiophene

CAPEX capital expenditures

CHPS Cold High Pressure Separator

COS carbonyl sulfide

CSTR continuous stirred-tank reactors

DBT Dibenzothiophene

DDS direct desulfurization

DMDS Dimethyldisulfide

EAC equivalent annual cost

ECA Emission Controlled Areas

EBR ebullated bed reactor

EDX energy-dispersive X-ray spectroscopy

EGCS exhaust gas purification systems

EMA Emission Controlled Area

EST ENI Slurry Technology

FCC Fluid Catalytic Cracking

GC gas chromatography

GC/MS gas chromatography - mass spectrometry

IMO International Maritime Organization

IFO Intermediate Fuel Oils

HDCCR Hydro-removal of Conradson Carbon

HDM hydrodemetallization

HDN hydrodenitrification

HDO hydrodeoxidization

List of Tables XIII

HDS hydrodesulfurization

HFO Heavy Fuel Oil

HPLC high-pressure liquid chromatography

HOK activated lignite (Herdofenkoks)

HS hydrogen sulfide

HSFO High Sulfur Fuel Oil

HYD Hydrogenation

LEL lower explosion limit

LNG Liquefied Natural Gas

LSF2020 Low Sulfur Fuel Regulations

LSFO Low Sulfur Fuel Oil

MARPOL International Convention for the Prevention of Pollution from Ships

(Maritime Pollution)

MDO Marine Diesel Oil

MGO Marine Gas Oil

ODS oxidative desulfurization

OPEX operational expenditures

PFD process flow diagram

PID piping and instrumentation diagram

RFCC Residue Fluid Catalytic Cracking

ROI return on invest

SEM scanning electron microscope

SCW supercritical water

TIC total investment cost

ULSFO Ultra Low Sulfur Fuel Oil

UEL upper explosion limit

VCC VEBA Combi Cracking

VGO Vacuum Gas Oil

VLSFO Very Low Sulfur Fuel Oil

VR Vacuum Residue

Symbols

Roman letters

Sign Description Unit

CCapacity t/a

cConcentration mol/l

EAActivation energy kJ/mol

IInvestment cost e

KwSelf-dissociation constant of water -

kpre-exponential factor kg/kg*h (de-

pending on

reaction order

MMolar mass g/mol

mMass kg

nH/Cmolar ratio H/C mol/mol

pPressure bar

RMolar gas constant J/mol*K

rrate of reaction kg/kg*h (de-

pending on

reaction order

TTemperature ◦C

tTime h

Greek letters

Sign Description Unit

δp Pressure drop bar

∆SDesulfurization kg/kg

τResidence time s

ωMass fraction kg/kg

List of Tables XV

Subscripts

Sign Description

0Initial value

AReacting specie

feed Feed

gas Gas fraction

istands for the different fractions residue, oil and gas

nNumber - denotes the number of a state, order of

reaction etc.

org atom chemically integrated into organic matter

oil Oil fraction

res Residue fraction

Chapter 1

Introduction

Though non-fossil fuels are increasingly gaining importance in the transportation

sector, combustion engines driven by crude oil based fuels are still dominant. Espe-

cially in the shipping sector, where mostly heavy oils are used, a strong increase in

demand is predicted, which is estimated at 200 Mt/a in 2020 [6,7]. Marine engines

mostly run on Heavy Fuel Oil (HFO) mainly composed of vacuum residue. De-

pending on the crude oil source, the fraction of vacuum residue varies, but globally

around 15% of the annual crude oil consumption is vacuum residue (750 Mt/a in

2020 [8]). Naturally, this heavy fraction of crude oil is high in undesired species,

especially in sulfur.

On January 1st, 2020 new Low Sulfur Fuel Regulations (LSF2020) came into force

reducing the sulfur cap from 3.5% to 0.5% [4]. These regulations are the latest in

a series of International Maritime Organization (IMO) measures to reduce marine

pollution in response to the threat of air pollution. The LSF2020 emissions legisla-

tion means that ships must significantly reduce their emissions both on open seas

and in coastal areas.

The question for both refineries and shipping industry now is how to comply with

these new regulations. Generally, three options are discussed:

•Switching to a different fuel with lower sulfur content - Marine Gas Oil (MGO)

or Marine Diesel Oil (MDO)

•Installation of exhaust gas purification systems (EGCS) on board the ship

•Switching to Liquefied Natural Gas (LNG) powered ships

All three options are associated with considerable costs and changes in infrastructure

and therefore present refineries as well as shipping companies with major economic

2 Chapter 1 Introduction

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

2005 2010 2015 2020 2025 2030

Sulfur limit (wt-%)

Yea r

Global

EMA (Emission Controlled Areas)

Figure 1.1: Implementation of sulfur limits by the IMO globally and in EMA

challenges. An alternative with fewer downsides would be of great benefit for the

sector.

A different option could be the removal of sulfur directly from the vacuum residue

which could then be further used as Very Low Sulfur Fuel Oil (VLSFO) comply-

ing with the new sulfur regulations. The complete infrastructure for fuel handling,

starting at the refinery and ending with the consumption of fuel in heavy engines

of ships, would not have to undergo any costly changes. Additionally, the precise

removal of sulfur from vacuum residue without conversion to lower boiling prod-

ucts could reduce H2consumption during the hydrotreating process thus reducing

cost for refineries. This could lead to a less expensive supply chain for maritime

fuel, compared to the scenario that the shipping industry switches to lower boiling

products like MGO.

Around the world, refineries are equipped with desulfurization units - mostly in the

form of hydrodesulfurization (HDS) processes using expensive, highly active fixed

bed catalysts - to meet the strict sulfur limits for low boiling fuels like diesel and

gasoline [9]. These processes enable the explicit removal of sulfur in the form of

H2S with minimal cracking of the feedstock. Due to the high viscosity and the

coke-forming tendencies, vacuum residue causes severe problems when being sent

over these fixed bed catalysts. Therefore, different process options are investigated

to remove sulfur from vacuum residue, ideally as presented in Fig. 1.2 with no

cracking and thus minimal H2consumption.

A potential procedure for desulfurization of vacuum residue is possible by alteration

of the Bergius-Pier process, which applies a high H2partial pressure of above 300 bar,

intermediate temperature of 400 ◦C-500 ◦C, together with a dispersed catalyst, in

Figure 1.2: The authors understanding of a hypothetical vacuum residue molecule with sulfur and

the ideal removal of sulfur via HDS (the author does not claim that these exact molecules exist in

reality)

a slurry reactor. The process was originally designed for the liquefaction of coal and

has been adopted for the conversion of other feedstocks including vacuum residue

[10]. In case of HDS, the aim is to keep conversion at minimum level, while maximiz-

ing the sulfur removal at the same time. Thermodynamically, the removal of sulfur

is preferred before rupture of simple carbon - carbon bonds, implying that theo-

retically, HDS should be possible at less severe conditions, compared to conversion

processes. Consequently, the choice of an adequate slurry phase catalyst together

with optimal choice of reaction conditions could display a route to desulfurization

of vacuum residue at minimal conversion and H2consumption.

This work focuses on the investigation of HDS of vacuum residue in a slurry reactor.

Optimization of the process conditions in a semi-continuous stirred tank reactor

was performed and global reaction kinetics of HDS and hydroconversion reactions of

high sulfur vacuum residue were studied and are presented in Chapter 6.1. Several

catalysts and additives were tested an their suitability for HDS of vacuum residue

was evaluated in Chapter 6.2. A novel catalyst for slurry phase HDS was developed

and the fundamental reaction mechanism is discussed in Chapter 6.4. In Chapter 7

4 Chapter 1 Introduction

the economic feasibility of applying this novel catalyst at large scale HDS of vacuum

residue is evaluated.

Chapter 2

Background

2.1 Crude oil

Crude oil is a natural product and consists of numerous different hydrocarbons rang-

ing from CH4to molecules with hundreds of carbon atoms. Because of its natural

offspring, the diverse structure also includes heteroatoms like oxygen, nitrogen, and

sulfur, as well as metals like vanadium and nickel [11].

The composition of crude oil may vary strongly, also affecting its physical properties

including the boiling range, carbon chain length as well as asphaltene content. In

its naturally occurring state, crude oil is of low value. It is therefore sent to a refin-

ery, where separation and upgrading into different marketable products like diesel,

gasoline, waxes, and heavy fuel oil are performed. In total, refinery products supply

a third to half of the worlds energy supply [7]. Removal of undesired heteroatoms,

of which sulfur is the most prominent, from crude oil is also one major task for

refiners. Marketable HFO, as one fraction resulting from the refining process, is the

main focus of the sulfur removal addressed in this work since HFO mainly consists

of vacuum residue [12]. Sulfur and sulfur chemistry found in crude oil, together

with the path molecules take within a refinery that end up in the vacuum residue

fraction, are essential for the understanding of the nature of the vacuum residue

and the complexity of the sulfur removal from this fraction. Therefore, in Sections

2.1.1 and 2.1.2 an overview on the nature of sulfur in crude oil and the properties

of vacuum residue are given, followed by a short description of a refinery in Section

2.2.

6 Chapter 2 Background

Thiole

Sulfide

Disulfide

Thiane

Thiophene

Benzothiophene

Dibenzothiophene

Figure 2.1: Sulfur containing structures in crude oil

2.1.1 Sulfur in crude oil

Of the different heteroatoms found in crude oil, sulfur is often the most concentrated

[13]. Sulfur content may vary significantly and correlates strongly with the gravity

of the feedstock ranging up to 20 wt%. Below a sulfur level of 1 wt% the crude is

referred to as low sulfur; above, as high sulfur. Sulfur may occur as elemental sulfur,

2.1 Crude oil 7

carbonyl sulfide (COS), inorganic sulfur, including hydrogen sulfide (HS) or, the

largest fraction, as heteroatoms within the carbon matrix. This largest fraction is

comprised of mercaptans, sulfides as well as polycyclic sulfides [14].

Figure 2.1 shows the main organic sulfur bond types existing in crude oil. While

most of the elemental sulfur, hydrogen sulfide, as well as cabonyl sulfide, are not

included in vacuum residue after distillation, the portion of sulfides, thioles, and

polycyclic sulfides increase. The predominant structures present in vacuum residue

are BT and DBT derivatives built into complex macro molecules [15].

2000%

4000%

6000%

8000%

10000%

12000%

0% 20% 40% 60% 80% 100%

Increasing aromaticity, decreasing hydrogen content

Increasing nitrogen sulfur and metals content

Increasing boiling point and molecule size

Distillable fractions Maltenes and Asphaltenes

Figure 2.2: Distribution of heteroatoms in crude oil according to the boiling point fraction

2.1.2 Vacuum residue

Vacuum residue is the crude oil fraction with the highest initial boiling point and

therefore the bottom product from the vacuum distillation step, which is further

described in Section 2.2.1. Depending on the conditions applied in the vacuum

distillation, the initial boiling point of the vacuum residue varies but it is mostly

above 500 ◦C. Because it is the bottom product of the distillation, it accumulates

most of the undesired and potentially catalyst impairing, as well as environmentally

polluting components like sulfur and heavy metals. Vacuum residue is of high vis-

cosity and almost solid at room temperature due to the large molecule sizes. Its

chemical composition is complex and because of low or not existing volatility of the

constituents, an exact characterization of the material is difficult. The quality of

8 Chapter 2 Background

the vacuum residue is specified by the quality of the crude oil used in the refinery

and with decreasing amounts of crude oil reservoirs the crude oil quality is more

and more becoming of heavier composition. Therefore, upgrading of the residue is

gaining higher importance.

Though the characterization of vacuum residue is difficult, one can distinguish dif-

ferent molecule classes present in heavy oils and residua according to their solubility.

Several, different procedures exist for the separation [16–20] and depending on the

applied method, the cut point between fractions varies. Therefore, a precise ex-

planation of the used method is important. The method applied in this work is

described in Section 5.4.2.2. A simplification indicating, which fraction is soluble in

which solvent is presented in Fig. 2.3.

Heavy oil

Asphaltenes /

Coke

Maltenes

Aromats ResinsSaturates Asphaltenes Coke

soluble insoluble

n-heptane/n-hexane

toluene

soluble

methanol

soluble

n-heptane

soluble soluble insoluble

toluene

Figure 2.3: Solvent fractionation

From the properties of the solvents, in which the different fractions dissolve, one can

conclude the character of the respective fraction.

2.1.2.1 Coke

Coke will mostly not be present in crude oil but can form when retrograde reactions

take place and no hydrogen is present to saturate the forming radicals. Coke is a

solid carbon structure insoluble in solvents. For most processes, formation of coke

is undesired. Especially in HDS reactions and processes, where catalysts are used,

the formation of coke is very problematic and process conditions are chosen in such

a manner that reactions leading to coke are minimized.

2.1 Crude oil 9

2.1.2.2 Asphaltenes

Asphaltenes are the heaviest fraction (molar mass of 500 - 3,000 g/mol) found in

crude oil and are composed of macromolecular carbon structures containing large

amounts of heteroatoms (sulfur, oxygen and nitrogen) and heavy metal (nickel, iron,

vanadium) giving them a polar character. They are predominantly black or dark

brown solids (non-volatile) with a very low H/C ratio, strongly influencing the high

viscosity of vacuum residue [21–27]. DIN 51595 defines asphaltenes as the insoluble

components of a crude oil sample diluted with 30 times the amount of a non-polar

solvent (n-heptane) at temperatures between 18 and 28 ◦C[16]. Asphaltenes are

therefore not an organic group of substances in the sense of the term, because they

are not defined chemically by their molecular or chemical structure, but by their

physical properties. Therefore, also the molecular and chemical structure of an as-

phaltene molecule is discussed intensively in literature and several different example

molecules with varying characteristics can be found in literature [28–34]. In Fig. 2.4

atomic force microscope (AFM) pictures are shown of different asphaltenes found in

crude oil.

Unraveling the Molecular Structures of Asphaltenes by Atomic Force

Microscopy

Bruno Schuler,*

,†

Gerhard Meyer,

†

Diego Pena,

‡

Oliver C. Mullins,

and Leo Gross*

,†

†

IBM Research −Zurich, Saumerstrasse 4, 8803 Ruschlikon, Switzerland

‡

CIQUS and Facultad de Química, Universidad de Santiago de Compostela, E-15782 Santiago de Compostela, Spain

Schlumberger-Doll Research, Cambridge, Massachusetts 02139, United States

SSupporting Information

ABSTRACT: Petroleum is one of the most precious and

complex molecular mixtures existing. Because of its chemical

complexity, the solid component of crude oil, the asphaltenes,

poses an exceptional challenge for structure analysis, with

tremendous economic relevance. Here, we combine atomic-

resolution imaging using atomic force microscopy and

molecular orbital imaging using scanning tunnelling micros-

copy to study more than 100 asphaltene molecules. The

complexity and range of asphaltene polycyclic aromatic

hydrocarbons are established in detail. Identifying molecular

structures provides a foundation to understand all aspects of petroleum science from colloidal structure and interfacial

interactions to petroleum thermodynamics, enabling a ﬁrst-principles approach to optimize resource utilization. Particularly, the

ﬁndings contribute to a long-standing debate about asphaltene molecular architecture. Our technique constitutes a paradigm shift

for the analysis of complex molecular mixtures, with possible applications in molecular electronics, organic light emitting diodes,

and photovoltaic devices.

■INTRODUCTION

In nature, molecules generally exist in mixtures. Petroleum is

probably the most prominent of such mixtures and one of the

most complex materials encountered with possibly over

100,000 distinct chemical constituents.

1,2

The primary

unresolved component of crude oil is asphaltene.

2−4

Under-

standing the structure of asphaltenes is of immense economic

importance

2,4−8

and a prerequisite to establishing the

structure−function relationship in petroleomics,

but their

molecular architecture has been subject to a long-standing

debate.

3,9−13

Speciﬁcally, some studies indicate that individual

asphaltene molecules contain primarily one polycyclic aromatic

hydrocarbon (PAH)

3,12

(“island”), while other studies indicate

that structures with multiple PAHs (“archipelago”) contrib-

ute.

Diﬃculties associated with resolving this issue include

formation of archipelago from island structures in experi-

ments

and the potential inability to disaggregate asphaltene

aggregates.

Recent experiments using laser desorption have

established disaggregation of asphaltenes and have obtained

dominance of island structures,

consistent with other mass

spectral measurements.

Asphaltenes are linked to many key economic issues in the

petroleum industry today. Unwanted asphaltene phase

transitions hinder petroleum production, transportation, and

reﬁning.

2,4

Asphaltene interfacial activity with rocks aﬀects

wettability, oﬀering a focal point for enhanced oil recovery.

New asphaltene thermodynamics is used to evaluate the extent

of ﬂuid equilibrium in reservoirs, indicating ﬂow connectivity,

the most important reservoir uncertainty.

In addition,

understanding the chemical processes that occur in oil

reservoirs

6,7

is improved by accurate structural characterization.

The capability to address all of these concerns is founded on an

accurate representation of asphaltene molecules. However, the

structure analysis of asphaltenes has posed an exceptional

challenge because of their chemical complexity that is only now

being resolved.

Scanning probe microscopy oﬀers the unique capability of

imaging single adsorbates at the atomic scale. The character-

ization of asphaltenes had been attempted with scanning

tunnelling microscopy (STM),

20,21

but to date no atomic-

resolution could be achieved on asphaltene molecules. Recent

progress in atomic force microscopy (AFM) enabled visual-

ization of the atomic structure of individual molecules in real

space.

This method was also used to analyze bond order,

identify the molecular structure of natural compounds

24,25

and

graphene nanoribbons,

and detect products of chemical

synthesis

and on-surface reactions.

28,29

By using STM and

ultrathin insulating ﬁlms as a substrate, one can also map

molecular orbitals.

30−32

Here, we present atomic-resolution low-temperature AFM

data of individual molecules of asphaltene, one of the most

complex and intriguing natural mixtures existing. Additionally,

orbital imaging with the STM is used to access the polycyclic

Received: April 20, 2015

Published: July 14, 2015

Article

pubs.acs.org/JACS

J. Am. Chem. Soc. 2015, 137, 9870−9876

Downloaded via TU BERLIN on November 29, 2019 at 12:04:03 (UTC).

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

the surface compared to the PAH.

46,47

The side-chains

sometimes also undergo tip-induced conformational changes

when being scanned in STM or AFM mode. As shown in

Figure 5a, they often appear as zigzag-like patterns, associated

with alkanes. In STM orbital images, these side-chains do not

contribute (see Figure 5e,f). A similar example is PA2 (Figure

5g−i), which has a side-chain of about 20 Å in length. Also

here, the PAH core alone composes the orbital, which allows

conclusions on its structure. Imaging orbitals are particularly

important for PA as the peripheral alkanes can make

interpretation of AFM images of PA diﬃcult. Typically, PAs

also exhibit more substituents and ﬁve-membered rings than

CAs do. Exemplary, PA3, shown in Figure 5j−l, features three

such pentagonal rings.

In summary, we ﬁnd that asphaltene molecules consist of a

central aromatic core with peripheral alkane chains. In some

cases, this central core is divided into several distinct PAHs

connected by a single bond, which proves the presence of

“archipelago”-type molecules. Nevertheless, a single aromatic

core with peripheral alkanes is the dominant asphaltene

molecular architecture, proving the main aspects proposed by

the Yen−Mullins model.

The diverse PAH architecture

(number and type of rings and overall shape) of CAs and

PAs is similar, despite signiﬁcant diﬀerences in their formation

and postprocessing. The main diﬀerence is the presence of

longer side groups in PA. Additionally, PA contains more

substituted rings. Both, CA and PA molecules are larger than

expected from previous studies.

3,12,44

The preparation procedure and single-molecule analysis

presented here might raise doubts concerning the signiﬁcance

for the vast set of molecule structures in the mixture. In the

following we address the main issues connected with the

preparation and discuss the beneﬁts and limitations related to

our single-molecule approach. Thermally evaporating the

molecules in UHV involves mainly two concerns: (i)

sublimation of volatile, light components (about 100 u or

less) in UHV at room temperature before evaporation onto the

sample and (ii) dissociation of large molecules, which crack at

temperatures lower than their sublimation temperature. The

latter is expected to occur (in part) for molecules of about 1000

u or more.

Therefore, very large asphaltenes will not be

detected, however, their relative abundance is known to be

small

3,9,12,16,17,48,49

and visually we do not observe any residuals

in the evaporator. Also for asphaltene samples we observe

molecules with masses above 500 u on the surface, e.g., CA6,

CA11,CA17,CA18,PA1, and PA14 indicating that molecules

of this size can be deposited without fragmentation. Non-

covalently bonded globular aggregates that were shown to

comprise a signiﬁcant part of asphaltenes

2,15,49−51

are most

likely disaggregated upon evaporation. Note that all molecules

that are evaporated in the solid angle of the surface will be

adsorbed because both the Cu and NaCl substrates have a

sticking coeﬃcient of unity at 10 K. On the surface we ﬁnd

almost exclusively individual molecules and no larger

aggregates. The observation of no signiﬁcant qualitative

diﬀerence of the molecular structures on Cu and NaCl

indicates that bond cleavage catalyzed by the Cu(111) surface

is not an issue. However, the conformation of the molecules is

inﬂuenced by the surface, as upon adsorption the molecules

usually adopt a planarized conformation with respect to the gas

phase. To study single molecules as opposed to an entire

ensemble not only bears limitations but also unique beneﬁts:

Most importantly the existence of a speciﬁc molecule within a

complex mixture can be provided even if the molecule is rare

and the mixture divers, as it is the case in our study. On the

downside, a universal quantiﬁcation such as the mean molecular

weight or the relative abundance is not feasible. These diﬀerent

advantages and challenges with respect to conventional

techniques for structure elucidation as nuclear magnetic

resonance, X-ray diﬀraction, and mass spectrometry render

AFM a valuable complementary tool for the characterization of

molecular mixtures.

■CONCLUSIONS

This study shows that complex, polydisperse molecule mixtures

can be analyzed by AFM and STM on a single-molecule basis

with atomic-resolution. Understanding the chemistry and

physics of asphaltenes, so important in many economic

settings, requires proper characterization of the molecular

architecture, and not just in terms of the mean molecular

properties but also the range of molecular architectures found.

Molecular imaging presented herein has resolved uncertainties

about the mean molecular structures and also provided the ﬁrst

direct measurement of the tremendous range of molecular

structures in asphaltenes.

More generally, this technique might be useful to study the

traceability of a sample, that is, to characterize and identify the

Figure 5. Petroleum asphaltenes PA1-PA3. (a,b) AFM images of PA1

on NaCl(2 ML)/Cu(111) at diﬀerent set-points zfrom (I= 1 pA, V=

0.2 V). (c) Laplace-ﬁltered image of (b). (d−f) STM images (I=1

pA) of PA1 recorded at 0.2, 1.95, and −2.4 V, respectively. (g,h) AFM

image of PA2 on NaCl(2 ML)/Cu(111) (g) and its Laplace-ﬁltered

version (h). (i) STM orbital image (I= 0.6 pA) of PA2 corresponding

to the LUMO resonance. On the left side of (g−i) there is a small

patch of third-layer NaCl. (j,k) AFM images of PA3 on NaCl(2 ML)/

Cu(111) at diﬀerent set-points zfrom (I= 1.8 pA, V= 0.2 V). (l)

Laplace-ﬁltered image of (k).

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.5b04056

J. Am. Chem. Soc. 2015, 137, 9870−9876

9874

argued before). Although not all structures can be assigned based

on their respective AFM image, they contain important

information on size, planarity, compactness of each molecule

and content and size of polycyclic aromatic hydrocarbons

(PAHs), aliphatic side chains, and methyl side groups. The

complete sets of AFM images and their analysis are detailed in

the Supporting Information.

Common to all samples is that the “island”-type architec-

ture,

17−20

i.e., a central aromatic core with attached side chains, is

predominant over “archipelago”-type molecules (that were also

detected but only with occurrences of less than 10% in all

samples, e.g., B2.1). On the structural moiety level, we frequently

observe ﬂuoranthene and ﬂuorene-type motifs. For the latter, we

can distinguish at least four diﬀerent AFM patterns, which we

tentatively assign to ﬂuorene, dibenzothiophene, ﬂuorenone, and

carbazole. Note, however, that cyclopentadienone and furan are

diﬃcult to diﬀerentiate by AFM

and both might be considered

as alternative options in the assignments of Figure 3. Several

methyl groups, identiﬁed by AFM before,

attached to the PAH

are common for all eight samples, too. Interestingly many

ﬂuorene-type moieties have such a methyl group at their 1 or 8

position, for instance molecules A1.4,A2.2,C1.1,C1.2,C1.4,

and D2.3, which might indicate a speciﬁc pathway for their

formation. Five-membered rings also occur in fused pairs (C1.5

and D2.3). Occasionally we ﬁnd saturated ﬁve-membered (A1.3,

C1.4, and C2.2) and six-membered (C2.1) rings fused to

aromatic rings. These groups could be assigned based on a recent

study of aliphatic model compounds.

If longer aliphatic chains

exist in a sample, they tend to be present either as a long

unbranched strand attached to a PAH (along with methyl

groups) (A1.6,B2.3 left, and D1.2 right) or just as an isolated

alkane (A1.2,A1.5,A2.1,B1.3,B1.4, and D1.4). Such chains can

be 5−70 Å long and are typically buckled or twisted in certain

reoccurring patterns. However, often a (at least partially) straight

conformation is adopted such that a characteristic zigzag pattern

can be recognized.

5,6

The isolated alkane chains represent

coprecipitated waxes that are not asphaltenes by deﬁnition but

can be found in asphaltene samples if they are not extensively

puriﬁed.

Next, we summarize some observations made for each pair of

samples based on AFM measurements of individual molecules.

In samples Awe ﬁnd the comparably largest diversity of diﬀerent

structures: molecules with an extended aromatic core, isolated

aliphatic chains, and more three-dimensional (“bulky”) mole-

cules with roughly equal proportionality. This was characteristic

for the Asamples whereas other samples were typically

dominated by one or two out of these three types of molecule

structures. In tendency, the deposit A2 contains less bulky

molecules compared to the crude A1, which is consistent with

earlier works showing that the average molar H/C ratio for A1 is

greater than the ratio for A2.

Particular for sample B1 is that its molecules can be essentially

classiﬁed into aromatics, among which we ﬁnd even fully

aromatic molecules such as B1.1 and isolated aliphatic chains

with similar abundance. The hydroprocessed sample B2 in

contrast contains more bulky molecules at the expense of the

aromatic fraction. This can be rationalized by the hydrotreat-

ment, which adds hydrogens to the molecules. Surprisingly, we

still ﬁnd longer aliphatic side chains (B2.3 left), indicating an

incomplete cracking process.

The steam cracker tar asphaltene C1 is very well suited for

AFM-based structure identiﬁcation as most of the molecules

were highly aromatic. Accordingly, a large portion of molecules

could be assigned completely. They are more homogeneous in

size compared to the other samples and typically feature methyl

substituted aromatic hydrocarbons. Nonetheless, we found also

Figure 2. AFM measurements. Characteristic selection of CO-tip AFM raw data of individual heavy oil molecules. The labels indicate the sample name

and molecule number. For molecule A1.6 the inset shows a part of the same frame recorded at a larger scan height. The asterisk indicates a manipulation

image where the molecule rotates around a pinning site. A more comprehensive set of measurements for each sample can be found in the Supporting

Information. Scale bars: 5 Å.

Energy & Fuels Article

DOI: 10.1021/acs.energyfuels.7b00805

Energy Fuels 2017, 31, 6856−6861

6858