1566 Energy & Fuels 2003, 17, 1566-1569

Consistency of Asphaltene Chemical Structures with

Pyrolysis and Coking Behavior
Murray R. Gray*
Department of Chemical and Materials Engineering, University of Alberta,
Edmonton, Alberta T6G 2G6, Canada
Received January 17, 2003. Revised Manuscript Received August 6, 2003

Molecular representations of asphaltenes have been proposed from a variety of analytical

methods; however, any useful representation must also consider the behavior of these mixtures
upon processing. Mild thermal cracking of asphaltenes gives products that have a wide range of
boiling points and contain many chemical types, including a range of aromatic, heteroatomic,
and aliphatic components. Under severe cracking, the content of aromatic carbon correlates with
the yield of solid residue, which can range from 35% to 75%. These process observations are
most consistent with representations of asphaltenes as a variety of aromatic groups joined by
bridges and substituted by aliphatic groups.

The development of new analytical methods and
microscopy and molecular simulation techniques has
generated considerable interest in the classic problem
of asphaltene chemistry. The most contentious issues
are the true molecular weight of the asphaltenes1 and
the presence (or absence) of various sizes of aromatic
groups and bridges.2-4 The complexity of both the
chemistry of these materials and their solution and
colligative properties require that effective models for
asphaltene chemical structure include data from a
variety of sources.2 Although asphaltenes are a major
concern in production operations, because of their role
in emulsion stabilization and fouling, this fraction of
crude oil is also important in the conversion of vacuum
residues by processes such as coking, catalytic cracking,
and hydroconversion. The purpose of this paper is to
set boundaries on proposed models for the asphaltene
chemical structure, using data from studies of residue
conversion chemistry. Cracking and pyrolysis are not
preferred analytical methods for defining chemical
structure; however, any model for the asphaltene chemi-
cal structure must be consistent with the observed
reaction characteristics in regard to upgrading the
processes.
Although thousands of metric tons of asphaltenes are
cracked, coked, or hydrogenated in refineries each day,
relatively few studies have considered the conversion
of asphaltenes under normal processing conditions. Most
studies have examined entire residues, for two excellent
reasons. The first reason is that the separation of
heptane-insoluble asphaltenes from residues or crude
oil for subsequent processing studies is difficult and
* Author to whom correspondence should be addressed. E-mail:
murray.gray@ualberta.ca.
(1) Groenzin, H.; Mullins, O. C. Energy Fuels 2000, 14, 677-684.
(2) Speight, J. G. The Chemistry and Technology of Petroleum, 2nd
ed.; Marcel Dekker: New York, 1991.
(3) Murgich, J.; Abanero, J. A.; Strausz, O. P. Energy Fuels 1999,
13, 278-286.
(4) Ralston, C. Y.; Mitra-Kirtley, S.; Mullins, O. C. Energy Fuels
1996, 10, 623-630.
expensive. The secondsand more importantsreason is
that the asphaltene content is a poor prediction of
performance in catalytic cracking and coking. In these
processes, the Conradson carbon residue (CCR) or
microcarbon residue (MCR) is a more useful predictor
of the coke-forming tendency. This point is illustrated
in the residue phase diagram proposed by Wiehe.5 As
illustrated in Figure 1, the asphaltenes cluster in a
particular region of molecular weight and hydrogen
content. The upper axis shows that the CCR content of
the asphaltenes can be in the range of 40-85 wt %,
based on the sample set of Wiehe.5 If we add additional
data from studies that incorporate a wider range of
heptane-insoluble asphaltenes,6,7 the domain of asphalt-
enes in the diagram then is extended and the range of
CCR or MCR contents for asphaltenes becomes even
wider. Given that the CCR or MCR content correlates
well with some refinery processes, Figure 1 illustrates
why these measures are used frequently instead of the
asphaltene content.
A simple reason for the wide range of MCR content
in heptane-insoluble material is that the liquid-phase
solubility in a solvent at low temperature is a poor
predictor of high-temperature behavior. Neither solubil-
ity nor reactivity at temperatures of >300 °C are
predicted by the solubility in heptane at 20 °C. Variables
that affect asphaltene precipitation include aromaticity
(and, indirectly, the H/C ratio and hydrogen content1,5),
molecular weight, and polar functional groups.2,9 The
role of hydrogen bonding, because of the acids and
hydroxyl groups, is diminished significantly at 100 °C10
and would be insignificant at refinery processing tem-
(5) Wiehe, I. A. Ind. Eng. Chem. Res. 1992, 31, 530-536.
(6) Calemma, V.; Rausa, R. J. Anal. Appl. Pyrolsis 1997, 40-41,
569-584.
(7) Rahmani, S.; McCaffrey, W. C.; Dettman, H. D.; Gray, M. R.
Energy Fuels 2003, 17, 1048-1056.
(8) Rogel, E. Energy Fuels 1997, 11, 920-925.
(9) McLean, J. D.; Kilpatrick, P. K. Energy Fuels 1997, 11, 570-
585.
(10) Pauling, L. The Chemical Bond: A Brief Introduction to Modern
Structural Chemistry; Cornell University Press: Ithaca, NY, 1967.

10.1021/ef030015t CCC: $25.00 © 2003 American Chemical Society

Published on Web 09/17/2003
Consistency of Asphaltene Structures
Figure 1. Residue phase diagram, following Wiehe.5 Curve a represents the boundary for heptane solubles (from Wiehe5), and
curve b represents the revised boundary based on expanded data set. Legend for the data from Wiehe5 is as follows: ([) toluene
insolubles, (0) asphaltenes, and (O) heptane-soluble resins. Legend for the data from Rogel8 is as follows: (9) asphaltenes and
(3) heptane-soluble resins. Legend for other asphaltene data is as follows: (2) from Calemma and Rausa6 and (b) from Rahmani
et al.7
Figure 2. Dependence of the carbon residue yield on the 13C
NMR aromatic carbon content of asphaltenes: (O) microcarbon
residue (MCR) data for n-C7 asphaltenes; (0) data for pyrolysis
residues from the thermogravimetric analysis of n-C7 asphalt-
enes, by Calemma and Rausa;6 and (4) MCR data from
Dettman et al.12 for n-C5 asphaltenes separated from residues
before and after visbreaking.
peratures. Given the multiple chemical functionalities
that influence solubility, the success of Wiehe’s phase
diagram in clustering asphaltenes in two dimensions
suggests that the molecular weight and aromaticity are
dominant characteristics of asphaltenes, which is con-
sistent with the observations of Speight2 and Groenzin
and Mullins.1 As illustrated in Figure 1, an extended
version of Wiehe’s diagram is consistent with data on
polymer materials that are soluble in toluene and
insoluble in heptane.11
Aromaticity and molecular weight are also key prop-
erties that control the formation of carbon residue, as
in the CCR and MCR determinations. Aromatic carbon
content, as determined by 13C NMR, correlates highly
with carbon residue, as illustrated in Figure 2. The data
of Calemma and Rausa6 were for asphaltenes from
(11) Francuskiewicz, F. Polymer Fractionation; Springer-Verlag,
Berlin, 1994.
Energy & Fuels, Vol. 17, No. 6, 2003 1567
Middle Eastern and North African crude oils. They
determined the carbon residue via a thermogravimetric
technique. The MCR data in Figure 2 were determined
by ASTM method D4530 for asphaltenes from Middle
Eastern, Mexican, and Chinese crude oils, and from
Athabasca bitumen. Data from Dettman et al.12 were
from residues from Athabasca, Kuwait, Shengli, and
Rubiales before and after visbreaking. The range of
MCR and pyrolysis residue values is broad, which is
consistent with Wiehe’s observations.5 The correlation
between aromaticity and carbon residue was much
better than the correlation with hydrogen content. This
observation is consistent with the reaction of nonvolatile
aromatic groups to form coke.13
The carbon residue data of Figure 2, and the phase
diagram in Figure 1, illustrate two extremely important
characteristics of asphaltenes. The first is that asphalt-
enes, from different crude oils, can exhibit a wide range
of behavior under coking or cracking conditions. De-
tailed studies on a single asphaltene material, or a
narrow range of samples with similar characteristics,
cannot be used to infer a general model for asphaltene
chemistry. The second point is that when the MCR
content of an asphaltene is 50%-60%, then, by neces-
sity, 40%-50% of the mass of the initial material is
evolved as vapors under pyrolysis conditions. This type
of pyrolysis is accompanied by molecular rearrange-
ments, dehydrogenation, and addition reactions; how-
ever, any structural model for petroleum asphaltenes
must allow for significant evolution of mass by the
thermal cracking of single bonds.
Two very different views of the asphaltene structure
have been proposed in the literature. The first, following
(12) Dettman, H. D.; Salmon, S. L.; Ross, A. M.; Patmore, D. J.
Physical Properties, Thermal Reactivity and Molecular Composition:
Comparison of Heavy Crudes and Bitumen from Different Geographi-
cal Sources. Presented at the 52nd Canadian Chemical Engineering
Conference, Vancouver, British Columbia, October 20-23, 2002; Paper
34.
(13) Wiehe, I. A. Energy Fuels 1994, 8, 536-544.
1568 Energy & Fuels, Vol. 17, No. 6, 2003
Figure 3. Proposed asphaltene structural models: (a) condensed aromatic cluster model14 and (b) bridged aromatic model.3
Yen,14 assumes extensive condensation of the aromatic
rings into large sheets. These sheets are assumed to be
soluble, because of the saturated rings and side chains
around the molecular periphery (Figure 3a). This type
of structure, with lower molecular weight, has been
proposed by Groenzin and Mullins,1 on the basis of
spectroscopic studies. A very different structural orga-
nization, illustrated in Figure 3b, was proposed by
Speight2 and by Strausz and co-workers (e.g., Murgich
et al.3), on the basis of pyrolytic and selective oxidation
studies to determine the building blocks of asphaltenes.
A molecule of the type shown in Figure 3a cannot give
significant mass yields of volatile product in a pyrolysis
experiment. The side chains would crack off readily, and
the naphthenic rings then would undergo a combination
of dehydrogenation (to form aromatics) and cracking (to
give mainly light ends). In contrast, a structure of the
type in Figure 3b can easily give a wide range of product
sizes, from methane to toluene-insoluble carbon residue,
depending on the balance between cracking, product
release, and molecular rearrangements. Consistent with
(14) Yen, T. F. Prepr.sAm. Chem. Soc., Div. Pet. Chem. 1972, 17,
(4), F102.
Gray
this chemical structure, Wiehe13 depicted maltenes and
asphaltenes in residue schematically as aromatic cores
that are linked by reactive bridges and substituted with
side chains. Under coking conditions, the larger core
aromatic groups would accumulate in the liquid phase
and contribute to coke formation, whereas sufficiently
small fragments would form the observed liquid prod-
ucts. In a complex mixture, structures such as Figure
3a cannot be ruled out, particularly when the fraction
of aromatic carbon approaches 70%, as in the extremum
in Figure 2; however, such a chemical structure cannot
account for the average amount of volatile product
evolution under the pyrolysis conditions from a range
of asphaltenes.
Data on the nature of products from the cracking of
asphaltenes give further support to the diversity of
asphaltenes from different crude oils and indicate that
the chemical structures in asphaltenes must be consis-
tent with the evolution of a significant yield of volatile
products during pyrolysis. The data of Figure 4 show
simulated distillation curves from cracked products from
asphaltenes prepared by the supercritical fluid extrac-
tion of Athabasca bitumen.15 The asphaltenes were
Consistency of Asphaltene Structures
Figure 4. Simulated distillation of products from the thermal
cracking of Athabasca asphaltenes at 430 °C for 60 min.
cracked in a closed reactor at 430 °C for 60 min,16,17 and
the products were analyzed by simulated distillation
following ASTM standard test method D2887, extended
to high temperature. The material that was cracked
under a hydrogen atmosphere gave a yield of 6.7%
toluene-insoluble solids, 4.2% gases, and 86.7% liquids.
As illustrated in Figure 4, the largest fraction of the
liquid product was in the gas oil range (343-524 °C). A
minor amount of product was in the range of naphtha
(177 °C), which would be the only expected distillate
product from the structure in Figure 3a. Similar obser-
vations of a very broad range of products from asphalt-
ene pyrolysis were reported by Savage and Klein.18 The
more recent structures that have been proposed by
Groenzin and Mullins,1 which are structurally similar
to those depicted in Figure 3a but with molecular
weights of 500-700, would also give predominantly
naphtha saturates and vacuum residue as liquid prod-
ucts.
Chemical characterization of the products from the
pyrolysis of asphaltenes depends strongly on the ana-
lytical methods selected and, possibly, on the chemistry
of the asphaltene selected. When the cracked products
are analyzed by gas chromatography (GC), the dominant
components in the cracked products are n-alkanes and
n-alkenes for a wide range of asphaltenes.19 For ex-
(15) Chung, K. H.; Xu, C. M.; Hu, X. Y.; Wang, R. N. Oil Gas J.
1997, 95, (1), 66-69.
(16) Chung, K. H.; Xu, S.; Gray, M. R.; Zhao, Y.; Kotlyar, L.; Sparks,
B. Rev. Process Chem. Eng. 1998, 1, 41-79.
(17) Zhao, Y.; Gray, M. R.; Chung, K. Energy Fuels 2001, 15, 751-
755.
(18) Savage, P. E.; Klein, M. T.; Kukes, S. G. Energy Fuels 1988, 2,
619-628.
(19) Strausz, O. P.; Mojelsky, T. W.; Faraji, F.; Lown, E. M.; Reng,
P. Energy Fuels 1999, 13, 207-227.
Energy & Fuels, Vol. 17, No. 6, 2003 1569
ample, Artok et al.20 used Curie-point pyrolysis to
analyze products from the asphaltenes from an Arabian
crude. Paraffins, olefins, and aromatics were identified
by gas chromatography-mass spectroscopy (GC-MS),
but only up to C21. In this range, n-alkane and n-alkene
products dominated. Calemma and Rausa6 used ther-
mogravimetric analysis and GC-MS to analyze a wide
range of asphaltenes. Of the identified compounds in
the pyrolysis products, the fraction of aromatic com-
pounds ranged from 10.3% to 23.6%. Given the difficulty
in resolving distinct peaks of aromatic isomers via GC,
these values must be taken as minimum estimates,
because significant amounts of pyrolysis product could
not be identified. To allow a more complete character-
ization, Payzant et al.21 used a rapid pyrolysis method
to generate gram quantities of product from Athabasca
asphaltenes. A total of 9.5% of the initial asphaltene
was evolved as saturates, 14% as aromatics, and 10%
as polar aromatics. These results cannot be reconciled
with a model on the basis of only large, highly condensed
aromatic cores with substituent side chains. If aromatic
groups can evolve from asphaltenes via the pyrolysis of
bridges, then these data, along with the MCR data and
boiling curves of Figure 4, are consistent with the
bridged asphaltene-type structure of Figure 3b.
None of these arguments address the molecular
weight of the asphaltene material: they address only
the chemical structure, in terms of bridged aromatics
versus large condensed aromatic cores. To agree with
experimental observations from cracking and pyrolysis,
the asphaltenes, and other components of vacuum
residue, must have a sufficient molecular weight such
that vaporization is unfavorable. This criterion requires
only that most of the components have a normal boiling
point of over ∼650 °C,22 as illustrated by the boiling
curve for Athabasca asphaltenes from supercritical
extraction in Figure 4.
Acknowledgment. This work was supported by the
NSERC/Syncrude Industrial Research Chair in Ad-
vanced Upgrading of Bitumen. The assistance of the
Syncrude Research Analytical Laboratory, in the analy-
sis of samples, is gratefully acknowledged.
EF030015T
(20) Artok, L.; Hirose, Y.; Hosokawa, M.; Murata, S.; Nomura, M.
Energy Fuels 1999, 13, 287-296.
(21) Payzant, J. D.; Lown, E. M.; Strausz, O. P. Energy Fuels 1991,
5, 445-453.
(22) Olmstead, W.; Freund, H. Thermal Conversion Kinetics of
Petroleum Residua. Presented at the American Institute of Chemical
Engineers (AIChE) Spring Meeting, New Orleans, LA, 1998; Paper
34b.