14 Oileld Review

Revealing Reservoir Secrets Through

Asphaltene Science
Downhole uid analysis of the heaviest components of petroleum can help unlock
information about reservoir structure. Understanding how asphaltenes associate in
oil columns permits scientists and engineers to use asphaltene concentration
gradients to determine the presence of sealing barriers. Production results have
conrmed the validity of this approach, which is being extended to address the
structure and dynamics of uids in complex reservoirs.
A. Ballard Andrews
Oliver C. Mullins
Andrew E. Pomerantz
Cambridge, Massachusetts, USA
Chengli Dong
Hani Elshahawi
Shell Exploration and Production
Houston, Texas, USA
David Petro
Marathon Oil Corporation
Houston, Texas
Douglas J. Seifert
Saudi Aramco
Dhahran, Saudi Arabia
Murat Zeybek
Dhahran, Saudi Arabia
Julian Y. Zuo
Sugar Land, Texas
Oileld Review Winter 2012/2013: 24, no. 4.
Copyright 2013 Schlumberger.
For help in preparation of this article, thanks to John Mainstone,
The University of Queensland, Brisbane, Australia.
InSitu Fluid Analyzer, LFA and MDT are marks of
Schlumberger.
INTERSECT is a joint mark of Schlumberger, Chevron
and Total.
Long before scientists grappled with the heaviest
component of petroleumasphalthumans
were putting it to use. In the ancient world,
Babylonians used asphalt as mortar, and
Egyptians employed it for mummication.
1

Asphalts ability to preserve and bind has been
carried through the intervening centuries to a
host of current applications that include paving,
roong, waterproong and insulation.
In the oil eld, the utility of asphalt is less
clear. Asphaltenes, the primary component of
asphalt, tar or bitumen, can create ow assur-
ance problems in the formation, production tub-
ing and pipeline.
2
Additionally, crudes with high
asphaltene levels are less valuable on world mar-
kets; their hydrogen deciency limits their yield
of liquid hydrocarbons and their sulfur and metal
content creates problems for rening.
3
1. Yen TF and Chilingarian GV (eds): Asphaltenes and
Asphalts, 2. Amsterdam: Elsevier Science BV,
Developments in Petroleum Science, 40B, 2000.
2. Kabir CS and Jamaluddin AKM: Asphaltene
Characterization and Mitigation in South Kuwaits Marrat
Reservoir, paper SPE 80285, presented at the SPE
Middle East Oil Show and Conference, Bahrain,
February 2023, 1999.
3. Allan D and Davis PE: Rening ReviewA Look Behind
the Fence, Oileld Review 19, no. 2 (Summer 2007): 1421.
4. Elshahawi H, Mullins OC, Hows M, Colacelli S,
Flannery M, Zuo J and Dong C: Reservoir Fluid Analysis
as a Proxy for Connectivity in Deepwater Reservoirs,
Petrophysics 51, no. 2 (April 2010): 7588.
5. This classication is typically labeled a SARA
analysissaturates, aromatics, resins and asphaltenes.
For more: Akbarzadeh K, Hammami A, Kharrat A,
Zhang D, Allenson S, Creek J, Kabir S, Jamaluddin A,
Marshall AG, Rodgers RP, Mullins OC and Solbakken T:
AsphaltenesProblematic but Rich in Potential,
Oileld Review 19, no. 2 (Summer 2007): 2243.
6. Black oil is used in reservoir modeling to describe oil in
place. The conventional black oil model uses three
components: water, oil and gas. For more on black oil
modeling: Huan G: The Black Oil Model for a Heavy Oil
Reservoir, paper SPE 14853, prepared for presentation at
the SPE International Meeting on Petroleum Engineering,
Beijing, March 1720, 1986.
7. Mullins OC: The Physics of Reservoir Fluids. Sugar Land,
Texas, USA: Schlumberger, 2008.
Zuo JY, Freed D, Mullins OC, Zhang D and Gisolf A:
Interpretation of DFA Color Gradients in Oil Columns
Using the Flory-Huggins Solubility Model, paper
SPE 130305, presented at the CPS/SPE International
Oil and Gas Conference and Exhibition, Beijing,
June 810, 2010.
>
Reservoir gradients. Measurements on a condensate oil from a reservoir
in Norway show that the formation pressure and temperature at the gas/oil
contact zone lie on an equation of state (EOS)generated bubblepoint line
dividing the liquid and two-phase regions. Composition data on reservoir
uids from this eld show large gradients. Composition gradients in the
reservoir depend on uid conditions, and as the reservoir temperature and
pressure approach the bubblepoint line and critical point, large
composition gradients develop.
P
r
e
s
s
u
r
e
,

b
a
r
Temperature, K
Increasing
composition
gradients
Liquid Vapor Two-phase region
700 600 500 400 300 200 100
100
200
300
400
500
0
800 900
Formation conditions
Critical point
Bubblepoint
Dewpoint
Winter 2012/2013 15 15
The high cost of offshore operations and the
trend toward deeper wells worldwide have
renewed the imperative for understanding reser-
voir uids at a molecular level. Operators can no
longer afford to view reservoirs as homogeneous
tanks of oil and gas. In addition to knowing uid
composition, they must also be able to assess res-
ervoir connectivity, particularly when costs dic-
tate a limited number of wells. Imaging and
pressure surveys are often insufcient to com-
pletely assess oil drainage patterns, so operators
are turning to downhole uid analysis (DFA) and
asphaltene science to better understand reser-
voir structures.
4
In the recent past, operators characterized
oil in reservoirs with a few parameters such as
specic gravity, gas/oil ratio (GOR) and a simple
chemical classication of the bulk oil.
5
However,
DFA measurements on oil columns from around
the world reveal that reservoir uids present a
much more complex picture, both vertically in
the oil column and laterally across the eld. Such
results, coupled with decades of analytical
research, are yielding a more complete picture of
asphaltene physical forms in the reservoir. These
research advances explain how and under what
conditions asphaltenes associate with each other
and allow all components of the reservoir uid
mixgas, liquids and solidsto be described by
equations based on thermodynamic principles.
The end result of this work enables use of pre-
dicted and observed asphaltene concentration
gradients to conrm or disprove uid drainage
connectivity in an oil column.
This article focuses on new asphaltene sci-
ence and covers its origins, development and
uses. Cases from deepwater Gulf of Mexico and
Middle Eastern elds illustrate how these devel-
opments are helping oileld scientists and engi-
neers learn more about connectivity in reservoirs
and the distribution of gases, liquids and solids in
the uids contained therein.
Reservoir FluidsA Complex Picture
A beaker of petroleum on a laboratory bench or
an open hatch on a stock tank presents a decep-
tively simple view of underground uidsthat an
entire reservoir consists of only black oil and gas.
6

Fluid property gradients, where present because
of reservoir conditions, may appear to affect only
the GOR. However, this view is inaccurate
because at actual reservoir conditions, composi-
tion gradients can exist not only for the GOR, but
also for asphaltenes and the individual compo-
nents of the oil (previous page).
7
N
16 Oileld Review
Asphaltenes in petroleum have been a focus
of study by oileld engineers and scientists for
decades. Much about asphaltenes has seemed
complex and inconclusive. Interest in these
compounds has taken on several dimensions
over time. In the early years of the industry,
downstream research was centered on optimiz-
ing uses for the asphalt by-products from ren-
ing operations. In the last half of the twentieth
century, that focus turned toward efcient con-
version of heavy ends and their asphaltene com-
ponent as reners sought to maximize the
production of transportation fuels. In upstream
exploration and production, the focus on
asphaltenes has almost always been on mitigat-
ing and avoiding their negative impacts. These
impacts include formation plugging because of
precipitation and the effects of high viscosity
during production and transportation (below).
8

However, new science developed over the last
decade has shown that asphaltene gradients in
the reservoir can provide valuable insights
about reservoir structure.
Asphaltenes found in reservoir uids are a
complex molecular mixture of particles colloi-
dally suspended in oil that have no single chemi-
cal identity. They are usually dened as a solubility
classthat is, those molecules that are insoluble
in n-heptane but soluble in toluene. Asphaltene
molecules are typically condensed aromatic rings
that can contain heteroatoms such as nitrogen
and sulfur as well as metals such as nickel and
vanadium. Almost every chemical property of
asphaltenes has been the subject of signicant
debate, except for their elemental composition.
An early controversy centered on the nature of
the covalently bound chemical groups versus
those that are associated in noncovalent aggre-
gates.
9
The wide range of molecular weights
obtained at that time1,700 to 500,000 g/mol
was attributed to varying aggregate sizes. Over
the last decade, research on asphaltenes has
encompassed multiple branches of analytical
chemical science to produce a much clearer pic-
ture of asphaltene properties and how individual
asphaltene molecules associate to form larger
particles (above).
10
Downhole Fluid Analysis
Downhole uid analysis helps scientists and engi-
neers examine reservoir uids in their native
environment. The DFA concept has evolved from
a technique for uid identication via openhole
sample acquisition to a means of analyzing reser-
voir uids and their spatial variations at forma-
tion conditions in real time. The concept is
simple: Following drilling, a cylindrical sampling
and analysis module is lowered into a well on
wireline, and uids are collected from the forma-
tion. This tool, the MDT modular formation
8. Mullins, reference 7.
Edgeworth R, Dalton BJ and Parnell T: The Pitch Drop
Experiment, European Journal of Physics 5, no. 4
(October 1984): 198200.
9. Dickie JP and Yen TF: Macrostructures of the Asphaltic
Fractions by Various Instrumental Methods, Analytical
Chemistry 39, no. 14 (December 1967): 18471852.
10. Mullins OC: The Modied Yen Model,
Energy & Fuels 24 (January 2010): 21792207.
11. Creek J, Cribbs M, Dong C, Mullins OC, Elshahawi H,
Hegeman P, OKeefe M, Peters K and Zuo JY:
Downhole Fluids Laboratory, Oileld Review 21, no. 4
(Winter 2009/2010): 3854.
>
Asphaltene viscosity. In 1927, researchers at
The University of Queensland, Australia, heated
a sample of pitch, or asphalt, and placed it in a
funnel that was subsequently sealed (inset ). The
asphalt was allowed to settle for three years at
room temperature before researchers cut the
funnel stem. Since that date, the asphalt has
dripped from the funnel, averaging one drop
every nine to ten years. In 2002, the ninth drop
was starting to form. While the viscosity of
heavy oils is not nearly as high as that of
asphalt, viscosity rises sharply with increasing
asphaltene content. Data on asphaltenes and
deasphalted oil from several crudes show a
rapid increase in viscosity with rising hexane
asphaltene content that spans six orders of
magnitude in viscosity. These data are
represented by a Pal-Rhodes viscosity model.
(Photograph courtesy of JS Mainstone, The
University of Queensland).
Hexane asphaltenes, wt %
V
i
s
c
o
s
i
t
y
,

P
a
.
s

a
t

6
0

C
0 10
10
1
10
2
10
3
10
4
10
5
10
6
10
7
10
8
10
9
20 30 40 50
Asphaltenes,
deasphalted oils
Pal-Rhodes
viscosity model
>
Asphaltene properties. During the past decade, advances in analytical
science have allowed a more consistent picture of asphaltene structure to
emerge. Estimates for the mean asphaltene molecular weight have been
reduced by several orders of magnitude and are now about 750 g/mol, and
the range is signicantly tighter. Similarly, scientists now know the median
number of fused rings per asphaltene polyaromatic hydrocarbon (PAH) is
about seven, with one PAH per molecule dominating. In addition, the
number of PAH stacks in an asphaltene nanoaggregate, unknown a decade
ago, is one. All of these developments have allowed researchers to
establish consistent physical models regarding asphaltene molecules and
to show how they associate with one another in reservoir uids.
Property
Mean asphaltene molecular weight
Number of PAHs per asphaltene
Number of fused rings per
asphaltene PAH
Number of PAH stacks
in a nanoaggregate
1 to 20
2 to 20
unknown 1
7 (average)
1 dominates
750 g/mol 10
3
to 10
6
g/mol
Reported Values, 1998 Reported Values, 2009
12. Optical density, measured by MDT spectroscopy, is
calculated from the degree of absorption in the visible
and near-infrared portion of the frequency bandfrom
wavelengths of about 400 to 2,000 nm. Components of
reservoir uids, such as asphaltenes, have
characteristic absorptions in this range that reect their
molecular structures. Optical density gives a
dimensionless numerical value to the color
characteristics of these uids. For more on downhole
optical density applications: Creek et al, reference 11.
13. Mullins OC, Andrews AB, Pomerantz AE, Dong C, Zuo JY,
Pfeiffer T, Latifzai AS, Elshahawi H, Barr L and Larter S:
Impact of Asphaltene Nanoscience on Understanding
Oileld Reservoirs, paper SPE 146649, presented at the
SPE Annual Technical Conference and Exhibition,
Denver, October 30November 2, 2011.
Winter 2012/2013 17
dynamics tester, contains a probe for sampling
reservoir uids and an array of sensors for analyz-
ing the sampled uids on a real-time basis
(above). An MDT tool congured for DFA can pro-
vide a long list of reservoir data ranging from gen-
eral properties such as GOR and pressure and
temperature at depth to specic attributes such
as density, composition and miscible sample con-
tamination by nonaqueous drilling uids.
11
In
addition to determining GOR and other proper-
ties, the MDT tool uses spectroscopy to measure
optical densityessentially oil colorwhich is
directly proportional to asphaltene concentra-
tion.
12
Fluid property variations interpreted from
DFA measurements made at several depth sta-
tions in a well can sometimes indicate nearby
sealing barriers (right).
13
Identifying compartments in a reservoir is not
as challenging as assessing oil drainage connec-
tivity within those compartments, especially
before production. Static pressure surveys may
fail to nd hard-to-image sealing barriers before
production starts because pressure equilibrium
and composition equilibrium are achieved over
different time scales. Composition equilibrium is
achieved slowly, and the difference between the
>
Modular formation dynamics tester. The MDT tool (above) contains a complex array of
instrumentation for downhole sampling and analysis. In a typical conguration (right), the MDT tool
components include a section for storing samples in addition to an InSitu Fluid Analyzer system and
LFA live uid analyzer system for real-time downhole uid analysis. Reservoir uids enter the tool at
the formation probe and are pumped in two directionsupward toward the InSitu Fluid Analyzer tool
and downward toward the LFA module. The InSitu Fluid Analyzer tool contains two spectrometers and
a uorescence detector for analysis of hydrocarbons, CO
2
, pH and uid color; it also contains
instruments for measuring density, resistivity, pressure and temperature. Reservoir uid from the
sampling probe that is pumped downward passes through the LFA module. This device employs an
absorption spectrometer to quantify and monitor the amount of reservoir and drilling uids that are
present. A gas refractometer (not shown) in the tool differentiates between gas and liquids.
Sample modules
InSitu Fluid Analyzer
system
LFA live fluid
analyzer system
Focused probe
Pump 2
Pump 1
>
Sealing barriers. Using DFA to reveal the
presence of uid density inversions can
sometimes help identify sealing barriers in a
reservoir. GOR data for two depth zones in an oil
column illustrate this concept. Using GOR as a
proxy for density in this column, scientists found a
low-GOR, high-density uid at Point A (left), above
a high-GOR, low-density uid at Point B (right).
This nding indicates the possible presence of a
sealing barrier between the two zones.
V
e
r
t
i
c
a
l

d
e
p
t
h
,

m
GOR, ft
3
/bbl
Possible sealing barrier
A
B
X,800
X,700
1,000 2,000 3,000 4,000 5,000
X,600
X,500
X,400
X,300
X,200
18 Oileld Review
time to reach pressure equilibrium and that to
reach composition equilibrium for the heaviest
fraction of crude can be several orders of magni-
tude (above).
14
Massive uid migration in the res-
ervoir is required to achieve compositional
equilibration, and for this to occur, there must be
good reservoir connectivity. In contrast, pressure
equilibration can be achieved with very small
mass transfer, which can occur through leaky
seals. Consequently, pressure equilibration is a
necessary but insufcient condition to establish
connectivity in the reservoir.
Nearly equilibrated asphaltene concentration
gradients between two zones are indicative of
connectivity. However, before that concept can
be implemented on a practical basis, it is neces-
sary to have a model for asphaltenes that
accounts for their thermodynamic characteris-
tics and how they associate with each other deep
in the reservoir.
Modeling Asphaltenes
Since 2000, advances in analytical instrumenta-
tion and science have allowed a much clearer
picture of asphaltene structure to emerge. Such
advances have narrowed the knowledge gap
about their properties and have led to a more
rened description of asphaltene science as
embodied in the modied Yen model.
15
This
model was later renamed the Yen-Mullins
model.
16
It envisions asphaltenes in crude oil as
existing in three distinct and separate formsas
asphaltene molecules, as nanoaggregates of indi-
vidual asphaltene molecules and as clusters of
nanoaggregates (below). The number of analyti-
cal methods employed over the last decade to
resolve the molecular weight, size and aggrega-
tion parameters in this model is extensive and
includes time-resolved uorescence depolariza-
tion and laser-based mass spectrometry for
molecular and aggregate size and weight deter-
mination. For most model parameters, such as
asphaltene molecular weight, scientists must
apply several techniques to reduce the
uncertainty.
The asphaltene molecule is at the rst level of
the Yen-Mullins model. The typical asphaltene
molecule consists of several fused aromatic rings
with peripheral alkane substituents, often with
scattered sulfur and nitrogen heteroatoms.
This molecule has a mean molecular weight of
750 g/mol with most of the population ranging
from 500 to 1,000 g/mol and a length of about
1.5 nm. In this model hierarchy, the asphaltene
>
The Yen-Mullins model of asphaltene nanoscience. At low concentrationstypical in condensates and volatile oils
asphaltenes are predicted to exist as a solution of molecules that measure about 1.5 nm (left). At higher concentrationsfound
in black oilsasphaltenes are dispersed as 2-nm nanoaggregates (center). At still higher concentrations, such as those found in
mobile heavy oils, asphaltenes are dispersed as clusters of 5 nm (right).
Molecule Nanoaggregate Cluster
~1.5 nm ~2 nm ~5 nm
N
>
Reservoir equilibration. Reservoir modeling gives insight to the time
required to reach equilibration. Modeling of a tilted sheet reservoir
with a low-permeability zone in the center shows that uid composition
equilibrationmeasured by density, methane or heavy fraction
is seven to eight orders of magnitude slower than the corresponding
pressure equilibration.
T
i
m
e

t
o

r
e
a
c
h

e
q
u
i
l
i
b
r
a
t
i
o
n
,

y
e
a
r
s
Black oil Volatile oil Condensate Gas
0
1
10
1
10
2
10
3
10
4
10
5
10
6
10
7
10
8
10
9
Pressure
Fluid density
Methane
Heavy fraction
Winter 2012/2013 19
14. Pfeiffer T, Reza Z, Schechter DS, McCain WD and
Mullins OC: Determination of Fluid Composition
Equilibrium Under Consideration of AsphaltenesA
Substantially Superior Way to Assess Reservoir
Connectivity than Formation Pressure Surveys, paper
SPE 145609, presented at the SPE Annual Technical
Conference and Exhibition, Denver, Colorado, USA,
October 30November 2, 2011.
15. Mullins, reference 10.
Mullins et al, reference 13.
16. Sabbah H, Morrow AL, Pomerantz AE and Zare RN:
Evidence for Island Structures as the Dominant
nanoaggregate is the next structure in size. These
particles represent an aggregation of about six
asphaltene molecules in a single disordered
stack about 2 nm in size. The asphaltenes in
nanoaggregates are tightly bound, and exterior
alkanes on the nanoaggregate particle project
outwardly. The largest particle in the Yen-Mullins
model is the cluster, which represents a group of
about eight nanoaggregates. Clusters, which are
loosely bound, are about 5 nm in size.
Although all of the forms envisioned by the
Yen-Mullins model may occur in any oil column,
the specic form depends largely on the asphal-
tene concentration. In wells that produce volatile
oils and condensates with high GOR, the asphal-
tene concentration will be less than 0.5 wt % and
the asphaltene particles will be 1 to 1.5 nm in
size. At higher asphaltene concentrations, such
as black oil columns with moderate GOR values,
the asphaltene concentration will usually be less
than 5 wt % and the asphaltene particles will be
principally 2-nm nanoaggregates. In even higher
asphaltene concentrations, as seen in mobile
heavy oils that have low GOR, asphaltene levels
will range from 5 to 35 wt %, with 5-nm clusters as
the primary asphaltene particle.
Tar mats may occur in formations with signi-
cant levels of mobile heavy oil and are areas of
nearly immobile asphaltenes usually found at the
base of an oil column near the oil/water contact.
There are two predominant forms of tar mats.
17

One type occurs at the base of a mobile heavy oil
column as a result of seemingly continuous
extension of a large asphaltene concentration
and viscosity gradient. The other type of tar mat
occurs at the base of a lighter oil column and is
discontinuous in asphaltene concentration.
The rst type of tar mat results from a subtle
destabilization of asphaltenes at the top of the oil
column followed by transport of asphaltenes to
the base of the oil column to form a mat. The sec-
ond type of tar mat may occur when there is a sig-
nicant gas charge into the top of a reservoir
containing black oil. As the gas diffuses down the
column, the GOR increases and causes asphal-
tene molecules and nanoaggregates to form clus-
ters. These clusters descend ahead of the diffusive
gas front, which moves lower in the column with
time. When the gas front reaches the bottom of
the column, the asphaltenes are expelled from
the oil to form the tar mat (above right).
18
>
Tar mat formation. One mechanism for tar mat formation (top) envisions a
stable black oil column (A) in which biogenic methane moves beneath an
upper seal (B). As the methane slowly diffuses down the oil column, large
GOR and asphaltene gradients are formed (C). These gradients can become
large enough that a dense, asphaltene-rich tar mat may form at the bottom
of the column (D). A thin section from a tar mat formed at the base of a high-
GOR column shows tar on the grains of a cemented sandstone (bottom).
Aquifer Aquifer
Aquifer Aquifer
Black oil
column
Methane
charging updip
A B
C D
Methane
diffusing downdip
Tar mat
Seal
Tar
Architecture of Asphaltenes, Energy & Fuels 25, no. 4
(2011): 15971604.
17. Mullins OC, Zuo JY, Dong C, Andrews AB, Elshahawi H,
Pfeiffer T, Cribbs ME and Pomerantz AE: Downhole
Fluid Analysis and Asphaltene Nanoscience for
Reservoir Evaluation Measurement, Transactions of the
SPWLA 53rd Annual Logging Symposium, Cartagena,
Colombia, June 1620, 2012, paper CCC.
18. Zuo JY, Elshahawi H, Mullins OC, Dong C, Zhang D,
Jia N and Zhao H: Asphaltene Gradients and Tar Mat
Formation in Reservoirs Under Active Gas Charging,
Fluid Phase Equilibria 315 (February 15, 2012): 9198.
20 Oileld Review
Correctly modeling asphaltenes requires a
two-pronged approach. The Yen-Mullins model
provides the solution to the rst challengea
useful framework for the asphaltene particles
that form in an oil column along with estimates of
particle size and molar volume. The second part
of the problem is to mathematically describe the
asphaltene concentration gradients for the vari-
ous asphaltene physical states as predicted by
the Yen-Mullins model.
In thermodynamic systems, a state variable is
a parameter such as temperature, pressure or
volume, which depends on the state of the system
but not the path used to get to that state. The
mathematical equation that relates state vari-
ables is called an equation of state (EOS). In
1834, Benoit Paul mile Clapeyron, a French
engineer and physicist, developed the ideal gas
law, an EOS that relates pressure, volume and
temperature. The ideal gas law is a rst-order
equation that ignores molecular volumes and
forces and is accurate only for weakly interacting
gases at moderate conditions. In 1873, van der
Waals developed a cubic EOS that approximates
the behavior of gases and liquids by taking into
account molecular forces and the size of mole-
cules. Since that time, many variants of the clas-
sic cubic EOS have been developed, and these
equations have been used for decades to model
uid behavior in oil columns. However, using
these equations for black oil modeling in reser-
voirs containing signicant levels of asphaltenes
is not satisfactory. Because asphaltenes lack a
gas phase or a critical point, they must be treated
as a pseudocomponent and handled empirically.
Although this approach is adequate to model
hydrocarbon gas-liquid equilibria and determine
parameters such as GOR, it is inadequate for
modeling molecular and colloidally suspended
particles such as asphaltenes, asphaltene nano-
aggregates and clusters of nanoaggregates.
The need to model solution behavior of mix-
tures containing solvents and large molecules
such as asphaltenes has existed for decades.
Much research in the 1940s focused on the ther-
modynamics and solution behavior of polymer
compounds and resulted in the Flory-Huggins
theory.
19
More recently, the Flory-Huggins approach
has been used to examine asphaltene instabil-
ity.
20
Recognizing the need for a rst principles
approach to describe asphaltene concentration
gradients in oil columns, scientists have devel-
oped the Flory-Huggins-Zuo EOS for this pur-
pose.
21
This equation incorporates a gravity term
for asphaltenes using their known size. This grav-
ity term is essential for modeling asphaltene
gradients. The equation was developed starting
with the free energy of a mixture of asphaltenes
and solvent as a function of the free energies
associated with gravity, solubility and entropy of
mixing. At equilibrium, the derivative of the free
energy sum is zero, and the solution of the result-
ing partial differential equations yields the Flory-
Huggins-Zuo EOS. In its original form, this
equation expresses the asphaltene concentration
gradient as a volume fraction of asphaltenes at
various depths in the oil column. Since oil color is
directly proportional to the asphaltene concen-
tration, the optical density ratio is usually substi-
tuted for the volume ratio for a more practical
measurement. The resulting equation gives the
asphaltene concentration in terms of optical den-
sity and is an exponential function of several
parameters (above).
22
The rst term in the Flory-Huggins-Zuo EOS
accounts for the effect of gravity and is the
most signicant term for asphaltenes in an oil
column for low-GOR oils (next page, top right).
Gravitational effects cause asphaltenes to accu-
mulate at the base of a column, although thermal
energy counteracts gravity to some extent. This
rst term expresses gravitational effects as the
buoyancy of an object in a liquidthe gravity
effectdivided by a function of the tempera-
>
Asphaltene equation of state (EOS). The Flory-Huggins-Zuo EOS (top) predicts asphaltene gradients in an oil column. Optical
density at two depths is predicted as an exponential function of three termsgravity, entropy and solubility. The gravity term
depends primarily on asphaltene particle size and depth. The entropy term is a measure of molecular randomness and depends
on molar volumes. The nal term in this equationsolubilitydepends on GOR, density and composition.
OD h
i
( ( optical density at depth h
i

a
h
i
( ( asphaltene concentration at depth h
i

v
a
asphaltene molar volume
v oil phase molar volume
g gravitational constant
density difference between asphaltenes and oil phase
T temperature
R ideal gas constant

a
asphaltene solubility parameter
oil phase solubility parameter
Fluid color Gravity term Entropy term Solubility term =

+
RT RT
exp = =
h
2
h
1
h
1
h
2
v
v
a
v
a
v
a
v
+

a OD



g v
a

2 2
a

a

( (( (

a
OD
( (( (( (
( (( (
h
2
h
2 h
2
h
1
h
1
h
1
Winter 2012/2013 21
turethe thermal effect. For large physical
forms of asphaltenes, such as clusters found in
heavy oils, the gravity term is signicant and
gives rise to high concentrations of asphaltenes
near the base of the oil column.
The remaining two terms in the new asphal-
tene EOS are similar to the original Flory-
Huggins terms for entropy and solubility. The
entropy is stated in terms of ratios of molar vol-
umes of asphaltenes and solvent at two depths.
The entropy effect tends to randomize the
asphaltene distribution and counteract gradi-
ents, but is usually not large for asphaltenes in
crude oils. The other factor in the Flory-Huggins-
Zuo EOS that essentially corresponds to the
original Flory-Huggins work is the solubility
term. For asphaltene gradients, this term is
expressed in solubility parameters that are cal-
culated from GOR or mass densities. This term
accounts for changes of asphaltene solubility in
the liquid phase and is important for high-GOR
oil that produces a low-density liquid, rich in
parafnic alkanes that decrease asphaltene
solubility. For low-GOR oils, however, the solu-
bility term is usually not signicant.
The end result of this new equation of state
for asphaltenes is the prediction of asphaltene
concentrations, directly proportional to uid
color, at any depth in the oil column. Almost all of
the parameters may be measured or estimated
from downhole uid analysis results of the bulk
oil at various depth stations. Those parameters
not directly measuredsuch as the solubility
parametersmay be obtained from correlations
of known properties.
The only adjustable parameter in the Flory-
Huggins-Zuo EOS is the asphaltene molar vol-
ume, which is related to particle size. The
particle size cannot be determined directly
from the downhole data, but there are other
ways to nd it. The rst method is to tune the
unknown size of the asphaltene particles to
match the downhole uid color data from
measure ments at different depths. This size is
then checked against the Yen-Mullins model
particle types to ensure it is within the boundar-
ies described by the model. The second method
is to assume that heavy ends in the oil column
are either asphaltene molecules, nanoaggre-
gates or clusters. In this case, the assumed size
is used to predict the downhole asphaltene gra-
dients in the oil column, which can be checked
against the actual data. If there is consistency,
then the data can be used to assess connectivity
and other reservoir properties. Analysis of the
data may not always suggest a single asphaltene
particle type because multiple particle types
may be involved (below).
19. Flory PJ: Thermodynamics of High Polymer Solutions,
Journal of Chemical Physics 10, no. 1 (January 1942):
5161.
Huggins ML: Thermodynamic Properties of Solutions of
Long-Chain Compounds, Annals of the New York
Academy of Sciences 43, no. 1 (March 1942): 132.
20. Buckley JS, Wang J and Creek JL: Solubility of the
Least Soluble Asphaltenes, in Mullins OC, Sheu EY,
Hammami A and Marshall AG (eds): Asphaltenes,
Heavy Oils, and Petroleomics, New York: Springer
Science+Business Media (2007): 401438.
21. Zuo JY, Mullins OC, Freed D, Elshahawi H, Dong C and
Seifert DJ: Advances in the Flory-Huggins-Zuo Equation
of State for Asphaltene Gradients and Formation
Evaluation, Energy & Fuels (in press).
22. Freed DE, Mullins OC and Zuo JY: Theoretical Treatment
of Asphaltene Gradients in the Presence of GOR
Gradients, Energy & Fuels 24, no. 7 (July 15, 2010):
39423949.
Zuo et al, reference 21.
>
Gravity effects. The effect of gravity depends on which asphaltene
physical form predominates in the well. For a 100-m [328-ft] oil column
containing mostly asphaltene clusters (black), gravity effects are large, as
evidenced by the dramatic increase of asphaltene content with depth. The
intermediate size nanoaggregates (blue) show a much more gradual
change, while the asphaltene molecules (red) show only a small change
from top to bottom of the column.
V
e
r
t
i
c
a
l

d
e
p
t
h
,

m
0 0.2 0.4 0.6 0.8 1.0
100
80
60
40
20
0
5.0-nm clusters
2.0-nm nanoaggregates
1.5-nm molecules
Asphaltene concentration at depth
Asphaltene concentration at 100 m
>
Multiple particle types. A black oil column that was subjected to a late gas
and condensate charge shows evidence that more than one asphaltene
particle type is present in the column. Analysis of DFA data using the
Flory-Huggins-Zuo EOS indicates that nanoaggregates alone would not
account for the increase in asphaltene concentrationas measured by
optical densitywith depth (left). In this example, the late gas charge
destabilized the asphaltenes, causing clusters to form; these clusters settled
toward the bottom of the oil column because of gravity (right). The presence
of large viscosity and asphaltene gradients characterized this oil column,
and production of this well proceeded with no signicant problems.
V
e
r
t
i
c
a
l

d
e
p
t
h
,

m
O
i
l

c
o
l
u
m
n
Optical density
Nanoaggregates
and clusters
Nanoaggregates
X,Y00
X,Z00
X,X50
X,Y50
X,Z50
0 0.5 1.5 2.5 1.0 2.0 3.0
EOS models
DFA data
Nanoaggregate Cluster
22 Oileld Review
Downhole uid analysis, the new Yen-Mullins
model and the Flory-Huggins-Zuo EOS can be used
together to model asphaltene gradients in actual
oil columns. The rst step is the use of DFA to give
experimental data on asphaltene concentration
via uid color, GOR and other physical parameters
at several depth stations in a well. The Yen-Mullins
model then provides a physical picture of the
asphaltene entities that may be present and allows
the operator to make reasonable assumptions on
particle size. That size is then used in the Flory-
Huggins-Zuo EOS to predict the asphaltene con-
centration gradient in the well. If this gradient
matches the experimental data, it can be used to
further assess reservoir connectivity. This analysis
is not a mere curve-tting exercise. The matching
of sizes computed by the new EOS and the Yen-
Mullins model gives the operator condence that
the system is in equilibrium.
Asphaltene Science and Complex Reservoirs
An example from a complex eld in the Gulf of
Mexico illustrates how asphaltene science is
used in answering practical questions. This eld,
operated by Marathon, included an area produc-
ing intermediate-GOR black oil that consisted of
six sand layers spanning 1,000 ft [300 m] of depth
and intersected by multiple wells.
23
The chal-
lenge for the operator was to develop an accurate
description of reservoir uid properties and
understand connectivity among the various sand
layers. The reservoir uids were analyzed by mul-
tiple methods. DFA was employed using the MDT
tool both to gather real-time information and
obtain samples for further PVT analysis in the
laboratory. Using advanced gas chromatographic
analysis, the operator also performed geochemi-
cal ngerprinting on collected samples. Although
the data covered multiple wells in the area of
interest, not all analyses were performed at all
depth stations; the most complete dataset came
from two wells in one of the sands. These data
and their analyses show how connectivity ques-
tions can be viewed through the lens of the new
asphaltene science.
Prior to use of asphaltene gradients to give
clues to connectivity in a reservoir sand layer,
operators often used data from bulk oil sampling
and formation pressure at several depths to make
judgments on connectivity. Data on GOR, stock-
tank oil density and formation pressure from the
two Marathon wells spanning about 500 ft
[152 m] of depth in Sand A show differences that
suggest barriers to connectivity. In particular, the
pressure gradients from the two wells do not
appear to coincide, which is indicative of a seal-
ing barrier. However, these differences may
reect either measurement imprecision or differ-
ences in the way the data were collected (above).
>
Fluid properties and formation pressure in a eld in the Gulf of Mexico. DFA data on GOR (left) and density (center) from two wells
in Sand A show variability that lies either within or very close to the measurement error bands; scientists can draw no denitive
conclusions about connectivity. Data on formation pressure (right) show a difference between the two wells, suggesting a potential
for a sealing barrier. However, since the pressure measurements on the two wells were conducted on different wireline runs, and
the runs used different pressure gauges, assessment of connectivity using pressure was also inconclusive.
1,000 1,100 1,200 X,200
V
e
r
t
i
c
a
l

d
e
p
t
h
,

f
t
Fluid GOR, ft
3
/bbl Stock-tank oil density, g/cm
3
Formation pressure, psi
X,275 X,350 1,300 0.85 0.86 0.87 0.88 0.89 0.90
X,000
X,050
X,100
X,150
X,200
X,250
X,300
X,350
X,400
X,450
X,500
Well 5 Water Well 1
23. Dong C, Petro D, Latifzai AS, Zuo J, Pomerantz AE and
Mullins OC: Evaluation of Reservoir Connectivity from
Downhole Fluid Analysis, Asphaltene Equation of State
Model and Advanced Laboratory Fluid Analyses, paper
SPE 158838, presented at the SPE Annual Technical
Conference and Exhibition, San Antonio, Texas, October
810, 2012.
24. Resins are a solubility class similar to asphaltenes and
are typied by polyaromatic hydrocarbon molecules.
25. Elshahawi H, Ramaswami S, Zuo JY, Dong C, Mullins OC,
Zhang D and Ruiz-Morales Y: Advanced Reservoir
Evaluation Using Downhole Fluid Analysis and
Asphaltene Flory-Huggins-Zuo Equation of State,
prepared for presentation at the SPWLA 54th Annual
Logging Symposium, New Orleans, June 2226, 2013.
26. The ability to absorb light and then uoresce is
characteristic of some light oils. Like optical density,
uorescence intensity is dimensionless. For more:
Creek et al, reference 11.
Winter 2012/2013 23
Using these data, Marathon engineers found it
difcult to determine whether Sand A is hydrauli-
cally connected between Wells 1 and 5.
In addition to these uid properties and for-
mation pressures, the operator also obtained
downhole optical density measurements at sev-
eral depth stations for the two wells in Sand A
(above). The agreement with the Flory-Huggins-
Zuo EOS model indicates that the asphaltenes
in the two wells are in equilibrium as 2-nm
asphaltene nanoaggregates; this analysis pre-
dicts connectivity in Sand A between the two
wells. Similar analyses of other sand layers in
this eld did not show equilibrium in some
cases, prompting the operator to conclude that
there was no connectivity between those sands.
Actual eld production data conrmed all
asphaltene analysisbased predictions regard-
ing connectivity between sands.
The new science on asphaltenes can also be
useful in analyzing lighter oils and even conden-
sates that contain essentially no asphaltenes but
have heavy resins.
24
A well in the Gulf of Mexico,
operated by Shell, illustrates this concept.
25
The
light oil column from this well has virtually no
asphaltenes and a large GOR variation: from
4,000 ft
3
/bbl [720 m
3
/m
3
] at the top of the column
to 2,600 ft
3
/bbl [463 m
3
/m
3
] at a depth 440 ft
[134 m] below it.
>
Fluid optical density in Sand A. The operator
obtained reservoir uid optical density
measurements at several DFA depth stations for
wells that penetrated Sand A. These data, which
showed a smooth increase with depth, t the
Flory-Huggins-Zuo EOS prediction using 2-nm
nanoaggregates as the asphaltene particle state.
0 0.1 0.2 0.3 0.4 0.5
V
e
r
t
i
c
a
l

d
e
p
t
h
,

f
t
Optical density
X,000
X,050
X,100
X,150
X,200
X,250
X,300
X,350
X,400
X,450
X,500
Well 5 Water Well 1 EOS model
The operator had to contend with three
issues: Describe the compositional variation of
the bulk oil in terms of GOR and other parame-
ters, identify the gradient of the heavy compo-
nents in the volatile oil and decide if the
production zone was connected. To answer these
questions, Shell obtained DFA data at ve depth
stations from the top to the bottom of the oil col-
umn. GOR and other properties were analyzed
using a traditional cubic EOS in conjunction with
established correlations. Results from this work
were tuned to experimental data to give a satis-
factory t. Shell engineers found it difcult, how-
ever, to quantify optical characteristics of the oil
in this column to answer questions about the
heavy-end gradient and connectivity.
In most crude oils, optical densities offer good
sensitivity for measuring the relative concentra-
tion of heavy ends. However, for nearly colorless
oils, such as this Shell light oil, optical density is
not sensitive enough, especially at very high GOR
levels and low heavy-end concentrations. The dif-
ference between colorless oils that have 100%
light transmission and almost colorless oils that
have 99% light transmission is difcult to discern
using only optical density. Fluorescence inten-
sity, however, is applicable to this type of sample
and may be correlated directly to the fraction of
heavy resin or asphaltenes.
26
In this case, both
optical methods were used to give a complete
color description of the resin concentration gra-
dient with depth (below). Because of the small
V
e
r
t
i
c
a
l

d
e
p
t
h
,

m
Fluorescence intensity
Optical density
540
500
460
420
380
340
0
0 0.04 0.08 0.12 0.16 0.20 0.24 0.28
0.3 0.2 0.1
Optical density
Fluorescence intensity
EOS model
>
Optical gradients in a light oil column. Shell engineers used uorescence intensity and optical
density to measure the concentration of heavy molecules in an oil column from a well in the Gulf of
Mexico. Application of the Flory-Huggins-Zuo EOS to these DFA data with a 1-nm particle size is a
good t and indicates the heavy resin molecules in the column are in equilibrium. The color
characteristics of this light oil are similar to those observed in prior work in which the source of the
blue color was identied as the ve-ring PAH perylene (inset). If the properties for perylene are used
in the EOS for these data, the calculated size of the heavy resin is 0.96 nm, suggesting that the source
of the color in this oil column is perylene-like molecules.
24 Oileld Review
1-nm resin particle size, the gravity term in the
Flory-Huggins-Zuo EOS is also small, and the
expression is dominated by the GOR effect on the
solubility term. The equilibrium distribution of
resin molecules indicates that this oil column is
connected, as conrmed by subsequent produc-
tion data. These results suggest that this
approach is useful not only for black oils but also
light oils and rich gas condensates. Extending
this methodology to mobile heavy oil in a large
Middle Eastern eld completes the picture.
A large anticlinal oil reservoir operated by
Saudi Aramco has proved challenging to describe
by conventional modeling.
27
The low GOR oil col-
umn in this eld is stratied and is characterized
by black oil at the crest and mobile heavy oil
below it, with a tar mat above the oil/water con-
tact at the bottom. Although the black oil portion
is manageable from a production viewpoint, the
asphaltene concentrations in the mobile heavy
oil increase sharply with depth, and the oil
reaches a viscosity of about 1,000 cP [1,000 mPa
.
s]
just above the tar mat. Conventional PVT model-
ing does not account for these viscosity gradient
observations, and the existence of these discrete
zones represents major challenges in developing
production plans. Saudi Aramco engineers
turned to asphaltene science to help them under-
stand and model the compositional gradients in
this reservoir.
DFA and laboratory data on the different com-
position zones in this reservoir were obtained
from eight wells around the circumference of the
eld (above). The data show that the top and
majority of the column is black oil with less than
5 wt % asphaltenes and little concentration gradi-
ent with depth. This is consistent with an interval
containing mostly 2-nm nanoaggregates. The
next portion of the column is mobile heavy oil
with an asphaltene content ranging from 5 to
35 wt % asphaltenes. Using the Yen-Mullins model
as a guide, scientists determined that mobile
heavy oil with this range of asphaltenes should
exist primarily as 5-nm clusters (next page). The
tar mat with an asphaltene concentration greater
than 35 wt % is at the bottom of the oil column.
Asphaltene concentrations in the tar mat are
irregular with depth, indicating that this zone is
not equilibrated. The tar mat and heavy oil sec-
tions of this reservoir resulted from gravitational
accumulation of asphaltenes at the base of the oil
column, possibly from a late gas charge.
The combination of the detailed DFA data on
asphaltene concentrations and viscosity, coupled
with the agreement with the asphaltene science,
is important in describing this complex reservoir.
These data on viscosity, connectivity and location
of the tar mat have a signicant impact on pro-
duction planning for this eld.
Determining oil drainage patterns and con-
nectivity in a specic area is an important out-
come but is only the beginning for asphaltene
science. Going from black oil, characterized by a
few simple properties, to oil columns and reser-
voirs with detailed compositions is one part of
that frontierbut there are other possible direc-
tions as well.
New Frontiers
Few compounds among the thousands found in
crude oil have evoked as much interest and avoid-
ance as asphaltenes. In the past, asphaltenes
often meant operating problems for producers
and difculties for reners because of their high
molecular weight, high viscosity, plugging char-
acteristics and high levels of molecular contami-
nants. Scientists and engineers, long fascinated
by these heavy molecules, have persevered in
their attempts to understand and characterize
them. The result is a new branch of asphaltene
chemistry that is changing the ways in which sci-
entists view connectivity of oil columns within
the same reservoir. Through the use of advanced
sampling and analysis techniques such as DFA,
scientists are able to extend these new ways of
looking at asphaltenes from single wells to adja-
cent wells and reservoirs. The next step is to
extend that view across entire producing basins.
Proper incorporation of diverse phenomena,
such as large GOR variations, pressure gradients,
asphaltene gradients and the presence of tar
mats, will aid operators in eld development and
planning. At the current stage, these analyses
apply to oil columns and reservoirs at equilib-
rium. Extending this type of analysis to the fac-
tors causing a reservoir to be out of equilibrium
allows the theory to be applied to a wider range of
situations, as has been shown in deepwater wells
in the Gulf of Mexico.
28
>
Downhole uid analysis testing. Characterization
of this Middle East reservoir was accomplished
by analysis of samples from eight wells around
its periphery. DFA and oil samples provided
useful data on the black oil and mobile heavy oil
zones, while data on the tar mat zone were
obtained through core analysis.
Well
Black oil
Heavy oil
Tar mat
Water
27. Seifert DJ, Zeybek M, Dong C, Zuo JY and Mullins OC:
Black Oil, Heavy Oil and Tar in One Oil Column
Understood by Simple Asphaltene Nanoscience, paper
SPE 161144, presented at the Abu Dhabi International
Petroleum Exhibition and Conference, Abu Dhabi, UAE,
November 1114, 2012.
28. Zuo et al, reference 18.
29. Edwards DA, Gunasekera D, Morris J, Shaw G, Shaw K,
Walsh D, Fjerstad PA, Kikani J, Franco J, Hoang V and
Quettier L: Reservoir Simulation: Keeping Pace
with Oileld Complexity, Oileld Review 23, no. 4
(Winter 2011/2012): 415.
Winter 2012/2013 25
In addition to advancements in understand-
ing equilibrium, ascertaining connectivity and
predicting oil column gradients, the new asphal-
tene science has spawned unexpected and poten-
tially useful applications for other areas such as
enhanced oil recovery. For some time, scientists
and engineers have known that asphaltenes have
certain interfacial characteristics that are simi-
lar to those of naturally occurring surfactants.
For example, asphaltenes can alter the balance
between oil-wet and water-wet zones in a reser-
voir. Because mixed wettability zones may
contain nearly one half of eld reserves in
large Middle East reservoirs, the capability of
asphaltenes to change wettability could result in
large increases in recovery.
Another branch of work on asphaltenes
applies to viscosity and its prediction. Large vis-
cosity gradients are a natural consequence of the
asphaltene concentration gradient. The ability to
predict gradients in asphaltenes and viscosity for
oil columns brings up an interesting possibility.
Advanced reservoir simulatorssuch as the
INTERSECT reservoir simulatornow use clus-
ters of parallel computers to solve the thousands
of equations necessary to model and predict the
properties of an entire eld.
29
These equations
simulate the material, energy and property bal-
ances for small cubic reservoir sectionscalled
cellsas a function of time and position in the
reservoir. Cell size in these simulators has contin-
ued to decrease as computational power has
increased, and modern simulators now handle
cells as small as 50 m [164 ft] in the large reser-
voirs of the Middle East. Geoscientists hope to
merge the new asphaltene science and gradient
predictions with reservoir simulation so that
asphaltene and viscosity predictions are made for
the entire eldvertically and horizontally.
These new reservoir simulators not only model
eld composition and properties but also include
modules for eld management and facilities
planning. The ability to make good predictions
for asphaltene gradients would be an additional
step in optimizing eld development.
Future possibilities for applying fundamental
knowledge about asphaltenes abound. Knowing
more about property and asphaltene gradients
throughout oil elds will not only aid operators in
making better decisions about eld development,
but may yield benets in areas as diverse as res-
ervoir connectivity, viscosity gradients and
enhanced oil recovery. DA
>
Mobile heavy oil characterization. Application of the Flory-Huggins-Zuo EOS for asphaltenes to the mobile heavy oil
data (left) yields a cluster size of 5.2 nm, conrming the expected size of 5 nm. For this mobile heavy oil zone, EOS
gravity is the only term needed to describe the six-fold variation in asphaltene content over the periphery of this eld.
The photograph shows a mobile heavy oil in the laboratory.
V
e
r
t
i
c
a
l

d
e
p
t
h
,

f
t
Asphaltene, wt %
Mobile heavy oil
Tar mat
Black oil
Y,100
Y,150
X,900
X,950
Y,050
X,800
X,850
X,700
X,750
Y,200
0 5 10 20 30 40 15 25 35 45
Y,000
EOS model
Laboratory data
Tar mat
40 45