Sunteți pe pagina 1din 8

Light-scattering study of petroleum asphaltene

aggregation

Yevgeniy G. Burya, Igor K. Yudin, Victor A. Dechabo, Victor I. Kosov, and Mikhail A. Anisimov

Dynamic light scattering with an original optical scheme has been used for the investigation of opaque
共strongly light-absorbing兲 asphaltene colloids in crude oils and hydrocarbon mixtures. Diffusion-limited
aggregation and reaction-limited aggregation as well as a crossover between these two regimes have been
observed. A simple interpolation for the crossover kinetics is proposed. Asphaltene colloidal struc-
tures, originally persisting in crude oils, have been detected. Addition of a precipitant above a threshold
induces asphaltene aggregation. Depending on the nature of the precipitant, different crude oils re-
spond differently on its addition: 共a兲 exponential-in-time growth of aggregates to huge flocks or 共b兲 fast
formation of stable-in-size particles. © 2001 Optical Society of America
OCIS codes: 000.2190, 000.2170.

1. Introduction troleum industry is to separate the crude oil into four


Optical spectroscopy provides a powerful methodol- chemically distinct fractions: saturates, aromatics,
ogy for studying molecular and supramolecular struc- asphaltene, and resins 共SARA兲. Two key parame-
tures and their dynamics. To investigate the ters that control the stability of asphaltene particles
colloidal properties of petroleum asphaltene particles in a crude oil are the ratio of aromatics to saturates
suspended in pure hydrocarbons and in crude oils, we and that of resins to asphaltenes. When this ratio
used static and dynamic light-scattering techniques. decreases, asphaltene particles may coalesce and
To elucidate the nature and the mechanism of the form larger aggregates.2
aggregation kinetics of colloidal asphaltene particles, Analysis of experimental data indicates that the
we performed real-time monitoring measurements. formation of associates in crude oil and petroleum
The general characterization of crude oils and as- liquid fractions is complicated and cannot be de-
phaltenes 共a heavy petroleum fraction兲 is an impor- scribed by a simple model. Non-Newtonian rheo-
tant task for the petroleum industry.1–3 Different logical behavior of petroleum systems 共especially
crude oils exhibit an enormous range of properties those with high-content asphaltenes and paraffins兲
that influence all phases of oil exploitation. Crude suggests the presence of a supramolecular meso-
oil is a complex mixture of hydrocarbons and hetero- scopic structure. However, the physical nature of
atomic organic compounds of varying molecular this structure is not well understood. It is com-
weight and polarity. A common practice in the pe- monly accepted that the mesoscopic structure is
somehow associated with asphaltenes. Asphaltenes
are defined as solidlike heavy fractions of crude oils
Y. G. Burya is with the I. M. Gubkin State University of Oil and
Gas, Leninsky Prospect 65, Moscow 117917, Russia, and the In-
that are completely miscible in aromatic hydrocar-
stitute for Physical Science and Technology, University of Mary- bons such as toluene or benzene but are insoluble in
land, College Park, Maryland 20742. I. K. Yudin, V. A. Dechabo, n-alkanes. This property is used for the extraction
and V. I. Kosov are with the Oil and Gas Research Institute, of asphaltenes.3 As asphaltenes are poorly charac-
Russian Academy of Sciences, Gubkina 3, Moscow 117971, Russia. terized, the above definition is in some sense arbi-
Mikhail A. Anisimov 共anisimov@glue.umd.edu兲 is with the Insti- trary. Petroleum asphaltenes are complex aromatic
tute for Physical Science and Technology and the Department of macrocycle molecules surrounded and linked by ali-
Chemical Engineering, University of Maryland, College Park,
phatic chains and heteroatoms.3 The molecular
Maryland 20742.
Received 26 September 2000; revised manuscript received 16
structure of asphaltenes varies significantly, depend-
January 2001. ing on their origin, method of petroleum recovery,
0003-6935兾01兾244028-08$15.00兾0 and history of extraction.2 This study is one of the
© 2001 Optical Society of America first attempts, to our knowledge, to find a reliable

4028 APPLIED OPTICS 兾 Vol. 40, No. 24 兾 20 August 2001


Fig. 1. Light-scattering geometries: 共a兲 backscattering and 共b兲 corner scattering.

correlation between asphaltene stability and the col- where kB is Boltzmann’s constant and T is the tem-
loidal properties of crude oils. perature.
The kinetics of the aggregation of colloidal particles Equation 共2兲 is valid for noninteracting particles.
has been a subject of extensive experimental and the- If the particles are involved in an aggregation pro-
oretical study. Numerous studies of the structure cess, they certainly interact. However, Eq. 共2兲 is still
and the aggregation of gold colloids, silica, polysty- applicable to the monitoring of the change in the
rene latex, and melanin particles have been report- apparent particle size if the characteristic time of
ed.4,5 The universality of the aggregation regimes aggregation is much greater than the time of a single
for these classical colloidal systems is discussed.4,6 – 8 measurement. It is worthy to note that, because of
Nevertheless, the mechanism of aggregation, the pro- the low resolution of particle sizing in asphaltene
cess of formation of fractal aggregates, and a cross- opaque systems, one can measure a mean particle
over between different regimes of aggregation, size only. The opacity of the asphaltene solution
especially for complex colloidal systems, still requires prevents reliable measurements of the angular dis-
further investigation. In this paper, we present the tribution of the light-scattering intensity as well.
results of a study of the aggregation kinetics of as- Detailed information concerning polydispersity, rota-
phaltene colloidal particles in hydrocarbon media by tional diffusion, and particle fractality4,11–14 for these
dynamic light scattering. very complex systems still cannot be obtained.
We used an original optical scheme that enables us
2. Method and Technique to perform dynamic light-scattering measurements
A photon-correlation spectrometer with an original in disperse petroleum systems and, in particular, to
optical scheme and a multifunctional single-board study aggregation phenomena in opaque solutions of
correlator to perform dynamic light-scattering mea- petroleum asphaltenes.9 Two suitable light-
surements was used in the experiments.9 scattering geometries were used for particle sizing in
The desired information about particle motion can opaque petroleum dispersions. In the backscatter-
be obtained from the correlation function of the ing geometry 关Fig. 1共a兲兴 a rectangular sample cell is
scattered-light intensity, which has a characteristic placed at an angle of 45° with respect to the incident
decay time tc. The diffusion coefficient D of the par- laser beam that emanates from a low-intensity
ticles suspended in liquid media is related directly to He–Ne laser 共2 mW, 632.8 nm兲 with the intensity of
the decay rate of the time-dependent correlation func- the laser beam further attenuated by a neutral glass
tion. For the homodyne spectrum the relation be- filter to avoid thermal lens effects. The photon-
tween D and tc reads10 detecting system is located such that its optical axis
is normal to the surface of the sample cell. The scat-
D ⫽ 1兾2t c k 2, (1) tered light is collected from the region where the laser
beam enters the sample. To obtain optimal condi-
where k is the wavenumber of the scattered light, k ⫽ tions for collecting scattered light and for suitable
共4␲n兾␭兲sin共⌰兾2兲, n is the refractive index of the me- alignment of the incident laser beam required the
dium, ␭ is the light wavelength in vacuum, and ⌰ is design of an optical plate. The position of the scat-
the scattering angle. tering volume can be adjusted to study light scatter-
The Stokes–Einstein equation interrelates the dif- ing near the wall of the cell 共approximately 200 ␮m
fusion coefficient D, the shear viscosity ␩ of the me- away from the wall to exclude the influence of the
dium, and the mean hydrodynamic radius of the surface兲 as well as in the bulk of the fluid. Such a
particles R by scheme allows one to measure particle size in any
opaque liquid. The corner geometry 关Fig. 1共b兲兴 is
D ⫽ k B T兾6␲␩R, (2) suitable for semitransparent dispersions. Such a

20 August 2001 兾 Vol. 40, No. 24 兾 APPLIED OPTICS 4029


scheme enables one to minimize stray light. We
take special precautions to prevent heterodyning that
is due to stray light. We used standard square
4-mm path-length optical-glass cells produced by
Hellma Cell, Inc.
Asphaltenes together with resins are the compo-
nents that make most crude oils black in color.2 So
in asphaltene solutions as well as in native crude oils
visible light is absorbed because of the presence of
those components. It is also important to emphasize
that, in these opaque systems, strong light absorption
by asphaltene particles practically eliminates multi-
ple light-scattering effects.
3. Experiment and Results
Although the complexity of the samples to be studied
imposes limitations on the experiment, we took sev-
eral precautions to improve the quality of measure- Fig. 2. Mean radius of the aggregates plotted as a function of time
ments. Petroleum samples were dehydrated and for initial asphaltene solutions of 1 g兾l 共squares兲, 10 g兾l 共circles兲,
filtered. Filtered n-heptane was used as a precipi- and 5 g兾l 共triangles兲 in toluene on adding the near-threshold
tant. To avoid a thermal lens effect, we used a nat- amount of n-heptane.
ural filter. Fresh samples were prepared for each
experiment. Immediately after the addition of
n-heptane particles sizing was started. To obtain a tion time ␶R. If ␶D ⬎⬎ ␶R the aggregation kinetics are
correlation function of reasonably good quality re- controled by diffusion and referred to as diffusion-
quired that the accumulation time for each measure- limited aggregation 共DLA兲. In the case of ␶R ⬎ ␶D
ment last from 1 to 15 min, depending on the rate of the aggregation kinetics are determined by the rate of
the process. The total duration of each experimen- the reaction and referred to as reaction-limited ag-
tal run varied from 10 min to several days. All the gregation17,18 共RLA兲. Following from a solution of
experiments were made at ambient conditions. the Smoluchowski equation, DLA and RLA kinetic
processes are described by the simple formulas
A. Toluene–n-Heptane Mixtures
N ⫽ 1 ⫹ t兾␶ D, R ⫽ R 0共1 ⫹ t兾␶ D兲 1兾df, (3)
The experimental procedure included both light-
scattering intensity measurements and the accumu- N ⫽ exp共t兾␶ R兲, R ⫽ R 0 exp共t兾␶ R d f 兲, (4)
lation of the autocorrelation function in the
homodyne regime. The resultant correlation func- respectively, where N is the average number of par-
tion was fitted to a single-exponential decay function ticles in an aggregate, R is the mean radius of an
with one cumulant, an effective diffusion coefficient aggregate, R0 is the initial particle size, and t is time.
was deduced, and the mean radius of the particles In Equations 共3兲 and 共4兲 df is the fractal dimension-
was calculated. ality according to the definition N ⫽ 共R兾R0兲df. It is
The asphaltenes used were extracted from important to note that df is different for DLA and
Karazhanbas crude oil 共Kazakhstan兲. Three initial RLA. We propose a simple equation to incorporate
solutions of asphaltenes in toluene that had concen- crossover behavior from RLA to DLA, as given by
trations of less than the critical micelle concentration
共CMC兲, of 10 g兾l 共well above the CMC兲, and of 5 g兾l dN兾dt ⫽ N兾共␶ D␥N ⫹ ␶ R兲, (5)
共slightly above the CMC兲 were prepared. The CMC where ␥ ⫽ 关R兾R0共D兲兴df is a constant. For ␶D␥N much
is 3 g兾l. For each initial concentration of asphalt- smaller than ␶R Eq. 共5兲 implies RLA. As N grows,
enes, we determined the threshold concentration of ␶D␥N ultimately becomes larger than ␶R, when ␶D␥N
n-heptane in the following way: Heptane was grad- ⬎⬎ ␶R, and DLA prevails. Integration of Eq. 共5兲 gives
ually added to the asphaltene solutions until the ag- the crossover equation for the aggregation number N
gregation of asphaltene started. We observed the as
growth of asphaltene aggregates for solutions with
the threshold concentration of n-heptane as well as 共N ⫺ 1兲共␶ D␥兾␶ R兲 ⫹ ln N ⫽ t兾␶ R, (6)
for solutions with a number of other concentrations of
n-heptane that were higher than the threshold con- and for the mean radius R as
centration. In all, aggregation kinetics for three ␶ D关R兾R 0共D兲兴 df ⫹ ␶ R d f ln关R兾R 0共R兲兴 ⫽ t. (7)
concentrations of n-heptane 共initial concentration of
asphaltene of 1 g兾l兲 and seven concentrations of Our experimental results for three initial concen-
n-heptane 共initial concentration of asphaltene 10 g兾l兲 trations of asphaltenes are represented in Fig. 2.
were studied.15,16 The solid curves represent the fitting of the data in
Generally, the aggregation process involves two accord with Eqs. 共3兲 for DLA, Eqs. 共4兲 for RLA, and
characteristic times: a diffusion time ␶D and a reac- Eq. 共7兲 for the crossover behavior.

4030 APPLIED OPTICS 兾 Vol. 40, No. 24 兾 20 August 2001


Fig. 4. Viscosity of crude oil–n-heptane mixtures of SkyBlue oil
Fig. 3. Scaled mean number of particles in an aggregate plotted 共squares兲 and Swanson River oil 共circles兲 at constant temperature.
as a function of the scaled time t* for the DLA 共dashed curve兲, the
RLA 共solid curve兲, and the crossover 共dotted curve兲 in a 5-g兾l as-
phaltene solution.
ratio, this oil differs from the others, so we can expect
different aggregation behavior.
To interpret our experimental data in the sense of
The two limiting regimes of aggregation and a particle size, we need to know the shear viscosity of
crossover between them for a 5-g兾l asphaltene solu- the oil–precipitant mixtures. The viscosity of the
tion exhibit a universal behavior, as shown in Fig. 3. oils and the oil–precipitant mixtures was measured
If one uses the scaled variables R兾R0 and t* ⫽ t兾␶char, by use of a Rheometrics fluids spectrometer 共Model
with the characteristic time of ␶char ⫽ ␶R for the RLA RFS II兲. The Model RFS II is a laboratory instru-
and the crossover curves and with ␶char ⫽ ␶D for the ment for studying the viscous and the viscoelastic
DLA curves, the three DLA curves and the seven properties of both structured and unstructured fluids.
RLA curves collapse into two master curves, and the The RFS II instrument applies a steady rotational or
crossover between the two master curves can clearly dynamic oscillatory shear, or each form of shear se-
be seen 共the solid circles兲. The scaling factor ␬ ⫽ 43 quentially, to the test material. By measuring the
is the ratio of the two characteristic times, ␶R兾␶D␥, torque transmitted through the material, one can de-
and it was found from the fitting of Eq. 共7兲 to the data rive the material’s storage 共or elastic兲 modulus, loss
that exhibit crossover behavior. For DLA N corre- 共or viscous兲 modulus, and viscosity.
sponds to the average number of asphaltene mole- As all the light-scattering experiments were car-
cules in an aggregate, whereas for RLA and for the ried out at room temperature, the viscosity was mea-
crossover aggregation the aggregation number N has sured at two different temperatures close to room
represents the average number of micelles in an ag- temperature. The viscosity was then extrapolated
gregate. These two quantities are related by the to a required temperature.
factor ␥. The procedure of viscosity measurements was as
follows: After the temperature of the cap and the
B. Crude Oil Samples: Viscosity Measurements bob were stabilized crude oil was transported into the
Three different samples of crude oils 共supplied by cap and mixed for some time. After that the crude
Shell Oil Company兲 were investigated. The SARA oil viscosity value was taken with the RFS II. Then
analysis for the crude oils is presented in Table 1. n-heptane was gradually added, and the viscosity of
According to information from Shell, these oils had mixtures at different oil:precipitant ratios was ob-
asphaltene-instability problems. In particular, the tained. The shear viscosity of the oil–precipitant
first oil caused troubles during oil recovery and trans- mixtures for the Swanson River and the Sky Blue oils
portation. With respect to the aromatics兾saturates is presented in Fig. 4.

Table 1. Representative SARA Values for the Crude Oil Samples

SARA Values

Crude Oil Saturates 共%兲 Aromatics 共%兲 Resins 共%兲 Asphaltenes 共%兲

Swanson River 53.1 35.2 7.2 4.53


European SkyBlue 27.13 56.72 9.4 6.75
Mars A-12 24.3 55.6 14.9 5.15

20 August 2001 兾 Vol. 40, No. 24 兾 APPLIED OPTICS 4031


Fig. 5. Size of the aggregates plotted as a function of the moni- Fig. 6. Mean radius of the aggregates plotted as a function of the
toring time in Unocal’s Swanson River crude oil. The squares monitoring time in Unocal’s Swanson River oil mixtures with dif-
represent the first run, and the circles the second run. ferent precipitant concentrations: 25% n-heptane 共filled squares兲,
18% n-heptane 共filled triangles兲, 9% n-heptane 共open diamonds兲,
5% n-heptane 共open circles兲.

C. Crude Oil Samples: Light-Scattering Measurements


The effective Brownian diffusion coefficient detected cess is similar to the process observed with pure oil.
in the Swanson River oil was found to change with The rest of the experimental data represents expo-
the monitoring time. Initially, the diffusion coeffi- nential growth with a characteristic time of 2300 min
cient was approximately 5 ⫻ 10⫺10 cm2 s⫺1, which in the 5% mixture 共Fig. 8兲. However, the initial
corresponds to a particle size of approximately 1 ␮m, stage of this exponential growth process is concealed
but within 5 days the particle size increased by a by the power-law process because larger particles
factor of 2 共Fig. 5兲. During the experiment, the laser dominate. This overlapping area prevents us from
beam constantly entered the sample, and the growth distinguishing between these two processes with dy-
of the particle size might somehow be caused by laser namic light scattering. The point of intercept of
radiation. However, this effect requires further in- these two processes is shifted to lower times with an
vestigation. To exclude any thermal lens effect, we increase in the precipitant concentration 共Fig. 6兲.
reduced the laser power with a neutral filter to ap- This Swanson River crude oil was found to be close
proximately 12 times the attenuation. to aggregation instability. We could not detect a
Spectacular aggregation kinetics were observed in threshold concentration of the precipitant for aggre-
the mixtures of the oil with n-heptane added. Typ- gation onset. Asphaltene particle growth was ob-
ical curves for oil–precipitant mixtures with different served in a solution with a concentrations of
n-heptane concentrations that represent a slow in- n-heptane as low as 3%.
crease of particle size in the double-logarithmic scale Unocal’s Swanson River oil clearly reveals an in-
are shown in Fig. 6. The scattered intensity was
found to increase during the experiments. The pro-
cess is considered a colloidal particle-aggregation pro-
cess, and it represents exponential growth of the
particle size. The rate of aggregation strongly de-
pends on the precipitant concentration 共Fig. 7兲. An
exponential fit agrees reasonably well with the ex-
perimental data and gives us a characteristic concen-
tration for this oil sample that is equal to
approximately 4% of the n-heptane concentration.
The addition of a precipitant reduces the energy bar-
rier of the reaction of aggregation, and a greater
amount of the precipitant results in a higher aggre-
gation rate. Such behavior corresponds to the RLA
observed in toluene– heptane solutions of asphalt-
enes.
The particle-size growth in mixtures with 5%, 9%,
and 18% n-heptane involves two different parts. Fig. 7. Exponential growth of the aggregation rate plotted as a
The initial experimental data obey a power law with function of the precipitant volume concentration for Unocal’s
an exponent close to 1兾3. This extremely slow pro- Swanson River oil–precipitant mixtures.

4032 APPLIED OPTICS 兾 Vol. 40, No. 24 兾 20 August 2001


Fig. 8. Mean radius of the asphaltene aggregates plotted versus
the monitoring time for Unocal’s Swanson River oil with 5% by
volume of the precipitant. Two dynamic processes, power-law
growth and exponential growth, can clearly be seen in the double-
logarithmic scale.

trinsic colloidal structure that appears unstable.


Mixtures of the oil and n-heptane show spectacular
aggregation phenomena in which the rate depends on
the oil:precipitant ratio.19 In accord with the SARA
analysis, the content of this oil is quite different from
Fig. 9. 共a兲 Mean radius and 共b兲 light-scattering 共LS兲 intensity for
that the other oils 关there is a large fraction of satu- SkyBlue oil with 44% by volume of the precipitant n-heptane plot-
rated hydrocarbons, and paraffins 共see Table 1兲兴. So ted as a function of the monitoring time.
the analysis gives us a good indication of crude oil
stability against asphaltene aggregation.
An effective diffusion coefficient of approximately period of more than 20 h for a solution with a 30%
2 ⫻ 10⫺9 cm2 s⫺1 was measured in the European Sky- concentration of n-heptane, for a time period of less
Blue crude oil and does not change with the monitoring than 10 h for a solution with a concentration of 44%
time for at least 24 h. This diffusion coefficient cor- n-heptane, for a duration of less than 6 h for a solu-
responds to particles of approximately 20 nm in diam- tion with a concentration of 49% n-heptane, and for a
eter. In a mixture with 20% of the precipitant added,
the diffusion coefficient increased to as high as 10⫺8
cm2 s⫺1, but the particle size remained the same. The
experiment with this mixture lasted for 24 h and
showed no significant change in the size of the parti-
cles or the scattered intensity during the experiment.
In solutions with concentrations of n-heptane of 25%
and higher two effective particle sizes were detected by
the analysis of the dynamic correlation function in dif-
ferent correlation-time scales. Although the smaller
particles remained in the system, larger aggregates
formed. Although the intensity of the scattered light
decreased with the monitoring time, the diffusion co-
efficients did not change by more than a factor of 2.
This change can be attributed to worse experimental
conditions, namely, a low scattered intensity. An ex-
ample of the process for a mixture with 44% n-heptane
in shown in Figs. 9共a兲 and 9共b兲. The size of the larger
aggregates depends on the amount of the precipitant Fig. 10. Average size of the asphaltene particles for various con-
added and grows to as high as several micrometers centrations of the precipitant n-heptane. The squares represent
共Fig. 10兲. small asphaltene aggregates in SkyBlue oil; the circles represent
Depending on the concentration of the precipitant, large aggregates of asphaltenes in SkyBlue oil; the triangles rep-
we were able to measure the particle size over a time resent aggregates of asphaltenes in Mars A12 oil.

20 August 2001 兾 Vol. 40, No. 24 兾 APPLIED OPTICS 4033


duration of more than 2 h for a solution with a con- the asphaltene solutions with concentrations that are
centration of 53% n-heptane. After the aforemen- slightly higher than the CMC threshold and a con-
tioned time periods the scattered intensity became so centration of the precipitant 共n-heptane兲 that is near
low that we could not take further measurements. the precipitation threshold a crossover between these
The effective particle sizes were restored after the two regimes has been observed. In terms of scaled
sample was shaken, and they disappeared within variables our data demonstrate a universal behavior
1.5 h. No slow change in particle size was observed that, we expect, may be present in other colloidal
in this crude oil sample. systems.
From the results described above, we can conclude We have also investigated the colloidal properties
that European SkyBlue oil has an intrinsic colloidal and the kinetics of asphaltene aggregation in crude
structure. The structure remains stable both in oil samples with a dynamic light-scattering method
pure oil and in solution with the addition of as much that was adapted to opaque fluids. For two oil sam-
as 20% of the precipitant. The size of the asphaltene
ples there exists an intrinsic colloidal-like structure
particles suspended in the oil media is approximately
with particles of approximately 0.02 and 1 ␮m in
20 nm. Addition of the precipitant at concentrations
higher than the threshold concentration initiates fast diameter. Our experiments show that, for one sam-
aggregation. Because the aggregates grow too fast, ple, this structure is somehow affected by laser radi-
their size change with time cannot be measured. ation, but the effect of photochemistry in crude oils
The addition of n-heptane causes the formation of requires further investigation. We intend to use a
large 共up to several micrometers兲 aggregates; how- He–Ne green laser and an infrared laser in future
ever, their size remains stable. Eventual sedimen- measurements to study the effect of the wavelength
tation follows aggregation, and the number of of the light.
particles in the scattering volume decreases with We assume that the particles observed are
time. The sedimentation rate depends on the con- asphaltene–resin colloidal particles. Using n-
centration of the precipitant. A higher precipitant heptane as the precipitant, we have studied the
concentration leads to a higher sedimentation rate stability of crude oil systems against asphaltene
and vice versa. The threshold concentration for this aggregation. One of the oils exhibits an extremely
oil sample is approximately 25% of the volume of slow aggregation process. The greater the amount
n-heptane. of the precipitant added, the faster the rate of ag-
No Brownian diffusion coefficient was detectable in gregation. There is no threshold concentration of
either the Mars A12 crude oil sample or in a solution of the precipitant for this sample. Two other samples
the crude oil with 20% n-heptane. A low diffusion exhibit different behavior. The second sample ex-
coefficient, which indicates large particles and does not hibits fast aggregation, which starts at the thresh-
change significantly with monitoring time, was ob- old concentration of the precipitant and lasts a
served in solutions of a 30% to a 60% concentration of short period of time. The process leads to the for-
n-heptane. The scattered intensity remained un- mation of a colloidal structure with stable asphalt-
changed as well. The diffusion coefficient was ap- ene aggregates. The third sample has no initial
proximately 6 ⫻ 10⫺10 cm2 s⫺1 for mixtures with colloidal structure, but it does occur in a solution
different oil:heptane ratios. However, the size of the with 20% n-heptane and remains stable. The ad-
particles grew from approximately 1 ␮m for a 33%
dition of more than 60% by volume of precipitant
volume concentration of n-heptane to 10 ␮m for a 60%
breaks the stability of the structure completely and
concentration. A greater than 60% amount of the
precipitant causes the formation of large clusters that causes the flocculation of particles into huge clus-
cannot be measured. The dependence of the size of ters, which cannot be measured by the light-
the aggregates on the concentration of the precipitant scattering method.
for the SkyBlue and the Mars A12 oils is shown in Fig. Our results have shown that the oils with high
10. SkyBlue oil and Mars A12 exhibit similar behav- aromatic:saturate ratios have a more stable structure
ior. The threshold concentration of aggregation is and exhibit no slow aggregation, whereas the oil with
clearly observed. a much lower ratio is on the edge of stability and
We concluded that the Mars A-12 oil has no detect- shows slow aggregation kinetics.
able colloidal structure. Adding n-heptane induces
the formation of stable colloidal particles. The The research at the University of Maryland is sup-
threshold concentration for its formation is approxi- ported by the Petroleum Research Fund of the Amer-
mately 20% n-heptane. A considerable amount of ican Chemical Society, grant 33899-AC9. We thank
the precipitant added 共more than 60%兲 breaks down Sada Iyer, Shell International, Westhollow Technol-
the stability of the system and causes the flocculation ogy Center, Houston, Texas, and Sheila Dubey,
of particles into huge clusters. Equilon Enterprises, Westhollow Technology Center,
Houston, Texas, for providing us with the crude oil
4. Conclusion samples and the SARA data. We are grateful to
We have demonstrated that the aggregation of as- E. E. Gorodetskii for fruitful discussions. We are
phaltene particles in hydrocarbon solutions obeys two particularly indebted to J. V. Sengers for help and
different kinetics laws, namely, DLA and RLA. For collaboration.

4034 APPLIED OPTICS 兾 Vol. 40, No. 24 兾 20 August 2001


References 11. P. Wiltzius, “Hydrodynamic behavior of fractal aggregates,”
1. E. Y. Shue and O. C. Mullins, eds., Asphaltenes: Fundamen- Phys. Rev. Lett. 58, 710 –713 共1987兲.
tals and Applications 共Plenum, New York, 1996兲. 12. Z. Y. Chen, P. Meakin, and J. M. Deutch, “Comment on ‘Hy-
2. O. C. Mullins and E. Y. Shue, eds., Structure and Dynamic of drodynamic behavior of fractal aggregates,’ ” Phys. Rev. Lett.
Asphaltenes 共Plenum, New York, 1998兲. 59, 2121 共1987兲.
3. R. B. Long, “The concept of asphaltenes,” in Chemistry of As- 13. P. N. Pusey, J. G. Rarity, R. Klein, and D. A. Weitz, “Comment
phaltenes, Vol. 195 of Advances in Chemistry Series, J. W. on ‘Hydrodynamic behavior of fractal aggregates,’ ” Phys. Rev.
Bunger and N. C. Li, eds. 共American Chemical Society, Wash- Lett. 59, 2122 共1987兲.
ington, D.C., 1981兲, pp. 17–27. 14. P. N. Pusey and R. J. A. Tough, “Particle interaction,” in Dy-
4. M. Y. Lin, H. M. Lindsay, D. A. Weitz, R. C. Ball, R. Klein, and namic Light Scattering, R. Pecora, ed. 共Plenum, New York,
P. Meakin, “Universal reaction-limited colloid aggregation,” 1985兲, pp. 85–179.
Phys. Rev. A 41, 2005–2020 共1990兲. 15. I. K. Yudin, G. L. Nikolaenko, E. E. Gorodeskii, V. I. Kosov,
5. J. S. Huang, J. Sung, M. Eisner, S. C. Moss, and J. Gallas, “The V. R. Melikyan, E. L. Markhashov, D. Frot, and Y. Briolant,
fractal structure and the dynamics of aggregation of synthetic “Mechanism of asphaltene aggregation in toluene– heptane
melanin in low pH aqueous solutions,” J. Chem. Phys. 90, mixtures,” J. Pet. Sci. Eng. 20, 297–301 共1998兲.
25–29 共1989兲. 16. I. K. Yudin, G. L. Nikolaenko, E. E. Gorodeskii, E. L.
6. M. Y. Lin, H. M. Lindsay, D. A. Weitz, R. C. Ball, R. Klein, and Markhashov, V. A. Agayan, M. A. Anisimov, and J. V. Sengers,
P. Meakin, “Universality in colloid aggregation,” Nature “Crossover kinetics of asphaltene aggregation in hydrocarbon
339共6223兲, 360 –362 共1989兲. solutions,” Physica A 251, 235–244 共1998兲.
7. M. Y. Lin, H. M. Lindsay, D. A. Weitz, R. C. Ball, R. Klein, and 17. D. A. Weitz, J. S. Huang, M. Y. Lin, and J. Sung, “Limits of the
P. Meakin, “Universality of fractal aggregates probed by light fractal dimension for irreversible kinetic aggregation of gold
scattering,” Proc. R. Soc. London Ser. A 423共1864兲, 71– 87 共1989兲. colloids,” Phys. Rev. Lett. 54, 1416 –1419 共1985兲.
8. M. Y. Lin, H. M. Lindsay, D. A. Weitz, R. C. Ball, R. Klein, and 18. D. A. Weitz and M. Y. Lin, “Dynamic scaling of cluster-mass
P. Meakin, “Universal diffusion-limited colloid aggregation,” J. distributions in kinetic colloid aggregation,” Phys. Rev. Lett.
Phys. Condens. Matter 2, 3093–3113 共1990兲. 57, 2037–2040 共1986兲.
9. I. K. Yudin, G. L. Nikolaenko, V. I. Kosov, V. A. Agayan, M. A. 19. I. K. Yudin, Y. G. Burya, V. A. Dechabo, E. E. Gorodetskii, V. I.
Anisimov, and J. V. Sengers, “A compact photon-correlation Kosov, and M. A. Anisimov, “Aggregation phenomena in pe-
spectrometer for research and education,” Int. J. Thermophys. troleum colloids studied by dynamic light scattering,” in Pho-
18, 1237–1248 共1997兲. ton Correlation and Scattering, Postconference Digest, Vol. 47
10. H. Z. Cummins and E. R. Pike, eds., Photon Correlation and of OSA Trends in Optics and Photonics 共Optical Society of
Light Beating Spectroscopy 共Plenum, New York, 1974兲. America, Washington, D.C., 2000兲, pp. 30 –32.

20 August 2001 兾 Vol. 40, No. 24 兾 APPLIED OPTICS 4035

S-ar putea să vă placă și