2728 Langmuir 2005, 21, 2728-2736

High-Q Ultrasonic Determination of the Critical

Nanoaggregate Concentration of Asphaltenes and the
Critical Micelle Concentration of Standard Surfactants
Gaëlle Andreatta, Neil Bostrom, and Oliver C. Mullins*
Schlumberger-Doll Research, Old Quarry Road, Ridgefield, Connecticut 06877

Received June 2, 2004. In Final Form: October 15, 2004

Asphaltenes are known to be interfacially active in many circumstances such as at toluene-water

interfaces. Furthermore, the term micelle has been used to describe the primary aggregation of asphaltenes
in good solvents such as toluene. Nevertheless, there has been significant uncertainty regarding the critical
micelle concentration (CMC) of asphaltenes and even whether the micelle concept is appropriate for
asphaltenes. To avoid semantic debates we introduce the terminology critical nanoaggregate concentration
(CNAC) for asphaltenes. In this report, we investigate asphaltenes and standard surfactants using high-Q,
ultrasonic spectroscopy in both aqueous and organic solvents. As expected, standard surfactants are shown
to exhibit a sharp break in sonic velocity versus concentration at known CMCs. To prove our methods,
we measured known surfactants with CMCs in the range from 0.010 g/L to 2.3 g/L in agreement with the
literature. Using density determinations, we obtain micelle compressibilities consistent with previous
literature reports. Asphaltenes are also shown to exhibit behavior similar to that of ultrasonic velocity
versus concentration as standard surfactants; asphaltene CNACs in toluene occur at roughly 0.1 g/L,
although the exact concentration depends on the specific (crude oil) asphaltene. Furthermore, using
asphaltene solution densities, we show that asphaltene nanoaggregate compressibilities are similar to
micellar compressibilities obtained with standard nonionic surfactants in toluene. These results strongly
support the contention that asphaltenes in toluene can be treated roughly within the micelle framework,
although asphaltenes may exhibit small levels of aggregation (dimers, etc.) below their CNAC. Furthermore,
our extensive results on known surfactants agree with the literature while the asphaltene CNACs reported
here are one to two orders of magnitude lower than most previously published results. (Previous work
utilized the terminology “micelle” and “CMC” for asphaltenes.) We believe that the previously reported
high concentrations for asphaltene CMCs do not correspond to primary aggregation; perhaps they refer
to higher levels of aggregation or perhaps to a particular surface structure.

I. Introduction mostly carbon and hydrogen (atomic ratio ∼1:1.2), and

Asphaltenes are the most aromatic and most enigmatic
component of crude oil.1-4 They have enormous impact in
all phases of petroleum resource utilization and, thus,
are of extreme interest. Their definition is operational
(e.g., insoluble in n-heptane, soluble in toluene) in part
because of the difficulty of assigning to them a proper
chemical classification. This definition ensures that a
single asphaltene sample is polydisperse in any property
of interest (e.g., molecular weight, polarity, number of
fused aromatic ring). Nevertheless, this definition does
provide sufficient chemical restriction to be useful in a
chemical sense in addition to an operational sense.
Predictive petroleum science rests in part on understand-
ing constituent molecular structures. As Crick advises,5
“to understand function, study structure.” A central focus
in our lab has been to establish asphaltene molecular
structures to then understand asphaltene function. Many
of the major molecular structural issues of asphaltenes
are now understood. In our first-principles approach, we
now launch our investigation into the primary aggregation
of asphaltenes.
Many molecular structural properties of asphaltenes
have been sorted out in recent years. Asphaltenes are
(1) Chilingarian, G. V., Yen, T. F., Eds. Bitumens, Asphalts, and Tar
Sands; Elsevier Scientific Publishing Co.: New York, 1978.
(2) Bunger, J. W., Li, N. C., Eds. Chemistry of Asphaltenes; American
Chemical Society: Washington, DC, 1984.
(3) Sheu, E. Y., Mullins, O. C., Eds. Asphaltenes: Fundamentals and
Applications; Plenum Pub. Co.: New York, 1995.
(4) Mullins, O. C., Sheu, E. Y., Eds. Structures and Dynamics of
Asphaltenes; Plenum Pub. Co.: New York, 1998.
(5) Crick, F. Phys. Today 2000, 53, 19.
∼40% of the carbon is aromatic. Asphaltenes also possess
heteroatoms (a few percents), mostly nitrogen and sulfur.
The characterization of sulfur6 and nitrogen7 moieties in
asphaltenes has been established using X-ray absorption
near-edge structure. Infrared spectroscopy elucidates the
oxygen moieties.8 The basic ring geometries of the aromatic
ring system have been shown using 13C NMR8 and X-ray
Raman spectroscopy.9 The type of alkane substituent has
been explored via NMR and IR8 and by pyrolysis gas
chromatography.10 The size of the aromatic ring systems
has been determined in numerous ways; scanning tun-
neling microscopy has been used to obtain direct images
of the aromatic portions of asphaltene molecules.11 High-
resolution transmission electron microscopy has also been
used to image aromatic ring systems in asphaltenes.12
Optical absorption and emission spectroscopy13 coupled
with molecular orbital calculations14 has been used to
interrogate this same issue. Finally, time-resolved fluo-
rescence depolarization (FD) studies also constrain the
(6) George, G. N.; Gorbaty, M. L. J. Am. Chem. Soc. 1989, 111, 3182.
(7) Mitra-Kirtley, S.; Mullins, O. C.; Chen, J.; van Elp, J.; George,
S. J.; Cramer, S. P. J. Am. Chem. Soc. 1993, 115, 252.
(8) Scotti, R.; Montanari, L. In ref 4, chapter 3.
(9) Bergmann, U.; Groenzin, H.; Mullins, O. C.; Glatzel, P.; Fetzer,
J.; Cramer, S. P. Chem. Phys. Lett. 2003, 369, 184.
(10) Strausz, O. P.; Mojelsky, T. W.; Lown, E. M. Fuel 1992, 71, 1355.
(11) Zajac, G. W.; Sethi, N. K.; Joseph, J. T. Scanning Microsc. 1994,
8, 463.
(12) Sharma, A.; Groenzin, H.; Tomita, A.; Mullins, O. C. Energy
Fuels 2002, 16, 490.
(13) Mullins, O. C. In ref 4, chapter 2.
(14) Ruiz-Morales, Y. J. Chem. Phys. A 2002, 106, 11283.

10.1021/la048640t CCC: $30.25 © 2005 American Chemical Society

Published on Web 02/25/2005
Asphaltene CNACs and Standard Surfactants CMCs
proposed size of asphaltene ring systems.15-18 Mass
spectroscopy is also providing stringent requirements for
proposed asphaltene structures.19,20 All results from these
disparate methods are in accord, providing a robust
answer. The number of fused rings in asphaltenes varies
from ∼4 to 10.
The molecular weight of asphaltenes was previously
one of the biggest unresolved issues in asphaltene science.
Because of agreement sited above about the basic unit or
subunit of asphaltene molecules, the debate over asphalt-
ene molecular weight essentially has been whether
asphaltenes are monomeric or polymeric. One proposed
asphaltene structure showed a polymer consisting of about
10 monomers,10 although the author of that paper advised
that his molecular weight determination was uncertain.
In fact, the molecular weight of asphaltenes had been the
subject of debate spanning decades, both in time and in
numerical magnitude. This debate is now resolved; FD
studies have shown that asphaltenes are small molecules,
essentially monomeric, not polymeric systems.15-18 Fur-
thermore, by knowing the proper molecular weight, one
can relate the molecular structure to function both for
asphaltene solubility17 and for molecular changes with
feedstock cracking.18 Mass spectroscopy has weighed on
this topic. Initial field ionization mass spectral studies19,20
showed that petroleum asphaltenes are about 700 g/mol
with a width in the mass distribution of about a factor of
2. Subsequently, many mass spectral studies using many
ionization techniques have corroborated these results.21-25
There are, however, some mass spectral reports utilizing
laser desorption that are at variance with all other mass
spectral methods.26
By comparing and contrasting crude oil and petroleum
asphaltenes, FD studies have shown that the key mo-
lecular parameters controlling asphaltene solubility are
intermolecular binding of the aromatic ring systems versus
steric disruption of this binding by alkane substituents.17
Furthermore, a direct correlation exists between solubility,
molecular size, and ring size.15-18,27 Index-of-refraction
studies clearly show the importance of van der Waals
interaction in the binding in asphaltene solubility.28,29
Within this point of view, the Yen model for primary and
secondary aggregation of asphaltenes is reasonable. In
this picture, primary aggregation occurs when aromatic
ring systems of different asphaltene molecules stack more
or less as pancakes to form primary nanoaggregates with
corresponding high binding energy. (We are attempting
to change the terminology per the reviewers’ request.
These nanoaggregates were called asphaltene “micelles”
in many previous reports.) However, alkane steric hin-
(15) Groenzin, H.; Mullins, O. C. J. Phys. Chem. A 1999, 103, 11237.
(16) Groenzin, H.; Mullins, O. C. Energy Fuels 2000, 14, 677.
(17) Buenrostro-Gonzalez, E.; Groenzin, H.; Lira-Galeana, C.; Mul-
lins, O. C. Energy Fuels 2001, 15, 972.
(18) Buch, L.; Groenzin, H.; Buenrostro-Gonzalez, E.; Andersen, S.
I.; Lira-Galeana, C.; Mullins, O. C. Fuel 2003, 82, 1075.
(19) Boduszynski, M. M. In ref 2, chapter 2.
(20) Boduszynski, M. M. Energy Fuels 1988, 2, 597.
(21) Miller, J. T.; Fisher, R. B.; Thiyagarajan, P.; Winans, R. E.;
Hunt, J. E. Energy Fuels 1998, 12, 1290.
(22) Yang, M.-G.; Eser, S. ACS Reprints, 281th ACS National Meeting,
New Orleans, LA,Aug 22-26, 1999; American Chemical Society:
Washington, DC, 1999; p 768.
(23) Qian, K.; Rodgers, R. P.; Hendrickson, C. L.; Emmett, M. R.;
Marshall, A. G. Energy Fuels 2001, 15, 492.
(24) Hughey, C. A.; Rodgers, R. P.; Marshall, A. G. Anal. Chem. 2002,
74, 4145.
(25) Cunico, R. I.; Sheu, E. Y.; Mullins, O. C. Pet. Sci. Technol. 2004,
22, 787.
(26) Suelvas, I.; Islas, C. A.; Millan, M.; Galmes, C.; Carter, J. F.;
Herod, A. A.; Kandiyoti, R. Fuel 2003, 82, 1.
(27) Groenzin, H.; Mullins, O. C.; Eser, S.; Mathews, J.; Yang, M.-G.;
Jones, D. Energy Fuels 2003, 17, 498.
Langmuir, Vol. 21, No. 7, 2005 2729
drance would restrict the number of molecules in the stack.
Further addition of asphaltene to the solution results in
new nanoaggregates but not in the size increase of a single
nanoaggregate. Eventually, at very high concentrations,
the nanoaggregates undergo clustering; the binding energy
of one nanoaggregate to another is much lower than the
binding interior to the nanoaggregate. This picture is
roughly in accord with the temperature dependence of
small-angle neutron scattering (SANS); the smallest
aggregates are always observed independent of concen-
tration.30 However, the smallest particle detected by SANS
(∼5 nm diameter) is larger than one might expect for the
nanoaggregates giving rise to some uncertainty. The huge
viscosity variation with temperature of asphaltic systems
such as roofing tar shows that the disruption of higher
asphaltene aggregation is not that high in binding
energy.31 Particle size analysis for asphaltene flocculation
was produced by solvent addition32 and live oil depres-
surization.33
Using some micellar formalisms for the primary as-
phaltene aggregation is plausible from a molecular
structural point of view. In particular, the concept that
nanoaggregate growth shuts off after reaching a certain
small size due to steric hindrance is consistent with
substantial observations; it needs to be tested. Surface
tension measurements have been employed to measure
the “asphaltene critical micelle concentration (CMC)” in
pyridine.34 A clear break in the surface tension data
occurred at ∼400 mg/L. Other surface tension studies have
reported “CMCs of asphaltene in toluene” to be much
higher in concentration, 10 g/L in one study35 and 1.7 g/L
in another study.36 There are kinetic issues associated
with surface tension measurements that may help explain
the large range of values reported for the asphaltene CMC.
Thermo-optical techniques suggested a possible “CMC”
for asphaltenes in toluene at ∼50 mg/L.37 Microcalorimetry
has also been applied to asphaltenes; reported “asphaltene
CMCs in toluene” are in the range of several grams per
liter.38 A recent report using microcalorimetry illustrates
effects from water on aggregation and labels the observed
transitions at ∼3 g/L asphaltene in toluene as apparent
“CMCs”.39 The authors also state they do not detect a
CMC without water being present.
High-Q, high-resolution ultrasonic spectroscopy is
perhaps the most direct and sensitive method to test the
formation of micelles. The speed of sound is a direct probe
of the bulk and so is not sensitive to surface issues, and
one can easily exclude transients in the measurements.
High-Q ultrasonic measurements have been used suc-
cessfully to monitor many types of phase transitions in
solution.40-46
(28) Buckley, J. Energy Fuels 1999, 13, 328.
(29) Buckley, J. S.; Hirasaki, G. J.; Liu, Y.; Von Drasek, S.; Wang,
J.-X.; Gill, B. S. Pet. Sci. Technol. 1998, 16, 251.
(30) Sheu, E. Y. In ref 3, chapter 1.
(31) Lin, M.-S.; Chaffin, J. M.; Davison, R. R.; Glover, C. J.; Bullin,
J. A. In ref 4, chapter 9.
(32) Anisimov, M. A.; Yudin, I. K.; Nikitin, V.; Nikolaenko, G.;
Chernoustan, A.; Toulhoat, H.; Frot, D.; Briolant, Y. J. Phys. Chem.
1995, 99, 9576.
(33) Joshi, N. B.; Mullins, O. C.; Jamaluddin, A.; Creek, J.; McFadden,
J. Energy Fuels 2001, 15, 979.
(34) Sheu, E. Y. J. Phys.: Condens. Matter 1996, 8, A125.
(35) da Silva Ramos, A. C.; Haraguchi, L.; Nostripe, F. R.; Loh, W.;
Mohamed, R. S. J. Pet. Sci. Eng. 2001, 32, 201.
(36) Bouhadda, Y.; Bendedouch, D.; Sheu, E. Y.; Krallafa, A. Energy
Fuels 2000, 14, 845.
(37) Acevedo, S.; Ranaudo, M. A.; Pereira, J. C.; Castillo, J.;
Fernandez, A.; Perez, P.; Caetano, M. Fuel 1999, 78, 997.
(38) Andersen, S. I.; Christensen, S. D. Energy Fuels 2000, 14, 38.
(39) Andersen, S. I.; del Rio, J. M.; Khvostitchenko, D.; Shakir, S.;
Lira-Galeana, C. Langmuir 2001, 17, 307.
2730 Langmuir, Vol. 21, No. 7, 2005 Andreatta et al.

Here, high-Q ultrasonic measurements are performed The density of the solution is deduced from the previous
on aqueous and toluene solutions containing standard equations (3 and 4):
surfactants and are compared against known literature
values when available to validate the methodology. The F ) F0 + (1 - ṽ1F0)c1 + (1 - ṽmF0)cm (5)
governing equations for the micelle phase equilibrium
model41,42 are given, and all data presented here are where F0 is the density of the solvent, c1 is the weight
interpreted within this framework. The surfactants used concentration of surfactant in the monomeric form, and
included SDS in water, C16TAB in water, Tween 80 in cm is the weight concentration of surfactant in the micellar
water and separately in toluene, and Brij 35 in toluene. form.
(Surfactant names are explained in the Experimental If assumed that the phase equilibrium model is valid
Section.) CMC determinations via ultrasonic spectroscopy here, then
are shown to agree well for known surfactants over a broad
range of CMCs. In particular, surfactants with very small for c < cmc, c1 ) c and cm ) 0
values of CMCs are treated without difficulty using high-Q
ultrasonic measurements. We employ these ultrasonic for c > cmc, c1 ) cmc and cm ) c - cmc
techniques to study several asphaltene-toluene systems
up to concentrations of several grams per liter. In addition, where c is the weight concentration of surfactant in the
with density measurements, the ultrasonic results provide solution and cmc is the numerical value at the CMC. That
a direct measure of monomer and micelle compressibilities. is, at CMC, cm ) 0 but any new increase in surfactant
Thus, asphaltene compressibilities can be compared with concentration corresponds to increasing the cm.
those of standard surfactants. For all of these solutions, For c < cmc
density measurements were made enabling the determi-
nation of micellar or nanoaggregate compressibilities in F ) F0 + (1 - ṽ1F0)c (6)
solution and, in some case, monomer compressibilities in
solution. Comparisons are emphasized between standard
and for c > cmc
surfactants and asphaltenes.

II. Theory F ) F0 + (ṽm - ṽ1)F0cmc + (1 - ṽmF0)c (7)

The ultrasound velocity u in liquid (solution) is related
To obtain the adiabatic compressibility of the solution κ
to the density of the liquid (solution) F and the adiabatic
as a function of the concentrations of surfactants in the
compressibility of the solution κ by the relationship
solution, the density is differentiated with respect to
pressure P at constant entropy.
1
u) (1)
xFκ 1 ∂F
κ) ( )
F ∂P S
(8)
The compressibility of the liquid can be considered as
adiabatic because the compressions and decompressions Differentiating eq 5,
in ultrasonic waves are too fast for heat dissipations.43

| |∂(1 - ṽ1F0)
|
For dilute surfactant solutions, the ultrasonic velocity ∂F ∂F0 ∂c1
and the density of the solution can be expressed as ) + c1 + (1 - ṽ1F0) +
functions of the concentration of surfactants.41 In general, ∂P S ∂P S ∂P S ∂P
surfactant molecules in solution coexist in monomeric and
micellar forms. Here, we follow the treatment given in ref
41. Considering a volume V, if there are w0 g of solvent,
∂(1 - ṽmF0)
∂P S |
cm + (1 - ṽmF0)
∂cm
∂P
(9)

w g of surfactant, w1 g of surfactants in the monomeric If it is considered that the concentration of monomers c1

form, and wm g of surfactants in the micellar form, and the concentration of micelles cm change with pressure
only through the changes in the volume of the solution,
w ) w1 + w m (2) it follows that

Furthermore, ∂c1
) c1κ (10)
∂P
V ) w0v0 + w1ṽ1 + wmṽm (3)
The adiabatic compressibility of the solvent is defined by
where v0 is the specific volume of the solvent, ṽ1 is the
apparent specific volume of the monomeric form, and ṽm ∂cm
is the apparent specific volume of the micellar form. The ) cmκ (11)
∂P
weight of the surfactant solution is equal to

FV ) w0 + w1 + wm (4) κ0 )
1 ∂F0
F0 ∂P( )| S
(12)

(40) Kudryashov, E.; Kapustina, T.; Morrissey, S.; Buckin, V.;

Dawson, K. J. Colloid Interface Sci. 1990, 203, 59. The apparent adiabatic compressibility of the surfactant
(41) Zielinski, R.; Ikeda, S.; Nomura, H.; Kato, S.; J. Colloid Interface in the monomeric form is defined by
Sci. 1987, 119, 398.

( )|
(42) Blandamer, M. J.; Cullis, P. M.; Soldi, L. G.; Engberts, J. B. N.
1 ∂ṽ1
F.; Kacperska, A.; Van Os, N. M.; Subha, M. C. S. Adv. Colloid Interface κ̃1 ) - (13)
Sci. 1995, 58, 171. ṽ1 ∂P S
(43) Buckin, V.; Smyth, C. Semin. Food Anal. 1999, 4, 113.
(44) Mukerjee, P.; Mysels, K. J. NSRDS-NBS 36; U.S. Department
of Commerce: Washington, DC, 1971. And the apparent adiabatic compressibility of the sur-
Asphaltene CNACs and Standard Surfactants CMCs Langmuir, Vol. 21, No. 7, 2005 2731

Figure 1. Chemical structure of Tween 80.

factant in the micellar form is defined by

κ̃m ) - ( )|
1 ∂ṽm
ṽm ∂P S
(14)

Then, from eq 9 and using the definitions 10-14,

Fκ ) F0κ0 + κ[(1 - ṽ1F0)c1 + (1 - ṽmF0)cm] +

F0c1ṽ1(κ̃1 - κ0) + F0cmṽm(κ̃m - κ0) (15)

From eqs 5 and 15, the adiabatic compressibility of the

solution is given by
κ ) κ0 + (κ̃1 - κ0)ṽ1c1 + (κ̃m - κ0)ṽmcm
Equation 1 relates the ultrasonic velocity in liquid to
density and adiabatic compressibility. For dilute solutions
(c1 , 1 and cm , 1),
Figure 2. Ultrasonic spectrum of water at 25 °C.
(16)
from Acros and Sigma-Aldrich. The chemical structure of Tween
80 is shown in Figure 1.
The ultrasonic measurements were performed on the HRUS
102 high-resolution ultrasonic spectrometer from Ultrasonic
Scientific, Ltd. The speed of sound is determined using a

[( ) ]
resonance technique in the range of frequencies between 2 and
u0 κ̃1 20 MHz; we used ∼5 MHz for our experiments, and the
u ) u0 + ṽ1 2 - - v 0 c1 + spectrometer can measure the speed of sound to one part in 5
2 κ0

[( ) ]
× 10-6. The measurements are made using two identical cells
u0 κ̃m filled with a volume from 1 to 2 mL, one filled with the analyzed
ṽm 2 - - v0 cm (17) solution and the other with the solvent (water or toluene). Both
2 κ0
cells are fixed together in the same block and are thermostated
at 25 ( 0.1 °C, enabling small differences in ultrasonic velocity
where v0 is the specific volume of the solvent and is equal to be determined.43 Each cell consists of a resonance cavity with
to the inverse of F0. a resonant glass chamber built with two lithium niobate
For c < cmc, transducers on two opposite sides of the cell; one transducer is

[( ) ]
used as the signal source, and the other is the receiver. Two
u0 κ̃1 main factors determine the resolution of the measurements: the
u ) u0 + ṽ1 2 - - v0 c (18) quality of the resonance (including a high factor of quality Q and
2 κ0 the absence of satellites of the resonance peaks) and the stability
of the resonances. The first factor is ensured by having a high
And for c > cmc, c1 ) c, and cm ) 0 precision in the parallel alignment of the cells and the quality
of the lithium niobate piezotransducers. The second factor

u ) u0 +
u0
2 [( ) (
ṽ1 2 -
κ̃1
κ0
- ṽm 2 -
κ̃m
κ0 )]
cmc +
requires a special construction of the resonator, insisting on
keeping the distance between the two transducers constant. To
maintain a perfect geometry, the resonator chamber has been
u0
2 [(
ṽm 2 -
κ̃m
κ0 ) ]
- v0 c (19)
built separately from the transducers. Of course, frequencies
can be measured to very high accuracies.43 In fact, for our
measurements exquisite precision is needed more than exquisite
accuracy. The ultrahigh-Q resonator coupled with the dual beam
It can be concluded from eqs 18 and 19 that the ultrasonic experimental configuration provides this objective. The conver-
velocity can be modeled by two straight lines in the plot sion of frequency to sound speed is performed using the equation
of ultrasonic velocity versus concentration; one line
segment below the CMC, the other above.41,44-46 The δu δfn
) (20)
apparent compressibilities of the monomeric form and the u fn
micellar form can be deduced from ultrasonic and density
measurements.41,42,44-46 where u is the speed of sound and fn is the frequency of the nth
resonance. This equation presumes plane wave propagation.43,48
III. Experimental Section A typical ultrasonic spectrum of water is shown in Figure 2.
Several of the sharp ultrasonic resonances are shown in Figure
The different surfactants used here are sodium dodecyl sulfate 3 at the frequency range used for this study. Figures 4 and 5
(SDS) from Sigma-Aldrich Chemicals, purity > 99%, hexade- show the comparable spectra for toluene. In these figures, the
cyltrimethylammonium bromide (C16TAB) from Sigma-Aldrich, amplitude of the output signal is shown as a function of the
purity > 99%, polyoxyethylene 23 lauryl ether (Brij 35; C12E23) frequency of the acoustical signal. In addition to the narrow
from Acros Chemicals, purity 99%, and polyoxyethylene sorbitan acoustic cell resonances, Figures 2 and 4 show broad resonances
monooleate (Tween 80) from Acros Chemicals, purity > 99%. (at ∼4, 7, and 10 MHz, for example) that are associated with
N-heptane asphaltenes UG8 and BG5 from Kuwaiti crude oils
were used; the extraction procedure is described elsewhere.17 (45) Bloor, D. M.; Gormally, J.; Wyn-Jones, E. J. Chem. Soc., Faraday
The aqueous solutions were made in Milli-Q water for the density Trans. 1 1984, 80, 1915.
measurements and in distilled water for the ultrasound mea- (46) Suárez, M. J.; López-Fontán, J. L.; Sarmiento, F.; Mosquera, V.
surements. The organic solutions were prepared in toluene 99.8% Langmuir 1999, 15, 5265.
2732 Langmuir, Vol. 21, No. 7, 2005
Figure 3. Several acoustic cell resonances in the ultrasonic
spectrum of water at 25 °C.
Figure 4. Ultrasonic spectrum of toluene at 25 °C.
Figure 5. Several acoustic cell resonances in the ultrasonic
spectrum of toluene at 25 °C.
ultrasonic resonances in the glass walls of the cells and with the
transducers. These spectral regions were avoided in all of our
experiments. For distilled water at 25 °C, the speed of sound is
u0 ) 1496.7 m/s with a resolution of 0.0075 m/s. In toluene at
25 °C, the speed of sound is u0 ) 1307.1 m/s with a resolution
of 0.0065 m/s.
All ultrasonic spectra were acquired by diluting solutions from
the highest concentrations, with stepwise concentration reduc-
tions. Each curve consisted of approximately at least 15-25
points. For each point, a quantitative dilution was performed;
the solution was stirred and allowed to equilibrate for 10-15
min prior to recording the ultrasonic frequency for that con-
centration. Integrated runs were typically 4-6 h. No difference
in the spectrum was observed if the equilibration time was
increased or decreased by a factor of 2. The reproducibility of the
measurements was checked and found to be very good. All the
ultrasonic titrations exhibit a break between two straight-line
segments in the curves, as expected from eqs 18 and 19. The
CMCs of the different surfactants were given by the intersection
of two straight-line segments.
The densities were measured at 25 ( 0.01 °C with an Anton
Paar DMA 4500 densitometer with a resolution of 5 × 10-5 g/cm3.
Each measurement was done twice and averaged. The densities
of milli-Q water and toluene were found to be close to the expected
Andreatta et al.
Figure 6. (a) Ultrasonic titration of SDS in aqueous solution
at 25 °C. (b) Density measurements of SDS in aqueous solution
at 25 °C.
Table 1. Speeds of Sound, Densities, and Adiabatic
Compressibilities at the Temperature of 25 °C of the
Solvents Used in This Study
κ0 adiabatic
speed of sound density v0 compressibility
solvent (m/s) (g/cm3) (cm3/g) (10-5/bar)
water 1496.7 0.997 04 1.002 97 4.48
toluene 1307.1 0.862 22 1.1598 6.79
values: 0.997 04 and 0.862 22 g/cm3, respectively. The apparent
specific volumes of the monomer and of the micelle of SDS, C16-
TAB, and Tween 80 in water and Tween 80 and Brij 35 in toluene
were calculated using eq 7, using a straight-line segment of the
points at concentrations above the CMC, which are more reliable
than the points at concentrations lower than the CMC because
of the resolution of the densitometer. For Tween 80 in water and
asphaltene in toluene, the apparent specific volume of the micelle
was determined from the slope of the density versus concentration
graph but the apparent specific volumes of monomer have not
been calculated because of insufficient measurement resolution
at the very low concentrations.
Table 1 presents the ultrasonic speed of sound and densities
measured for our two solvents, water and toluene. The derived
compressibility from eq 1 is also given.
IV. Results and Discussion:
Ionic Surfactants. Two well-known ionic surfactants
were studied here in aqueous solutions: SDS (anionic)
and C16TAB (cationic). The CMCs were deduced from the
ultrasonic velocity versus concentration plots, were taken
as the intersections of the two straight-line segments
(above and below the CMC), and were found to be close
to the CMC given in the literature.40,44,49 Figure 6a gives
the solution ultrasonic velocity versus SDS concentration,
(47) De Lisi, R.; Milioto, S.; Verrall, R. E. J. Solution Chem. 1990,
19 (7), 665.
Asphaltene CNACs and Standard Surfactants CMCs Langmuir, Vol. 21, No. 7, 2005 2733

Figure 7. (a) Ultrasonic titration of C16TAB in aqueous solution Figure 8. (a) Ultrasonic titration of Tween 80 in water at 25
at 25 °C. (b) Density measurements of C16TAB in aqueous °C. (b) Density measurements of Tween 80 in water at 25 °C.
solution at 25 °C.

Table 2. Values of the CMCs at 25 °C of the Ionic The results are in excellent agreement with the
Surfactants Used in This Study references.40,45-47,49 κ̃1 is large and negative for the ionic
surfactant CMC (this work) (g/L) CMC [refs] (g/L)
surfactants SDS and C16TAB in water40,41,46,47 while κ̃m is
large and positive.40,41,45-47 This high molar compressibility
2.393 [44] for the surfactants in the micellar state can be associated
SDS 2.573
2.408 [49]
with the compressibility of the internal core of micelles
0.328 [44]
C16TAB 0.335
0.334 [40] and is close to the adiabatic compressibility of pure
hydrocarbon liquids with the same length of the hydro-
Table 3. Apparent Specific Volumes and carbon chain.40
Compressibilities of the Monomeric and Micellar Forms Nonionic Surfactants in Aqueous and Organic
of the Ionic Surfactants Used in This Study at a Solvents. Tween 80 (see Figure 1) is a nonionic am-
Temperature of 25 °C
phiphile composed of 20 oxyethylene groups on an
ṽ1 ṽm κ̃1 κ̃m oxocyclopentane core which are the hydrophilic part of
surfactant (cm3/g) (cm3/g) (10-5/bar) (10-5/bar)
the molecule, while the hydrocarbon chain is the hydro-
SDS (this work) 0.822 0.873 -1.97 4.02 phobic part of the molecule. It was studied here both in
SDS [refs] 0.813 [50] 0.854 [50] -1.93 [46] 4 [46] water and in toluene. Brij 35 is a polyoxyethylene dodecyl
4.3 [45]
C16TAB (this work) 0.864 0.983 -1.64 4.33
ether, which is a nonionic surfactant with a C12 hydro-
C16TAB [refs] 0.962 [40] 0.989 [40] -0.039 [40] 4.28 [40] phobic alkyl chain and a hydrophilic chain of 23 poly-
0.964 [47] 0.988 [47] -1.25 [47] 4.24 [47] oxyethylene subunits. It was studied in toluene.
1.002 [50] Figure 8a shows the solution ultrasonic velocity of
Tween 80, and Figure 8b shows the corresponding density
and Figure 6b gives the solution density versus SDS curve. At the very low concentration of 8 mg/L, Tween 80
concentration. Figure 7a gives the solution ultrasonic exhibits a clear change in the ultrasonic velocity curve.
velocity versus C16TAB concentration, and Figure 7b gives This agrees with the literature determination of the
the solution density versus C16TAB concentration. Table CMC.51 Thus, the ultrasonic method for CMC determi-
2 presents the CMCs obtained here at 25 °C for SDS and nation is established over 2.5 orders of magnitude in
C16TAB and also gives literature values for these CMCs. concentration. Two nonionic surfactants were run in
Note the excellent agreement. toluene; Figure 9a shows for Tween 80 the solution
The density measurements provide apparent specific ultrasonic velocity versus Tween 80 concentration; Figure
volumes; these combined with the ultrasonic measure- 9b shows the corresponding density curve. Figure 10a
ments allow us to derive the apparent adiabatic com- shows for Brij 35 the solution ultrasonic velocity versus
pressibilities of the monomer and micelle. Our apparent
specific volumes and derived compressibilities for SDS (48) Mason, W. P.; Thurston, R. N. Physical Acoustic. In High-
and C16TAB are given in Table 3 and are compared with Frequency Continuous Wave Ultrasonics; Bolef, D. I., Miller, J. G., Eds.;
Academic Press: New York, 1971; Vol. VIII, chapter 3.
literature values. Note the generally excellent agreement (49) Priev, A.; Zalipsky, S.; Cohen, R.; Barenholz, Y. Langmuir 2002,
for all parameters. 18, 612.
2734 Langmuir, Vol. 21, No. 7, 2005
Figure 9. (a) Ultrasonic titration of Tween 80 in toluene at
25 °C. (b) Density measurements of Tween 80 in toluene at 25
°C.
Figure 10. (a) Ultrasonic titration of Brij 35 in toluene at 25
°C. (b) Density measurements of Brij 35 in toluene at 25 °C.
Brij 35 concentration; Figure 10b shows the corresponding
density curve. For these nonionic surfactants in toluene,
it is not surprising that there is not a clear break in the
Andreatta et al.
Table 4. CMCs for Nonionic Surfactants in Aqueous and
Organic Solutions at 25 °C
surfactant solvent CMC (our work; g/L) CMC [refs] (g/L)
Tween 80 water 0.008 0.013 [51]
Tween 80 toluene 7.4
Brij 35 toluene 17
Table 5. Apparent Specific Volumes and
Compressibilities of the Monomeric and Micellar Forms
of the Nonionic Surfactants in Toluene and Water at 25
°C Used in This Study
ṽ1 ṽm κ̃1 κ̃m
surfactant (cm3/g) (cm3/g) (10-5/bar) (10-5/bar)
Tween 80 in water 0.9227 1.37
Tween 80 in toluene 0.8895 0.9162 4.38 3.52
Brij 35 in toluene 0.9269 0.9309 5 3.94
velocity curves (even though we fit sections of the curve
with straight lines.) We interpret the changes in slopes
in the ultrasonic velocity curves as effective CMCs for
Figures 9a and 10a. Nonionic surfactants in toluene do
not exhibit a sharp limit to the size of the micelles; this
leads to curvature of the ultrasonic plots (Figures 9a and
10a). Table 4 lists the CMCs determined here for standard
nonionic surfactants. Using measured densities, apparent
specific volumes and apparent compressibilities are
obtained and are listed in Table 5.
Comparison of the ultrasonic curves for ionic surfactants
in water versus nonionic surfactants in water shows very
different behavior. The differential quantity, the com-
pressibility, is much more sensitive and, thus, accounts
for this change much more than the integral quantity, the
density. The ionic surfactants in water have much different
apparent compressibilities than the nonionic surfactants
in toluene. In particular, the ionic surfactants exhibit
negative apparent compressibilities (in water) for the
monomeric form with a very large increase in apparent
compressibilities upon formation of the micelle. These
micelles have a nonpolar core which is anticipated to be
rather compressible. On the other hand, the nonionic
surfactants in toluene exhibit very large positive apparent
compressibilities in toluene for the monomeric form and
show a reduction in apparent compressibilities upon
micelle formation. The core for these micelles is polar and
is anticipated to be more rigid. Simple heuristics account
for these systematics and are useful for comparison with
asphaltene results.
Asphaltenes. UG8 Asphaltenes. UG8 Asphaltenes have
been used by essentially every technique we have of
analyzing asphaltenes. Their properties are common; thus,
they represent “typical” asphaltenes. Figure 11a shows
the solution ultrasonic velocity versus concentration for
UG8 asphaltenes in toluene. A clear break in this curve
is observed; this gives a “critical nanoaggregate concen-
tration” (CNAC) for asphaltenes of 0.164 g/L. The velocity
versus concentration curve for asphaltene is very similar
to that of other nonionic surfactants in toluene, strength-
ening the CMC interpretation. Furthermore, at concen-
trations higher than the CMC, there is not another break
in the curve even up to concentrations of 2 g/cm3 asphaltene
in toluene. Either there is no other change in aggregates
up to this concentration or any further change in ag-
gregation has no effect on ultrasonic velocity (because the
binding energy is too low to change the compressibility).
From the density measurements (Figure 11b), we can
calculate the apparent specific volume of nanoaggregate
from the slope of the density versus concentration graph
but we cannot get the apparent specific volume of monomer
because the concentrations are too low for the accuracy
Asphaltene CNACs and Standard Surfactants CMCs
Figure 11. (a) Ultrasonic titration of asphaltenes UG8 in
toluene at 25 °C. (b) Density measurements of asphaltenes UG8
in toluene at 25 °C.
Table 6. Values of the CMC, Apparent Specific Volumes
of the Nanoaggregates, and Apparent Adiabatic
Compressibilities of the Nanoaggregates for Different
Asphaltenes in Toluene at 25 °C
asphaltenes CMC (g/L) ṽm (cm3/g) κ̃m (10-5/bar)
UG8 0.164 0.8814 3.95
BG5 0.048 0.8582 3.45
of the densitometer. The adiabatic apparent compress-
ibility in the nanoaggregate form is then calculated from
the ultrasonic data with eq 19. The results are presented
in Table 6. The apparent compressibility of the asphaltene
nanoaggregate is close in magnitude to that of nonionic
surfactants, again lending credence to the CNAC inter-
pretation for the asphaltenes.
BG5 Asphaltenes. BG5 Asphaltenes were obtained from
Kuwait Burgan5 crude oil. They too have been subjected
to many different kinds of investigation, and they too have
“typical” characteristics. Asphaltenes from UG8 and BG5
might be called “vanilla” asphaltenes. Figure 12a shows
the solution ultrasonic velocity versus concentration for
BG5 asphaltenes. A break in this curve is evident; this
gives a CNAC of 0.048 g/L. The CNAC for BG5 is lower
than that of UG8 asphaltenes; nevertheless, a CNAC is
evident in both cases. Again, no other change is observed
in the ultrasonic velocity up to 3 g/cm3. As in the case of
UG8 asphaltenes, we can calculate the apparent specific
volume of the nanoaggregate from the slope of the solution
density versus concentration (Figure 12b), but we cannot
get the apparent specific volume of monomer because the
concentration is too low. From the ultrasonic data, we can
calculate the apparent adiabatic compressibility in the
micellar form. The results are presented in Table 6. Again,
we get agreement between asphaltene micelle apparent
compressibilities with those of other nonionic surfactants.
Langmuir, Vol. 21, No. 7, 2005 2735
Figure 12. (a) Ultrasonic titration of asphaltenes BG5 in
toluene at 25 °C. (b) Density measurements of asphaltenes BG5
in toluene at 25 °C.
Figure 13 shows an expanded scale of the ultrasonic
velocity versus concentration to make the CNAC clear.
Figure 13a expands the low concentration range of Figure
11a while Figure 13b expands the low concentration range
of Figure 12a. The break in the ultrasonic velocity curve
is quite clear; we interpret this break to be the CNAC. At
higher concentrations than the CNAC, there is no change
in the ultrasonic slope. This indicates that the nanoag-
gregates are not changing at these concentrations, just
that there are more of them at higher concentration. That
is, nanoaggregate growth shuts off. We also note that while
a clear break is evident at the CNAC, the ultrasonic data
cannot rule out formation of dimers or trimers at
concentrations below the CNAC.
Different asphaltenes exhibit CNACs at similar con-
centrations but with some variability in the exact value:
CNACs ∼ 50-150 mg/L. The apparent compressibilities
of the asphaltene nanoaggreagtes are similar to each other
and similar to other apparent micelle compressibilities
for other nonionic surfactants in toluene. The CNACs
determined here are 1-2 orders of magnitude lower than
those in literature reports for asphaltene-toluene systems
obtained by other techniques. In our view, the other
techniques are not recording proper CNACs. They may
be recording some higher-level aggregation phenomenon.
Many techniques do not measure an explicit parameter
such as apparent nanoaggregate compressibility that can
then be checked against known surfactants as we do.
Rather, some of the other techniques only interpret a
change in some property as the CNAC. If these techniques
are not sufficiently sensitive to detect CNACs at 100 mg/
L, then corresponding data will be subject to misinter-
pretation.
The governing chemical principles of asphaltenes that
determine solubility and, thus, define asphaltenes have
been shown to be van der Waals attraction of aromatic
2736 Langmuir, Vol. 21, No. 7, 2005
Figure 13. (a) Ultrasonic titration of asphaltenes UG8 in
toluene at 25 °C. (b) Ultrasonic titration of asphaltenes BG5
in toluene at 25 °C.
ring systems versus steric repulsion of alkane chains;17
essentially asphaltene molecules are shaped “like your
hand” with the ring systems being the palm and the alkane
chains being the fingers. We believe that the same forces
are operative in determining the asphaltene nanoaggre-
gate. The idea is that the first few asphaltene molecules
to associate have fairly clear access to intermolecular
interaction of the fused ring system. However, subsequent
aggregation of more molecules becomes constrained by
alkane substituents, thereby impeding further stacking.
At some point, the interaction between the nanoaggregate
and an additional molecule becomes rather weak, probably
as a result of steric repulsion. At this point new nano-
aggregates form with increasing concentration. These
ideas are central to the Yen model; our data is in general
in agreement with this well-known model.52
The low values of the CNACs help explain why
colligative techniques such as vapor pressure osmometry
(VPO) always record “molecular” weights that are too high.
Andreatta et al.
VPO and other colligative methods are performed at
concentrations that are significantly in excess of the
asphaltene CNAC; consequently, VPO provides an ag-
gregate weight. VPO is often in error by a factor of ∼5 for
molecular weight determination; consequently, VPO along
with our asphaltene CNAC results here show that the
aggregation number in an asphaltene micelle is ∼5.
It is important to note that for nonionic surfactants and
asphaltenes in organic solvents trace water may influence
results significantly.39 In this report we have not attempted
to dry the toluene, so our results may be influenced by
trace water. First, we note that most previous determina-
tions of the asphaltene CMC by various techniques also
did not dry the toluene. To obtain a direct comparison
with our work and previous work, we have used similar
methodology. Our CNACs differ substantially from these
previous studies, as has been noted. Second, in natural
systems, water is frequently present; we are interested in
performing experiments on a system simple enough to
hopefully understand while, at the same time, providing
some indication of what is to be expected in natural
systems. In the future, it will be important for us to assess
the effect of dissolved water on our data.
V. Conclusions
High-Q ultrasonic spectroscopy has proven to be a very
valuable tool in the characterization of micelle formation
for known surfactants and of nanoaggregate formation of
asphaltenes. CMCs of ionic and nonionic surfactants are
easily measured in high and low concentration ranges.
For standard surfactants, we obtain excellent agreement
between our measurements and literature values of CMCs,
apparent specific volumes, and apparent compressibilities.
Asphaltenes in toluene exhibit CNACs at ∼100 mg/L. The
phase equilibrium model for micelles applies readily to
all data presented here, surfactant and asphaltene data
alike. Furthermore, derived parameters such as the
compressibility of nanoaggregates of asphaltenes and of
micelles of nonionic surfactants are very similar, thereby
strengthening the conclusion that asphaltenes are non-
ionic surfactants that form nanoaggregates and exhibit
CNACs. The literature reports with much larger concen-
trations for asphaltene CNACs (of CMCs) are probably
not measuring CNACs but perhaps some higher order
aggregation. Our asphaltene CNACs explain why VPO
measurements of asphaltene molecular weights are
consistently too high.
LA048640T
(50) Corkill, J. M.; Goodman, J. F.; Walker, T. Trans. Chem. Soc.
1967, 63, 768.
(51) www.sigma-aldrich.com (accessed 2004).
(52) Yen, T. F. Prepr. Pap.sAm. Chem. Soc., Div. Pet. Chem. 1990,
35, 314.