Langmuir 2006, 22, 575-581 575

Molecular Assembly of Metallacarboranes in Water: Light Scattering

and Microscopy Study
Pavel Matějı́ček,*,† Petr Cı́gler,‡,§ Karel Procházka,† and Vladimı́r Král‡
Department of Physical and Macromolecular Chemistry, Charles UniVersity in Prague,
HlaVoVa 2030, 128 40 Prague 2, Department of Analytical Chemistry, Institute of Chemical Technology,
Technicka 5, 166 28 Prague 6, and Institute of Organic Chemistry and Biochemistry, Academy of Science
of the Czech Republic, FlemingoVo n.2, 166 10 Praha 6, Czech Republic

ReceiVed August 12, 2005. In Final Form: October 27, 2005

Polyhedral metallacarboranes are used mainly as ion-pairing agents and recently have been recognized as potent
inhibitors of HIV protease. They are characterized by exceptional hydrophobicity, rigid geometry, delocalized negative
charge, ion-pairing behavior, and strong acidity of their conjugated acids. The completely novel phenomenon, association
of these promising pharmaceutical tectons in aqueous solutions, is described here. The behavior of two structural types
of metallacarboranes, [bis(1,2-dicarbollide)cobaltate(1-)] and bis[(3)-1,2-dicarbollylcobalt]-(3,6)-1,2-dicarbacanastide-
(2-)], in aqueous solution was studied by a combination of static and dynamic light scattering and microscopy
methods. Spherical aggregates with radii of ca. 100 nm and fairly monodisperse nanostructures were found in aqueous
solutions. The behavior of nanoaggregates is fairly complex and depends on the concentration and aging of the
solutions. The particles are stabilized in the solution by counterions. The formation of larger clusters upon dilution
of bis(1,2-dicarbollide)cobaltate(1-) solutions was observed. The secondary aggregation can be suppressed by addition
of NaCl. Gel permeation chromatography measurements of sodium bis(1,2-dicarbollide)cobaltate(1-) show that the
majority of matallacarborane molecules form nanoaggregates and only a small amount of the metallacarborane remains
molecularly soluble or forms small oligomers.

Introduction truly “soluble” in water. This fanciful group of inorganic

A large group of pharmacologically active compounds could
be characterized by a common atributeseither their hydrophobic
or their amphiphilic nature. As a result, these compounds tend
to form different assemblies (aggregates) in aqueous solutions1,2
which could influence significantly their biological activity. The
incorporation of a drug into an aggregate affects its physico-
chemical properties, pharmacokinetics, and interaction with
biological membranes.3 Moreover, in contrast to single molecules,
the aggregates of several small compounds were shown as unique
nonspecific and promiscuous inhibitors of enzymes.4-6 The
biological activity was determined in this case mainly by the size
of the aggregate and its hydrophobicity. Such nonspecifity is a
key problem in high-throughput screening of large libraries of
chemical compounds in early-stage drug discovery.7
The recent finding of strong anti-HIV activity of polyhedral
metallacarboranes8 led us to question if these potential drugs are
* To whom correspondence should be addressed. Phone: +420221951300.
Fax: +420224919752. E-mail: matej@vivien.natur.cuni.cz.
† Charles University in Prague.
‡ Institute of Chemical Technology.
§ Academy of Science of the Czech Republic.
(1) Atwood, D.; Florence, A. T. Surfactant Systems: Their Chemistry, Pharmacy
and Biology; Chapman and Hall: New York, 1983; Chapter 4.
(2) Attwood, D. AdV. Colloid Interface Sci. 1995, 55, 271-303.
(3) Schreier, S.; Malheiros, S. V. P.; de Paula, E. Biochim. Biophys. Acta 2000,
1508, 210-234.
(4) McGovern, S. L.; Helfand, B. T.; Feng, B.; Shoichet, B. K. J. Med. Chem.
2003, 46, 4265-4272.
(5) Seidler, J.; McGovern, S. L.; Doman, T. N.; Shoichet, B. K. J. Med. Chem.
2003, 46, 4477-4486.
(6) McGovern, S. L.; Caselli, E.; Grigorieff, N.; Shoichet, B. K. J. Med. Chem.
2002, 45, 1712-1722.
(7) Feng, B. Y.; Shelat, A.; Doman, T. N.; Guy, R. K.; Shoichet, B. K. Nat.
Chem. Biol. 2005, 1, 146-148.
(8) Cigler, P.; Kozisek, M.; Rezacova, P.; Brynda, J.; Otwinowski, Z.; Pokorna,
J.; Plesek, J.; Gruener, B.; Doleckova, L.; Masa, M.; Sedlacek, J.; Bodem, J.;
Kraeusslich, H.-G.; Kral, V.; Konvalinka, J. Proc. Natl. Acad. Sci. U.S.A. 2005,
102, 15394-15399.
compounds was discovered by Hawthorne in the 1960s,9-11 and
their synthetic chemistry has been explored extensively.12,13
Metallacarboranes are based on a three-dimensional skeleton
built up from triangular B-H and C-H facets and a central
metal atom(s) that is (are) sandwiched by polyhedral subclusters.
They are characterized by exceptional hydrophobicity, delocalized
negative charge, rigid geometry, low nucleophilicity, strong
acidity of their conjugated acids, and remarkable thermal and
chemical stability.12,14
The above-mentioned physicochemical properties are reflected
in the ion-pairing behavior of metallacarboranes, which in turn
led to their known applications in extraction chemistry.14-16 In
several works concerning ion exchange reactions, the surface
activity of the [1]- anion17,18 (see Chart 1) and the hexachlorinated
derivative of the [1]- anion19 at the interface of two immiscible
liquids was reported. Thus, the tendency of metallacarboranes
to associate in aqueous solutions could be easily expected, even
when they lack an amphiphilic topology. We can learn from the
solution properties of a closely related group of carboranes without
a coordinated metal ion, 1,2-dicarba-closo-dodecaboranes de-
(9) Francis, J. N.; Hawthorne, M. F. J. Am. Chem. Soc. 1968, 90, 1663-1664.
(10) Hawthorne, M. F.; Francis, J. N. Inorg. Chem. 1971, 10, 863-864.
(11) Hawthorne, M. F.; Young, D. C.; Wegner, P. A. J. Am. Chem. Soc. 1965,
87, 1818-1819.
(12) Sivaev, I. B.; Bregadze, V. I. Collect. Czech. Chem. Commun. 1999, 64,
783-805.
(13) Bregadze, V. I.; Timofeev, S. V.; Sivaev, I. B.; Lobanova, I. A. Russ.
Chem. ReV. 2004, 73, 433-453.
(14) Plesek, J. Chem. ReV. (Washington, D.C.) 1992, 92, 269-278.
(15) Rais, J.; Gruner, B. In SolVent extraction; Marcus, I., SenGupta, A. K.,
Eds.; Marcel Dekker: New York, 2005; p 243.
(16) Makrlik, E.; Rais, J.; Base, K.; Plesek, J.; Vanura, P. J. Radioanal. Nucl.
Chem. 1995, 198, 359-365.
(17) Popov, A.; Borisova, T. J. Colloid Interface Sci. 2001, 236, 20-27.
(18) Davion Van Mau, N.; Issaurat, B.; Amblard, G. J. Colloid Interface Sci.
1984, 101, 1-9.
(19) Chaumont, A.; Galand, N.; Schurhammer, R.; Vayssiere, P.; Wipff, G.
Russ. Chem. Bull. 2004, 53, 1459-1465.

10.1021/la052201s CCC: $33.50 © 2006 American Chemical Society

Published on Web 12/16/2005
576 Langmuir, Vol. 22, No. 2, 2006
Chart 1. Structures of Investigated Sandwich-Type
Polyhedral Metallacarboranes Characterized by Peculiar
Peanut-like ([1]-) and Banana-like ([2]2-) Shapes
rivatized with two glucosyl units, that they behave as amphiphilic
compounds, forming aggregates in aqueous solution.20 The
aggregation behavior was reported for other hydrophobic
spherically shaped molecules, e.g., fullerene derivatives;21,22
however, no studies were done on negatively charged polyhedral
metallacarboranes.
We have focused our research on the formation and behavior
of metallacarborane colloidal particles in an aqueous solution of
two “very water soluble” metallacarborane inhibitors of HIV
protease, bis(1,2-dicarbollide)cobaltate(1-) and bis[(3)-1,2-
dicarbollylcobalt]-(3,6)-1,2-dicarbacanastide(2-) ions, [1]- and
[2]2-, respectively (see Chart 1), which exhibit surface activity
as mentioned above. We investigate their size, shape, and behavior
by a combination of light scattering and microscopy techniques
complemented by conductometry and gel permeation chroma-
tography. To our best knowledge, this is the first report on the
molecular assembly of polyhedral metallacarboranes in solution.
The presented study is a part of a more complex project concerning
the physical-chemical behavior of metallacarboranes in aqueous
solution and their interactions with some important proteins such
as HIV protease or serum albumins (papers in preparation and
already published8).
Experimental Section
Chemicals and Solutions. Cesium salts of metallacarboranes
were a kind gift of Dr. Bohumı́r Grüner and Dr. Jaromı́r Plešek
(Institute of Inorganic Chemistry, Academy of Science of the Czech
Republic, Řež near Prague). They were characterized using mass
spectrometry and 1H and 11B NMR spectroscopy.
Cs[1] (MW 456.65): MS m/z ) 323.7 (100) [M]- (calcd 323.740);
1H{11B} NMR (acetone-d ) δ 3.93 (CH1,2), 3.37 (H8), 2.97 (H10),
6 the incident light. Hydrodynamic radii, RH, were evaluated from the
2.69 (H4,7), 1.92 (H9,12), 1.63 (H6), 1.56 (H5,11); 11B NMR
(acetone-d6) δ 6.4 (d, B8) [142.4], 1.2 (d, B10) [138.5], -5.5 (d,
B9,12) [overlap], -6.2 (d, B4,7) [overlap], -17.4 (d, B5,11) [154.0],
-22.9 (d, B6) [165.5].
Cs2[2] (MW 769.08): MS m/z ) 251.6 (100) [M]2- (calcd
251.635); 1H{11B} NMR (acetone-d6) δ (for atom numbering, see
Chart 1): 4.56 (CH1,2), 4.45 (H8,10), 3.42 (CH1′,2′,1′′,2′′), 3.10
(H4,5,7,11), 3.06 (H8′,8′′), 2.66 (H10′,10′′), 2.42, 1.74, 1.69 (H9,-
12, 4′,7′,9′,12′,4′′,7′′,9′′,12′′) (overlap), 1.52 (H6′,6′′), 1.41 (H5′,-
11′,5′′,11′′); 11B NMR (acetone-d6) δ 21.4 (d, B8,10) [143.5], 4.3
(d, B8′,8′′) [143.4], 0.0 (d, B4,5,7,11) [149.5], -3.6 (d, B10′,10′′)
[140.4], -7.9 (B9,12, 4′,7′,9′,12′,4′′,7′′,9′′,12′′) [overlap], -19.4
(d, B5′,11′,5′′,11′′) [155.7], -24.2 (d, B6′,6′′) [167.9].
Cesium salts were converted to sodium salts using the extraction
procedure described in ref 23. The aqueous solution of the cesium
salt acidified by sulfuric acid was extracted two times by diethyl
(20) Morandi, S.; Ristori, S.; Berti, D.; Panza, L.; Becciolini, A.; Martini, G.
Biochim. Biophys. Acta 2004, 1664, 53-63.
(21) Nakamura, E.; Isobe, H. Acc. Chem. Res. 2003, 36, 807-815.
(22) Tsao, C. S.; Lin, T. L.; Jeng, U. S. J. Phys. Chem. Solids 1999, 60,
1351-1353.
Matějı́ček et al.
ether. The organic layer was shaken with a 10% aqueous solution
of sodium carbonate, reextracted by water, and dried using anhydrous
sodium sulfate. The solvent was evaporated, and solid residues were
dried in a vacuum. The sodium salts of metallacarboranes were
obtained in nearly quantitative yields.
All metallacarborane solutions were prepared by dilution of a
stock solution (c ) 10 g/L for Na[1] and Na2[2] and c ) 0.25 g/L
for Cs[1]) with deionized and filtered water.
Techniques. NMR Spectroscopy and Mass Spectrometry. 1H
(160.380 MHz) and 11B (500 MHz) NMR spectroscopy measurements
were performed on a Varian Unity-500 instrument equipped with
the appropriate decoupling accessories. All NMR spectra were
recorded from CD3COCD3 solutions at 25 °C. Chemical shift values
for 11B NMR spectra were referenced to external F3B‚OEt2, and
those for 1H{11B} NMR spectra were referenced to Si(CH3)4.
Chemical shifts δ are reported in units of parts per million downfield
from the reference. All coupling constants are reported in hertz. The
NMR data are presented in the following format: (11B NMR) 11B
chemical shifts δ(11B) (ppm), J(11B-1H) in parentheses; (1H{11B}
NMR) selectively decoupled δ(1H)-{11B selective} chemical shifts.
Mass spectrometry measurements were performed on a Bruker
Esquire 3000 instrument using electrospray ionization. Negative
ions were detected. Samples dissolved in acetonitrile (concentrations
approximately 100 ng/mL) were introduced to the ion source by
infusion of 0.25 mL/h, the drying temperature was 300 °C, the drying
gas flow 4 L min-1, and the nebulizing gas pressure 8 psi.
Dynamic Light Scattering (DLS). The light scattering setup (ALV,
Langen, Germany) consisted of a 633 nm He-Ne laser, an ALV
CGS/8F goniometer, an ALV high QE APD detector, and an ALV
5000/EPP multibit, multi-τ autocorrelator. The solutions of the highest
concentration were filtered through 0.45 µm Acrodisc filters, and
those of lower concentrations were prepared by mixing with water
filtered through 0.20 µm Acrodisc filters. The measurements were
carried out for different concentrations and different angles at 20 °C.
The data analysis was performed by fitting the measured normalized
autocorrelation function g2(t) ) 1 + β|g1(t)|2, where g1(t) is the
electrical field correlation function, t is the lag time, and β is a factor
accounting for deviation from the ideal correlation. An inverse
Laplace transform of g1(t) with the aid of a constrained regularization
algorithm (CONTIN) provides the distribution of relaxation times,
τA(τ):24
g1(t) ) ∫τA(τ) exp(-t/τ) d ln τ (1)
Diffusion coefficients were calculated from individual diffusion
modes as D ) Γ/q2, where Γ ) 1/τ and q ) (4πn0/λ) sin(θ/2) is the
magnitude of the scattering vector. Here θ is the scattering angle,
n0 the refractive index of the pure solvent, and λ the wavelength of
effective diffusion coefficients using the Stokes-Einstein formula
RHeff ) kT/(6pDeffh), where for Deff we can write25
Deff ) D(1 + CRg2q2)(1 + kDc + ...) (2)
where D is the angle- and concentration-independent diffusion
coefficient, Rg is the radius of gyration, c is the sample concentration,
C is a constant depending on the polydispersity and the slowest
internal mode of motion in the scattering particle, and kD is the
diffusion virial coefficient. Therefore, to obtain a true RH, the apparent
hydrodynamic radius has to be extrapolated to the zero scattering
angle.
Static Light Scattering (SLS). The measurements were performed
on the same instrument, which was employed for DLS measurements.
(23) Plesek, J.; Base, K.; Mares, F.; Hanousek, F.; Stibr, B.; Hermanek, S.
Collect. Czech. Chem. Commun. 1984, 49, 2776-2789.
(24) Berne, B. J.; Pecora, R. Dynamic Ligtht Scattering; J. Wiley-Interscience:
New York, 1976.
(25) Burchard, W.; Schmidt, M.; Stockmayer, W. H. Macromolecules 1980,
13, 1265-1272.
Molecular Assembly of Metallacarboranes in Water
Data were treated by the standard method using the Zimm equation26
Kc
)
1 1
Rcor(q,c) p(q,c) Mw (
+ 2A2c + ... ) (3)
where K ) 4π2n2(dn/dc)2/λ4NA is a constant containing the refractive
index n of the solvent, refractive index increment of the colloidal
particles with respect to the solvent, dn/dc, wavelength of the incident
light, λ, and Avogadro constant, NA. The other symbols are defined
as follows: Rcor(q,c) is the corrected Rayleigh ratio, which depends
on the sample concentration and on the scattering vector q, Mw is
the weight average molar mass of scattering particles, and A2 is the
“light-scattering-weighted” second virial coefficient of the concen-
tration expansion.
The function p(q,c) ) Icor(q)/Icor(q ) 0) takes into account both
intraparticle and interparticle interference effects. We assume that
the scattering function p(q,c) may be reasonably expressed as a
product of the single particle form factor P(q) and the solution
structure factor S(q,c), i.e., p(q,c) ) P(q) S(q,c).27 The structure
factor was equal to 1 for all concentrations and samples studied
herein. Since the values of the refractive index increment dn/dc and
the actual concentration of scattering particles are unknown in our
system, SLS data were used only to obtain the form factor P(q) by
dividing extrapolated values of Kc/R(c) to the zero angle by apparent
values of Kc/R(q,c), where c is concentration of metallacarboranes
in solution and the dn/dc value is arbitrary. From a Guinier region
of the form factor the radius of gyration, Rg, was calculated.
All solutions of metallacarboranes, measured by light scattering,
were colored (yellow for [1]- or deep red for [2]2-); however, the
wavelength of the used laser (633 nm) does not coincide with
absorption bands of metallacarboranes in water. Therefore, all SLS
and DLS results are reliable.26
Atomic Force Microscopy (AFM). All measurements were
performed in the tapping mode under ambient conditions using a
commercial scanning probe microscope, Digital Instruments Nano-
Scope dimensions 3, equipped with a Nanosensors silicon cantilever,
typical spring constant 40 N m-1. Metallacarborane particles were
deposited on a fresh (i.e., freshly peeled out) mica surface (flogopite,
theoretical formula KMg3AlSi3O10(OH)2, Geological Collection of
Charles University in Prague, Czech Republic) by a fast dip coating
in an aqueous solution of sample. After the evaporation of water,
the samples for AFM were dried in a vacuum oven at ambient
temperature for ca. 5 h.
Scanning Electron Microscopy (SEM). SEM measurements were
carried out using a scanning electron microscope coupled with an
energy-dispersive X-ray analyzer (Hitachi S-4700 + Noran D-6823).
Metallacarborane particles were deposited on glass slides by a fast
dip coating in an aqueous solution of sample. After the evaporation
of water, the slides were dried in a vacuum oven at ambient
temperature for ca. 5 h. The surface was than gold coated, and the
micrographs were taken at an acceleration voltage of 10 or 15 kV.
Conductometry. The conductivity of aqueous metallacarborane
solutions was measured with precision conductivity equipment
consisting of a Radiometer apparatus CDM210 and a conductivity
cell. The conductivity was measured for solutions of different
concentrations, which were prepared by dilution of a stock solution
(1 g/L) with deionized water. All measurements were carried out
at a temperature of 25 °C.
Gel Permeation Chromatography (GPC). The GPC equipment
(Watrex, Prague, Czech Republic) consisted of a Delta Chrom SDS
030 pump, a Delta Chrom UVD 200 detector, and a Rheodyne
injection valve, model 7125 (Cotati, CA) with a 20 µL sample loop.
Signal acquisition and data handling were performed with the
DataApex Clarity 2.1 software. A commercially available steel
column, Watrex GMB 10000 (column length 250 mm), particle size
7 µm, and deionized water as the mobile phase were used. The
detection wavelength was 214 nm. The measurements were carried
(26) Kratochvil, P. In Classical Light Scattering from Polymer Solutions,
Polymer Science Library 5; Jennkins, A. D., Ed.; Elsevier: Amsterdam, 1987.
(27) Klein, R.; D’Aguanno, B. In Light Scattering. DeVelopments and Principles;
Brown, W., Ed.; Clarendon Press: Oxford, England, 1996; p 30.
Langmuir, Vol. 22, No. 2, 2006 577
Figure 1. DLS correlation spectra of Na[1] aggregates 2 days after
preparation for different concentrations: (curve 1) 0.05 g/L, (curve
2) 0.3 g/L, (curve 3) 0.5 g/L, and (curve 4) 5 g/L.
out at the laboratory temperature (22 ( 2 °C). A mixture of standards,
MIX A (consisting of glucose and three dextrans), was used for
relative calibration.
Results and Discussion
Characterization of Metallacarborane Solutions by Light
Scattering. The solution behavior and aggregation of metal-
lacarborane were studied by static and dynamic light scattering.
DLS measurement reveals the presence of relatively large colloidal
particles containing a large number of aggregated metallacar-
borane molecules in all studied aqueous solutions of Na[1], Na2-
[2], and Cs[1]. The solutions are stable and do not precipitate.
In this paper we focused mainly on the behavior of Na[1] in
aqueous solutions; however, the behavior of Na2[2] and Cs[1]
is discussed for comparison to show a more general property of
metallacarborane aggregates in water.
Distributions of hydrodynamic radii RH of aggregates of Na-
[1] in aqueous solution, measured 2 days after the preparation,
are plotted in Figure 1 for several concentrations. The position
and width of the peaks obtained from the DLS autocorrelation
function vary with concentration. The aggregates are fairly
monodisperse at low concentrations (curve 1). Smaller aggregated
particles form at medium concentrations (curves 2 and 3), and
the fraction of large particles increases substantially in the region
of high concentrations (curve 4). Our study showed that this
trend is very reproducible. The particle polydispersity increases
with concentration, but it changes also with time. The changes
in particle size with time are depicted in Figure 3 and in the inset
in Figure 2. Values of the hydrodynamic radius RH and radius
of gyration Rg of aggregates 1 month after filtration of the solutions
are shown in Figure 2 as functions of Na[1] concentration. In
the inset, hydrodynamic radii RH of several concentrations of
Na[1] measured at different times after preparation are shown.
In the concentration region above 1 g/L, the curves increase with
increasing concentration and the polydispersity of the aggregates
is very high immediately after preparation. The polydispersity
of particles in the solution remains high, even if the solution has
been filtered and a certain fraction of the sample has been found
to adsorb on the filter. This fact indicates that large aggregates
or clusters form in the solution immediately after filtration. The
radius of particles decreases during aging of the solutions (several
weeks after the preparation) and finally remains constant in the
concentration region above 0.3 g/L. This indicates that unstable
clusters, which form shortly after the preparation of solutions,
disappear, because they transform slowly into relatively mono-
disperse stable ones.
The behavior of very dilute solutions of Na[1] is surprising.
The aggregates are fairly monodisperse, and their radius increases
578 Langmuir, Vol. 22, No. 2, 2006
Figure 2. Hydrodynamic radius RH (curve 1) and radius of gyration
Rg (curve2) as a function of the Na[1] concentration 1 month after
preparation. Inset: RH of Na[1] aggregates as a function of the
concentration 1 day (curve 1, hollow spheres), 1 week (curve 2,
black spheres), 2 weeks (curve 3, hollow squares), and 1 month
(curve 4, black squares) after preparation. The solid lines connecting
the data are drawn only to guide the eye.
Figure 3. Hydrodynamic radius RH (curve 1) and radius of gyration
Rg (curve 2) of Na[1] aggregates as a function of time. The Na[1]
concentration is 0.05 g/L. The solid lines connecting the data are
drawn only to guide the eye.
with time (Figure 3). The increase at early times is rather fast,
and then it slows, but we found that it continues at least 1 month
after preparation. The system seems to reach equilibrium (i.e.,
the radius of the aggregates does not change) after several months.
The size of the aggregates increases with decreasing concentration
below 0.3 g/L. It is worth mentioning that all time evolutions
were found to be reproducible.
The behavior of metallacarboranes is fairly complex, and its
understanding requires knowledge of the shape of the particles
and their fraction in the solution. Because the intensity of scattered
light is very sensitive to the presence of large aggregates, the
molecularly soluble metallacarboranes and their oligomers do
not almost contribute to the light scattering signal, and the
estimation of the fraction of aggregates in the solution cannot
be obtained from SLS and DLS data. This is why we have to
use a combination of several other methods (e.g., GPC; see later).
The behavior of Cs[1] was studied for comparison. It was
found that the solubility of Cs[1] is low (about 0.25 g/L) and the
sizes of the aggregates increase strongly with dilution. The
concentration dependence of RH is similar to that of Na[1] (not
shown). Furthermore, all distributions of DLS correlation times
for Cs[1] solutions were monomodal and fairly narrow in the
whole concentration range.
It is worth mentioning that no steep increase in particle size
with decreasing metallacarborane concentration was observed
in solutions of Na2[2]. Consequently, the hydrodynamic radius
Matějı́ček et al.
Figure 4. Inverse form factor P-1 of Na[1] aggregates as a function
of q2Rg2 1 h (curve 1), 3 h (curve 2), 1 day (curve 3), 2 days (curve
4), and 1 week (curve 5) after preparation. The Na[1] concentration
is 0.05 g/L. The solid curves organize the data visually but do not
represent an attempt to model the data.
of aggregates in Na2[2] solutions does not change with
concentration and yields an RH of ca. 110 nm. This metallac-
arborane is fairly soluble in water. It contains two Na+ ions in
each molecule, and it seems that Na2[2] aggregates are sufficiently
stable in water and therefore do not form secondary clusters.
Information on the shape of the aggregates can be obtained
from the form factor, P(q), in combination with the ratio Rg/RH.
The static light scattering measurements did not show any
evidence of interparticle interactions, which are frequently
observed in charged colloidal solutions at low ionic strength.
Hence, the structure factor is expected to equal 1, and only the
form factor (intraparticle effects) influences the light scattering
function p(q,c) (see eq 3). The ratios Rg/RH ) 0.785 ( 0.112,
0.736 ( 0.060, and 0.704 ( 0.112 for Na[1], Cs[1], and Na2[2],
respectively, have been obtained from SLS and DLS measure-
ments in aqueous solutions. For monodisperse hard spheres the
Rg/RH ratio equals 0.775, for hard ellipsoids this number goes
sharply up, and for hollow spheres the value is about 1.28 The
measured values suggest that metallacarborane solutions contain
roughly spherical particles.
An additional proof of this assumption is the shape of the
inverse form factor, P-1(q), as a function of the product Rg2q2
(Figure 4).26 The function turns sharply up and exhibits a
maximum at Rg2q2 ) ca. 12, which is predicted by theory for
homogeneous spheres. The shape of the function P-1(q) can be
influenced by various factors, e.g., by polydispersity in sizes,
and does not always give unambiguous information on the particle
shape. However, the observed shape of the function P-1(q) with
a clearly pronounced maximum provides a fair indication of a
more or less spherical shape. Curves measured at different times
(Figure 4) demonstrate the growth of spherical particles with
time. For relatively small particles existing at early times (curves
1-4), the product Rg2q2 does not reach enough large values to
show a maximum and the maximum is detectable only for
solutions more than 1 week old (curve 5). To yield convincing
proof for the spherical shape of metallacarborane particles, we
investigated the system also by other techniques (e.g., SEM and
AFM; see later).
Assuming the spherical shape of aggregates, the number of
aggregated metallacarborane molecules can be estimated from
the particle diameter (measured by DLS) and density (obtained
from the literature29,30). Unfortunately, there are no data available
(28) Burchard, W. In Light Scattering. DeVelopments and Principles; Brown,
W., Ed.; Clarendon Press: Oxford, England, 1996; p 439.
(29) Zalkin, A.; Hopkins, T. E.; Templeton, D. H. Inorg. Chem. 1967, 6,
1911-1915.
Molecular Assembly of Metallacarboranes in Water
Figure 5. Dependence of relaxation rates Γ obtained by DLS on
the magnitude of the scattering vector of Na[1] aggregates in water
with concentrations of 10 g/L (curve 1), 0.27 g/L (curve 2), and 0.07
g/L (curve 3). Linear dependence is assumed in the case of a diffusive
motion of particles.
for Na[1], but the density of this compound can be expected to
be in the range 1-2 g/cm3. It should be kept in mind that the
density of the aggregates could be lower than that in the solid
state because of hydration and repulsion of anions (counterions
partially release toward the bulk solution). Therefore, only a
rough estimate of the number of molecules in each particle,
106-107, was done.
To ensure that the measured hydrodynamic radii are reliable,
we have to prove that the motion of metallacarborane particles
in aqueous solutions corresponds to diffusion.28 The proof of the
diffusion-controlled motion is the linear dependence of the
relaxation rate Γ on q2, i.e., Γ ) Dq2, where D is the diffusion
coefficient. This dependence is depicted in Figure 5 for three
different concentrations of Na[1] in water. The agreement is
reasonably good within the range of experimental errors;
nevertheless, the slope of the log(Γ)-log(q) dependence, which
should be 2, is slightly higher and gradually increases with q (up
to ca. 2.25) especially for clusters below the concentration limit
at 0.3 g/L. The experimental slope may be slightly higher than
2, because it contains a contribution of a slow internal mode
proportional to q3. Its origin can be attributed to the motion of
primary aggregates in the secondary clusters, mainly in solutions
below the limit of 0.3 g/L.31
The performed light scattering measurements suggest that
aqueous solutions of metallacarborane contain aggregates of
roughly spherical shape. The behavior of the aggregates is
complex and depends on the concentration and aging of the
solution. Large clusters of primary aggregates are formed, when
solutions of Na[1] and Cs[1] are diluted below ca. 0.3 g/L. This
observation may be rationalized as follows. Individual aggregates
are stabilized by counterions. After dilution, the concentration
of counterions near the surface of metallacarborane particles
decreases, because they escape in bulk solution and gain
translational entropy. Consequently, the system is destabilized
and starts to aggregate to decrease the overall surface of the
aggregates. This means that the clustering of aggregates is a
result of the entropy-to-enthalpy interplay. Hydrophobic parts of
metallacarborane molecules try to avoid the polar aqueous
medium, and the primary aggregates form large clusters at low
dilution when small ions do not condense at their surface and
do not provide sufficient stability.
The assumption that equally charged colloidal particles
surrounded by small ions can effectively interact via attractive
forces has been discussed in the literature since the 1930s (for
(30) St. Clair, D.; Zalkin, A.; Templeton, D. H. Inorg. Chem. 1969, 8, 2080-
2086.
(31) Wu, C.; Ngai, T. Polymer 2004, 45, 1739-1742.
Langmuir, Vol. 22, No. 2, 2006 579
example, by Langmuir in 193832). The classical DLVO theory
(Derjaguin-Landau-Verwey-Overbeek)33,34 assumes only re-
pulsive electrostatic interactions between two equally charged
particles. To explain the experimental results, which cannot be
explained by the DLVO model, another pairwise potential was
proposed, which contains a long-range attractive part.35 The
tendency to form clusters or concentrated domains of ordered
particles and dilute regions was observed in solutions of silica
or latex spheres.36-39 The destabilization of a mixture of colloidal
particles differing in size was observed and explained by depletion
forces and phase separation at low ionic strength.40,41 Furthermore,
many experimental studies performed over several past decades
pointed out the importance of counterions.42 The concept of so-
called “thermodynamically bound” counterions explains the phase
transition at low ionic strength43 and the formation of temporal
aggregates44 in colloidal systems. In dilute solutions, macroion
clusters stabilized by counterions have usually been observed.45
The behavior cannot be explained by a simple pairwise potential.
It is necessary to consider the net electrostatic interaction of the
assembly of large colloidal particles and small ions.42
The above conclusions can be supported by the following
experiment concerning the destabilization of metallacarborane
aggregates upon dilution. If Na[1] solution is diluted by 0.1 M
NaCl instead of pure water, no cluster formation occurs in the
low concentration region in salted solution and the hydrodynamic
radius of the aggregates remains the same as that in the high
concentration region in pure water (ca. 110 nm). Hence, the
addition of salt suppresses the formation of secondary aggregates
in dilute solutions as predicted by theory.
The conductivity measurements on dilute Na[1] solutions were
carried out to obtain the critical micellar concentration, cmc.
The specific conductivity of associating systems as a function
of concentration shows typically one or two breaks corresponding
to the critical micellar and premicellar concentrations.2,46 We
did not observe any breaks up to 2 g/L (conductometer limit),
despite the existence of aggregates in Na[1] solutions. The
dependence of specific conductivity on concentration is almost
linear (not shown). It confirms our conclusion from light scattering
measurements that aggregates exist in the whole measured
concentration range. Therefore, the detection of the cmc is below
the limits of both light scattering and conductivity measurements.
We have to keep in mind that the detection of the cmc is
complicated by the observed formation of large clusters in very
dilute solutions. It can be suppressed by the addition of salt, but
this precludes conductometry measurements. The surface tension
measurements of very dilute solutions might be more appropriate
even after addition of the salt as pointed out by a reviewer of
this paper. These experiments have not yet been carried out.
(32) Langmuir, I. J. Chem. Phys. 1938, 6, 873-896.
(33) Deryagin, B.; Landau, L. Acta Physicochim. URSS 1941, 14, 633-662.
(34) Verwey, E. J. W.; Overbeek, J. Th. G. Theory of the Stability of Lyophobic
Colloids; Elsevier Publishing Co., Inc.: New York, 1948.
(35) Sogami, I.; Ise, N. J. Chem. Phys. 1984, 81, 6320.
(36) Ito, K.; Ieki, T.; Ise, N. Langmuir 1992, 8, 2952-2956.
(37) Yoshida, H.; Ise, N.; Hashimoto, T. J. Chem. Phys. 1995, 103, 10146-
10151.
(38) Yoshida, H.; Ise, N.; Hashimoto, T. Langmuir 1995, 11, 2853-2855.
(39) Dosho, S.; Ise, N.; Ito, K.; Iwai, S.; Kitano, H.; Matsuoka, H.; Nakamura,
H.; Okumura, H.; Ono, T. Langmuir 1993, 9, 394-411.
(40) Henderson, D.; Boda, D.; Chan, K. Y.; Wasan, D. T. Mol. Phys. 1998,
95, 131-135.
(41) Xu, W.; Nikolov, A. D.; Wasan, D. T. AIChE J. 1997, 43, 3215-3222.
(42) Schmitz, K. S. Langmuir 1997, 13, 5849-5863.
(43) Lin, S. C.; Lee, W. I.; Schurr, J. M. Biopolymers 1978, 17, 1041-1064.
(44) Clark, N. A.; Ackerson, B. J.; Hurd, A. J. Phys. ReV. Lett. 1983, 50,
1459-1462.
(45) Groehn, F.; Antonietti, M. Macromolecules 2000, 33, 5938-5949.
(46) Hunter, R. J. Foundations of Colloid Science; Clarendon Press: Oxford,
1995; Vol. I.
580 Langmuir, Vol. 22, No. 2, 2006 Matějı́ček et al.

Characterization of Metallacarborane Aggregates by

AFM and SEM. Light scattering measurements revealed the
existence of aggregates in metallacarborane solutions, but a
detailed study of their structure requires more direct experimental
techniques, e.g., AFM and SEM. Although the shape of the
colloidal particles is often deformed after their deposition on the
substrate, the microscopy study yields valuable information on
the shape and polydispersity of aggregates and also on the structure
of detected clusters. By preparing samples for microscopy
measurements from dilute solutions, we ensure that particles are
evenly spread and do not overlap on the substrate. Hence, we
can relate the original shape of the particles in solution and that
on the surface.
The AFM scan of Na2[2] aggregates on a mica surface (Figure
6a) shows fairly monodisperese particles with a spherical plan
of view deposited from the solution with a concentration of 0.5
g/L. The aggregates of Na[1] deposited from solutions in the
concentration range 0.25-10 g/L are very similar. On the contrary,
the aggregates of Na[1] deposited on a mica surface from very
dilute solutions (0.03 g/L) are larger and polydisperse (Figure
6b). In Figure 6c, typical profiles of Na[1] aggregates for different
concentrations are shown. The dependence of aggregate sizes on
concentration is similar to that in the solutions (compare Figure
2). Nevertheless, the clusters deposited on the surface are less
uniform than those in the solution. By scanning a large mica
surface, we find both regions of small spherical aggregates and
their clusters. This observation supports the assumption concern-
ing the phase separation in very dilute solutions of metallac-
arboranes. It is necessary to take into account that large clusters
on mica may split into smaller fragments due to deformation
forces acting at the surface.
The SEM pictures of the gold-coated Na[1] sample deposited
on glass from the solution with a concentration of 0.175 g/L also
show (Figure 7a,b) both individual spherical aggregates and
clusters formed by smaller particles for the same sample. This
observation supports the suggested mechanism of secondary
aggregation (clustering) below the concentration limit of 0.3
g/L. It should be mentioned that the coexistence of small
aggregates and their clusters was observed for all concentrations
of Na[1] solutions aged several days (0.03, 0.175, 2.0, and 10.0
g/L). This indicates that clusters form in the whole concentration
range with time. However, they are less stable in the concentration
region above ca. 0.3 g/L (see the light scattering section).
Both microscopy techniques image metallacarborane ag-
gregates as relatively spherical and monodisperse particles.
Experiments with samples deposited from solutions differing in
aging time show that metallacarboranes undergo secondary
aggregation with time and dilution.
Fraction of Na[1] Aggregates in Aqueous Solutions. To
obtain information on the fraction of aggregated metallacarborane
molecules, a gel permeation chromatography study of Na[1]
solutions was carried out. All chromatograms for concentrations
between 1 and 10 g/L are fully reproducible and show two peaks
at retention times of about 6.5 and 13.0 min (Figure 8). For
solutions with c < 1 g/L, the second peak is hidden in the noise
and cannot be analyzed. The first peak can be assigned to
aggregates and the second one to the fraction of Na[1] molecules Figure 6. (a) AFM scan on a mica surface of Na2[2] aggregates
or small oligomers (dextran calibration), which are detectable adsorbed from solution with a concentration of 0.5 g/L. (b) AFM
neither by light scattering nor by microscopy. scan on a mica surface of Na[1] aggregates adsorbed from solution
with a concentration of 0.03 g/L. (c) Typical profiles of AFM scans
Because the kinetics of aggregate formation is relatively slow of Na[1] aggregates deposited on a mica surface from solution with
(see the light scattering section), the ratio of large aggregates to concentrations of (curve 1) 0.03 g/L, (curve 2) 0.1 g/L, and (curve
small oligomers should remain unchanged during the separation 3) 0.25 g/L.
Molecular Assembly of Metallacarboranes in Water
Figure 7. SEM scans of Na[1] aggregates deposited from aqueous
solution with a concentration of 0.175 g/L on a glass surface and
gold coated: (a) individual particles of aggregates and (b) clusters
of aggregates.
Figure 8. Gel permeation chromatogram of a Na[1] aqueous solution
with a concentration of 2 g/L.
in the column.47 Nevertheless, the mobility of molecularly soluble
metallacarborane or small oligomers (based on the “second”
peak) is slower than the water flow. This is probably due to
specific interactions of metallacarborane molecules with the
(47) Prochazka, K.; Bednar, B.; Tuzar, Z.; Kocirik, M. J. Liq. Chromatogr.
1988, 11, 2221-2239.
Langmuir, Vol. 22, No. 2, 2006 581
stationary phase. Hence, the equilibrium between molecules and
aggregates of metallacarboranes may shift in the column, and
the amount of aggregates in solution can be underestimated by
GPC. The fraction of Na[1] in the aggregates was calculated
from the measured chromatograms as the ratio of the first peak
area divided by the area of both peaks. Its value does not change
with concentration, and it yields 0.82. This means that the majority
of Na[1] molecules (at least 80%) form in relatively large
aggregates.
Conclusions
The metallacarboranes studied in this work show a pronounced
amphiphilic character, which results in formation of aggregates
in aqueous solutions. Their size, shape, and behavior were studied
by a combination of light scattering (SLS and DLS) and
microscopy (AFM and SEM) techniques with the help of
chromatographic measurements. The association of metallac-
arboranes in water was observed for the first time.
Several weeks after the preparation, the hydrodynamic radii
of the aggregates reach “equilibrium” values around 115 nm for
Na[1] above the concentration limit of 0.3 g/L and 110 nm for
Na2[2] in the whole concentration range studied. Angular SLS
measurements, together with values of the Rg/RH ratio and
microscopy scans, suggest that the aggregates are spherical.
Individual particles of Na[1] and Cs[1] undergo secondary
aggregation when the solution is diluted below 0.3 g/L. The size
of the clusters formed at high dilution increases steeply with
decreasing concentration. The clustering is efficiently suppressed
by the addition of an indifferent electrolyte, e.g., NaCl.
The fraction of Na[1] molecules which form the aggregates,
determined by GPC, is approximately 80%. This means that the
majority of metallacarborane molecules form aggregates in
aqueous solutions and only a relatively small fraction of molecules
dissolve molecularly or form small aggregates.
Acknowledgment. We express our gratitude to Dr. Jaromı́r
Plešek and Dr. Bohumı́r Grüner (Institute of Inorganic Chemistry,
Academy of Science of the Czech Republic, Řež near Prague,
Czech Republic) for providing the samples of metallacarboranes
and for helpful discussion. We also thank Dr. Eva Tesařová
(Department of Physical and Macromolecular Chemistry, Charles
University in Prague, Czech Republic) for GPC measurements
and data analysis, Ing. Milena Špı́rková (Institute of Macro-
molecular Chemistry, Academy of Science of the Czech Republic
in Prague) for AFM measurements, and Dr. Jan Konvalinka and
Milan Kožı́šek (Institute of Organic Chemistry and Biochemistry,
Academy of Science of the Czech Republic, Prague, Czech
Republic) for fruitful research cooperation following this project.
This work was supported by grants from the Ministry of Education
of the Czech Republic (MSM 6046137307, and LC 512), Grants
303/05/0336, 203/03/0900, 203/03/0716, and 301/04/1315 from
the Grant Agency of the Czech Republic, and EU Grant CIDNA
NMP4-CT-2003-505669 to V.K. K.P. and P.M. acknowledge
the financial support of the Grant Agency of the Czech Republic
(Grant 203/04/0490). This work was partly supported by the
grant from Ministry of Education (MSMT) of Czech Republic
within Programme 1M0508 “Research Centre for new Antivirals
and Antineoplastics” to P.C.
LA052201S