ARTICLE

Centrifugal Field-Induced Colloidal

Assembly: From Chaos to Order
Mengdi Chen, Helmut Cölfen,* and Sebastian Polarz*

Department of Chemistry, University of Konstanz, Universitätsstrasse 10, 78457 Konstanz, Germany

ABSTRACT Given a certain binary phase, its crystal structure can

usually be rationalized by thermodynamics and packing rules.
However, there is still no fundamental understanding of how
metastable solids form and how this is influenced by kinetic factors.
Furthermore, the very early stages of crystallization remain vague.
We present an experimental study using a binary mixture of colloidal
particles as a model. Particle assembly is provoked by centrifugation,
and the degree of separation can be adjusted precisely. Between the
scenarios of random mixture (glassy state) and complete phase
separation, conditions could be realized to facilitate the occurrence
of various periodic structures in a single experiment. Some of these
structures can only be explained by a dynamic, nonequilibrium and dissipative state. Eventually, the formation of certain metastable lattices can be
interpreted as an emergent, systemic phenomenon. Thus, centrifugation can be used for studying the chaos
order transition in complex particle systems.

KEYWORDS: systems phenomena . colloidal model systems . nonclassical nucleation . analytical ultracentrifugation . self-assembly

T
he structure of any solid at the unit- Furthermore, a glass is also a rather vague
cell dimension is a prevailing factor term. What could make two glasses differ-
determining its properties, which is ent and how to control the difference is still
best represented by materials existing in an open question.
different modifications (polymorphs, glasses), In conclusion, it is highly desirable to
e.g., carbon as an extreme case. It is highly obtain deep insights into the formation of
important to understand, in depth, how and glasses and metastable solids on the basis
which modification forms and how to ac- of experimental data. Because the latter is
cess potentially undiscovered forms.1 For a currently beyond reach for atomic systems,
system defined in composition, only one we follow a different approach here. It is
distinct modification can be explained sole- well-known from literature that monodis-
ly by thermodynamic considerations; all perse colloidal particles can act as models
other forms are the result of more or less for solid-state phenomena.6
8 The advan-
marked, kinetic contributions. To this day, tage is the much easier accessibility of
insights are rather empirical because the data as the colloidal particles used are typi-
crystal structure is determined at the very cally several orders or magnitudes larger
early phases of the process, either at solidi- compared to atoms. Electron microscopy
faction from a dispersion or for amorphous
becomes a powerful tool for the investi-
crystalline transitions. The lifespan of these gation of the resulting solid materials.
early stages is short; they are far from equi- Most of the past attention was devoted to * Address correspondence to
librium with highly dynamic character.2,3 materials consisting of one size of spheri- helmut.coelfen@uni-konstanz.de,
sebastian.polarz@uni-konstanz.de.
There is very limited information about cal particles due to their special, photonic
the factors governing the very early stages properties.9,10 Different methods for colloi- Received for review February 17, 2015
of crystallization.4,2 Whereas numerous meth- dal assembly are known starting from col- and accepted July 12, 2015.
ods exist for investigating the crystal- loidal dispersions like evaporation-induced
Published online July 13, 2015
line state, with diffraction methods leading self-assembly, electrophoresis, or centrifu- 10.1021/acsnano.5b01116
the way, the precise characterization of gation.11
19 Using colloidal particles as
an amorphous solid is extremely difficult.5 supra-atoms enables the study of a range C 2015 American Chemical Society

CHEN ET AL. VOL. 9 ’ NO. 7 ’ 6944–6950 ’ 2015 6944

www.acsnano.org
 ARTICLE
of binary systems;7,20,21 see also the excellent review
articles by Möhwald et al. and by Stein et al.22,23 In
2003, Kitaev and Ozin presented intriguing phases
formed by polystyrene latex particles differing strongly
in size (e.g., 1.3 μm þ 145 nm).24 Park et al. have
described a layer-by-layer technique for the genera-
tion of thin films of binary colloidal phases.25 One
should also note the work of Murray et al., who have
prepared highly ordered, binary lattices using in-
organic nanocrystals and also those with nonspherical
shape.7,26 The resulting solids may then be interpre-
ted as alloys, with glasses and solid solutions as the
chaotic state and crystalline binary structures as or-
Figure 1. Φ dependency of the sedimentation coefficients.
dered states. Circles = monodisperse dPS1, squares = dPS1 in binary
In the current study, we use colloidal spheres system dPS1 þ dPS2, triangles = monodisperse dPS2; solid
(monodisperse, anionically stabilized polystyrene latex lines = theoretical data (eq 1).
particles prepared using emulsion polymerization)27
combined with centrifugation as a model system for found sedimentation coefficients fit perfectly to the
studying the peculiarities of binary phases over a range relationship given by eq 1. It seems that there are no
of compositions. An advantage of this approach is that, special features apart from the expected nonideality
unlike a classical crystallization experiment comprising even though two different sizes of the colloidal parti-
atoms or ions, various parameters can be varied in- cles are present.30
dependent from each other, including kinetic param- With the data in Figure 1, it should be possible to
eters. The size d of the individual colloidal particles differentiate between strong and weak separation
(dPS1= 150 nm, dPS2 = 300 nm) and the quotient conditions by the differences in the sedimentation
(QPF = dPS1/ dPS2) thereof was kept constant. First, we coefficients of the two latices (w Δs), depending on
varied the total volume fraction of colloidal particles the combination of parameters cPS1,2, Φ, and F and
(Φ = φPS1 þ φPS2) keeping their relative number ratio centrifugation time tCF (see also S-3, Supporting
(N = φPS1/φPS2) constant because Φ comes closest to Information). The results are shown in Figure 2. When
changing concentration in a classical crystallization separation is weak, only at the top of the centrifugation
experiment. Then, the centrifugal force F was changed vial does one find a region consisting of separated
as a kinetic parameter influencing how fast the parti- smaller spheres PS1, which assemble into close pack-
cles sediment. The emerging structures were investi- ing (Figure 2a,b). The remaining, bottom part of the
gated using scanning electron microscopy (SEM) and sample exists as a mixture of both sphere types in a
were compared to classical solid-state crystals when- disordered, glassy state. The presence of larger spheres
ever possible. Finally, the outcome of an experiment in the matrix of the smaller spheres does obviously
with two successive steps with different centrifugal strongly prohibit the incurrence of crystalline domains,
force was investigated. which fits to experimental and theoretical observations
on binary colloidal systems made by others.31,32 Closer
RESULTS AND DISCUSSION inspection of the SEM images shows that there is a
First, it is important to gain control over sedimenta- shallow gradient regarding the relative concentration
tion and separation behavior of the binary system. of the two types of particles (see also S-4, Supporting
Analytical ultracentrifugation (AUC) is a powerful tool Information).
to obtain in situ information about the sedimentation The situation changes when conditions for strong
of the particles.28 AUC was performed for dispersions separation conditions are selected (Figure 2e,f). PS1
varying in total volume fraction, and the monodisperse and PS2 have formed separate domains, each predo-
situation was compared to the binary case (Figure 1; minantly close packed (face-centered cubic; see S-5,
see also S-2, Supporting Information). Several impor- Supporting Information). Between these two extremes
tant conclusions can be extracted from the presented there is a situation shown in Figure 2c in which at the
data. According to sedimentation theory, the concen- top and bottom of the centrifugation vial are densest
tration dependency of the sedimentation coefficient packed zones. In between, a glassy, binary system can
s for repulsive interaction between the particles is29 be found. The gradient in the particle size distribution
function is more pronounced than before (S-4, Sup-
s0
s ¼ (1) porting Information). However, close to the medium
(1 þ ks c)
separation scenario, one can find systems in which the
with s0 = the sedimentation coefficient at infinite dilu- entire binary phase is not completely disordered.
tion and ks = nonideality constant. The experimentally Several binary zones possess periodic order (Figure 2d).

CHEN ET AL. VOL. 9 ’ NO. 7 ’ 6944–6950 ’ 2015 6945

www.acsnano.org
 ARTICLE
Figure 3. SEM micrographs of disordered binary structures
obtained for large values of F (a) and ordered binary
structures obtained for small values of F (b). Scale bars =
200 nm.

(see Figure SI-7a, Supporting Information). However,

from the side, it becomes clear that only in the case of
NaCl the centers of small and large spheres are placed
in plane, and for CsCl the small spheres are located
below the top layer. From closer inspection of the SEM
images (Figure 3 and Figure SI-7b, Supporting
Information) it can be proven unequivocally that the
rocksalt structure is present here. The emergence of
the NaCl lattice is not surprising as it can be explained
purely by packing arguments. QPF = 0.5 is very close to
the ideal value (0.414) for the NaCl structure.33 It is the
thermodynamic optimum for this particular combina-
tion of particles. Therefore, the NaCl structure can even
be obtained using only earth's gravitational field (S-8,
Supporting Information).
Figure 2. Schematic drawing and SEM micrographs show-
ing different concentration gradients of the two sphere However, for medium separation conditions and
types (dPS1, dPS2) in the binary system caused by the medium values of centrifugal force F (= 3.0  103 g),
adjustment of separation conditions (a, b = weak; c, complexity explodes. A number of periodic, thermo-
d = medium; e,f = strong). (d) Emergence of periodic binary
features near the disorder f order transition. Scale bars = dynamically less stable structures (Figures 4 and 5) are
1 μm. observed. The emergence of such a variance of struc-
tures is in good agreement with a prediction made by
Because we wanted to explore the latter phenom- Heras et al. in a theoretical paper in 2013, where the
enon in further depth, the remaining parameters like authors described phase diagrams of colloidal mix-
QPF, N, and most importantly the centrifugal force were tures under gravity.34 The structure shown in Figure 4a
varied. A subtle interplay between Δs and centrifugal is in best agreement with a lattice based on a densely
force F is the central order parameter in the system. It is packed layer of the larger particles PS2. On top of this,
obvious that F directly influences the sedimentation the smaller spheres PS1 form a layer by regular filling of
rate. When sedimentation occurs very fast due to high the voids of the PS2-layer. The structure is finalized by
centrifugal force (e.g., F = 1.2  104 g; g = earth's an additional PS2 layer on top. Comparison to struc-
gravitational acceleration 9.81 m/s2), the systems tures related to atomic lattices shows that Figure 4a
remain in the glassy state independent of Δs (see indicates a 1:2 stoichiometry akin to aluminum dibor-
Figure 3a). ide (AlB2; space group P6/mmm). Figure 4b can be
For small centrifugal force (e.g., F = 1.2  102 g), a rationalized by a small distortion involving a move-
significant fraction of the material exhibits the forma- ment of the PS1 spheres closer to each other in a
tion of a periodic (PS1)(PS2) structure shown in pairwise fashion (see S-9 Scheme 1, Supporting In-
Figure 3b, consistent with the NaCl lattice (space group formation). The crystal domain shown in Figure 4b is
Fm3m) (see S-6, Supporting Information). However, it slightly less symmetric, but it still has the (PS1)2(PS2)1
must be noted that the assignment of crystal structures composition. The structure shown in Figure 4c cannot
(3D information) from only one crystal plane [hkl] (2D be related to AlB2, which is hexagonal in nature with a
information) is difficult and often not precise, and hash-shaped a,b unit cell basis. The edges indicated in
whenever possible one should take adjacent lattice Figure 4c (see also S-9, Scheme 2, Supporting Information)
planes into consideration. For instance, when seen allow identification of the a, b, c directions. The (001) plane
exclusively from the [100] perspective, NaCl and CsCl is composed of four PS2 particles arranged in a quadratic
look almost identical with a square unit pattern fashion (a = b = dPS2 = 300 nm), centered by one PS1

CHEN ET AL. VOL. 9 ’ NO. 7 ’ 6944–6950 ’ 2015 6946

www.acsnano.org
 ARTICLE
Figure 5. SEM micrographs of the nondense structures.
Atomic analogue: sphalerite (a) and wurtzite structure
(b). Scale bars = 200 nm.

obtained. The possibility for the occurrence of thermo-

dynamically less stable structures is best represented
by a periodic structure characterized by the same
stoichiometry compared to the NaCl structure (1:1)
but with a much lower packing factor shown in
Figure 5. Whereas the coordination mode in NaCl is
(6/6), a tetrahedral (4/4) coordination can be clearly
identified. The observed phase is in good agreement
to the sphalerite structure (F43m). In addition to
sphalerite a pattern can also be found that fits to
the hexagonal analogue wurtzite (P63mc) (Figure 5b).
QPF (= 0.5) is much too large compared to the ideal
value for the sphalerite or wurtzite (0.225). Thus, the
emergence of the shown structures with (4/4) co-
Figure 4. SEM micrographs and schematic of the crystal ordination represents presumably the most aston-
lattice for different ordered binary phases obtained via
centrifugation: (a) layered structure akin to AlB2; (b) dis- ishing and, thus, most significant result of our current
torted AlB2 structure; (c) the tetragonal A3B variant; the study. It should be noted that the prevalence of the
tetragonal A2B variant before (d) and after distortion structures is very different. Whereas relatively large
(e). Scale bars = 200 nm.
areas (>10 μm) associated with the more dense NaCl
sphere. Thus, the symmetry is no longer hexagonal as and AlB2 structure can be found (see Figure S-11,
in AlB2. The c-extension = 430 nm indicates that a Supporting Information), the other phases are pre-
tetragonal system is adopted. The isolated unit cell is sent more seldom and are smaller. However, this
shown in S-10 (Supporting Information) and corre- result is rather expected. A precise quantification,
sponds to a (PS1)3(PS2)1 stoichiometry. which could help in assessing the relevance of
The ordered domains observed in Figure 4d can thermodynamic versus kinetic factors, is not possible
be explained starting from the latter structure by with our methods because the structures are present
removing two symmetry equivalent spheres from the side by side, and SEM only gives information about
PS1-layer (S-9, Scheme 2 Supporting Information). the surface. It is not known at which position a
This changes the stoichiometry to (PS1)2(PS2)1. Finally, domain is seen and how far the domain extends
because there is now more space in the [002] layer it in 3-D.
can be observed that the PS1 particles are slightly Whereas all previous lattices (Figures 3b and 4) could
changing their position (S-9, Scheme 3, Supporting be explained by packing considerations and stoichi-
Information), and the structure shown in Figure 4e is ometry influenced by the shallow concentration

CHEN ET AL. VOL. 9 ’ NO. 7 ’ 6944–6950 ’ 2015 6947

www.acsnano.org
 ARTICLE
gradient, despite centrifugation as a force for densifi-
cation, sphalerite and wurtzite are clearly less dense
structures. Furthermore, for hard colloidal particles
with repulsive, spherical interaction potentials it is very
hard to explain the low coordination number with local
tetrahedral symmetry. The latter results are first indica-
tions that the mechanism leading to the various struc-
tures might be more complex than anticipated in the Figure 6. Schematic illustration of the effect of the insertion
first place. of a spherical guest particle of size R1 into a colloidal
dispersion of host particles with size R2. The host particles
Further evidence for the latter hypothesis comes are excluded from a volume that is larger than that of the
from the inspection of the entire length of the sedi- guest particle volume V1 by a factor of (1 þ R2/R1)3.
mentation body (Figure 2c,e and S-12, Supporting
Information). At the bottom of the centrifugation tube anymore but a dispersion of host latexes and the
a closest packed PS2 phase exists. This finding is situation becomes more complex.
unusual. Prior to centrifugation there is a homoge- For concentrated binary dispersions, it has been
neous mixture of the two sphere types in the entire pointed out that the buoyant force exerted on a sedi-
volume. When centrifugation and sedimentation be- menting particle can be strongly affected by other com-
gins, although the sedimentation coefficients sPS1 and ponents in the solution.30 Our binary suspension can be
sPS2 are different (Figure 1), small particles close to the regarded as a host (PS2) suspension containing guest
bottom of the vial are expected to sediment together particles (PS1). This situation is illustrated in Figure 6.
with the bigger ones to form a binary mixture (S-13, Since the guest particle is surrounded by a depletion
Supporting Information). However, the absence of PS1 layer of water with a density, which is lower than that of
in this region indicates that there must also be an the surrounding host dispersion, it is understandable
upward movement, which removes the smaller parti- that a flotation of the guest particle can occur depend-
cles from the bottom zone. This upward movement ing on the thickness of the depletion layer.30 Indeed, it
is not understandable in terms of the standard buoy- has been shown that very dense gold nanoparticles
ancy term (1
hνF) used in the evaluation of (ultra)- floated in a dispersion of polymer latexes. However, for
centrifugation experiments with hν = partial specific a generic description, the simple excluded volume
volume of the sample (0.949 mL/g for polystyrene) model can not be used and the buoyancy force can
and F = solvent density since this predicts sedimenta- be calculated using density functional theory (DFT)35
tion of the polystyrene latexes because their density is and the Mansoori
Carnahan
Starling equation of
higher than that of the solvent water. However, in our state for hard sphere mixtures yielding an equation
case, the solvent is not a one-component system for the effective density F* of the host suspension30

F
F0 6
2(2
3q)(1
Φ2 )
3(1
q2 )(1
Φ2 2 ) þ (1
q)2 (2 þ q)(1
Φ2 )3

¼ (2)
m 2 0 n2 2(4
Φ2 )Φ2 þ (1
Φ2 )4
with
movement (flotation). This is the case at Φ2 > 0.05 for
m2 0 ¼ (F2
F0 )vPS2 (3)
our investigated conditions, a concentration exceeded
for the entire length of the column in the tube. The
NPS2
n2 ¼ (4) larger the particle size difference is, the more pro-
V
nounced the flotation of the guest particles.30 A yet
dPS2 unsolved key problem is to quantitatively experimen-
q ¼ (5) tally observe the radial concentration gradients of both
dPS1
latexes in the mixture in the Analytical Ultracentrifuge.
where F0 = density of the pure solvent, F2 = density of This would in principle be possible if the differently
PS2 particle, m2 = mass of PS2 particle vPS2 = volume sized nanoparticles have different UV
vis markers.
of a PS2 particle, n2 = concentration of PS2 particles, Detection of the concentration gradients would allow
NPS2 = number of PS2, V = total volume of the quantitative calculation of the radially dependent
suspension, Φ2 = volume fraction of the PS2 particles effective density F* by eqs 2
5 and therefore the
with Φ2 = 1
Φ1, and dPS1 and dPS2 are the diameters calculation of the entire concentration gradients. How-
of two kinds of particles, respectively. ever, the strong light scattering of the PS2 prevents this
When the volume fraction of the host particles even when using 60% sucrose to match the PS refrac-
exceeds a certain value, the effective density F* of tive index as closely as possible. The radial concentra-
the solvent turns to be equal and higher than the tion gradients at high particle concentrations also
density of the guest particles leading to their upward hinder the calculation of a Peclet number for each

CHEN ET AL. VOL. 9 ’ NO. 7 ’ 6944–6950 ’ 2015 6948

www.acsnano.org
 ARTICLE
the smaller particles PS1. The dynamic character, com-
bined with the large number of interacting species and
the steady supply with energy, due to the ongoing cen-
trifugation, enables a dissipative state far from equi-
librium, and this state is located near the disorder

order transition.

CONCLUSIONS
Figure 7. SEM micrographs for the two-step sedimentation In this study, we were able to show that it is possible
experiment at F = 1.2  104 g for 20 min first (a) followed by to create a structural gradient of a binary latex mixture
F = 1.2  102 g for 5 h (b). Scale bars = 200 nm.
including different glassy states but also different
particle size as is possible for diluted dispersions.36 The crystalline structures in a single centrifugation experi-
reason is the concentration-dependent diffusion coef- ment. With the help of analytical ultracentrifugation as
ficient, which changes locally in the centrifuge tube an in situ observation technique, we could precisely
and with it the Peclet number, which relates convec- adjust the degree of separation of the colloidal parti-
tion to diffusion. cles with different size and, thus, also a concentration
The system seems to be more dynamic than we had gradient. It was found that the centrifugal force is a
expected. To test this, we performed a two-step sedi- crucial order parameter in the system. Besides the
mentation experiment. First, the dispersion was cen- expected thermodynamically most stable structure
trifuged at F = 1.2  104 g for 20 min and then the same and glassy states, a large number of different meta-
vial was centrifuged at F = 1.2  102 g for 5 h. As stable polymorphs varying in stoichiometry were
described before, the first step leads to a binary, glassy observed side by side in one experiment. Most impor-
state (Figure 7a). Next, although all colloidal particles tantly, considering that centrifugation is a process
have already sedimented, the second step leads to leading to densification, the occurrence of less-dense,
complete reorganization (Figure 7b). A periodic binary low-coordinated structures like Sphalerite was very
phase is observed. surprising. It could be shown that the emergence of
Finally, we can compare the system presented here periodic order is an emergent phenomenon resulting
to other cases, for which the emergence of order from from nonequilibrium, dissipative conditions, which are
chaos has been reported. One famous example is the commonly not considered in crystallization theories,
so-called Bénard experiment.37 A thin film of a liquid is but can be found for instance in systems like the
exposed to a temperature gradient, which is parallel to Bénard cell known for the description of ordering
the direction of gravitation. Gravity tends to pull the phenomena in complex systems. Therefore, our cur-
denser liquid from top to bottom, while viscous damp- rent study poses the stimulating question, if besides
ing force acts as a counter force. Depending on ΔT packing factors and preference for coordination geo-
(the order parameter in the Bénard system), different metry also nonequilibrium, systemic features could
and ordered patterns can emerge in the liquid. If ΔT is influence the result of kinetically controlled crystal-
increased to be larger, the system would become more lization process not only at the colloidal but also on the
complex and chaotic. The situation in the presented atomic scale. It seems that in the initial periods of a
centrifugation experiments of binary colloidal disper- crystallization process in a binary system, even for a
sions is similar. There is an overall sedimentation move- fixed size ratio of the particles, several polymorphs can
ment of the particles caused by the centrifugal field. form parallel to each other. Indeed, for CaCO3, it was
The densification of the dispersion at the bottom of the recently found that several nucleation pathways are
centrifugation vial leads to an upward movement of simultaneously operative.38

METHODS direction along the centrifugal force was subsequently investi-

gated by using a Zeiss 249 CrossBeam 1540XB scanning electron
The monodisperse, negatively charged polystyrene micro-
microscope. A Beckman Optima XL-I analytical ultracentrifuge
spheres were synthesized by emulsifier-free emulsion poly-
and a UV/vis multiwavelength analytical ultracentrifuge39 devel-
merization27 and purified by dialysis. Binary colloidal structures
oped as an open-source instrument40 were used to measure the
of 150 nm (PS1) and 300 nm (PS2) latexes were fabricated with
sedimentation coefficient of polystyrene spheres.
the L-70 ultracentrifuge (Beckman instruments) using the Beck-
man swing out rotor SW 55 Ti and ultra clear (5  41 mm) Conflict of Interest: The authors declare no competing
centrifuge tubes to maintain sedimentation as undisturbed as financial interest.
possible. After centrifugation, the remaining solvent at the top
of the centrifuge tube was removed before the centrifuge tube Acknowledgment. M.C. is funded by a Chinese Scholarship
was placed in a desiccator over silica gel for drying. The resulting Council stipend.
structures were examined by cutting the dried samples with a Supporting Information Available: ζ-Potential measurements
razor blade along the long axis of the sample column, and the for the different latex particles, concentration dependency of

CHEN ET AL. VOL. 9 ’ NO. 7 ’ 6944–6950 ’ 2015 6949

www.acsnano.org
 ARTICLE
sedimentation behavior studies by analytical ultracentrifugation, 20. Redl, F. X.; Cho, K. S.; Murray, C. B.; O'Brien, S. Three-
adjustment of different separation conditions, photographic Fimensional Binary Superlattices of Magnetic Nanocryst-
images showing the development of separation with centrifuga- als and Semiconductor Quantum Dots. Nature 2003, 423,
tion time, schematic representation of the different concentra- 968–971.
tion gradients obtained due to different separation conditions, 21. Bartlett, P.; Ottewill, R. H.; Pusey, P. N. Superlattice Forma-
transformation of different binary structures, investigation of the tion in Binary-Mixtures of Hard-Sphere Colloids. Phys. Rev.
entire length of the sedimentation body, and schematic repre- Lett. 1992, 68, 3801–3804.
sentation of the sedimentation process expected at the bottom 22. Wang, D. Y.; Möhwald, H. Template-Directed Colloidal Self-
of the centrifugation tube. The Supporting Information is avail- Assembly - The Route to 'Top-Down' Nanochemical En-
able free of charge on the ACS Publications website at DOI: gineering. J. Mater. Chem. 2004, 14, 459–468.
10.1021/acsnano.5b01116. 23. Li, F.; Josephson, D. P.; Stein, A. Colloidal Assembly: The
Road from Particles to Colloidal Molecules and Crystals.
Angew. Chem., Int. Ed. 2011, 50, 360–388.
REFERENCES AND NOTES 24. Kitaev, V.; Ozin, G. A. Self-Assembled Surface Patterns of
1. Jansen, M. A Concept for Synthesis Planning in Solid-State Binary Colloidal Crystals. Adv. Mater. 2003, 15, 75–78.
Chemistry. Angew. Chem., Int. Ed. 2002, 41, 3746–3766. 25. Kim, M. H.; Im, S. H.; Park, O. O. Fabrication and Structural
2. Pienack, N.; Bensch, W. In-Situ Monitoring of the Forma- Analysis of Binary Colloidal Crystals with Two-Dimensional
tion of Crystalline Solids. Angew. Chem., Int. Ed. 2011, 50, Superlattices. Adv. Mater. 2005, 17, 2501–2505.
2014–2034. 26. Dong, A. G.; Chen, J.; Vora, P. M.; Kikkawa, J. M.; Murray,
3. Demichelis, R.; Raiteri, P.; Gale, J. D.; Quigley, D.; Gebauer, D. C. B. Binary Nanocrystal Superlattice Membranes Self-
Stable Prenucleation Mineral Clusters are Liquid-Like Ionic Assembled at the Liquid-Air Interface. Nature 2010, 466,
Polymers. Nat. Commun. 2011, 2, 590. 474–477.
4. Gebauer, D.; Cölfen, H. Prenucleation Clusters and Non- 27. Reese, C. E.; Guerrero, C. D.; Weissman, J. M.; Lee, K.; Asher,
Classical Nucleation. Nano Today 2011, 6, 564–584. S. A. Synthesis of Highly Charged, Monodisperse Polystyr-
5. Greaves, G. N.; Sen, S. Inorganic Glasses, Glass-Forming ene Colloidal Particles for the Fabrication of Photonic
Liquids and Amorphizing Solids. Adv. Phys. 2007, 56, Crystals. J. Colloid Interface Sci. 2000, 232, 76–80.
1–166. 28. Planken, K. L.; Cölfen, H. Analytical Ultracentrifugation of
6. Leunissen, M. E.; Christova, C. G.; Hynninen, A. P.; Royall, Colloids. Nanoscale 2010, 2, 1849–1869.
C. P.; Campbell, A. I.; Imhof, A.; Dijkstra, M.; van Roij, R.; van 29. Harding, S. E., Rowe, A. J., Horton, J. C., Eds. Analytical
Blaaderen, A. Ionic Colloidal Crystals of Oppositely Ultracentrifugation in Biochemistry and Polymer Science;
Charged Particles. Nature 2005, 437, 235–240. Royal Society of Chemistry: London, 1992; p 629.
7. Shevchenko, E. V.; Talapin, D. V.; Kotov, N. A.; O'Brien, S.; 30. Piazza, R.; Buzzaccaro, S.; Secchi, E.; Parola, A. On the
Murray, C. B. Structural Diversity in Binary Nanoparticle General Concept of Buoyancy in Sedimentation and Ultra-
Superlattices. Nature 2006, 439, 55–59. centrifugation. Phys. Biol. 2013, 10, 045005.
8. Yethiraj, A.; van Blaaderen, A. A Colloidal Model System 31. Sentjabrskaja, T.; Babaliari, E.; Hendricks, J.; Laurati, M.;
with an Interaction Tunable from Hard Sphere to Soft and Petekidis, G.; Egelhaaf, S. U. Yielding of Binary Colloidal
Dipolar. Nature 2003, 421, 513–517. Glasses. Soft Matter 2013, 9, 4524–4533.
9. Joannopoulos, J. D.; Villeneuve, P. R.; Fan, S. H. Photonic 32. Voigtmann, T. Multiple Glasses in Asymmetric Binary Hard
Crystals: Putting a New Twist on Light. Nature 1997, 386, Spheres. Epl 2011, 96, 36006.
143–149. 33. Chen, M.; Cölfen, H.; Polarz, S. The Effect of Centrifugal
10. Wong, S.; Kitaev, V.; Ozin, G. A. Colloidal Crystal Films: Force on the Assembly and Crystallization of Binary Col-
Advances in Universality and Perfection. J. Am. Chem. Soc. loidal Systems: Towards Structural Gradients. Z. Natur-
2003, 125, 15589–15598. forsch., B: J. Chem. Sci. 2013, 68, 103–110.
11. Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Synthesis and 34. de las Heras, D.; Schmidt, M. The Phase Stacking Diagram
Characterization of Monodisperse Nanocrystals and Close- of Colloidal Mixtures under Gravity. Soft Matter 2013, 9,
Packed Nanocrystal Assemblies. Annu. Rev. Mater. Sci. 8636–8641.
2000, 30, 545–610. 35. Hansen, J. P.; I.R, M. Theory of Simple Liquids, 3rd ed.;
12. Holland, B. T.; Blanford, C. F.; Do, T.; Stein, A. Synthesis of Elsevier: Amsterdam, 2006.
Highly Ordered, Three-Dimensional, Macroporous Struc- 36. Benes, K.; Tong, P.; Ackerson, B. J. Sedimentation, Peclet
tures of Amorphous or Crystalline Inorganic Oxides, Phos- number, and Hydrodynamic Screening. Phys. Rev. E 2007,
phates, and Hybrid Composites. Chem. Mater. 1999, 11, 10.1103/PhysRevE.76.056302.
795–805. 37. Cross, M. C.; Hohenberg, P. C. Pattern-Formation outside of
13. Trau, M.; Saville, D. A.; Aksay, I. A. Assembly of Colloidal Equilibrium. Rev. Mod. Phys. 1993, 65, 851–1112.
Crystals at Electrode Interfaces. Langmuir 1997, 13, 6375– 38. Nielsen, M. H.; Aloni, S.; De Yoreo, J. J. In situ TEM Imaging
6381. of CaCO3 Nucleation Reveals Coexistence of Direct and
14. Trau, M.; Saville, D. A.; Aksay, I. A. Field-Induced Layering of Indirect Pathways. Science 2014, 345, 1158–1162.
Colloidal Crystals. Science 1996, 272, 706–709. 39. Strauss, H. M.; Karabudak, E.; Bhattacharyya, S.; Kretzschmar,
A.; Wohlleben, W.; Colfen, H. Performance of a Fast Fiber
15. Holgado, M.; Garcia-Santamaria, F.; Blanco, A.; Ibisate, M.;
Based UV/Vis Multiwavelength Detector for the Analytical
Cintas, A.; Miguez, H.; Serna, C. J.; Molpeceres, C.; Requena,
Ultracentrifuge. Colloid Polym. Sci. 2008, 286, 121–128.
J.; Mifsud, A.; Meseguer, F.; Lopez, C. Electrophoretic
40. Cölfen, H.; Laue, T. M.; Wohlleben, W.; Schilling, K.;
Deposition to Control Artificial Opal Growth. Langmuir
Karabudak, E.; Langhorst, B. W.; Brookes, E.; Dubbs, B.;
1999, 15, 4701–4704.
Zollars, D.; Rocco, M. The Open AUC Project. Eur. Biophys. J.
16. Denkov, N. D.; Velev, O. D.; Kralchevsky, P. A.; Ivanov, I. B.;
2010, 39, 347–359.
Yoshimura, H.; Nagayama, K. 2-Dimensional Crystalliza-
tion. Nature 1993, 361, 26–26.
17. Denkov, N. D.; Velev, O. D.; Kralchevski, P. A.; Ivanov, I. B.;
Yoshimura, H.; Nagayama, K. Mechanism of Formation of
2-Dimensional Crystals from Latex- Particles on Sub-
strates. Langmuir 1992, 8, 3183–3190.
18. Jiang, P.; Bertone, J. F.; Colvin, V. L. A Lost-Wax Approach to
Monodisperse Colloids and Their Crystals. Science 2001,
291, 453–457.
19. Jiang, P.; Bertone, J. F.; Hwang, K. S.; Colvin, V. L. Single-
Crystal Colloidal Multilayers of Controlled Thickness.
Chem. Mater. 1999, 11, 2132–2140.

CHEN ET AL. VOL. 9 ’ NO. 7 ’ 6944–6950 ’ 2015 6950

www.acsnano.org