Upload
jun-xia
View
214
Download
1
Embed Size (px)
Citation preview
Self-Assembly of Two Agents in a Core-Shell-Corona
Multicompartment Micelle Studied by Dissipative
Particle Dynamics Simulations
Jun Xia, Chongli Zhong*
Department of Chemical Engineering, The Key Lab of Bioprocess of Beijing, Beijing University of Chemical Technology,Beijing 100029, P. R. ChinaFax: þ86-10-64419862; E-mail: [email protected]
Received: June 14, 2006; Revised: July 25, 2006; Accepted: July 26, 2006; DOI: 10.1002/marc.200600411
Keywords: block copolymers; computer modeling; density distribution; dissipative particle dynamics; multicompartmentmicelle
Introduction
Multicompartment micelles, introduced about ten years
ago, can provide interesting morphologies for nanotechno-
logy.[1–3] To date various multicompartment micelles have
been designed and investigated,[3–8] and increasing atten-
tion is being paid to their potential applications as nano-
containers to solubilize two or more incompatible agents
within separate nanoscopic compartments.[9] The core-
shell-corona micelle is a simple morphology of multi-
compartment micelles that can be formed from linear ABC
triblock copolymers. These ‘three-layer’ micelles have
been used in the development of new nanotechnological
applications, such as drug delivery and catalysis.[10–12]
To date, investigations on the simultaneous storage of two
kinds of hydrophobic agents in core-shell-corona micelles
are very scarce, to which molecular simulation is a powerful
tool that can give insight into molecular-level details of the
extent and the locus of solubilization. However, to date, no
such simulation studies have been performed.
To have a molecular understanding of the solubilization
of two or more agents in multicompartment micelles, the
Summary: Dissipative particle dynamics simulations areperformed on the distributions of two agents in a core-shell-corona multicompartment micelle. The simulated resultsshow that when the agents are weakly hydrophobic, theirdistributions in the multicompartment micelle are largelyaffected by the interactions between the agents and theblocks; while for strongly hydrophobic agents, the self-assembly of solubilized species in the micelle is also affected
largely by the interactions between the species. This workconfirms that a multicompartment micelle can store twoagents within separate nanoscopic compartments simulta-neously, and shows that the distributions of the agents can betailored easily by changing the interactions presented. Thisprovides molecular-level information that is useful for thefuture rational design of new micellar systems with tailoredproperties.
Simulated cross sections of the multicompartment micelles with strongly hydrophobicsolubilized agents (the solvent and block A are omitted for clarity, block B is dark gray,block C is light gray, agent P is white, and agent Q is black).
Macromol. Rapid Commun. 2006, 27, 1654–1659 � 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1654 DOI: 10.1002/marc.200600411 Communication
dissipative particle dynamics (DPD) technique is employed
here to study the self-assembly of two agents in a core-shell-
corona multicompartment micelle, to which particular
attention is paid to the influences of the interactions bet-
ween the agents as well as between the agents and the
blocks.
Experimental Part
Dissipative Particle Dynamics (DPD) Method
The DPD method, originally developed by Hoogerbrugge andKoelman,[13,14] is a mesoscopic simulation technique that canbe used to study systems over greater length and time scalesthan are accessible in classical molecular dynamics and MonteCarlo simulations, and this larger scale simulation is veryimportant for investigating the behavior of polymers. TheDPD method has been successfully used to study the micro-structures and properties of polymers in the bulk state and insolvent.[15–24] Details of the DPD method are given by Grootand co-workers.[15,25]
Models and Parameters
In this work, multicompartment micelles formed from linearABC triblock copolymers in water were studied. Block A wasdefined to be hydrophilic, block B weakly hydrophobic, andblock C strongly hydrophobic. The ABC copolymer chain wasmodeled as a 10þ 8þ 2 spring-bead chain that was denotedA10B8C2. The two hydrophobic agents, denoted as P and Q, aswell as the solvent (denoted as S) were modeled as single DPDbeads. Furthermore, agent P was compatible with block B,while agent Q was compatible with block C.
There were six components in the system, and thus there werea total of 15 pairs of DPD parameters (aij) to describe theinteractions between the components in the system. As a result,some parameters had to be fixed and the study focused on theeffects of the other ones. In this work, the influence of theinteractions between the agents as well as between the agentsand the blocks were of interest. Therefore, the interactionsbetween the blocks and those between the blocks and the solventwere fixed. This was equivalent to studying the assembly ofvarious agents in a given micelle. The DPD parameters for themwere selected asaAS¼ 25,aBC¼ 60,aAB¼ aBS¼ 40, andaAC¼aCS¼ 90 to make the A10B8C2 copolymers assemble into core–shell–corona multicompartment micelles in water. Since agentP was compatible with block B, and agent Q was compatiblewith block C, aBP¼aCQ¼ 25 was set. According to the study ofGroot and Warren,[25] the bead density in the simulations can beeither 3 or 5. It was set to 3 in this work so that less computationalefforts were required. To make sure this selection did not affectthe simulation results, some parallel simulations using a beaddensity of 5 were carried out to demonstrate that the choice ofbead density did not affect the main results and the conclusionsderived. For a bead density of 3, the repulsion parameterbetween like particles was aii¼ 25 according to the work ofGroot and Madden.[15]
Simulation Details
The DPD simulations were performed in a cubic cell of size25� 25� 25 rc
3, which contained a total of about 47 000 DPDbeads. Periodic boundary conditions were applied and thevolume fraction of the copolymer was set to be 0.1 to ensure thesystem could form enough micelles. The volume fractions ofthe two hydrophobic agent beads were identical and they wereboth set to 0.01. For convenience, the cut-off radius rc, theparticle mass m, and kBTwere all taken as unity. The time stepDt was taken as 0.05, and adjacent particles in the polymerchain interacted via a linear spring with a harmonic springconstant of 4.0. The number of DPD steps carried out for a DPDsimulation in this work was (2–4)� 105, and depended on thesystem concerned as previously reported.[16–19]
Results and Discussion
Validation of the DPD Method
Although the DPD technique has been used successfully to
describe the formation of multicompartment micelles from
star triblock copolymers[18] and linear pentablock copoly-
mers in water,[19] as well as the cooperative self-assembly
of nanoparticle mixtures in block copolymers[17] in our
previous work, it is further validated for describing solubili-
zation in micelles by comparison with the Monte Carlo
(MC) simulations of Xing and Mattice[26] in this work. They
investigated the solubilization of insoluble agents (IA) in
A5B10A5 micelles where the volume fraction of solubilizate
varied from 0.0078 to 0.125, the polymer volume fraction
was 0.04, and a total of 128 chains were considered. Their
simulations show that a transition of a micellar structure
into a droplet microemulsion occurs as more IA is added to
the system. Our DPD simulations can reproduce their
observations under similar conditions. This, together with
our previous work,[17–19] indicates that the DPD technique
is reliable for studying the self-assembly of agents in multi-
compartment micelles.
Self-Assembly of Two Agents inMulticompartment Micelles
Here the focus is on the investigation of the effects of the
interactions between the agents as well as between the
agents and the blocks on the self-assembly of the two agents
in multicompartment micelles. Even so, there are a total of 7
pairs of DPD parameters to be considered. Therefore, case
studies must be performed, and two such cases are explored:
one is the case of weakly hydrophobic agents (weak solubi-
lization), and the other that of strongly hydrophobic agents
(strong solubilization).
Case 1. Weakly Hydrophobic Agents
Aweakly hydrophobic agent is defined here by setting their
interactions with the hydrophilic species (block A and
Self-Assembly of Two Agents in a Core-Shell-Corona Multicompartment Micelle Studied by . . . 1655
Macromol. Rapid Commun. 2006, 27, 1654–1659 www.mrc-journal.de � 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
water) as equal to that of the interactions of the weakly
hydrophobic block B with the hydrophilic species, thus,
aAP¼aPS¼ aAQ ¼ aQS¼ 40. To study the effects of the
interactions between the solubilizates and the blocks, two
series of simulations are performed: one is the case that the
repulsion strength between the two solubilizates is weak
(aPQ¼ 30), and the other is the case that strong repulsion
exists between the two agents (aPQ¼ 60). The repulsion is
varied between the two agents and the incompatible blocks
from 30 to 60, that is, aBQ¼ aCP¼ 30–60.
The simulated density profiles for the weak repulsion of
the two agents (aPQ¼ 30) and aBQ¼ aCP ¼ 30 (also weak
repulsion between the two agents with the incompatible
blocks) are shown in Figure 1a. Obviously, agent Q distri-
butes throughout block C with slight accumulation at the
interface of blocks B and C. Agent P, on the other hand,
distributes throughout block B with preferential occupation
of the center of the shell formed by block B. A direct
visualization of the simulated multicompartment micelle
with the two agents is given in Figure 1c. When the repul-
sion between the agents and the incompatible blocks
is increased to 60, that is aBQ¼ aCP¼ 60, although agent
Q still distributes throughout block C, it accumulates in the
center of the core formed by block C (Figure 1b and 1d)
instead of the interface as before (Figure 1c). The distri-
bution of agent P, however, is not largely influenced,
although agent P is no longer observed in block C because
of the strong repulsion.
Furthermore, DPD simulations of a strong repulsion be-
tween the two agents (aPQ¼ 60) are performed by varying
aBQ¼ aCP from 30 to 60. The simulated density profiles are
nearly identical to those of weak repulsion between the two
agents (aPQ¼ 30). It seems for weakly hydrophobic agents
(weak solubilization) that the localization of the solubili-
zates in multicompartment micelles is more significantly
affected by the chemistries of the blocks, and the effects of
the chemistries of the agents are less important because of
the small quantities of the solubilizates.
Case 2: Strongly Hydrophobic Agents
The self-assembly of strongly hydrophobic agents is investi-
gated by setting the interactions of the two agents with the
hydrophilic species equal to that of the strongly hydrophobic
Figure 1. a,b) Simulated density profiles for weakly hydrophobic agents solubilized in a core–shell–corona multicompartment micelle. c,d) Simulated cross sections of the multicompartmentmicelles with solubilized agents. (c) corresponds to (a) and (d) corresponds to (b) (the solvent andblock Awere omitted for clarity, block B is dark gray, block C is light gray, agent P is white, and agentQ is black).
1656 J. Xia, C. Zhong
Macromol. Rapid Commun. 2006, 27, 1654–1659 www.mrc-journal.de � 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
block C with the hydrophilic species, that is, aAP¼ aPS¼aAQ¼ aQS¼ 90. Again, two cases are studied: weak (aPQ¼30) and strong (aPQ¼ 60) repulsion between the two agents.
The simulated density profiles for aPQ¼ aBQ¼ aCP¼ 30,
aPQ¼ 30, and aBQ¼ aCP¼ 60 are shown in Figure 2a and b,
respectively. Figure 2a shows that agent Q distributes uni-
formly in block C, while agent P concentrates more
preferably at the interface of blocks B and C. This is quite
different from Figure 1a, where weakly hydrophobic agents
are considered, which indicates that the hydrophobicity of
the solubilizates significantly affects their distributions in
multicompartment micelles. When the repulsion strength
between the two agents and the incompatible blocks is
increased to aBQ¼ aCP¼ 60, as depicted in Figure 2b, the
concentration of block C in the center of the core decreases
with an increase in concentration of agent Q in the center.
Meanwhile, for agent P, the change in distribution is not
large, which is similar to the case of the weakly hydro-
phobic agents. A comparison of Figure 1b and Figure 2b
shows that the distribution of agent Q is more sensitive to
the interactions within the micellar system. To have a direct
visualization of the structures of the multicompartment
micelles with agents, the cross sections of the micelles
corresponding to Figure 2a and b are shown in Figure 3a and
b, respectively.
DPD simulations for a strong repulsion between the two
agents (aPQ¼ 60) have been further performed by varying
aBQ¼ aCP from 30 to 60. In the case of aBQ¼ aCP¼ 30, as
shown in Figure 2c, agent Q concentrates more in the center
of the core, a feature that is quite different from the distri-
bution shown in Figure 2a. This illustrates that for strongly
hydrophobic agents, the repulsion between the agents also
influences their distributions, possibly mainly because of
the large quantities of solubilizates present. Although agent
P also tends to accumulate in the center of block B, the
change in distribution is not significant as compared to the
one in Figure 2a.
The DPD simulation results for aBQ¼ aCP¼ 60 are
shown in Figure 2d. Obviously, by increasing the repulsion
strength between the agents and the incompatible blocks,
agent Q accumulates strongly in the center of the core with
C blocks being pushed away from the center. This illustrates
that in the case of strong solubilization, a transition of
a micellar structure into a droplet microemulsion may
occur in multicompartment micellar systems like the
one observed in normal micellar systems by Xing and
Mattice.[26] As shown in Figure 2d and 3d, in the case of
strong repulsion between the agents and the incompatible
blocks, agent P also tends to concentrate in the center of the
shell, that is, both agents can accumulate in the centers of
the compatible blocks.
The above simulations show that for strongly hydro-
phobic agents, in contrast to case 1, the self-assembly of the
agents are affected largely by both the interactions between
the agents and that between the agents and the incompatible
blocks.
Figure 2. Simulated density profiles for strongly hydrophobic agents solubilized in a core–shell–corona multicompartment micelle.
Self-Assembly of Two Agents in a Core-Shell-Corona Multicompartment Micelle Studied by . . . 1657
Macromol. Rapid Commun. 2006, 27, 1654–1659 www.mrc-journal.de � 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Evolution of a Multicompartment Micelle withTwo Agents Solubilized
It is both interesting and important to reveal the formation
process of multicompartment micelles with solubilized
agents, which can be realized by examining the snapshots of
the density profiles of the evolution process. In this work,
the evolution of the multicompartment micelle that corres-
ponds to Figure 3d is adopted as an example, and the
corresponding simulated snapshots of density profiles are
shown in Figure 4. Obviously, at the beginning all the
hydrophobic components aggregate in the central area of
the ‘nucleus’, whereas the solvent in the central region of
the nucleus remains high in density and the hydrophilic
block (block A) has a low density in the central region
(t¼ 25). The B blocks are then pushed away from the center
Figure 3. Simulated cross sections of the multicompartment micelles with solubilizedagents corresponding to Figure 2 (the solvent and block Awere omitted for clarity, block B isdark gray, block C is light gray, agent P is white, and agent Q is black).
Figure 4. Evolution process of the multicompartment micelle that corresponds to Figure 2d and 3d.
1658 J. Xia, C. Zhong
Macromol. Rapid Commun. 2006, 27, 1654–1659 www.mrc-journal.de � 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
by C blocks and agent Q to form the inner shell of the micelle
(t¼ 100), and the block B compatible agent (agent P) moves
to the shell with further evolution (t¼ 250). During this pro-
cess, water molecules are pushed away from the center and
the hydrophilic blocks gradually form the corona. Further-
more, agent Q pushes C blocks away from the center of the
core to accumulate in the center, and C blocks form a shell-
like structure that surrounds the accumulated agent Q with a
low concentration in the center to form the structure shown in
Figure 3d.
Conclusion
The DPD simulations show that a multicompartment
micelle can store two or more agents within separate nano-
scopic compartments simultaneously, and the distributions
of the agents can be tailored by changing the interactions
between incompatible species. For weakly hydrophobic
agents, the interactions between the agents and the blocks
play an important role, while for strongly hydrophobic
agents, the interactions between the agents also has a large
influence. The simulated results elucidate the relation be-
tween the distributions of agents in a multicompartment
micelle to the interactions present in the system, which
provides useful information for developing new nanocon-
tainers that may deliver two or more agents simultaneously.
Acknowledgements: The financial support of the NSFC(20476003) and the Specialized Research Fund for the DoctoralProgram ofHigher Education of China (20040010002) are greatlyappreciated.
[1] A. Laschewsky, Curr. Opin. Colloid Interface Sci. 2003, 8,274.
[2] J.-F. Lutz, A. Laschewsky, Macromol. Chem. Phys. 2005,206, 813.
[3] S. Kubowicz, J.-F. Baussard, J.-F. Lutz, A. F. Thunemann,H. von Berlepsch, A. Laschewsky, Angew. Chem. Int. Ed.2005, 44, 5262.
[4] Z. Li, E. Kesselman, Y. Talmon, M. A. Hillmyer, T. P. Lodge,Science 2004, 306, 98.
[5] Z. Li, M. A. Hillmyer, T. P. Lodge, Macromolecules 2006,39, 765.
[6] Z. Zhou, Z. Li, Y. Ren, M. A. Hillmyer, T. P. Lodge, J. Am.Chem. Soc. 2003, 125, 10182.
[7] S. Kubowicz, A. F. Thunemann, R. Weberskirch, H. Mohwald,Langmuir 2005, 21, 7214.
[8] A. F. Thunemann, S. Kubowicz, H. von Berlepsch, H.Mohwald, Langmuir 2006, 22, 2506.
[9] T. P. Lodge, A. Rasdal, Z. Li, M. A. Hillmyer, J. Am. Chem.Soc. 2005, 127, 17608.
[10] J.-F. Gohy, N. Willet, S. Varshney, J.-X. Zhang, R. Jerome,Angew. Chem. Int. Ed. 2001, 40, 3214.
[11] Y. Tang, S. Y. Liu, S. P. Armes, N. C. Billingham,Biomacromolecules 2003, 4, 1636.
[12] R. S. Underhill, G. Liu, Chem. Mater. 2000, 12, 3633.[13] P. J. Hoogerbrugge, J. M. V. A. Koelman, Europhys. Lett.
1992, 19, 155.[14] J. M. V. A. Koelman, P. J. Hoogerbrugge, Europhys. Lett.
1993, 21, 363.[15] R. D. Groot, T. J. Madden, J. Chem. Phys. 1998, 108,
8713.[16] D. Liu, C. Zhong, Macromol. Rapid Commun. 2005, 26,
1960.[17] D. Liu, C. Zhong,Macromol. Rapid Commun. 2006, 27, 458.[18] J. Xia, C. Zhong, Macromol. Rapid Commun. 2006, 27,
1110.[19] D. Liu, C. Zhong, unpublished results.[20] H.-J. Qian, Z.-Y. Lu, L.-J. Chen, Z.-S. Li, C.-C. Sun,
Macromolecules 2005, 38, 1395.[21] S. Yamamotoa, S.-a. Hyodo, J. Chem. Phys. 2005, 122,
204907.[22] S. G. Schulz, H. Kuhn, G. Schmid, C. Mund, J. Venzmer,
Colloid Polym. Sci. 2004, 283, 284.[23] P. Prinsen, P. B. Warren, M. A. J. Michels, Phys. Rev. Lett.
2002, 89, 148302.[24] L. Rekvig, M. Kranenburg, J. Vreede, B. Hafskjold, B. Smit,
Langmuir 2003, 19, 8195.[25] R. D. Groot, P. B. Warren, J. Chem. Phys. 1997, 107,
4423.[26] L. Xing, W. L. Mattice, Langmuir 1998, 14, 4074.
Self-Assembly of Two Agents in a Core-Shell-Corona Multicompartment Micelle Studied by . . . 1659
Macromol. Rapid Commun. 2006, 27, 1654–1659 www.mrc-journal.de � 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim