Romanian version

SIM-NANOPORE

Molecular simulations for penetrable fluids confined in nanopore systems

Program Capacities/Module III – Bilateral cooperation  462 / 18.02.2011

Type of project:                Bilateral

Partner countries:              Romania – South Korea

Period of collaboration:     March 2011 – December 2011

PARTNERS

Coordinators:
Dr. Viorel Chihaia Institute of Physical Chemistry “Ilie Murgulescu”, Romanian Academy Splaiul Independentei 202, 060021, Bucharest Phone (+40) 021-316-7912 e-mail: vchihaia@icf.ro	Prof. Soong-Hyuck Suh Keimyung University, Department of Chemical Engineering #1,000 Shindang-dong, Dalseo-ku, Daegu, 704-701, Republic of Korea e-mail: shsuh@kmu.ac.kr
Teams:
Dr. Popa Vlad	Prof. Jae-Young Bae
Dr. Munteanu Gabriel	Prof. Kwan-Kyu Park
Dr. Stanica Nicolae	Prof. Young_A Son
Dr. Rares Scurtu	Prof. Tae-Young Kim
Dr. Dascalu Myrella Izabella

PROJECT GENERAL OBJECTIVE

Science is probably one of the first trades globalized over the world. Regardless of nationality, scientists can make cooperative work successfully based on their common research interests and complementary expertise. The so-called ‘mobility’ program can be a good example. In this direction, our general aim may be described into one goal: to develop the effective research network between Korea and Romania. We believe that broad visions for constructing the bilateral research networking would be one of the main keys to the success of this KICOS/ANCS research program.

Contrary to other groups, our participating teams in Keimyung University, Korea, and Institute of Physical Chemistry, Romania, have the relatively long history of mutual relationships. In fact, two research groups between Prof. Suh and Dr. Chihaia have started their exchanging activities over the last decade since 1999. The implementation of a Blade Cluster in the Institute of Physical Chemistry, within the frame of the project CAPACITATI PNI-II 84/2007 founded by ANCS, is an important result of the two groups collaboration, were Prof. Suh helped the Romanian group with his knowledge about High Performance Computation Centers.

Furthermore, starting form 2005, we have organized and served as the representative committee members for the Korea/Romania Joint Workshop: Molecular Science and Engineering, which has been held every odd-year in Korea and even-year in Romania, respectively. This workshop endeavors to provide an effective bilateral forum between Korea and Romania in the areas of chemistry, physics, molecular science, and its related engineering. In addition to the Korea/Romania collaborations, we are currently conducting cooperative research work with other EU countries including Spain, German, Hungary, and Bulgaria. We are also aiming to creating the EU-FP7 research project throughout our successful research partnerships to be made in this program.

PROJECT DESCRIPTION

Soft materials are materials with physical states that are characterized by an energy scale comparable with room temperature thermal energy. In spite of the various forms of these materials (liquids, colloids, surfactants, polymers, foams, gels, microemulsions, granular materials, membranes and some of biological materials), many of their properties have common physicochemical origins, such as a large number of internal degrees of freedom, weak interactions between structural elements, and a delicate balance between entropic and enthalpic contributions to the free energy. Such systems have a complex structure that is difficult to be described directly from its atomic or molecular consituents. The constituent mesoscopic particles, may contain thousands or even millions of atoms, interconnected to one another in complicated ways. These systems self-organizes very often into mesoscopic physical structures (with typical sizes 1 nm-1m) that are much larger than the microscopic scale, and yet are much smaller than the macroscopic (overall) scale of the material. They come in a variety of forms, such as linear chain, branched, star-shaped, dendritic, and copolymer, as well as taking on a variety of functions, e.g., as steric stabilizers, additives, and depletants. Additional flexibility arises from it being possible to influence the structural and phase behaviour of polymer solutions by changing the solvent quality.

The properties and interactions of these mesoscopic structures may determine the macroscopic behavior of the material. Their high flexibility constitutes a major advantage of soft-matter systems in comparison to their atomic counterparts is that one can engineer the constituent particles at the molecular level. In this way, an enormous variety of architectures can be achieved, leading to a corresponding richness in the structural and phase behavior of such systems. Typically, soft materials have weak interactions among molecular or supramolecular components and are often either amorphous or can self-assemble from the liquid state. There are often many levels of complexity with heirarchical, supramolecular structures that can be cooperative and far from equilibrium.

Because computational modeling and simulations are not constrained by the current limit of our ability to fabricate the specific nanostructures, it can explore and examine the potential targets for the design and construction of a wide range of atomically defined nanostructures, which can be a fundamental goal for nanotechnology. The underlying theme in this project is to develop molecular simulation approaches which accurately describe nanostructured pore systems of interest. We will employ various molecular simulation methods together with statistical thermodynamic approaches to determine the detailed molecular properties, which are often difficult or impossible to measure experimentally. Comparison with experimental data can be used to investigate model nanopore systems, and the ultimate goal is to use molecular simulations to carry out experiments that cannot be undertaken in the real laboratory.

        We are most often concerned with the structural arrangments, viscoelastic rheology, and/or mechanical behavior of these materials. The properties of such materials may be diferent when they are confined inside cavities or porous materials and this is our main target of the present project to find informations about the soft materials inside nanoporous sistems. The experimental investigation of such confined sistems is dificult. Computer simulations have also played a very important role in understanding the conformations of confined mesoscopic particles, where the role of short-range attractions appears to play a minor role in the overall behavior of the macroscopic systems. The investigation asisted by computer numerical simulations of the phenomena and the characterization of the properties of the soft materials inside nanoporous sistems is the main goal of our project. For the various mesoscopic particles architectures the effective interactions between suitably chosen coordinates are shown to be ultrasoft, i.e., they either remain finite or diverge very slowly at zero separation. As a consequence, the fluid phases have unusual characteristics, including anomalous pair correlations and mean-field like thermodynamic behaviour.

        From this point of view, as an intermediate between theory and experiment, the equilibrium and transport properties obtained from molecular simulations for confined penetrable fluids can provide an invaluable tool against statistical thermodynamic approximations in the literature. We will also focus on the development of algorithms and computing methods for the structural characterization of a given system based on the interatomic connectivity, e.g., clustering, layering, ring-shaped topology, cage/channel formation, etc. Besides such structural criteria, another methodological attention will be given to the algorithm development for the characterization of dynamic properties. A special atention will be payd to the topologic characterization of the hydrogen bonded large periodic systems.

        The equilibrium and transport properties of soft-condensed systems via Monte Carlo and molecular dynamics simulations will be important for the molecular design of nanoporous materials and applications with an aid of molecular-based computer simulations. In addition to such classical computations, specific model potentials employed for nanoporous systems will be parameterized by comparing the experimental data with the simulation data obtained from semiempirical and ab-initio quantum calculations.

        Nanoscience and nanoengineering can be the dominant direction for cutting-edge technologies in this new century. In this field of nanotechnology, it often requires research collaborations on the part of many specialists to gain detailed perspective approaches, just as fashion designers rely on weavers, tailors, and cutters to accomplish new designs. It is therefore desirable to promote bilateral research work to reach optimal solutions for nanotechnology. This situation will open the possibility for interconnections between various topics in nanotechnology to bring about new chances for further applications.

        Prof. Soong-Hyuck Suh’s research interests are associated with statistical mechanical theories and molecular-based computer simulations, with particular emphasis on the equilibrium and transport properties of fluids and mixtures confined within model nanopore systems. His work in this project is a continuation of his current research work on model nanopore systems at the atomic and molecular level, where theoretical and computational approaches have demonstrated that such a route to molecular science and engineering is quite viable. Research areas in this direction can open the way to the effective design and the improved fabrication for nanoporous and nanocomposite materials to be used in advanced nanotechnologies. He implemented ultrasoft and bounded potentials in a simulation code developed in his laboratory, enabling the users to increase the complexity of systems that can be modeled in the fields of nanomaterials, molecular crystals, polymers and organic chemistry. His software provides a coarse-grained approach that allows the combination of soft and hard materials in a single calculation.

        Dr. Chihaia has developed various computational algorithms. His studies in this direction include the rigid-body dynamics, the optimal searching routine of rings and cages with arbitrary size based on the atom-to-atom connectivity, the hydrogen bond topology, and the automatic building of the potential energy surface for the adsorption and diffusion phenomena. Recently, he has initiated and pioneered the research project related to the high computing system in Romania, entitled as the High Performance Computer Cluster for Computational Material Science in Romania. He has experiences on both the quantum and the classical molecular simulations, e.g., the adsorption and diffusion of molecules and small clusters on various surfaces, the characterization of the structural and dynamic features of the channel- and cage-like compounds (carbon nanotube, porous silicon, buckyball, water clusters and clathrate hydrates) using quantum chemistry and solid states physics, and the embedding methods for the extended atomic and molecular systems.

        The two groups are complementing each other by their abilities, knowldges and their infrastuctures. The results of their collaboration demonstrate the efficiency of their work together. Both countries of Korea and Romania, based on our experience of previous research cooperation, have their own strength and advantage: the fundamental science to the Romanian side and the engineering application to the Korean side. In order to be globally competitive, it need to have feedbacks and interactions among scientists and engineers in both countries. Theoretical and computational methods to be investigated in this work will play a leading role in the development of nanosystems, guiding the fundamental issues in molecular design and modeling.

REZULTS

To test the installation codes so we chose the system soft materials (soft materials) that are material physical state characterized by energy with corresponding heat values ​​close to room temperature. Despite various forms of these materials (liquids, colloids, surfactants, polymers, foams, gels, microemulsions, granular materials, membranes and some biological matriale) many of their properties have common physicochemical origins, such as a large number of degrees of freedom, weak interactions between structural elements and a delicate balance between entropy and enthalpy contributions to their free energy. Such systems have a complex structure difficult to describe directly the atomic or molecular their constituents. Constituent particles may contain thousands and even millions of atoms, interconnected in a complicated way. These systems self-organize into mesoscopic physical structures (typical size between 1 nm to 1μm) are higher as microscopic scale but much smaller than macroscopic scale the material. These meso-objects have various forms such as linear strings, branched type star, and dendritic copolymers. They have a variety of functions such as steric stabilizers, additives and dispersion.

Their flexibility can influence the structure and polymer solutions comporatarea by changing the solvent quality. Structural properties and interactions of these can cause mesoscopic macroscopic behavior of the material. Their high flexibility is an advantage compared to their opponents because they allow manipulation of their atomic constituent particles at the molecular level. This can get a wide variety of architectures, with different behaviors.
We list some examples of soft materials: liquids, colloids, polymers, foams, gels, granular materials as well as more complex systems such as bacterial suspensions and biological polymers. Usually soft materials have weak interactions between molecular or supramolecular components and are usually amorphous form or may be self-assembled from licide states. Often these materials has many levels of complexity with hierarchical supramolecular structures that can be competitive in states far from equilibrium. Of great interest to our teams surfactant systems, particularly those formed by sodium sulfate. Sodium dodecyl sulfate (SDS - Sodium Sulfate DODECYL) is an anionic detergent and distorting proteins that confer a negative electrical charge SDS-protein complex, which allows us to control migration of protein molecular function of their mass and their electrical charge. It is very useful in separating nano-objects such as graphene, nanotubes or fullerenes and nano-systems formation control and meso-porous. Below we present results obtained in model systems containing it some water solutions - SDS.

Sodium dodecyl sulphate - SDS

Sodium dodecyl sulfate is an organic compound with the formula NaC12H25SO4 or CH3 (CH2) 11OSO3Na, consisting of a chain of 12 carbon atoms attached to a sulfate group SO3 through one of oxygen. It is an anionic surfactant. To characterize the energy stability of this system, we performed DFT calculations using quantum type and exchange correlation potential PW91 GGA type and a numerical DNP basis. Conditions for achieving convergence were very strict geometric optimization: 10-6 Ha energy, energy gradient 10-3 Ha / A, 0.005 A. The energy shift of about sodium atom chain of CH3 (CH2) 11OSO3 is given by relationship

BE (Na) = E (CH3 (CH2) 11OSO3Na) - [E (CH3 (CH2) 11OSO3-) + E (Na +)]

where E (CH3 (CH2) 11OSO3Na) is the energy of the whole system, E (CH3 (CH2) 11OSO3-) is the energy chain CH3 (CH2) 11OSO3-and E (Na +) is the energy of ion sodium alone (104 015 kcal / mol). The most stable conformers of CH3 (CH2) 11OSO3-is presented in Table 1. Sodium ion binds to the sulfate group in various positions at a distance S-Na of about 2.72 A. The most stable conformers is called CH3 (CH2) 11OSO3Na, Na - down in Table 1.

Table 1. Various associated structures and their energies SDS system connectors (BE).

If dimers isolated can be seen as the most stable isomer is the called Dimer Head-Head Parallel 1 for the two molecules SDS are parallel with the same orientation and the two ions Na + are located on one side and the other to two sulfate groups. It is possible that immersion in water system to modify the polarization of SDS molecules which would affect the stability of isomers but do not expect to change the location and orientation of two molecules parallel. An abinitio study would be very expensive due to the large number of molecules and configurations that should be investigated. SDS molecules form micelles that water to concentrations higher than a critical value suggests a parallel arrangement of molecules preferentially SDS.

                Coarse Grain parameterization of force fields for SDS systems, water and allotropic forms of carbon

    Potential energy surface for fields of type d effort coarse grain is much smoother than those of all-atoms (AA). The main requirement of these fields is to reproduce the overall movement of atomic groups that are represented by pseudoparticulele beads. CG potential is usually parameterized to reproduce the radial distribution function of different pairs of elements in the system or by reversing the equation molecular dynamics simulations Boltzann of AA. Therefore these potential parameters depend on the molecular dynamics simulations run all-atoms including the time step used for integrating the equation of motion and system temperature. CGMD simulation time step can be considered for 5 to 20 times higher than that used for AA simulations, allowing the study of phenomena that occur at larger time scales. Also reduces the number of particles between the forces of interaction is calculated and whose equation of motion is integrated. Thus the calculation effort is much reduced, allowing the study of larger systems and larger stroke. On the other hand, CG forces and potentials have mathematical forms simpler, generally short distance, usually because the potentials are truncated. A class of potential high-use is the potential to parameterization Martini.

    These potential interactions consist of connection type, usually with a quadratic form of the difference distance between two particles and the value of the equilibrium distance and interactions of nelegatura, usually represented by Lennard-Jonnes potential truncated. In this study we have taken from literature force field parameters of type GC Martini 2.0 [SJ Marrink, H.J. Risselada, S. Yefimov, D.T. Tielman and A.H. de Vries, J.Phys.Chem.B, 111, 2007, 7812] and I used to interesting systems after a consistency check parameters. Thus we performed molecular dynamics calculations of GC systems in Table 2. The potential Martini four water molecules are represented by a single particle pseudo-P4, the SDS molecule a fragment of four carbons and 4 or 5 hydrogen is represented by a particle C2, SO4-sulfate group is represented by a particle Qa charged -1, and sodium ion by a charged particle Qd; in carbon systems, two linked carbon atoms are represented by a particle sc4. For GC systems of water tend to freeze at room temperature, we replaced BP4 particles between 5 and 10% of P4 particles. They are designed to disrupt the arrangement of water molecules in ordered structures and prevent,''freezing system.

    We built two amorphous systems consisting of water and SDS a total of 20 000 molecules in representing AA and carbon systems presented in Table 2. I turned AA systems in CG equivalent systems by replacing groups of atoms with their corspondentii CG. To reduce internal tensions have optimized CG structures mantinand first fixed shape and size of cells and then I completely relaxed systems. Simulations were performed in NVT thermodynamic ensembles for a temperature of 300 K, maintained by a Nose thermostat with a parameter Q = 100. They tried different time steps, the optimal for all systems is 25 fs, which is an increase in the time scale of about 50 times the AA simulations. Systems were balanced for 20 ns and thermodynamic values ​​were averaged from the other 10 ns. The results were compared with those obtained from AA simulations for a time step of 0.5 fs for a period of equilibration of 200 ps. To see a satisfactory reproduction of radial distribution functions. Simulations were repeated for NPT ensembles, under a pressure of 1 atm using a barometer Berensen. Simualrile network constants obtained in AA are reproduced in a 5% margin of error in CG computations.

Table 2. CG representation of the atomic systems of interest in this study

SDS micelles immersed in water

I checked reproduction density values of pure water and water-SDS solution by molecular dynamics simulations for 20 ns GC, using a statistical ensemble under the NPT (T = 300K, P = 1 atm, Berendsen barostat). Initial system density was 2 g/cm3 and or obtained after a period of 500 ps equilibrium densities of 1 to 2 g/cm3. In water-SDS solution of SDS molecules form agglomerations in which the molecules are oriented with sodium ions in the direction changed in comparison. From other studies in literature show that for more than a critical value of SDS concentration, these molecules are quasi-spherical clusters gather the SDS molecules are aligned radial ends where ions Na + are present, outside clusters. These clusters are stable over time and are in fact micelles.

Figure 2. Systems used in the study of reproductive density of pure water and water-SDS solution.

Figure 3. The structure of micelles of SDS immersed in water. Pentriu clarity water molecules are not represented.

For a sodium concentration of 2 to 3 molecules of water molecules have formed a stable Miceli remains almost spherical after 50 ns of simulation for a CG NVT thermodynamic ensemble. T = 300K (see Figure 3). Using the balanced structure we have built with double SDS Miceli a dimer structure Miceli-SDS. Initial distance between the centers of gravity of the two micelles is 8.46 nm. I relaxed for 1ns system thus formed a NPT ensemble (T = 300K, P = 1 atm, Berendsen barostat). Note that at the end of the simulation the two micelles is close to an average distance of 5.60 nm (see Figure 4). The two micelles change their shape slightly, initially spherical. We continued the simulation in NVT ensemble by fixing the size of simulation box obtained in NPT simulations and keeping the other conditions of previous CG simuarii. Note that after 5ns micelles elongate and two-and change the orientation, approaching the two ends. After about 9.5 ns distance between the two endings is reduced and more SDS molecules and diffusion occurs from Miceli to another. A migration of molecules from these SDS micelles one of two in favor downsize their other micelles (after about 15 ns). After 25 ns, no diffusion is observed between the two molecules SDS micelles. The two micelles holds its shape and orientation of the end of the 40 ns simulation. It is important to mention that during these simulations SDS molecules retains most of the time orienttarea with hydrophilic end (Na +) outward toward the water.

Figure 4. Evolution of two micelles system immersed in water in molecular dynamics simulations in NPT and NVT thermodynamic ensembles. Pentriu clarity water molecules are not represented.

Diffusion through amorphous nanoporous carbon

Sieves with nanometer size windows are very important in selectivity and production of ordered systems, playing in their preparation mask. We assembled in the same cell of two layers of graphene simulation, a layer of water-SDS solution and an amorphous carbon mesh windows / average pore size of 13 A (see Figure 5). The three systems are balanced separately by CG molecular dynamics simulations at a temperature of 300 K and amorphous carbon graphene layers are separated by empty space. Graphene layers are designed to separate water-SDS solution and any water molecules and SDS would migrate through porous carbon windows.

Figure 5. Images from top and front of the system components investigated.

Figure 6. Evolution of the system during molecular dynamics simulations, in terms of imposing rigid porous system uniui for average pore size of 13 A.

To see the role played by rigid carbon structures we performed two simulations: i) maintaining fixed positions pseud, carbon particles and ii) letting them move freely according to Newton dynamics. Both simulations Sund NVT thermodynamic ensembles performed by keeping fixed the basic cell and temperature T = 300K, controlled with a Nose thermostat with parameter Q = 100. While maintaining a fixed pseudo-particles of carbon is found that during the first 50 ps, the water-SDS solution, stick 'porous carbon and a few water molecules penetrate inside windows and SDS (see Figure 6) and remain localized there throughout the simulation (50ns). Only a few water molecules get in the free zone, no SDS molecule but not fall in this area.

Figure 7. Details of the structures formed by water molecules (green spheres isolated) and SDS (chains, in which sodium ions are blue heads) located near the amorphous carbon (left) and that adsorbed on the top layer of graphene (right).

Upper graphene surface a small number of SDS molecules adsorb water and forming a layer. SDS molecules adsorb parallel to the layer of graphene, a single molecule is perpendicular to the layer. Some water molecules enter and leave the pores SDS or change their orientation. Pore ​​size is large enough to access 4-5 SDS molecules and several water molecules. Water molecules are forced to leave shortly pores (1-2 ps) after they penetrate. SDS molecules that penetrate the pores have a hydrophilic end orientation to the top of the pores, leaving the hydrophobic end-oriented solution. Water-SDS solution layer changes its original structure by rearranging SDS molecules in solution: molecules in the pores are oriented vertically with hydrophobic end to the pore, in the vicinity of the hydrophobic ends of SDS molecules entering the pores and the remaining SDS molecules orienteza hydrophobic ends hydrophilic ends of SDS molecules to close pores. SDS molecules in the lower layer solution is oriented almost horizontally oriented heads trying to take parts of the same type. It is assumed that this restructuring in size, shape and spacing of pores plays an important role as well as temperature, thickness and concentration of SDS solution.

Figure 7. Evolution of the system during molecular dynamics simulations, in terms of imposing a porous flexible system for average pore size of 13 A.

If full flexibilisation carbon structures in the first 50 ps the system evolution is very similar to the case of simulation with rigid carbon structures but another 10 ps after the two layers of graphite and carbon curls porous due to vibrations extends vertically about 2-3%. Within a few nanoseconds water-SDS solution is constrained between the deoua carbon structures. Here there may be some artifacts in the simulation regarding the initial conditions, ie random choice of initial rates of carbon atoms can give these systems the rotation movements in all conditions, friction 'between the carbon layers images to transform these movements internal vibration and rotation here to have these excessive wane graphene layers. However, these artifacts have no essential influence on molecules permeated into the pores. SDS molecules permeated into the pores have the same orientation as for the system with rigid carbon structures which requires a similar behavior of water molecules in the SDS solution as with the previous simulation.