pubs.acs.org/Macromolecules Published on Web 10/06/2009 r 2009 American Chemical Society 8260 Macromolecules 2009, 42, 8260–8270 DOI: 10.1021/ma901584w Multiscale Computer Simulation Studies of Water-Based Montmorillonite/Poly(ethylene oxide) Nanocomposites Radovan Toth, † Dirk-Jan Voorn, ‡ Jan-Willem Handgraaf, § Johannes G. E. M. Fraaije, §,^ Maurizio Fermeglia, † Sabrina Pricl, † and Paola Posocco* ,† † Molecular Simulation Engineering (MOSE) Laboratory, Department of Chemical Engineering (DICAMP), University of Trieste, 34127 Trieste, Italy, ‡ Fluor Consultants B.V., 3088EA Rotterdam, The Netherlands, § Culgi B.V., 2300 AN Leiden, The Netherlands, and ^ Leiden Institute of Chemistry, Soft Matter Chemistry, Gorlaeus Laboratories, Universiteit Leiden, 2333 CC Leiden, The Netherlands Received July 20, 2009; Revised Manuscript Received September 8, 2009 ABSTRACT: This work presents a multiscale computational approach to probe the behavior of polymer/ clay nanocomposites based on poly(ethylene oxide) (PEO)/montmorillonite (MMT) as obtained from water intercalation. In details, our modeling recipe is based on four sequential steps: (a) atomistic molecular dynamics simulations to derive interaction energy values among all system components; (b) mapping of these values onto mesoscale dissipative particle dynamics parameters; (c) mesoscopic simulations to determine system density distributions and morphologies (i.e., intercalated vs exfoliated); (d) simulations at finite- element levels to calculate the relative macroscopic properties. The entire computational procedure has been applied to four PEO/MMT systems with PEO chains of different molecular weight (750, 1100, 2000, and 5000), and thermal and electrical characteristics were predicted in excellent agreement with the available experimental data. Importantly, our methodology constitutes a truly integrated multiscale modeling approach, in which no “learning against experiment” has been performed in any step of the computational recipe. 1. Introduction In recent years, polymer nanocomposites based on layered silicates, or polymer-clay nanocomposites (PCNs), have at- tracted great industrial and academic interest as they often exhibit remarkable improvement in materials properties with respect to virgin polymers or conventional micro/macro composites. These enhanced features include high mechanical moduli, increased strength and heat resistance, decreased gas permeability and flammability, and increased biodegradability in case of bio- degradable polymers. 1 Fabricating polymer clay nanocomposites (PCNs) in an efficient and cost-effective manner, however, poses significant synthetic challenges. As the ultimate properties of these hybrid systems commonly depend on their structure, it is of particular interest to establish the morphology of the final composite. To this purpose, the development of theories and the application of computer simulation techniques have opened avenues for the design of these materials, and the a priori prediction/optimization of their structures and properties. 2 The commonly used clay materials for the preparation of PCNs belong to the same general family of 2:1 layered silicates, or phyllosilicates, montmorillonite (MMT) being a prime exam- ple of these minerals. Their crystal structure consists of layers made up of two tetrahedrally coordinated silicon atoms fused to an edge-shared octahedral sheet of either aluminum or magne- sium hydroxide. The layer thickness is around 1 nm, and the lateral dimension may vary from 30 nm up to several micrometers or larger, depending on the particular mineral. Stacking of the layers leads to a regular van der Waals gap between the layers called the interlayer space or gallery. Isomorphic substitution within the layers (for example, Al 3þ replaced by Mg 2þ or Fe 2þ , or Mg 2þ replaced by Li þ ) results in an excess of negative charge, which is counterbalanced by alkali and alkaline earth cations located inside the galleries. This type of layered silicate is characterized by a moderate surface charge known as the cation exchange capacity (CEC), generally expressed as mequiv/100 g. Generally speaking, mixing a polymer and a clay may not result in a nanocomposite material. 1 Indeed, in their pristine state layered silicates are only directly miscible with hydrophilic polymers, such as poly(ethylene oxide) (PEO) 3 or poly(vinyl alcohol) (PVA). 4 To render layered silicates compatible with other polymer matrices, one must convert the normally hydro- philic silicate surface to an organophilic one, making the inter- calation of many engineering polymers possible. Depending on the strength of interfacial interactions between the polymer matrix and the clay (modified or not), two main types of PCNs can be thermodynamically achieved: (i) inter- calated nanocomposites, in which the insertion of a polymer matrix into the clay galleries occurs in a crystallographically regular fashion, regardless of the clay to polymer ratio, and (ii) exfoliated nanocomposites, where the individual clay layers are separated in a continuous polymer matrix by an average distances that depends on clay loading. The two architectures described above can be practically produced by (i) in situ polymerization of a given monomer in the presence of the layered silicate, (ii) solution intercalation, where both the polymer matrix and clay are dispersed in a common solvent followed by precipitation, or (iii) melt processing, which involves the mechanical mixing of the polymeric matrix and the inorganic filler. 1b MMT/PEO-based PNCs are hybrid structures with improved electrical properties for electronic applications in solid-state electrolyte batteries. 5-7 The intercalation of water-soluble PEO molecules between the clay galleries can be obtained by mixing the clay with an aqueous dispersion of PEO (i), or by direct intercalation from the melt (ii). 8 In the latter case, the organic *To whom correspondence should be addressed. E-mail: posocco@ dicamp.units.it.