Molecular Dynamics Simulations of Nanoporous Systems: Mechanism of Enhanced Dynamics of Ions
Introduction of Studies for Porous Systems
Porous materials have important roles in the recent development of nanotechnology. The possibility of the application of porous materials is widely spreading [1-5] because of their unique characteristics such as lightness and large surface area. In the research of nanoscience fields as well as in industry, porous materials are used as supports for catalysts [1]. They are also expected to be used for drug deliveries [2], where the solubility of the drug can be changed by porous adsorbents. Furthermore, porous materials are good candidates for electric devices, such as electrodes in solid-state batteries [3]. They are also expected to be useful for the storage of Li ions [4] and/or other materials. Porous structures also can be used as sheaves for the separation of materials [5]. Not only in applications, but porous materials can also be a target to examine the physical and chemical properties of materials with controlling or modifying them [6, 7], in fundamental research fields.
Molecular Dynamics of Nanostructures and Nanoionics: Simulations in Complex Systems Junko Habasaki
Copyright © 2021 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-77-8 (Hardcover), 978-1-003-04490-1 (eBook) www.jennystanford.com
At the beginning of the study of porous materials, the silica system was a target to be examined by molecular dynamics simulations. Studies of silica and related alkali silicate are useful for understanding behaviors of the porous materials and give fundamental understandings for using them as devices and/or storages. In this chapter, some methods used to study porous silica in molecular dynamics simulations are introduced and then results of such modeling used for silicate systems are explained.
In nanoporous materials and their composites, enhancement of ionic conductivity has been reported as shown in Chapter 3 and possible mechanisms having different origins have been proposed so far. Especially, it is noteworthy that lithium silicate glass grown within pores and in composites also shows the enhancement as shown in Chapter 3.
In our previous works [6, 7], effects of introducing pores were examined by focusing enhanced ionic motion of Li ions in porous lithium silicates examined by molecular dynamics (MD) simulation in the constant volume and energy (NVE) conditions. Interestingly, enhanced dynamics having comparable characteristics to the composites shown in Chapter 3 were predicted. Therefore, the focus of the chapter is on clarifying the mechanism of enhanced diffusion found by MD simulations, although the mechanism found here is not necessarily the same as those in composites.
Usually, the high performance of porous materials tends to be attributed to their extremely high surface areas or structures of boundaries combined with the formation of pores [8]. However, there are many possibilities to be removed to decide what mechanism is important.
Since the introducing pore usually results in the structural changes of materials in the nanoscale and hence details of the structures and roles of pores need to be clarified. In this situation, MD simulations can play roles to clarify the mechanisms and structures in atomistic levels. It will be useful to show how possible mechanisms of the enhancement of the dynamics can be distinguished from each other in MD simulations in the case of the porous lithium disilicate for further examination of related systems.
In applications, the diameter of channels (pores) is an important factor to control the properties of porous materials. Porous materials can be subdivided into three categories by sizes of pores, set out by IUPAC as follows [9] as already mentioned in Chapter 2.
In the report, pores with widths exceeding about 50 nm are called macropores; pores of widths between 2 and 50 nm are called mesopores, and pores with widths not exceeding about 2 nm are called micropores. In Chapters 6 and 7, porous systems treated are mainly mesoporous and microporous ones.
Mesoporous silicate is experimentally prepared in several works. In the review by Alothman [10], the utility of silicate mesoporous materials for several purposes has been discussed. The purposes include supports for catalysis, separation, selective adsorption, and storage of gas or other guest molecules. It was argued that in silicate materials, well-defined pore sizes of about 2.0-10.0 nm and their range are better than that of microporous zeolites (<2.0 nm) for such applications. Systems are expected to take roles in environmental pollution control, especially for the uptake of heavy metals. Porous systems show a variety of morphology, which might affect the macroscopic behavior of the system. The properties of porous materials are affected by many factors besides the morphology.
Modeling of Nanoporous Materials
Characteristics of pores introduced in the systems depend on the methods of preparation. It is pointed out that channel-like structures are obtained by sol-gel methods [11], while more isolated pores can be obtained by chemical vapor deposition [12, 13], experimentally. In [1], three methods are compared to modify the size and/or shapes of pores. In the gel obtained from colloidal silica (see Chapters 8 and 9), we have examined a structure of gel, where networks of Si04 units are three-dimensionally connected infinitely by periodic boundary condition in MD simulations. In such cases, not only structures of silica networks, but also the structures and dynamics of the water trapped in the pore can be examined in parallel. By such analyses, the existence of open-pore structures is clarified. As found in this example, MD simulations are useful techniques to examine structures in porous materials as well as dynamics within them.
First, several methods for preparing porous models [14-18] used in the studies of porous silica will be summarized. After that, results of our MD simulations will be explained.
Modeling of Porous Silica
Kieffer and Angell [14] prepared porous silica by the expansion of the dense silica system, where both lengths of basic cell and positions of atoms are scaled. The method is performed under the NVE (constant number of particles, constant volume, and constant energy) conditions and was later refined by Nakano et al. [15] using NPT (constant number of particles, constant pressure, and constant temperature) condition to allow relaxing of the system volume during the run. The difference of porous structures in silicate in different extended conditions of MD runs will be discussed later.
A different method, a charge scaling procedure, was used for the preparation of a nanoporous silica model [16]. As in experiments, porous systems can be prepared by the aggregation process [17] and by a spontaneous gelation process [18]. A comparison of structures and mechanical properties obtained by several methods is found in ref. [19].
In ref. [18], starting from the nano-colloidal silica-water-salt system, spontaneous formation of the gel has been reproduced using fully atomistic MD simulations. The spontaneous formation of gel seems to provide a way to prepare the natural structures of the porous systems. The resultant porous structure of the silica part in the silica gel is shown in Fig. 6.1. In this figure, the formation of 3-D channels is clearly found from the structure without solvent. Besides, the structures of a solvent part filled in the channel can be used to learn the structure of pores and/or channels (see Chapters 8 and 9 for the details of the structure of silica and water parts).
The structure of the silica part can be used for the further analysis of porous systems. The porous structures after eliminating water and salt are found to be stable at least during several ns runs in both NVE and NPT conditions.
Another possible modeling method of porous materials in MD/MC may be to use a template of structures. The method using a random or systematic removal of atoms or units from the system can also be possible ways, although the modification of conditions is required to continue stable runs of the system.
Figure 6.1 An example of nanoporous silica taken from the gel spontaneously formed from nanocolloidal-silica—water-NaCl system obtained by MD simulation (see Chapters 8 and 9). Structures of water and NaCl are omitted in the present figure. Red and blue particles are for oxygen and silicon, respectively. Each SiC>4 unit is shown by a tetrahedron.
Modeling of Porous Silicate
Introducing pore is often accompanied with the changes of structures from ordered ones to disordered ones. Dynamics of the system will be affected by such structural changes and the effect is quite large. Actually, diffusion coefficients of ions in the ionically conducting materials in the crystalline state are known to be several orders slower than that in the corresponding glasses having the same composition [20-22] and this remarkable difference between glass and crystal has been confirmed by MD simulations of the lithium silicate systems [23], which are typical ionically conducting systems.
In the experimental situation of systems with nanocrystals or nanoparticles, the contribution of disorder introduced in the system with mixing other components will affect dynamics considerably. Therefore, before discussing the effects caused by the modification of the system by pores, elimination of the effects in different ordering is necessary.
For this purpose, we prepared the porous system with starting from the system in the glassy state [6, 7], so that the effects of introducing disorder is removed a priori. That is, porous lithium disilicate systems were prepared by the expansion of the glass, which was obtained by the rapid cooling of the corresponding melt. By this procedure, one can eliminate effects of different contents of the crystalline phase and glassy phase, which might affect the enhancement. The functional form and description of the potential models used in the MD simulations are shown in the Appendix.
In Chapters 6 and 7, density p is given in g cm'3 and time is in ps, unless otherwise stated. We may use non-SI units, case by case, for several reasons. One of the reasons is the convenience to treat problems in intuitively easier ways. For example, lengths are given in A (1 A = 0.1 nm), because it is convenient to treat atomistic sizes. Porosity (defined by the ratio of the volume of space to the total volume, that is the fraction of void), P0 (in %) was obtained from MD simulations (after equilibration at each density) and is represented as the function of each density, p for porous systems, including the system before expansion.
