Molecular dynamics simulations of melting and the glass transition of nitromethane

Molecular dynamics simulations have been used to investigate the thermodynamic melting point of the crystalline nitromethane, the melting mechanism of superheated crystalline nitromethane, and the physical properties of crystalline and glassy nitromethane. The maximum superheating and glass transition temperatures of nitromethane are calculated to be 316 and 160 K, respectively, for heating and cooling rates of 8.9 x 10(9) Ks. Using the hysteresis method [Luo et al., J. Chem. Phys. 120, 11640 (2004)] and by taking the glass transition temperature as the supercooling temperature, we calculate a value of 251.1 K for the thermodynamic melting point, which is in excellent agreement with the two-phase result [Agrawal et al., J. Chem. Phys. 119, 9617 (2003)] of 255.5 K and measured value of 244.73 K. In the melting process, the nitromethane molecules begin to rotate about their lattice positions in the crystal, followed by translational freedom of the molecules. A nucleation mechanism for the melting is illustrated by the distribution of the local translational order parameter. The critical values of the Lindemann index for the C and N atoms immediately prior to melting (the Lindemann criterion) are found to be around 0.155 at 1 atm. The intramolecular motions and molecular structure of nitromethane undergo no abrupt changes upon melting, indicating that the intramolecular degrees of freedom have little effect on the melting. The thermal expansion coefficient and bulk modulus are predicted to be about two or three times larger in crystalline nitromethane than in glassy nitromethane. The vibrational density of states is almost identical in both phases.     

Molecular dynamics simulations of shock waves in oriented nitromethane single crystals

The structural relaxation of crystalline nitromethane initially at T = 200 K subjected to moderate (~15 GPa) supported shocks on the (100), (010), and (001) crystal planes has been studied using microcanonical molecular dynamics with the nonreactive Sorescu-Rice-Thompson force field [D. C. Sorescu, B. M. Rice, and D. L. Thompson, J. Phys. Chem. B 104, 8406 (2000)]. The responses to the shocks were determined by monitoring the mass density, the intermolecular, intramolecular, and total temperatures (average kinetic energies), the partitioning of total kinetic energy among Cartesian directions, the radial distribution functions for directions perpendicular to those of shock propagation, the mean-square displacements in directions perpendicular to those of shock propagation, and the time dependence of molecular rotational relaxation as a function of time. The results show that the mechanical response of crystalline nitromethane strongly depends on the orientation of the shock wave. Shocks propagating along [100] and [001] result in translational disordering in some crystal planes but not in others, a phenomenon that we refer to as plane-specific disordering; whereas for [010] the shock-induced stresses are relieved by a complicated structural rearrangement that leads to a paracrystalline structure. The plane-specific translational disordering is more complete by the end of the simulations (~6 ps) for shock propagation along [001] than along [100]. Transient excitation of the intermolecular degrees of freedom occurs in the immediate vicinity of the shock front for all three orientations; the effect is most pronounced for the [010] shock. In all three cases excitation of molecular vibrations occurs more slowly than the intermolecular excitation. The intermolecular and intramolecular temperatures are nearly equal by the end of the simulations, with 400-500 K of net shock heating. Results for two-dimensional mean-square molecular center-of-mass displacements, calculated as a function of time since shock wave passage in planes perpendicular to the direction of shock propagation, show that the molecular translational mobility in the picoseconds following shock wave passage is greatest for [001] and least for the [010] case. In all cases the root-mean-square center-of-mass displacement is small compared to the molecular diameter of nitromethane on the time scale of the simulations. The calculated time scales for the approach to thermal equilibrium are generally consistent with the predictions of a recent theoretical analysis due to Hooper [J. Chem. Phys. 132, 014507 (2010)].     

Molecular dynamics simulations of the melting of 1,3,3-trinitroazetidine

Physical properties of condensed-phase 1,3,3-trinitroazetidine (TNAZ) have been computed with molecular dynamics (MD) and a nonreactive, fully flexible force field formulated by combining the intramolecular interactions obtained from the Generalized AMBER Force Field and the rigid-molecule force field developed by Sorescu-Rice-Thompson [J. Phys. Chem. B 1997, 101, 798] (AMBER-SRT). The results are compared with MD calculations, using the AMBER force field. The predicted densities of crystalline TNAZ from both force fields are about 10% lower than the experimental value. The calculated thermodynamic melting point at 1 atm from the AMBER-SRT force field is 390 K, in good agreement with the measured value of 374 K, while the AMBER force field predicts a thermodynamic melting point of 462 K. The lattice parameters and the molecular and crystal structures calculated with the AMBER-SRT force field are in excellent agreement with experiment. Simulations with the AMBER-SRT force field were also used to generate the isotherm of TNAZ up to 4 GPa and the bulk modulus and its pressure derivative.     

Molecular dynamics simulations of the melting of aluminum nanoparticles

Molecular dynamics simulations are performed to determine the melting points of aluminum nanoparticles of 55-1000 atoms with the Streitz-Mintmire [Phys. Rev. B 1994, 50, 11996] variable-charge electrostatic plus potential. The melting of the nanoparticles is characterized by studying the temperature dependence of the potential energy and Lindemann index. Nanoparticles with less than 850 atoms show bistability between the solid and liquid phases over temperature ranges below the point of complete melting. The potential energy of a nanoparticle in the bistable region alternates between values corresponding to the solid and liquid phases. This bistability is characteristic of dynamic coexistence melting. At higher temperatures, only the liquid state is stable. Nanoparticles with more than 850 atoms undergo a sharp solid-liquid-phase transition characteristic of the bulk solid phase. The variation of the melting point with the effective nanoparticle radius is also determined.     

Molecular dynamics simulations of shock waves in oriented nitromethane single crystals: plane-specific effects

Molecular dynamics simulations of supported shock waves (shock pressure P(s) ～ 15 GPa) propagating along the [110], [011], [101], and [111] directions in crystalline nitromethane initially at T = 200 K were performed using the nonreactive Sorescu-Rice-Thompson force field [D. C. Sorescu, B. M. Rice, and D. L. Thompson, J. Phys. Chem. B 104, 8406 (2000)]. These simulations, combined with those from a preceding study of shocks propagating along [100], [010], and [001] directions in nitromethane for similar conditions of temperature and shock pressure [L. He, T. D. Sewell, and D. L. Thompson, J. Chem. Phys. 134, 124506 (2011)], have been used to study the post-shock relaxation phenomena. Shocks along [010] and [101] lead to a crystal-crystal structure transformation. Shocks propagating along [011], [110], [111], [100], and [001] exhibit plane-specific disordering, which was characterized by calculating as functions of time the 1D mean square displacement (MSD), 2D radial distribution function (RDF), and 2D orientation order parameter P(2)(θ) in orthogonal planes mutually perpendicular to the shock plane; and by calculating as functions of distance behind the shock front the Cartesian components of intermolecular, intramolecular, and total kinetic energies. The 2D RDF results show that the structural disordering for shocks along [100], [110], and [111] is strongly plane-specific; whereas for shocks along [001] and [011], the loss of crystal structural order is almost equivalent in the orthogonal planes perpendicular to the shock plane. Based on the entire set of simulations, there is a trend for the most extensive disordering to occur in the (010) and (110) planes, less extensive disordering to occur in the (100) plane, and essentially no disordering to occur in the (001) plane. The 2D P(2)(θ) and 1D MSD profiles show, respectively, that the orientational and translational disordering is plane-specific, which results in the plane-specific structural disordering observed in the 2D RDF. By contrast, the kinetic energy partitioning and redistribution do not exhibit plane specificity, as shown by the similarity of spatial profiles of the Cartesian components of the intermolecular, intramolecular, and total kinetic energies in orthogonal planes perpendicular to the shock plane.     

Molecular dynamics simulations of melting of perfect crystalline hexahydro-1,3,5-trinitro-1,3,5-s-triazine

The melting mechanism of superheated perfect crystalline hexahydro-1,3,5-trinitro-1,3,5-s-triazine (alpha-RDX) has been investigated using molecular dynamics simulations with the fully flexible force field developed by Smith and Bharadwaj [J. Phys. Chem. B 103, 3570 (1999)]. Sequential 50 ps equilibration simulations of the constant stress-constant temperature ensemble were performed at 10 K intervals over the range of 300-650 K, corresponding to a heating rate of 2.0 x 10(11) Ks. A solid-solid phase transition is observed between 480 and 490 K, followed by melting, which occurs between 500 and 510 K. The solid-solid phase transition, both displacive and rotational, is characterized by an abrupt decrease in the lengths of the unit cell edges a and b and an increase of the length of edge c. The molecular conformation in the new phase is AAE, although the axial nitro groups have different changes: one shift is more axial and the other is more equatorial. Phases other than alpha-RDX have been observed experimentally, however, there are insufficient data for comparisons to ascertain that the new phase observed here corresponds to a real phase. At the high heating rate (2.0 x 10(11) Ks) used in the simulations, the melted RDX reaches full orientational disorder at about 540 K and translational freedom at around 580 K. If the simulation at the melting temperature (510 K) is run sufficiently long complete rotational freedom is achieved in a few hundreds of picoseconds, while complete translational freedom requires much longer. These results show that given a sufficiently high heating rate, the system can exist for significant periods of time in a near-liquid state in which the molecules are not as free to rotate and diffuse as in the true liquid state. The bond lengths and bond angles undergo little change upon melting, while there are significant changes in the dihedral angles. The molecular conformation of RDX changes from AAE to EEE upon melting. The ramification of this for formulating force fields that accurately describe melting is that it is important that the torsional motions are accurately described.     

Molecular dynamics simulations of semicrystalline polymers: Crystallization,melting, and reorganization

Large‐scale molecular dynamics (MD) simulations of semicrystalline entangled polymers are carried out to explore crystallization and melting processes. Semicrystalline polymers are obtained from disordered melts via homogeneous nucleation. In the early stage of the crystallization process, the collective scattering does not show the emergence of nuclei seeds. Although the crystallization process is thermodynamically simple, the melting process is complex resulting in multiple‐peaked melting endotherms. The molecular origin is found to be the different thermal stabilities of microcrystalline domains (MCDs). Coexistence of melting and growth of different MCDs during sufficiently slow heating enlarges the difference of their thermal stabilities. An increase of stem length close to the melting point is assisted by disorder effects in particular in the surface regions of the MCDs. The number of trans–trans states is decreasing, which increases the flexibility and mobility of the crystalline stems. We have also investigated self‐seeding processes, and we show how these can be used to obtain single lamellar crystals in MD simulations. © 2010 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys, 2010     

Molecular dynamics simulations of the melting mechanisms of perfect and imperfect crystals of dimethylnitramine

The melting mechanisms of perfect and imperfect crystalline dimethylnitramine have been studied using molecular dynamics simulations. The imperfect crystal was created by removing approximately 10% of the molecules from the center of the simulation cell. The density, diffusion coefficient, translational and orientational order parameters, and void size were calculated as functions of temperature and simulation time. Upon melting, the volume of the imperfect crystal slowly decreases with time due to the shrinkage of the void then suddenly decreases to a minimum value due to collapse of the structure around the void with concomitant diffusion of molecules into the void. The simulation cell volume then increases as the liquid nucleus formed at the void expands. The melting of perfect crystals must occur by a different mechanism. As the temperature of the perfect crystal reaches the maximum superheating temperature, there is an increase in the thermal motions of the molecules that result in the formation of liquid centers (characterized by translational order parameter consistent with the liquid phase) at random locations. The liquid centers rapidly grow, resulting in a complete transition to the liquid phase. The increases in orientational and translational freedom occur simultaneously in the imperfect crystal, and in the perfect crystal, orientational freedom significantly precedes translational freedom.     

Molecular dynamics simulations of a hard sphere crystal and reaction-like mechanism for homogeneous melting

Molecular dynamics simulations of a hard sphere crystal are performed for volume fractions ranging from solidification point to melting point. A local bond order parameter is chosen to assign a nature, liquid or solid, to a particle. The probability for a liquid or solid particle to change state presents a typical sigmoid shape as the nature of its neighbors changes. Using this property, I propose a reaction-like mechanism and introduce a small number of rate constants. A mean-field approach to melting and a kinetic Monte Carlo algorithm on a lattice are derived from these chemical processes. The results of these models successfully compare with molecular dynamics simulations, proving that the main properties of melting can be captured by a small number of dynamical parameters.     

Density-dependent liquid nitromethane decomposition: molecular dynamics simulations based on ReaxFF

The decomposition mechanism of hot liquid nitromethane at various compressions was studied using reactive force field (ReaxFF) molecular dynamics simulations. A competition between two different initial thermal decomposition schemes is observed, depending on compression. At low densities, unimolecular C-N bond cleavage is the dominant route, producing CH(3) and NO(2) fragments. As density and pressure rise approaching the Chapman-Jouget detonation conditions (～30% compression, >2500 K) the dominant mechanism switches to the formation of the CH(3)NO fragment via H-transfer and/or N-O bond rupture. The change in the decomposition mechanism of hot liquid NM leads to a different kinetic and energetic behavior, as well as products distribution. The calculated density dependence of the enthalpy change correlates with the change in initial decomposition reaction mechanism. It can be used as a convenient and useful global parameter for the detection of reaction dynamics. Atomic averaged local diffusion coefficients are shown to be sensitive to the reactions dynamics, and can be used to distinguish between time periods where chemical reactions occur and diffusion-dominated, nonreactive time periods.     

Early chemistry in hot and dense nitromethane: molecular dynamics simulations

We report density functional molecular dynamic simulations to determine the early chemical events of hot (T=3000 K) and dense (rho=1.97 g/cm(3), V/V(0)=0.68) nitromethane (CH(3)NO(2)). The first step in the decomposition process is an intermolecular proton abstraction mechanism that leads to the formation of CH(3)NO(2)H(+) and the aci ion H(2)CNO(2) (-). This event is also confirmed to occur in a fast annealing simulation to a final temperature of 4000 K at rho=2.20 g/cm(3). An intramolecular hydrogen transfer that transforms nitromethane into the aci acid form, CH(2)NO(2)H, accompanies this event. To our knowledge, this is the first confirmation of chemical reactivity with bond selectivity for an energetic material near the Chapman-Jouget state of the fully reacted material. We also report the decomposition mechanism followed up to the formation of H(2)O as the first stable product. We note that similarities in the global features of reactants, intermediates, and products of the reacting fluid seem to indicate a threshold for similar chemistry in the range of high densities and temperatures reported herein.     

Molecular dynamics study of the crystallization of nitromethane from the melt

The crystallization of nitromethane, CH(3)NO(2), from the melt on the (100), (010), (001), and (110) crystal surfaces at 170, 180, 190, 200, 210, and 220 K has been investigated using constant-volume and -temperature (NVT) molecular dynamics simulations with a realistic, fully flexible force field [D. C. Sorescu, B. M. Rice, and D. L. Thompson, J. Phys. Chem. B 104, 8406 (2000)]. The crystallization process and the nature of the solid-liquid interface have been investigated by computing the molecular orientations, density, and radial distribution functions as functions of time and location in the simulation cell. During crystallization the translational motion of the molecules ceases first, after which molecular rotation ceases as the molecules assume proper orientations in the crystal lattice. The methyl groups are hindered rotors in the liquid; hindrance to rotation is reduced upon crystallization. The width of the solid-liquid interface varies between 6 and 13 ? (about two to five molecular layers) depending on which crystal surface is exposed to the melt and which order parameter is used to define the interface. The maximum rate of crystallization varies from 0.08 molecules ns(-1) ?(-2) for the (010) surface at 190 K to 0.41 molecules ns(-1) ?(-2) for the (001) surface at 220 K.     

Molecular dynamics simulations of halide glasses

Summary Halide glasses have been extensively studied in recent years because of their potential application as infrared transmitting fibre optic materials. They are believed to be more ionic than glasses based on silica and should therefore be more amenable to molecular dynamics simulation using simple two-body potentials. Here the main features of structural models derived using such techniques are described and compared with available structural data. Possible future applications of this approach are outlined.     

Molecular dynamics simulations of peptide-surface interactions

Proteins, which are bioactive molecules, adsorb on implants placed in the body through complex and poorly understood mechanisms and directly influence biocompatibility. Molecular dynamics modeling using empirical force fields provides one of the most direct methods of theoretically analyzing the behavior of complex molecular systems and is well-suited for the simulation of protein adsorption behavior. To accurately simulate protein adsorption behavior, a force field must correctly represent the thermodynamic driving forces that govern peptide residue-surface interactions. However, since existing force fields were developed without specific consideration of protein-surface interactions, they may not accurately represent this type of molecular behavior. To address this concern, we developed a host-guest peptide adsorption model in the form of a G(4)-X-G(4) peptide (G is glycine, X is a variable residue) to enable determination of the contributions to adsorption free energy of different X residues when adsorbed to functionalized Au-alkanethiol self-assembled monolayers (SAMs). We have previously reported experimental results using surface plasmon resonance (SPR) spectroscopy to measure the free energy of peptide adsorption for this peptide model with X = G and K (lysine) on OH and COOH functionalized SAMs. The objectives of the present research were the development and assessment of methods to calculate adsorption free energy using molecular dynamics simulations with the GROMACS force field for these same peptide adsorption systems, with an oligoethylene oxide (OEG) functionalized SAM surface also being considered. By comparing simulation results to the experimental results, the accuracy of the selected force field to represent the behavior of these molecular systems can be evaluated. From our simulations, the G(4)-G-G(4) and G(4)-K-G(4) peptides showed minimal to no adsorption to the OH SAM surfaces and the G(4)-K-G(4) showed strong adsorption to the COOH SAM surface, which is in agreement with our SPR experiments. Contrary to our experimental results, however, the simulations predicted a relatively strong adsorption of G(4)-G-G(4) peptide to the COOH SAM surface. In addition, both peptides were unexpectedly predicted to adsorb to the OEG surface. These findings demonstrate the need for GROMACS force field parameters to be rebalanced for the simulation of peptide adsorption behavior on SAM surfaces. The developed methods provide a direct means of assessing, modifying, and validating force field performance for the simulation of peptide and protein adsorption to surfaces, without which little confidence can be placed in the simulation results that are generated with these types of systems.     

Molecular dynamics simulations of lipid bilayers

Computer simulation methods are becoming increasingly widespread as tools for studying the structure and dynamics of lipid bilayer membranes. The length scale and time scale accessible to atomic-level molecular dynamics simulations are rapidly increasing, providing insight into the relatively slow motions of molecular reorientation and translation and demonstrating that effects due to the finite size of the simulation cell can influence simulation results. Additionally, significant advances have been made in the complexity of membrane systems studied, including bilayers with cholesterol, small solute molecules, and lipid-protein and lipid-DNA complexes. Especially promising is the progress that continues to be made in the comparison of simulation results with experiment, both to validate the simulation algorithms and to aid in the interpretation of existing experimental data.     

Molecular dynamics simulations of biotin carboxylase

Biotin carboxylase catalyzes the ATP-dependent carboxylation of biotin and is one component of the multienzyme complex acetyl-CoA carboxylase that catalyzes the first committed step in fatty acid synthesis in all organisms. Biotin carboxylase from Escherichia coli, whose crystal structures with and without ATP bound have been determined, has served as a model system for this component of the acetyl-CoA carboxylase complex. The two crystal structures revealed a large conformational change of one domain relative to the other domains when ATP is bound. Unfortunately, the crystal structure with ATP bound was obtained with an inactive site-directed mutant of the enzyme. As a consequence the structure with ATP bound lacked key structural information such as for the Mg2+ ions and contained altered conformations of key active-site residues. Therefore, nanosecond molecular dynamics studies of the wild-type biotin carboxylase were undertaken to supplant and amend the results of the crystal structures. Specifically, the protein-metal interactions of the two catalytically critical Mg2+ ions bound in the active site are presented along with a reevaluation of the conformations of active-site residues bound to ATP. In addition, the regions of the polypeptide chain that serve as hinges for the large conformational change were identified. The results of the hinge analysis complemented a covariance analysis that identified the individual structural elements of biotin carboxylase that change their conformation in response to ATP binding.     

Molecular dynamics simulations of polyelectrolyte adsorption

We have performed molecular dynamics simulations of polyelectrolyte adsorption at oppositely charged surfaces from dilute polyelectrolyte solutions. In our simulations, polyelectrolytes were modeled by chains of charged Lennard-Jones particles with explicit counterions. We have studied the effects of the surface charge density, surface charge distribution, solvent quality for the polymer backbone, strength of the short-range interactions between polymers and substrates on the polymer surface coverage, and the thickness of the adsorbed layer. The polymer surface coverage monotonically increases with increasing surface charge density for almost all studied systems except for the system of hydrophilic polyelectrolytes adsorbing at hydrophilic surfaces. In this case the polymer surface coverage saturates at high surface charge densities. This is due to additional monomer-monomer repulsion between adsorbed polymer chains, which becomes important in dense polymeric layers. These interactions also preclude surface overcharging by hydrophilic polyelectrolytes at high surface charge densities. The thickness of the adsorbed layer shows monotonic dependence on the surface charge density for the systems of hydrophobic polyelectrolytes for both hydrophobic and hydrophilic surfaces. Thickness is a decreasing function of the surface charge density in the case of hydrophilic surfaces while it increases with the surface charge density for hydrophobic substrates. Qualitatively different behavior is observed for the thickness of the adsorbed layer of hydrophilic polyelectrolytes at hydrophilic surfaces. In this case, thickness first decreases with increasing surface charge density, then it begins to increase.     

Molecular dynamics simulations of cardiolipin bilayers

Cardiolipin is a key lipid component in the inner mitochondrial membrane, where the lipid is involved in energy production, cristae structure, and mechanisms in the apoptotic pathway. In this article we used molecular dynamics computer simulations to investigate cardiolipin and its effect on the structure of lipid bilayers. Three cardiolipin/POPC bilayers with different lipid compositions were simulated: 100, 9.2, and 0% cardiolipin. We found strong association of sodium counterions to the carbonyl groups of both lipid types, leaving in the case of 9.2% cardiolipin virtually no ions in the aqueous compartment. Although binding occurred primarily at the carbonyl position, there was a preference to bind to the carbonyl groups of cardiolipin. Ion binding and the small headgroup of cardiolipin gave a strong ordering of the hydrocarbon chains. We found significant effects in the water dipole orientation and water dipole potential which can compensate for the electrostatic repulsion that otherwise should force charged lipids apart. Several parameters relevant for the molecular structure of cardiolipin were calculated and compared with results from analyses of coarse-grained simulations and available X-ray structural data.     

Molecular dynamics simulations of supramolecular polymer rheology

Using equilibrium and nonequilibrium molecular dynamics simulations, we studied the equilibrium and rheological properties of dilute and semidilute solutions of head-to-tail associating polymers. In our simulation model, a spontaneous complementary reversible association between the donor and the acceptor groups at the ends of oligomers was achieved by introducing a combination of truncated pseudo-Coulombic attractive potential and Lennard Jones repulsive potential between donor, acceptor, and neighboring groups. We have calculated the equilibrium properties of supramolecular polymers, such as the ring/chain equilibrium, average molecular weight, and molecular weight distribution of self-assembled chains and rings, which all agree well with previous analytical and computer modeling results. We have investigated shear thinning of solutions of 8- and 20-bead associating oligomers with different association energies at different temperatures and oligomer volume fractions. All reduced viscosity data for a given oligomer length can be collapsed into one master curve, exhibiting two power-law regions of shear-thinning behavior with an exponent of -0.55 at intermediate ranges of the reduced shear rate β and -0.8 (or -0.9) at larger shear rates. The equilibrium viscosity of supramolecular solutions with different oligomer lengths and associating energies is found to obey a power-law scaling dependence on oligomer volume fraction with an exponent of 1.5, in agreement with the experimental observations for several dilute or semidilute solutions of supramolecular polymers. This implies that dilute and semidilute supramolecular polymer solutions exhibit high polydispersity but may not be sufficiently entangled to follow the reptation mechanism of relaxation.