Volume 137, Issue 14, 14 October 2012

Thin films of water under nanoscopic confinement are prevalent in natural and manufactured materials. To investigate the equilibrium and dynamic behavior of water in such environments, we perform molecular dynamics simulations of water confined between atomistically detailed hydrophobic plates at T = 298 K for pressures (−0.1) ⩽ P ⩽ 1.0 GPa and plate separations of 0.40 ⩽ d ⩽ 0.80 nm. From these simulations, we construct an expanded Pdphase diagram for confined water, and identify and characterize a previously unreported confined monolayerice morphology. We also study the decompressioninduced sublimation of bilayer ice in a d = 0.6 nm slit, employing principal component analysis to synthesize lowdimensional embeddings of the drying trajectories and develop insight into the sublimation mechanism. Drying is observed to proceed by the nucleation of a bridging vapor cavity at one corner of the crystalline slab, followed by expansion of the cavity along two edges of the plates, and the subsequent recession of the remaining promontory of bilayer crystal into the bulk fluid. Our findings have implications for the understanding of diverse phenomena in materials science, nanofluidics, and protein folding and aggregation.
 COMMUNICATIONS


Communication: Direct determination of triplepoint coexistence through cell model simulation
View Description Hide DescriptionIn simulations of fluidsolid coexistence, the solid phase is modeled as a constrained system of WignerSeitz cells with one particle per cell. This model, commonly referred to as the constrained cell model, is a limiting case of a more general cell model, which is formed by considering a homogeneous external field that controls the number of particles per cell and, hence, the relative stability of the solid against the fluid phase. The generalized cell model provides a link that connects the disordered, fluid phase with the ordered, solid phase. In the present work, the phase diagram of this model is investigated through multicanonical simulations at constant pressure and histogram reweighting techniques for a system of 256 LennardJones particles. The simulation data are used to obtain an estimate of the triple point of the LennardJones system. The triplepoint pressure is found to be higher compared to previous work. The likely explanation for this discrepancy is the highly compressible nature of the gas phase.

Communication: An obligatory glass surface
View Description Hide DescriptionTheory predicts, and experiments have shown, that dynamics is faster at glass surfaces than in the bulk, allowing the glass to settle into deeper energy landscape minima, or “age more.” Is it possible that a glass surface could survive at temperatures where the bulk crystallizes, or that it could remain glassy after the bulk is heated all the way to its melting temperature and recooled? We image in realtime and with subnanometer resolution the twostate surface dynamics on a ceriumbased glass surface, from deep within the glassy regime to above the crystallization temperature. Unlike other surfaces that we have studied, this glass surface remains amorphous even after the bulk recrystallizes. The surface retains noncrystalline structure and two state dynamics of cooperatively rearranging regions even after heat annealing to just below the bulk melting temperature. The heatannealed cooperatively rearranging regions are larger than originally, a sign that the surface is well aged. The surface dynamics depends weakly on temperature, showing no sign of the superexponential increase in bulk dynamics expected near T _{ g }.
 Top

 ARTICLES

 Theoretical Methods and Algorithms

Local unitary transformation method for largescale twocomponent relativistic calculations. II. Extension to twoelectron Coulomb interaction
View Description Hide DescriptionThe local unitary transformation (LUT) scheme at the spinfree infiniteorder Douglas–KrollHess (IODKH) level [J. Seino and H. Nakai, J. Chem. Phys.136, 244102 (2012)10.1063/1.4729463], which is based on the locality of relativistic effects, has been extended to a fourcomponent Dirac–Coulomb Hamiltonian. In the previous study, the LUT scheme was applied only to a oneparticle IODKH Hamiltonian with nonrelativistic twoelectron Coulomb interaction, termed IODKH/C. The current study extends the LUT scheme to a twoparticle IODKH Hamiltonian as well as oneparticle one, termed IODKH/IODKH, which has been a real bottleneck in numerical calculation. The LUT scheme with the IODKH/IODKH Hamiltonian was numerically assessed in the diatomic molecules HX and X_{2} and hydrogen halide molecules, (HX)_{ n } (X = F, Cl, Br, and I). The total Hartree–Fock energies calculated by the LUT method agree well with conventional IODKH/IODKH results. The computational cost of the LUT method is reduced drastically compared with that of the conventional method. In addition, the LUT method achieves linearscaling with respect to the system size and a small prefactor.

New implementation of the configurationbased multireference second order perturbation theory
View Description Hide DescriptionWe present an improved version of the configurationbased multireference secondorder perturbation approach (CBMRPT2) according to the formulation of Lindgren on perturbation theory of a degenerate model space. This version involves a reclassification of the perturbation functions and new algorithms to calculate matrix elements in the perturber energy expressions utilizing the graphical unitary group approach and the holeparticle symmetry. The diagonalizethenperturb (DP), including RayleighSchrödinger and BrillouinWigner, and diagonalizethenperturbthendiagonalize (DPD) modes have been implemented. The new CBMRPT2 method is applied to several typical and interesting systems: (1) the vertical excitation energies for several states of CO and N_{2}, (2) energy comparison and timing of the ground state of C_{4}H_{6}, (3) the quasidegeneracy of states in LiF, (4) the intruder state problems of AgH, and (5) the relative energies of dicopperoxygenammonia complex isomers. The results indicate that the computational accuracy and efficiency of the presented methods are competitive and intruderfree. It should be emphasized that the DPD method rectifies naturally the shortcomings of LiF potential energy curves constructed by the original second order complete active space perturbation theory (CASPT2), without having to recourse to the socalled state mixture. Unlike CASPT2, the new methods give the same energy ordering for the two dicopperoxygenammonia isomers as the previous multireference configuration interaction with single and double excitations methods. The new CBMRPT2 method is shown to be a useful tool to study small to mediumsized systems.

Bayesian uncertainty quantification and propagation in molecular dynamics simulations: A high performance computing framework
View Description Hide DescriptionWe present a Bayesian probabilistic framework for quantifying and propagating the uncertainties in the parameters of force fields employed in molecular dynamics (MD) simulations. We propose a highly parallel implementation of the transitional Markov chain Monte Carlo for populating the posterior probability distribution of the MD forcefield parameters. Efficient scheduling algorithms are proposed to handle the MD model runs and to distribute the computations in clusters with heterogeneous architectures. Furthermore, adaptive surrogate models are proposed in order to reduce the computational cost associated with the large number of MD model runs. The effectiveness and computational efficiency of the proposed Bayesian framework is demonstrated in MD simulations of liquid and gaseous argon.

Theoretical study of vaporliquid homogeneous nucleation using stability analysis of a macroscopic phase
View Description Hide DescriptionStability analysis is generally used to verify that the solution to phase equilibrium calculations corresponds to a stable state (minimum of the free energy). In this work, tangent plane distance analysis for stability of macroscopic mixtures is also used for analyzing the nucleation process, reconciling thus this analysis with classical nucleationtheories. In the context of the revised nucleationtheory, the driving force and the nucleation work are expressed as a function of the Lagrange multiplier corresponding to the mole fraction constraint from the minimization problem of stability analysis. Using a van der Waals fluid applied to a ternary mixture, Lagrange multiplier properties are illustrated. In particular, it is shown how the Lagrange multiplier value is equal to one on the binodal and spinodal curves at the same time as the driving force of nucleation vanishes on these curves. Finally, it is shown that, on the spinodal curve, the nucleation work from the revised and generalized nucleationtheories are characterized by two different local minima from stability analysis, irrespective of any interfacial tension models.

Inferring bulk selfassembly properties from simulations of small systems with multiple constituent species and small systems in the grand canonical ensemble
View Description Hide DescriptionIn this paper, we generalize a methodology [T. E. Ouldridge, A. A. Louis, and J. P. K. Doye, J. Phys.: Condens. Matter22, 104102 (2010)] for dealing with the inference of bulk properties from small simulations of selfassembling systems of characteristic finite size. In particular, schemes for extrapolating the results of simulations of a single selfassembling object to the bulk limit are established in three cases: for assembly involving multiple particle species, for systems with one species localized in space and for simulations in the grand canonical ensemble. Furthermore, methodologies are introduced for evaluating the accuracy of these extrapolations. Example systems demonstrate that differences in cluster concentrations between simulations of a single selfassembling structure and bulk studies of the same model under identical conditions can be large, and that convergence on bulk results as system size is increased can be slow and nontrivial.

Selfconsistent phonons revisited. I. The role of thermal versus quantum fluctuations on structural transitions in large LennardJones clusters
View Description Hide DescriptionThe theory of selfconsistent phonons (SCP) was originally developed to address the anharmonic effects in condensed matter systems. The method seeks a harmonic, temperaturedependent Hamiltonian that provides the “best fit” for the physical Hamiltonian, the “best fit” being defined as the one that optimizes the Helmholtz free energy at a fixed temperature. The present developments provide a scalable unified framework that accounts for anharmonic effects in a manybody system, when it is probed by either thermal (ℏ → 0) or quantum fluctuations (T → 0). In these important limits, the solution of the nonlinear SCP equations can be reached in a manner that requires only the multiplication of 3N × 3N matrices, with no need of diagonalization. For short range potentials, such as LennardJones, the Hessian, and other related matrices are highly sparse, so that the scaling of the matrix multiplications can be reduced from to . We investigate the role of quantum effects by continuously varying the deBoer quantum delocalization parameter Λ and report the NΛ (T = 0), and also the classical NT (Λ = 0) phase diagrams for sizes up to N ∼ 10^{4}. Our results demonstrate that the harmonic approximation becomes inadequate already for such weakly quantum systems as neon clusters, or for classical systems much below the melting temperatures.

NonMarkovian reduced dynamics based upon a hierarchical effectivemode representation
View Description Hide DescriptionA reduced dynamics representation is introduced which is tailored to a hierarchical, Morichain type representation of a bath of harmonic oscillators which are linearly coupled to a subsystem. We consider a spinboson system where a single effective mode is constructed so as to absorb all systemenvironment interactions, while the residual bath modes are coupled bilinearly to the primary mode and among each other. Using a cumulant expansion of the memory kernel, correlation functions for the primary mode are obtained, which can be suitably approximated by truncated chains representing the primaryresidual mode interactions. A series of reduceddimensional bath correlation functions is thus obtained, which can be expressed as FourierLaplace transforms of spectral densities that are given in truncated continuedfraction form. For a master equation which is second order in the systembath coupling, the memory kernel is reexpressed in terms of localintime equations involving auxiliary densities and auxiliary operators.

Building band structures for long finite chains in presence of an electric field
View Description Hide DescriptionAt first, three different methods devoted to extracting the periodicsystem band structures from the results of finitesystem calculations are being discussed. Special emphasis is put on the performance of these methods for a system that is exposed to an external electrostatic field. Using model calculations we found that the three methods converge to the same results in the absence of the electrostatic field, but they lead to small differences when the field is included, which can be ascribed to distortions of the spatial extensions of the orbitals generated by the field. These small differences become important when addressing the issue whether it is possible to identify uniquely a set of band structures for the infinite, periodic system exposed to an electrostatic field. This question, related to the question whether the phases of the periodicsystem orbitals are of physical importance, has so far not been answered. Applying the results of the first part we ultimately suggest that it is not possible to identify a unique set of band structures when the periodic system is exposed to an electrostatic field.
 Advanced Experimental Techniques

Observation of proximities between spin1/2 and quadrupolar nuclei: Which heteronuclear dipolar recoupling method is preferable?
View Description Hide DescriptionWe have recently shown that the dipolarmediated heteronuclear multiplequantum coherence (DHMQC) method allows observing throughspace proximities between spin1/2 (^{1}H, ^{13}C, ^{31}P…) and quadrupolar (^{23}Na, ^{27}Al…) nuclei. However, the DHMQC effectiveness depends on the choice of the heteronuclear dipolar recoupling sequence. Here, we compare the efficiency and the robustness of four rotorsynchronized sequences: the symmetrybased ones, and its supercycled version, , and two schemes based on simultaneous amplitude and frequency modulations, denoted SFAM1 and SFAM2. For the SFAM methods, we point out efficient recoupling conditions that facilitate their experimental optimization and we introduce analytical expressions for the buildup of DHMQC signal in the case of an isolated spin pair. We show that the main differences between these four sequences lie in the number of adjustable parameters and in their robustness with respect to chemical shift and homonuclear dipolar interactions. The relative performances of these four recoupling sequences are analyzed using average Hamiltonian theory, numerical simulations, and ^{27}Al{^{31}P} DHMQC experiments on crystalline aluminophosphate.
 Atoms, Molecules, and Clusters

Optical absorbance of doped Si quantum dots calculated by timedependent density functional theory with partial electronic selfinteraction corrections
View Description Hide DescriptionThe optical properties of Si quantum dots(QDs) with phosphorous and aluminumdopants have been calculated with the recently tested HeydScuseriaErnzerhof (HSE) density functionals to ascertain the effect of functional corrections to electronic selfinteraction. New results have been obtained for 20 crystalline and amorphous structures of Si_{29} and Si_{35}quantum dots and are compared to our previous results obtained using the PW91/PW91 functionals. The bandgaps are greater in magnitude and shifted to higher energies in HSE calculations compared to PW91 calculations, and the absorption spectrum is blueshifted in HSE. Trends in the shifts of absorbances due to doping are similar for both sets of calculations, with dopedQDs absorbing at lower photon energies than undoped QDs. Consistent with previous results, the bandgaps of QDs are found to decrease as the size of the QD increases, and the absorption spectra of amorphous QDs are redshifted compared to those of crystalline structures. The molecular orbitals involved in the transitions with the largest oscillator strengths show that the electron density moves towards the surface of the quantum dot as the structure is excited. The lifetimes of photoexcited states were found to differ substantially between the two functionals due to their sensitivity to the overlaps of initial and final orbitals. Comparison with available experimental and independent theoretical results supports the conclusion that the HSE functional better matches experimental results due to the partial inclusion of HartreeFock exchange.

Photofragment translational spectroscopy of CH_{3}I at 225 nm—with the high excitation of the symmetric stretch vibration of CH_{3} fragment
View Description Hide DescriptionThe photodissociation dynamics of CH_{3}I at 225 nm is studied on our high resolution miniTOF photofragment translational spectrometer. The photofragment translational spectra of the I* and the I channels via parallel (∥) and perpendicular (⊥) transitions, i.e., of the four pathways ^{3}Q_{0}, ^{3}Q_{0} ← ^{1}Q_{1}, ^{1}Q_{1}, and ^{1}Q_{1} ← ^{3}Q_{0}, are obtained with both the symmetric stretch (ν_{1}) and the umbrella (ν_{2}) vibrational modes of the CH_{3} fragments partially resolved. The strong excitation of the symmetric stretch mode (ν_{1}) is revealed in both the I and the I* channels. The branching fractions for the four pathways (0.09 for ^{3}Q_{0}, 0.03 for ^{3}Q_{0} ← ^{1}Q_{1}, 0.34 for ^{1}Q_{1}, and 0.54 for ^{1}Q_{1} ← ^{3}Q_{0}) show that the parallel transition (^{3}Q_{0} ← X) is the major and the I channel is dominant in the photodissociation of CH_{3}I at 225 nm. The curvecrossing probability is found to be 0.86 for ^{1}Q_{1} ← ^{3}Q_{0} but 0.08 for ^{3}Q_{0} ← ^{1}Q_{1}.

A new method to derive electronegativity from resonant inelastic xray scattering
View Description Hide DescriptionElectronegativity is a wellknown property of atoms and substituent groups. Because there is no direct way to measure it, establishing a useful scale for electronegativity often entails correlating it to another chemical parameter; a wide variety of methods have been proposed over the past 80 years to do just that. This work reports a new approach that connects electronegativity to a spectroscopic parameter derived from resonant inelastic xray scattering. The new method is demonstrated using a series of chlorinecontaining compounds, focusing on the Cl 2p ^{−1}LUMO^{1} electronic states reached after Cl 1s → LUMO core excitation and subsequent KL radiative decay. Based on an electrondensity analysis of the LUMOs, the relative weights of the Cl 2p _{ z } atomic orbital contributing to the Cl 2p _{3/2} molecular spinorbit components are shown to yield a linear electronegativity scale consistent with previous approaches.

Highresolution photoelectron spectroscopy of linear ← bent polyatomic photodetachment transitions: The electron affinity of CS_{2}
View Description Hide DescriptionA combination of highresolution velocitymapimaging photoelectron spectroscopy and isotopic substitution is used to show that precise electron affinities can be obtained from polyatomic photodetachmentspectra, even for cases involving significant changes in equilibrium geometry between the molecular neutral and anion. The chosen example CS_{2} (linear) ← (X ^{2} A _{1}) (bent) photodetachment transition is found to preferentially access highlyexcited v _{2} (bending) levels of the neutral, with no observation possible of the lowestv _{2} bands. Nevertheless, through ^{13}C isotopic substitution, the v _{2} numbering is established unambiguously and the adiabatic electron affinity of CS_{2} is found to be 4456(10) cm^{−1} [0.5525(13) eV], by far the most precise value reported to date.

The infrared spectrum of HOOH^{+} trapped in solid neon
View Description Hide DescriptionWhen a Ne:H_{2}O_{2} mixture is codeposited at 4.3 K with a beam of neon atoms that have been excited in a microwave discharge, three new, photosensitive absorptions appear which can be assigned to the three infraredactive vibrational fundamentals of transHOOH^{+}. When the Ne:H_{2}O_{2} deposition system is pretreated with the vapors of D_{2}O, the product absorptions include new peaks which can be attributed to vibrational fundamentals of transHOOD^{+} and transDOOD^{+}. Density functional calculations of the vibrational fundamentals of the three hydrogen peroxide cation isotopologues support the proposed assignments. Broad, photosensitive product absorptions also appear near the positions of vibrational transitions of O_{3} ^{−}, and may be contributed by a weakly bound complex of that species with H_{2}O.

Vibrational modes of methane in the structure H clathrate hydrate from ab initio molecular dynamics simulation
View Description Hide DescriptionVibrational spectra of guest molecules in clathrate hydrates are frequently measured to determine the characteristic signatures of the molecular environment and dynamical properties of guesthost interactions. Here, we present results of our study on the vibrational frequencies of methane molecules in structure H clathrate hydrates, namely, in the 5^{12} and 4^{3}5^{6}6^{3} cages, as the frequencies of stretching vibrational modes in these environments are still unclear. The vibrational spectra of methane molecules in structure H clathrate hydrate were obtained from ab initio molecular dynamics simulation and computed from Fourier transform of autocorrelation functions for each distinct vibrational mode. The calculated symmetric and asymmetric stretching vibrational frequencies of methane molecules were found to be lower in the 4^{3}5^{6}6^{3} cages than in the 5^{12} cages (3.8 cm^{−1} for symmetric stretching and 6.0 cm^{−1} for asymmetric stretching). The C–H bond length and average distance between methane molecules and hostwater molecules in 4^{3}5^{6}6^{3} cages were slightly longer than those in the 5^{12} cages.

Electronic effects on the melting of small gallium clusters
View Description Hide DescriptionMotivated by experimental reports of higherthanbulk melting temperatures in small gallium clusters, we perform firstprinciples molecular dynamics simulations of Ga_{20} and using parallel tempering in the microcanonical ensemble. The respective specific heat (C_{ V }) curves, obtained using the multiple histogram method, exhibit a broad peak centered at approximately 740 and 610 K—well above the melting temperature of bulk gallium (303 K) and in reasonable agreement with experimental data for . Assessment of atomic mobility confirms the transition from solidlike to liquidlike states near the C_{ V } peak temperature. Parallel tempering molecular dynamics simulations yield lowenergy isomers that are ∼0.1 eV lower in energy than previously reported ground state structures, indicative of an energy landscape with multiple, competing lowenergy morphologies. Electronic structureanalysis shows no evidence of covalent bonding, yet both the neutral and charged clusters exhibit greaterthanbulk melting temperatures.
 Liquids, Glasses, and Crystals

A computational investigation of the phase behavior and capillary sublimation of water confined between nanoscale hydrophobic plates
View Description Hide DescriptionThin films of water under nanoscopic confinement are prevalent in natural and manufactured materials. To investigate the equilibrium and dynamic behavior of water in such environments, we perform molecular dynamics simulations of water confined between atomistically detailed hydrophobic plates at T = 298 K for pressures (−0.1) ⩽ P ⩽ 1.0 GPa and plate separations of 0.40 ⩽ d ⩽ 0.80 nm. From these simulations, we construct an expanded Pdphase diagram for confined water, and identify and characterize a previously unreported confined monolayerice morphology. We also study the decompressioninduced sublimation of bilayer ice in a d = 0.6 nm slit, employing principal component analysis to synthesize lowdimensional embeddings of the drying trajectories and develop insight into the sublimation mechanism. Drying is observed to proceed by the nucleation of a bridging vapor cavity at one corner of the crystalline slab, followed by expansion of the cavity along two edges of the plates, and the subsequent recession of the remaining promontory of bilayer crystal into the bulk fluid. Our findings have implications for the understanding of diverse phenomena in materials science, nanofluidics, and protein folding and aggregation.

Experimental studies of Debyelike process and structural relaxation in mixtures of 2ethyl1hexanol and 2ethyl1hexyl bromide
View Description Hide DescriptionBinary solutions of 2ethyl1hexanol (2E1H) with 2ethyl1hexyl bromide (2E1Br) are investigated by means of dielectric, shear mechanical, nearinfrared, and solvation spectroscopy as well as dielectrically monitored physical aging. For moderately diluted 2E1H the slow Debyelike process, which dominates the dielectric spectra of the neat monohydroxy alcohol, separates significantly from the αrelaxation. For example, the separation in equimolar mixtures amounts to four decades in frequency. This situation of highly resolved processes allows one to demonstrate unambiguously that physical aging is governed by the αprocess, but even under these ideal conditions the Debye process remains undetectable in shear mechanical experiments. Furthermore, the solvation experiments show that under constant charge conditions the microscopic polarization fluctuations take place on the time scale of the structural process. The hydrogenbond populations monitored via nearinfrared spectroscopy indicate the presence of a critical alcohol concentration, x _{ c } ≈ 0.5–0.6, thereby confirming the dielectric data. In the pure bromide a slow dielectric process of reduced intensity is present in addition to the main relaxation. This is taken as a sign of intermolecular cooperativity probably mediated via halogen bonds.