The dynamics of the third level (upper curve) and the almost rational impasse su te (lower curve). Each of the zero-order states of (CHj)jCCCH molecule is characterized by 3IV - 6 * 42 vibrational quantum number, |clt oj, _c«j).
C hapter 2
Superexchange Mechanism of Intramolecular Vibrational Relaxation : Localization Properties
- Introduction
- Tier system and co n n ectivity o f vibrational states
- R esu lts
- C onclusion
- A cknow ledgem ents
In the present paper, the localization properties of the level system of (CHs^YCCH molecules, Y = C or Si, are studied with a method originally developed for random solids by Skinner and co-workers[ll, 12]. The coupling of the light state in the C compound extends to much longer distances along the plane coordinate than in the Si compound.
B ibliography
PE values for (CH3)3CCCH
C hapter 3
- Introd u ction
- C alculational D etails
- A nalysis of th e Spectra
- Concluding Remarks
In this case, the distribution of energy in the molecule is assumed to be "microcanonical" with the width of the distribution AE E. For example, the value for the interaction of the stretch with the CCH bend is the largest and is « 21cm-1 . This is the origin of the large splittings of about 21 cm-1 seen in the spectrum.
Because there is some uncertainty in the values of Xi*, we have made several calculations with different values for the anharmonic constants. 3, but rescaled so that the difference in the statistical weights of the peaks arising from v\ = 1 and v\ = 0 can be more easily observed.
C h ap ter 4
In the spectroscopic description, the highly off-resonant nature of the modes directly coupled to the bright mode leads to a narrow spectral line. The energy of each of the zero-order states in the basis set is calculated using the expression6. In Figure 6 we give a typical example of the small splitting that occurs due to the quartic couplings and the finite d a (note the significantly expanded, logarithmic, scale).
This result implies that the coupling of the bright state to the states in the initial levels is increased. The small peaks are due to direct (and weak) coupling of the bright state to states in the initial levels that are usually highly decoupled.
C h ap ter 5
A Periodic Boundary Conditions Formulation for Aqueous Solvent Dynamics
A P eriod ic B oundary C onditions im p lem en tation o f Solvent D ynam ics
In trod u ction
The nature of the solvent structure and its effect on and response to chemical change is a fundamental problem in chemical physics. For fast chemical processes in solution, the molecular nature of the solvent can be important at small distances from the solute. Some of the initial calculations used groups of particles with two-body Lennard-Jones type interactions to characterize the solvent.
In the PBC formalism, the molecules representing the solute and solvent are usually confined to a cubic cell and periodic boundary conditions apply. This isotropy occurs because the unit cell replicates infinitely, causing an unphysical sign change in the polarization of the solvent at the edge of the cell.
Equations of M otion
The main technical question that remains is the choice of a physically accurate potential for the interaction between particles and an integrator to propagate the above equations of motion. Although equation 5.1 is formally correct for use as a propagation equation, other algorithms give more accurate results, since appropriate linear combinations of similar equations for x(t — At ), x( t — 2A t. This fact is important as it allows larger values of At to be used for the actual calculations.
For the reaction field calculations, the MolDy package was significantly modified to solve the equations of motion for molecules confined to a spherical cavity surrounded by a dielectric continuum. The propagation equations used for both types of calculations are a variant of the well-known Verlet algorithm[20].
E xperim ental B ackground
For a rigid molecule with n atoms treated as mass points, the forces and torques can be determined from the forces exerted by each atom at each time step.
Generalized Solvent Coordinate
C om putational Strategies
- E quilibration from Initial C onfiguration
- Solvent S tru ctu re
- Solvent D y n a m ics from M D trajectories
The initial orientations of individual molecules are usually chosen randomly, and the initial velocities are obtained from the Maxwell-Boltzmann distribution. One simple measure that a system is in equilibrium is obtained by monitoring the total potential and kinetic energy of the system after scaling has stopped. Similar relationships exist between other equilibrium correlation functions and solvent response functions obtained from other step functional changes in the electrical properties of the solute.
It is also possible to obtain S{t) directly from MD calculations by simulating the step function change in the electronic property of the solute that is analogous to the experimental change in the solute caused by the interaction of the laser with the solute dust would be caused. . In this change, the solvent molecules are no longer in equilibrium with the new electrostatic state of the solute, as they were in equilibrium with the ground (or other initial) state of the solute.
Interm olecular P otential
Having obtained all the analytical machinery for propagating the dynamical equations and for analyzing the data, we now define the potential function we use to govern the dynamics. For a system of charges, the electrostatic potential is written as a sum of short- and long-range contributions, as in equation 13, where the former are written in real space and the latter in reciprocal space, where the periodicity of the MD cell is used. This choice of solvent–solvent interaction allows a comparison between the present results and those previously obtained by them.
In the previous calculations, a large spherical cluster of water molecules was used with a solute that had to be at the center of the cluster to model the solute-solvent system. In the calculations described here, periodic boundary conditions are used without any restrictions on the position of the solute.
R esu lts and D iscussion
- E nergy C onservation
- R ad ial D istrib u tion functions
- S olven t D ynam ics
Radial distribution functions were calculated from the instantaneous configurations of the molecules at different times during each trajectory. Figure 5.2 shows the temperature of the system monitored during a 50 ps simulation with 256 solvent molecules and one solute. We find that there is a slight upward increase in the total kinetic energy (or temperature) as a function of time, but even with a 50ps trajectory, the kinetic energy is conserved to within 5% of the total.
Conservation of total potential energy is even better than conservation of kinetic energy. Figure 5.3 shows the total potential energy of the system monitored during the 50 ps simulation.
- C onclusion
In this figure, some of the characteristic features of the solvent dynamics around a small, uncharged solute are clearly seen. The initial part of decay represents the "inertial" or rotational ("libration") motion of the solvent molecules around the solute and the last part represents the diffusive motion. It is also observed that the time-correlation functions for charged solutes are not very sensitive to the size of the solute[15].
Briefly, it is determined that the short time behavior of the solvent response function by the. In particular, the oscillations seen in Figure 5.11 after ~ 4 0 0 /s are an artefact of the small number of individual not.
C hapter 6
A Reaction Field Formulation of Solvent Dynamics
Introduction
This method is also more physically intuitive as it allows the separation of the solvent into the nearby molecular part and the distant dielectric part. Finally, with future advances in theory, the use of actual complete dielectric dispersion curves, e(w), for the continuum part of the solvent in MD simulations may be possible with this formulation. Some of these differences are in the potential energy function for the cluster, the boundary conditions to be imposed, methodology to calculate the radial distribution functions, and the position of the solute within the cluster.
In particular, the Ewald sum method, which was used to calculate the long-range Coulomb potential, replaced the reaction field potential. This implementation of the reaction field method differs from the one reported here because we have chosen to avoid using periodic boundary conditions.
R eaction Field P otential
There is no closed-form solution for the above Vrxn and our first approach was to use the Pade approximation [8] to the infinite series by explicitly calculating the first few (« 8) terms. It is useful to note that the above formulation of Vrxn is simply the generating function for Legendre polynomials, except for the multiplicative term, which is a function of e and I. If you wish to exploit the generating function for Legendre polynomials. nomiali, the multiplicative expression was simplified by expansion.
An analytical expression can be obtained for the two infinite sums in the previous equation. After some simple mathematical manipulations, we get The electrostatic forces at site j with charge qj are obtained directly from the potential using
L ennard-Jones Potential
B oundary C onditions
A simple formulation of this method would be to use a spherical cavity with reflecting boundary conditions, with the continuum starting immediately outside the cavity. From a technical point of view, if the continuum starts at the point to which the charged particles are allowed to arrive, then the reaction field potential becomes singular when the r r ,/a2. In the first method, the cavity with the water molecules was further surrounded by a shell of water molecules that were frozen in their positions.
The shell prevented the particles inside the cavity from approaching the continuum boundary, thus allowing the equations of motion for unfrozen particles to be propagated successfully. This approach was based on exploiting the excluded volume of molecules in the continuum, which prevents
R esu lts
- Solvational and Stru ctu ral Q uantities C alcu lated
- Frozen Shell R e su lts
- Stru ctu ral R esu lts from ‘LJ confinem ent’ sim ulations
The main output of the simulation is the time-dependent electrostatic potential acting on the solute due to the solvent, V (t) = V£t(t), or the solute-solvent interaction energy (Eint{t) = Efnt (t) + Ej^t (t)). The calculation method used was similar to the one described for the PBC system, slightly modified[1 2] due to the spherical nature of the cavity. In Figure 6.9, the convergence of the results as a function of cavity size is still evident.
The individual dipoles, approaching the solvent molecules, are more likely to be aligned perpendicular to the line joining the position of the dipole to the center of the cavity (cos# = 0) than to be aligned parallel to it (cos # = ±1) . First, the 0 - 0 RDF (<7oo(r)) is shown when there is a neutral or positively charged solute in the center of the cavity.
H RDFs from reaction field simulations 1.6
- Solvational R esu lts
^(r) is a bulk quantity that depends on all pairwise distances between individual oxygen atoms in the solvent molecules and should be negligibly affected by the presence of a single solute particle. The fact that these two curves shown above are almost identical indicates that, at least for cavities of size 9 A or larger, the solutes do not affect the properties of the bulk solvent. 1 2 show the extent to which a charged solute changes the orientational distribution function of the solvent (h(cos9, defined above)) and oxygen solute (O-X) RDFs when these quantities are compared to those derived from a neutral solute.
The first solvation shell is significantly more prominent and well defined for the charged solute when compared to the uncharged one. The orientational structure of the solvent molecules around the solute is made non-symmetric by the positively charged solute.
- C om parison o f R F R esu lts w ith Other results
- C onclusion
In Figure 6.13, the t c f for a neutral solute fixed in the center of the cavity is shown. The tc f shows all the characteristics of the solvent response function of an "atomic" type of solute. The later, slower and oscillatory decay of C(t) to 0 in s corresponds to the diffusive (or translational) motion of the solvent.
The tcf for the charged solute, shown in Figure 6.14, shows a slower decay than that of the neutral solute. The analysis of solvent dynamics for a moving solute in a cavity of limited size is not transparent.
B ib liograp h y
