Solvation Models and 2' Combined QM MM Methods - PowerPoint PPT Presentation

1 / 22
About This Presentation
Title:

Solvation Models and 2' Combined QM MM Methods

Description:

Some use atom-centered charges, such as Mulliken charges, considered to be at the center of a sphere representing each atom as in the vdW model ... – PowerPoint PPT presentation

Number of Views:69
Avg rating:3.0/5.0
Slides: 23
Provided by: chem224
Category:

less

Transcript and Presenter's Notes

Title: Solvation Models and 2' Combined QM MM Methods


1
Solvation Models and 2. Combined QM / MM
Methods
  • See review article on Solvation by Cramer and
    Truhlar Chem. Rev. 99, 2161-2200 (1999)

2
Part 1. Solvation Models
  • Some describe explicit solvent molecules
  • Some treat solvent as a continuum
  • Some are hybrids of the above two
  • Treat first solvation sphere explicitly while
    treating surrounding solvent by a
    continuum model
  • These usually treat inner solvation shell quantum
    mechanically, outer solvation shell classically

Each of these models can be further subdivided
according to the theory involved classical (MM)
or quantum mechanical
3
Explicit QM Water Models
  • Sometimes as few as 3 explicit water molecules
    can be used to model a reaction adequately

Could use HF, DFT, MP2, CISD(T) or other theory.
4
Explicit Force Field Water Models
  • 3 types
  • rigid model, with interactions described by
    pairwise Coulombic and Lennard-Jones potentials
  • flexible models
  • polarizable models
  • TIP3P, TIP4P, and TIP5P rigid water models

(transferable intermolecular potentials, three
parameter)
5
Continuum (Reaction Field) Models
  • Consider solvent as a uniform polarizable medium
    of fixed dielectric constant e having a solute
    molecule M placed in a suitably shaped cavity.

e
6
Continuum Models
e
  • Creation of the cavity costs energy (i.e., it is
    destabilizing), whereas dispersion interactions
    between solute and solvent molecules add
    stabilization. The electronic charge of solute M
    polarizes the medium, inducing charge moments,
    which add electrostatic stabilization
  • DGsolvation DGcavity DGdispersion
    DGelectrostatic

7
Models Differ in 5 Aspects
e
  • Size and shape of the solute cavity
  • Method of calculating the cavity creation and the
    dispersion contributions
  • How the charge distribution of solute M is
    represented
  • Whether the solute M is described classically or
    quantum mechanically
  • How the dielectric medium e is described.

(these 5 aspects will be considered in turn on
the following slides)
8
1a. Solute Cavity Size and Shape
  • Spherical Ellipsoidal van der
    Waals
  • (Born) (Onsager)
    (Kirkwood)

r
r
9
1b. Solvent-Accessible Surface
  • A solvent-accessible surface is made by
    connecting the center of a rolling a sphere
    (e.g., 1.5 Å radius) around the vdW (electron
    density iso-) surface of the molecule.
  • This method excludes pockets
    that are inaccessible even to

    small solvent molecules.

10
2. Cavity Dispersion Energy
  • The energy required to create the cavity (entropy
    factors and loss of solvent-solvent vdW
    interactions) and the stabilization due to
    dispersion (vdW interactions, including some
    repulsion) are usually assumed to be proportional
    to the surface area of the solute M and the
    surface tension of the solvent.
  • These contributions may be treated as one term
    (proportional to the entire molecular volume) or
    as a sum of terms with each atom type
    having a different proportionality constant, such
    parameters derived by best fit to experimental
    solvation energies.

11
3a. Charge Distribution of Solute
  • Some use atom-centered charges, such as Mulliken
    charges, considered to be at the center of a
    sphere representing each atom as in the vdW model
  • Other approaches involve a dipole or multipole
    expansion (the simplest of these for a neutral
    molecule involves only the dipole moment).
  • Multipole expansions methods often need several
    orders (dipole, quadrupole, hexapole, octupole,
    decapole, etc.) for best results.

12
3b. Charge Distribution of Solute
(assuming vdW sized spheres for each atom)
r
(dipole polarizability)
(charge q)
(dipole m)
(this calculation is summed over all atoms)
13
4. Description of Solute M
  • Solute molecule M may be described by
  • classical molecular mechanics (MM)
  • semi-empirical quantum mechanics (SEQM),
  • ab initio quantum mechanics (QM)
  • density functional theory (DFT), or
  • post Hartree-Fock electron correlation methods
    (MP2 or CISDT).

14
5. Describing the Dielectric Medium
  • Usually taken to be a homogeneous static medium
    of constant dielectric constant e
  • May be allowed to have a dependence on
    the distance from the solute molecule M.
  • In some models, such as those used to model
    dynamic processes, the dielectric may depend
    on the rate of the process (e.g., the response of
    the solvent is different for a fast process
    such as an electronic excitation than for a
    slow process such as a molecular
    rearrangement.)

15
Example SM5.4/A in Titans AM1
(Chris Cramer, U. Minn)
  • Employs a generalized Born approximation with
    semiempirical parameter sets to represent
    solute-solvent interaction.
  • Cavity is made of interlocking spheres ( vdW
    surface)
  • Charge distribution of the solute is represented
    by atom-centered Mulliken charges

Born
16
Hybrid Solvation Model
Red highest level of theory (MP2, CISDT) Blue
intermed. level of theory (HF, AM1,
PM3) Black lowest level of theory (MM2, MMFF),
or Continuum
17
How Good are Solvation Models?
  • For neutral solutes, experimental free energies
    of solvation between the range of 5 to -15
    kcal/mol are measurable to an accuracy of 0.1
    kcal/mol.
  • Continuum models of solvation can calculate
    energies of solvation to within 0.7 kcal/mol on a
    large data set of neutral molecules.
  • The solvation energy of charged species can be
    measured only to accuracies of 5 kcal/mol.
    Computed solvation energies have similar errors.

18
Part 2. Hybrid QM/MM Methods
  • For problems such as modeling the mechanism of an
    enzyme, MM is not good enough and QM is too
    costly. A hybrid approach offers the best
    solution.
  • This method is essentially the same approach as
    is used in the hybrid model of solvation.

19
Hybrid QM/MM Methods
  • Hybrid QM/MM methods may employ any combination
    of high level theory (HF, MP2, DFT) to model a
    small, select part of the molecule, and any type
    of MM (MM3, MMFF) to model the rest of the
    structure.
  • the ONIUM method (next slide) in Gaussian 03
    allows several layers e.g., CISD(T), then HF or
    AM1, then MM for outside.

20
ONIUM (layering) Method
Red highest level of theory (MP2, CISDT) Blue
intermed. level of theory (HF, AM1,
PM3) Black lowest level of theory (MM2, MMFF)
catalytic triad of carboxypeptidase
21
Problems of Hybrid Approaches
  • The biggest problem is how to adequately model
    the interface or boundary between the QM-modeled
    region and the MM-modeled region.
  • In some respects it is the same problem faced in
    using explicit solvent molecules for the modeling
    the first solvation shell and a continuum model
    for more distant solvent molecules in the hybrid
    model.

22
Problems of Hybrid Approaches
  • One promising approach is to cleave bonds that
    occur at the interface between the layers and
    cap each end of the bond with a hypothetical
    hydrogen atom.
  • Hybrid QM/MM methodology such as the ONIUM method
    is experiencing increasing use and remarkable
    success in the solution of complex biochemical
    problems.
Write a Comment
User Comments (0)
About PowerShow.com