Built-in Models

(TD)DFT with Turbomole

Implementations of the interface between turbomole and mudslide. Turbomole provides electronic parameters such as energies, gradients, NAC coupling, etc to mudslide and mudslide performs molecular dynamics calculations

class TMModel(states: Type[ndarray], run_turbomole_dir: str = '.', workdir_stem: str = 'run_turbomole', representation: str = 'adiabatic', reference: Optional[Any] = None, expert: bool = False, turbomole_modules: Optional[Dict] = None)

Bases: ElectronicModel_

A class to handle the electronic model for excited state Turbomole calculations

call_turbomole(outname='turbo.out') → None

Call Turbomole to run the calculation

clone()

Create a copy of the electronics object that owns its own resources, including disk

Major difference from deepcopy is that clone should prepare any resources that are needed for the object to be used, for example, disk access, network access, and so on. deepcopy should only copy the object in memory, not acquire resources like disk space.

compute(X, couplings: Optional[Any] = None, gradients: Optional[Any] = None, reference: Optional[Any] = None) → None

Calls Turbomole/Turboparse to generate electronic properties including gradients, couplings, energies. For this to work, this class needs to know where the “original” files/data sit, so that they can get properly passed to Turbomole. (__init__() can get these file locations.)

setup_coords()

Setup the coordinates for the calculation

class TurboControl(control_file='control', workdir=None)

Bases: object

A class to handle the control file for turbomole

add_point_charges(coords: Type[ndarray], charges: Type[ndarray])

Add point charges to the control file

point_charges data group has the structure: $point_charges nocheck list pcgrad <x> <y> <z> <q> …

adg(dg, data, newline=False)

Convenience function to run add data group (adg) on a control

check_turbomole_is_installed()

Check that turbomole is installed, raise exception if not

cpc(dest)

Copy the control file and other files to a new directory

get_masses(symbols)

Get the masses of the atoms in the system

read_coords()

Read the coordinates from the control file

Returns:

(symbols, coords) where symbols is a list of atomic symbols and

coords is a numpy array of shape (n_atoms * 3) with coordinates in angstroms

read_hessian()

Projected Hessian has a structure of $hessian (projected) 1 1 0.000 0.000 0.000 0.000 0.000 1 2 0.000 0.000 0.000 0.000 0.000 2 1 0.000 0.000 0.000 0.000 0.000 …

read_point_charge_gradients()

Read point charges and gradients from the control file

point charges in dg $point_charges gradients in dg $point_charge_gradients

Returns: (coords, charges, gradients) where coords is a numpy array of shape (nq, 3) with coordinates in Bohr

run_single(module, stdout=<_io.TextIOWrapper name='<stdout>' mode='w' encoding='UTF-8'>)

Run a single turbomole module

sdg(dg, file=None, show_keyword=False, show_body=False, show_filename_only=False, discard_comments=True, quiet=False)

Convenience function to run show data group (sdg) on a control

use_weight_derivatives(use=True)

Check if weight derivatives are used in the control file

where_is_dg(dg, absolute_path=False)

Find which file a data group is in

Molecular Mechanics with OpenMM

class OpenMM(pdb, ff, system)

Bases: ElectronicModel_

OpenMM interface

compute(position: Type[ndarray], couplings: Optional[Any] = None, gradients: Optional[Any] = None, reference: Optional[Any] = None)

Compute energy and forces

openmm_is_installed()

Check if OpenMM is installed

Diabatic Scattering Models

Implementations of the one-dimensional two-state models Tully demonstrated FSSH on in Tully, J.C. <I>J. Chem. Phys.</I> 1990 <B>93</B> 1061.

class LinearVibronic(representation: str = 'adiabatic', reference: Optional[Any] = None, mass: float = [243.6078782, 134.6412667, 99.93022402, 66.33593369, 3475.98736], E1: float = 8.5037, E2: float = 9.4523, lamb: float = 0.3289, r0sqrtw5mh: float = 4.35, om: float = array([0.1117, 0.2021, 0.2723, 0.4102]), k1: float = array([-0.0456, 0.0399, -0.2139, -0.0864]), k2: float = array([-0.0393, 0.0463, 0.2877, -0.1352]), An: float = array([1.4823, -0.2191, 0.0525, -0.0118]))

Bases: DiabaticModel_

class ShinMetiu(representation: str = 'adiabatic', reference: Optional[Any] = None, nstates: int = 3, L: float = 19.0, Rf: float = 5.0, Rl: float = 3.1, Rr: float = 4.0, mass: float = 1836.0, m_el: float = 1.0, nel: int = 128, box: Optional[Any] = None)

Bases: AdiabaticModel_

V(R: Type[ndarray]) → Type[ndarray]

\(V(x)\)

dV(R: Type[ndarray]) → Type[ndarray]

\(\nabla V(x)\)

class Subotnik2D(representation: str = 'adiabatic', reference: Optional[Any] = None, a: float = 0.2, b: float = 0.6, c: float = 0.015, d: float = 0.3, f: float = 0.05, g: float = 0.3, w: float = 2.0, mass: float = 2000.0)

Bases: DiabaticModel_

V(r: Type[ndarray]) → Type[ndarray]

\(V(x)\)

dV(r: Type[ndarray]) → Type[ndarray]

\(\nabla V(x)\)

class SubotnikModelS(representation: str = 'adiabatic', reference: Optional[Any] = None, a: float = 0.015, b: float = 1.0, c: float = 0.005, d: float = 0.5, xp: float = 7.0, mass: float = 2000.0)

Bases: DiabaticModel_

V(x: Type[ndarray]) → Type[ndarray]

\(V(x)\)

dV(x: Type[ndarray]) → Type[ndarray]

\(\nabla V(x)\)

class SubotnikModelX(representation: str = 'adiabatic', reference: Optional[Any] = None, a: float = 0.03, b: float = 1.6, c: float = 0.005, xp: float = 7.0, mass: float = 2000.0)

Bases: DiabaticModel_

V(x: Type[ndarray]) → Type[ndarray]

\(V(x)\)

dV(x: Type[ndarray]) → Type[ndarray]

\(\nabla V(x)\)

class SuperExchange(representation: str = 'adiabatic', reference: Optional[Any] = None, v11: float = 0.0, v22: float = 0.01, v33: float = 0.005, v12: float = 0.001, v23: float = 0.01, mass: float = 2000.0)

Bases: DiabaticModel_

V(x: Type[ndarray]) → Type[ndarray]

\(V(x)\)

dV(x: Type[ndarray]) → Type[ndarray]

\(\nabla V(x)\)

class TullyDualAvoidedCrossing(representation: str = 'adiabatic', reference: Optional[Any] = None, a: float = 0.1, b: float = 0.28, c: float = 0.015, d: float = 0.06, e: float = 0.05, mass: float = 2000.0)

Bases: DiabaticModel_

Tunneling through a double avoided crossing used in Tully’s 1990 JCP

\[\begin{split}V_{11} &= 0 \\ V_{22} &= -A e^{-Bx^2} + E_0 \\ V_{12} &= V_{21} = C e^{-D x^2}\end{split}\]

V(X: Type[ndarray]) → Type[ndarray]

\(V(x)\)

dV(x: Type[ndarray]) → Type[ndarray]

\(\nabla V(x)\)

class TullyExtendedCouplingReflection(representation: str = 'adiabatic', reference: Optional[Any] = None, a: float = 0.0006, b: float = 0.1, c: float = 0.9, mass: float = 2000.0)

Bases: DiabaticModel_

Model with extended coupling and the possibility of reflection. The most challenging of the models used in Tully’s 1990 JCP

\[\begin{split}V_{11} &= A \\ V_{22} &= -A \\ V_{12} &= \left\{ \begin{array}{cr} B e^{Cx} & x < 0 \\ B \left( 2 - e^{-Cx} \right) & x > 0 \end{array} \right.\end{split}\]

V(X: Type[ndarray]) → Type[ndarray]

\(V(x)\)

dV(x: Type[ndarray]) → Type[ndarray]

\(\nabla V(x)\)

class TullySimpleAvoidedCrossing(representation: str = 'adiabatic', reference: Optional[Any] = None, a: float = 0.01, b: float = 1.6, c: float = 0.005, d: float = 1.0, mass: float = 2000.0)

Bases: DiabaticModel_

Tunneling through a single barrier model used in Tully’s 1990 JCP

\[\begin{split}V_{11} &= \left\{ \begin{array}{cr} A (1 - e^{Bx}) & x < 0 \\ -A (1 - e^{-Bx}) & x > 0 \end{array} \right. \\ V_{22} &= -V_{11} \\ V_{12} &= V_{21} = C e^{-D x^2}\end{split}\]

V(X: Type[ndarray]) → Type[ndarray]

\(V(x)\)

dV(x: Type[ndarray]) → Type[ndarray]

\(\nabla V(x)\)