4 |
|
\section{Introduction} |
5 |
|
|
6 |
|
%applications of langevin dynamics |
7 |
< |
As an excellent alternative to newtonian dynamics, Langevin |
8 |
< |
dynamics, which mimics a simple heat bath with stochastic and |
9 |
< |
dissipative forces, has been applied in a variety of studies. The |
10 |
< |
stochastic treatment of the solvent enables us to carry out |
11 |
< |
substantially longer time simulation. Implicit solvent Langevin |
12 |
< |
dynamics simulation of met-enkephalin not only outperforms explicit |
13 |
< |
solvent simulation on computation efficiency, but also agrees very |
14 |
< |
well with explicit solvent simulation on dynamics |
15 |
< |
properties\cite{Shen2002}. Recently, applying Langevin dynamics with |
16 |
< |
UNRES model, Liow and his coworkers suggest that protein folding |
17 |
< |
pathways can be possibly exploited within a reasonable amount of |
18 |
< |
time\cite{Liwo2005}. The stochastic nature of the Langevin dynamics |
19 |
< |
also enhances the sampling of the system and increases the |
20 |
< |
probability of crossing energy barrier\cite{Banerjee2004, Cui2003}. |
21 |
< |
Combining Langevin dynamics with Kramers's theory, Klimov and |
22 |
< |
Thirumalai identified the free-energy barrier by studying the |
23 |
< |
viscosity dependence of the protein folding rates\cite{Klimov1997}. |
24 |
< |
In order to account for solvent induced interactions missing from |
25 |
< |
implicit solvent model, Kaya incorporated desolvation free energy |
26 |
< |
barrier into implicit coarse-grained solvent model in protein |
27 |
< |
folding/unfolding study and discovered a higher free energy barrier |
28 |
< |
between the native and denatured states. Because of its stability |
29 |
< |
against noise, Langevin dynamics is very suitable for studying |
30 |
< |
remagnetization processes in various |
31 |
< |
systems\cite{Palacios1998,Berkov2002,Denisov2003}. For instance, the |
32 |
< |
oscillation power spectrum of nanoparticles from Langevin dynamics |
33 |
< |
simulation has the same peak frequencies for different wave |
34 |
< |
vectors,which recovers the property of magnetic excitations in small |
35 |
< |
finite structures\cite{Berkov2005a}. In an attempt to reduce the |
36 |
< |
computational cost of simulation, multiple time stepping (MTS) |
37 |
< |
methods have been introduced and have been of great interest to |
38 |
< |
macromolecule and protein community\cite{Tuckerman1992}. Relying on |
39 |
< |
the observation that forces between distant atoms generally |
40 |
< |
demonstrate slower fluctuations than forces between close atoms, MTS |
41 |
< |
method are generally implemented by evaluating the slowly |
42 |
< |
fluctuating forces less frequently than the fast ones. |
43 |
< |
Unfortunately, nonlinear instability resulting from increasing |
44 |
< |
timestep in MTS simulation have became a critical obstruction |
45 |
< |
preventing the long time simulation. Due to the coupling to the heat |
46 |
< |
bath, Langevin dynamics has been shown to be able to damp out the |
47 |
< |
resonance artifact more efficiently\cite{Sandu1999}. |
7 |
> |
As alternative to Newtonian dynamics, Langevin dynamics, which |
8 |
> |
mimics a simple heat bath with stochastic and dissipative forces, |
9 |
> |
has been applied in a variety of studies. The stochastic treatment |
10 |
> |
of the solvent enables us to carry out substantially longer time |
11 |
> |
simulations. Implicit solvent Langevin dynamics simulations of |
12 |
> |
met-enkephalin not only outperform explicit solvent simulations for |
13 |
> |
computational efficiency, but also agrees very well with explicit |
14 |
> |
solvent simulations for dynamical properties\cite{Shen2002}. |
15 |
> |
Recently, applying Langevin dynamics with the UNRES model, Liow and |
16 |
> |
his coworkers suggest that protein folding pathways can be possibly |
17 |
> |
explored within a reasonable amount of time\cite{Liwo2005}. The |
18 |
> |
stochastic nature of the Langevin dynamics also enhances the |
19 |
> |
sampling of the system and increases the probability of crossing |
20 |
> |
energy barriers\cite{Banerjee2004, Cui2003}. Combining Langevin |
21 |
> |
dynamics with Kramers's theory, Klimov and Thirumalai identified |
22 |
> |
free-energy barriers by studying the viscosity dependence of the |
23 |
> |
protein folding rates\cite{Klimov1997}. In order to account for |
24 |
> |
solvent induced interactions missing from implicit solvent model, |
25 |
> |
Kaya incorporated desolvation free energy barrier into implicit |
26 |
> |
coarse-grained solvent model in protein folding/unfolding studies |
27 |
> |
and discovered a higher free energy barrier between the native and |
28 |
> |
denatured states. Because of its stability against noise, Langevin |
29 |
> |
dynamics is very suitable for studying remagnetization processes in |
30 |
> |
various systems\cite{Palacios1998,Berkov2002,Denisov2003}. For |
31 |
> |
instance, the oscillation power spectrum of nanoparticles from |
32 |
> |
Langevin dynamics simulation has the same peak frequencies for |
33 |
> |
different wave vectors, which recovers the property of magnetic |
34 |
> |
excitations in small finite structures\cite{Berkov2005a}. |
35 |
|
|
36 |
|
%review langevin/browninan dynamics for arbitrarily shaped rigid body |
37 |
|
Combining Langevin or Brownian dynamics with rigid body dynamics, |
38 |
< |
one can study the slow processes in biomolecular systems. Modeling |
39 |
< |
the DNA as a chain of rigid spheres beads, which subject to harmonic |
40 |
< |
potentials as well as excluded volume potentials, Mielke and his |
41 |
< |
coworkers discover rapid superhelical stress generations from the |
42 |
< |
stochastic simulation of twin supercoiling DNA with response to |
43 |
< |
induced torques\cite{Mielke2004}. Membrane fusion is another key |
44 |
< |
biological process which controls a variety of physiological |
45 |
< |
functions, such as release of neurotransmitters \textit{etc}. A |
46 |
< |
typical fusion event happens on the time scale of millisecond, which |
47 |
< |
is impracticable to study using all atomistic model with newtonian |
48 |
< |
mechanics. With the help of coarse-grained rigid body model and |
49 |
< |
stochastic dynamics, the fusion pathways were exploited by many |
38 |
> |
one can study slow processes in biomolecular systems. Modeling DNA |
39 |
> |
as a chain of rigid beads, which are subject to harmonic potentials |
40 |
> |
as well as excluded volume potentials, Mielke and his coworkers |
41 |
> |
discovered rapid superhelical stress generations from the stochastic |
42 |
> |
simulation of twin supercoiling DNA with response to induced |
43 |
> |
torques\cite{Mielke2004}. Membrane fusion is another key biological |
44 |
> |
process which controls a variety of physiological functions, such as |
45 |
> |
release of neurotransmitters \textit{etc}. A typical fusion event |
46 |
> |
happens on the time scale of millisecond, which is impractical to |
47 |
> |
study using atomistic models with newtonian mechanics. With the help |
48 |
> |
of coarse-grained rigid body model and stochastic dynamics, the |
49 |
> |
fusion pathways were explored by many |
50 |
|
researchers\cite{Noguchi2001,Noguchi2002,Shillcock2005}. Due to the |
51 |
< |
difficulty of numerical integration of anisotropy rotation, most of |
52 |
< |
the rigid body models are simply modeled by sphere, cylinder, |
53 |
< |
ellipsoid or other regular shapes in stochastic simulations. In an |
54 |
< |
effort to account for the diffusion anisotropy of the arbitrary |
51 |
> |
difficulty of numerical integration of anisotropic rotation, most of |
52 |
> |
the rigid body models are simply modeled using spheres, cylinders, |
53 |
> |
ellipsoids or other regular shapes in stochastic simulations. In an |
54 |
> |
effort to account for the diffusion anisotropy of arbitrary |
55 |
|
particles, Fernandes and de la Torre improved the original Brownian |
56 |
|
dynamics simulation algorithm\cite{Ermak1978,Allison1991} by |
57 |
|
incorporating a generalized $6\times6$ diffusion tensor and |
58 |
|
introducing a simple rotation evolution scheme consisting of three |
59 |
|
consecutive rotations\cite{Fernandes2002}. Unfortunately, unexpected |
60 |
< |
error and bias are introduced into the system due to the arbitrary |
61 |
< |
order of applying the noncommuting rotation |
60 |
> |
errors and biases are introduced into the system due to the |
61 |
> |
arbitrary order of applying the noncommuting rotation |
62 |
|
operators\cite{Beard2003}. Based on the observation the momentum |
63 |
|
relaxation time is much less than the time step, one may ignore the |
64 |
< |
inertia in Brownian dynamics. However, assumption of the zero |
64 |
> |
inertia in Brownian dynamics. However, the assumption of zero |
65 |
|
average acceleration is not always true for cooperative motion which |
66 |
|
is common in protein motion. An inertial Brownian dynamics (IBD) was |
67 |
|
proposed to address this issue by adding an inertial correction |
73 |
|
dynamics for nonskew bodies without translation-rotation coupling by |
74 |
|
separating the translation and rotation motion and taking advantage |
75 |
|
of the analytical solution of hydrodynamics properties. However, |
76 |
< |
typical nonskew bodies like cylinder and ellipsoid are inadequate to |
77 |
< |
represent most complex macromolecule assemblies. These intricate |
76 |
> |
typical nonskew bodies like cylinders and ellipsoids are inadequate |
77 |
> |
to represent most complex macromolecule assemblies. These intricate |
78 |
|
molecules have been represented by a set of beads and their |
79 |
< |
hydrodynamics properties can be calculated using variant |
80 |
< |
hydrodynamic interaction tensors. |
79 |
> |
hydrodynamic properties can be calculated using variants on the |
80 |
> |
standard hydrodynamic interaction tensors. |
81 |
|
|
82 |
|
The goal of the present work is to develop a Langevin dynamics |
83 |
< |
algorithm for arbitrary rigid particles by integrating the accurate |
84 |
< |
estimation of friction tensor from hydrodynamics theory into the |
85 |
< |
sophisticated rigid body dynamics. |
83 |
> |
algorithm for arbitrary-shaped rigid particles by integrating the |
84 |
> |
accurate estimation of friction tensor from hydrodynamics theory |
85 |
> |
into the sophisticated rigid body dynamics algorithms. |
86 |
|
|
87 |
|
\section{Computational Methods{\label{methodSec}}} |
88 |
|
|
89 |
|
\subsection{\label{introSection:frictionTensor}Friction Tensor} |
90 |
|
Theoretically, the friction kernel can be determined using the |
91 |
< |
velocity autocorrelation function. However, this approach become |
92 |
< |
impractical when the system become more and more complicate. |
91 |
> |
velocity autocorrelation function. However, this approach becomes |
92 |
> |
impractical when the system becomes more and more complicated. |
93 |
|
Instead, various approaches based on hydrodynamics have been |
94 |
< |
developed to calculate the friction coefficients. In general, |
94 |
> |
developed to calculate the friction coefficients. In general, the |
95 |
|
friction tensor $\Xi$ is a $6\times 6$ matrix given by |
96 |
|
\[ |
97 |
|
\Xi = \left( {\begin{array}{*{20}c} |
99 |
|
{\Xi _{}^{tr} } & {\Xi _{}^{rr} } \\ |
100 |
|
\end{array}} \right). |
101 |
|
\] |
102 |
< |
Here, $ {\Xi^{tt} }$ and $ {\Xi^{rr} }$ are translational friction |
103 |
< |
tensor and rotational resistance (friction) tensor respectively, |
104 |
< |
while ${\Xi^{tr} }$ is translation-rotation coupling tensor and $ |
105 |
< |
{\Xi^{rt} }$ is rotation-translation coupling tensor. When a |
106 |
< |
particle moves in a fluid, it may experience friction force or |
107 |
< |
torque along the opposite direction of the velocity or angular |
108 |
< |
velocity, |
102 |
> |
Here, $ {\Xi^{tt} }$ and $ {\Xi^{rr} }$ are $3 \times 3$ |
103 |
> |
translational friction tensor and rotational resistance (friction) |
104 |
> |
tensor respectively, while ${\Xi^{tr} }$ is translation-rotation |
105 |
> |
coupling tensor and $ {\Xi^{rt} }$ is rotation-translation coupling |
106 |
> |
tensor. When a particle moves in a fluid, it may experience friction |
107 |
> |
force or torque along the opposite direction of the velocity or |
108 |
> |
angular velocity, |
109 |
|
\[ |
110 |
|
\left( \begin{array}{l} |
111 |
|
F_R \\ |
119 |
|
\end{array} \right) |
120 |
|
\] |
121 |
|
where $F_r$ is the friction force and $\tau _R$ is the friction |
122 |
< |
toque. |
122 |
> |
torque. |
123 |
|
|
124 |
< |
\subsubsection{\label{introSection:resistanceTensorRegular}\textbf{The Resistance Tensor for Regular Shape}} |
124 |
> |
\subsubsection{\label{introSection:resistanceTensorRegular}\textbf{The Resistance Tensor for Regular Shapes}} |
125 |
|
|
126 |
|
For a spherical particle with slip boundary conditions, the |
127 |
|
translational and rotational friction constant can be calculated |
142 |
|
\end{array}} \right) |
143 |
|
\] |
144 |
|
where $\eta$ is the viscosity of the solvent and $R$ is the |
145 |
< |
hydrodynamics radius. |
145 |
> |
hydrodynamic radius. |
146 |
|
|
147 |
< |
Other non-spherical shape, such as cylinder and ellipsoid |
148 |
< |
\textit{etc}, are widely used as reference for developing new |
149 |
< |
hydrodynamics theory, because their properties can be calculated |
150 |
< |
exactly. In 1936, Perrin extended Stokes's law to general |
151 |
< |
ellipsoids, also called a triaxial ellipsoid, which is given in |
152 |
< |
Cartesian coordinates by\cite{Perrin1934, Perrin1936} |
147 |
> |
Other non-spherical shapes, such as cylinders and ellipsoids, are |
148 |
> |
widely used as references for developing new hydrodynamics theory, |
149 |
> |
because their properties can be calculated exactly. In 1936, Perrin |
150 |
> |
extended Stokes's law to general ellipsoids, also called a triaxial |
151 |
> |
ellipsoid, which is given in Cartesian coordinates |
152 |
> |
by\cite{Perrin1934, Perrin1936} |
153 |
|
\[ |
154 |
|
\frac{{x^2 }}{{a^2 }} + \frac{{y^2 }}{{b^2 }} + \frac{{z^2 }}{{c^2 |
155 |
|
}} = 1 |
156 |
|
\] |
157 |
|
where the semi-axes are of lengths $a$, $b$, and $c$. Unfortunately, |
158 |
|
due to the complexity of the elliptic integral, only the ellipsoid |
159 |
< |
with the restriction of two axes having to be equal, \textit{i.e.} |
159 |
> |
with the restriction of two axes being equal, \textit{i.e.} |
160 |
|
prolate($ a \ge b = c$) and oblate ($ a < b = c $), can be solved |
161 |
|
exactly. Introducing an elliptic integral parameter $S$ for prolate |
162 |
< |
: |
162 |
> |
ellipsoids : |
163 |
|
\[ |
164 |
|
S = \frac{2}{{\sqrt {a^2 - b^2 } }}\ln \frac{{a + \sqrt {a^2 - b^2 |
165 |
|
} }}{b}, |
166 |
|
\] |
167 |
< |
and oblate : |
167 |
> |
and oblate ellipsoids: |
168 |
|
\[ |
169 |
|
S = \frac{2}{{\sqrt {b^2 - a^2 } }}arctg\frac{{\sqrt {b^2 - a^2 } |
170 |
< |
}}{a}. |
170 |
> |
}}{a}, |
171 |
|
\] |
172 |
|
one can write down the translational and rotational resistance |
173 |
|
tensors |
174 |
< |
\[ |
188 |
< |
\begin{array}{l} |
174 |
> |
\begin{eqnarray*} |
175 |
|
\Xi _a^{tt} = 16\pi \eta \frac{{a^2 - b^2 }}{{(2a^2 - b^2 )S - 2a}}. \\ |
176 |
|
\Xi _b^{tt} = \Xi _c^{tt} = 32\pi \eta \frac{{a^2 - b^2 }}{{(2a^2 - 3b^2 )S + |
177 |
|
2a}}, |
178 |
< |
\end{array} |
193 |
< |
\] |
178 |
> |
\end{eqnarray*} |
179 |
|
and |
180 |
< |
\[ |
196 |
< |
\begin{array}{l} |
180 |
> |
\begin{eqnarray*} |
181 |
|
\Xi _a^{rr} = \frac{{32\pi }}{3}\eta \frac{{(a^2 - b^2 )b^2 }}{{2a - b^2 S}}, \\ |
182 |
< |
\Xi _b^{rr} = \Xi _c^{rr} = \frac{{32\pi }}{3}\eta \frac{{(a^4 - b^4 )}}{{(2a^2 - b^2 )S - 2a}} \\ |
183 |
< |
\end{array}. |
200 |
< |
\] |
182 |
> |
\Xi _b^{rr} = \Xi _c^{rr} = \frac{{32\pi }}{3}\eta \frac{{(a^4 - b^4 )}}{{(2a^2 - b^2 )S - 2a}}. |
183 |
> |
\end{eqnarray*} |
184 |
|
|
185 |
|
\subsubsection{\label{introSection:resistanceTensorRegularArbitrary}\textbf{The Resistance Tensor for Arbitrary Shapes}} |
186 |
|
|
192 |
|
from all possible ellipsoidal spaces, $r$-space, to all possible |
193 |
|
combination of rotational diffusion coefficients, $D$-space, is not |
194 |
|
unique\cite{Wegener1979} as well as the intrinsic coupling between |
195 |
< |
translational and rotational motion of rigid body, general |
195 |
> |
translational and rotational motion of rigid bodies, general |
196 |
|
ellipsoids are not always suitable for modeling arbitrarily shaped |
197 |
< |
rigid molecule. A number of studies have been devoted to determine |
198 |
< |
the friction tensor for irregularly shaped rigid bodies using more |
199 |
< |
advanced method where the molecule of interest was modeled by |
200 |
< |
combinations of spheres(beads)\cite{Carrasco1999} and the |
197 |
> |
rigid molecules. A number of studies have been devoted to |
198 |
> |
determining the friction tensor for irregularly shaped rigid bodies |
199 |
> |
using more advanced methods where the molecule of interest was |
200 |
> |
modeled by a combinations of spheres\cite{Carrasco1999} and the |
201 |
|
hydrodynamics properties of the molecule can be calculated using the |
202 |
|
hydrodynamic interaction tensor. Let us consider a rigid assembly of |
203 |
< |
$N$ beads immersed in a continuous medium. Due to hydrodynamics |
203 |
> |
$N$ beads immersed in a continuous medium. Due to hydrodynamic |
204 |
|
interaction, the ``net'' velocity of $i$th bead, $v'_i$ is different |
205 |
|
than its unperturbed velocity $v_i$, |
206 |
|
\[ |
256 |
|
B_{ij} = \delta _{ij} \frac{I}{{6\pi \eta R}} + (1 - \delta _{ij} |
257 |
|
)T_{ij} |
258 |
|
\] |
259 |
< |
where $\delta _{ij}$ is Kronecker delta function. Inverting the $B$ |
260 |
< |
matrix, we obtain |
259 |
> |
where $\delta _{ij}$ is the Kronecker delta function. Inverting the |
260 |
> |
$B$ matrix, we obtain |
261 |
|
\[ |
262 |
|
C = B^{ - 1} = \left( {\begin{array}{*{20}c} |
263 |
|
{C_{11} } & \ldots & {C_{1N} } \\ |
266 |
|
\end{array}} \right), |
267 |
|
\] |
268 |
|
which can be partitioned into $N \times N$ $3 \times 3$ block |
269 |
< |
$C_{ij}$. With the help of $C_{ij}$ and skew matrix $U_i$ |
269 |
> |
$C_{ij}$. With the help of $C_{ij}$ and the skew matrix $U_i$ |
270 |
|
\[ |
271 |
|
U_i = \left( {\begin{array}{*{20}c} |
272 |
|
0 & { - z_i } & {y_i } \\ |
275 |
|
\end{array}} \right) |
276 |
|
\] |
277 |
|
where $x_i$, $y_i$, $z_i$ are the components of the vector joining |
278 |
< |
bead $i$ and origin $O$. Hence, the elements of resistance tensor at |
278 |
> |
bead $i$ and origin $O$, the elements of resistance tensor at |
279 |
|
arbitrary origin $O$ can be written as |
280 |
|
\begin{eqnarray} |
281 |
|
\Xi _{}^{tt} = \sum\limits_i {\sum\limits_j {C_{ij} } } \notag , \\ |
283 |
|
\Xi _{}^{rr} = - \sum\limits_i {\sum\limits_j {U_i C_{ij} } } U_j. \notag \\ |
284 |
|
\label{introEquation:ResistanceTensorArbitraryOrigin} |
285 |
|
\end{eqnarray} |
303 |
– |
|
286 |
|
The resistance tensor depends on the origin to which they refer. The |
287 |
< |
proper location for applying friction force is the center of |
287 |
> |
proper location for applying the friction force is the center of |
288 |
|
resistance (or center of reaction), at which the trace of rotational |
289 |
|
resistance tensor, $ \Xi ^{rr}$ reaches a minimum value. |
290 |
|
Mathematically, the center of resistance is defined as an unique |
291 |
|
point of the rigid body at which the translation-rotation coupling |
292 |
< |
tensor are symmetric, |
292 |
> |
tensors are symmetric, |
293 |
|
\begin{equation} |
294 |
|
\Xi^{tr} = \left( {\Xi^{tr} } \right)^T |
295 |
|
\label{introEquation:definitionCR} |
296 |
|
\end{equation} |
297 |
|
From Equation \ref{introEquation:ResistanceTensorArbitraryOrigin}, |
298 |
< |
we can easily find out that the translational resistance tensor is |
298 |
> |
we can easily derive that the translational resistance tensor is |
299 |
|
origin independent, while the rotational resistance tensor and |
300 |
|
translation-rotation coupling resistance tensor depend on the |
301 |
|
origin. Given the resistance tensor at an arbitrary origin $O$, and |
341 |
|
|
342 |
|
\subsection{Langevin Dynamics for Rigid Particles of Arbitrary Shape\label{LDRB}} |
343 |
|
|
344 |
< |
Consider a Langevin equation of motions in generalized coordinates |
344 |
> |
Consider the Langevin equations of motion in generalized coordinates |
345 |
|
\begin{equation} |
346 |
|
M_i \dot V_i (t) = F_{s,i} (t) + F_{f,i(t)} + F_{r,i} (t) |
347 |
|
\label{LDGeneralizedForm} |
358 |
|
in body-fixed frame and converted back to lab-fixed frame by: |
359 |
|
\[ |
360 |
|
\begin{array}{l} |
361 |
< |
F_{f,i}^l (t) = A^T F_{f,i}^b (t), \\ |
362 |
< |
F_{r,i}^l (t) = A^T F_{r,i}^b (t). \\ |
361 |
> |
F_{f,i}^l (t) = Q^T F_{f,i}^b (t), \\ |
362 |
> |
F_{r,i}^l (t) = Q^T F_{r,i}^b (t). \\ |
363 |
|
\end{array} |
364 |
|
\] |
365 |
|
Here, the body-fixed friction force $F_{r,i}^b$ is proportional to |
398 |
|
\tau _{r,i}^b = \tau _{r,i}^b +r_{MR} \times f_{r,i}^b |
399 |
|
\end{equation} |
400 |
|
where $r_{MR}$ is the vector from the center of mass to the center |
401 |
< |
of the resistance. Instead of integrating angular velocity in |
402 |
< |
lab-fixed frame, we consider the equation of motion of angular |
403 |
< |
momentum in body-fixed frame |
401 |
> |
of the resistance. Instead of integrating the angular velocity in |
402 |
> |
lab-fixed frame, we consider the equation of angular momentum in |
403 |
> |
body-fixed frame |
404 |
|
\begin{equation} |
405 |
|
\dot j_i (t) = \tau _{t,i} (t) = \tau _{s,i} (t) + \tau _{f,i}^b (t) |
406 |
|
+ \tau _{r,i}^b(t) |
421 |
|
{\bf j}\left(t + h / 2 \right) &\leftarrow {\bf j}(t) |
422 |
|
+ \frac{h}{2} {\bf \tau}^b(t), \\ |
423 |
|
% |
424 |
< |
\mathsf{A}(t + h) &\leftarrow \mathrm{rotate}\left( h {\bf j} |
424 |
> |
\mathsf{Q}(t + h) &\leftarrow \mathrm{rotate}\left( h {\bf j} |
425 |
|
(t + h / 2) \cdot \overleftrightarrow{\mathsf{I}}^{-1} \right). |
426 |
|
\end{align*} |
427 |
|
In this context, the $\mathrm{rotate}$ function is the reversible |
432 |
|
/ 2) \cdot \mathsf{G}_x(a_x /2), |
433 |
|
\end{equation} |
434 |
|
where each rotational propagator, $\mathsf{G}_\alpha(\theta)$, |
435 |
< |
rotates both the rotation matrix ($\mathsf{A}$) and the body-fixed |
435 |
> |
rotates both the rotation matrix ($\mathsf{Q}$) and the body-fixed |
436 |
|
angular momentum (${\bf j}$) by an angle $\theta$ around body-fixed |
437 |
|
axis $\alpha$, |
438 |
|
\begin{equation} |
439 |
|
\mathsf{G}_\alpha( \theta ) = \left\{ |
440 |
|
\begin{array}{lcl} |
441 |
< |
\mathsf{A}(t) & \leftarrow & \mathsf{A}(0) \cdot \mathsf{R}_\alpha(\theta)^T, \\ |
441 |
> |
\mathsf{Q}(t) & \leftarrow & \mathsf{Q}(0) \cdot \mathsf{R}_\alpha(\theta)^T, \\ |
442 |
|
{\bf j}(t) & \leftarrow & \mathsf{R}_\alpha(\theta) \cdot {\bf |
443 |
|
j}(0). |
444 |
|
\end{array} |
470 |
|
{\bf \tau}^{s}(t + h) &\leftarrow {\bf u}(t + h) |
471 |
|
\times \frac{\partial V}{\partial {\bf u}}, \\ |
472 |
|
% |
473 |
< |
{\bf \tau}^{b}(t + h) &\leftarrow \mathsf{A}(t + h) |
473 |
> |
{\bf \tau}^{b}(t + h) &\leftarrow \mathsf{Q}(t + h) |
474 |
|
\cdot {\bf \tau}^s(t + h). |
475 |
|
\end{align*} |
476 |
|
Once the forces and torques have been obtained at the new time step, |
490 |
|
\section{Results and Discussion} |
491 |
|
|
492 |
|
The Langevin algorithm described in previous section has been |
493 |
< |
implemented in {\sc oopse}\cite{Meineke2005} and applied to the |
494 |
< |
studies of kinetics and thermodynamic properties in several systems. |
493 |
> |
implemented in {\sc oopse}\cite{Meineke2005} and applied to studies |
494 |
> |
of the static and dynamic properties in several systems. |
495 |
|
|
496 |
|
\subsection{Temperature Control} |
497 |
|
|
502 |
|
the algorithm. In order to verify the stability of this new |
503 |
|
algorithm, a series of simulations are performed on system |
504 |
|
consisiting of 256 SSD water molecules with different viscosities. |
505 |
< |
Initial configuration for the simulations is taken from a 1ns NVT |
506 |
< |
simulation with a cubic box of 19.7166~\AA. All simulation are |
505 |
> |
The initial configuration for the simulations is taken from a 1ns |
506 |
> |
NVT simulation with a cubic box of 19.7166~\AA. All simulation are |
507 |
|
carried out with cutoff radius of 9~\AA and 2 fs time step for 1 ns |
508 |
< |
with reference temperature at 300~K. Average temperature as a |
508 |
> |
with reference temperature at 300~K. The average temperature as a |
509 |
|
function of $\eta$ is shown in Table \ref{langevin:viscosity} where |
510 |
|
the temperatures range from 303.04~K to 300.47~K for $\eta = 0.01 - |
511 |
|
1$ poise. The better temperature control at higher viscosity can be |
540 |
|
Langevin dynamics with viscosity of 0.1 poise demonstrates a faster |
541 |
|
scaling to the desire temperature. In extremely lower friction |
542 |
|
regime (when $ \eta \approx 0$), Langevin dynamics becomes normal |
543 |
< |
NVE (see green curve in Fig.~\ref{langevin:temperature}) which loses |
544 |
< |
the temperature control ability. |
543 |
> |
NVE (see orange curve in Fig.~\ref{langevin:temperature}) which |
544 |
> |
loses the temperature control ability. |
545 |
|
|
546 |
|
\begin{figure} |
547 |
|
\centering |
560 |
|
273~K. It has an equivalent implicit solvent system containing only |
561 |
|
two banana shaped molecules with viscosity of 0.289 center poise. To |
562 |
|
calculate the hydrodynamic properties of the banana shaped molecule, |
563 |
< |
we create a rough shell model (see Fig.~\ref{langevin:roughShell}), |
563 |
> |
we created a rough shell model (see Fig.~\ref{langevin:roughShell}), |
564 |
|
in which the banana shaped molecule is represented as a ``shell'' |
565 |
|
made of 2266 small identical beads with size of 0.3 \AA on the |
566 |
|
surface. Applying the procedure described in |
567 |
|
Sec.~\ref{introEquation:ResistanceTensorArbitraryOrigin}, we |
568 |
< |
identified the center of resistance at $(0, 0.7482, -0.1988)$, as |
569 |
< |
well as the resistance tensor, |
568 |
> |
identified the center of resistance at $(0 \AA, 0.7482 \AA, |
569 |
> |
-0.1988 \AA)$, as well as the resistance tensor, |
570 |
|
\[ |
571 |
|
\left( {\begin{array}{*{20}c} |
572 |
|
0.9261 & 0 & 0&0&0.08585&0.2057\\ |
588 |
|
%\end{array}} \right). |
589 |
|
%\] |
590 |
|
|
591 |
< |
Curves of velocity auto-correlation functions in |
591 |
> |
Curves of the velocity auto-correlation functions in |
592 |
|
Fig.~\ref{langevin:vacf} were shown to match each other very well. |
593 |
|
However, because of the stochastic nature, simulation using Langevin |
594 |
< |
dynamics was shown to decay slightly fast. In order to study the |
595 |
< |
rotational motion of the molecules, we also calculated the auto- |
596 |
< |
correlation function of the principle axis of the second GB |
597 |
< |
particle, $u$. |
594 |
> |
dynamics was shown to decay slightly faster than MD. In order to |
595 |
> |
study the rotational motion of the molecules, we also calculated the |
596 |
> |
auto-correlation function of the principle axis of the second GB |
597 |
> |
particle, $u$. The discrepancy shown in Fig.~\ref{langevin:uacf} was |
598 |
> |
probably due to the reason that the viscosity using in the |
599 |
> |
simulations only partially preserved the dynamics of the system. |
600 |
|
|
601 |
|
\begin{figure} |
602 |
|
\centering |
617 |
|
\centering |
618 |
|
\includegraphics[width=\linewidth]{vacf.eps} |
619 |
|
\caption[Plots of Velocity Auto-correlation Functions]{Velocity |
620 |
< |
auto-correlation functions in NVE (blue) and Langevin dynamics |
621 |
< |
(red).} \label{langevin:vacf} |
620 |
> |
auto-correlation functions of NVE (explicit solvent) in blue and |
621 |
> |
Langevin dynamics (implicit solvent) in red.} \label{langevin:vacf} |
622 |
|
\end{figure} |
623 |
|
|
624 |
|
\begin{figure} |
626 |
|
\includegraphics[width=\linewidth]{uacf.eps} |
627 |
|
\caption[Auto-correlation functions of the principle axis of the |
628 |
|
middle GB particle]{Auto-correlation functions of the principle axis |
629 |
< |
of the middle GB particle in NVE (blue) and Langevin dynamics |
630 |
< |
(red).} \label{langevin:twoBanana} |
629 |
> |
of the middle GB particle of NVE (blue) and Langevin dynamics |
630 |
> |
(red).} \label{langevin:uacf} |
631 |
|
\end{figure} |
632 |
|
|
633 |
|
\section{Conclusions} |