next up previous
Next: Simulation Up: Spherocylinders with attraction Previous: Spherocylinders with attraction

THE ANGLE DEPENDENT PAIR POTENTIAL

  In the previous section we used perturbation theory to estimate the phase behavior of the Asakura-Oosawa model for rod-polymer mixtures. The crucial assumption in the perturbation theory is that we can use the properties of the hard spherocylinder reference fluid to compute the free-volume fraction at finite fugacity. To check if this assumption is justified, we have to perform simulations of the full Asakura-Oosawa model at finite z. Although this is certainly feasible for individual state points, such calculations would become prohibitively expensive if we wish to compute phase diagrams for different values of q and different aspect ratios of the rods. The reason is that, in a full (off-lattice) simulation of the AO model, every polymer has to be simulated separately. The number of particles needed for such simulation is quite large, certainly for the small polymers at high fugacities. As an example we can estimate the number of polymers in a nematic phase of spherocylinders of =5 at a density =0.5 in equilibrium with a reservoir with polymers of diameter =0.15. As can be seen in figure 6.2, the polymer fugacity can easily be equal to =10. The free volume fraction for this density and q-value is approximately 0.4. This corresponds to an equilibrium polymer density in the nematic phase of 4. As the volume of the simulation box has to be larger than to avoid multiple overlaps of the spherocylinders, the minimum number of polymers needed in such is simulation is . This is a lower bound, as the polymers tend to drive the spherocylinders together and create more space for themselves. Moreover, due to the large particle concentration, the acceptance probability of moving a spherocylinder in a sea of polymers is dropping dramatically.

To avoid simulating the full polymer-colloid model, while retaining the effect of the depletion interaction on the structure of the colloidal suspension, we constructed a model of hard spherocylinders with a pairwise additive attractive ``depletion'' potential that approximates the real potential of mean force W as introduced in chapter 1. The most important feature of this potential is that it must be dependent on relative orientation of the rods. The depletion attraction is strong () when the spherocylinders are parallel, but much weaker () for perpendicular rods.

The most direct way to compute the depletion pair potential is to calculate the overlap volume between the depletion zones of the two spherocylinders as a function of distance and relative orientation, and multiply this volume by the polymer osmotic pressure. This procedure yields an effective pair potential that can be used in simulations of the phase diagram. In the case of spherical colloids mixed with small polymers, this approach predicts a phase diagram that is in reasonable agreement with the ``exact'' phase diagram [10,6].

 

Figure: Geometrical representation of the computation of the effective depletion interaction between two spherocylinders (eqn. 6.1). On the left, the situation for two perpendicular spherocylinders is depicted. The and axes denote the distance from the center of mass along a cylinder axis. The circle indicates the points for which the distance equals . Inside this circle the distance is smaller than . As the integral in eqn. 6.1 is bounded by the length of the spherocylinders, the total interaction is proportional to the cross section of the square and the ellipse. On the right, the angle between the spherocylinder is less than 90 degrees. The circle therefore changes into an ellipse, with its major axes along the lines and . The interaction is still proportional to the area bounded by the ellipse and the condition and .

 

Figure: Generalized square-well interaction potentials , as a function of distance r for various orientations and center-of-mass distances of two spherocylinders with =5. One spherocylinder is fixed at the origin oriented along the z-axis, while the position and orientation of the second is varied. a) Parallel spherocylinders, with the center of mass of the second spherocylinder shifted along the z-direction. From the bottom to the top the curves denote potentials for a shift of z/D =0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5 and 5.0. b) Perpendicular spherocylinders, with the center of mass of the second spherocylinder shifted along the z-direction. From the bottom to the top the curves denote potentials for a shift of z/D =0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5 and 5.0. c) Two spherocylinders with zero z-shift, the second one rotated around the line through the centers of mass over an angle . From the bottom to the top the curves denote potentials for an angle of = 0, 10, 20, 30, 40, 50, 60, 70, 80 and 90 degrees. d) Two spherocylinders with zero z-shift, the second one rotated over an angle in the plane through the line connecting the centers of mass and the z-axis. From the bottom to the top the curves denote potentials for an angle of = 0, 10, 20, 30, 40, 50, 60, 70, 80 and 90 degrees.


Although the depletion pair-potential for spherocylinders could be computed in the same way, the resulting function would not be cheap to compute during a simulation. As the computation of the pair potential is the most time-consuming part of the program, it is useful to devise an effective pair potential that resembles the depletion interaction, but is computationally cheap. To this end, we assume that the spherocylinder can be considered as a continuous distribution of (overlapping) spheres of diameter D, with their centers distributed uniformly on a line segment of length L. The individual spheres interact through a square-well potential. The total pair potential of the two spherocylinders is then a sum (or actually, an integral) of all the individual contributions, and is expressed as:

 

Here denotes the relative position of the centers of mass of the two hard spherocylinders i and j, with orientations specified by the unit vectors and , respectively. The range of depletion interaction () is a measure for the polymer diameter . The well depth can be interpreted as a measure for the polymer concentration. The Heaviside step-function is defined in the usual manner:

The evaluation of the integral in eqn. 6.1 can be reduced to a simple geometrical problem. First, consider the interaction between two infinitely long cylinders. If the cylinders are mutually perpendicular, the integral will reduce to the calculation of the surface of a circle with radius , where is the minimum distance between the lines that run through the cylinder axes. As the angle between the cylinders becomes smaller, this circle changes into an ellipse by scaling one of the ellipse axes by . The integral diverges for parallel, infinite cylinders, because every point along the cylinder gives a finite contribution. However, the integral in eqn. 6.1 is bounded by the finite length of the spherocylinders. This is translated in the geometrical picture as cuts through the ellipse, as shown in figure 6.3. If we multiply the area enclosed by the ellipse and/or the cuts of the bounds by the well depth we obtain the total interaction potential.

A quantity of particular interest is the maximum interaction between two spherocylinders. The interaction between two particles is largest when they are oriented parallel with their centers of mass at the minimum distance (D).

 

In figure 6.4 we show the behavior of this pair potential for a few relative orientations and distances of the two spherocylinders.



next up previous
Next: Simulation Up: Spherocylinders with attraction Previous: Spherocylinders with attraction



Peter Bolhuis
Tue Sep 24 20:44:02 MDT 1996