next up previous
Next: Conclusion Up: What is the Previous: Normal periodic boundaries

SHIFTED PERIODIC BOUNDARIES

The transition could be second order and the hysteresis might be due to a slow relaxation to the equilibrium state. The rate of relaxation is determined by the exchange of molecules between the smectic layers during compression from the nematic phase . This diffusion becomes inefficient at higher pressure. To circumvent this problem we have applied the shifted periodic boundary conditions to the Onsager system (see section 5.3.2.3) and determined the compression and expansion curves again. The results are plotted in figure 5.17. These equations of state seem to indicate that the nematic smectic transition is continuous in the Onsager limit. The continuity also follows from figure 5.18 where , our measure for nematic order, is plotted against the reduced density.
However, the shifted periodic boundaries suppress the the herringbone structured smectic fluctuations phase and inclusion of these fluctuations could make the transition (weakly) first order again.

 

Figure:  Equations-of-state around the nematic to smectic transition of the spherocylinder system in the limit . The solid line is the equation-of-state for the expansion from a smectic, the dotted line is the compression curve started from the nematic. The hysteresis has disappeared due to the use of shifted periodic boundaries.

 

Figure:  The measure for nematic order in an onsager spherocylinder system as a function of reduced denstity for a compression from the nematic (filled circles) and a expansion from the smectic phase (open circles). The hysteresis has disappeared due to the use of shifted periodic boundaries.

In short, we cannot establish with certainty the order of the N-S transition in the limit . To establish the nature of the transition, a full finite size scaling analysis is required. As the shifted periodic boundary conditions simulation offer better statistics, we tend to believe that the N-S transition is continuous. Since the transition is clearly first order for =5 there should be a tricritical point at intermediate . This tricritical point has been the subject of theoretical studies. In refs. [102,103] is estimated that the tricritical point corresponds to 5 while the theoretical analysis in ref [104] suggest that it should occur at =50. The first prediction is clearly incompatible with our simulation results and those of ref [27]

 

Figure:  The phase diagram of hard spherocylinders plotted as a function of . In this way we can establish the link between the finite length region and the Onsager limit at .



next up previous
Next: Conclusion Up: What is the Previous: Normal periodic boundaries



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