next up previous
Next: About this document Up: Rescaling of diffusion coefficients Previous: The Selkov model

VI. The FIS reaction

We now analyze the case of the FIS reaction, which was originally studied in gif and later in gif, where a variety of patterns was observed. We will consider the four variable model presented in gif, which is given by the following set of reactions

       eqnarray1146

where tex2html_wrap_inline2472tex2html_wrap_inline2474tex2html_wrap_inline2476 and tex2html_wrap_inline2478 . Including diffusion, we get from this model the following dynamical equations:

     eqnarray1175

where

     eqnarray1197

and

     eqnarray1214

In writing Eqs. (94)-(97) we are assuming that the species V and A are fed into the system and this happens at the same rate at which all of the species are removed ( tex2html_wrap_inline2484 ).

In this Section we apply the methods of Sec. V of the paper to Eqs. (86)-(89) to obtain the equations describing the slow time dynamics for this model. In order to make this calculation more specific, we consider the parameter values that are used in the experiments when replicating spot pattern is observed: tex2html_wrap_inline2486tex2html_wrap_inline2488tex2html_wrap_inline2490tex2html_wrap_inline2492tex2html_wrap_inline2494tex2html_wrap_inline2496 . It is not completely clear what the values of tex2html_wrap_inline2498 and tex2html_wrap_inline2500 actually are inside the gel where the reaction takes place. We will consider tex2html_wrap_inline2502 and tex2html_wrap_inline2504 . Inspired by the reduction of the ODE's performed by Gáspár and Showalter, we seek a reduction of the PDE's (86)-(89) in which the variables a and z are eliminated in favor of u and v. To this end we follow the steps of Sec. V of the paper considering tex2html_wrap_inline2516tex2html_wrap_inline2518tex2html_wrap_inline2520 and tex2html_wrap_inline2522 . We find:

   eqnarray1245

   eqnarray1266

Since the Z species is iodine, which binds to the gel, its diffusion coefficient may be neglected. For the sake of simplicity, we will also neglect the diffusion term of A. Inserting the expansions tex2html_wrap_inline2528 and tex2html_wrap_inline2530 in Eqs. (86)-(87) we find:

   eqnarray1294

where we have defined tex2html_wrap_inline2532tex2html_wrap_inline2534tex2html_wrap_inline2536tex2html_wrap_inline2538tex2html_wrap_inline2540tex2html_wrap_inline2542 and tex2html_wrap_inline2544 . Now, the whole calculation is consistent provided that tex2html_wrap_inline2546 and tex2html_wrap_inline2548 . As we may see from Eqs. (99) and (101), both tex2html_wrap_inline2550 and tex2html_wrap_inline2552 contain terms that are proportional to tex2html_wrap_inline2554 and tex2html_wrap_inline2556 . These time derivatives may get large given that Eqs. (102)-(103) contain terms which are proportional to tex2html_wrap_inline2558 . However, the existence of more than two timescales is of help in this case. Assuming that the right-hand-side of Eqs. (99), (101), (102) and (103) are dominated by the terms proportional to tex2html_wrap_inline2558 , we may rewrite the conditions tex2html_wrap_inline2546 and tex2html_wrap_inline2548 as:

   eqnarray1396

For the parameter values we are considering it is tex2html_wrap_inline2566 and tex2html_wrap_inline2568 . Thus, provided that v/u does not become too large, the conditions (105) and (106) are satisfied and the calculation is self-consistent.


next up previous
Next: About this document Up: Rescaling of diffusion coefficients Previous: The Selkov model
Silvina Ponce Dawson

Fri Aug 27 03:50:25 ART 1999