What do you mean by stress constant? I know that the solid mechanics module can give you the stress field map for your structure, but if you are trying to find the birefringence dependence on the stress, I believe that COMSOL is not a good way to do that.

I say that because this is a material property and COMSOL usually gets this information as an input. Usually people use COMSOL to obtain the expected effect of a material property in a certain geometry-physics combination.For instance, in my case I calculate how stress and boundary movement change optical properties (dispersion) of a given structure. For the stress contribution I must create a material with photo-elastic properties, so I have to know that to start the simulation.

Now, if you want the stress field map for a PCF mechanical mode, you should give the solid mechanics module a try. You may start with a 1/N section of the fiber, where N is the rotation symmetry of your structure (N=6 for an hexagon), in a 2-D geometry. Set symmetry boundaries and search for eigenfrequencies around the expected value for your modes of interest. With that you filter out many modes with polarization orthogonal to the ones your are interested in.

Regards,

Gustavo

Thank you for the reply.

I have tried with the structural mechanics module in COMSOL. There I have created a 2D PCF structure which as you know is combination of air holes and silica material. I have defined the youngs modulus and poisson ratio of silica. But for air, i have given the bulk modulus (since youngs modulus not there) and if given zero it will give a singularity error.

Further a fixed constraint is provided to half of the boundry (semicircle) and a boundary load to the rest of the boundary (next semicircle).

After simulation, I am taking the stress coefficient along x and y axis i.e sigma-x and sigma_y. The difference of this sigma_x and sigma_y gives the birefringence due to stress.

B(due to stress)= (C1-C2)(SIGMA_X-SIGMA_Y)

dB/dP=(B0-Bstress)/P, B0 is intrinsic birefringence and P is the pressure. This gives the birefringence dependence on stress.

Is this the right way to go? Please comment

I don't work with PCF, so I will make some questions that may seem obvious to you.

Do you need to have air in your simulation? Or better yet, are you also simulating the optical field? If not, I would draw the PCF without material in the holes, with that you need only the silica proporties. For the mechanical structure module that is enough.

I think I understood what you are doing. So you really need to determine the stress field to find the birefringence. Can't you do the same with the eigenfrequency solver? Or are you really interested in the effects of an externally applied pressure on your fiber?

About the birefringence calculation, it seems OK for me. But if you need more help with COMSOL, I will be glad to help. I have been using it for a while now and I am well familiar with the mechanical models, specially for eigenfrequency problems

Thank you ,

The presence of asymmetric holes in the PCF structure is the reason for the increased stress, this is what i think. So, compromising holes from the structure  cannot be a good idea as different arrangement of hole gives different birefringence stress sensitivity. I am trying to calculate the variation of birefringence with respect to the external pressure applied so that this fiber can be fabricated and put for pressure sensing using PM-PCF SAGNAC.

I have no idea about the eigen frequency solver. I am also new to the structural mechanics module, as I have been working with the RF module only for optical design.

If you giving me a little bit idea about the eigen frequency solver and its role in my application.

It would have been great.

I understand, your PCF on COMSOL is not uniform. So there will be no easy way using symmetry to reduce the problem.

I presume you used the eigenmode solver for the RF model. The eigenfrequency is basically the same, but instead of searching for the eigenmode (eigenvector) you search for the eigenfrequencies (eigenvalues). In the former you have to guess a propagation constant (effective index) for a given frequency/eavelength, while in the latter you guess a frequency for a given propagation constant.

Both solvers give basically the same solution, the eigenmodes and/or eigenfrequencies of the structure, be it for the mechanical or the EM problem. In your case I don't think this solvers will help, as a steady pressure usually will "excite" a combination of various eigenmodes (Fourier decomposition).

But what about the need to have air in the simulation. Are you using multiphysics, simulating both mechanical and optical fields and combining them? If you are simulating them separately, it would facilitate to remove the material in the holes for the mechanical case. For what you said you can estimate the stress-dependent birefringence only from the mechanical simulation, so that's what I would do.

Also, if you expect to have small displacement due to the applied pressure, you may as well use perturbation theory. I recommend Steven Johnson's paper about EM perturbation theory for moving boundary optical structures: http://journals.aps.org/pre/abstract/10.1103/PhysRevE.65.066611

I am simulating the mechanical model seperately and using the stress constants from there, I am doing some numerical calculations for finding out the birefringence pressure sensitivity.