Discussion Forum

Hi Karima,

I've never worked with fluorescent materials, but this sounds really interesting! Both of your ideas seems good. Did you had any success applying them on your model?

Another approach that may help you, is to add an imaginary part to the refractive index of the fluorescent material. Normally, this imaginary part accounts for the absorption losses of the material. But, you can switch the sign so it accounts for gain instead of losses. The problem is, this approach will amplify every wavelength impinging on your fluorescent material. But there are tweaks to solve this. So please tell me, which kind of source are you use (or need) to excite the fluorescence?

Best!

Hii Felipe,

Sorry for this late response! I was having some difficulties with COMSOL and WaveOptics, I did manage to add a complex refractive index to my material. Which did as you predicted, all wavelengths were amplified in this manner. However I also want the material to absorb one part of the spectrum (UV light) and emit the light in another part (red or blue), the so-called Stokes Shift. The conceptual idea is to excite the material with sunlight and have the fluorescent material absorb the UV and re-emit it in the visible region, preferably blue or red. Currently I use a port condition as exciting port and in the pre-defined Wavelength Domain module of the WaveOptics, Frequency domain I use a built in sweep in wavelengths to imitate the solar spectrum, ranging it from 300nm to 700 nm. Hope this paints the picture better of what I'm trying to achieve with COMSOL.

What were your ideas on tweaking this problem with the imaginary k of the refractive index? I'm very curious!!

Karima,

you can make the index a function of the wavelength. Would that help?

Cheers Edgar

Hi Edgar,

Thank you for your response! I did actually make my refractive index wavelength dependent currently in the non-fluorescent material using experimental data from my research group. The main issue I think for me for the fluorescent material is how to incorporate the fact that the absorption takes place in a different part of the spectrum than the emission.

Could this be done in two subsequent study steps? First step calculates absorption and second step calculates emission using a Stoke's-shifted index derived from the first step. It seems that the problem is one-way coupled only?

Aha! That sounds like a bright idea Edgar, if I understood it correctly: I would create first a positive imaginary k (maybe as a file under materials-rfi?) which would absorb in the UV region and then create a negative imaginary k which would simulate emission in the red/blue region? However how do I tell each Study Step what k to use?

There are different ways to achieve that. You can use a variable k, that may be defined by an interpolation function for absorption and emission in two separate variable nodes for the two ks. Same variable name in both nodes. In the study step you can then tick the 'Modify model configuration for study step' check box and activate/deactivate the respective variable node. You can also set up the physics node twice with suitable k-assignments and run one physics in one node, the other in the second node. There may be more options like using boolean switches, such as: absorptionk1(freq) + emissionk2(freq). You set the parameters absorption and emission to 0 or 1 as needed.

The optimal way may depend on other requirements in your model.

Hi Edgar,

I tried out one of your ideas today, namely the one where I create two functions for different k's. However I stumbled upon a problem when I created the second study step, since this only allows me to turn on/off the nodes under the Wave Optics node. I think this is because I have defined my luminescent material as a geometrical layer in the section Geometry and specified its refractive index under Materials, both of these nodes (so geometry and materials) I am unable to (de)-activate when I select the 'Modify model configuration for study step' check box in the second study step.

Would you have any other suggestions for me to work around this issue?

Hi Karima,

you can also put the material property directly into the respective physics node. You may need to add a separate physics node for this layer. This will then overide the settings in the material

Hi Karima and Edgar,

I'm enjoying a lot reading this thread. Thanks for sharing!

Karima, can you please share the functions you are using with us? And also, can you tell us if these functions come from (i.e. theory or a experimental fit)?

Best

Hii Felipe and Edgar,

Filipe, welcome back to the thread haha:) I am currently using simplified versions for the absorption/emission functions, my research group has experimental data of the actual refractive index (both n and k). But I first wanted to try with simple Gaussian distributed functions, its parameters (mu and sigma) are not completely made up: the mean is obtained from the experiments I mentioned before and the FWHM as well. The only difference is that the experimental data I will use at a later stage does not follow a perfect Gaussian distribution like I'm simulating now. I used an analytic function to define these k absorbing/emitting, which look like: .

@Edgar creating a second "EM Wave, Frequency Domain" node and only including the fluorescent layer as a domain did not work for me. Since COMSOL does not seems to understand what I mean when I put in mat3.rfi.k_emit as material property under the second physics node, it only understands mat3.rfi.ki_iso. I did define both material functions k_abs and k_emit but chose k_abs as the ki_iso under the material node, since this is needed for the first physics node. So it seems that when I did that, the expression mat3.rfi.k_emit became unaccessable. This made no sense to me and this is when I stopped.

Karima,

not sure what is going wrong there. Things may be easier if you can post your model and the expressions you want to use on the respective domains and study steps.

Cheers Edgar

Hi Edgar,

Of course, I attached the file below! Thank you for your help, it means a lot:)

I read up on some literature and I think that maybe using a k_emit is not the accurate way to mimic the actual physical process at hand. Since with this method the impinging light is amplified for a specific wavelength (or range of wavelengths), it will not return any output any if the incoming light does not contain that specified wavelength. I don't want to amplify incoming light, but actually 'create' it. So in theory, when a monochromatic beam with a wavelength in the UV region hits my material slab, it should be absorbed and translated to an emission in the visible spectrum. I don't think this can be achieved with a k_emission as a material property in a second physics node. I thought of maybe having the fluorescent material as a whole still absorb the UV light but then have COMSOL 'create' the emission light with an electric dipole point, do you think this would be possible in COMSOL?

Hi Karima,

I modified your model to show how you can run study steps and switch variable and physics nodes in the study step settings. I put the n_layer and k_layer variable for the fluorescent layer definition in two global variable nodes. They are a bit easier to reference then in the wee2 and wee3 physics nodes for the fluorescent layer. The ewfd physics node contains two wee-nodes, one for absorption and one for emission.

I study step1 for absorption you see the respective variable node and wee-physics node for absorption active. Analog for study step2 for emission.

You are right, this not yet solving the complete problem. One approach may be to determine absorption as a function of wavelength in the first step and excite the transmit port in the second step by a function that is a wavelength shifted function some way derived from absorption. This may then as well require to switch between two port definitions in the study sequence. There is certainly more than one option to achieve the goal. I guess you will now need some time to play and try things.

Cheers Edgar