Learning Forum

[smajety]

Hello,

I am running a FDTD simulation of a waveguide with dipole source (broadband). I am estimating the amount of emission coupled to the fundamental TE mode of the waveguide at each wavelength. I am aware that the dipole power needs to be normalized to account for the purcell enhancement in a non-homogenous medium. I am using the ratio (dipolepower/sourcepower) to normalize. But when I do a broadband simulation, there are certain wavelengths for which the transmission is larger than 1 which is physically not possible. Is the normalizing factor accurate? Is there any other approach for broadband simulation?

Thank you.

[Guilin Sun]

 

Did you divide the transmission by the Purcell factor (dipolepower/sourcepower) or multiplied it?  It should be divided. If the result is correctly renormalized, there might be two causes:

1: the dipole power may not be accurate when the dipole is located in lossy material, or in small mesh size. Please search dipolepower in Lumerical website to get more information, as the current forum has difficulty with a website link.

2: When you use broadband dipole to extract excited mode information, caution ay be taken that, the mode profile is frequency dependent.

I guess your result might be related to the mode profiles that are at one frequency but the decomposition is for broadband.

 

 

 

[Guilin Sun]

For #2, please use mode expansion monitor and set the recording frequency to be broadband.

[smajety]

My elaborate comments with plots are not posting for some reason.

Yes, I am using a mode expansion monitor and dividing by purcell factor. When I use a finer mesh, the transmission values are still greater than one.

I would also like to point out that in the Lumerical youtube video "Dipole Source in Ansys Lumerical FDTD – Lesson 6, Part 1", they mention that the purcell factor needs to be multiplied for renormalization. What do you think about that?

I was unable to find the setting "set the recording frequency to be broadband" in my mode expansion monitor. It would be helpful if you could share a screenshot of the settings window where I can find that. 

Thank you.

[smajety]

Hello Guilin

[Guilin Sun]

For Purcell factor, if you use the result from the dipole source, or using dipolepower/sourcepower, when the mesh is finer, or the dipole is located at a lossy material, the analytical dipolepower is not accurate: https://optics.ansys.com/hc/en-us/articles/360034925293-dipolepower-Script-command

You will need to use box monitor to modify the actual dipole power. If it is in lossy material, the box monitor dimension needs to be carefully set: too large the dimension will lower the actual radiated power due to abosorption; too small the dimension may affacts the accuracy of power integegration.

As for the mode expansion monitor, there is a box "frequency points" where you can specify the number of frequency points to match those in DFT monitor. Sometimes the forum does not accept the image so I will try to paste it in next reply.

 

 

[Guilin Sun]

 

You may also set the Global monitor frequency points which will ensure all the frequency-domain monitors, include the expansion monitor to have the same frequency points.

[smajety]

Unable to post responses here

[smajety]

Only very short responses are allowed. Please advise.

[Guilin Sun]

Now it should work properly. Please try and let me know if there is still any issue.

 

[smajety]

So, you mention that in a lossless dielectric (k=0) the value obtained from the dipolesource(f) is the accurate value and transmission box is not needed. I think my optimization sweeps agree with that statement (I am using a dipole source with a fixed emission frequency not a range of frequencies). But my initial question was with regards to a broadband dipole source, in which case the transmission values at certain wavelengths were >1 and sometimes > 2 as well. Why do you think that might arise? In that case, what would be the approach you suggest?

Thank you.

[Guilin Sun]

When injecting broadband dipole source, if the material has dispersion, then the analytical solution might be not accurate.  Please use the box monitors.

Another possibility is the interaction of Dipole with PML reflection. Please have sufficient distance betweent the dipole and the PML, as well as the sufficient thickness of the PML, in particular if you use override region, or higher mesh accuracy.

 

[smajety]

When I doubled the distance between the PML layers and the edge of the photonic structure, the analytical value changed by less than 0.1%, so I am assuming the PML distance is sufficient. I am using 16 layers for my PML and a mesh accuracy = 2.

My other question is: is the analytical result accurate if I have a distributed Bragg reflector structure with only dielectrics and no lossy, dispersive material? The dipole source is in one of the dielectric.

[Guilin Sun]

The website does not give details, but it should work for any structure. Although it specifically say loss is zero, I thought any change of refracgtive index may affect the result. You can try to use a small dielectric volume at the dipole location, eg, 3by3 meshes, constant refractive index to ovedride the original material. You can use an averged index from the original material.

The other possobility if, if there is resonance. If so, you may need longer simulation time and smaller autoshutoff.

 

Please send your simulation file to me if there is still an issue.

[sagar]

I also faced the same problem with Transmission > 1. As you suggested renormalization of transmission by dividing with Purcell factor. Renormalized_T=T/PF, where source power will be replaced by dipole power. The problem with the division is it will give the maximum values if T and PF both have nearly the same values. It does not matter whether the values are small or larges. For the diameter sweep of the dielectric, I found T=0.067 and PF=0.14 at 100 nm dia of the structure, and at 300 nm dia, I got T=0.98 and PF=2.31 leads to a maximum transmission, while doing renormalization I am getting Re_T= 0.47 for 200nm dia and 0.42 for 300 nm dia. Which seems to be an illogical situation. 

Maybe renormalization is a solution but should we use some kind of filter like renormalization is necessary only when PF>1. Please comment on it. Thank you.