Learning Forum

[evadi]

Hi, I am having trouble understanding the output power of a time monitor. Let me briefly explain the simulation:

I am simulating a H0 PhC cavity coupled to a W1 waveguide. The cavity has a loaded Q-factor of 14,000 and an intrinsic Q-factor (no waveguide) of 500,000. Inserting a waveguide that close to the cavity will break the cavity symmetry and I expect the intrinsic Q-factor to massively drop and this is what I want to calculate. The calculation want to do require me to calculate the total power flux (over time) in all directions.

So, I am exciting the system with a narrowband dipole placed in the cavity to excite only the mode of interest, and I have included a 2D time monitor in the YZ plane cutting through the output waveguide, and a 2D time monitor in the XY plane next to the top PML layer. Finally, I am recording the output power of the time monitors denoted as Px and Pz, which is plotted in semilogy scale in the bottom left plots.

Looking at these plots, we can see that the time trace of the output power is single-moded. What I expect is that the total power in the WG recorded in Px will be higher, since most of the light will be coupled to the WG. A smaller portion of light will leak vertically to the Pz monitor.

The problem:

Going back to the 'Power vs Time' plots: each trace should be in the form of ' I_0 * cos^2(2π ω_0 t + φ) * exp(-Γt) ', however what we see is that the power do not reach zero (or close to that) for neither Px or Pz. What is even stranger is that the minima and the amplitude of the Px and Pz monitor greatly vary. How is this explained? Is there any way to fix that? As you can see there shouldn't be a problem of sampling. What I want is to estimate the total power Px and Pz?

I would like to thank you in advance!

Evangelos

[Guilin Sun]

It seems this cavity has regular waveguides as input and output, I would suggest that you use the regular mode source instead of the dipole source to excite the source.
In addition, it seems you set the source in time domain but the "off set" is too small. We will need a time signal that can gradually increase first, reach the peak and then decrease.
Since now the excitation is not monochromatic, so this assumption 'I_0 * cos^2(2¤Ç ¤ë_0 t + ¤å) * exp(-╬ôt)' might not be valid. It is the pulse, although its spectrum is narrow, that contains frequencies other than w_0. This is why it does not reach zero.
If you want to see its change with about a single wavelength, please create a quasi-monochromatic time signal, like in this example: 1D cavity laser using 4-level 2-electron material the cw_source.lsf can be modified for your chosen w_0.
A side note: in most cases we do not need to see the time-domain power evolution as we usually need frequency-domain, cw result.

[evadi]

Hi gsun I am measuring Q-factors, so the time evolution of the power is important to me. So I cannot use a quasi-monochromatic/CW source. I want to excite the system and observe the decay. Either way, I just want a narrowband source (narrower than the mode spacing of the H0) to excite the one mode I am interested in - the rest of the frequencies should decay much faster than the mode of interest. Looking at the time traces of the power, they look fairly monochromatic - not too much interference. Still I have some questions:
I don't understand how a regular mode source will change anything in my system. Can you elaborate on that?
I am also not sure why changing the pulse offset will change the amplitude variation of Px,z(t). I tried it, just to be sure and I didn't observe any differences.
My main problem still remains: Why the monitor Pz has this lower max-low variation than Px. At the end I want to compare Pz/Px and I am not sure if I should consider the power average over a circle, or just the top power for each circle. If I take the average, Pz is larger than Px which do not make sense physically. Any thought on that?
Thanks!
Evangelos

[evadi]

[Guilin Sun]

A1: the mode source excitation will excite the true cavity mode that you can measure later.
A2: Without proper offset, the source will create higher frequency that may cause simulation diverging, and interfere with your pulse. Please calculate the spectrum, or better you do fft and compare the results. This is just my suggestion and you do not need to follow.
A3: When you expect that the P should reach zero, I guess you are based on the regular cavity. Now it is Bloch modes not the plane wave or regular waveguide modes. I believe the result is correct but you may need to explore more about its physics. In this article I gave the bloch modes https://support.lumerical.com/hc/en-us/articles/360041567854-Equi-frequency-contours-of-photonic-crystal and I am sure you are familiar with Bloch modes. However understanding its unique properties takes time. Your waveguide seems along x axis, I am not sure why you expect Px and Pz behave the same. This is a more theoretical issue other than simulations, which is out of scope of this forum. Please review this forum guideline Guidelines for Posting on Ansys Learning Forum ÔÇö Ansys Learning Forum