The relation between Q value and spectrum in the Qanalysis


#1

I want to know whether the microcavity spectrum is accurate when the Q value fluctuates too much.
for example,in the photonic crystals cavity with HIgh-Q,its spectrum is below,

,
and the output is:

Resonance 1:
frequency = 202.126THz, or 1483.2 nm
Q = -38347.8 +/- 3.64374e+006
Resonance 2:
frequency = 219.134THz, or 1368.08 nm
Q = 82523.3 +/- 4.54105e+006
Resonance 3:
frequency = 185.127THz, or 1619.39 nm
Q = 24377.2 +/- 527449

as you see,the Q value fluctuates too much,but the spectrum is the result of what I want. whether there is a such possibility that the spectrum is correct even though the Q value is wrong? I have doubts about this q algorithm in the more extreme structrue and setting ,at the same time, my structure and setting may not be incorrect.
thanks anyhelp


L3 Cavity Mode Simulation
#2

Hi @phd.wangbo

The High-Q Analysis Group uses a Fourier transform of time domain data to calculate the resonance modes and as long as you have enough signal you should expect to get precise result about the frequency of the optical modes. Calculating the quality factor, however, is a bit challenging as the analysis group tries to fit the slope of time signal data. A logical procedure is to:

  1. run a simulation with a few broadband source and monitors to locate the resonant frequencies
  2. perform a second simulation with narrow frequency source to find the exact frequency and also the mode profile to locate the anti-nodes at resonant frequencies
  3. optimising step where you need to locate your source and and Q analysis group at the anti-node of the resonant modes, and the simulation time to make sure that you are recording enough data to fit

I just run some sample examples form our available High-Q analysis examples in KB and I learned that simulation time plays an important role (even if you have properly placed the source and Q analysis groups) on getting a proper value for Q and also the error bars. If you find tips that helped you to get more robust results, feel free to share with us. Also, I am willing to take a look at your file if the problem didn’t solve.