Heater design in interconnect


I would like to attach a heater to a microring resonator to observe the change in spectrum using Interconnect. Is there a way from where I can import a heater element in Interconnect and observe the spectrum using an ONA?

Any help will be appreciated.

Hi @lekha.rane,

Thank you for the question. Do you know how the effective index of the ring resonator waveguide changes with temperature? If you do, you can use the effective index temperature sensitivity property to define how much the effective index changes with a change in temperature:

You can then vary the temperature setting of the Root element, and the effective index change will automatically be included in the ring resonator.

If you don’t know how the effective index changes with temperature I would recommend taking a look at the thermal switch example to see how this can be determined with a MODE simulation.

I hope this helps. Let me know if you have any questions.

Hello @kjohnson,

I would like to calculate the relationship between applied voltage to the heater and change in the power at through and drop ports of the microring in Lumerical Interconnect.

I have implemented the model you suggested by varying the effective index and have observed the power values at the through and drop ports. Now, I would like to calculate the voltage applied to the heater to obtain the same power values at through and drop ports.

Unfortunately, I am unable to find a heater design or a P-N junction component that can help me implement this.

Is there any model in Interconnect that I might be able to use?


Hello @kjohnson,

According to thermal variations, there is a shift in the resonant wavelength of the resonator with change in temperature. After substituting the effective index values and keeping the other parameters constant, I am unable to observe the shift and the spectrum is a linear graph rather than a resonating one.

The sensitivity I obtain after running the thermal switch example the sensitivity values seems to be 2 x 10^-4/K. But I am assuming it is for some specific ring dimensions which are for the thermal switch.

My final goal is to obtain the power at through and drop ports at different temperatures which can help me find a relation between the amount of voltage required for the heater to shift the resonant frequency back to its original value.

Sorry for all the trouble!


No, there is no built-in model in Interconnect for this. As you can see in the thermal switch example, there are several steps required for this simulation:

  1. Applied voltage to change in temperature (HEAT simulation)
  2. Change in temperature to change in effective index (MODE simulation)
  3. Change in effective index to change in transmission (INTERCONNECT simulation)

Putting all of these results together, you obtain the relationship between applied voltage and change in transmission.

Once you have the relationship between voltage and change in effective index from steps 1 and 2, you could use the method I outlined in my previous post to model the change in resonance. You could also use a ring modulator element. With this element, you can set up a table of applied voltages and the corresponding changes in index (which you would have from steps 1 and 2).

Could you attach your simulation file so I can take a look?

hello @kjohnson,

I have already used the thermal switch as the example and obtained the relationship of neff vs temperature. For eg: at 300K neff = 2.45285 and 350K = 2.46285. If i substitute these values of effective index in my ring, I obtain the attached image which is not the transmission spectrum I am looking for.

I was under the impression that by changing the effective index I will be able to see the shift in the resonance frequency. Although, by changing the effective index i still obtain the same graph. If I just run the example of ring resonator using interconnect, I can clearly see the transmission spectrum which i wish to observe using interconnect.
I think that I need to change other parameters of the ring such as group index and such to observe the total change.


This is the graph I wish to observe.

The graph I obtain when I use the ring resonator element

The transmission spectrum will depend on the ring parameters (radius, effective/group index, coupling coefficients, etc.). I would recommend that you make sure that the resonant wavelengths are within the spectrum you are measuring with the ONA. It would probably be a good idea to increase the bandwidth the ONA is measuring as well, right now it is only 1 nm which might not be large enough to see the resonance spectrum.

Hello @kjohnson,

Sorry, I understood my mistake immediately but somehow could not delete my previous post. Regarding the ring modulator, is it possible to observe the changes in the resonant wavelength by changing the DC bias of the ring?

I have seen the ring resonator (interconnect) example, but it seems very difficult.I would like to obtain the same results (peak resonance at different voltages) for a particular configuration of the ring modulator.

I was under the impression that by changing the voltage of the DC source attached at the modulator, I could observe the shift in resonance, but it seems that is not the case.

Yes, that should be how it works. You can see an example in step 5 of the ring modulator example. It is possible that the shift in the transmission spectrum is very small for the parameters you are using. Can you please share your simulation files so I can take a look?

Hello @kjohnson,

I have changed the radius, ng, neff and central frequency for the ring and ONA. I obtain the same spectrum no matter the change in the DC source. Attached is the file I am working on.Demo1.icp (115.5 KB)

There is a change in the transmission, it is just very small with these element properties:

If you want to see a larger change, you can set larger index perturbations in the “measurement” property of the ring modulator, and vary the DC amplitude from 0 to 1:

Hello @kjohnson,

Can you tell me the basis on which these index perturbations depend on? In theory I have studied that the more voltage applied to the ring, the more difference you can observe in the shift. While here it seems the voltage changes only from 0 to 1au.

While my ring resonates at 1550nm for 0V, I would like to calculate the voltage at which it will resonate at 1546.6nm and 1553.4. Is there a way to achieve this?

The default effective index perturbation values for the ring modulator element are just meant as an example. You have to determine the index perturbation as a function of applied voltage for your device (this is done in steps 3 and 4 in the ring modulator example) and update the ring modulator element with your values.

Once you have updated the ring modulator with your index perturbation values, you can vary the applied voltage and see the variation in the resonance frequency by plotting the transmission spectrum. The resonant frequencies are at the minima of the transmission spectrum.

Hello @kjohnson,

following the ring modulator example, step 4, point 5, I am unable to find the export to interconnect button in my solver window. Can you please let me know where it is?

Also, while doing the frequency sweep, if I write the value of effective index in my effective index tab, the value remains constant over the entire sweep.

In the “Frequency Analysis” tab, select “Data Export” from the “options” dropdown menu. The “Export for INTERCONNECT” button is at the bottom:


There could be a couple of reasons for this. Depending on the waveguide and the frequency range, it is possible that the effective index of the modes are actually not changing very much. Also, depending on which index you put for the effective index, it is possible that the modes being found are not the mode of interest, possibly unbound or unphysical modes.

I would recommend that you find the modes and select the mode you are interested in before you perform the frequency sweep. Make sure the “track selected mode” option is selected. The “effective index” setting will then be unavailable, because the solver will use the effective index of the selected mode as the starting point: