Model reflectivity modulated external cavity lasers using TWLM



I am trying to understand and use the TWLM in INTERCONNECT as the SOA core of the directly modulated external cavity lasers. I have the following two questions:

  1. Could I know what reference papers TWLM use in the modelling so that I could read and get familiar with?
  2. I am trying to directly modulate the output mirror reflectivity (not the normal injection current), could this feature be realized using the current TWLM?

Many Thanks


Hi @gyaliu,

Thank you for the questions.

  1. Here is a reference paper for the TW Laser model:
    A.J Lowery, “New dynamic semiconductor laser model based on the transmission-line modelling method”, IEE PROCEEDINGS, Vol. 134, Pt. J, No. 5, OCTOBER 1987

  2. Could you let me know more details about what you are trying to do here? How do you want to modulate the mirror reflectivity and how do you expect this to affect the lasing pattern?

Hope this could help.


Dear Guanhui,

Thank you for the reply.

I am trying to build up the model according to the paper “Directly Reflectivity Modulated Laser” 01.pdf (1.5 MB) with correct representation of both static tunable lasing spectrum, dynamic eye diagram and small signal EO response. So I proceed the simulation in time domain using Transient Sample Mode (TSM) solver since I want to get the dynamic response of the circuit.

I have had the following circuit model (RSOA + phase shifter+Vernier ring filter (self built compact model)+ Michelson Interferometer Mirror (MIM, self built compact model))with reasonable static tunable laser spectrum and dynamic eye diagram.

Static lasing spectrum:

Dynamic eye diagram: (25Gbps OOK)

But I have a problem when trying to use the ENA to get the EO response of the laser modulation (modulation signal is applied to either arm of the MIM) matching the experimental result. My main target is to get the EO response down to KHz ~ MHz range to confirm the ~100KHz laser relaxation frequency :

Experimental result:

Another theoretical calculation of the laser relaxation frequency:

Current simulation result:

I think the problem is partially coming from the ENA simulation bandwidth linked to the circuit sampling rate (which is quite high 10s THz for current laser simulation), so that I could not get enough resolution at the small frequency range. And the total simulation time increases dramatically when I increase the “number of points” parameter in the ENA. And it seems to me the minimum resolution of ENA is ~0.1GHz.

I have also thought about sending sinusoidal wave (with different frequency) into the circuit and capture the optical oscillation using OVA/OOSC, but the simulation time for a 100KHz frequency wave will be also long.

Do you have any recommendations for getting the EO response down to KHz frequency resolution while maintaining the limited simulation time/memory?




Hi @gyaliu,

Nice design!! Could you please attach your file here and I will take a look. Thanks.