This application example will simulate a quarter-wave-shifted index-coupled distributed feedback (DFB) laser and compare results to the literature. To study the performance of the laser, different parameters affecting the side-mode suppression ratio (SMSR) such as gain profile and bias current will be studied and the light-current (L-I) curve will be derived. In the end, we demonstrate some of the new features released in the 2020a r5 version, such as variation of the facet phase and multiple phase slips.
Unlike the conventional Fabry-Perot lasers with discrete mirrors at the ends of the optical cavity, DFB lasers use gratings in the body of the laser as a wavelength-selective element.
INTERCONNECT’s TWLM element can have gratings incorporated to design a complex DFB laser. A variety of grating structures can be designed using different settings such as chirp, multiple phase slips and apodization.
This application example will simulate a quarter-wave-shifted index-coupled DFB laser with the parameters given in reference . Characterization and performance study steps will be described. We will plot DFB laser’s multi-mode spectrum vs current and spectral gain width, find SMSR as a function of current and also plot the L-I curve obtained from the simulation for the laser and compare it with that obtained in reference . In addition, in the “Taking the results further” section, we illustrate the effects of multiple phase slips and facet phase variation.
Run and results
Instructions for running the model and discussion of key results
Study change in SMSR by changing bias current and gain shape
- Open [[spectrum.lsf]] file and run the script.
The spectrum of the DFB laser for a flat gain (FWHM = 1000nm) and a bias current of 28mA is measured by the ‘OSA_1’ and plotted below. The simulation result corresponds well to the result reported in the reference .
- In the script, increase the bias current to 63mA and run the script again.
In good agreement with the reference, increasing the bias current to 63mA, the side mode suppression ratio (SMSR), which is the power difference between the main mode and the side mode, is decreased. This is due to the spatial hole burning at a high bias current, which reduces the difference in threshold gain between the main mode and side modes.
- For a more realistic gain shape, decrease the FWHM to 10nm in the script and run again.
The side mode is weakened and SMSR is improved. SMSR in this case is measured to be ~43dB that agrees reasonably with the published value of 38dB .
SMSR and optical power in terms of input current
The calculation of SMSR as a function of input current can be automatized by using Lumerical's script language in addition to the sweep object that was used to sweep over current and save power and spectrum results. The steps to generate SMSR as a function of input current are the following:
- Open and run [[SMSR_vs_current.lsf]]. This script runs the sweep object “CurrentSweep” to obtain spectra for different input currents. It then calculates SMSR in dB by subtracting magnitudes of the two most dominant peaks.
The result for the sweep with 10nm gain FWHM is given in the figure below. As can be seen the SMSR increases as the bias current increases
- Open and run the script file [[plotLI.lsf]].
The LI results from the “CurrentSweep” object as well as the reference  (LIcurveDataRefFig2a.txt) are plotted below for comparison. Both show reasonably similar threshold currents and slope efficiencies, validating the accuracy of our TWLM models for the design of DFB lasers.
Important model settings
Description of important objects and settings used in this model
For transient sample mode simulations, the sample rate must be sufficiently high for accurate results. Since the sample rate is equal to bandwidth, it should be large enough to contain the relevant portions of the gain and cavity spectral shapes. Also, the time step (inverse of the sampling rate) and cavity section length are related via the group velocity, hence the sample rate should be large enough to have a sufficient spatial resolution to accommodate the required number of sections.
OSA resolution can be enabled to reduce the noise of the measured signal. Here, to make the measured signal smooth in order to be able to calculate SMSR easier, the OSA resolution is set to Gaussian function and a frequency of 1GHz is selected.
A simple Lorentzian gain shape is used to describe the gain profile in this example. Changing the gain shape quality factor by changing the FWHM is used to study for different gain profiles.
A simple grating is used in this example with a phase slip of 90 degrees introduced in the middle. No chirp function is used and apodization is set to uniform.
Updating the model with your parameters
Instructions for updating the model based on your device parameters
Design a complex grating
The gratings incorporated in the TWLM element can be set up for a complex grating structure. Chirp, apodization, and phase slips can be enabled and set in the grating section of the TWLM element to design for a sophisticated DFB grating.
Changing gain properties
To update the gain region properties, active region geometry can be modified in the “standard” section and other gain properties can be updated in the “waveguide/gain” section in the TWLM.
Changing waveguide properties
To update the waveguide properties, the “waveguide/ mode 1” section in the TWLM properties can be modified.
Taking the model further
Information and tips for users that want to further customize the model.
Realistic gain spectrum
As seen in this example, having a realistic gain profile is important for the accurate characterization of the designed laser. To accurately model the gain medium, use the MQW gain solver. Check MQW gain solver and MQW edge-emitting laser examples.
To study the performance of the laser having phase slips, add multiple phase slips at different lengths of the cavity, and study laser performance.
- In the TWLM element add three-phase slips of pi/2 (1.57 rad) in normalized locations of 0.25, 0.5 and 0.75 in the “grating phase slips table” and set “grating phase slip” to zero (this latter option sets the phase slip at 0.5, which can also be done through the phase slip table).
- In the “Diagnostic” section in TWLM, set “record photon density profile” to true.
- Run the simulation.
- Run the script [[photon_density.lsf]].
The script takes an average of the profile over time for each position and plots the result. As you can see below, having the three-phase slips results in a photon density profile to have three peaks. Spreading the photon distribution in space with this technique, reduces the spatial hole burning effect.
HR/AR coating on facets
For a DFB laser with HR/AR coating, the change in the facet phase of the reflective facet due to variability in distance between the facet and the gain layer, e.g. from cleaving the structure, can affect the spectrum. To study this effect on the peak wavelength, follow these steps for an example case:
- In the TWLM element, change the left facet reflectivity from 0 to 0.9.
- Make sure there are no phase slips introduced to the TWLM element.
- Run the sweep “FacetPhaseSweep”. This will sweep the values for the left facet phase and save the measured spectrum from ‘OSA_1’
- Run the script file [[peakwl_vs_facetphase.lsf]].
The script plots the peak wavelength vs facet phase (left facet) as shown below:
Here we can see that by changing the facet phase, there will be mode competition between two modes in the Bragg gratings main band with a period of 180 degrees. The figure below shows the change in the spectrum by changing the facet phase for the first period. At the top of the figure below, the spectrum shows the two main modes from the Bragg grating main band. We can see how changing the facet phase, moves the peak wavelength between these two modes.
Additional documentation, examples and training material
- Arthur J. Lowery, Adrian Keating, and Casper N. Murtonen, “Modeling the Static and Dynamic Behavior of Quarter-Wave-Shifted DFB Lasers”, IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 28, NO. 9. SEPTEMBER 1992