"PICWave provides probably the fastest way of simulating PICs available today" - Dr. Dominic Gallagher, Photon Design C.E.O.

PICWave is a photonic integrated circuit (pic) simulator capable of modeling both passive and active components efficiently using the TWTD time domain algorithm. The simulator is ideal for studying the interaction of optical components in a larger circuit and can model devices that are even meters in length. It can for example model a 2mm diameter ring resonator in a very short time to resolutions of a few MHz in optical frequency. In addition, it includes a detailed multi-model SOA model.

  • Efficient optical circuit simulator
  • Travelling Wave Time Domain (TWTD) optical model
  • Supports multiple transverse modes, e.g. for polarization diversity studies
  • Optical spectra - time averaged and time-evolving
  • Power spectra - time averaged and time-evolving
  • Wide variety of instruments - measuring optical power, wavelength, current, carrier density, threshold etc
  • Integrated high performance fully vectorial 2D+z Finite Difference mode-solver
  • Arbitrary time domain input signals controlling optical power, wavelength, drive voltages, drive currents
  • Integrated grating solver - real and gain gratings
  • Eye diagrams
  • S-matrix import from FIMMPROP or other component simulator


PC: x86+x64: Win2000/XP/Vista, 1GB RAM, 2GHz or better recommended.

Active Module

Physical Models

  • Lorentzian optical phase and intensity noise model
  • Electrical noise model
  • Travelling wave electrode model
  • Longitudinal hole burning
  • Lateral hole burning
  • Carrier diffusion
  • Non-linear gain
  • Auger processes
  • Thermal effects
  • Advanced Multi-Lorentzian Gain model import from HAROLD


  • PI and PV curves
  • MQWs
  • Quantum efficiency
  • Chirp simulation
  • RIN spectra
  • Material database system
  • Import gain tables
  • Electro-absorb modulator model


  • Photonic integrated circuits (PICs)
  • Tunable laser diodes
  • Large ring resonators
  • Mach-Zehnder modulators
  • Travelling wave SOAs
  • Electro-absorption modulators

Example: An active 2R Optical regenerator

One of PICWave's strengths is its ability to combine passive and active components. This optical regenerator consists of a large passive optical circuit totaling over 10,000um of waveguides plus two SOAs.

The regenerator is shown schematically above, consisting of a Mach-Zehnder interferometer with an SOA in each arm. Initially a steady state signal is injected into the device at B and passes through both arms of the MZI then recombines on the right. When a data signal is present at A, this upsets one of the SOAs causing destructive interference when the B signal recombines, turning off the output. By this means we can regenerate the signal on/off levels, amplify the signal and convert the signal to another wavelength.

The regenerator was simulated with PICWave using an NRZ pseudo-random input bit pattern with an on/off ratio of 5:1 and a rise time of 100ns. To the right is shown eye diagram of the output signal. The output signal has a much higher on/off ratio compared to the input, and also 20x amplification. However in this case the SOAs have added significant noise to the output.

PICWave includes an extensive model for noise sources present in semiconductor devices, modeling carrier fluctuations, phase noise and intensity noise.

Example: Modeling a large ring resonator

PICWave can easily and efficiently model ring resonators of 100's um diameter. The algorithm is many orders of magnitude more efficient than e.g. FDTD for this application, and for example can compute a triple 200um diameter ring resonator down to a wavelength resolution of 50MHz and spectral range of 50nm in just a couple of minutes.

As a benchmark, a single 60um diameter ring resonator for which the theoretical result is known was simulated and the result shown in the plot above. PICWave yielded a result in just a few seconds, while FDTD needed 14 hours. Note also that the accuracy obtained by PICWave is better than that obtained by FDTD. As you can see from the graph, PICWave agreement with the theoretical solution is remarkable, especially considering that it only takes a few seconds.

The difference in calculation time is even bigger for larger ring resonators. The reason is that PICWave is able to use a large time step, and so can handle simulation durations in the range of several ns, yielding spectral resolutions finer than 1GHz. In fact, MHz resolutions are attainable. This degree of resolution is important when modeling large ring resonators.

Example: Mode hopping in a Fabry-Perot laser

The figures below show mode hopping in a Fabry-Perot laser simulated with PICWave.

Current is increased steadily from 0mA to 100mA in 20ns. The top plot shows the optical output power. The middle plot is the mode gain which gives an indication of the carrier density. The lower plot shows the time-evolving spectrum.

The shifting mode frequencies below threshold can be seen until lasing locks the carrier density. Mode hops are accompanied by a temporary chaotic phase. Note also that because of spatial hole burning the gain (middle plot) decreases gradually with increasing optical power.