• Photonics Research
  • Vol. 10, Issue 4, 939 (2022)
Jinlong Xiang, Yujia Zhang, Yaotian Zhao, Xuhan Guo*, and Yikai Su
Author Affiliations
  • State Key Laboratory of Advanced Optical Communication Systems and Networks, Department of Electronic Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
  • show less
    DOI: 10.1364/PRJ.445954 Cite this Article Set citation alerts
    Jinlong Xiang, Yujia Zhang, Yaotian Zhao, Xuhan Guo, Yikai Su. All-optical silicon microring spiking neuron[J]. Photonics Research, 2022, 10(4): 939 Copy Citation Text show less
    (a) Microscope image of the silicon microring used in the experiment. (b) Schematic figure of the measurement setup for both characterization of the microring and investigation of its spiking dynamics. TLS, tunable laser source; PC, polarization controller; DUT, device under test; OSA, optical spectrum analyzer; CR, circulator; EDFA, erbium-doped fiber amplifier; BPF, bandpass filter; PD, photodetector; AWG, arbitrary waveform generator; OSC, oscilloscope; EOM, electro-optical modulator. (c) Non-normalized transmission of the microring. (d) Transmission of one single resonance and its Lorentz curve.
    Fig. 1. (a) Microscope image of the silicon microring used in the experiment. (b) Schematic figure of the measurement setup for both characterization of the microring and investigation of its spiking dynamics. TLS, tunable laser source; PC, polarization controller; DUT, device under test; OSA, optical spectrum analyzer; CR, circulator; EDFA, erbium-doped fiber amplifier; BPF, bandpass filter; PD, photodetector; AWG, arbitrary waveform generator; OSC, oscilloscope; EOM, electro-optical modulator. (c) Non-normalized transmission of the microring. (d) Transmission of one single resonance and its Lorentz curve.
    (a) Measured real-time output waveform of self-pulsation for different input powers. (b) Frequency of self-pulsation in relation to pump power.
    Fig. 2. (a) Measured real-time output waveform of self-pulsation for different input powers. (b) Frequency of self-pulsation in relation to pump power.
    (a) Measured real-time output waveform of excitability in response to perturbation pulses of increasing power. (b) Normalized strength of the “negative” spike as a function of perturbation power. The threshold area is indicated by the shaded region.
    Fig. 3. (a) Measured real-time output waveform of excitability in response to perturbation pulses of increasing power. (b) Normalized strength of the “negative” spike as a function of perturbation power. The threshold area is indicated by the shaded region.
    (a) Simulated and (b) measured microring neuron spikes for different wavelength detuning values of the pump and perturbation light.
    Fig. 4. (a) Simulated and (b) measured microring neuron spikes for different wavelength detuning values of the pump and perturbation light.
    (a) Distribution of the excitability threshold for the microring neuron over different pump powers and wavelength detuning. (b) For Pin=1 mW and δr=−20 pm in region I, the microring neuron exhibits a typical refractory period with Ptr=0.6 mW, which is successfully suppressed by increasing the perturbation power to Ptr=2 mW. (c) For Pin=0.5 mW and δr=−40 pm in region II, the microring neuron remains silent when the perturbation power is the same as the pump power of Ptr=0.5 mW, but is excited by two perturbation pulses of Ptr=3 mW with a time duration of 80 ns.
    Fig. 5. (a) Distribution of the excitability threshold for the microring neuron over different pump powers and wavelength detuning. (b) For Pin=1  mW and δr=20  pm in region I, the microring neuron exhibits a typical refractory period with Ptr=0.6  mW, which is successfully suppressed by increasing the perturbation power to Ptr=2  mW. (c) For Pin=0.5  mW and δr=40  pm in region II, the microring neuron remains silent when the perturbation power is the same as the pump power of Ptr=0.5  mW, but is excited by two perturbation pulses of Ptr=3  mW with a time duration of 80 ns.
    Measured real-time output waveform of refractory period with varying time intervals between two perturbation pulses. Each pulse is strong enough to excite the silicon microring neuron to emit a complete spike.
    Fig. 6. Measured real-time output waveform of refractory period with varying time intervals between two perturbation pulses. Each pulse is strong enough to excite the silicon microring neuron to emit a complete spike.
    Measured real-time output waveform of temporal integration with varying time intervals between two perturbation pulses. Only a subthreshold spike can be triggered by each single pulse.
    Fig. 7. Measured real-time output waveform of temporal integration with varying time intervals between two perturbation pulses. Only a subthreshold spike can be triggered by each single pulse.
    Measured real-time output waveform of inhibitory dynamics with varying time intervals between the excitatory and inhibitory pattens.
    Fig. 8. Measured real-time output waveform of inhibitory dynamics with varying time intervals between the excitatory and inhibitory pattens.
    PropertiesIntegratorsResonators
    BifurcationSaddle-node on invariant circleSaddle-nodeSubcritical Andronov–HopfSupercritical Andronov–Hopf
    ExcitabilityClass IClass IIClass II
    BistabilityNoYesYesNo
    ThresholdWell definedMay not be defined
    Subthreshold oscillationsNoYes
    All-or-none spikingYesNo
    Frequency preferenceNoYes
    Table 1. Summary of Major Neurocomputational Properties [4]
    Jinlong Xiang, Yujia Zhang, Yaotian Zhao, Xuhan Guo, Yikai Su. All-optical silicon microring spiking neuron[J]. Photonics Research, 2022, 10(4): 939
    Download Citation