A traveling wave mach-zehnder modulator distributed modeling method based on verilog-a

Abstract

This invention discloses a distributed modeling method for traveling-wave Mach-Zehnder modulators based on Verilog-A. Specifically, in the SPICE environment, a complex envelope is used to represent the optical signal, and the real part, imaginary part, and wavelength of the optical signal are transmitted as node variables. The modulator is divided into multiple micro-segments, and a cascaded RF distributed parameter network and a third-order polynomial active phase shifter model are used. Simultaneously, a dual-time-delay coupling mechanism between optical waves and microwaves is introduced. The constructed model is used for the joint simulation of steady-state, frequency domain, and transient responses of the Mach-Zehnder modulator. This invention is beneficial for improving the continuity and convergence of the model under large-signal conditions. Through micro-segmentation, RF distributed parameter networks, and the dual-time-delay coupling mechanism between optical waves and microwaves, this invention can describe the steady-state, frequency domain, and transient responses of the traveling-wave Mach-Zehnder modulator in a unified solution environment.

Description

Technical Field

[0001] This invention relates to silicon photonic modulator modeling and simulation technology, and more particularly to a distributed modeling method for traveling wave Mach-Zehnder modulators based on Verilog-A. Background Technology

[0002] With the continuous improvement of optical interconnect speeds in data centers, modern high-speed optoelectronic transceiver systems have gradually evolved into complex hybrid systems composed of photonic devices and large-scale transistor circuits. For Mach-Zehnder modulators, their operation involves multiple factors such as optical field interference, carrier depletion effect, and traveling wave electrode distribution parameters. Device design is no longer limited to single physical field analysis, but needs to take into account device-level characteristics, circuit-level excitation, and system-level dynamic response.

[0003] While existing field simulation methods can accurately obtain waveguide mode distributions, carrier depletion characteristics, and high-frequency electrical parameters, their results are mostly in the form of discrete field data and numerical curves, making them difficult to directly embed into integrated circuit design environments. They also struggle to meet the requirements of system-level simulation for model reusability, computational efficiency, and convergence. Especially when performing joint analysis of long code patterns, large-signal transients, or link-level processes, traditional methods typically rely on fine meshes and small time steps, resulting in significant computational overhead. Furthermore, existing optical system simulation platforms usually employ relatively independent solver engines, making it difficult to directly call standard CMOS SPICE model libraries. This means that joint verification between photonic devices and electronic circuits still requires additional data conversion and interface adaptation.

[0004] Furthermore, in SPICE-type solvers, if the amplitude and phase of the optical signal are directly transmitted in polar coordinates, the optical phase may frequently cross the principal value interval when the Mach-Zehnder modulator is driven by a large-amplitude microwave signal. This can cause abrupt changes in the phase signal, affecting the Jacobian matrix solution in the Newton iteration process and thus reducing simulation convergence. Therefore, it is necessary to propose a modeling method for traveling-wave Mach-Zehnder modulators that can be invoked in a unified circuit environment and takes into account both nonlinearity and distributed effects. Summary of the Invention

[0005] To address the challenges of directly embedding existing field simulation methods into SPICE circuit environments, the difficulty in unifying and jointly utilizing photonic devices and electronic circuits, and the insufficient model convergence under large signal conditions, this invention provides a distributed modeling method for traveling wave Mach-Zehnder modulators based on Verilog-A.

[0006] The present invention provides a distributed modeling method for traveling wave Mach-Zehnder modulators based on Verilog-A, comprising the following steps:

[0007] Step 1: Obtain the device parameters of the traveling wave Mach-Zehnder modulator. The device parameters include at least the PN junction parasitic parameters, the effective refractive index variation parameters, and the traveling wave electrode distribution parameters.

[0008] Step 2: In the Verilog-A language environment, establish optical network nodes in the SPICE (Simulation Program with Integrated Circuit Emphasis) circuit simulation environment, use complex envelopes to represent optical signals, and map the real and imaginary parts of the complex envelopes to parallel node variables for transmission.

[0009] Step 3: In the Verilog-A language environment, establish a basic optical device behavior model based on optical network nodes. The basic optical device behavior model includes at least waveguides, beam splitters, beam combiners, and active electro-optic phase shifters.

[0010] Step 4: In the Verilog-A language environment, establish the electro-optic response model of the active electro-optic phase shifter, and fit the relationship between the bias voltage and the additional phase shift, as well as the relationship between the bias voltage and the additional loss.

[0011] Step 5: Divide the active electro-optic phase shifter into multiple micro-segments along the light propagation direction, and cascade the RF distributed parameter network and the active electro-optic phase shifter model in each micro-segment to construct a distributed traveling wave Mach-Zehnder modulator model.

[0012] Step 6: Introduce optical propagation delay and radio frequency propagation delay into the electro-optic response model of the active electro-optic phase shifter to characterize the time delay coupling between light waves and microwaves.

[0013] Step 7: In the unified SPICE circuit solver, call the distributed traveling wave Mach-Zehnder modulator model to perform steady-state simulation, frequency domain simulation and / or transient simulation of the traveling wave Mach-Zehnder modulator.

[0014] Furthermore, optical network nodes establish optical port rules through custom nature and discipline, enabling optical nodes and electrical nodes to participate in unified solutions within the same circuit simulation environment. Optical signals are transmitted through an array of optical ports of length 3, where the first and second components are used to transmit the real and imaginary parts of the complex envelope, respectively, and the third component is used to transmit the optical carrier wavelength parameter.

[0015] Furthermore, the behavioral model of the basic optical device also includes a continuous wave source model and a photodetector model. The continuous wave source model is used to output a complex optical field envelope with a set optical power and wavelength, while the photodetector model is used to introduce a parasitic electrical network based on square-law detection.

[0016] Furthermore, in the electro-optic response model of the active electro-optic phase shifter, both the additional phase shift and the additional loss are established using polynomial fitting, where the polynomial is a third-order polynomial. The electro-optic response model of the active electro-optic phase shifter does not directly use the instantaneous input voltage as the modulation quantity, but first extracts the effective RF voltage after the RF propagation delay to complete the optical field modulation calculation, and then applies the optical propagation delay to the modulated output optical field.

[0017] Furthermore, the RF distributed parameter network includes series resistance, series inductance, parallel parasitic capacitance, substrate leakage resistance, and junction capacitance and junction resistance corresponding to the micro-segment, obtained by length scaling.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] 1. This invention can transform the discrete results obtained from parameter extraction into behavioral models that can be directly called in the SPICE circuit environment.

[0020] 2. The present invention adopts a complex envelope and Cartesian coordinate representation method, which is beneficial to improving the continuity and convergence of the model under large signal conditions.

[0021] 3. This invention describes the steady-state, frequency domain, and transient responses of traveling wave Mach-Zehnder modulators in a unified solution environment through micro-element segmentation, radio frequency distributed parameter networks, and optical wave and microwave dual-time-delay coupling mechanisms. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the overall distributed modeling of the traveling wave Mach-Zehnder modulator based on Verilog-A in this invention.

[0023] Figure 2 This is a diagram of the Lumerical INTERCONNECT link structure used for frequency domain verification in this invention.

[0024] Figure 3 This is a comparison chart of the frequency domain simulation results of the Verilog-A model of this invention and the Lumerical INTERCONNECT platform. Detailed Implementation

[0025] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0026] The principle of the distributed modeling method for traveling wave Mach-Zehnder modulators based on Verilog-A of the present invention is as follows: Figure 1 As shown, the specific steps include:

[0027] Step 1: Obtain the device parameters of the traveling wave Mach-Zehnder modulator. The device parameters include at least the PN junction parasitic parameters, the effective refractive index variation parameters, and the traveling wave electrode distribution parameters.

[0028] Step 2: In the Verilog-A language environment, establish optical network nodes for the SPICE circuit simulation environment, use complex envelopes to represent optical signals, and map the real and imaginary parts of the complex envelopes to parallel node variables for transmission.

[0029] Optical network nodes establish optical port rules through custom nature and discipline settings, enabling optical nodes and electrical nodes to participate in unified solutions within the same circuit simulation environment. Optical signals are transmitted through a 3-element optical port array, where the first and second components are used to transmit the real and imaginary parts of the complex envelope, respectively, and the third component is used to transmit the optical carrier wavelength parameter.

[0030] Step 3: In the Verilog-A language environment, establish a basic optical device behavior model based on optical network nodes. The basic optical device behavior model includes at least waveguides, beam splitters, beam combiners, and active electro-optic phase shifters.

[0031] The basic optical device behavior model also includes a continuous wave source model and a photodetector model. The continuous wave source model is used to output a complex optical field envelope with a set optical power and wavelength, while the photodetector model is used to introduce a parasitic electrical network based on square-law detection.

[0032] Step 4: In the Verilog-A language environment, establish the electro-optic response model of the active electro-optic phase shifter, and fit the relationship between the bias voltage and the additional phase shift, as well as the relationship between the bias voltage and the additional loss.

[0033] In the electro-optic response model of the active electro-optic phase shifter, both the additional phase shift and the additional loss are established using polynomial fitting, where the polynomial is a third-order polynomial. The active electro-optic phase shifter electro-optic response model does not directly use the instantaneous input voltage as the modulation quantity. Instead, it first extracts the effective radio frequency voltage after the radio frequency propagation delay to complete the optical field modulation calculation, and then applies the optical propagation delay to the modulated output optical field.

[0034] Step 5: Divide the active electro-optic phase shifter into multiple micro-segments along the light propagation direction, and cascade the RF distributed parameter network and the active electro-optic phase shifter model in each micro-segment to construct a distributed traveling wave Mach-Zehnder modulator model.

[0035] The RF distributed parameter network includes series resistance, series inductance, parallel parasitic capacitance, substrate leakage resistance, and junction capacitance and junction resistance corresponding to the micro-segment, obtained by length scaling.

[0036] Step 6: Introduce optical propagation delay and radio frequency propagation delay into the electro-optic response model of the active electro-optic phase shifter to characterize the time delay coupling between light waves and microwaves.

[0037] Step 7: In the unified SPICE circuit solver, call the distributed traveling wave Mach-Zehnder modulator model to perform steady-state simulation, frequency domain simulation and / or transient simulation of the traveling wave Mach-Zehnder modulator.

[0038] Example 1:

[0039] Modeling of optical network nodes and basic devices.

[0040] First, optical network nodes are established in the Verilog-A language environment. Considering the high absolute carrier frequency of optical signals in the communication band, directly performing transient integration on high-frequency optical waves in the time-domain solver would significantly increase simulation time and reduce solution convergence. Therefore, this embodiment uses complex envelopes to represent optical signals, and only transmits complex envelope information between network nodes. Since Verilog-A itself does not directly support complex data types, the complex envelope is split into two parallel real variables, real and imaginary, for transmission.

[0041] Furthermore, optical port rules are established by customizing nature and discipline, and each optical port is bound to an array of length 3, where indices [0] and [1] are used to transmit the real and imaginary parts of the complex envelope, respectively, and index [2] is used to transmit the optical carrier wavelength parameters across modules. Based on this unified port form, a continuous wave source model, waveguide model, beam splitter model, beam combiner model, and photodetector model are established as the basic units for the subsequent construction of the distributed model of the traveling wave Mach-Zehnder modulator.

[0042] Example 2:

[0043] Construction of a distributed model for a traveling wave Mach-Zehnder modulator.

[0044] An active electro-optic phase shifter model is established based on the aforementioned basic device model. This model uses the bias voltage as the input variable and fits the additional phase shift and additional losses separately, preferably using a third-order polynomial for characterization. To describe the impact of the mismatch between the optical group velocity and the radio frequency phase velocity on the device response under high-speed conditions, optical propagation delay and radio frequency propagation delay are introduced into the model. The effective radio frequency voltage after a preset radio frequency propagation delay is first extracted, and then the optical propagation delay is applied to the modulated output optical field.

[0045] To further characterize the distributed propagation characteristics of the traveling wave electrode, the active electro-optic phase shifter is divided into multiple micro-segments along the light propagation direction. Within each micro-segment, a radio frequency distributed parameter network and an optical phase shifter micro-element model are alternately cascaded. The radio frequency distributed parameter network includes series resistance, series inductance, parallel parasitic capacitance, substrate leakage resistance, and junction capacitance and junction resistance loads, thereby approximately describing the distributed propagation process of light waves and microwaves in a pure circuit solver.

[0046] Example 3:

[0047] Frequency domain verification.

[0048] like Figure 2 As shown, to verify the characterization ability of the established Verilog-A model for multi-port complex optical field interference, wavelength dispersion dependence, and spatial phase accumulation mechanism, a system-level link based on an asymmetric Mach-Zehnder interferometer was constructed, and its results were compared and verified with the commercial Lumerical INTERCONNECT platform. The frequency domain verification described herein is an embodiment verification method of the model of this invention, and not a limitation on the scope of protection of this invention.

[0049] At the input end, a continuous light source or optical network analysis module is used to inject an optical signal within the wavelength range to be scanned. The input light is split into two paths by a beam splitter and then enters the two arms of a Mach-Zehnder modulator. Different optical path lengths and bias conditions are set in the two arms to form an asymmetric interference structure. At the output end, a beam combiner is used to complete coherent superposition, and the transmission rate curve and logarithmic gain curve of the system output end are extracted.

[0050] In the SPICE circuit simulation environment, the input light wavelength is defined as a global scanning parameter. After performing a frequency domain scan, the output spectrum of the established Verilog-A model is obtained. This result is then compared with the frequency domain simulation results obtained from the Lumerical INTERCONNECT platform to verify the model's ability to characterize optical field interference, wavelength response, and spatial phase accumulation. A comparison of the frequency domain simulation results of the Verilog-A model and the Lumerical INTERCONNECT platform is provided. Figure 3 As shown, within the investigated wavelength range, the transmission rate curves and logarithmic gain curves obtained by the Verilog-A model and the Lumerical INTERCONNECT platform exhibit largely consistent overall trends, with good correspondence between the main peak and valley positions. For the transmission rate results, the interference peak and valley positions obtained by the two models largely overlap, and the variation patterns near the valley values ​​are consistent. For the gain results, the logarithmic gain spectra obtained by the two models maintain good consistency at the main maxima and minima positions, and the periodic characteristics of the gain curves changing with wavelength are consistent.

[0051] In summary, the Verilog-A model established in this invention can well characterize the frequency domain response characteristics of the traveling wave Mach-Zehnder modulator and can be used for subsequent steady-state, frequency domain and transient joint simulation analysis.

Claims

1. A method for distributed modeling of a traveling wave Mach-Zehnder modulator based on Verilog-A, characterized in that, Includes the following steps: Step 1: Obtain the device parameters of the traveling wave Mach-Zehnder modulator. The device parameters include at least the PN junction parasitic parameters, the effective refractive index variation parameters, and the traveling wave electrode distribution parameters. Step 2: In the Verilog-A language environment, establish optical network nodes in the SPICE circuit simulation environment, use complex envelopes to represent optical signals, and map the real and imaginary parts of the complex envelopes to parallel node variables for transmission. Step 3: In the Verilog-A language environment, establish a basic optical device behavior model based on the optical network node. The basic optical device behavior model includes at least waveguides, beam splitters, beam combiners, and active electro-optic phase shifters. Step 4: In the Verilog-A language environment, establish the electro-optic response model of the active electro-optic phase shifter, and fit the relationship between the bias voltage and the additional phase shift and the relationship between the bias voltage and the additional loss respectively; Step 5: Divide the active electro-optic phase shifter into multiple micro-segments along the light propagation direction, and cascade the RF distributed parameter network and the active electro-optic phase shifter model in each micro-segment to construct a distributed traveling wave Mach-Zehnder modulator model. Step 6: Introduce optical propagation delay and radio frequency propagation delay into the electro-optic response model of the active electro-optic phase shifter to characterize the time delay coupling between light waves and microwaves; Step 7: In the unified SPICE circuit solver, call the distributed traveling wave Mach-Zehnder modulator model to perform steady-state simulation, frequency domain simulation and / or transient simulation of the traveling wave Mach-Zehnder modulator.

2. The distributed modeling method of a traveling wave Mach-Zehnder modulator based on Verilog-A according to claim 1, characterized in that, The optical network nodes establish optical port rules by customizing nature and discipline, enabling optical nodes and electrical nodes to participate in unified solutions in the same circuit simulation environment.

3. The distributed modeling method of a traveling wave Mach-Zehnder modulator based on Verilog-A according to claim 1, characterized in that, The optical signal is transmitted through an optical port array of length 3, wherein the first and second components are used to transmit the real and imaginary parts of the complex envelope, respectively, and the third component is used to transmit the optical carrier wavelength parameter.

4. The distributed modeling method of a traveling wave Mach-Zehnder modulator based on Verilog-A according to claim 1, wherein, The basic optical device behavior model also includes a continuous wave source model and a photodetector model. The continuous wave source model is used to output a complex optical field envelope with a set optical power and wavelength, and the photodetector model is used to introduce a parasitic electrical network based on square-law detection.

5. The distributed modeling method of a traveling wave Mach-Zehnder modulator based on Verilog-A according to claim 1, wherein, In the electro-optic response model of the active electro-optic phase shifter, the additional phase shift and additional loss are both established using polynomial fitting, where the polynomial is a third-order polynomial.

6. The distributed modeling method of a traveling wave Mach-Zehnder modulator based on Verilog-A according to claim 1, wherein, The active electro-optic phase shifter electro-optic response model does not directly use the instantaneous input voltage as the modulation quantity. Instead, it first extracts the effective radio frequency voltage after the radio frequency propagation delay to complete the optical field modulation calculation, and then applies the optical propagation delay to the modulated output optical field.

7. The distributed modeling method of a traveling wave Mach-Zehnder modulator based on Verilog-A according to claim 1, wherein, The RF distributed parameter network includes series resistance, series inductance, parallel parasitic capacitance, substrate leakage resistance, and junction capacitance and junction resistance corresponding to the micro-segment, obtained by length scaling.

Images