SOA segmented modeling simulation design method based on space-time staggered half-step subdivision

Abstract

The invention discloses an SOA segmented modeling simulation design method based on space-time staggered half-step subdivision, and relates to the field of semiconductor laser amplifiers and semiconductor lasers, and the method comprises the steps: in the light amplification direction of an SOA active region, according to the carrier concentration and the electric field component of an optical signal in a space-time staggered half-step subdivision mode, carrying out the space-time staggered half-step subdivision mode; segmenting the SOA active region step by step, and constructing an SOA segmentation model based on spatio-temporal staggered half-step subdivision; respectively simulating the plurality of parameter groups to be simulated by utilizing the constructed SOA segmentation model based on the space-time staggered half-step subdivision to obtain a simulation result of the SOA segmentation model based on the space-time staggered half-step subdivision on each parameter group; and performing data analysis on the simulation result of each parameter group, and determining one or more parameter groups of which the performance meets the requirements as the simulation optimization result of the SOA device. According to the method, a first-order error term can be eliminated, and a model and a method are improved to two orders, so that the simulation precision and the calculation efficiency are improved in order of magnitude.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor laser amplifiers, and in particular to a SOA (Semiconductor Optical Amplifier) ​​segmented modeling simulation design method based on staggered half-step partitioning in time and space. Background Art

[0002] The papers "Semiconductor optical amplifiers: recent advances and applications", "SOA for future PONs" and "Linear and nonlinear semiconductor optical amplifiers" all point out that broadband and short-distance fiber-optic communications include data center optical interconnection and fiber-optic access networks, and SOA and SOA-based semiconductor lasers are widely used in these applications.

[0003] According to the different application categories of SOA in optical communication and optical sensing, a series of numerical calculation and simulation models have been produced. For example, the frequency domain steady-state model proposed in the paper "Wideband semiconductor optical amplifier steady-state numerical model" has been widely used in the basic modeling and calculation of SOA. The detailed time domain models proposed in the papers "Detailed dynamic model for semiconductor optical amplifiers and their crosstalk and intermodulation distortion" and "Modeling and measurement of longitudinal gain dynamics in saturated semiconductor optical amplifiers of different length" are used to simulate and analyze time domain nonlinear effects such as signal crosstalk and gain saturation, and are extended to broadband models in the paper "Wideband dynamic numerical model of a tapered buried ridge stripesemiconductor optical amplifier gate".

[0004] The frequency domain-based model can provide fast simulation results of the longitudinal coupling and evolution of multi-channel signal power and noise power within the signal bandwidth along the SOA, but it cannot provide the time domain evolution information of waveform-level signal amplitude and phase. Coherent optical communications have been commercially available since the 2010s. Because it uses both amplitude and phase to transmit information, it improves the receiving sensitivity compared to incoherent optical communications and reduces the channel spacing between wavelength division multiplexing channels. Coherent optical communications usually use higher-order modulation formats such as 16-QAM (Quadrature Amplitude Modulation) to further improve the communication transmission rate. The SOA model based on the time domain has the ability to finely simulate the time domain evolution of amplitude and phase, making it an essential tool for the design of coherent optical communication-related devices and system optimization.

[0005] High-order modulation formats such as N-QAM can modulate M = log2 N bits of information into a waveform. For M input waveforms with different amplitudes and phases, when the SOA is in a linear state, the amplitudes and phases of the M SOA output waveforms remain relatively unchanged, and the system performance is only affected by ASE (Amplified Spontaneous Emission) noise, but at this time the SOA only works in the small signal amplification range, the signal-to-noise ratio is small, and the system performance is poor. When the SOA is in a deep gain saturation state, the signal-to-noise ratio is large, but at this time the amplitude and phase of the SOA output waveform produce large distortions, greatly reducing system performance. Using the time-domain-based SOA model, the time-domain evolution of the waveform signal including amplitude and phase when the SOA is in a gain saturation state can be accurately simulated to find the balance point between the signal-to-noise ratio and waveform distortion, and optimize system performance.

[0006] Because SOA and semiconductor lasers based on SOA have multiple geometric, physical, and signal parameters, and undergo multiple physical processes such as current injection, carrier transition, and depletion, there are many modeling and design parameters. The joint optimization of these parameters is a major challenge in SOA modeling and design: for example, when there are three parameters that need to be jointly optimized, and each parameter takes 10 values, the simulation calculation required to traverse all combinations reaches 10. 3 That is 1000 times, which requires a huge amount of calculation and computing time.

[0007] Different from the frequency domain steady-state SOA model, the simulation calculation time of the time domain large signal SOA model is relatively long, which seriously restricts the device design efficiency and system optimization effect of SOA. In view of the above phenomenon, there is an urgent need for a new modeling and simulation design method for large signal SOA time domain model with a leap-forward improvement in the calculation efficiency, so as to achieve high-speed traversal of parameter space, which is of great significance not only for SOA and semiconductor lasers based on SOA, but also for the modeling, simulation and optimization design of amplifiers and lasers in other fields, such as amplifiers and lasers based on semiconductors or non-semiconductors. Summary of the invention

[0008] The present invention provides a SOA segmented modeling simulation design method based on staggered half-step partitioning in time and space, which overcomes the problems of low large signal time domain simulation calculation efficiency and long optimization design time consumption of SOA, semiconductor laser based on SOA and other devices.

[0009] The SOA segmentation modeling simulation design method based on time-space staggered half-step segmentation of the present invention comprises: in the light amplification direction of the SOA active region, the SOA active region is segmented step by step according to the time-space staggered half-step segmentation of the carrier concentration and the electric field component of the optical signal, and a SOA segmentation model based on time-space staggered half-step segmentation is constructed; using the constructed SOA segmentation model based on time-space staggered half-step segmentation, a plurality of parameter groups to be simulated are simulated respectively, and a simulation result of each parameter group by the SOA segmentation model based on time-space staggered half-step segmentation is obtained; and data analysis is performed on the simulation result of each parameter group, and one or more parameter groups whose performance meets the requirements are determined as the simulation optimization result of the SOA device.

[0010] Preferably, the method further comprises: acquiring parameter data required for the SOA segmentation model based on the staggered half-step partitioning in time and space, so as to directly use the parameter data for the constructed SOA segmentation model based on the staggered half-step partitioning in time and space.

[0011] Preferably, the parameter data include: bandgap parameter data, optical waveguide parameter data, confinement factor and electro-optical gain parameter data. The bandgap parameter data include carrier concentration at equilibrium, and the acquisition method thereof includes: in a constructed SOA segmentation model based on staggered half-step segmentation in time and space, a simulation is performed with zero input of optical signal as input signal, and the value when the model reaches equilibrium is used as the carrier concentration at equilibrium. The electro-optical gain parameter data include group velocity of the signal, and the acquisition method thereof includes one of the following methods: acquisition from known literature or database; extraction through laboratory measurement.

[0012] Preferably, in the light amplification direction of the SOA active region, the SOA active region is segmented step by step according to the staggered half-step segmentation of the carrier concentration and the electric field component of the optical signal in time and space, and the SOA segmentation model based on the staggered half-step segmentation in time and space is constructed, which includes: using the pre-acquired equilibrium carrier concentration as the initial value of the carrier concentration of the semiconductor laser amplifier optical waveguide; dividing the semiconductor laser amplifier optical waveguide device according to the transmission direction of the optical signal to obtain multiple sections of semiconductor laser amplifier optical waveguides, each with a length of Δz; sampling the input optical signal according to the sampling time interval Δt to obtain M sampling points, wherein Δz=v g Δt,v g is the group velocity of the optical signal propagating in the optical waveguide of the semiconductor laser amplifier; according to the differential equation satisfied by the SOA physical model, the iterative equation of the calculation point of the optical signal is determined, wherein the carrier concentration N[z,t] on the corresponding segment of the semiconductor laser amplifier optical waveguide contained in the iterative equation of the calculation point of the optical signal is the carrier concentration on the segment of the semiconductor laser amplifier optical waveguide after Δt / 2.

[0013] Preferably, in the optical amplification direction of the SOA active region, the SOA active region is segmented step by step in a manner of staggered half-step segmentation of the carrier concentration and the electric field component of the optical signal in time and space, and constructing a SOA segmentation model based on staggered half-step segmentation in time and space also includes: determining an iterative equation for the calculation point of the carrier concentration according to a rate equation satisfied by the semiconductor carrier concentration, wherein the signal intensity E[z,t] on the corresponding segment of the semiconductor laser amplifier optical waveguide contained in the iterative equation for the calculation point of the carrier concentration is the average value of the signal intensity at the two end points of the segment of the semiconductor laser amplifier optical waveguide after Δt / 2.

[0014] Preferably, the multiple parameter groups to be simulated are obtained through the following steps: selecting adjustable and optimizable parameters from pre-acquired parameters of semiconductors or other electro-optical gain materials as parameters to be simulated; generating multiple parameter groups to be simulated based on the parameters to be simulated and their ranges, wherein each parameter group to be simulated includes all parameters to be simulated and parameter values ​​within their ranges.

[0015] Preferably, the data analysis of the simulation results of each parameter group to determine one or more parameter groups whose performance meets the requirements as the simulation optimization results of the SOA device includes: calculating the error vector amplitude of the simulation results of each parameter group; selecting and saving one or more parameter groups whose error vector amplitude meets the requirements from multiple parameter groups as the simulation optimization results of the SOA device.

[0016] Preferably, the method further includes: assigning one or more parameter groups whose error vector amplitude meets the requirements to the SOA contained in the system model in sequence, and simulating the system model to obtain the system simulation result of the system model for each parameter group whose error vector amplitude meets the requirements; calculating the bit error rate of the system simulation result of each parameter group, and selecting one or more parameter groups whose bit error rate meets the requirements as the simulation optimization result of the system model.

[0017] The SOA segmentation model based on staggered half-step partitioning in time and space constructed by the present invention can eliminate the first-order error terms of a simple segmentation model with first-order accuracy, and improve the model and method to the second order, thereby improving the simulation accuracy and calculation efficiency by orders of magnitude and reducing the time consumption of optimization design. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of the SOA segmented modeling and simulation design method based on staggered half-step partitioning in time and space;

[0019] Figure 2 This is a schematic diagram of the structure of a common bulk semiconductor laser amplifier;

[0020] Figure 3 It is a schematic diagram of SOA segmentation and signal sampling;

[0021] Figure 4 It is a schematic diagram of the simulation algorithm for SOA;

[0022] Figure 5 is a schematic diagram of a system example of system optimization;

[0023] Figure 6 It is a flow chart of the system optimization design of semiconductor laser amplifier;

[0024] Figure 7 It is a comparison chart of the output waveforms of the two partitioning methods and the true solution;

[0025] Figure 8 It is a comparison of the global errors of the results of the two subdivision methods. DETAILED DESCRIPTION

[0026] The specific implementation of the present invention is described in detail below. It is necessary to point out that the following implementation is only used for further explanation of the present invention and cannot be understood as limiting the scope of protection of the present invention. Some non-essential improvements and adjustments made to the present invention by persons skilled in the art based on the above-mentioned content of the present invention still fall within the scope of protection of the present invention.

[0027] See also Figure 1 The SOA segmented modeling and simulation design method based on staggered half-step partitioning in time and space can include the following steps:

[0028] Step S101: In the light amplification direction of the SOA active region, the SOA active region is segmented step by step according to the staggered half-step segmentation of the carrier concentration and the electric field component of the optical signal in time and space, and a SOA segmentation model based on the staggered half-step segmentation in time and space is constructed;

[0029] Step S102: using the constructed SOA segmentation model based on the staggered half-step partitioning in time and space, respectively simulating multiple parameter groups to be simulated, and obtaining the simulation result of each parameter group based on the SOA segmentation model based on the staggered half-step partitioning in time and space;

[0030] Step S103: performing data analysis on the simulation results of each parameter group, and determining one or more parameter groups whose performance meets the requirements as simulation optimization results of the SOA device.

[0031] Further, the step S102 includes: using the pre-acquired carrier concentration at equilibrium as the initial value of the carrier concentration of the semiconductor laser amplifier optical waveguide; dividing the semiconductor laser amplifier optical waveguide device according to the transmission direction of the optical signal to obtain multiple sections of semiconductor laser amplifier optical waveguides, each with a length of Δz; sampling the input optical signal according to the sampling time interval Δt to obtain M sampling points, wherein Δz=v g Δt,v g is the group velocity of the optical signal propagating in the optical waveguide of the semiconductor laser amplifier; according to the differential equation satisfied by the SOA physical model, the iterative equation of the calculation point of the optical signal is determined; and according to the rate equation satisfied by the semiconductor carrier concentration, the iterative equation of the calculation point of the carrier concentration is determined. Among them, the carrier concentration N[z,t] on the corresponding section of the semiconductor laser amplifier optical waveguide contained in the iterative equation of the calculation point of the optical signal is the carrier concentration on the section of the semiconductor laser amplifier optical waveguide after Δt / 2, and the signal intensity E[z,t] on the corresponding section of the semiconductor laser amplifier optical waveguide contained in the iterative equation of the calculation point of the carrier concentration is the average value of the signal intensity at the two end points of the section of the semiconductor laser amplifier optical waveguide after Δt / 2.

[0032] Furthermore, after step S104, the method further includes: assigning one or more parameter groups whose error vector amplitude meets the requirements to the SOA contained in the system model in sequence, simulating the system model to obtain the system simulation result of each parameter group, performing data analysis on the system simulation result of each parameter group, and determining one or more parameter groups whose performance meets the requirements as the simulation optimization result of the system model.

[0033] The present invention provides a new time-space staggered half-step partitioning SOA modeling and high-speed and efficient simulation design method for optical communication semiconductor laser amplifiers and semiconductor laser technologies. The method has second-order accuracy. Compared with the simple segmented model with first-order accuracy used in the prior art, the present invention eliminates the first-order error term by partitioning the active area of ​​the amplifier in time and space, thereby improving the simulation accuracy and computational efficiency by orders of magnitude, allowing more efficient modeling and faster simulation of complex SOA devices, and increasing the number of optimization iterations and the range of traversed parameter space by orders of magnitude within a limited time, which can greatly improve the relevant electrical and optical performance indicators of the designed optical communication semiconductor laser amplifiers, semiconductor lasers and similar semiconductor devices.

[0034] Below through Figures 2 to 8 Provide detailed explanation.

[0035] See also Figure 2 The structure of a common bulk semiconductor laser amplifier mainly includes an electrode 1 for providing bias current, an active region 2, an anti-reflection surface 3, an insulating layer 4, etc., wherein the z direction is the propagation direction of the signal light, and L is the total length of the SOA device. After the optical signal is coupled into the SOA, the high-energy carriers in the active region amplify the optical signal due to stimulated radiation. At the same time, due to the pumping of the bias current, the consumed high-energy carriers are continuously replenished to ensure the continuation of the amplification process. When the optical signal reaches the anti-reflection layer, most of the signal is used as the output signal, and only a very small part of the signal becomes a reverse propagation signal and propagates in the reverse direction of the SOA. Generally speaking, even after amplification, the proportion of the reflected signal in the final output light is still less than the proportion of the spontaneous radiation noise in the output light, so it can be ignored in practice. The present invention still takes into account the reflected light signal and its amplification, and the present invention can accurately simulate this effect. Therefore, the modeling, simulation, and optimization design method of the semiconductor laser amplifier optical waveguide based on the staggered half-step segmentation in time and space of the present invention involves the following four main steps:

[0036] Step (1): extracting electrical, optical and electro-optical coupling parameter data of semiconductors or other electro-optical gain materials and SOA devices, such as limitation factors and electro-optical gain parameter data.

[0037] Wherein, step (1) is specifically as follows:

[0038] Step (1.1): Model and simulate the electrical properties of semiconductors or other electro-optical gain materials and extract bandgap parameter data. The parameters extracted in this step include: recombination coefficient (Shockley–Read–Hall recombination coefficient, Bimolecular recombination coefficient, Auger recombination coefficient, etc.), effective mass of light and heavy holes, bandgap width, carrier lifetime, carrier concentration at equilibrium, etc. The intrinsic parameters of these materials are detailed in relevant papers and databases such as "Wideband Semiconductor Optical Amplifier Steady-State Numerical Model". In industrial design applications, these parameters can also be obtained from laboratory tests. The relevant parameters after other process modifications such as doping can be found in the calculation formula or model in the relevant semiconductor physics textbooks, or obtained from laboratory tests. For the carrier concentration at equilibrium, an estimated value can be used. In the model established in the following step (2), the optical signal zero input is used as the input signal. After a period of simulation, the value when the system reaches equilibrium is set as the carrier concentration at equilibrium.

[0039] Step (1.2): Modeling and simulating the optical waveguide composed of semiconductor or other electro-optical gain materials and extracting the optical waveguide parameter data. The parameters extracted in this step specifically include: SOA length, area of ​​active area, confinement factor, scattering loss coefficient of waveguide, reflectivity of incident and exit interfaces.

[0040] Among them, the SOA length and the area of ​​the active region are parameters that can be manually determined during the SOA design process and are usually set to a certain range. The relevant parameter values ​​are traversed through the subsequent steps (3) and (4) to design an SOA device that achieves system optimization.

[0041] Among them, the book "Physics of Optoelectronic Devices" gives the definition of the limiting factor Γ:

[0042]

[0043] Where E and H represent the electric field and magnetic field distribution at the signal input end surface, respectively. is the unit vector along the signal propagation direction, and ds is the tiny surface element on the signal input end face.

[0044] Therefore, there are two methods to calculate the restriction factor: ① Through the relevant FDTD electromagnetic simulation modeling, the light mode of the corresponding SOA structure is simulated, and the restriction factor of the corresponding structure can be calculated using the above formula. ② For SOA structures with analytical solutions or approximate analytical solutions, the restriction factor can be calculated using analytical / numerical calculations.

[0045] Step (1.3): Extract the limiting factor and electro-optical gain parameter data of semiconductor or other electro-optical gain materials. The parameters extracted in this step specifically include: differential gain parameter and its frequency response, carrier concentration in transparency and its frequency response, group velocity of the signal and its frequency response, and gain-phase coupling parameter.

[0046] Among them, the differential gain dg / dN at the frequency of the input signal is determined using the following gain parameter formula in the paper "Wideband Semiconductor Optical Amplifier Steady-State Numerical Model" and the parameters obtained in step (1.1).

[0047]

[0048] The letters represent: g-SOA gain, ν-carrier frequency, N-carrier concentration, c-speed of light, n1-refractive index of SOA active area, τ s -Radiative carrier recombination lifetime, m e / m hh -Effective mass of conduction band electrons / valence band heavy holes, E g -Band gap width, f c / f v -Fermi distribution function in the conduction band / valence band, T0-average lifetime of electrons under the action of an external field.

[0049] The parameters described in step (1.1) above, in addition to the acquisition methods and steps described above, can also be obtained from published literature, material databases and laboratory tests.

[0050] Among them, the carrier concentration and its frequency response during transparency can be obtained from simulation and laboratory tests, the group velocity of the signal and its frequency response can be obtained from optical waveguide calculations, and the gain-phase coupling parameters can be obtained from laboratory measurements.

[0051] In one example, parameters are extracted from relevant literature, databases or laboratory measurements and recorded in Table 1. For parameters that can be artificially controlled and to be optimized, such as the length, width and geometry of the active region, the length of the device, etc., they are set to a series of values ​​to be simulated.

[0052] Table 1. The correspondence between the values ​​and letters of a set of parameters selected in the example and the physical quantities.

[0053]

[0054] Step (2): The light amplification direction of the SOA active region is segmented step by step in a staggered half-step manner in time and space, a time-domain mathematical and physical model of the light field and material gain segmentation is established, and a simulation formula and algorithm based on the model are obtained, and simulation calculations are implemented by hardware or software according to the above method.

[0055] Wherein, step (2) comprises:

[0056] Step (2.1): Use the carrier concentration at equilibrium given in step (1) to determine the initial value of the carrier concentration of the semiconductor laser amplifier optical waveguide.

[0057] Specifically, the carrier concentration at equilibrium obtained in step (1) is taken as N p＝0 =N p＝0.5 p is the subscript used in the calculation of carrier concentration, i.e., the time index (N p＝0 With N p＝0.5 They represent the carrier concentrations at time 0 and time 0.5Δt respectively, where Δt is the simulation time step).

[0058] Step (2.2): Mathematically divide the semiconductor laser amplifier optical waveguide (device) into multiple segments, such as N segments, each with a length of Δz, in the direction of optical signal transmission, such as the z direction. Use software or hardware to generate thousands of bits of 16-QAM signals, and mathematically divide the input optical signal to be simulated (or the optical signal to be input) into multiple segments, such as M segments, at time t, each with a length of Δt, and sample at the beginning of each segment to obtain {E(mΔt)}, or, sample the optical signal to be input at intervals of Δt in time to obtain M sampling points in total, and obtain an array of M complex numbers as the input signal for simulation. At the same time, ensure that Δt = Δz / v g ,v g represents the group velocity of the optical signal in the semiconductor medium, obtained by step (1).

[0059] See also Figure 3 The upper part is a schematic diagram of the subdivision of a SOA device of length L along the light propagation direction (z direction), and the length of each subdivided segment is Δz. The lower part is the sampling of the input signal in time, with the horizontal axis representing time and the vertical axis representing the instantaneous power of the signal, and the sampling interval is Δt. Ensure that Δz = v g Δt,v g is the group velocity of the signal in the device.

[0060] So far, the mathematical model is accurate for the physical process: in each Δz segment, the optical signal propagates forward and backward in the waveguide, and is amplified under the action of stimulated recombination of unbalanced carriers. The unbalanced carrier density evolves under the multiple effects of applied current, optical signal and spontaneous recombination. However, the above mathematical model has no analytical solution for any signal, so the SOA is first divided into several small segments of length Δz for numerical calculation. Then proceed to the next step. In academic research and product development so far, the calculation point of the optical signal is selected as the midpoint or end point of the Δz segment of each SOA in space, and the calculation point of the carrier concentration is selected as the midpoint of the SOA Δz segment. At the same time, in time, the calculation points of the two are selected at the same time. In the present invention, the above numerical modeling method is called the synchronous segmentation modeling method, and its essence is to regard the amplification of the optical signal on each SOA Δz segment as the result of stimulated radiation of the carrier concentration on the active area of ​​the segment at the same time. Moreover, the change in carrier concentration on each SOAΔz segment is calculated as the effect of the average value of the signal strength at the center or two end points of the Δz segment at the same time. The above approximation of simulating continuous distributed amplification as segmented amplification is obvious, because this changes the real physical transmission mechanism, thus causing calculation simulation errors. When Δz and Δt become smaller, the simulation error decreases and the simulation accuracy improves. When Δz tends to 0, the simulation error gradually disappears. However, when Δz and Δt become smaller, the total number of steps required for calculation becomes larger. When Δz and Δt decrease and tend to 0, the total number of steps required for calculation tends to infinity, making calculation, simulation, and optimization impossible to achieve. In the research of the present invention, it was found that the synchronous segmentation modeling method described above is a first-order method, that is, every time Δz decreases by 2 times, the simulation error decreases by 2 times. To achieve a certain simulation high precision, the required Δz is generally very small, so that the number of calculation steps is large, resulting in a long calculation simulation time and low efficiency.

[0061] Step (2.3): Establish the calculation point of the optical signal to be calculated It represents the electric field of the optical signal at z = nΔz at the time t = mΔt. Correspondingly, the value range of z is an integer from 0 to N+1. The differential equation satisfied by the recognized physical model of semiconductor laser amplifier is as follows: Among them, E + It represents the electric field of the signal transmitted in the forward direction (z direction), E - Represents the signal electric field transmitted in the reverse direction (negative direction of z), Γ, a, N0, α H and α L They are respectively the confinement factor, differential gain, carrier concentration in transparency, gain-phase coupling parameter and waveguide scattering loss. Based on the above differential equations, the iterative equations of the calculation points of the optical signal can be obtained: In the above formula, N[z,t] is preferably N[(n±0.5)Δz,(m+0.5)Δt]. In the common method, N[z,t] is selected as N[nΔz,mΔt] or N[(n±0.5)Δz,mΔt].

[0062] Step (2.4): Establish calculation points for carrier concentration It represents the carrier concentration at the moment t = pΔt and z = qΔz. Among them, p and q are both half integers greater than 0 (i.e. 0.5, 1.5, 2.5...). Calculate the carrier concentration of each section on the SOA after the light signal enters the active area Δt / 2 time. At this time, the light signal only enters the active area Δz / 2, and the impact on the active area is negligible. Therefore, the carrier concentration remains in a dynamic balance between injection and recombination of the electric pump and remains unchanged, that is, N 0.5Δt =N0, N0 is the carrier concentration at equilibrium, which is the starting point of iteration and can also be called the initial carrier concentration.

[0063] Step (2.5), according to, we can get the iterative equation of carrier concentration.

[0064] Among them, the rate equation satisfied by the semiconductor carrier concentration is as follows:

[0065]

[0066] Among them, I, e, and V represent the SOA driving current, elementary charge, and volume of the SOA active area, respectively. c is the cross-sectional area of ​​the SOA active region, Represents the energy of the photon; R is a composite term related to the carrier concentration, specifically in the form of R(N)=AN+BN 2 +CN 3 , A, B, and C are three composite coefficients. The remaining letters have the same meanings as those in the equations related to the electric field in the previous text.

[0067] The iterative equation for carrier concentration is as follows:

[0068]

[0069] In the above formula, E[z,t] is selected as The E can represent E + or E - The usual method selects the time index p of the carrier concentration N as an integer rather than a half-integer, and E ± [z,t] is or That is, the electric field component of the optical signal and the calculation point of the carrier concentration have the same time index.

[0070] The process of selecting the calculation points of the optical signal electric field and carrier concentration in the above steps (2.3), (2.4) and (2.5) is called time-space staggered half-step subdivision. The closer the time-space step length Δz and Δt are to 0, the closer they are to the actual situation.

[0071] See also Figure 4 Figure a) shows a common arrangement of calculation points after segmentation. The horizontal axis represents the direction of light signal propagation, the vertical axis represents the time direction, and the black dots represent the carrier concentration (N), the electric field of the forward propagating signal (E + ) and the electric field of the reverse propagating signal (E - ) calculation point. The brackets in the figure indicate the time and space position of the calculation point. For example, N(4,1) indicates the carrier concentration calculation point located at the time 4Δz, Δt in space. Figure 4 b) in the figure shows the arrangement of calculation points used in the present invention. The horizontal axis represents the direction of light signal propagation, and the vertical axis represents the time direction. The black dots represent the electric field (E + ) and the electric field of the reverse propagating signal (E - ), and the hollow dots represent the calculation points of the carrier concentration (N). The cross box and two rounded rectangles in the figure represent the calculation points involved in the update equations of the carrier concentration, the electric field of the forward propagating signal, and the electric field of the reverse propagating signal, respectively. Among them, the carrier concentration below the cross box and the electric field values ​​at the left and right ends are used to calculate the carrier concentration at the next moment, that is, above the cross box; the electric field value at the lower left corner of the rounded rectangle and the carrier concentration at the midpoint are used to calculate the electric field value at the next moment, that is, the upper right corner. For the reverse propagating signal, the dotted rounded rectangle box is used for update calculation. The asynchronous (half-step) decomposition method of the Δz segment proposed in the present invention is a second-order method, which advances the calculation point of the carrier concentration by Δt / 2 in time, and calculates the optical signal amplified on each section of the semiconductor laser amplifier optical waveguide as the influence of the carrier concentration on the section of the waveguide after Δt / 2. In addition, the change in carrier concentration on each section of the semiconductor laser amplifier optical waveguide is calculated as the average value of the signal intensity at the two end points of the waveguide after Δt / 2. The above-mentioned modeling method for mathematical model division of the semiconductor laser amplifier and the update calculation method of the electric field value and the carrier concentration are collectively referred to as the optimal semiconductor laser amplifier modeling simulation method with staggered half-step division in time and space.

[0072] Depend on Figure 4It can be seen that compared with the synchronous segmentation method, the present invention has a similar computational complexity when having the same number of calculation points and calculation amount, that is, when the device length d is equal and Δz is equal. However, it has been verified by computer simulation calculation that the half-step segmentation modeling method of the present invention has second-order accuracy. Compared with the first-order accuracy of the synchronous modeling method described in the paper "Wideband dynamic numerical model of a tapered buried ridge stripe semiconductor optical amplifier gate" (IEE Proceedings-Circuits, Devices and Systems, VOL.149, NO.3, JUNE 2002), the present invention changes the spatiotemporal position of the calculation point of the carrier concentration and performs the update calculation of the electric field value and the carrier concentration as described above, thereby offsetting the first-order error term and improving the simulation accuracy by orders of magnitude. Therefore, the computational efficiency can be improved by orders of magnitude by reducing the number of segments N under the same simulation accuracy. Specifically, for the second-order method, every time Δz decreases by 2 times, the simulation error is reduced by 2. 2 =4 times; when Δz decreases by 10 times, the simulation error decreases by 10 2 = 100 times. To achieve a certain simulation high precision, the required Δz can be larger, resulting in fewer calculation steps, resulting in shorter calculation simulation time, and the calculation efficiency is greatly improved compared to the first-order method.

[0073] Step (2.6): Iteratively calculate the optical signal and carrier concentration using the input optical signal, the initial carrier concentration, and the spatial boundary response to the optical signal to obtain as output signal.

[0074] In short, in step (2), the prior art adopts the method of synchronous integer division of carrier concentration and electric field time, while the present invention adopts the method of staggered half-step division of carrier concentration and electric field time and space for the first time. Compared with the prior art, the number of calculation points of the present invention is not increased, and the amount of calculation is the same. However, the present invention eliminates the first-order error term by staggered half-step division of semiconductor laser amplifier optical waveguide in time and space, thereby improving the simulation accuracy by orders of magnitude. The calculation efficiency can be improved by orders of magnitude by reducing the number of segments N at the same simulation accuracy.

[0075] It should be pointed out that the scope of the present invention is not limited by the above description of segmentation. For example, the present invention is also applicable to segmentation of other staggered steps of carrier concentration and electric field.

[0076] Step (3): Perform data analysis and extraction on the device simulation results to obtain device and system design parameter results, including input signal power, SOA geometric parameters such as length, pump current and other device optimization parameter results.

[0077] Specifically, for the multiple input parameters in step (1), simulations are performed in sequence, the corresponding EVM (error vector magnitude) is calculated, and several parameter combinations with better performance are selected as the candidate parameter space of step (4).

[0078] Step (4): Perform system-level simulation and calculation using a system model including a semiconductor laser amplifier to obtain system simulation and optimization design results.

[0079] In the system simulation of practical applications, for example Figure 5 In the coherent optical communication system based on M-order quadrature amplitude modulation QAM shown in the figure, the optical fiber input power is a system-level candidate parameter that needs to be optimized, and its corresponding SOA device candidate parameter can be the pump current. Applying the modeling of step (2), the candidate parameter space obtained in step (3), such as the size of the pump current, is traversed by system simulation to determine the final optimization result.

[0080] The SOA system optimization process of steps (3) and (4) is shown in Figure 6 , specifically including:

[0081] Step ①: Determine the optimization parameters (such as the length and geometry of the active region, the length of the device, the pump current, etc.) and their possible ranges, and generate multiple parameter groups to be simulated with appropriate step sizes based on the ranges;

[0082] Step ②: Determine whether device simulation has been performed for all parameter groups, if yes, execute step ⑥, if not, execute step ③;

[0083] Step 3: Use the method in step (2) to simulate the device and calculate the EVM (error vector magnitude) of the output result;

[0084] Step ④: When it is determined that the EVM of the parameter group meets the requirements, execute step ⑤;

[0085] Step ⑤: Save the parameter group to the parameter group of system simulation;

[0086] Step ⑥: Perform system simulation and calculate the BER (bit error rate) of the output result;

[0087] Step ⑦: Determine whether the BER of the parameter group meets the requirements, if yes, execute step ⑧, if not, execute step ⑥;

[0088] Step ⑧: Save the parameter group to the simulation optimization results.

[0089] The following is a numerical simulation verification and comparison of the traditional synchronous subdivision method and the segmented modeling simulation design method of the semiconductor laser amplifier optical waveguide based on the half-step subdivision of the present invention.

[0090] Using the parameters given in Appendix 1, a numerical simulation model is established. A set of binary numbers randomly generated by a computer is used as the transmitted data. The set of binary numbers is encoded into 512 symbols in the 16QAM modulation format. The signal interval Δt corresponding to each symbol is sampled into several discrete values ​​to form a one-dimensional array. Simulations are performed in the manner of synchronous splitting "split 1" and half-step splitting "split 2", and the optical signal at the end of the semiconductor laser amplifier optical waveguide is recorded as the output signal E out , note that its numerical accuracy depends on the chosen Δz. As described in the paper "Symmetrized SSF scheme to controlglobal simulation accuracy in fiber optic communication systems" (IEEE JOURNALOF LIGHTWAVE TECHNOLOGY, VOL. 26, NO. 2, JANUARY 2008), the normalized standard deviation (nsd) is used as the global error metric, but is defined as the difference between the approximate solution for a particular step size Δz and the approximate solution for a very small Δz or a very large number of steps per symbol N steps The difference between the "true solutions" obtained.

[0091] See also Figure 7 , Split1 is the output result of synchronous splitting, Split2 is the output result of staggered half-step splitting, and the True solution is to use the split1 method (i.e., synchronous splitting format) to take the number of steps N used for each symbol steps is the "true solution" obtained in 4096 steps. The output signal waveforms obtained by the two dissection methods are compared with the true solution. steps The result of half-step decomposition is closer to the true solution.

[0092] The number of steps N used for each symbol steps Global error with nsd The relationship is as follows Figure 8As shown, the horizontal axis is the number of steps used for each symbol, and the vertical axis is the global error. FO-FDM is a first-order finite difference method, i.e., a common synchronous partitioning format, and SO-FDM is a second-order finite difference method, i.e., the staggered half-step partitioning of the present invention. Among them, the half-step partitioning method, i.e., split 2, is proved to be a second-order method by simulation results: when the step size is halved, the nsd error is reduced to 1 / 4. The synchronous partitioning method, i.e., split 1, is proved to be a first-order method, and in order to achieve the same nsd, it requires more simulation steps. At the accuracy or error level of , compared with the synchronous partitioning method, the half-step partitioning method requires only 1 / 3 of the number of partitions of the former. Since the partitions are performed in both time and space, the number of calculation steps used is only 1 / 9 of the former. At the accuracy level of , compared with the synchronous splitting method, the half-step splitting method requires about 1 / 10 of the number of splits, that is, 1 / 100 of the number of calculation steps, which can improve the calculation efficiency by about 100 times. In the case of optical communication using high-capacity error correction codes, the accuracy level is required to be higher. In this case, the staggered half-step splitting method may speed up more than the synchronous splitting method.

[0093] In summary, the present invention has the following technical effects:

[0094] (1) In the simulation calculation of semiconductor laser amplifiers, semiconductor lasers, and various electrical and optical devices based on the physical mechanism or mathematical model of semiconductor laser amplifiers, the present invention solves the problem of low calculation efficiency and long optimization design time caused by the current common first-order accuracy modeling simulation calculation method, and provides a new calculation point distribution method for the high-speed simulation field of optical communication amplifiers;

[0095] (2) The present invention offsets the calculation points of the segmented semiconductor laser amplifier, so that the calculation points of the optical signal electric field and the calculation points of the carrier density are offset by half the segmentation width in time and space. This offset is the key to achieving second-order accuracy and is a significant performance improvement over existing first-order accuracy modeling and simulation technology.

[0096] (3) The present invention eliminates the first-order error term by offsetting the calculation points of the segmented semiconductor laser amplifier, thereby improving the simulation accuracy and calculation efficiency by orders of magnitude, allowing more realistic large-signal modeling and faster large-signal simulation of complex semiconductor laser amplifiers, and increasing the number of optimization iterations by orders of magnitude within a limited time, which can greatly improve the performance indicators of the designed optical communication high-speed semiconductor laser amplifier.

[0097] Although exemplary embodiments of the present invention have been described for the purpose of illustration, it will be understood by those skilled in the art that various modifications, additions and substitutions may be made in form and detail without departing from the scope and spirit of the invention disclosed in the appended claims, and all such changes shall fall within the scope of protection of the appended claims of the present invention, and the various steps in the various departments and methods of the products claimed by the present invention may be combined together in any combination. Therefore, the description of the embodiments disclosed in the present invention is not intended to limit the scope of the present invention, but is used to describe the present invention. Accordingly, the scope of the present invention is not limited by the above embodiments, but is defined by the claims or their equivalents.

Claims

1. A SOA segmented modeling simulation design method based on staggered half-step partitioning in time and space, characterized in that: include: In the light amplification direction of the SOA active region, the SOA active region is segmented step by step according to the staggered half-step segmentation of the carrier concentration and the electric field component of the optical signal in time and space, and a SOA segmentation model based on the staggered half-step segmentation in time and space is constructed; Using the constructed SOA segmentation model based on staggered half-step partitioning in time and space, multiple parameter groups to be simulated are simulated respectively, and the simulation results of each parameter group based on the SOA segmentation model based on staggered half-step partitioning in time and space are obtained; Data analysis is performed on the simulation results of each parameter group to determine one or more parameter groups whose performance meets the requirements as simulation optimization results of the SOA device.

2. The method according to claim 1, characterized in that The method further comprises: The parameter data required by the SOA segmentation model based on the staggered half-step segmentation in time and space are obtained, so as to directly use the parameter data in the constructed SOA segmentation model based on the staggered half-step segmentation in time and space.

3. The method according to claim 2, characterized in that The parameter data include: band gap parameter data, optical waveguide parameter data, confinement factor and electro-optical gain parameter data.

4. The method according to claim 3, characterized in that The bandgap parameter data includes carrier concentration at equilibrium, and the acquisition method includes: in the constructed SOA segmented model based on staggered half-step segmentation in time and space, simulation is performed with zero input of optical signal as input signal, and the value when the model reaches equilibrium is used as the carrier concentration at equilibrium.

5. The method according to claim 3, characterized in that: The electro-optical gain parameter data includes the group velocity of the signal, and the acquisition method includes one of the following methods: acquiring from known literature or database; extracting through laboratory measurement.

6. The method according to claim 1, characterized in that The SOA active region is segmented step by step in the light amplification direction of the SOA active region according to the staggered half-step segmentation of the carrier concentration and the electric field component of the optical signal in time and space, and the SOA segmentation model based on the staggered half-step segmentation in time and space is constructed, including: The carrier concentration at equilibrium obtained in advance is used as the initial value of the carrier concentration of the semiconductor laser amplifier optical waveguide; According to the transmission direction of the optical signal, the semiconductor laser amplifier optical waveguide device is divided to obtain a plurality of semiconductor laser amplifier optical waveguides with each section having a length of Δz; The input optical signal is sampled at a sampling time interval Δt to obtain M sampling points, where Δz = v g Δt,v g is the group velocity of the optical signal propagating in the optical waveguide of the semiconductor laser amplifier; According to the differential equation satisfied by the SOA physical model, an iterative equation of the calculation point of the optical signal is determined, wherein the carrier concentration N[z,t] on the corresponding segment of the semiconductor laser amplifier optical waveguide contained in the iterative equation of the calculation point of the optical signal is the carrier concentration on the segment of the semiconductor laser amplifier optical waveguide after Δt / 2.

7. The method according to claim 6, characterized in that In the light amplification direction of the SOA active region, the SOA active region is segmented step by step according to the staggered half-step segmentation of the carrier concentration and the electric field component of the optical signal in time and space, and the construction of the SOA segmentation model based on the staggered half-step segmentation in time and space also includes: According to the rate equation satisfied by the semiconductor carrier concentration, an iterative equation for the calculation point of the carrier concentration is determined, wherein the signal intensity E[z,t] on the corresponding segment of the semiconductor laser amplifier optical waveguide contained in the iterative equation for the calculation point of the carrier concentration is the average value of the signal intensity at the two end points of the segment of the semiconductor laser amplifier optical waveguide after Δt / 2.

8. The method according to claim 1, characterized in that: The multiple parameter groups to be simulated are obtained by the following steps: Selecting adjustable and optimizable design parameters from pre-acquired parameters of semiconductors or other electro-optical gain materials as parameters to be simulated; According to the parameters to be simulated and their ranges, a plurality of parameter groups to be simulated are generated, wherein each parameter group to be simulated includes all the parameters to be simulated and parameter values ​​within the range thereof.

9. The method according to claim 1, characterized in that: The step of performing data analysis on the simulation results of each parameter group and determining one or more parameter groups whose performance meets the requirements as the simulation optimization results of the SOA device includes: Calculate the error vector magnitude of the simulation results for each parameter group; One or more parameter groups whose error vector amplitude meets the requirements are selected from the multiple parameter groups and saved as simulation optimization results of the SOA device.

10. The method according to any one of claims 1 to 9, characterized in that: The method further comprises: sequentially assigning one or more parameter groups for which the error vector amplitude meets the requirements to the SOA included in the system model, and simulating the system model to obtain a system simulation result of the system model for each parameter group for which the error vector amplitude meets the requirements; The bit error rate of the system simulation result of each parameter group is calculated, and one or more parameter groups whose bit error rates meet the requirements are selected as the simulation optimization results of the system model.

Images