A terahertz waveguide directional coupler with low amplitude unevenness and implementation method

Abstract

The application relates to a terahertz waveguide directional coupler with low amplitude unevenness and an implementation method, which comprises the following steps: A1: selecting a standard rectangular waveguide, expanding the wide side a in the transverse direction to form upper and lower wide rectangular waveguides; A2: periodically inserting N rectangular branch sections, including wide branch sections and narrow branch sections, into the two-layer waveguides, and the sections share a height c; A3: constructing M-stage stepped sections at the two ends of the wide waveguide; A4: performing cooperative optimization on the variable group {w, c, x, X}; and A5: forming a four-port terahertz waveguide directional coupler. Through the wide waveguide structure, the multi-stage stepped section and the coupling branch section design, the amplitude consistency in the 185GHz-200GHz frequency band is improved, and the amplitude unevenness is reduced. The application solves the problem of large amplitude unevenness of the traditional waveguide directional coupler under the wideband condition, and has a good engineering application prospect.

Description

Technical Field

[0001] The present invention relates to the field of terahertz technology, and in particular to a waveguide directional coupler with low amplitude unevenness designed for the terahertz frequency band and an implementation method thereof. Background Art

[0002] A directional coupler is a passive device used in the RF and microwave fields. It has functions such as power distribution and synthesis, and can also be used for power detection and signal component analysis. There are many forms of directional couplers. In the low-frequency microwave band, microstrip lines are often used. However, in the terahertz band, the loss of microstrip and other transmission lines is too large, so waveguide transmission lines are often used. Waveguides are closed metal cavities with the advantages of high power capacity, high Q value, and low loss. They are the most commonly used transmission lines in the terahertz band. The terahertz waveguide directional coupler based on waveguide also has the above advantages. The waveguide directional coupler is a four-port device. Compared with the three-port power splitter, it can meet the requirements of full port matching, reciprocity, and losslessness, while the three-port device can only meet two of them.

[0003] In the terahertz frequency band, the most common use of waveguide directional couplers is the power splitting / combining function. The traditional waveguide directional coupler structure is as follows: Figure 1 As shown in the figure, input is taken from input port 1 and is equally divided at output port 2 and coupled port 3, meaning the power is split in half. The opposite is true for combined power. Isolation port 4 is an isolated port and is connected to a matched load during operation. Amplitude non-flatness refers to the difference in power amplitude between ports 2 and 3. Ideally, it would be best to distribute power equally between output port 2 and coupled port 3. However, in practice, there is a difference in power between these two ports, and this difference should be minimized to ensure low amplitude non-flatness.

[0004] Problems with traditional structures and other inventions:

[0005] Figure 1 The amplitude non-flatness of the two ports (output port 2 and coupling port 3) of the traditional terahertz waveguide directional coupler is poor, and it is impossible to achieve both broadband and low amplitude non-flatness at the same time.

[0006] The invention (publication number: CN114725644A, hereinafter referred to as invention [1]) proposes a structure that, near the 192 GHz frequency band, has an amplitude non-flatness that is superior to that of a conventional waveguide directional coupler with the same operating bandwidth. Among its S parameters, S11 is superior to -15 dB, the amplitude non-flatness |S21-S31| is less than 0.06 dB, and S41 is superior to -15 dB; whereas the conventional directional coupler has an S11 superior to -25 dB, an amplitude non-flatness |S21-S31| is less than 0.5 dB, and S41 superior to -25 dB. It can be seen that, in addition to the amplitude non-flatness being superior to that of the conventional structure, the other parameters S11 and S41 of the invention [1] have deteriorated by about 10 dB. In addition, the phase non-flatness of the invention [1] has also deteriorated to 90 ± 1.5°, while that of the conventional directional coupler is approximately 90 ± 0.5°.

[0007] The structure of the invention application (publication number: CN116345104A) operates in the W band and also improves the traditional waveguide orientation. The amplitude non-flatness is better than 0.2dB, but the S11 also deteriorates to -15dB and the phase non-flatness deteriorates to 90±3°.

[0008] The amplitude non-flatness of the traditional terahertz waveguide directional coupler at two ports (2, 3) is poor, making it difficult to achieve both broadband and low amplitude non-flatness. Invention [1] (publication number: CN114725644A) proposed a new structure. Near the 192GHz frequency band, although the amplitude non-flatness was significantly improved and better than the traditional waveguide directional coupler, other S parameters such as S11 and S41 deteriorated by about 10dB, and the phase non-flatness also increased. Another invention application (publication number: CN116345104A) works in the W band and also improves the amplitude non-flatness of the traditional waveguide directional coupler, but also causes the S11 parameter to deteriorate to -15dB and the phase non-flatness to increase to 90±3°. Although these improvements have made progress in amplitude non-flatness, other performance parameters such as S11 and phase non-flatness have been greatly degraded. Summary of the Invention

[0009] The present invention aims to overcome the shortcomings of the prior art and provide a low-amplitude-non-flatness waveguide directional coupler and its implementation method. The invention adopts a multi-step ladder and wide waveguide structure to reduce the amplitude non-flatness of the waveguide directional coupler. At a frequency band of around 192 GHz and with the same operating bandwidth, the amplitude non-flatness is superior to that of conventional waveguide directional couplers.

[0010] The object of the present invention is achieved through the following technical solutions:

[0011] A method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness comprises the following steps:

[0012] A1 selects a standard rectangular waveguide as the original matrix, and widens its wide side a to wide side w in the horizontal direction, forming two layers of wide rectangular waveguides;

[0013] A2 inserts N rectangular branch nodes periodically along the z direction between the common narrow walls of the two layers of wide waveguides. The rectangular branch nodes include wide branch nodes and narrow branch nodes, where the thickness of the wide branch node is , the thickness of the narrow branch node is , the two share the same height c, and the width of all branch nodes is equal to the waveguide width w; the set {w,c, , }Set as optimization variable group;

[0014] A3 constructs M-level step gradient sections at both ends of the wide waveguide. The width of each step decreases continuously. The length of each step is x, and the total length is X.

[0015] A4 For the variable group {w,c, , ,x,X} for collaborative optimization;

[0016] A5 uses oxygen-free copper CNC milling to process the entire optimized structure. The isolated end is used to install the matching load to form a four-port terahertz waveguide directional coupler.

[0017] Terahertz waveguide directional couplers are based on the propagation characteristics of electromagnetic waves in waveguide structures. When terahertz waves pass through the waveguide, the geometric structure and material properties of the waveguide will affect the propagation path and energy distribution of the wave. Directional couplers use these principles to achieve wave splitting and directional transmission. By selecting a standard rectangular waveguide and laterally widening its wide edge to form two layers of upper and lower wide rectangular waveguides, this design allows electromagnetic waves to propagate in a specific mode in the waveguide. Rectangular branch nodes are periodically inserted between the two layers of wide waveguides. These branch nodes include wide branch nodes and narrow branch nodes. The thickness and height of these branch nodes are optimized to achieve a specific coupling effect. The size difference between the wide branch nodes and the narrow branch nodes results in different coupling efficiencies for electromagnetic waves of different frequencies, thereby achieving frequency selectivity.

[0018] Stepped transitions are constructed at both ends of the wide waveguide. These gradually change the waveguide width through a continuously decreasing width, helping to achieve a smooth wave transition and reduce reflections and scattering, thereby improving coupler performance. Parameters such as the waveguide width, branch section size, and step-step transition length are collaboratively optimized to ensure good directional coupling performance within the specific terahertz frequency band.

[0019] The optimized waveguide directional coupler structure is fabricated using oxygen-free copper CNC milling, with matching loads installed at the isolation end to reduce reflections. The resulting four-port terahertz waveguide directional coupler is capable of splitting the input signal into two parts and directing them to different output ports while maintaining signal directionality and low loss.

[0020] As a preferred embodiment, the original matrix is ​​a WR5.1 (UG-387) standard rectangular waveguide, where a=1.295mm, b=0.648mm, and the width ratio w / a is between 1.35 and 1.45 to stimulate high-order modes and suppress undesired modes.

[0021] As a preferred embodiment, the thickness and height of the rectangular branch section satisfy 0.2≤ / c≤0.6 and 0.5≤ / c≤1.5 to strike a balance between coupling strength and processing feasibility.

[0022] As a preferred embodiment, the size of the rectangular branch section is c=0.267mm. =0.146mm, =0.39mm.

[0023] As a preferred embodiment, the number of steps M=10, the length of each step x=0.5 mm, and the total length X=5 mm.

[0024] As a preferred embodiment, the step gradient section is only reduced in the width direction, and its reduction rule adopts an inverse linear gradient, so that the waveguide width transitions step by step from w to a.

[0025] As an optimal method, sensitivity analysis is introduced in the optimization process to analyze the variable group {w, c, , ,x,X} is expanded by the first order Taylor to determine the key parameters that have the greatest impact on the amplitude unevenness |S21-S31|.

[0026] As a preferred embodiment, the stepped gradient section implements an E-plane stepped transition on the wide side, so that the wide side gradually shrinks from w to the wide side a of the standard rectangular waveguide, while the narrow side b remains unchanged; a circular arc transition with a radius r = 0.2mm is provided at the junction of the step side and the standard waveguide to reduce discontinuity and suppress high-order modes and reflections.

[0027] A terahertz waveguide directional coupler with low amplitude non-flatness, comprising:

[0028] The upper wide rectangular waveguide and the lower wide rectangular waveguide have a wide side of w and a narrow side of b;

[0029] The coupled branch node array is composed of N rectangular copper columns, each of which includes a wide branch node with a thickness of , narrow branch section, narrow branch section thickness , common height c;

[0030] M-level stepped gradient sections are set at both ends of the wide waveguide, with each level length x and a total length of X.

[0031] As a preferred embodiment, the wide rectangular waveguide has a wide side w=1.8 mm, a narrow side b=0.648 mm, a stepped gradient section number M=10, a length of each section x=0.5 mm, and a total length X=5 mm.

[0032] The present invention has at least the following beneficial effects: the terahertz waveguide directional coupler with low amplitude non-flatness and the preparation method thereof provided by the present invention have significant technical advantages over the existing technology. By introducing a wide waveguide structure and a multi-step gradient section, and combining the design of a coupling branch section, the amplitude consistency of the coupler in the 185GHz-200GHz frequency band is effectively improved. While maintaining excellent isolation and wide-band characteristics, the amplitude non-flatness is significantly reduced, which is better than the existing structure. The present invention not only solves the problem of large amplitude non-flatness of traditional waveguide directional couplers under broadband conditions, but also provides a high-performance core component for high-precision power allocation / synthesis, signal detection and system integration in the terahertz frequency band, and has good engineering application prospects and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] To reveal the technical details of the embodiments of the present invention, the following is a brief introduction to the drawings involved in the embodiments. It should be emphasized that these drawings only illustrate several embodiments of the present invention and should not be considered as defining the scope of the invention. Those skilled in the art can deduce other relevant drawings based on these drawings without engaging in creative work.

[0034] Figure 1 Schematic diagram of the structure of a traditional waveguide directional coupler;

[0035] Figure 2 This is a schematic diagram of the structure of the terahertz waveguide directional coupler with low amplitude non-flatness proposed by the present invention;

[0036] Figure 3 The structural parameters of the terahertz waveguide directional coupler with low amplitude non-flatness proposed by the present invention;

[0037] Figure 4 A comparison diagram of amplitude non-flatness simulation results of a conventional waveguide directional coupler and the waveguide directional coupler of the present invention;

[0038] Figure 51. A comparison chart of S11 simulation results of a conventional waveguide directional coupler and the waveguide directional coupler of the present invention;

[0039] Figure 6 A comparison chart of S41 simulation results of a conventional waveguide directional coupler and the waveguide directional coupler of the present invention;

[0040] Figure 7 A comparison diagram of phase non-flatness simulation results of a conventional waveguide directional coupler and the waveguide directional coupler of the present invention;

[0041] Figure 8 A simulation interface for the waveguide directional coupler of the present invention;

[0042] Figure 9 This is the electric field diagram of the waveguide directional coupler of the present invention;

[0043] In the figure, 1-input end, 2-output end, 3-coupling end, 4-isolation end. DETAILED DESCRIPTION

[0044] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0045] In the following, embodiments of the present disclosure are described in detail with the aid of accompanying drawings. However, please be aware that the present disclosure is not limited to the specific forms shown herein. Rather, it should be understood to encompass various variations, equivalents, and / or alternatives to the embodiments of the present disclosure. In describing the drawings, the same reference numerals will be used to indicate similar components.

[0046] In this disclosure, terms are used to illustrate specific embodiments and do not constitute limitations of this disclosure. In this context, the use of the singular also encompasses the plural, unless the text clearly indicates otherwise. In the process of explanation, it should be understood that terms such as "including" or "having" are intended to indicate the presence of a feature, quantity, step, operation, structural component, part, or combination thereof, and do not preclude the possibility or addition of one or more other features, quantities, steps, operations, structural components, parts, or combinations thereof.

[0047] It should be understood that while the following description provides extensive specific details intended to facilitate a comprehensive understanding of the example embodiments, those skilled in the art will appreciate that the example embodiments can be implemented without these specific details. For example, systems may be presented in block diagram form to avoid excessive detail that would obscure the clarity of the examples. In other cases, unnecessary details regarding well-known processes, structures, and techniques may be omitted to maintain clarity of the examples.

[0048] like Figure 2 As shown, a method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness is provided. The coupler includes an input terminal 1, an output terminal 2, a coupling terminal 3, and an isolation terminal 4. Figure 1 As shown, the following steps are included:

[0049] A1 selects a standard rectangular waveguide as the original matrix, and widens its wide side a to wide side w in the horizontal direction, forming two layers of wide rectangular waveguides;

[0050] A2 inserts N rectangular branch nodes periodically along the z direction between the common narrow walls of the two layers of wide waveguides. The rectangular branch nodes include wide branch nodes and narrow branch nodes, where the thickness of the wide branch node is , the thickness of the narrow branch node is , the two share the same height c, and the width of all branch nodes is equal to the waveguide width w; the set {w,c, , }Set as optimization variable group;

[0051] A3 constructs M-level step gradient sections at both ends of the wide waveguide. The width of each step decreases continuously. The length of each step is x, and the total length is X.

[0052] A4 For the variable group {w,c, , ,x,X} for collaborative optimization;

[0053] The optimized structure of A5 is processed by oxygen-free copper CNC milling. The isolation end 4 is used to install the matching load to form a four-port terahertz waveguide directional coupler.

[0054] In this embodiment, the design parameters {w, c, , ,x,X}, successfully developed a terahertz waveguide directional coupler with low amplitude non-flatness, which not only greatly improved the performance of the waveguide and the stability of signal transmission, but also effectively controlled the electromagnetic field distribution inside the waveguide by periodically inserting rectangular branch nodes between two layers of wide waveguides, achieving precise power distribution and signal coupling. In addition, the design of the stepped gradient section significantly reduced the reflection and standing waves at the waveguide port, improving the transmission efficiency and signal quality of the entire waveguide system. The use of oxygen-free copper material for CNC milling ensures the high precision and good conductivity of the waveguide, further improving the overall performance of the waveguide directional coupler. At the same time, the coupler achieves broadband matching of the waveguide by constructing stepped gradient sections at both ends of the wide waveguide, allowing the coupler to maintain good working performance over a wide frequency range. Finally, the optimized waveguide directional coupler has a compact structure, is easy to integrate and install, and is very suitable for a variety of applications in the terahertz frequency band.

[0055] Preferably, the width of each step adopts the "inverse linear gradient" rule, so that the waveguide characteristic impedance (≈ the inverse of the width) changes linearly along the axial direction, that is, the width of the kth step is:

[0056] ,in, is the wide side width of the k-th level section (unit: mm), is the starting broadside width (in ), is the ending broadside width (at ), is the total number of steps ( , integer), It is the current step number, ranging from 0 to M, with a total of M+1 section positions.

[0057] The reciprocal linear gradient makes the characteristic impedance of the waveguide (≈ the reciprocal of the wide side) change linearly along the axial direction, with the shortest impedance transition path. The reflection coefficient can be suppressed to below -30dB within the broadband, and the adjacent level differences are continuous and without mutations, taking into account both low loss and easy processing.

[0058] Level 10 case (M=10):

[0059] =1.800mm, =1.295mm, M=10

[0060] calculate:

[0061] 1 / =0.555556 , 1 / =0.772201 , Δ=0.216645

[0062] Table 1 and Correspondence table

[0063]

[0064] Linear gradient in reciprocal space, at the beginning ( When the value is smaller), the width decreases faster; as As the value increases, the change in each step gradually decreases; therefore, the adjacent differences form a decreasing sequence. The adjacent step differences are uniform, with no sudden changes throughout the entire process, and are suitable for the 185-200GHz terahertz waveguide gradient section.

[0065] Reciprocal linear gradient in When the width is small (the initial section), the drop in width is the largest, and the step difference gradually decreases with each subsequent step, forming a decreasing sequence of adjacent step differences. The large width change in the broadband, low-reflection initial section quickly completes the main impedance transition, canceling the reflected energy "in one go" over a very short distance. Subsequent small-step fine-tuning further refines the matching, reducing the reflection coefficient to below -30dB across the entire 185-200GHz band. The rapid change in the initial section suppresses high-order modes, equivalent to a "slowly varying taper," quickly pushing the cutoff frequency of high-order modes out of the operating band and reducing parasitic resonances. Subsequent sections remain virtually unchanged, preventing the excitation of new high-order fields and ensuring single-mode transmission.

[0066] This embodiment offers a friendly machining tolerance, with a large step difference at the front end, making it easy for the CNC to distinguish; and a small step difference at the end, perfectly within the machine tool's minimum feed rate, preventing "cutting" and loss of precision due to excessive detail. Short overall length and low loss: Because most impedance transitions are concentrated in the first 2-3 steps, the overall length can be controlled to 5mm (10 steps x 0.5mm), with an insertion loss of <0.05dB, far superior to traditional arithmetic or geometric gradients. In short, the "fast at the front, slow at the back" decreasing step difference allows for rapid energy matching at the front end and precise finishing at the back end, achieving the three key performance characteristics of broadband, low reflection, and easy machining.

[0067] Preferably, the variable set {w, c, , ,x,X} for collaborative optimization.

[0068] In a preferred embodiment of the present invention, a collaborative optimization process for devices in the 185 GHz to 200 GHz frequency band is proposed to improve their amplitude consistency, input matching performance, and isolation performance. The process includes the following steps:

[0069] First, a multi-objective optimization function is set in the A4-1 stage. Based on the S parameters obtained from the full-wave finite element simulation as a benchmark, the following three objective functions are established within the frequency band of interest: =max(|S21–S31|), which represents the unevenness of the transmission amplitude between output ports; =max(|S11|), which represents the return loss of the input port; =max(|S41|), which represents the signal leakage level between the isolated port and the signal source. The optimization goal is to minimize the above three indicators, and set the constraints separately: ≤0.1dB, ≤–30dB, ≤–20dB.

[0070] In the A4-2 stage, the decision variables and their value intervals are defined to form the variable group {w, c, , ,x,X}. Where w represents the waveguide width, and its value range is [1.70mm,1.90mm]; c represents the coupling column height, and its value range is [0.20mm,0.35mm]; is the thickness of the narrow branch node, and its value range is [0.10mm, 0.20mm]; is the thickness of the wide branch node, ranging from [0.30 mm, 0.50 mm]; x is the length of a single step, ranging from [0.40 mm, 0.60 mm]; X is the total length, defined as M times x, where M is an integer and ranges from [6, 14].

[0071] Subsequently, in the A4-3 phase, a three-dimensional full-wave electromagnetic model was constructed. This was performed using HFSS (High Frequency Structure Simulator) electromagnetic simulation software, with all structural parameters controlled parametrically. To ensure simulation accuracy, the model mesh was refined, with a minimum mesh size of less than λ / 10. A swept frequency calculation covered the entire target frequency band, with 201 frequency points set. The average simulation time for a single point was approximately 2.5 minutes.

[0072] In stage A4-4, an evolutionary algorithm was deployed to perform multi-objective optimization tasks. The Non-Dominated Sorting Genetic Algorithm II (NSGA-II) was selected and implemented in Matlab. Simulations were performed using Matlab and HFSS, with parallel computing performed on a 24-core server. A total of 2,400 simulations were completed, optimizing the parameters corresponding to each simulation result.

[0073] Entering the A4-5 stage, the compromise solution is selected from the obtained simulation results as the optimal design solution. The final parameter values ​​are determined as follows: w*=1.800mm, c*=0.267mm, *=0.146mm, *=0.390mm, x*=0.500mm, M*=10, so the total length X*=5.000mm. The parameters marked with * in the manual are fixed parameters.

[0074] In the A4-6 phase, the performance of the optimal solution was verified. Substituting the determined parameters into the original full-wave model for re-simulation, the results showed that the maximum difference between |S21–S31| was 0.08dB, the maximum value of |S11| was –32dB, and the maximum value of |S41| was –23dB. All indicators met the preset optimization goals.

[0075] Finally, during the A4-7 phase, engineering feasibility was confirmed. Based on the optimized dimensions, a prototype device was manufactured using oxygen-free copper through CNC precision machining, with a tolerance of ±5μm. A 3μm-thick electroless silver plating was applied to the surface to enhance conductivity. Field measurements showed that the deviation between simulated and measured data did not exceed 0.02dB, fully verifying the effectiveness and engineering practicality of the optimization process.

[0076] In a preferred embodiment, the original matrix is ​​a WR5.1 (UG-387) standard rectangular waveguide, wherein a=1.295mm, b=0.648mm, the wide side w=1.8mm after widening, and the width ratio w / a is between 1.35 and 1.45 to stimulate high-order mode and suppress the undesired mode. When the original waveguide width a=1.295mm for WR5.1 (UG-387), The cut-off frequency is about 116GHz. About 231GHz. After widening the wide side to w=1.8mm: The cutoff frequency drops to about 166 GHz, which is below the 185–200 GHz operating band. The mode is excited. The width ratio w / a≈1.39, between 1.35–1.45, is high enough to make and The phase constant difference increases, the coupling efficiency improves; if the value is not too large, The cutoff is still above 250GHz, The higher order modes are still cut off in the whole working band and thus suppressed. In short, in the range of 185–200 GHz, there are only and controllable The existence of this mode not only ensures broadband coupling but also suppresses spurious modes.

[0077] 、 、 The type of transmission mode in a rectangular waveguide, specifically the transverse electric (TE) mode. Electromagnetic waves propagating in a waveguide can have different modes, which describe the distribution of the electric and magnetic fields across the waveguide's cross section.

[0078] Mode: This is the most commonly used primary mode. In this mode, the electric field varies in only one dimension (usually the broadside direction) while remaining constant in the other (the narrowside direction). It has the lowest cutoff frequency and is therefore often used in practical applications.

[0079] Mode: The electric field of this mode has two and a half cycles of changes in the broadside direction. mode, which has a higher cutoff frequency.

[0080] Mode: The electric field of this mode changes in the direction of the wide side for three and a half cycles, and its cutoff frequency is The model is even higher.

[0081] In a preferred embodiment, the thickness and height of the rectangular branch section satisfy 0.2≤ / c≤0.6 and 0.5≤ / c≤1.5 to strike a balance between coupling strength and processing feasibility.

[0082] The thickness of the rectangular branch section 、 Together with the height c, the coupling column determines the equivalent susceptance at the common narrow wall of the upper and lower wide rectangular waveguides. In the terahertz frequency band, the coupling column can be approximated as a capacitor-inductor composite unit connected in parallel between the upper and lower wide rectangular waveguides. Its susceptance is proportional to the product of the column's cross-sectional area and height, and inversely proportional to the waveguide's equivalent impedance. Therefore, the greater the column height c, the longer the column's "effective length" in the electric field direction, the greater the susceptance, and the stronger the coupling; the thicker the column, the greater the susceptance. 、 The larger it is, the larger the cross-sectional area of ​​the column is, which also increases the susceptance.

[0083] In order to meet the dual requirements of "sufficient coupling" and "machinability", the present invention sets the following boundaries for the thickness-height ratio: 0.2≤ / c≤0.6 and 0.5≤ / c≤1.5. This limit defines the range of the coupling column's "width-to-thickness ratio": the lower limit ensures the coupling column still has sufficient cross-sectional area to provide the necessary coupling power; the upper limit prevents the column from being too thick, which could introduce high-order modes or excessive reflections, while also preventing the thickness from approaching the minimum tool limit of the machining machine, ensuring reliable manufacturing within the CNC tolerance of ±5µm.

[0084] By using the above proportional constraints, the coupling column maintains a monotonic and controllable coupling change throughout the entire operating frequency band. At the low end of the band, the column can still provide sufficient coupling to make the power division error less than 0.1dB. At the high end of the band, the column will not excessively perturb the main mode, suppressing and above undesirable modes, thereby maintaining the amplitude non-flatness, return loss and isolation standards.

[0085] In a preferred embodiment, the preferred size of the rectangular branch section is c=0.267 mm. =0.146mm, =0.39mm.

[0086] In a preferred embodiment, the number of steps M=10, the length of each step x=0.5 mm, and the total length X=5 mm.

[0087] In a preferred embodiment, the step gradient section is only decreased in the width direction, and its decreasing rule adopts an inverse linear gradient, so that the waveguide wide side gradually transitions from w to a; this rule makes the waveguide characteristic impedance change linearly in the inverse space, thereby achieving an equal ripple reflection coefficient in the entire 185-200GHz band.

[0088] In one embodiment, the directional coupler structure parameters (see Figure 3 ) are shown in the following table, where the number of steps is 10.

[0089] The dimensions a and b are based on the WR5.1 (UG-387) standard waveguide model. The remaining parameters are the preferred parameters at this frequency and are not limited to these dimensions. The dimensions can be changed according to different application requirements. The actual product can be processed using CNC technology. The processing tolerance in this example is ±0.005mm. The preferred parameters are as follows: a=1.295mm, b=0.648mm, c=0.267mm, =0.146mm, =0.39mm,X=5mm,w=1.8mm,l=7.5mm.

[0090] In a preferred embodiment, sensitivity analysis is introduced in the optimization process to analyze the variable set {w, c, , ,x,X} is subjected to a first-order Taylor expansion to determine the key parameters that have the greatest impact on the amplitude non-flatness |S21–S31|. The amplitude non-flatness |S21-S31| refers to the absolute value of the difference in transmission coefficients between the input terminal 1 to the output terminal 2 (S21) and the transmission coefficients from the input terminal 1 to the coupling terminal 3 (S31) of the terahertz waveguide directional coupler within a specific frequency band, which is used to measure the amplitude consistency between the two ports. Lower amplitude non-flatness indicates higher amplitude consistency and better signal distribution performance. This embodiment proposes a collaborative optimization method based on sensitivity analysis combined with a multi-objective genetic algorithm to improve the amplitude flatness, input matching performance and isolation of devices in the 185GHz to 200GHz frequency band. The method includes the following two key stages:

[0091] First, perform variable sensitivity analysis in the A4-a stage. Before formally entering the global optimization, the decision variable group {w, c, , ,x,X} are normalized, and the amplitude non-flatness ΔA = |S21–S31| at the center frequency of 192 GHz is used as the objective function to perform the first-order Taylor expansion.

[0092] The partial derivatives of each variable are calculated by finite difference method, and are sorted according to their absolute values, so as to quantify the sensitivity of each variable to the amplitude flatness. The analysis results show that: The absolute value of w is the largest, which indicates that the wide side w and the wide branch thickness d2 are the main factors affecting the amplitude flatness; The absolute value of w is the largest, which indicates that the wide side w and the wide branch thickness d2 are the main factors affecting the amplitude flatness; The sensitivity of the remaining variables is one order of magnitude lower, and only serves as a fine tuning term.

[0093] This sensitivity analysis step only requires a small amount of gradient calculation, which can effectively reduce the search space and divide the original 6-dimensional variables into 3 high-weight main variables (w, , c) and 3 low-weight fine tuning variables (x, , X). This step serves as a "pre-processing" process before NSGA-II optimization, providing prior information for subsequent optimization and improving search efficiency.

[0094] Subsequently, multi-objective global optimization based on NSGA-II is carried out in the A4-b stage. According to the sensitivity analysis results, the crossover and mutation probabilities of each variable in the genetic algorithm are reconfigured, the high-weight main variables (w, , c) are given higher search weights, and the low-weight fine tuning variables (x, , X) are appropriately reduced in weight to reduce invalid iterations. In the optimization process, the Non-dominated Sorting Genetic Algorithm II (NSGA-II) is used, combined with full-wave finite element simulation data, using HFSS and Matlab joint simulation, running in parallel on a 24-core server, a total of 2400 simulation tasks are completed, and the simulation results converge.

[0095] The above two-stage optimization method constitutes a progressive process of "screening before searching, pre-processing + main optimization". The first stage locks the key variables through fast sensitivity analysis, significantly reducing the search dimension; the second stage introduces variable weight mechanism in the genetic algorithm, improving the convergence speed and stability. The synergistic effect of the two not only retains the global optimization ability, but also avoids the waste of resources caused by blind search, achieving efficient and stable multi-objective optimization.

[0096] In a preferred embodiment,

[0097] The stepped transition section implements E-plane stepped transition on the wide side, so that the wide side is gradually reduced to the standard rectangular waveguide wide side a in steps, while the narrow side b remains unchanged; a radius r=0.2mm arc transition is provided at the junction of the stepped side surface and the standard waveguide to reduce discontinuity and suppress high-order modes and reflections.

[0098] In a preferred embodiment, the CNC milling tolerance of the integral copper material is ±0.005 mm, and after milling, the inner wall of the coupling cavity is subjected to chemical silver plating, with a silver layer thickness of ≥3 μm and a surface roughness Ra of <0.4 μm.

[0099] In a preferred embodiment, the matching load is a conical absorber, and its front end cone angle θ satisfies tanθ=Δw / L, where Δw is the maximum width difference of the step gradient section, L=5mm, to ensure that the reflected power of the isolation end 4 is lower than -30dB.

[0100] A terahertz waveguide directional coupler with low amplitude non-flatness, comprising:

[0101] The upper wide rectangular waveguide and the lower wide rectangular waveguide have a wide side of w and a narrow side of b;

[0102] The coupled branch node array is composed of N rectangular copper columns, each of which includes a wide branch node with a thickness of , narrow branch section, narrow branch section thickness , common height c;

[0103] M-level stepped gradient sections are set at both ends of the wide waveguide, with each level length x and a total length of X;

[0104] Matching load, installed at the isolation terminal 4;

[0105] Calibration interface, located at each port, for connecting to a vector network analyzer.

[0106] This invention provides a terahertz waveguide directional coupler with low amplitude non-flatness, belonging to the field of terahertz communication and microwave device technology. The coupler comprises two upper and lower wide rectangular waveguides, each with a wide side dimension w and a narrow side dimension b, for directional coupling between the primary and secondary waveguides. Between the two waveguides lies an array of coupling branches composed of N rectangular copper pillars, each containing a wide branch and a narrow branch, to control the coupling strength and bandwidth.

[0107] Furthermore, the coupler features a stepped gradient section for gradual waveguide impedance matching, reducing reflection loss and extending the operating bandwidth. A matched load is connected to the isolation port 4 to absorb uncoupled signals and improve port isolation. Furthermore, each port features a calibration port for easy connection to a vector network analyzer, enabling high-precision measurements and system calibration.

[0108] The terahertz waveguide directional coupler with low amplitude non-flatness achieves the comprehensive performance of high directivity, wide bandwidth, low loss, high isolation and good stability through the coordinated design of coupling branch node array, step gradient section and column structure (the entire coupler is a copper structure), and has broad application prospects and practical value.

[0109] In a preferred embodiment, the wide rectangular waveguide has a wide side w=1.8 mm, a narrow side b=0.648 mm, a stepped gradient section number M=10, a length of each section x=0.5 mm, and a total length X=5 mm.

[0110] In a preferred embodiment, the system comprises a symmetrical four-port waveguide device that can be expanded into a 2×2 or 4×4 power distribution network through modular stacking for use in terahertz imaging or communication front-ends. This modular design allows for easy expansion into 2×2 or 4×4 power distribution networks through stacking, enhancing the system's flexibility and scalability. Suitable for terahertz imaging or communication front-ends, this device can meet the demands of high-frequency, high-power applications, improving system performance and application range. The symmetry of the overall structure ensures uniform signal distribution and transmission, reduces signal loss and interference, and improves system stability and reliability.

[0111] Analyze various directional coupler indicators. Note: The S11 level represents the return loss of the directional coupler. The smaller the S11, the better. For example, an S11 of -20dB is better than an S11 of -15dB. Amplitude flatness is the absolute value of the difference between S21 and S31. A smaller value indicates better flatness. The S41 level represents the isolation of the directional coupler. Similarly, the smaller the S41, the better. For example, an S41 of -20dB is better than an S41 of -15dB.

[0112] The present invention reduces the amplitude unevenness of the waveguide directional coupler. Figure 4 This is a comparison chart of the amplitude non-flatness simulation results of a traditional waveguide directional coupler and the waveguide directional coupler of the present invention. The simulation results show that at the center frequency, the amplitude non-flatness of the traditional waveguide directional coupler is about 0.4dB, which means that there is a 0.4dB difference in the power values ​​of the output signals of the two ports; the structure proposed by the present invention has an amplitude non-flatness of less than 0.1dB at the center frequency, the amplitude non-flatness is reduced by 75%, and the overall in-band amplitude non-flatness level is reduced.

[0113] S11 represents the level of device return loss. The smaller the S11, the lower the signal reflection of the device and the higher the transmission efficiency. When S11 is -20dB, it means that only 1% of the signal is reflected. Figure 5 This is a comparison chart of the S11 simulation results of a traditional waveguide directional coupler and the waveguide directional coupler of the present invention. The S11 level of the traditional waveguide directional coupler is better than -25dB. The structure proposed in the present invention has an S11 level better than -30dB while significantly reducing the amplitude unevenness, and has better performance. Although the existing technology has taken certain measures to reduce the amplitude unevenness, its S11 level has deteriorated to -15dB. Compared with the comparative documents (CN114725644A and CN116345104A), the structure proposed in the present invention has been optimized by 15dB.

[0114] S41 represents the isolation of the device. Figure 6 This is a comparison chart of the S41 simulation results of a traditional waveguide directional coupler and the waveguide directional coupler of the present invention. The S41 level of the traditional waveguide directional coupler is better than -25dB. The structure proposed in the present invention maintains an S41 level better than -20dB while significantly reducing the amplitude flatness. The S41 level in the comparative document (CN114725644A) deteriorated to -15dB, and the structure proposed in the present invention is optimized by 5dB compared with the comparative document.

[0115] Figure 7 This is a comparison chart of the phase non-flatness simulation results of a traditional waveguide directional coupler and the waveguide directional coupler of the present invention. The phase non-flatness of the traditional waveguide directional coupler is about 90±0.5°, while the phase non-flatness of the structure proposed in the present invention is also 90±0.5°, showing no deterioration. The comparative document (CN114725644A) deteriorates to 90±1.5°, and the comparative document (CN116345104A) deteriorates to 90±3°.

[0116] Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as covering the preferred embodiments and all changes and modifications that fall within the scope of the invention. The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness, characterized in that: The following steps are involved: A1: Select a standard rectangular waveguide as the original matrix, and extend its wide side a to wide side w in the horizontal direction to form two layers of wide rectangular waveguides; A2: N rectangular branch nodes are periodically inserted along the z direction between the common narrow walls of the two layers of wide waveguides. The rectangular branch nodes include wide branch nodes and narrow branch nodes. The thickness of the wide branch node is , the thickness of the narrow branch node is , the two share the same height c, and the width of all branch nodes is equal to the waveguide width w; the set {w,c, , }Set as optimization variable group; A3: Construct M steps of gradient sections at both ends of the wide waveguide. The width of each step decreases continuously. The length of each step is x, and the total length is X. The gradient section decreases only in the width direction, and the decreasing rule adopts an inverse linear gradient, so that the width of the waveguide gradually transitions from w to a. The width of each step adopts an inverse linear gradient rule, so that the characteristic impedance of the waveguide changes linearly along the axial direction, that is, the width of the kth step is: ;in, is the broadside width of the k-th level section, is the starting broadside width, is the ending broadside width, is the total number of steps, is the current step number, ranging from 0 to M, with a total of M+1 section positions; A4: For the variable group {w,c, , ,x,X} for collaborative optimization; sensitivity analysis is introduced in the optimization process to analyze the variable group {w,c, , ,x,X} is expanded by the first order Taylor to determine the amplitude non-flatness The most influential key parameters; A5: The optimized structure is processed by oxygen-free copper CNC milling. The isolated end is used to install the matching load to form a four-port terahertz waveguide directional coupler.

2. The method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness according to claim 1, characterized in that: The original matrix is ​​a WR5.1 / UG-387 standard rectangular waveguide, wherein a=1.295 mm, b=0.648 mm, and the width ratio w / a is between 1.35 and 1.

45.

3. The method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness according to claim 1 or 2, characterized in that: The thickness and height of the rectangular branch section meet 0.2≤ / c≤0.6 and 0.5≤ / c≤1.

5.

4. The method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness according to claim 3, characterized in that: The size of the rectangular branch section is c=0.267mm, =0.146mm, =0.39mm.

5. The method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness according to claim 1, characterized in that: The number of steps M=10, the length of each step x=0.5 mm, and the total length X=5 mm.

6. The method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness according to claim 1, characterized in that: The step gradient section implements an E-plane step transition on the wide side, so that the wide side gradually shrinks from w to the wide side a of the standard rectangular waveguide, while the narrow side b remains unchanged; a circular arc transition with a radius r=0.2mm is provided at the junction of the step side and the standard waveguide.

7. A terahertz waveguide directional coupler with low amplitude non-flatness, wherein the terahertz waveguide directional coupler with low amplitude non-flatness is obtained by using the method for realizing a terahertz waveguide directional coupler with low amplitude non-flatness according to claim 1, characterized in that: include: The upper wide rectangular waveguide and the lower wide rectangular waveguide have a wide side of w and a narrow side of b; The coupled branch node array is composed of N rectangular copper columns, each of which includes a wide branch node with a thickness of , narrow branch section, narrow branch section thickness , common height c; M-level stepped gradient sections are set at both ends of the wide waveguide, with each level length x and a total length of X.

8. The low amplitude non-flatness terahertz waveguide directional coupler according to claim 7, characterized in that: The wide rectangular waveguide has a wide side w=1.8 mm and a narrow side b=0.648 mm; The number of step gradient stages M=10, the length of each stage x=0.5mm, and the total length X=5mm.

Images