A printed circuit quality monitoring device and method based on dual-wavelength synchronous transmission and reflection detection

Abstract

The application provides a printed circuit quality monitoring device and method based on dual-wavelength synchronous transmission and reflection detection, which realizes real-time online monitoring of the conductivity, size precision and geometric deformation of the printed circuit. In the production process of the printed circuit, a metamaterial printed quality detection pattern with specific reflection and transmission performance is introduced, and the transmission characteristics of the pattern to terahertz waves are monitored to comprehensively evaluate key functional parameters such as the conductivity, thickness and diffusion characteristics of the printed ink. The system detects the production deviation in real time and automatically adjusts the production parameters, optimizes the production conditions, thereby significantly reduces the output of unqualified products, improves the operation efficiency and economy of the production line, reduces resource waste and lowers the production cost.

Description

Technical Field

[0001] This invention relates to the field of quality monitoring in the production process of printed electronics, and more specifically, to a printed circuit quality monitoring device and detection method based on dual-wavelength synchronous transmittance and reflectance detection. Background Technology

[0002] Printed electronics is a cutting-edge technology that applies traditional printing processes to the manufacture of electronic components and devices. It enables the direct printing of electronic devices onto various flexible substrates, achieving low-cost, high-efficiency production of electronic products. This technology uses multiple methods, including flexographic printing, inkjet printing, gravure printing, and screen printing, to directly print photonic devices and electronic products onto various substrates such as fabrics, paper, and plastics. This technology has been widely applied in wearable devices, the sports industry, and flexible displays. Due to the fast production pace, the quality of the printing process can vary significantly.

[0003] The entire process of printed circuit board manufacturing generally consists of four main stages: (a) ink and substrate selection, (b) printing, (c) sintering, and (d) print quality inspection. The first step is selecting suitable ink and substrate. Among the various available substrates, flexible substrates are particularly favored due to their unique flexibility, stretchability, and torsionability while maintaining good electrical properties. A variety of conductive, semi-conductive, and dielectric inks have been developed to meet different printing needs, with conductive inks receiving particular attention. These include three main categories: metallic inks, carbon-based inks, and composite inks. Metallic inks, with gold, silver, copper, and other metal particles as the conductive core, are widely used in commercial applications due to their superior conductivity, which can improve signal transmission efficiency, reduce latency, and decrease energy consumption in electronic circuits.

[0004] Once the substrate and ink are determined, the next step is printing. Printing technologies are selected based on specific needs, including flexographic printing, gravure printing, screen printing, and inkjet printing, all of which are widely used in the market. In inkjet printing, functional inks are prepared by uniformly dispersing functional materials (such as metal nanoparticles, conductive polymers, or carbon nanotubes) in a suitable solvent to form a stable dispersion system. The uniformity of the ink directly determines the quality and performance of the printed electronic circuits.

[0005] In addition, in the process of printed electronics manufacturing, the quality of ink directly determines the performance indicators of the final product, including power consumption, response speed and overall electrical performance. Therefore, the quality of ink must be strictly controlled during the manufacturing process to ensure the production of efficient and reliable electronic products.

[0006] Throughout the production process, printing quality inspection is extremely critical, especially in a fast-paced production environment where quality can vary significantly during printing. This presents a major challenge for the printed electronics industry. Unlike the traditional graphic printing industry, where quality control primarily focuses on color, resolution, registration, and surface finish, printed electronics manufacturing must not only ensure the stability of the process itself but also guarantee the functionality of the devices.

[0007] However, achieving real-time, online process feedback in mass production currently faces many challenges, yet it is crucial for industrial manufacturing. Manufacturers often face high costs when the production process goes out of control. Problems are often only discovered after the entire roll of product has been produced, and evaluating the functionality of devices or materials on the substrate through off-line testing can lead to a large amount of material being scrapped due to defects, resulting in significant economic losses. To achieve the economic benefits of mass production, scrap rates must be kept low, requiring real-time, non-destructive online quality control during production. In printed circuit board (PCB) production, functional and aesthetic inspections are indispensable. Currently, machine vision-based inspection methods are mainly used to identify aesthetic defects. However, electrical performance testing, as one of the key indicators of PCBs, has traditionally relied on contact methods such as multimeters or four-probe systems. These methods are not only difficult to integrate with automated production but can also sometimes damage devices. Furthermore, contact testing methods generally have poor accuracy and repeatability; therefore, finding a non-contact, real-time measurement solution is particularly important. Summary of the Invention

[0008] The first objective of this invention is to provide a printed circuit quality monitoring device based on dual-wavelength synchronous transmission and reflection detection, comprising a computer, a terahertz receiver 1, a terahertz receiver 2, a terahertz transmitter 1, a terahertz transmitter 2, a quality detection pattern, a horn antenna, a coupling lens, a collimation, beam expansion and focusing module, and a lock-in amplifier.

[0009] The terahertz transmitter 1 and terahertz transmitter 2 are used to generate terahertz waves of two wavelengths.

[0010] The horn antenna is used to gradually extend terahertz waves from the waveguide into free space.

[0011] The coupling lens is used to adjust or focus the electromagnetic waves transmitted / received by the antenna;

[0012] The collimation, beam expanding and focusing module consists of a Kepler-type beam expander composed of two lenses and a focusing lens.

[0013] Furthermore, terahertz transmitter 1 and terahertz transmitter 2 generate two terahertz waves of different wavelengths. The terahertz waves are uniformly propagated into space through the horn antenna, and the propagation direction and divergence angle of the beams are optimized by the coupling lens placed in front.

[0014] Next, after the two terahertz waves pass through a semi-reflective lens, the propagation directions of the two terahertz waves are adjusted to the same direction. Then, after passing through the collimation, beam expansion and focusing module, the beam spread and divergence angle are further reduced, the stability and directionality of the terahertz waves in space are improved, and the two terahertz waves are focused on the printed pattern sample to be tested.

[0015] A portion of the beam is transmitted through the sample and named beam 1, while the other portion is reflected and named beam 2. Both beams 1 and 2 are adjusted to the direction of the two terahertz receivers via mirrors, and then reach terahertz receiver 1 and terahertz receiver 2 respectively through focusing lenses. Subsequently, these two terahertz receivers transmit the signal to a lock-in amplifier to enhance the signal accuracy and signal-to-noise ratio. Then, the amplified and processed signal is transmitted to a computer for acquisition and processing. Finally, the computer outputs the final predicted conductivity data, as well as the dimensional accuracy and geometric deformation data of the printed circuit, after judging by a pre-trained prediction model.

[0016] The second objective of this invention is to provide a printed circuit quality inspection method based on dual-wavelength synchronous transmission and reflection detection. By utilizing dual-wavelength synchronous transmission and reflection measurement technology, this invention effectively improves the accuracy and repeatability of printed circuit conductivity detection, while monitoring the dimensional accuracy and geometric deformation of printed circuits, thus optimizing the printed electronics production quality monitoring method and reducing the output of defective products.

[0017] The manufacturing of printed circuit boards includes the following steps:

[0018] S1. Select ink and substrate;

[0019] S2. Printing is carried out;

[0020] S3, sintering treatment;

[0021] S4. Printing quality inspection.

[0022] Specifically, in step S1, the ink is conductive ink and the substrate is a flexible substrate.

[0023] Furthermore, the conductive ink is a metallic ink, a carbon-based ink, or a composite ink.

[0024] Furthermore, the metallic ink is a metallic ink with gold, silver, and copper metal particles as the main conductive components.

[0025] Specifically, in step S2, the printing includes flexographic printing, gravure printing, screen printing, and inkjet printing.

[0026] Specifically, in step S4, we propose a printed circuit quality inspection method based on dual-wavelength synchronous transmission and reflection detection to achieve printed circuit quality inspection.

[0027] Furthermore, in this detection method, the dimensional parameters of the printing quality detection pattern (open-ring metamaterial) are designed and optimized to generate two distinct resonance peaks in the terahertz band. The frequencies of these two resonance peaks are further tuned to 200 GHz and 300 GHz by optimizing the open-ring structural parameters. Since both resonance peaks respond to both conductivity and metamaterial structural dimensions, the difference or ratio between them is analyzed to avoid conductivity measurement errors caused by variations in the incident position of electromagnetic waves on the quality control pattern or differences in the quality control pattern during the printing process.

[0028] Furthermore, in step S4, a multidimensional data array of material conductivity, resonance peak position, resonance peak intensity, and resonance peak Q factor is established as a calibration model for conductivity estimation through theoretical simulation or experimental calibration. After obtaining the characteristic spectrum of the open-loop resonant array, the conductivity of the printing ink to be tested is estimated by searching for the best-matching conductivity in the calibration data or model. Simultaneously, a calibration model for the dimensional accuracy and geometric deformation of the printed circuit can be established using the same method to estimate the dimensional accuracy and geometric deformation data of the printed circuit.

[0029] Due to the adoption of the above technical solution, the advantages of the present invention compared with the prior art are:

[0030] (1) By adopting a non-contact testing method, compared with existing methods on the market that rely on contact methods such as multimeters or four-probes for printed circuit quality monitoring, our method avoids sample damage or test interference that may be caused by physical contact, making it particularly suitable for the production of flexible electronics with high precision and high sensitivity requirements. In addition, real-time monitoring technology ensures that any quality deviation can be detected and dealt with immediately, greatly reducing the generation of defective products and improving the economic efficiency of the production process;

[0031] (2) Compared to traditional single-wavelength measurement, this invention utilizes a dual-wavelength transmission and reflection synchronous measurement method for the roll-to-roll printed circuit production process. This method optimizes traditional single-wavelength measurement by simultaneously monitoring the reflection and transmission spectra at two specific frequencies (wavelengths), effectively analyzing changes in the electrical properties and physical structure of the material. This technology can significantly improve the accuracy of detection and achieve strong suppression of environmental fluctuations and mechanical errors during the detection process;

[0032] (3) The data processing system of the present invention can analyze detection data (such as resonance peak position and resonance peak intensity and Q factor) in real time and automatically adjust production parameters based on data feedback. This real-time feedback and adjustment mechanism is significantly better than the usually delayed feedback in the prior art, thus improving production efficiency and product quality;

[0033] (4) This invention establishes a database containing multiple test indicators (such as resonance peak position, intensity, Q factor, etc.), and uses a comparison method based on the database and formula calculation to achieve accurate estimation of the conductivity, dimensional accuracy and geometric deformation data of printed circuits. Compared with the traditional direct measurement method, it not only improves the accuracy and efficiency of measurement, but also enables dynamic control and optimization of the production process. By comparing the actual measurement data with theoretical or previous experimental data in real time, this method can quickly identify and correct any deviations in production, greatly reducing the output of defective products and improving economic efficiency. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the synchronous transmission and reflection system in Example 1;

[0035] Figure 2 This is a schematic diagram of the dual-reflection system in Example 1;

[0036] Figure 3 This is a schematic diagram of the dual-transmission system in Example 1;

[0037] Figure 4 This is a flowchart of the quality inspection process for the synchronous transmission and reflection system in Example 1;

[0038] Figure 5 This is a schematic diagram of the THz wave incident on the open-loop array in Example 2;

[0039] Figure 6 This is a schematic diagram of the open-ring metamaterial in Example 2;

[0040] Figure 7 This is a schematic diagram of the double resonance peak in Example 2;

[0041] Figure 8 This is a schematic diagram of the conductivity change in Example 2;

[0042] Figure 9 This is a schematic diagram of the geometric dimension (line width t) changes in Example 2. Detailed Implementation

[0043] Example 1

[0044] like Figures 1 to 4As shown in the figure, this invention provides a printed circuit quality monitoring device based on dual-wavelength synchronous transmission and reflection detection. As shown in the figure below, we have designed three measurement systems for printed circuit quality detection: a synchronous transmission and reflection system, a dual reflection system, and a dual transmission system.

[0045] During the online inspection of roll-to-roll printed circuits, the terahertz spot may drift at the incident position on the quality control pattern due to a series of mechanical rotations or translations. Therefore, signal errors may arise from the different incident positions of the terahertz wave, affecting the accuracy of the inspection results. This error will be reflected in the reflection or transmission spectrum, potentially influencing the position and intensity of the resonance peak. To avoid reduced repeatability caused by the terahertz incident position, we propose a method based on simultaneous transmission and reflection measurements to improve inspection accuracy and optimize repeatability.

[0046] The components of the transmission and reflection system are as follows: 901 is the computer, 902 is the terahertz receiver 1, 903 is the terahertz receiver 2, 904 is the terahertz transmitter 1, 905 is the terahertz transmitter 2 (904 and 905 are used to generate terahertz waves of two wavelengths (frequency)), 906 is the quality detection pattern, 907 is the horn antenna (used to gradually extend the terahertz wave from the waveguide into free space), 908 is the coupling lens (used to adjust or focus the electromagnetic waves transmitted / received by the antenna), 909 is the collimation, beam expander and focusing module, which consists of a Kepler-type beam expander composed of two lenses and a focusing lens, and 910 is the lock-in amplifier.

[0047] Taking the synchronous transmission and reflection measurement system as an example, the overall workflow is as follows: Terahertz transmitters 904 and 905 generate two terahertz waves of different wavelengths. The terahertz waves propagate uniformly into space via a horn antenna, and the propagation direction and divergence angle of the beams are optimized by a coupling lens placed in front. Then, after the two terahertz waves pass through a semi-reflective mirror, the propagation direction of the two terahertz waves is adjusted to the same direction. Then, the collimation, beam expanding and focusing module 909 further reduces the beam spread and divergence angle, improves the stability and directionality of the terahertz waves during propagation in space, and focuses the two terahertz waves onto the printed pattern sample to be inspected. A portion of the beam is transmitted through the sample and named beam 1, while the other portion is reflected and named beam 2. Both beams 1 and 2 are adjusted to the direction of the two terahertz receivers via mirrors, and then reach terahertz receivers 902 and 903 respectively through focusing lenses. Subsequently, these two terahertz receivers transmit the signal to lock-in amplifier 910 to enhance the signal accuracy and signal-to-noise ratio. Then, the amplified and processed signal is transmitted to computer 901 for acquisition and processing. Finally, the computer outputs the final predicted conductivity data, as well as the dimensional accuracy and geometric deformation data of the printed circuit, after judging by the pre-trained prediction model in the computer.

[0048] This invention employs a synchronous measurement technology based on reflection and transmission, which helps to effectively screen for anomalies during printed circuit board (PCB) quality inspection. The synchronous transmission and reflection measurement system ensures consistent correlation between the two signals by simultaneously measuring the reflection and transmission spectra of the open-ring metamaterial. If this correlation is not met or the deviation exceeds a predetermined range during actual testing, the test result is considered abnormal and will not be considered a valid PCB conductivity test. Furthermore, the system automatically stores these abnormal data, facilitating subsequent troubleshooting and resolution of problems encountered during production. Moreover, this synchronous transmission and reflection measurement technology is based on dual-wavelength synchronous measurement, enabling the system to comprehensively analyze the electrical properties and physical structure of the material, avoiding the accuracy errors that may result from using a single wavelength measurement, and further improving the accuracy and sensitivity of the measurement. Simultaneously, this dual-wavelength synchronous transmission and reflection measurement technology effectively eliminates the influence of environmental fluctuations and mechanical errors, ensuring the stability of the test results. Due to the synchronicity of the two signals, any errors caused by different incident positions or external interference can be effectively suppressed, significantly improving the repeatability and accuracy of the test.

[0049] Furthermore, this dual-wavelength transmittance and reflectance synchronous measurement system can provide real-time feedback on any deviations during the production process. When a problem is detected, the system automatically adjusts production parameters (such as ink flow rate and sintering temperature) to further optimize the production process, improve efficiency, and reduce scrap rates. The entire measurement system is compatible with automated production and provides a real-time, high-precision quality inspection solution for printed circuit board manufacturing, thereby ensuring efficient and accurate production and consistent product quality.

[0050] Example 2

[0051] Printed electronics is a cutting-edge technology that applies traditional printing processes to the manufacture of electronic components and devices. It enables the direct printing of electronic devices onto various flexible substrates, achieving low-cost, high-efficiency production of electronic products. This technology uses multiple methods, including flexographic printing, inkjet printing, gravure printing, and screen printing, to directly print photonic devices and electronic products onto various substrates such as fabrics, paper, and plastics. This technology has been widely applied in wearable devices, the sports industry, and flexible displays. Due to the fast production pace, the quality of the printing process can vary significantly.

[0052] The manufacturing process of printed electronics generally consists of four main stages: (a) ink and substrate selection, (b) printing, (c) sintering, and (d) printing quality inspection. The first step is selecting suitable ink and substrate. Among the various available substrates, flexible substrates are particularly favored due to their unique flexibility, stretchability, and torsionability while maintaining good electrical properties. A variety of conductive, semi-conductive, and dielectric inks have been developed to meet different printing needs, with conductive inks receiving particular attention. These include three main categories: metallic inks, carbon-based inks, and composite inks. Metallic inks, with gold, silver, copper, and other metal particles as the conductive core, are widely used in commercial applications due to their superior conductivity, which improves signal transmission efficiency, reduces latency, and decreases energy consumption in electronic circuits.

[0053] Once the substrate and ink are determined, the next step is printing. Printing technologies are selected based on specific needs, including flexographic printing, gravure printing, screen printing, and inkjet printing, all of which are widely used in the market. In inkjet printing, functional inks are prepared by uniformly dispersing functional materials (such as metal nanoparticles, conductive polymers, or carbon nanotubes) in a suitable solvent to form a stable dispersion system. The uniformity of the ink directly determines the quality and performance of the printed electronic circuits.

[0054] In addition, in the process of printed electronics manufacturing, the quality of ink directly determines the performance indicators of the final product, including power consumption, response speed and overall electrical performance. Therefore, the quality of ink must be strictly controlled during the manufacturing process to ensure the production of efficient and reliable electronic products.

[0055] Throughout the production process, printing quality inspection is extremely critical, especially in a fast-paced production environment where quality can vary significantly during printing. This presents a major challenge for the printed electronics industry. Unlike the traditional graphic printing industry, where quality control primarily focuses on color, resolution, registration, and surface finish, printed electronics manufacturing must not only ensure the stability of the process itself but also guarantee the functionality of the devices.

[0056] However, achieving real-time, online process feedback in mass production currently faces many challenges, yet it is crucial for industrial manufacturing. Manufacturers often face high costs when the production process goes out of control. Problems are often only discovered after the entire roll of product has been produced, and evaluating the functionality of devices or materials on the substrate through off-line testing can lead to a large amount of material being scrapped due to defects, resulting in significant economic losses. To achieve the economic benefits of mass production, scrap rates must be kept low, requiring real-time, non-destructive online quality control during production. In printed circuit board (PCB) production, functional and aesthetic inspections are indispensable. Currently, machine vision-based inspection methods are mainly used to identify aesthetic defects. However, electrical performance testing, as one of the key indicators of PCBs, has traditionally relied on contact methods such as multimeters or four-probe systems. These methods are not only difficult to integrate with automated production but can also sometimes damage devices. Furthermore, contact testing methods generally have poor accuracy and repeatability; therefore, finding a non-contact, real-time measurement solution is particularly important.

[0057] This invention aims to overcome the shortcomings of existing technologies and provide a non-contact testing method, structure, and apparatus for printed circuit quality. This method is mainly used to detect the conductivity, diffusion characteristics, and dimensional accuracy of conductive inks during the printing process, thereby improving the product qualification rate in roll-to-roll printing. This testing method not only accelerates the quality control process but also avoids physical damage to the printed circuit through its non-contact nature, ensuring the integrity and functionality of the circuit.

[0058] This invention proposes an innovative method that introduces a quality control pattern based on metamaterials into the printed circuit manufacturing process, such as... Figure 5 and Figure 6 As shown, considering the unique modulation effect of conductive materials commonly used in printed electronics on terahertz waves, this invention utilizes this characteristic to achieve non-contact online measurement of the conductivity of printed circuits and the dimensional accuracy and deformation of printed lines. Specifically, we designed metamaterial control patterns with specific reflection and transmission properties. These patterns can comprehensively evaluate key functional parameters such as the conductivity, thickness, and diffusion characteristics of printing inks through their transmission characteristics (such as resonant frequency, intensity, and Q factor).

[0059] By deeply studying the resonant characteristics of open-loop resonant ring array metamaterials, especially the influence of material conductivity on these characteristics, we found that when the material's conductivity changes, the intensity at its resonant frequency exhibits a significant change. Specifically, higher conductivity results in greater resonant intensity and a higher Q-factor of the characteristic peak. This discovery means that by analyzing the intensity of the resonant peak and the Q-factor of the characteristic peak in real time, it is possible to achieve real-time monitoring of the conductivity of printing inks. Therefore, this phenomenon provides a valuable physical basis for non-contact detection of the conductivity of printed circuits, helping to monitor the electrical performance of printed circuits in real time and ensure product quality. Furthermore, the application potential of this method is not limited to printed circuits but may also extend to other fields involving the evaluation of conductive material performance.

[0060] To achieve the goal of real-time monitoring of the conductivity of printing ink, one approach is to employ a spectral scanning method based on a time-domain terahertz system. This method has the advantage of comprehensively detecting the characteristic peaks and overall reflection and transmission spectra of metamaterials. However, its biggest limitation lies in the fact that this method requires mechanical scanning to sample the time-domain delay of the terahertz pulses, thus presenting a drawback in terms of real-time performance. This undoubtedly restricts its application in automated production lines for online real-time monitoring. For example, in current roll-to-roll printing processes, printing speeds often reach 50 m / min, while the scanning speed of conventional time-domain terahertz systems is insufficient to meet these requirements.

[0061] To address the aforementioned challenges, this invention proposes a printed circuit detection method based on dual-frequency excitation and detection. Taking an open-ring metamaterial based on conductive silver paste as an example, we designed and optimized the size parameters of the open ring to generate two distinct resonance peaks in the terahertz band. Since both resonance peaks respond to both conductivity and the metamaterial's structural dimensions, we can analyze the difference or ratio between them to avoid conductivity measurement errors caused by variations in the incident position of electromagnetic waves on the quality control pattern or differences in the quality control pattern during printing. This significantly improves the accuracy of conductivity measurement; therefore, compared to single-frequency measurement, our dual-frequency method has higher measurement accuracy and sensitivity. Furthermore, to further enhance detection accuracy and robustness, we optimized the open-ring structure by precisely adjusting the key size parameters of the open ring, gradually modulating the frequencies of the two resonance peaks to 200 GHz (0.2 THz) and 300 GHz (0.3 THz), respectively. Figure 7As shown. The selection of these two frequencies is primarily based on the following considerations: Currently, commercially available single-frequency sources exhibit high radiated power at these frequencies, and the related detection technologies are mature and sophisticated, capable of providing high-precision signal measurement and real-time monitoring. This allows for stronger signal quality and lower noise levels when conducting experiments and tests within these frequency ranges. This optimized design not only enhances the sensitivity of the detected signal but also makes the system more compatible with existing commercial terahertz detection equipment, thereby improving overall practicality and operability.

[0062] Figure 8 The changes in resonance peaks with varying conductivity are shown. When the conductivity increases from 6000 S / m to 60000 S / m, the transmittance of the resonance peak at 200 GHz changes from 0.325 to 0.062, and the transmittance at 300 GHz changes from 0.556 to 0.289, with both resonance intensities increasing. Therefore, a matching relationship or mapping matrix between the resonance peak characteristics (resonance peak intensity 1, resonance peak intensity 2) and the material conductivity of the open-ring metamaterial can be established through theoretical simulation or test calibration. Alternatively, a deep learning model can be built to correlate the resonance peak input characteristics with the conductivity output. This approach can improve the accuracy and efficiency of predictions, thus providing strong technical support for the quality control and performance optimization of printed circuits.

[0063] In this example implementation, the present invention proposes a conductivity estimation method based on optimal matching. First, a multi-dimensional data array of material conductivity, resonance peak position, resonance peak intensity, and resonance peak Q factor is established as a calibration model for conductivity estimation through theoretical simulation or experimental calibration. During actual testing, after obtaining the characteristic spectrum of the open-ring resonator array, the conductivity of the printing ink under test is estimated by searching for the optimal matching conductivity in the calibration data or model. To establish the relationship between the intensity of resonance peak 1 and resonance peak 2 and the conductivity, a quadratic regression model is proposed to fit the data in this example. Assuming the intensity of resonance peak 1 is x1, the intensity of resonance peak 2 is x2, and the conductivity is the dependent variable y, a regression model containing linear, quadratic, and cross-terms of the independent variables is constructed as follows:

[0064]

[0065] The residual sum of squares (RSS) is the sum of the squared residuals of all sample points, representing the total error of the model fit. Its formula is:

[0066]

[0067] The core objective of the least squares method is to minimize the sum of squared residuals by adjusting the regression coefficients β0, β1, ..., β5. Specifically, the core principle of the least squares method is to find the best-fit curve by minimizing the sum of squared residuals across all data points, where the residuals refer to the differences between the actual observed values ​​and the model's predicted values. Through this process, we obtain a set of regression coefficients that optimize the model's fit to the data. Based on this, we can predict the conductivity of materials using the regression model and further analyze the influence of the resonance peak intensity on conductivity, revealing the linear and nonlinear relationships between the two. This regression analysis provides an effective mathematical tool for a deeper understanding of the conductivity properties of materials and also provides a theoretical basis for subsequent material design and performance optimization. In practical applications, we will judge the quality of the fitted model by testing the mean squared error and the adjusted coefficients of determination, while avoiding overfitting.

[0068] The training data (calibration spectrum) of the above regression model can be measured in the laboratory and stored in the database. When training the model parameters, the parameters in the database can be read directly. At the same time, the operating data in the production line can also be processed and added to the database to adjust the test accuracy of the model and improve the model performance.

[0069] In another implementation, we use calibration data to train a Support Vector Regression (SVR)-based model to predict the conductivity of printing inks. The model's input variables include the intensities of resonance peaks 1 and 2, and its output is the conductivity of the printing ink. By fine-tuning the model's hyperparameters, we can optimize its performance, thereby significantly improving the accuracy of the printed ink conductivity estimation. Furthermore, on actual production lines, we periodically integrate newly collected data into the training set. This not only helps the model adapt to changes in the production environment but also further enhances the accuracy of predictions and the model's generalization ability. This continuous data updating and model iteration is a key strategy for ensuring long-term stability and improving predictive performance.

[0070] Furthermore, since the modulation characteristics of the quality control pattern to terahertz waves depend not only on conductivity but also on another important quality indicator of the printed pattern (size linewidth and geometric deformation), to avoid the impact of dimensional errors or deformation of the quality control pattern on the conductivity detection accuracy during printing, and to achieve real-time monitoring of the dimensional accuracy and deformation of the printed lines, we propose two auxiliary detection schemes to optimize the accuracy and reliability of the detection process: 1. The first scheme is based on resonance design. Utilizing the dependence of its resonant frequency on parameters such as the size and linewidth of the resonant ring, the size of the printed ink can be effectively detected by monitoring its resonant position. For example, our research found that when the linewidth t of the resonant ring changes from 40μm to 80μm, the center position of the resonant frequency of one resonant peak shifts from 186GHz to 199GHz, and the other resonant peak shifts from 283GHz to 297GHz. Figure 9 As shown. Therefore, by monitoring the position of its resonant frequency, the linewidth accuracy and diffusion function of the printed pattern can be effectively detected. 2. The second method is based on machine vision, which uses image processing algorithms to detect and obtain the dimensional accuracy and deformation characteristics of the printed pattern. By updating the geometric linewidth parameters in the printing ink conductivity estimation model in real time, the conductivity detection accuracy is improved. When the system detects that the conductivity or dimensional accuracy of the printing ink exceeds a reasonable range, it will automatically adjust parameters such as ink flow rate, thickness, concentration or proportion of conductive particles, and sintering temperature to optimize the production process and improve the yield of roll-to-roll printing. These measures ensure the continuity of the production process and the consistency of product quality.

[0071] Furthermore, this invention provides two environmental interference elimination methods based on dual-frequency analysis. The first method involves optimizing the metamaterial structure so that changes in the conductivity of the printing material only affect a specific frequency range, while having almost no impact on other frequency bands. The response ratio of the characteristic frequency band and the reference frequency band cleverly provides information on the transmission characteristics of the studied parameters to the structure. Since the two are measured synchronously, the ratio can achieve a normalization effect, effectively suppressing interference from environmental fluctuations or other uncertain factors. Alternatively, the second method can be used. By optimizing the structure of the open-loop resonant ring or metamaterial, two or more resonance peaks can be obtained over a wide bandwidth. These resonance peaks respond to changes in conductivity and size structure. By analyzing the difference or ratio between the two, environmental fluctuations and other factors can be effectively suppressed. This avoids conductivity measurement errors caused by different positions of the THz wave incident on the quality control pattern or differences in the quality control pattern during the printing process, thus optimizing the conductivity measurement accuracy. This is also why we previously optimized the open-loop structure to generate two distinctive resonance peaks.

[0072] Using the methods described above, we can inspect the electrical properties (such as conductivity) and appearance indicators (such as dimensional accuracy and geometric deformation) of printed circuits online. This is particularly crucial for roll-to-roll printing production. These technologies not only help prevent the production process from spiraling out of control and generating large amounts of waste, thus avoiding waste and losses, but are also extremely beneficial for the development of new products or processes, significantly shortening the development cycle for testing new equipment and debugging new processes. In summary, the non-destructive, non-contact, online inspection method based on metamaterials proposed in this invention provides an effective new approach to printing quality control in the field of printed electronics, and is expected to be widely applied in the quality inspection and control of printed electronics.

[0073] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A printed circuit quality monitoring device based on dual-wavelength synchronous transmittance and reflectance detection, characterized in that, Includes a computer, terahertz receiver 1, terahertz receiver 2, terahertz transmitter 1, terahertz transmitter 2, quality detection pattern, horn antenna, coupling lens, collimation, beam expander and focusing module and lock-in amplifier; The terahertz transmitter 1 and terahertz transmitter 2 are used to generate terahertz waves of two wavelengths, with frequencies of 200 GHz and 300 GHz, respectively. By designing and optimizing the size parameters of the printing quality detection pattern, two obvious resonance peaks are generated in the terahertz band, and the frequencies of the two resonance peaks are tuned to 200 GHz and 300 GHz, respectively, by optimizing the open-loop structure parameters. The horn antenna is used to gradually extend terahertz waves from the waveguide into free space. The coupling lens is used to adjust or focus the electromagnetic waves transmitted / received by the antenna; The collimation, beam expanding and focusing module consists of a Kepler-type beam expander composed of two lenses and a focusing lens; Terahertz transmitter 1 and terahertz transmitter 2 generate two terahertz waves of different wavelengths. The terahertz waves are uniformly propagated into space through a horn antenna and the propagation direction and divergence angle of the beams are optimized by a coupling lens placed in front. Next, after the two terahertz waves pass through a semi-reflective lens, the propagation directions of the two terahertz waves are adjusted to the same direction. Then, after passing through the collimation, beam expansion and focusing module, the beam spread and divergence angle are further reduced, the stability and directionality of the terahertz waves in space are improved, and the two terahertz waves are focused on the printed pattern sample to be tested. A portion of the beam is transmitted through the sample and named beam 1, while the other portion is reflected and named beam 2. Both beams 1 and 2 are adjusted to the direction of the two terahertz receivers via mirrors, and then reach terahertz receiver 1 and terahertz receiver 2 respectively through focusing lenses. Subsequently, these two terahertz receivers transmit the signal to a lock-in amplifier to enhance the signal accuracy and signal-to-noise ratio. Then, the amplified and processed signal is transmitted to a computer for acquisition and processing. Finally, the computer outputs the final predicted conductivity data, as well as the dimensional accuracy and geometric deformation data of the printed circuit, after judging by a pre-trained prediction model.

2. The printed circuit quality inspection method based on dual-wavelength synchronous transmittance and reflectance detection according to claim 1, characterized in that, The manufacturing of printed circuit boards includes the following steps: S1. Select ink and substrate; S2. Printing is carried out; S3, sintering treatment; S4. Printing quality inspection; In step S4, since both resonance peaks respond to conductivity and metamaterial structure size, the difference or ratio between them is analyzed to avoid conductivity measurement errors caused by different incident positions of electromagnetic waves on the quality control pattern and differences in the quality control pattern during printing.

3. The printed circuit quality inspection method based on dual-wavelength synchronous transmittance and reflectance detection according to claim 2, characterized in that, In step S1, the ink is conductive ink and the substrate is a flexible substrate.

4. The printed circuit quality inspection method based on dual-wavelength synchronous transmittance and reflectance detection according to claim 3, characterized in that, The conductive ink is a metallic ink, a carbon-based ink, or a composite ink.

5. A printed circuit quality inspection method based on dual-wavelength synchronous transmittance and reflectance detection according to claim 4, characterized in that, The metallic ink is a metallic ink with gold, silver, and copper metal particles as the main conductive components.

6. The printed circuit quality inspection method based on dual-wavelength synchronous transmittance and reflectance detection according to claim 2, characterized in that, In step S2, the printing includes flexographic printing, gravure printing, screen printing, and inkjet printing.

7. A printed circuit quality inspection method based on dual-wavelength synchronous transmittance and reflectance detection according to claim 2, characterized in that, In step S4, a multidimensional data array of material conductivity, resonance peak position, resonance peak intensity, and resonance peak Q factor is established as a calibration model for conductivity estimation through theoretical simulation or experimental calibration. After obtaining the characteristic spectrum of the open resonant ring array, the conductivity of the printing ink to be tested is estimated by searching for the best matching conductivity in the calibration data or model. At the same time, a calibration model for the dimensional accuracy and geometric deformation of the printed circuit is established using the same method to estimate the dimensional accuracy and geometric deformation data of the printed circuit.

Images