Method for measuring internal transmission of a highly transmissive optical material

Abstract

The application discloses a kind of high transmission optical material inner transmittance measurement method, it is related to optical material performance parameter measurement field.The present application comprehensively introduces bayesian optimization and physical information neural network algorithm, realizes the high-precision inversion and prediction of material transmittance performance;Among them, high time resolution measurement is realized by using pulsed light cavity ring-down technology, ensure the accurate evaluation of high inner transmittance material in very low loss range, provide a feasible path for non-destructive, traceable measurement of ultra-high transmittance material;Solve the problem that spectrophotometric method cannot accurately measure the inner transmittance of optical material 99.0% / cm and above, accurate measurement of higher than 99.0% / cm inner transmittance of high-transmission optical material is realized by using pulsed light cavity ring-down technology, and the highest measurement accuracy can reach 0.0001% / cm.

Description

Technical Field

[0001] This invention relates to the field of optical material performance parameter measurement, specifically a method for measuring the internal transmittance of high-transmittance optical materials. Background Technology

[0002] Internal transmittance is a fundamental technical parameter of optical materials, and accurate measurement is a basic prerequisite for determining whether an optical material can be applied to a specific optical system. Spectrophotometry is typically used to directly measure the internal transmittance of optical materials. The national standard GB / T 7962.12-2010, "Test Methods for Colorless Optical Glass Part 12: Spectral Internal Transmittance," recommends using a dual-path spectrophotometric method to measure the transmittance of thin optical elements (5 mm thick) and thick optical elements (15 mm) separately. The ratio of the transmittance of the thick optical element to that of the thin optical element is the internal transmittance of the corresponding optical material. Typical measurement accuracies for internal transmittance are ±0.3% (wavelength range: 200~780 nm) and ±0.5% (780~2500 nm). To reduce the impact of light source intensity fluctuations on measurement results, Chinese patent application 201711238029.3, "Method for Measuring Internal Transmittance of Glass Spectrum," proposes using thin optical elements to calibrate the reference and test optical paths, and then measuring the transmittance of thin and thick optical elements separately to calculate the internal transmittance of the material being tested. Meanwhile, Chinese patent application 202410801428.X, "A Method for Measuring Internal Transmittance of Colored Glass Material," employs a single-channel spectrophotometric method to measure the external transmittance and reflectance of optical elements of a specific thickness, and then calculates the internal transmittance of the material being tested through data analysis and formulas. Both of these internal transmittance measurement methods are based on spectrophotometry, with a typical absolute measurement accuracy of ±0.3% and a repeatability accuracy of approximately ±0.1%. Clearly, an internal transmittance measurement accuracy of ±0.3% cannot meet the testing requirements for the internal transmittance of high-transmittance optical materials (above 99.0% / cm). Especially when the internal transmittance of optical materials exceeds 99.7% / cm, the measurement accuracy should reach at least ±0.01% to achieve accurate measurement of high internal transmittance. Therefore, spectrophotometry cannot meet the testing requirements for high internal transmittance of highly transparent optical materials.

[0003] When the internal transmittance of an optical material is higher than 99.0% / cm, the volume loss coefficient of the optical material can be accurately measured. α (Unit: 1 / cm), thus accurately determining the internal transmittance of optical materials (1- α(Unit: 1 / cm). Cavity ring-down technology, as a high-precision optical loss measurement technique for optical materials, can measure optical losses as low as 0.0001%, enabling accurate measurement of extremely low optical loss coefficients (as low as 0.0001% / cm) within optical materials. Furthermore, since cavity ring-down technology measures optical loss by measuring the attenuation time of the beam within the ringing cavity, accurate measurement of the ringing time is crucial for determining the optical loss. The ringing time is approximately inversely proportional to the total optical loss within the ringing cavity, with an upper limit of about 1%. Therefore, the applicable optical loss measurement range for cavity ring-down technology is 0.0001% to 1.0%. Depending on the thickness of the optical element being measured, the applicable range for measuring the optical loss coefficient within the optical element is 0.0001% to 5.0% / cm, corresponding to an internal transmittance measurement range of 95.0% to 99.9999% / cm.

[0004] For optical components made from the tested optical materials, their surfaces lack anti-reflection coatings and exhibit significant surface reflection, typically ranging from 3% to 6%. When measuring the optical loss of these components using cavity ring-down technology, the measuring laser beam must be incident perpendicularly to the surface of the tested optical component, ensuring that the reflected beam continues to propagate within the ring-down cavity without contributing to additional optical loss. However, research results indicate that the presence of surface reflection significantly impacts the measured optical loss, leading to a higher measured optical loss value than the actual loss of the tested optical component. Chinese patent application 202410162030.6, "A data processing method for eliminating the influence of surface reflection on optical loss measurement in cavity ring-down technology," eliminates the influence of surface reflection on optical loss measurement by establishing an approximately linear relationship between the measured optical loss value and the actual loss value of the optical element under test. However, this data processing method is only applicable when the optical loss value of the optical element under test is low (0.0001%~0.1%). When the actual loss value of the optical element is greater than 0.1% (typical range 0.1%~1%), the data processing method based on linear fitting has a large calculation error, which affects the measurement accuracy of the actual optical loss of the optical element under test. Summary of the Invention

[0005] This invention proposes a method for measuring the internal transmittance of high-transmittance optical materials, aiming to solve the problems of insufficient measurement accuracy of existing spectrophotometric methods, large surface reflection errors in optical cavity ringback methods, difficulty in handling nonlinear relationships, and limited measurement range.

[0006] One method for measuring the internal transmittance of a high-transmittance optical material includes the following steps:

[0007] S1. Prepare at least two planar optical elements with different thicknesses made of the high-transmittance optical material to be tested, measure the actual ring-down time of each planar optical element, and calculate the measured loss value based on the initial cavity ring-down time and the actual ring-down time of each planar optical element;

[0008] S2. Based on the measured actual ring-down time, establish a theoretical ring-down signal model for the test cavity containing planar optical elements; convolve the theoretical ring-down signal output by the model with the time-domain response function of the photodetector to obtain the simulated detection signal; and perform least-squares fitting on the simulated detection signal according to the single exponential decay form to extract the theoretical ring-down time, and calculate the corresponding theoretical loss value based on the theoretical ring-down time.

[0009] S3. Using the calculated measured loss value and theoretical loss value as input data, a loss mapping model is constructed by combining the Bayesian optimization algorithm, and the actual loss value after correction for multiple planar optical elements with different thicknesses is output through the loss mapping model.

[0010] S4. Based on the actual loss values ​​after correction of multiple planar optical elements with different thicknesses, and combined with the sample thickness difference, the volume optical loss rate per unit length is extracted using a linear fitting method.

[0011] S5. Using the extracted bulk optical loss rate and the basic optical parameter information of the optical material under test as input parameters, a physical information neural network model is constructed to predict the transmittance inversion; and by combining the physical information neural network model with the physical prior constraint relationship, the internal transmittance of the high-transmittance optical material under test is output.

[0012] The beneficial effects of the invention are:

[0013] (1) This invention achieves high-precision inversion and prediction of material transmittance performance by constructing a multi-step process that combines theoretical modeling and experimental measurement, and comprehensively introduces Bayesian optimization and physical information neural network algorithms. Among them, the interference of interface reflection and system response error is effectively suppressed by multiple ring-down tests of samples with thickness differences. Combined with the sample design with high parallelism and low surface roughness, the measurement stability and repeatability are significantly improved. At the same time, the pulsed cavity ring-down technology is used to achieve high time resolution measurement, ensuring accurate evaluation of high internal transmittance materials within an extremely low loss range, and providing a feasible path for non-destructive and traceable measurement of ultra-high transmittance materials.

[0014] (2) This invention solves the problem that the spectrophotometric method cannot accurately measure the internal transmittance of optical materials above 99.0% / cm. It uses pulsed cavity ring-down technology to achieve accurate measurement of the internal transmittance of high-transmittance optical materials above 99.0% / cm, with the highest measurement accuracy reaching 0.0001% / cm. Attached Figure Description

[0015] Figure 1 This is a flowchart of a method for measuring the internal transmittance of a high-transmittance optical material according to Embodiment 1 of the present invention;

[0016] Figure 2 This is a graph showing the relationship between the actual loss value and the measured loss value of the optical material under test, provided in Embodiment 3 of the present invention.

[0017] Figure 3 The relationship between the actual loss value and thickness of optical elements of different thicknesses, as measured by the pulsed cavity ring-down method, and its linear fitting graph are provided in Embodiment 3 of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the scope of protection of the present invention is not limited to the following description.

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention; that is, the described embodiments are only a part of the embodiments of the invention, and not all of them. The components of the embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention. It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0021] Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or machine that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or machine. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or machine that includes said element.

[0022] The features and performance of the present invention will be further described in detail below with reference to embodiments.

[0023] Example 1

[0024] like Figure 1 A method for measuring the internal transmittance of a high-transmittance optical material, comprising the following steps:

[0025] S1. Prepare at least two planar optical elements of different thicknesses made of the high-transmittance optical material to be tested, measure the actual ring-down time of each planar optical element, and calculate the measured loss value based on the initial cavity ring-down time and the actual ring-down time of each planar optical element, wherein the initial cavity ring-down time is obtained according to the initial cavity configuration of the pulsed cavity ring-down technology;

[0026] S2. Based on the measured actual ring-down time, establish a theoretical ring-down signal model for the test cavity containing planar optical elements; convolve the theoretical ring-down signal output by the model with the time-domain response function of the photodetector to obtain the simulated detection signal; and perform least-squares fitting on the simulated detection signal according to the single exponential decay form to extract the theoretical ring-down time, and calculate the corresponding theoretical loss value based on the theoretical ring-down time.

[0027] S3. Using the calculated measured loss value and theoretical loss value as input data, a loss mapping model is constructed by combining the Bayesian optimization algorithm, and the actual loss value after correction for multiple planar optical elements with different thicknesses is output through the loss mapping model.

[0028] S4. Based on the actual loss values ​​after correction of multiple planar optical elements with different thicknesses, and combined with the sample thickness difference, the volume optical loss rate per unit length is extracted using a linear fitting method.

[0029] S5. Using the extracted bulk optical loss rate and the basic optical parameter information of the optical material under test as input parameters, a physical information neural network model is constructed to predict the transmittance inversion; and by combining the physical information neural network model with the physical prior constraint relationship, the internal transmittance of the high-transmittance optical material under test is output.

[0030] Specifically, for step S1, cavity ring-down (CRD) technology is used to test high-transmittance optical material samples of different thicknesses. The workflow is as follows: First, two or more planar optical elements with significantly different thicknesses but consistent surface quality are fabricated. All elements are made from the material being tested and must be double-sided polished with a surface parallelism better than 2′ and a surface roughness no greater than 1 nm to ensure minimal scattering loss in the optical path and obtain highly repeatable data. During the experiment, the initial cavity ring-down time is measured first in the empty cavity (without sample insertion). Then, each optical element is inserted into the cavity center one by one, and the corresponding actual ring-down time is measured. By comparing the changes in ring-down time, the additional loss introduced by the optical element is derived, reflecting the material's absorption and scattering ability of light within the cavity at a specific thickness.

[0031] Furthermore, for step S2, a theoretical ring-down curve is established based on the time and loss model of light propagation in the ring-down cavity. Considering the multiple reflections of light between the sample and the mirror and their propagation loss, a theoretical pulse train signal model is constructed. The following workflow illustrates this: Since the actual detection system response has a certain bandwidth and time constant, the theoretical ring-down curve needs to be convolved with the time-domain response function of the photodetector to obtain a simulated detection signal close to the actual experimental environment. Subsequently, the simulated signal is fitted with a single exponential decay model using least squares to obtain the fitted theoretical ring-down time, and the theoretical loss value is further calculated.

[0032] Furthermore, for step S3, since systematic errors are unavoidable in experimental measurements, such as imperfect mirrors, scattering, and inconsistent coupling efficiencies, it is necessary to establish a correction relationship between the measured values ​​and the theoretical values. This step introduces the Gaussian Process Regression (GPR) method from the Bayesian optimization algorithm to construct a nonlinear mapping model between the theoretical loss value and the measured loss value. Through this model, the measurement results of each sample can be finely corrected to obtain a more realistic loss value that conforms to physical reality.

[0033] Furthermore, in step S4, the corrected actual loss values ​​of samples with different thicknesses are linearly fitted, with thickness as the independent variable and loss as the dependent variable. The slope of the fitted value is the volume optical loss rate per unit length. This step is based on Beer-Lambert's law in the light propagation process, which assumes that material absorption / scattering is linearly related to the optical path. This step can remove factors unrelated to thickness, such as end-face reflection, and obtain the intrinsic absorption / scattering properties of the material, which is a key physical quantity for subsequent transmittance inversion.

[0034] Furthermore, in step S5, a Physical Information Neural Network (PINN) is constructed by using the bulk optical loss rate and material optical parameters (such as refractive index, group velocity, and wavelength) as inputs. This PINN incorporates physical laws (such as Beer-Lambert's law and propagation formulas). Unlike traditional neural networks, this network introduces physical constraints during training, ensuring that the transmittance per unit length output by the network conforms to fundamental physical laws. The final output is the true transmittance of the high-transmittance material at a specified wavelength, obtained through inversion calculations, which can be used for optical design and material evaluation.

[0035] Furthermore, the planar optical elements with different thicknesses have the same surface roughness and surface quality, are double-polished, and the parallel error between the two polished surfaces of the element does not exceed 2′, and the surface roughness is not greater than 1nm.

[0036] Furthermore, the two surfaces of the planar optical elements with different thicknesses are parallel, with a thickness range of [0.5mm~20mm], and the thickness difference between different samples is greater than 0.5mm.

[0037] Furthermore, in step S1, the actual ring-down time of each planar optical element is measured, and based on the initial cavity ring-down time and the actual ring-down time of each planar optical element: the initial cavity ring-down time is obtained through pulsed cavity ring-down technology; each planar optical element is inserted into the center of the cavity one by one, and the actual ring-down time of the test cavity formed by them in the cavity is measured.

[0038] Furthermore, the output beam of the pulsed laser is a TEM. 00 The pulse width is less than 100 ns, and the response time of the photodetector is less than 100 ns.

[0039] Furthermore, the internal transmittance of the high-transmittance optical material ranges from 95.0% to 99.9999% / cm.

[0040] Furthermore, step S6 is also included: when the internal transmittance of the optical material being tested is in the range of 95.0%~99.0% / cm, the thickness range of the optical element to be used is 0.5mm~10mm; when the internal transmittance of the optical material being tested is in the range of 99.0%~99.9999% / cm, the thickness range of the optical element to be used is 2.0mm~20mm.

[0041] Furthermore, step S3 specifically includes the following sub-steps:

[0042] S301. Obtain the measured loss values ​​of multiple planar optical elements with different thicknesses in the test cavity. And the theoretical loss value calculated from the theoretical ringing time. Construct training sample pairs , wherein Indicates the thickness of the corresponding sample;

[0043] S302. Based on the input features consisting of sample thickness and theoretical loss value, a loss mapping model is established using the Gaussian process regression method to fit the measured loss value. Compared with theoretical loss value Nonlinear relationship between them;

[0044] S303. Using the trained loss mapping model, input the theoretical loss value of the thickness of a sample, predict and output the corrected actual loss value. .

[0045] Specifically, step S3 establishes a loss correction mapping model based on Bayesian optimization and Gaussian process regression to achieve high-precision conversion between theoretical predictions and actual measurements. Due to various systematic errors in the experiment, such as laser collimation errors, cavity interference, and environmental disturbances, there is a significant deviation between the theoretical loss value and the actual measurement results. Simple linear correction is insufficient to meet the accuracy requirements; therefore, Bayesian nonparametric modeling technology is used to model the error distribution. In the sample construction stage, the sample thickness and theoretical loss value are used as input vectors, and the actual measured loss rate is used as the output target, forming training sample pairs. A nonlinear mapping function is established through Gaussian process regression, and a kernel function is used to project the input feature space in a high dimension, allowing even small systematic shifts and nonlinear fluctuations to be identified and fitted in the model. This model not only outputs the corrected loss value but also includes confidence estimation and uncertainty evaluation, giving the prediction results a reliable range. In practical applications, by inputting the theoretical loss value and thickness of any new sample, the true loss estimate after model correction can be quickly obtained, enhancing the applicability of the data in subsequent inversion models.

[0046] Furthermore, step S4 specifically includes the following sub-steps:

[0047] S401. Construct a set of data pairs based on the thickness values ​​of planar optical elements of different thicknesses and the corresponding corrected actual loss values;

[0048] S402. Based on the actual loss values, including bulk optical loss rate and two-sided surface loss rate, establish a linear relationship model, which is expressed as follows:

[0049] ;

[0050] The surface loss rate includes the surface / subsurface defect absorption rate and the surface / subsurface scattering loss rate caused by surface processing, expressed as:

[0051] ;

[0052] Among them, the The volume loss rate per unit thickness is represented by the following. The thickness of the i-th sample is represented by the following: Indicates the surface / subsurface loss value, the This represents the absorption rate of surface / subsurface defects caused by surface processing. Represents the surface scattering loss rate, the Indicates the error term;

[0053] S403. Using the least squares method, the bulk optical loss rate and the surface / subsurface loss rate are jointly fitted, the model parameters are solved to obtain the best-fit bulk optical loss rate and surface loss rate, and the fitted bulk optical loss rate is output.

[0054] Furthermore, step S5 specifically includes the following sub-steps:

[0055] S501. The obtained bulk optical loss rate and the basic optical parameters of the material under test are used as model inputs;

[0056] S502. Construct a physical information neural network structure based on input variables, wherein the hidden layer of the physical information neural network structure integrates known light propagation theory relationships and Beer-Lambert law expressions as physical prior constraints to guide the network to learn the mapping relationship between loss and transmittance;

[0057] S503. The trained physical information neural network is used to predict the transmittance value per unit length of the corresponding high-transmittance optical material at a given wavelength, and the output result is used as the final measurement value of the high-transmittance optical material.

[0058] Specifically, step S5 uses the material thickness, volume loss rate, wavelength parameters, and material refractive index optical constant extracted in the preceding steps as input parameters, and introduces a physically constrained neural network structure to achieve inverse prediction of transmittance. Since there is a highly nonlinear relationship between material transmittance and multiple physical properties, and conventional models struggle to maintain universality across material conditions, a Physics-Informed Neural Network (PINN) mechanism is introduced when designing the neural network structure. By embedding the Beer-Lambert law as part of the loss function during model training, it ensures that the network output meets physical constraints, avoiding physically meaningless transmittance predictions. In its implementation, the network structure employs a multi-layer fully connected form and integrates residual modules to enhance deep feature learning capabilities. Simultaneously, Dropout and L2 regularization prevent overfitting, enabling the model to not only adapt to the current data but also generalize to other material systems. After the model is trained, its theoretical transmittance can be quickly calculated by inputting any set of optical characteristic parameters of high-transmittance materials, and the output results have significant physical consistency. It is especially suitable for high-throughput screening of new low-loss materials, verification of the optical performance of high-frequency light-transmitting elements, or evaluation of device-level light transmission efficiency, which significantly improves the efficiency of material research from experiment to engineering application.

[0059] Example 2

[0060] Furthermore, as a preferred embodiment of the above embodiments, a method for measuring the internal transmittance of high-transmittance optical materials is proposed. This method simplifies the method in Embodiment 1. Specifically: the optical element under test is inserted into the middle position of an optical resonant cavity to form a test cavity. The pitch and yaw angles of the optical element under test are adjusted so that its two parallel surfaces are perpendicular to the propagation direction of the light beam, thereby maximizing the amplitude and maximizing the decay time of the ringing-down signal in the test cavity. The ringing-down time of the test cavity is obtained by fitting the ringing-down signal detected by the photodetector according to a single exponential decay function. And calculate the measurement loss value of the optical element under test, using the following formula: ,in, To measure the loss value, , and These represent the refractive index and thickness of the optical element under test, respectively. L 0 The initial cavity length, c The speed of light;

[0061] Calculating the surface reflectivity of a high-transmission optical material using its refractive index parameter R 3 By combining the ring-down cavity structure parameters and the photodetector response characteristics theory, the measurement error of the loss value introduced by the surface reflection of the optical element under test is calculated. A mathematical relationship between the measured loss value and the actual loss value is established, and a correction factor for the measured loss value is obtained. The measured loss value is then corrected to obtain the actual loss value of the optical element under test with different thicknesses.

[0062] Example 3

[0063] Furthermore, as a preferred embodiment of the above embodiments, a step for obtaining the correction factor corresponding to the method proposed in Embodiment 2 is provided, specifically including the following steps:

[0064] Step (a): Establish the theoretical cavity ring-down signal of the pulsed cavity ring-down test cavity, considering the influence of surface reflection of the optical element under test. The expression is as follows:

[0065] ;

[0066] The following example is provided for the above content:

[0067] ;

[0068] ;

[0069] ;

[0070] ;

[0071] The The intensity of the first output pulse. This represents the number of times the pulse beam reflects back and forth within the test cavity. To input the distance between the plano-concave high-reflectivity cavity mirror and the optical element under test, This is the distance between the output concave high-reflectivity cavity mirror and the optical element under test;

[0072] When the optical element under test is located between the resonant cavities , To input the reflectivity of the plano-concave high-reflectivity cavity mirror To output the reflectivity of the plano-concave high-reflectivity cavity mirror, This represents the actual loss value of the optical element under test. (For time); the theoretical output cavity ring-down signal of the test cavity is convolved with the response function of the photodetector, and the result is fitted with a single exponential decay function to obtain the theoretical ring-down time of the test cavity. .

[0073] Step (b): The theoretical ringing time of the test cavity obtained from step (a) With formula The theoretical loss value of the optical element under test is calculated. The theoretical loss value is obtained by nonlinearly fitting the assumed actual loss value with a second-order function to obtain the theoretical loss value. ) and actual loss value ( The relationship between ) is expressed as: ,in c 0 , c 1 , c 2 This is a constant obtained from a second-order nonlinear fitting. The correction factor is determined using the relationship between the obtained theoretical loss value and the actual loss value.

[0074] The relationship between the actual loss value and the measured loss value of the tested optical material is as follows: Figure 2 As shown.

[0075] Measure the actual loss values ​​of at least two optical elements with different thicknesses. When the number of samples is 2, assume that the thicknesses of the two optical elements are respectively... and The actual optical loss values ​​are respectively and The bulk optical loss rate of the tested optical material is obtained. for (Unit: 1 / cm), internal permeability is (1- α (Unit: 1 / cm);

[0076] When the number of samples is 3 or more, a linear fit is performed on the relationship between the actual loss value and the thickness of samples with different thicknesses, such as... Figure 3 As shown, the slope of the linear fitting line is obtained. α , which is the bulk optical loss rate of the optical material being measured, and the internal transmittance (1- α (Unit: 1 / cm).

[0077] The above embodiments address the problem that existing technologies cannot accurately measure the internal transmittance (99.0%~99.9999% / cm) of highly transparent optical materials. Through embodiments two and three above, it is proposed to use pulsed cavity ring-down technology to achieve accurate measurement of the high internal transmittance of highly transparent optical materials. By measuring the optical loss of optical elements of different thicknesses made from the optical material under test, and using an improved data processing method to eliminate the influence of surface reflection on the measurement of optical loss of optical elements, accurate measurement of the high internal transmittance of highly transparent optical materials can be achieved.

[0078] In addition, Example 1 has the following advantages over Examples 2 and 3:

[0079] For the fixed quadratic polynomial mentioned in Examples 2 and 3 This is a static parameter model, which has limited adaptability to errors under different materials or measurement conditions. In contrast, the fitting model used in Example 1 does not have fixed parameters; instead, it is dynamically adjusted based on data performance during the actual measurement process. By introducing a global optimization mechanism based on probability modeling through Bayesian optimization-driven dynamic nonlinear correction, the function structure and parameters of the correction factor are automatically optimized according to the distribution of loss values ​​and experimental errors during each nonlinear fitting process. This effectively adapts to fitting deviations caused by variations in the surface reflectivity of different optical materials and changes in cavity structure.

[0080] In Examples 2 and 3, the bulk optical loss rate and internal transmittance are directly calculated by linear fitting or difference between thickness and loss value, ignoring the nonlinear influence of optical material physical properties (such as dispersion, nonlinear absorption, and local defects) on loss. In Example 1, however, a Physical Information Neural Network (PINN) is used to introduce material physical parameters, such as refractive index, bulk absorption coefficient, and scattering coefficient, and embed them into the loss rate prediction model. This simultaneously adheres to measurement data constraints and physical laws, improving the physical rationality of the model. Furthermore, the optimal thickness combination is inferred from the neural network, enhancing measurement sensitivity and responsiveness to loss changes.

[0081] Example 4

[0082] Furthermore, as a preferred embodiment of Example 2, the connection relationship of the apparatus used in the method is as follows: The pulsed laser beam output from the pulsed laser is injected into a stable optical resonant cavity composed of two plano-concave high-reflectivity cavity mirrors (input plano-concave high-reflectivity cavity mirror and output plano-concave high-reflectivity cavity mirror), referred to as the initial cavity, with a cavity length of [missing information]. An optical isolator is used to prevent the laser beam reflected from the input plano-concave high-reflectivity cavity mirror from re-entering the pulsed laser. The laser beam output from the output plano-concave high-reflectivity mirror is focused by a lens and collected by a photodetector. The initial cavity ring-down time is obtained by fitting the cavity ring-down signal measured by the photodetector to a single exponential decay function. .

[0083] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A method for measuring the internal transmittance of a high-transmittance optical material, characterized in that, Includes the following steps: S1. Prepare at least two planar optical elements with different thicknesses made of the high-transmittance optical material to be tested, measure the actual ring-down time of each planar optical element, and calculate the measured loss value based on the initial cavity ring-down time and the actual ring-down time of each planar optical element; S2. Based on the measured actual ring-down time, establish a theoretical ring-down signal model for the test cavity containing planar optical elements; convolve the theoretical ring-down signal output by the model with the time-domain response function of the photodetector to obtain the simulated detection signal; and perform least-squares fitting on the simulated detection signal according to the single exponential decay form to extract the theoretical ring-down time, and calculate the corresponding theoretical loss value based on the theoretical ring-down time. S3. Using the calculated measured loss value and theoretical loss value as input data, a loss mapping model is constructed by combining the Bayesian optimization algorithm, and the actual loss value after correction for multiple planar optical elements with different thicknesses is output through the loss mapping model. S4. Based on the actual loss values ​​after correction of multiple planar optical elements with different thicknesses, and combined with the sample thickness difference, the volume optical loss rate per unit length is extracted using a linear fitting method. S5. Using the extracted bulk optical loss rate and the basic optical parameter information of the optical material under test as input parameters, a physical information neural network model is constructed to predict transmittance inversion. The internal transmittance of the measured high-transmittance optical material is output by combining the physical information neural network model with the physical prior constraint relationship. The basic optical parameters include: material thickness, volume loss rate, wavelength parameters, and material refractive index optical constant.

2. The method for measuring the internal transmittance of a high-transmittance optical material as described in claim 1, characterized in that, The planar optical elements of different thicknesses have the same surface roughness and surface quality, are double-polished and the parallel error between the two polished surfaces of the element does not exceed 2′, and the surface roughness is not greater than 1nm.

3. The method for measuring the internal transmittance of a high-transmittance optical material as described in claim 1, characterized in that, The two surfaces of the planar optical elements with different thicknesses are parallel, and the thickness ranges from 0.5mm to 20mm. The thickness difference between different samples is greater than 0.5mm.

4. The method for measuring the internal transmittance of a high-transmittance optical material as described in claim 1, characterized in that, In step S1, the specific process for measuring the actual ring-down time of each planar optical element is as follows: the initial cavity ring-down time is obtained through pulsed cavity ring-down technology; each planar optical element is inserted into the center of the cavity one by one, and the actual ring-down time of the test cavity formed by them in the cavity is measured.

5. The method for measuring the internal transmittance of a high-transmittance optical material as described in claim 4, characterized in that: The output beam of the pulsed laser is TEM. 00 The pulse width is less than 100 ns, and the response time of the photodetector is less than 100 ns.

6. The method for measuring the internal transmittance of a high-transmittance optical material according to claim 1, characterized in that: The internal transmittance of the high-transmittance optical material ranges from 95.0% to 99.9999% / cm.

7. The method for measuring the internal transmittance of a high-transmittance optical material as described in claim 1, characterized in that, It also includes step S6: when the internal transmittance of the optical material being tested is in the range of 95.0%~99.0% / cm, the thickness range of the optical element to be used is 0.5mm~10mm; when the internal transmittance of the optical material being tested is in the range of 99.0%~99.9999% / cm, the thickness range of the optical element to be used is 2.0mm~20mm.

8. The method for measuring the internal transmittance of a high-transmittance optical material as described in claim 1, characterized in that, Step S3 specifically includes the following sub-steps: S301. Obtain the measured loss values ​​of multiple planar optical elements with different thicknesses in the test cavity. And the theoretical loss value calculated from the theoretical ringing time. Construct training sample pairs , wherein Indicates the thickness of the corresponding sample; S302. Based on the input features consisting of sample thickness and theoretical loss value, a loss mapping model is established using the Gaussian process regression method to fit the measured loss value. Compared with theoretical loss value Nonlinear relationship between them; S303. Using the trained loss mapping model, input the theoretical loss value of the thickness of a sample, predict and output the corrected actual loss value. .

9. The method for measuring the internal transmittance of a high-transmittance optical material as described in claim 1, characterized in that, Step S4 specifically includes the following sub-steps: S401. Construct a set of data pairs based on the thickness values ​​of planar optical elements of different thicknesses and the corresponding corrected actual loss values; S402. Based on the actual loss values, including bulk optical loss rate and two-sided surface loss rate, establish a linear relationship model, which is expressed as follows: ； The surface loss rate includes the surface / subsurface defect absorption rate and the surface / subsurface scattering loss rate caused by surface processing, expressed as: ； Among them, the The volume loss rate per unit thickness is represented by the following. The thickness of the i-th sample is represented by the following: Indicates the surface / subsurface loss value, the This represents the absorption rate of surface / subsurface defects caused by surface processing. Represents the surface scattering loss rate, the Indicates the error term; S403. Using the least squares method, the bulk optical loss rate and the surface / subsurface loss rate are jointly fitted, the model parameters are solved to obtain the best-fit bulk optical loss rate and surface / subsurface loss rate, and the fitted bulk optical loss rate is output.

10. The method for measuring the internal transmittance of a high-transmittance optical material as described in claim 1, characterized in that, Step S5 specifically includes the following sub-steps: S501. The obtained bulk optical loss rate and the basic optical parameters of the material under test are used as model inputs; S502. Construct a physical information neural network structure based on input variables, wherein the hidden layer of the physical information neural network structure integrates known light propagation theory relationships and Beer-Lambert law expressions as physical prior constraints to guide the network to learn the mapping relationship between loss and transmittance; S503. The trained physical information neural network is used to predict the transmittance value per unit length of the corresponding high-transmittance optical material at a given wavelength, and the output result is used as the final measurement value of the high-transmittance optical material.

Images