Spectrum detection method and system for modified ultra-high molecular plastic

Abstract

The invention provides a spectrum detection method and system for modified ultra-high molecular plastic, and relates to the technical field of high molecular spectrum detection. The method comprises the following steps: firstly, constructing a scattering correction model based on scattering characteristics of an ultraviolet-visible spectrum in a non-absorption wave band, and carrying out accurate baseline correction on an infrared spectrum to obtain an intrinsic infrared absorption spectrum; then analyzing the intrinsic infrared absorption spectrum to obtain a relative content parameter of a modifier component, and introducing the relative content parameter as a constraint condition into an unmixing process of a Raman spectrum so as to effectively separate and extract a Raman characteristic signal of the modifier. And finally, carrying out complementary analysis and cross validation on the intrinsic infrared absorption spectrum and the Raman characteristic signal to determine key modification structure information such as chemical bonding state and spatial distribution of the modifier in the material. The method is deployed in a spectrum detection sensor, and high-sensitivity and high-reliability detection and analysis of low-content or complex-distribution modified components in the modified ultra-high molecular plastic can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of polymer spectral detection technology, and more specifically, to a spectral detection method and system for modified ultrapolymer plastics. Background Technology

[0002] Currently, the performance optimization and functional modification of polymer materials are of paramount importance for materials research and development, process control, and product quality assurance. Traditionally, the analysis of modified polymer materials has relied primarily on single spectroscopic techniques. Infrared spectroscopy is widely used for functional group identification and quantitative analysis. However, traditional baseline correction methods are highly dependent on operator experience, lack objective physical evidence, and suffer from poor repeatability of correction results. This severely interferes with the subsequent identification and quantification accuracy of weak absorption signals, such as those with low-content modifiers.

[0003] Furthermore, while Raman spectroscopy offers low sample preparation requirements, high spatial resolution, and complementary information with infrared spectroscopy, the Raman signals of modifiers are often weak when analyzing complex modified systems. These signals are easily masked or overlapped by strong fluorescence backgrounds and strong signals from the matrix polymer. Direct spectral analysis is extremely difficult. Commonly used unmixing algorithms (such as principal component analysis (PCA) and independent component analysis (ICA) often produce non-unique unmixing results in the absence of effective prior information, making it difficult to accurately attribute and separate the pure spectral signals of specific modifiers.

[0004] To overcome the limitations of single technologies, in recent years, the industry has begun to explore strategies for combining multiple spectra or data fusion, such as performing measurements at the same location using infrared microscopy and Raman spectrometry; or collecting multiple spectral data separately and then processing them together using multivariate statistical methods (such as multivariate curve resolution). However, most of these approaches remain at the level of "parallel data processing" or "post-hoc correlation and comparison," failing to achieve deep collaboration and real-time feedback between different spectral data streams in the preprocessing and core resolution stages. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a spectral detection method and system for modified ultra-high molecular weight plastics, aiming to solve the subjectivity and instability of traditional empirical baseline correction.

[0006] In a first aspect, the present invention provides a method for spectroscopic detection of modified ultrapolymer plastics, comprising:

[0007] Acquire the ultraviolet-visible spectral data, infrared spectral data, and Raman spectral data of the sample to be tested;

[0008] A scattering correction model is constructed based on the ultraviolet-visible spectral data, and the infrared spectral data is baseline-corrected using the scattering correction model to obtain the intrinsic infrared absorption spectrum.

[0009] The intrinsic infrared absorption spectrum is analyzed to determine the relative content parameters of the modifier components. These relative content parameters are then used as constraints in the Raman spectroscopy unmixing process to extract the Raman characteristic signals of the modifier from the Raman spectral data.

[0010] Based on the intrinsic infrared absorption spectrum and the Raman characteristic signal, the modified structural information of the sample to be tested is determined.

[0011] Furthermore, the ultraviolet-visible spectral data, infrared spectral data, and Raman spectral data are from the same detection area of ​​the sample to be tested or adjacent areas with the same representativeness.

[0012] Furthermore, the construction of the scattering correction model includes:

[0013] The wavelength range in the ultraviolet-visible spectral data where the absorbance is below a preset threshold is identified as the non-absorption band.

[0014] Extract the attenuation trend of the spectral signal in the non-absorption band as a function of wavelength, and use the attenuation trend as a response feature;

[0015] Based on the attenuation trend, a function model is established with the wavelength of the ultraviolet-visible spectral data as the independent variable and the absorbance value of the non-absorption band as the dependent variable, which serves as a scattering correction model for baseline correction of the infrared spectral data.

[0016] Furthermore, the baseline correction of the infrared spectral data using the scattering correction model to obtain the intrinsic infrared absorption spectrum includes:

[0017] The functional relationship of the scattering correction model is mapped to the wavenumber domain of the infrared spectral data through coordinate transformation to obtain the mapped functional relationship;

[0018] Based on the mapped functional relationship, the infrared scattering background baseline is calculated and generated;

[0019] The infrared spectral data is differentially analyzed with the infrared scattering background baseline to obtain the baseline-corrected spectral signal, which is used as the intrinsic infrared absorption spectrum.

[0020] Further, the step of mapping the functional relationship of the scattering correction model to the wavenumber domain of the infrared spectral data through coordinate transformation includes:

[0021] Based on the inverse relationship between wavelength and wavenumber, the functional relationship of the scattering correction model is converted into a functional model with the wavenumber of the infrared spectrum as the independent variable, which is then used as the mapped functional relationship.

[0022] Furthermore, the step of analyzing the intrinsic infrared absorption spectrum to determine the relative content parameters of the modifier components includes:

[0023] The characteristic variables of the intrinsic infrared absorption spectrum are extracted, and the characteristic variables include at least the characteristic absorption signal intensity of the modifier and the polymer matrix of the sample to be tested;

[0024] The characteristic variables are input as input variables into the chemometrics quantitative model. The chemometrics quantitative model projects the input variables to obtain latent variables and generates regression coefficients through a partial least squares regression algorithm.

[0025] The relative content parameters of the modifier components are obtained by performing linear regression on the latent variables using regression coefficients.

[0026] Furthermore, the extraction of characteristic variables from the intrinsic infrared absorption spectrum, wherein the characteristic variables include at least the characteristic absorption signal intensities of the modifier and the polymer matrix of the sample to be tested, including:

[0027] The characteristic spectral bands of the modifier and the matrix polymer of the sample to be tested are identified in the intrinsic infrared absorption spectrum, and the intensity of the characteristic absorption signal is obtained by quantifying the signal amplitude of the characteristic spectral bands.

[0028] Furthermore, the step of introducing the relative content parameter as a constraint condition into the Raman spectroscopy unmixing process, and extracting the Raman characteristic signal of the modifier from the Raman spectral data, includes:

[0029] The relative content parameter is used as the initialization condition for the Raman spectrum unmixing process. The Raman spectrum data is iteratively unmixed using a non-negative matrix factorization algorithm to separate the pure Raman spectra of each chemical component.

[0030] Based on the characteristic variables of the intrinsic infrared absorption spectrum, the spectral matching benchmark of the modifier is determined;

[0031] The spectral matching reference includes: the wavenumber range in which the characteristic absorption peak of the modifier is identified in the intrinsic infrared absorption spectrum, and the relative proportion of the characteristic absorption signal intensity of the modifier and the matrix polymer in the wavenumber range;

[0032] The degree of fit is calculated based on the pure Raman spectra of each chemical component and the spectral matching benchmark. Pure Raman spectra with a degree of fit reaching a preset threshold are extracted as the Raman characteristic signals of the modifier.

[0033] Furthermore, determining the modified structural information of the sample to be tested based on the intrinsic infrared absorption spectrum and the Raman characteristic signal includes:

[0034] Based on the complementarity and correlation between the intrinsic infrared absorption spectrum and the Raman characteristic signal, cross-validation is performed to determine the modified structural information of the modifier in the sample to be tested. The modified structural information includes at least: chemical bonding state and spatial distribution information.

[0035] In a second aspect, the present invention provides a spectroscopic detection system for modified ultra-high molecular weight plastics, used to implement the spectroscopic detection method for modified ultra-high molecular weight plastics described in the first aspect, comprising:

[0036] The data acquisition unit is used to acquire ultraviolet-visible spectral data, infrared spectral data, and Raman spectral data of the sample to be tested;

[0037] The correction unit is used to construct a scattering correction model based on the ultraviolet-visible spectral data, and to perform baseline correction on the infrared spectral data using the scattering correction model to obtain the intrinsic infrared absorption spectrum.

[0038] The demixing unit is used to analyze the intrinsic infrared absorption spectrum, determine the relative content parameters of the modifier components, and introduce the relative content parameters as constraints into the Raman spectroscopy demixing process to extract the Raman characteristic signals of the modifier from the Raman spectral data.

[0039] The output unit is used to determine the modified structure information of the sample under test based on the intrinsic infrared absorption spectrum and the Raman characteristic signal.

[0040] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0041] This invention utilizes the physical scattering response characteristics of the ultraviolet-visible spectrum in the non-absorption band to construct a scattering correction model, and uses this model to perform baseline correction on the infrared spectrum. This overcomes the subjectivity and instability of traditional empirical baseline correction, and can more accurately remove the scattering background to obtain a pure intrinsic infrared absorption spectrum. By first analyzing the corrected infrared spectrum, the relative content parameters of the modifier can be reliably obtained, and these parameters are introduced as key prior constraints into unmixing algorithms such as nonnegative matrix factorization of Raman spectroscopy. This effectively avoids unmixing ambiguity or non-convergence problems caused by a lack of information in the algorithm.

[0042] This invention, through the deep fusion and mutual verification of complementary information provided by infrared spectroscopy and Raman spectroscopy, can more comprehensively and reliably determine key modified structural information such as the chemical bonding state, dispersion uniformity, and phase interface interaction of the modifier in the plastic matrix, and significantly improves the confidence level of the analysis results. Attached Figure Description

[0043] Figure 1 A flowchart illustrating a spectroscopic detection method for modified ultrapolymer plastics provided in an embodiment of the present invention;

[0044] Figure 2 A flowchart illustrating the baseline correction method based on ultraviolet and infrared spectroscopy provided in an embodiment of the present invention;

[0045] Figure 3 This is a flowchart of the Raman spectroscopy demixing method provided in an embodiment of the present invention;

[0046] Figure 4 This is a schematic diagram of a spectroscopic detection system for modified ultra-high molecular weight plastics provided in an embodiment of the present invention. Detailed Implementation

[0047] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0048] It should be noted that modified polymer plastics are formed by adding various modifiers to optimize or functionalize the performance of the matrix resin. The modifiers include: reinforcing fillers, toughening agents, flame retardants, compatibilizers, stabilizers, and functional additives.

[0049] Spectroscopic detection enables qualitative and quantitative analysis of modified polymer plastics, including: identifying and confirming the chemical composition and functional group structure of modifiers and matrix, determining the relative content and distribution uniformity of modifiers in the matrix, and elucidating the interaction mechanism between modifiers and matrix, such as chemical bonding, physical adsorption, or phase separation.

[0050] Example 1

[0051] See Figure 1 This embodiment provides a spectroscopic detection method for modified ultra-high molecular weight plastics, the method comprising steps S101 to S104, wherein:

[0052] S101: Acquire the ultraviolet-visible spectral data, infrared spectral data, and Raman spectral data of the sample to be tested;

[0053] S102: Construct a scattering correction model based on the ultraviolet-visible spectral data, and use the scattering correction model to perform baseline correction on the infrared spectral data to obtain the intrinsic infrared absorption spectrum;

[0054] S103: Analyze the intrinsic infrared absorption spectrum to determine the relative content parameters of the modifier components;

[0055] S104: The relative content parameter is introduced as a constraint condition into the Raman spectroscopy unmixing process to extract the Raman characteristic signal of the modifier from the Raman spectral data;

[0056] S105: Based on the intrinsic infrared absorption spectrum and the Raman characteristic signal, determine the modified structure information of the sample to be tested.

[0057] Regarding step S101:

[0058] As is generally known, spectroscopic detection primarily utilizes ultraviolet-visible spectroscopy, infrared spectroscopy, and Raman spectroscopy. Ultraviolet-visible spectroscopy data reflects molecular electronic energy level information; its response in the non-absorption band is mainly determined by the physical scattering characteristics of the sample, and the abscissa of the data is wavelength. Infrared spectroscopy reflects molecular vibrational energy level information and is the primary means of identifying functional groups and performing quantification; its abscissa is typically wavenumber. Raman spectroscopy also reflects molecular vibrational information, but it is complementary to infrared spectroscopy, has lower requirements for sample preparation, and offers high spatial resolution, although it is susceptible to fluorescence interference.

[0059] In this embodiment, the ultraviolet-visible spectral data, infrared spectral data, and Raman spectral data are from the same detection area of ​​the sample to be tested or adjacent areas with the same representativeness.

[0060] For example, a representative detection area is selected on the sample surface.

[0061] As an optional implementation method, for samples with known modification methods, regions that can reflect the overall average composition of the material are preferentially selected. At the same time, regions with stable spectral characteristics that can represent the main chemical components of the sample are selected as fixed detection locations to ensure that the local information of the analysis points can be effectively extrapolated to the overall evaluation of the sample.

[0062] First, complete the ultraviolet-visible spectral acquisition of the area. Then, in the same area or adjacent areas that have been confirmed to have the same appearance and composition characteristics, complete the infrared and Raman spectral acquisition in sequence.

[0063] In practice, preprocessing is performed separately for the three different types of spectra, based on their respective physical principles and main sources of interference.

[0064] As an optional implementation, the main part of the UV-Vis spectral preprocessing is noise smoothing.

[0065] For example, the Savitzky-Golay convolution smoothing method is used to suppress random noise while preserving spectral contour features.

[0066] Wavelength calibration is performed on the ultraviolet-visible spectrum to ensure that the wavelength axes of data collected by different instruments or different batches are precisely aligned.

[0067] As an optional implementation, infrared spectral preprocessing is performed to remove environmental background, such as removing the characteristic absorption of carbon dioxide and water vapor, and signal averaging is also performed to improve the signal-to-noise ratio.

[0068] The spectrum is truncated and focused on key wavenumber ranges containing the characteristic absorptions of the modifier and the matrix polymer.

[0069] As an alternative implementation, Raman spectral preprocessing is performed by employing an adaptive reweighted penalized least squares method to estimate and subtract the slowly changing baseline generated by sample fluorescence, thereby highlighting the Raman scattering signal;

[0070] The Raman spectrum is subjected to cosmic ray peak removal. An algorithm is used to automatically identify sharp peaks with abnormally high intensity and extremely narrow spectral width, and the abnormal data points are replaced by interpolation of the surrounding data points, thereby eliminating the interference of such transient noise.

[0071] Regarding step S102:

[0072] See Figure 2 This embodiment provides a method for baseline correction of infrared spectral data based on the ultraviolet-visible spectral data, including:

[0073] S201: Identify the non-absorption bands in ultraviolet-visible spectral data and establish a function model with wavelength as the independent variable based on its response characteristics as a scattering correction model;

[0074] S202: Map the functional relationship of the scattering correction model from the wavelength domain of the ultraviolet-visible spectrum to the wavenumber domain of the infrared spectrum;

[0075] S203: Based on the functional relationship mapped to the wavenumber domain, calculate and generate the scattering background baseline for the infrared spectral data;

[0076] S204: Perform a differential operation between the infrared spectral data and the calculated scattering background baseline to achieve differential correction of the infrared spectral data;

[0077] S205: Use the spectral signal obtained after differential correction as the intrinsic infrared absorption spectrum.

[0078] In specific implementation, wavelength ranges in the ultraviolet-visible spectral data with absorbance below a preset threshold are identified as non-absorption bands;

[0079] For example, by performing first or second derivative calculations on the ultraviolet-visible spectrum, a broad and gentle absorption peak can be distinguished from a gently sloping background; wherein the absorption characteristics are zero-crossing points or extreme points, while the slow changing trend caused by scattering is significantly suppressed into a flat region close to zero; the continuous wavelength range in which the derivative signal fluctuates slightly near zero and the original spectral absorbance value is relatively low and changes gently is regarded as the non-absorption band.

[0080] The preset threshold can be limited based on the chemical structure of the modified ultra-high molecular weight plastic matrix and the known modifier.

[0081] Extract the attenuation trend of the spectral signal in the non-absorption band as a function of wavelength, and use the attenuation trend as a response feature;

[0082] Based on the attenuation trend, a function model is established with the wavelength of the ultraviolet-visible spectral data as the independent variable and the absorbance value of the non-absorption band as the dependent variable; this model serves as a scattering correction model for baseline correction of the infrared spectral data.

[0083] For example, a function model of the wavelength and the absorbance value is constructed using a least squares fitting algorithm;

[0084] In specific implementation, based on the morphological characteristics of the ultraviolet-visible spectrum, a function form is selected for fitting. The fitting function includes power function form, exponential function form, or polynomial form of second order or lower.

[0085] The coefficients of each term in the function are calculated by the least squares method, thereby obtaining an explicit functional expression describing the relationship between the scattering attenuation trend of the sample under test and the wavelength, which is the completed scattering correction model.

[0086] In a specific implementation, based on the functional relationship of the scattering correction model, the coordinate transformation is used to map the infrared spectral data to the wavenumber domain to obtain the mapped functional relationship;

[0087] As an optional implementation, based on the conversion relationship between wavelength and wavenumber, the scattering correction model is converted into a function model with the wavenumber of the infrared spectrum as the independent variable;

[0088] Ultraviolet-visible spectra are expressed using wavelength, while infrared spectra are expressed using wavenumber.

[0089] Wavelength describes the spatial periodicity of light waves, while wavenumber describes the frequency characteristics of light. The conversion relationship between wavelength and wavenumber is inversely proportional.

[0090] Based on the aforementioned inverse relationship, the wavelength, which is used as the independent variable in the model function, is replaced with the wave number;

[0091] A new functional expression with infrared spectral wavenumber as the independent variable is obtained, which serves as a scattering correction model mapped to the wavenumber domain of infrared spectral data.

[0092] Based on the scattering correction model mapped to the wavenumber domain of infrared spectral data, an infrared scattering background baseline is calculated and generated;

[0093] The infrared spectral data is differentially analyzed with the infrared scattering background baseline to obtain the baseline-corrected spectral signal, which is used as the intrinsic infrared absorption spectrum.

[0094] In practice, the wavenumber range of the infrared spectral data to be processed is determined;

[0095] The starting and ending values ​​of the wavenumber range are defined based on one or more of the following methods: preset standard spectral characteristic regions, prior knowledge of the sample's chemical structure, and signal-to-noise ratio evaluation of the original data.

[0096] As an optional implementation, a series of discrete, uniformly spaced wavenumber points are generated within the wavenumber range as a sequence of abscissas.

[0097] For example, the wavenumber value of each wavenumber point is used as an input parameter and substituted into the scattering correction model mapped to the wavenumber domain of infrared spectral data.

[0098] The function of the scattering correction model mapped to the wavenumber domain of the infrared spectral data calculates the predicted intensity value corresponding to each wavenumber.

[0099] All calculated wavenumber values ​​and their corresponding predicted intensity values ​​are connected in wavenumber order to form a smooth curve that is either continuous or segmented, which serves as the calculated infrared scattering background baseline.

[0100] In practice, the difference between the infrared spectral data and the infrared scattering background baseline is calculated.

[0101] Ensure that the wavenumber coordinate point sequence of the infrared spectral data is completely consistent with the wavenumber coordinate point sequence of the infrared scattering background baseline;

[0102] If the infrared spectral data and the wavenumber coordinate point sequence of the infrared scattering background baseline are inconsistent, interpolation or resampling processing is performed on one of the data sets to make the wavenumber coordinate points of the two correspond one-to-one.

[0103] As an optional implementation, an arithmetic subtraction operation is performed for each corresponding wavenumber coordinate point;

[0104] Subtract the predicted intensity value of the infrared scattering background baseline at the same wavenumber coordinate point from the original measured intensity value of the infrared spectral data at the wavenumber coordinate point.

[0105] The subtraction operation is performed sequentially on all coordinate points within the wavenumber range.

[0106] For example, after performing the subtraction operation on all wavenumber coordinate points, a new set of data pairs is obtained;

[0107] The new data pair consists of wavenumber coordinate values ​​and their corresponding intensity values ​​after differential operations;

[0108] The set of the new data pairs is output as the baseline-corrected spectral signal, which is the intrinsic infrared absorption spectrum.

[0109] Regarding step S103:

[0110] It is well known that qualitative analysis based on infrared and Raman spectroscopy, by comparing with standard spectral libraries or relying on the known absorption / scattering peak positions of characteristic functional groups, can identify and confirm the chemical composition and structure of the matrix polymer and various modifiers. Furthermore, quantitative analysis can be performed by establishing mathematical models of spectral response and component concentration.

[0111] In this embodiment, characteristic variables of the intrinsic infrared absorption spectrum are extracted and quantitatively analyzed.

[0112] As an optional implementation, the characteristic variables include at least: the characteristic absorption signal intensity of the modifier and the polymer matrix of the sample to be tested;

[0113] The characteristic variables may also include reference parameters for correcting spectral shift and deformation, wherein the reference parameters are selected from the peak position or peak width of the stable characteristic peaks of the matrix polymer;

[0114] For example, the characteristic bands of the modifier and the matrix polymer of the sample to be tested are identified in the intrinsic infrared absorption spectrum, and the intensity of the characteristic absorption signal is obtained by quantifying the signal amplitude of the characteristic bands;

[0115] The quantization method for the signal amplitude includes selecting the absorbance value at a specific wavenumber point or calculating the integral value of the peak area of ​​the characteristic spectral band.

[0116] In practice, the identification of the characteristic spectral bands is carried out through one or more of the following methods: comparison of known chemical structures with standard spectral libraries, empirical assignment of vibrational frequencies of characteristic functional groups, and differential spectral analysis of control samples.

[0117] The comparison of the known chemical structure with the standard spectrum library is based on the pre-defined chemical information of the matrix polymer and modifier, and the peak position, peak shape and relative intensity are matched with the authoritative infrared standard spectrum database.

[0118] The empirical attribution of the vibrational frequencies of the characteristic functional groups is based on the characteristic frequency table of functional groups in infrared spectroscopy, and the spectrum is checked to see if it matches the expected functional group absorption in a specific wavenumber range.

[0119] The differential spectral analysis of the control sample used a pure matrix polymer as a control, and the characteristic absorption bands mainly caused by the modifier were highlighted by computer differential spectral processing.

[0120] As an optional implementation, after successfully identifying the characteristic spectral band, the signal amplitude of the characteristic spectral band is quantized;

[0121] The quantization is carried out by selecting the absorbance value of the characteristic spectral band at a specific wavenumber point, or by calculating the peak area integral value of the characteristic spectral band within a defined wavenumber interval;

[0122] The intensity of the characteristic absorption signal is obtained through the quantization operation.

[0123] As an optional implementation, the feature variables are input as input variables into the chemometrics quantitative model;

[0124] The chemometric quantitative model projects the input variables to obtain latent variables using a partial least squares regression algorithm, and the chemometric quantitative model includes regression coefficients;

[0125] In specific implementation, the partial least squares regression algorithm iteratively finds the direction with the largest covariance between the feature variable space and the content data space and projects it onto the original high-dimensional feature variables to transform them into fewer and more independent latent variables.

[0126] The regression coefficients are obtained during the model training phase by minimizing the prediction error.

[0127] The training phase uses a standard sample dataset with known modifier content, and the regression coefficients define a linear mapping relationship between latent variables and target content.

[0128] For example, during the model application phase, the feature variable data of the sample to be tested are transformed into the same latent variable space according to the same projection direction determined during the training phase.

[0129] As an optional implementation, the relative content parameters of the modifier components are obtained by performing linear regression on the latent variables using regression coefficients.

[0130] For example, linear regression is performed by linearly combining the latent variables obtained from the transformation with the regression coefficients;

[0131] The calculation result of the linear regression operation is output as the relative content parameter of the modifier component, which characterizes the relative concentration ratio of the modifier in the sample to be tested.

[0132] Regarding step S104:

[0133] See Figure 3 Here is a flowchart of the Raman spectroscopy demixing method provided in this embodiment, wherein:

[0134] The relative content parameter is used as the initialization condition for the Raman spectrum unmixing process. The Raman spectrum data is iteratively unmixed using a non-negative matrix factorization algorithm to separate the pure Raman spectra of each chemical component.

[0135] The Raman spectroscopy unmixing process is a calculation process that decomposes the Raman spectral data into several pure spectra representing a single chemical component and their corresponding concentration distributions.

[0136] The nonnegative matrix factorization algorithm is used for iterative unmixing. During the iteration process, the nonnegativity constraint of spectral intensity and concentration distribution is enforced. The pure spectral matrix and concentration distribution matrix are updated by alternately minimizing the reconstruction error.

[0137] In practice, the iteration process includes convergence assessment;

[0138] For example, when the change in the objective function value between two consecutive iterations is less than the preset tolerance or the maximum number of iterations is reached, the algorithm terminates and outputs a pure Raman spectrum.

[0139] As an optional implementation, the spectral matching reference of the modifier is determined based on the characteristic variables of the intrinsic infrared absorption spectrum;

[0140] The spectral matching criteria include: the wavenumber range in which the characteristic absorption peak of the modifier is identified in the intrinsic infrared absorption spectrum, and the relative proportion of the characteristic absorption signal intensity of the modifier and the matrix polymer within the wavenumber range.

[0141] In specific implementation, the degree of fit between the pure Raman spectra of each chemical component and the spectral matching benchmark is compared, and the pure Raman spectra with a degree of fit reaching a preset threshold are extracted as the Raman characteristic signals of the modifier.

[0142] As an optional implementation, the fit is quantified by calculating the degree of matching between the signal distribution pattern of the pure Raman spectrum in the wavenumber range and the relative proportion;

[0143] As an optional implementation, the correlation coefficient or cosine similarity between the pure Raman spectrum and the spectral matching reference can be calculated;

[0144] Set a preset threshold, which is determined based on empirical statistical values ​​from a large number of simulated unmixing experiments on known standard samples, or is set through theoretical analysis;

[0145] Candidate pure Raman spectra of candidates whose fit calculation values ​​reach or exceed the preset threshold are extracted as Raman characteristic signals of the modifier.

[0146] Regarding step S105:

[0147] Perform a spectral verification operation on the extracted Raman feature signal to obtain the spectral verification result;

[0148] As an optional implementation, the Raman characteristic signal is cross-compared with the Raman spectrum of the modifier to calculate their spectral similarity index;

[0149] For example, the spectral similarity index is obtained by calculating the correlation coefficient and root mean square error between the Raman feature signal and the standard Raman spectrum of the modifier, and then normalizing and weighting them.

[0150] The spectral similarity index is used to quantify the overall degree of agreement between the currently extracted signal and the standard reference spectrum; the higher the value, the better the consistency.

[0151] If the spectral verification result does not reach the preset confidence threshold, the parameters of the Raman spectral unmixing process are adjusted, and the unmixing is performed again.

[0152] The preset credibility threshold is set based on the weighted score of the comparison and inspection items;

[0153] In practice, if the spectral verification result does not reach the preset confidence threshold, the current unmixing result is determined to be unreliable. At this time, the initialization parameters or regularization parameters of the non-negative matrix factorization algorithm are adjusted, and the iterative unmixing process is restarted until a verified Raman feature signal is obtained or the maximum number of repetitions is reached.

[0154] Based on the complementarity and correlation between the intrinsic infrared absorption spectrum and the Raman characteristic signal, cross-validation is performed to determine the modified structural information of the modifier in the sample to be tested. The modified structural information includes at least: chemical bonding state and spatial distribution information.

[0155] In specific implementation, the modified structure information of the sample to be tested is determined based on the intrinsic infrared absorption spectrum and the Raman characteristic signal.

[0156] The determination process is based on cross-validation of the complementarity of the intrinsic infrared absorption spectrum and the Raman characteristic signal in terms of molecular vibrational information and their correlation in terms of chemical origin.

[0157] As an optional implementation, the cross-validation includes analysis of the chemical bonding state;

[0158] Specifically, the peak position and shape of the characteristic functional group absorption peaks of the modifier in the intrinsic infrared absorption spectrum are compared with the peak position, peak shape and relative intensity of the corresponding vibrational modes in the Raman characteristic signal;

[0159] If the two are highly consistent in their corresponding characteristics and there is no significant shift or new peak, then it is determined that the modifier and the matrix are mainly physically blended or have weak interactions.

[0160] If a new vibrational peak corresponding to the possible formation of chemical bonds appears in the Raman characteristic signal, or if there is a systematic shift or change in peak shape compared with the infrared characteristic peak, then the presence of a specific chemical bonding effect can be inferred by combining the information from both.

[0161] For example, the cross-validation also includes the analysis of spatial distribution information;

[0162] Based on the Raman characteristic signal and the concentration distribution information obtained synchronously during the unmixing process, a two-dimensional or three-dimensional distribution map of the modifier in the micro-region of the sample to be tested is generated.

[0163] Simultaneously, the intrinsic infrared absorption spectra collected from the same point or region are analyzed, and the overall average chemical information reflected by them is correlated with the Raman distribution map;

[0164] By comparing the uniformity and aggregation morphology of the Raman distribution map with the stability of the overall infrared spectrum signal, the dispersion uniformity of the modifier in the sample, the presence of phase separation or local enrichment, and thus comprehensively determining its spatial distribution information.

[0165] It should be noted that in specific implementations, the present invention can be deployed or integrated into a smart sensor, which can be designed as a device integrating a multispectral light source (such as an LED or laser diode), a detector array, an embedded processor, and a communication unit.

[0166] For example, the method provided by the present invention can run in an embedded system inside the sensor, and acquire spectra by controlling the start and stop of the multispectral light source and adjusting its parameters; perform preprocessing of spectral signals, feature extraction and modifier calculation locally on the sensor, and output structured modifier content and distribution information.

[0167] Example 2

[0168] Based on the same inventive concept, this embodiment provides a spectroscopic detection system for modified ultra-high molecular weight plastics, used to implement the spectroscopic detection method for modified ultra-high molecular weight plastics described in Embodiment 1. Since the principle of solving the problem in this embodiment is similar to the spectroscopic detection method for modified ultra-high molecular weight plastics described in Embodiment 1, the implementation of the system in this embodiment can refer to the implementation of the method in Embodiment 1, and the repeated parts will not be described again.

[0169] See Figure 4This embodiment provides a spectroscopic detection system for modified ultra-high molecular weight plastics, including: a data acquisition unit 10, a calibration unit 20, a demixing unit 30, and an output unit 40, wherein:

[0170] The data acquisition unit 10 is used to acquire the ultraviolet-visible spectral data, infrared spectral data and Raman spectral data of the sample to be tested;

[0171] The correction unit 20 is used to construct a scattering correction model based on the ultraviolet-visible spectral data, and to perform baseline correction on the infrared spectral data using the scattering correction model to obtain the intrinsic infrared absorption spectrum.

[0172] The demixing unit 30 is used to analyze the intrinsic infrared absorption spectrum, determine the relative content parameter of the modifier component, and introduce the relative content parameter as a constraint condition into the Raman spectroscopy demixing process to extract the Raman characteristic signal of the modifier from the Raman spectral data.

[0173] Output unit 40 is used to determine the modified structure information of the sample to be tested based on the intrinsic infrared absorption spectrum and the Raman characteristic signal.

[0174] Those skilled in the art will understand that, in the methods described above in the specific embodiments, the order in which the steps are written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic. It should be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0175] In the description of this specification, the terms "exemplary," "for example," "specifically," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

Claims

1. A spectroscopic detection method for modified ultrapolymer plastics, characterized in that, The method includes: Acquire the ultraviolet-visible spectral data, infrared spectral data, and Raman spectral data of the sample to be tested; A scattering correction model is constructed based on the ultraviolet-visible spectral data, and the infrared spectral data is baseline-corrected using the scattering correction model to obtain the intrinsic infrared absorption spectrum. The intrinsic infrared absorption spectrum was analyzed to determine the relative content parameters of the modifier components; The relative content parameter is introduced as a constraint condition into the Raman spectroscopy unmixing process to extract the Raman characteristic signal of the modifier from the Raman spectral data. Based on the intrinsic infrared absorption spectrum and the Raman characteristic signal, the modified structural information of the sample to be tested is determined.

2. The spectroscopic detection method for modified ultrapolymer plastics according to claim 1, characterized in that, The ultraviolet-visible spectral data, infrared spectral data, and Raman spectral data are from the same detection area of ​​the sample to be tested or adjacent areas with the same representativeness.

3. The spectroscopic detection method for modified ultrapolymer plastics according to claim 1, characterized in that, The construction of the scattering correction model includes: The wavelength range in the ultraviolet-visible spectral data where the absorbance is below a preset threshold is identified as the non-absorption band. Extract the attenuation trend of the spectral signal in the non-absorption band as a function of wavelength, and use the attenuation trend as a response feature; Based on the attenuation trend, a function model is established with the wavelength of the ultraviolet-visible spectral data as the independent variable and the absorbance value of the non-absorption band as the dependent variable, serving as a scattering correction model for baseline correction of infrared spectral data.

4. The spectroscopic detection method for modified ultrapolymer plastics according to claim 1, characterized in that, The infrared spectral data were baseline-corrected using the scattering correction model to obtain the intrinsic infrared absorption spectrum, which includes: The functional relationship of the scattering correction model is mapped to the wavenumber domain of the infrared spectral data through coordinate transformation to obtain the mapped functional relationship; Based on the mapped functional relationship, the infrared scattering background baseline is calculated and generated; The infrared spectral data is differentially analyzed with the infrared scattering background baseline to obtain the baseline-corrected spectral signal, which is used as the intrinsic infrared absorption spectrum.

5. The spectroscopic detection method for modified ultrapolymer plastics according to claim 4, characterized in that, The step of mapping the functional relationship of the scattering correction model to the wavenumber domain of the infrared spectral data through coordinate transformation includes: Based on the inverse relationship between wavelength and wavenumber, the functional relationship of the scattering correction model is converted into a functional model with the wavenumber of the infrared spectrum as the independent variable, which is then used as the mapped functional relationship.

6. The spectroscopic detection method for modified ultrapolymer plastics according to claim 1, characterized in that, The process of analyzing the intrinsic infrared absorption spectrum to determine the relative content parameters of the modifier components includes: The characteristic variables of the intrinsic infrared absorption spectrum are extracted, and the characteristic variables include at least the characteristic absorption signal intensity of the modifier and the polymer matrix of the sample to be tested; The characteristic variables are input as input variables into the chemometrics quantitative model. The chemometrics quantitative model projects the input variables to obtain latent variables and generates regression coefficients through a partial least squares regression algorithm. The relative content parameters of the modifier components are obtained by performing linear regression on the latent variables using regression coefficients.

7. The spectroscopic detection method for modified ultrapolymer plastics according to claim 6, characterized in that, The process of obtaining the intensity of the characteristic absorbed signal includes: Identify the characteristic spectral bands of the modifier and the matrix polymer of the sample under test in the intrinsic infrared absorption spectrum; The intensity of the characteristic absorption signal is obtained by quantifying the signal amplitude of the characteristic spectral band.

8. The spectroscopic detection method for modified ultrapolymer plastics according to claim 1, characterized in that, The Raman characteristic signals of the modifier extracted from the Raman spectral data include: The relative content parameter is used as the initialization condition for the Raman spectrum unmixing process. The Raman spectrum data is iteratively unmixed using a non-negative matrix factorization algorithm to separate the pure Raman spectra of each chemical component. Based on the characteristic variables of the intrinsic infrared absorption spectrum, the spectral matching benchmark of the modifier is determined; The spectral matching reference includes: the wavenumber range in which the characteristic absorption peak of the modifier is identified in the intrinsic infrared absorption spectrum, and the relative proportion of the characteristic absorption signal intensity of the modifier and the matrix polymer in the wavenumber range; The degree of fit is calculated based on the pure Raman spectra of each chemical component and the spectral matching benchmark. Pure Raman spectra with a degree of fit reaching a preset threshold are extracted as the Raman characteristic signals of the modifier.

9. The spectroscopic detection method for modified ultrapolymer plastics according to claim 1, characterized in that, The determination of the modified structural information of the sample to be tested includes: Based on the complementarity and correlation between the intrinsic infrared absorption spectrum and the Raman characteristic signal, cross-validation is performed to determine the modified structural information of the modifier in the sample to be tested. The modified structural information includes at least: chemical bonding state and spatial distribution information.

10. A spectroscopic detection system for modified ultra-high molecular weight plastics, used to implement the spectroscopic detection method for modified ultra-high molecular weight plastics according to any one of claims 1-9, characterized in that, The system includes: The data acquisition unit is used to acquire ultraviolet-visible spectral data, infrared spectral data, and Raman spectral data of the sample to be tested; The correction unit is used to construct a scattering correction model based on the ultraviolet-visible spectral data, and to perform baseline correction on the infrared spectral data using the scattering correction model to obtain the intrinsic infrared absorption spectrum. The demixing unit is used to analyze the intrinsic infrared absorption spectrum, determine the relative content parameters of the modifier components, and introduce the relative content parameters as constraints into the Raman spectroscopy demixing process to extract the Raman characteristic signals of the modifier from the Raman spectral data. The output unit is used to determine the modified structure information of the sample under test based on the intrinsic infrared absorption spectrum and the Raman characteristic signal.

Images