A method and system for analyzing the composition uniformity of a perovskite thin film

Abstract

The present application belongs to the technical field of uniformity analysis, and particularly relates to a perovskite film composition uniformity analysis method and system, to solve the technical problem that traditional technology is difficult to standardize quality monitoring of composition uniformity of large-area perovskite film. The analysis method comprises: S1, collecting photoelectric response signals returned by each to-be-tested point; S2, constructing a graph model based on the spatial positions of each to-be-tested point and the corresponding harmonic entropy; S3, calculating the clustering coefficient of each node in the graph model, and multiplying the clustering coefficient of the node by the harmonic entropy corresponding to the node to obtain a perturbation vector; S4, calculating the kurtosis of the perturbation vector, and linearly combining the kurtosis with the reference parameter to obtain an index representing the composition uniformity of the perovskite film. The analysis method provides an objective and reliable evaluation basis for quality control and process optimization of perovskite film.

Description

Technical Field

[0001] This invention belongs to the technical field of uniformity analysis, specifically relating to a method and system for analyzing the compositional uniformity of perovskite thin films. Background Technology

[0002] Perovskite solar cells, as an emerging photovoltaic technology, have attracted much attention due to their high photoelectric conversion efficiency, low composition, and simple fabrication process. The perovskite thin film, as the core light-absorbing layer of a perovskite solar cell, directly determines the device's performance and stability. The compositional uniformity, crystal quality, and morphological uniformity of the perovskite thin film are key factors affecting charge generation, transport, and collection efficiency. During the large-area fabrication of perovskite thin films, factors such as solvent evaporation rates, changes in ambient temperature and humidity, and the instability of the precursor solution itself can easily lead to problems such as compositional segregation, uneven grain size, pinholes, or defects within the perovskite film. This inhomogeneity can become recombination centers for charge carriers, thereby increasing leakage current, reducing open-circuit voltage and fill factor, and thus limiting the efficiency and lifespan of large-area perovskite solar cell modules. Therefore, developing a rapid, accurate, and non-destructive analytical method to evaluate the compositional uniformity of perovskite thin films is of paramount importance.

[0003] Currently, traditional techniques for analyzing the uniformity of perovskite thin films mainly employ scanning electron microscopy (SEM), atomic force microscopy (AFM), X-ray diffraction (XRD), and steady-state / transient photoluminescence (PL) imaging. SEM and AFM are primarily used to observe the surface morphology and roughness of perovskite films, while XRD is used to analyze crystal structure and crystal orientation. PL imaging can indirectly reflect the spatial differences in defect density and carrier dynamics by detecting the distribution of luminescence intensity after excitation of the perovskite film.

[0004] However, microscopy and diffraction techniques can only analyze "points" or "lines" in tiny areas, making it difficult to quickly obtain information on the overall homogeneity of large-area samples. Although PL imaging can characterize large areas, the resulting light intensity distribution map is qualitative or semi-quantitative, making it difficult to provide objective and accurate numerical standards, and thus impossible to perform standardized, online quality monitoring and control. Summary of the Invention

[0005] This invention provides a method and system for analyzing the compositional uniformity of perovskite thin films, thereby solving the technical problem that traditional techniques are difficult to use for standardized quality monitoring of the compositional uniformity of large-area perovskite thin films.

[0006] In a first aspect, the present invention provides a method for analyzing the compositional uniformity of perovskite thin films, comprising the following steps:

[0007] S1, apply a dual-frequency periodic modulation excitation signal to multiple test points on the perovskite thin film, and collect the photoelectric response signal returned by each test point;

[0008] S2. Calculate the harmonic entropy of each test point using the harmonic components of the photoelectric response signal, and calculate the average value of the harmonic entropy of all test points, using this average value as the reference parameter. Construct a graph model based on the spatial location of each test point and its corresponding harmonic entropy, with each test point being a node in the graph model. When the absolute value of the difference in harmonic entropy between any two nodes is less than the product of the reference parameter and the preset coefficient, an edge is established between the two nodes.

[0009] S3, calculate the clustering coefficient of each node in the graph model, and multiply the clustering coefficient of the node by the harmonic entropy corresponding to the node to obtain the perturbation vector;

[0010] S4. Calculate the kurtosis of the disturbance vector and linearly combine the kurtosis with the reference parameter to obtain an index representing the compositional uniformity of the perovskite thin film.

[0011] Furthermore, in S1, an intensity-adjustable LED light source is used as the excitation light source; the test point is set as an N×N grid array covering the surface of the perovskite thin film; the dual-frequency periodic modulation excitation signal is the superposition of two sine signals; and a photodetector and a data acquisition card are used to acquire the photoelectric response signal of each test point at a set sampling rate.

[0012] Furthermore, in S2, the harmonic entropy of each test point is calculated using the harmonic components of the photoelectric response signal, including:

[0013] The acquired photoelectric response signal was subjected to a Fast Fourier Transform to extract the amplitudes of the DC component, fundamental frequency component, higher harmonic components, and intermodulation sum and difference frequency components. There are a total of M+1 amplitude values, where s = 0, 1, ..., M;

[0014] Calculate the proportion of each amplitude in the total amplitude. Its calculation formula is ;

[0015] Calculate the harmonic entropy of each measurement point using the Shannon entropy formula. Harmonic entropy The calculation formula is: .

[0016] Furthermore, in S2, a graphical model is constructed based on the spatial location of each point to be measured and its corresponding harmonic entropy, including:

[0017] Each test point on the perovskite thin film is taken as a node in the graphical model; for any two nodes i and j in the graphical model, if the harmonic entropy of the node is... and An undirected edge is established between nodes i and j if the following conditions are met: ;

[0018] in, The average harmonic entropy of all points to be measured. These are the preset coefficients used in the construction of the graph model.

[0019] Furthermore, in S3, the clustering coefficient of each node in the graph model is calculated, including:

[0020] For any node i in the graph model, let the number of its neighboring nodes be denoted as . The actual number of edges between neighboring nodes is Clustering coefficient of node i Calculated using the following formula:

[0021] .

[0022] Furthermore, in S3, the clustering coefficient of a node is multiplied by the harmonic entropy corresponding to the node to obtain the perturbation vector, which includes:

[0023] For the i-th node (i=1, 2, ..., ...) in the graph model ), clustering coefficients With the corresponding harmonic entropy Multiply by the following formula to obtain the disturbance component. : ;

[0024] The perturbation components of all nodes constitute a The perturbation vector V = [ ].

[0025] Further, in S4, the kurtosis is linearly combined with the reference parameter to obtain an index representing the compositional uniformity of the perovskite thin film, including:

[0026] Let K be the kurtosis of the perturbation vector, and the reference parameter be... ; For reference parameters Normalization is performed to obtain dimensionless reference parameters. : Where M+1 represents the number of frequency components;

[0027] The index U is calculated using the following formula, which is a weighted sum of dimensionless parameters: ;in, is the weighting coefficient for kurtosis K.

[0028] Secondly, the present invention provides a perovskite thin film composition uniformity analysis system, comprising the following modules:

[0029] The acquisition module applies a dual-frequency periodic modulation excitation signal to multiple test points on the perovskite thin film and acquires the photoelectric response signal returned by each test point.

[0030] The module constructs a harmonic entropy for each test point by utilizing the harmonic components of the photoelectric response signal, and calculates the average value of the harmonic entropy of all test points, using this average value as a reference parameter. Based on the spatial location of each test point and its corresponding harmonic entropy, a graph model is constructed, with each test point as a node in the graph model. When the absolute value of the difference in harmonic entropy between any two nodes is less than the product of the reference parameter and a preset coefficient, an edge is established between the two nodes.

[0031] The calculation module calculates the clustering coefficient of each node in the graph model and multiplies the clustering coefficient of the node with the harmonic entropy corresponding to the node to obtain the perturbation vector.

[0032] The module obtains the kurtosis of the perturbation vector and linearly combines the kurtosis with the reference parameter to obtain an index representing the compositional uniformity of the perovskite thin film.

[0033] Furthermore, in the acquisition module, an intensity-adjustable LED light source is used as the excitation light source; the test point is set as an N×N grid array covering the surface of the perovskite thin film; the dual-frequency periodic modulation excitation signal is the superposition of two sinusoidal signals; and a photodetector and a data acquisition card are used to acquire the photoelectric response signal of each test point at a set sampling rate.

[0034] Furthermore, in the construction module, the harmonic entropy of each test point is calculated using the harmonic components of the photoelectric response signal, including:

[0035] The acquired photoelectric response signal was subjected to a Fast Fourier Transform to extract the amplitudes of the DC component, fundamental frequency component, higher harmonic components, and intermodulation sum and difference frequency components. There are a total of M+1 amplitude values, where s = 0, 1, ..., M;

[0036] Calculate the proportion of each amplitude in the total amplitude. Its calculation formula is ;

[0037] Calculate the harmonic entropy of each measurement point using the Shannon entropy formula. Harmonic entropy The calculation formula is: .

[0038] The beneficial effects are as follows: The perovskite thin film composition uniformity evaluation method of this invention combines the nonlinear harmonic characteristics of photoelectric response with graph theory analysis. By calculating harmonic entropy, it can more sensitively detect the nonlinear response characteristics caused by minute fluctuations in composition within the perovskite thin film, improving the sensitivity and depth of detection. This invention utilizes graph theory analysis to fuse the spatial location information of the test point with harmonic entropy. By constructing a graph model and calculating clustering coefficients, it displays the spatial correlation and local aggregation characteristics of non-uniform regions. This invention provides an objective and reliable evaluation basis for the quality control and process optimization of perovskite thin films. Attached Figure Description

[0039] Figure 1 This is a flowchart of a method for evaluating the compositional uniformity of perovskite thin films. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Those skilled in the art should understand that the embodiments described below are only some, not all, of the embodiments disclosed. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0041] Examples of the perovskite thin film composition uniformity evaluation method provided by the present invention:

[0042] like Figure 1 As shown, a method for evaluating the compositional uniformity of perovskite thin films includes the following steps:

[0043] S1 applies a dual-frequency periodic modulation excitation signal to multiple test points on the perovskite thin film and collects the photoelectric response signal returned by each test point.

[0044] In an optional embodiment, in S1, an intensity-adjustable LED light source is used as the excitation light source; the test point is set as an N×N grid array covering the surface of the perovskite thin film; the dual-frequency periodic modulation excitation signal is the superposition of two sinusoidal signals; and a photodetector and a data acquisition card are used to acquire the photoelectric response signal of each test point at a set sampling rate.

[0045] For example, the wavelength of the LED light source is 450nm, and N is 10; of the two sinusoidal signals, the frequency of the first sinusoidal signal is 100Hz, and the frequency of the second sinusoidal signal is 230Hz; the sampling rate is set to 20kHz.

[0046] Specifically, a 450nm LED light source is prepared, and its intensity can be adjusted according to experimental needs. A 10×10 grid is planned on the surface of the perovskite thin film sample to be tested, and the 100 grid intersections are the test points. A dual-frequency periodic modulation signal, consisting of a 100Hz sine wave and a 230Hz sine wave, is generated by a signal generator. This dual-frequency periodic modulation signal drives the LED light source, emitting modulated light whose intensity changes periodically with the dual-frequency periodic modulation signal.

[0047] The modulated light was sequentially focused onto 100 test points on the surface of the perovskite thin film. Each time the modulated light illuminated a test point, a photodetector received the photoelectric response signal from that point. The photodetector was connected to a data acquisition card, which converted the photoelectric response signal into a digital signal and recorded it at a sampling rate of 20 kHz. This process was repeated for all 100 test points to obtain 100 sets of raw photoelectric response signal time-series data corresponding to different spatial locations, providing a basis for subsequent uniformity analysis.

[0048] S2. Calculate the harmonic entropy of each test point using the harmonic components of the photoelectric response signal, and calculate the average value of the harmonic entropy of all test points, using this average value as the reference parameter. Construct a graph model based on the spatial location of each test point and its corresponding harmonic entropy, with each test point as a node in the graph model. When the absolute value of the difference in harmonic entropy between any two nodes is less than the product of the reference parameter and the preset coefficient, an edge is established between the two nodes.

[0049] In an optional embodiment, in S2, the harmonic entropy of each test point is calculated using the harmonic components of the photoelectric response signal, including:

[0050] The acquired photoelectric response signal was subjected to a Fast Fourier Transform to extract the amplitudes of the DC component, fundamental frequency component, higher harmonic components, and intermodulation sum and difference frequency components. There are a total of M+1 amplitude values, where s = 0, 1, ..., M;

[0051] Calculate the proportion of each amplitude in the total amplitude. Its calculation formula is ;

[0052] Calculate the harmonic entropy of each measurement point using the Shannon entropy formula. Harmonic entropy The calculation formula is: .

[0053] For example, M is 10, and the number of extracted frequency components is M+1=11; the DC component is 0Hz; the fundamental frequency components are the two original input frequencies, namely 100Hz and 230Hz as mentioned above; the higher harmonic components are integer multiples of the two fundamental frequency components, here 200Hz, 300Hz, and 460Hz are selected; the intermodulation sum and difference frequency components are linear combinations of the addition or subtraction of the two fundamental frequency components, here 130Hz (i.e., 230-100Hz) and 360Hz are selected. 0Hz (i.e., 2×230-100Hz), 330Hz (i.e., 100+230Hz), 430Hz (i.e., 2×230+100Hz), and 560Hz (i.e., 100+2×230Hz), which includes a DC component of 0Hz and 10 main frequency components of 100Hz, 130Hz, 200Hz, 230Hz, 300Hz, 330Hz, 360Hz, 430Hz, 460Hz, and 560Hz.

[0054] Specifically, select one set of data from the 100 sets obtained in the previous step, for example, the time series signal corresponding to the first test point. Perform a Fast Fourier Transform algorithm on this set of time series signals to transform it from the time domain to the frequency domain, obtaining a frequency-amplitude spectrum. From the frequency-amplitude spectrum, find the aforementioned 11 frequency components and read their amplitudes.

[0055] Assuming the 11 extracted amplitude values ​​are as follows until Add all 11 amplitude values ​​together to obtain a total amplitude value. Divide the amplitude value of each component by the total amplitude value to obtain the proportion of that component, denoted as . until .For example, equal Divide by the total amplitude. Substitute these 11 percentages into the Shannon entropy calculation formula, that is, for each percentage... Multiply by base 2 The logarithm of the result is taken, the 11 results are summed, and the negative number is taken. The resulting value is the harmonic entropy H of the measured point. This complete calculation process needs to be repeated for all 100 measured points to obtain 100 harmonic entropy values.

[0056] In an optional embodiment, in S2, a graphical model is constructed based on the spatial location of each point to be measured and its corresponding harmonic entropy, including:

[0057] Each test point on the perovskite thin film is taken as a node in the graphical model; for any two nodes i and j in the graphical model, if the harmonic entropy of the node is... and An undirected edge is established between nodes i and j if the following conditions are met: ;

[0058] in, The average harmonic entropy of all points to be measured. These are the preset coefficients used in the construction of the graph model.

[0059] The preset coefficient is determined based on engineering judgment after experimental testing and parameter scanning of multiple batches of perovskite thin films. Its function is to set a maximum allowable difference ratio, ensuring that the harmonic entropy difference between two test points is less than the global average uniformity. Only when the coefficient is multiplied by a factor of 1 is the data considered similar and connected as an edge. For example, the preset coefficients in graph model construction... It is 0.05.

[0060] Specifically, a graphical model is established to represent the similarity of harmonic entropy among different test points on the perovskite thin film: the 100 harmonic entropy values ​​H calculated in the previous step are added together, and then divided by 100 to obtain the average harmonic entropy of all test points. Assuming the calculation yields... The value is 2.5. A connection threshold is calculated using the formula: 2.5 × 0.05 = 0.125. This connection threshold will be used as the criterion for determining whether two nodes are similar.

[0061] Begin constructing the graphical model: The 100 nodes in the graphical model correspond to 100 test points on the perovskite thin film. Taking nodes i and j as examples, extract their corresponding harmonic entropy values. and Assuming It is 2.55. The value is 2.48. The absolute value of the difference between the two harmonic entropy values ​​is calculated, resulting in 0.07. Since 0.07 is less than the connection threshold of 0.125, an undirected edge is created between node i and node j to indicate that they are connected. If the harmonic entropy difference between another pair of nodes is greater than 0.125, no line is drawn between them. This process is repeated for all 4950 distinct pairs of nodes to form a complete undirected graph.

[0062] S3 calculates the clustering coefficient of each node in the graph model, and multiplies the clustering coefficient of the node with the harmonic entropy corresponding to the node to obtain the perturbation vector.

[0063] In an optional embodiment, in S3, the clustering coefficient of each node in the graph model is calculated, including:

[0064] For any node i in the graph model, let the number of its neighboring nodes be denoted as . The actual number of edges between neighboring nodes is The clustering coefficient of node i Calculated using the following formula: .

[0065] Specifically, after constructing the graph model, the local connectivity characteristics of each node in the graph model are analyzed one by one. Taking a node i in the graph as an example, all nodes directly connected to node i are identified; these nodes are called the neighbor nodes of node i. The total number of neighbor nodes is counted and denoted as . For example, if node i is directly connected to 4 other nodes, then =4.

[0066] Examine the four neighboring nodes themselves: focusing only on the relationships between these four neighbors, check how many edges exist between them. Assuming that three pairs of neighbors are connected, count the actual number of edges between them. =3. According to the formula for calculating the clustering coefficient, when the number of neighboring nodes... When ≥2, ×2, then divide by and The product of minus 1, the clustering coefficient The result is 0.5. If a node has fewer than two neighboring nodes, such as only one or zero, its clustering coefficient is directly set to 0. The above calculation process is performed on all 100 nodes in the graph model to obtain the clustering coefficient value for each node.

[0067] In an optional embodiment, in S3, the clustering coefficient of a node is multiplied by the harmonic entropy corresponding to the node to obtain a perturbation vector, including:

[0068] For the i-th node in the graph (i=1, 2, ... ), clustering coefficients With the corresponding harmonic entropy Multiply by the following formula to obtain the disturbance component. : ;

[0069] The perturbation components of all nodes constitute a The perturbation vector V = [ ].

[0070] For example, the total number of nodes here. .

[0071] Specifically, there are two datasets: one containing 100 harmonic entropies and the other containing 100 corresponding clustering coefficients. The data of these 100 nodes are processed in sequence.

[0072] Taking the first node as an example, find its harmonic entropy. and clustering coefficient Assuming It is 2.6. It is 0.5. Multiplying the two values ​​together gives the first disturbance component. That is, 1.3. The same operation is performed on the second node to change the harmonic entropy. With clustering coefficient Multiply to get This process is repeated until the perturbation components for all 100 nodes are calculated. 100 calculated disturbance components ,until Arrange them in order to form a one-dimensional array containing 100 elements. The one-dimensional array is the 100-dimensional perturbation vector V.

[0073] S4. Calculate the kurtosis of the disturbance vector and linearly combine the kurtosis with the reference parameter to obtain an index representing the compositional uniformity of the perovskite thin film.

[0074] Specifically, the mean and standard deviation of the perturbation vector are calculated. The mean is subtracted from each element of the perturbation vector, and then the result is divided by the standard deviation to obtain a standardized value. Each standardized value is then raised to the fourth power, and the average of all fourth-powered values ​​is calculated. Subtracting 3 from this average yields the kurtosis value of the perturbation vector, which reflects the sharpness of the perturbation vector data distribution.

[0075] In an optional embodiment, in S4, the kurtosis is linearly combined with the reference parameter to obtain an index representing the compositional uniformity of the perovskite film, including:

[0076] Let K be the kurtosis of the perturbation vector, and the reference parameter be... The reference parameters are normalized to obtain dimensionless reference parameters. : Where M+1 represents the number of frequency components;

[0077] The index U is calculated using the following formula, which is a weighted sum of dimensionless parameters: ;in, is the weighting coefficient for kurtosis K.

[0078] For example, M+1=11, consistent with the number of frequency components mentioned above. When selecting weighting coefficients... At that time, by trying different Candidate values ​​were selected, and the index U was calculated and compared with the experimental quality of perovskite films for verification. The final selected value is the weighting coefficient that maximizes the sensitivity and discrimination of index U to the inhomogeneity characteristics of perovskite films and has the best correlation with actual quality. It is 0.6.

[0079] Specifically, statistical analysis is performed on the 100-dimensional perturbation vector V obtained in the previous step to calculate the kurtosis K. Kurtosis K is a statistic that represents the steepness or flatness of the data distribution; a standard normal distribution has a kurtosis of 3. Assume the kurtosis K calculated for the perturbation vector V is 4.2. The average value of the harmonic entropy of all the measured points calculated previously is used. As a baseline parameter, assuming It is 2.5.

[0080] For reference parameters Normalization is performed to make the reference parameter a dimensionless value. By... Dividing by a constant, resulting in an approximate value of 3.459, the normalized baseline parameter... Approximately 0.723. The kurtosis K and the normalized baseline parameter are then used. Weighted summation is performed to obtain the uniformity index U. According to the formula, U = 2.8092. The index U integrates the global average characteristics and local structural characteristics of the perovskite film, providing an evaluation index for the overall uniformity of the perovskite film.

[0081] Examples of the perovskite thin film composition uniformity analysis system provided by the present invention:

[0082] The perovskite thin film composition uniformity analysis system includes the following modules:

[0083] The acquisition module applies a dual-frequency periodic modulation excitation signal to multiple test points on the perovskite thin film and acquires the photoelectric response signal returned by each test point.

[0084] The module constructs a harmonic entropy for each test point by utilizing the harmonic components of the photoelectric response signal, and calculates the average value of the harmonic entropy of all test points, using this average value as a reference parameter. Based on the spatial location of each test point and its corresponding harmonic entropy, a graph model is constructed, with each test point as a node in the graph model. When the absolute value of the difference in harmonic entropy between any two nodes is less than the product of the reference parameter and a preset coefficient, an edge is established between the two nodes.

[0085] The calculation module calculates the clustering coefficient of each node in the graph model and multiplies the clustering coefficient of the node with the harmonic entropy corresponding to the node to obtain the perturbation vector.

[0086] The module obtains the kurtosis of the perturbation vector and linearly combines the kurtosis with the reference parameter to obtain an index representing the compositional uniformity of the perovskite thin film.

[0087] In an optional embodiment, the acquisition module uses an intensity-adjustable LED light source as the excitation light source; the test point is set as an N×N grid array covering the surface of the perovskite thin film; the dual-frequency periodic modulation excitation signal is the superposition of two sinusoidal signals; and a photodetector and a data acquisition card are used to acquire the photoelectric response signal of each test point at a set sampling rate.

[0088] In an optional embodiment, the construction module calculates the harmonic entropy of each test point using the harmonic components of the photoelectric response signal, including:

[0089] The acquired photoelectric response signal was subjected to a Fast Fourier Transform to extract the amplitudes of the DC component, fundamental frequency component, higher harmonic components, and intermodulation sum and difference frequency components. There are a total of M+1 amplitude values, where s = 0, 1, ..., M;

[0090] Calculate the proportion of each amplitude in the total amplitude. Its calculation formula is ;

[0091] Calculate the harmonic entropy of each measurement point using the Shannon entropy formula. Harmonic entropy The calculation formula is: .

[0092] In addition, in the description of this specification, "multiple" means at least two, such as two, three or more, etc., unless otherwise expressly and specifically defined.

Claims

1. A method for analyzing the compositional uniformity of perovskite thin films, characterized in that, Includes the following steps: S1, apply a dual-frequency periodic modulation excitation signal to multiple test points on the perovskite thin film, and collect the photoelectric response signal returned by each test point; S2. Calculate the harmonic entropy of each test point using the harmonic components of the photoelectric response signal, and calculate the average value of the harmonic entropy of all test points, using this average value as the reference parameter. Construct a graph model based on the spatial location of each test point and its corresponding harmonic entropy, with each test point being a node in the graph model. When the absolute value of the difference in harmonic entropy between any two nodes is less than the product of the reference parameter and the preset coefficient, an edge is established between the two nodes. S3 calculates the clustering coefficient for each node in the graph model, including: For any node i in the graph model, let the number of its neighboring nodes be denoted as . The actual number of edges between neighboring nodes is Clustering coefficient of node i Calculated using the following formula: The clustering coefficient of the node is multiplied by the harmonic entropy corresponding to the node to obtain the perturbation vector. S4. Calculate the kurtosis of the disturbance vector and linearly combine the kurtosis with the reference parameter to obtain an index representing the compositional uniformity of the perovskite thin film.

2. The method for analyzing the compositional uniformity of perovskite thin films according to claim 1, characterized in that, In S1, an intensity-adjustable LED light source is used as the excitation light source; the test point is set as an N×N grid array covering the surface of the perovskite thin film; the dual-frequency periodic modulation excitation signal is the superposition of two sine signals; The photoelectric response signal of each test point is acquired using a photodetector and a data acquisition card at a set sampling rate.

3. The method for analyzing the compositional uniformity of perovskite thin films according to claim 1 or 2, characterized in that, In S2, the harmonic entropy of each test point is calculated using the harmonic components of the photoelectric response signal, including: The acquired photoelectric response signal was subjected to a Fast Fourier Transform to extract the amplitudes of the DC component, fundamental frequency component, higher harmonic components, and intermodulation sum and difference frequency components. There are a total of M+1 amplitude values, where s = 0, 1, ..., M; Calculate the proportion of each amplitude in the total amplitude. Its calculation formula is ; Calculate the harmonic entropy of each measurement point using the Shannon entropy formula. Harmonic entropy The calculation formula is: .

4. The method for analyzing the compositional uniformity of perovskite thin films according to claim 3, characterized in that, In S2, a graphical model is constructed based on the spatial location of each point to be measured and its corresponding harmonic entropy, including: Each test point on the perovskite thin film is taken as a node in the graphical model; for any two nodes i and j in the graphical model, if the harmonic entropy of the node is... and An undirected edge is established between nodes i and j if the following conditions are met: ; in, The average harmonic entropy of all points to be measured. These are the preset coefficients used in the construction of the graph model.

5. The method for analyzing the compositional uniformity of perovskite thin films according to claim 4, characterized in that, In S3, the clustering coefficient of a node is multiplied by the harmonic entropy corresponding to the node to obtain the perturbation vector, which includes: For the i-th node (i=1, 2, ..., ...) in the graph model ), clustering coefficients With the corresponding harmonic entropy Multiply by the following formula to obtain the disturbance component. : ; The perturbation components of all nodes constitute a The perturbation vector V = [ ].

6. The method for analyzing the compositional uniformity of perovskite thin films according to claim 5, characterized in that, In S4, the kurtosis is linearly combined with the reference parameter to obtain an index representing the compositional uniformity of the perovskite film, including: Let K be the kurtosis of the perturbation vector, and the reference parameter be... ; For reference parameters Normalization is performed to obtain dimensionless reference parameters. : Where M+1 represents the number of frequency components; The index U is calculated using the following formula, which is a weighted sum of dimensionless parameters: ;in, is the weighting coefficient for kurtosis K.

7. A perovskite thin film composition uniformity analysis system, characterized in that, Includes the following modules: The acquisition module applies a dual-frequency periodic modulation excitation signal to multiple test points on the perovskite thin film and acquires the photoelectric response signal returned by each test point. The module constructs a harmonic entropy for each test point by utilizing the harmonic components of the photoelectric response signal, and calculates the average value of the harmonic entropy of all test points, using this average value as a reference parameter. Based on the spatial location of each test point and its corresponding harmonic entropy, a graph model is constructed, with each test point as a node in the graph model. When the absolute value of the difference in harmonic entropy between any two nodes is less than the product of the reference parameter and a preset coefficient, an edge is established between the two nodes. The computation module calculates the clustering coefficients of each node in the graph model, including: For any node i in the graph model, let the number of its neighboring nodes be denoted as . The actual number of edges between neighboring nodes is Clustering coefficient of node i Calculated using the following formula: The clustering coefficient of the node is multiplied by the harmonic entropy corresponding to the node to obtain the perturbation vector. The module obtains the kurtosis of the perturbation vector and linearly combines the kurtosis with the reference parameter to obtain an index representing the compositional uniformity of the perovskite thin film.

8. The perovskite thin film composition uniformity analysis system according to claim 7, characterized in that, In the acquisition module, an intensity-adjustable LED light source is used as the excitation light source; the test point is set as an N×N grid array covering the surface of the perovskite thin film; the dual-frequency periodic modulation excitation signal is the superposition of two sinusoidal signals; The photoelectric response signal of each test point is acquired using a photodetector and a data acquisition card at a set sampling rate.

9. The perovskite thin film composition uniformity analysis system according to claim 7 or 8, characterized in that, In the construction module, the harmonic entropy of each test point is calculated using the harmonic components of the photoelectric response signal, including: The acquired photoelectric response signal was subjected to a Fast Fourier Transform to extract the amplitudes of the DC component, fundamental frequency component, higher harmonic components, and intermodulation sum and difference frequency components. There are a total of M+1 amplitude values, where s = 0, 1, ..., M; Calculate the proportion of each amplitude in the total amplitude. Its calculation formula is ; Calculate the harmonic entropy of each measurement point using the Shannon entropy formula. Harmonic entropy The calculation formula is: .

Images