An organic solar cell parameter optimization method based on light field numerical simulation

Abstract

This application discloses a parameter optimization method for organic solar cells based on optical field numerical simulation, belonging to the field of computer-aided simulation optimization technology. It includes: performing optical field distribution simulation on a tandem organic solar cell to obtain the photon absorptivity distribution corresponding to each functional layer; obtaining the theoretically calculated value of the short-circuit current density of each sub-cell; performing global collaborative optimization on the parameter space formed by the thickness of each functional layer to obtain an initial optimal parameter combination; determining the perturbation amount based on the variation range of the thickness of each functional layer in the actual fabrication process, and applying the perturbation amount to the initial optimal parameter combination to generate a perturbed parameter combination; selecting robust parameter combinations that meet the preset robust threshold constraint based on the current matching error and short-circuit current density attenuation amplitude under each perturbed parameter combination; and generating a global collaborative process guidance scheme for the fabrication of tandem organic solar cells.

Description

Technical Field

[0001] This application relates to the field of computer-aided simulation optimization technology, and more specifically, to a method for optimizing the parameters of organic solar cells based on numerical simulation of light fields. Background Technology

[0002] Organic solar cells have attracted widespread attention in the field of novel photovoltaic devices in recent years due to their advantages such as flexibility, low manufacturing cost, and suitability for large-area fabrication. Among them, tandem organic solar cells, through their multi-layered photoactive structure, synergistically absorb different wavelengths of light, effectively improving light energy utilization efficiency, and thus have become an important research direction for improving the photoelectric conversion performance of devices.

[0003] In tandem organic solar cells, the thickness parameters of each functional layer directly affect the light propagation path, light field distribution, and photon absorption efficiency within the device, and further influence the current matching relationship between the sub-cells. Due to the complex optical coupling relationships between the functional layers, the thickness parameters of each layer usually exhibit significant global synergistic characteristics, and changes in the parameters of a single functional layer often have a ripple effect on the overall device performance.

[0004] In existing technologies, device design often employs empirical trial-and-error, local parameter scanning, or single-layer independent optimization methods, making it difficult to simultaneously consider the optical field coupling relationship and overall current matching requirements of multilayer film systems. Although some technologies introduce optical simulation methods to analyze the device's optical field, most only perform static calculations on theoretical light absorption capabilities, lacking comprehensive consideration of multi-parameter collaborative optimization, the impact of process fluctuations, and device performance stability. This results in a significant deviation between theoretical optimization results and actual fabrication effects.

[0005] Furthermore, in actual thin-film deposition processes, the thickness of the functional layer is easily affected by factors such as deposition rate fluctuations and equipment control errors, leading to device performance fluctuations and impacting the consistency and mass production stability of tandem organic solar cells. Current technologies lack a parameter optimization method that combines numerical simulation of the light field, global parameter co-optimization, and process robustness analysis, making it difficult to meet the high-precision optimization requirements of complex tandem structures.

[0006] In summary, how to achieve global synergistic optimization and stable fabrication of tandem organic solar cells while considering the light propagation coupling relationship of multilayer optical film systems and the impact of actual process fluctuations has become an urgent technical problem to be solved. Summary of the Invention

[0007] To overcome a series of shortcomings in existing technologies, the purpose of this application is to provide a method for optimizing the parameters of organic solar cells based on numerical simulation of light fields, comprising the following steps: Using the thickness of each functional layer as a global collaborative optimization variable, and based on a complex refractive index database, the light field distribution of the stacked organic solar cell is simulated to obtain the photon absorptivity distribution of each functional layer. Based on the photon absorptivity distribution, the short-circuit current density distribution corresponding to each functional layer and the overall stacked structure is calculated to obtain the theoretical calculated value of the short-circuit current density of each sub-cell. Global collaborative optimization is performed on the parameter space formed by the thickness of each functional layer to obtain the initial optimal parameter combination; The perturbation amount is determined based on the thickness fluctuation range of each functional layer in the actual fabrication process, and the perturbation amount is applied to the initial optimal parameter combination to generate a perturbation parameter combination; Based on the current matching error and short-circuit current density attenuation amplitude under various perturbation parameter combinations, robust parameter combinations that meet the preset robust threshold constraints are selected. A mapping relationship between robust parameter combinations and actual thin film deposition process parameters is established to generate a global collaborative process guidance scheme for the fabrication of tandem organic solar cells.

[0008] In some embodiments, the method for simulating the light field distribution of tandem organic solar cells is as follows: Based on the electromagnetic field boundary conditions at the interface of each functional layer and the complex refractive index optical parameters of each functional layer, the interface matrix and the propagation matrix are constructed respectively. The interface matrix and the propagation matrix are concatenated according to the light propagation order to construct the total transmission matrix that characterizes the overall optical response relationship. Based on the relationship between the thickness of the functional layer and the coherence length, coherent and incoherent propagation are distinguished for each functional layer, and the local optical field distribution along the thickness direction inside each functional layer is solved based on the total transmission matrix. Calculate the photon absorption power density at discrete positions along the thickness direction of each functional layer, and integrate the photon absorption characteristics of each functional layer along the thickness direction. Based on the photon absorption characteristics of each functional layer, light absorption simulation results are generated to characterize the optical response capability of tandem organic solar cells.

[0009] In some embodiments, the method for obtaining the theoretically calculated value of the short-circuit current density of each sub-cell is as follows: The incident spectral data within the target wavelength range is acquired and photon flux conversion processing is performed to obtain the corresponding photon flux distribution. Based on the light absorption characteristics of each functional layer, the photon flux distribution is calculated by light absorption coupling to obtain the effective absorbed photon distribution of each functional layer. Based on the photoelectric conversion characteristics, the carrier conversion of the effectively absorbed photon distribution is calculated to obtain the short-circuit current density corresponding to each functional layer. The short-circuit current density of each functional layer is summarized and calculated to obtain the theoretical short-circuit current density of each sub-cell. The theoretical short-circuit current density is corrected based on the parasitic absorption loss of the non-photosensitive layer to obtain the theoretical calculated value of the short-circuit current density of each sub-cell.

[0010] In some embodiments, the method for global collaborative optimization of the parameter space formed by the thicknesses of each functional layer is as follows: Parameter initialization sampling is performed within the parameter space composed of the thickness parameters of each functional layer to generate an initial parameter population for multi-objective optimization. Based on the functional layer thickness combination corresponding to each initial parameter population, optical and electrical response simulation calculations are performed to obtain the objective function evaluation results corresponding to each thickness combination. The initial parameter population is subjected to non-dominated sorting, and selection, crossover and mutation operations are performed sequentially on the sorted initial parameter population to generate a new generation of parameter population; During the population iteration process, the current matching constraint is used to determine the constraint of each thickness combination, and the fitness of the corresponding individual is adjusted according to the constraint satisfaction. The population iteration process is terminated based on preset convergence conditions, and after termination, the combination of thickness parameters with the best comprehensive performance is selected from the Pareto optimal solution set to obtain the global collaborative optimization result.

[0011] In some embodiments, the method for determining the disturbance amount is as follows: Obtain the thickness control parameters of the thin film deposition process corresponding to each functional layer, and determine the thickness perturbation range of each functional layer based on the thickness fluctuation characteristics corresponding to different deposition processes. Based on the thickness perturbation range corresponding to each functional layer, the perturbation level of the thickness parameters of each functional layer is set, and the corresponding combination of thickness perturbation parameters is generated by experimental design method. Based on the number of functional layers, the combination of thickness perturbation parameters is optimized to obtain a representative combination of thickness parameters for performance analysis. The representative thickness parameters are combined and input into the preset optical and electrical response simulation model for calculation to obtain the corresponding current matching error and short-circuit current density. Based on the current matching error and short-circuit current density corresponding to each parameter combination, the robustness of the thickness parameters of each functional layer is evaluated, and the target thickness parameter combination that meets the preset performance requirements is selected.

[0012] In some embodiments, the method for selecting robust parameter combinations that satisfy the preset robustness threshold constraint is as follows: For each combination of thickness disturbance parameters, the current matching error and short-circuit current density are calculated, and the calculation results are compared with the performance results corresponding to the initial optimal parameter combination. Based on the current matching error and short-circuit current density results corresponding to various thickness disturbance parameter combinations, current matching robustness index and current performance robustness index are constructed respectively. Statistical analysis of performance fluctuations is performed on candidate parameter combinations that meet the preset robustness conditions to obtain statistical characteristic parameters that characterize performance stability. A comprehensive evaluation is conducted based on the average performance, performance stability, and worst-case performance of each candidate parameter combination, and robust parameter combinations that meet the preset robustness threshold constraints are selected.

[0013] In some embodiments, the method for establishing the mapping relationship between robust parameter combinations and actual thin film deposition process parameters is as follows: For different types of thin film deposition processes, a thickness mapping model between process parameters and thin film deposition thickness is constructed, and the model parameters of the thickness mapping model are determined based on experimental calibration data. Based on the target thickness parameters corresponding to each functional layer, the thickness mapping model of the corresponding deposition process is called to solve the corresponding combination of process parameters; For situations where multiple candidate deposition processes correspond to the same functional layer, the candidate deposition processes are matched and screened based on the target thickness requirements, thickness uniformity requirements, and deposition efficiency requirements to determine the target deposition process. Based on the target deposition process, process window constraint information is introduced into the corresponding thickness mapping model, and the adjustable range and parameter sensitivity characteristics of each process parameter are determined. Based on the target deposition process and the corresponding combination of process parameters, a mapping relationship is established between the robust parameter combination and the actual thin film deposition process parameters.

[0014] In some embodiments, the organic solar cell parameter optimization method further includes: During the execution of the process guidance scheme, online thickness monitoring data of each functional layer is collected simultaneously, and the thickness parameters are updated accordingly based on the thickness deviation. At the same time, the corresponding optical field simulation results and short-circuit current density prediction values ​​are calculated in real time.

[0015] In some embodiments, the method for equivalently updating the thickness parameter based on the thickness deviation is as follows: After the deposition of each functional layer is completed, the actual thickness data of the corresponding functional layer output by the online thickness monitoring system is obtained; Based on the actual thickness data, the theoretical thickness parameters of the corresponding functional layers in the simulation model are replaced and updated, and the propagation matrix parameters of the corresponding functional layers are updated synchronously. The thickness parameters of the functional layers that have been deposited are locked, and the thickness parameters of the functional layers that have not been deposited are retained as variables to be optimized, so as to construct a hybrid parameter model that includes the measured thickness parameters and the thickness parameters to be optimized. Based on the hybrid parameter model, the optical and electrical response simulation calculations were re-executed to obtain the updated light field distribution, short-circuit current density of each sub-cell, and current matching error results. Based on the updated current matching error results, the thickness parameters of subsequent functional layers are adjusted to compensate for the errors, and the equivalent update process and corresponding calculation results are recorded to complete the dynamic equivalent update of the thickness parameters.

[0016] In some embodiments, the organic solar cell parameter optimization method further includes: Based on the optical field simulation results and the predicted short-circuit current density, it is determined whether the current fabrication state meets the preset performance optimization standards. If the conditions are not met, the initial optimal parameter combination will be locally adaptively updated based on the current online film thickness deviation data, and incremental global collaborative optimization will be performed. If the conditions are met, the parameters of the current functional layer are locked, and the process guidance scheme is continued to be executed while maintaining the consistency of global collaborative constraints.

[0017] Compared with the prior art, this application has the following beneficial effects: This application uses the thickness of each functional layer of a tandem organic solar cell as a global co-optimization variable, and combines optical field numerical simulation, robust perturbation analysis, and actual process mapping to achieve synergistic optimization of the photoelectric performance and fabrication stability of the tandem organic solar cell while taking into account both current matching accuracy and tolerance to fabrication process fluctuations. This improves the theoretical optimization accuracy, actual fabrication consistency, and mass production feasibility of the device. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a method for optimizing organic solar cell parameters based on numerical simulation of light field, as disclosed in an embodiment of this application.

[0019] Figure 2 This is a schematic diagram of the global collaborative optimization and robustness screening process in the embodiments of this application.

[0020] Figure 3 This is a schematic diagram of the cross-sectional structure of the stacked organic solar cell in the embodiments of this application.

[0021] Figure 4 This is a schematic diagram illustrating the principle of transmission matrix calculation in the embodiments of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some embodiments of this invention, but not all embodiments.

[0023] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] The embodiments and directional terms described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0025] like Figure 1 As shown, a method for optimizing the parameters of organic solar cells based on numerical simulation of light fields includes the following steps: Acquire complex refractive index spectral data of each functional layer material in the target wavelength range, and establish a complex refractive index database characterizing the optical response properties of each functional layer material; Using the thickness of each functional layer as a global collaborative optimization variable, and based on a complex refractive index database, the transmission matrix method is used to simulate the light field distribution of tandem organic solar cells, and obtain the photon absorptivity distribution corresponding to each functional layer. Based on the photon absorptivity distribution, the short-circuit current density distribution corresponding to each functional layer and the overall stacked structure is calculated to obtain the theoretical calculated value of the short-circuit current density of each sub-cell. With the goal of maximizing the sum of short-circuit current densities of each sub-cell as the optimization objective and with the constraint that the current matching error between sub-cells does not exceed a preset matching threshold, global collaborative optimization is performed on the parameter space formed by the thickness of each functional layer to obtain the initial optimal parameter combination. The perturbation amount is determined based on the thickness fluctuation range of each functional layer in the actual fabrication process, and the perturbation amount is applied to the initial optimal parameter combination to generate a perturbation parameter combination; Based on the current matching error and short-circuit current density attenuation amplitude under various perturbation parameter combinations, robust parameter combinations that meet the preset robust threshold constraints are selected. Establish a mapping relationship between robust parameter combinations and actual thin film deposition process parameters to generate a global collaborative process guidance scheme for the fabrication of tandem organic solar cells; During the execution of the process guidance scheme, online thickness monitoring data of each functional layer are collected simultaneously, and the thickness parameters are updated equivalently based on the thickness deviation. At the same time, the corresponding optical field simulation results and short-circuit current density prediction values ​​are calculated in real time. Based on the optical field simulation results and the predicted short-circuit current density, it is determined whether the current fabrication state meets the preset performance optimization standards. If the conditions are not met, the initial optimal parameter combination will be locally adaptively updated based on the current online film thickness deviation data, and incremental global collaborative optimization will be performed. If the conditions are met, the parameters of the current functional layer are locked, and the process guidance scheme is continued to be executed while maintaining the consistency of global collaborative constraints.

[0026] The aforementioned method for optimizing organic solar cell parameters based on optical field numerical simulation in this application establishes a complex refractive index database for each functional layer material and combines it with the transfer matrix method to simulate the optical field distribution, thereby achieving high-precision analysis of the internal light propagation process and photon absorption characteristics of tandem organic solar cells. By using the thickness of each functional layer as a global collaborative optimization variable and combining it with short-circuit current density and current matching constraints for global optimization, the overall photoelectric conversion performance of the tandem structure is improved. By introducing thickness perturbation analysis and robustness screening mechanisms, the impact of actual fabrication process fluctuations on device performance stability is reduced. On this basis, by combining online thickness monitoring, dynamic equivalent updates, and incremental collaborative optimization, real-time parameter compensation and process collaborative control are achieved during the fabrication process of tandem organic solar cells, thereby significantly improving the device's optical field modulation capability, parameter optimization accuracy, fabrication consistency, and mass production stability.

[0027] In some embodiments, the method for establishing a complex refractive index database is as follows: Spectral ellipsometry was performed on samples of each functional layer material to obtain the relationship between ellipsometry response data and wavelength variation data of each functional layer material under different incident conditions. Physical constraint consistency verification is performed on the elliptic response data, abnormal measurement data is identified and processed, and corrected optical measurement data is obtained. Based on the optical dispersion characteristics of different material types, the complex refractive index parameterization modeling and fitting solution are performed on the corrected optical measurement data to obtain the corresponding complex refractive index spectral data. The complex refractive index spectral data is reconstructed and discretized in the wavelength dimension to obtain a standardized complex refractive index dataset that meets the requirements of numerical simulation calculations. The standardized complex refractive index dataset is stored in a structured manner and linked with material preparation and measurement metadata to construct a traceable and updatable complex refractive index database.

[0028] For example, taking the hole transport layer PEDOT:PSS, the front cell active layer PM6:Y6, the electron transport layer PDINN, and the rear cell active layer D18:L8-BO in a tandem organic solar cell as examples, the corresponding material films are first prepared on a quartz substrate using a spin-coating process. The thickness of the PEDOT:PSS film is controlled to be about 30 nm, the thickness of the PM6:Y6 active layer film is controlled to be about 100 nm, and the thickness of the PDINN film is controlled to be about 10 nm. After the film preparation is completed, annealing is performed at 100℃~120℃ for 10 min~20 min to improve the film uniformity and material stability.

[0029] Subsequently, ellipsometry was used to measure the ellipsometry of each functional layer material. The measurement wavelength range was set to 300 nm to 900 nm, the wavelength sampling interval was set to 2 nm to 5 nm, and the incident angles were set to 55°, 65°, and 75°, respectively, to obtain the corresponding ellipsometry parameters under different incident angles. and Data curves as a function of wavelength. To reduce the impact of environmental noise and local film thickness inhomogeneity on the measurement results, multiple measurement areas were selected on each type of thin film for repeated measurements, and the measurement results were averaged.

[0030] After obtaining the original ellipsometric response data, the data consistency was first verified based on the Kramers-Kronig physical constraint relationship, and anomalies were identified and eliminated for measurement points with obvious jumps, abnormal drifts, or low signal-to-noise ratios. Subsequently, corresponding complex refractive index models were established based on the optical dispersion characteristics of different materials. For relatively transparent functional layer materials such as PEDOT:PSS, the Cauchy dispersion model can be used for fitting; for active layer materials with obvious absorption peaks such as PM6:Y6 and D18:L8-BO, the Tauc-Lorentz model can be used for parameterized fitting; for functional layer materials with free carrier absorption characteristics, a Drude term can be further introduced for joint fitting.

[0031] During the model fitting process, the least squares iterative algorithm is used to optimize the model parameters so that the mean square error between the theoretical elliptic response result and the actual measurement result converges to below a preset threshold, for example, the mean square error is less than 0.01. After the fitting is completed, the complex refractive index spectral data corresponding to each functional layer material are obtained, that is, the correspondence between the real refractive index n and the imaginary extinction coefficient k as a function of wavelength.

[0032] Furthermore, to meet the requirements of subsequent numerical simulation of the light field, the obtained complex refractive index spectral data were resampled and discretized according to a uniform wavelength step. For example, a standardized data table was generated according to a 1nm wavelength interval, and spline interpolation was used to complete the missing wavelength regions. Then, the standardized complex refractive index data was associated with and stored with meta-information such as material name, material ratio, film thickness, annealing temperature, solvent type, preparation date, measurement equipment number, and measurement environment temperature and humidity, forming a traceable and updatable complex refractive index database for subsequent simulation of light field distribution, light absorption analysis, and parameter co-optimization of tandem organic solar cells.

[0033] The method for establishing the complex refractive index database described in this application achieves high-precision acquisition of the optical response characteristics of materials by measuring the spectral ellipticity of each functional layer material and obtaining ellipticity response data; improves the accuracy and reliability of optical measurement data by performing physical constraint consistency verification and anomaly data processing on the ellipticity response data; achieves accurate characterization of the complex refractive index spectral data of each functional layer material by combining the optical dispersion characteristics of different material types for parametric modeling and fitting solution; and on this basis, constructs a traceable and updatable complex refractive index database by standardizing and reconstructing the complex refractive index spectral data, structuring its storage, and associating it with meta-information, thereby significantly improving the data accuracy, parameter consistency, and reliability of subsequent device optical optimization analysis in optical field numerical simulation.

[0034] In some embodiments, the method for simulating the light field distribution of tandem organic solar cells is as follows: Based on the electromagnetic field boundary conditions at the interface of each functional layer and the complex refractive index optical parameters of each functional layer, the interface matrix and the propagation matrix are constructed respectively. The interface matrix and the propagation matrix are concatenated according to the light propagation order to construct the total transmission matrix that characterizes the overall optical response relationship. Based on the relationship between the thickness of the functional layer and the coherence length, coherent and incoherent propagation are distinguished for each functional layer, and the local optical field distribution along the thickness direction inside each functional layer is solved based on the total transmission matrix. Based on the local light field distribution and the material absorption coefficients of each functional layer, the photon absorption power density at discrete positions along the thickness direction of each functional layer is calculated, and the photon absorption characteristics of each functional layer are obtained by integrating along the thickness direction. Based on the photon absorption characteristics of each functional layer, light absorption simulation results are generated to characterize the optical response capability of tandem organic solar cells.

[0035] For example, taking a dual-junction tandem organic solar cell as an example, such as Figure 3As shown, its structure sequentially includes a glass substrate, an ITO anode layer, a PEDOT:PSS hole transport layer, a front subcell active layer PM6:Y6, an intermediate composite layer, a rear subcell active layer D18:L8-BO, a PDINN electron transport layer, and an Ag cathode layer; wherein, the thickness of the PEDOT:PSS layer is 30nm, the thickness of the front subcell active layer is 100nm, the thickness of the intermediate composite layer is 10nm, the thickness of the rear subcell active layer is 120nm, and the thickness of the PDINN layer is 8nm.

[0036] First, the complex refractive index parameters n and k of each functional layer material in the wavelength range of 300 nm to 900 nm are retrieved from a pre-established complex refractive index database. Based on the Maxwell electromagnetic field boundary continuity condition, the interface matrix of each layer is established. Specifically, for the interface between any adjacent functional layer i and functional layer i+1, the reflection coefficient and transmission coefficient can be calculated based on the difference in complex refractive index of the materials on both sides, and the corresponding interface matrix is ​​constructed. Simultaneously, based on the thickness of each functional layer and its corresponding complex refractive index parameters, a propagation matrix corresponding to the light propagation within each layer is established to characterize the phase change and amplitude attenuation characteristics of light waves during propagation within the functional layers.

[0037] Subsequently, following the order in which light propagates from the glass substrate into the device interior, the interface matrix and propagation matrix corresponding to each functional layer are cascaded layer by layer, such as... Figure 4 As shown, the overall transfer matrix corresponding to the entire stacked structure is constructed. Based on the overall transfer matrix, the electric field distribution at various locations inside the device under different wavelength incident conditions can be solved.

[0038] Furthermore, for the PEDOT:PSS layer, PM6:Y6 layer, and D18:L8-BO layer, whose thickness is less than the optical coherence length, a coherent propagation model can be used to preserve the interference effect in the multilayer film system. For functional layers with larger thickness or significant scattering effects, an incoherent propagation model can be used for averaging to improve the stability of simulation results under complex structures. Subsequently, based on the total transfer matrix, the interior of each functional layer is spatially discretized along the thickness direction with a step size of 1nm to 2nm to obtain the local optical field intensity distribution at each discrete location under different wavelength conditions.

[0039] After obtaining the local light field distribution, the absorption coefficients of the corresponding materials in each functional layer are combined. The photon absorption power density at each discrete location is calculated; whereby the photon absorption power density can be expressed as the product of the local optical field intensity and the material absorption coefficient. Subsequently, the photon absorption power density along the thickness direction of the functional layer is integrated to obtain the optical absorption characteristics of each functional layer under different wavelength conditions.

[0040] For example, simulation results show that the active layer of the PM6:Y6 sub-cell has a higher light absorption peak in the wavelength region around 620 nm, while the active layer of the D18:L8-BO sub-cell exhibits stronger near-infrared light absorption in the wavelength region around 780 nm. Simultaneously, analysis of the local light field distribution reveals a significant light field enhancement region near the interface between the active layers of the front and rear sub-cells, which is beneficial for improving photon absorption efficiency in the corresponding wavelength range. Finally, based on the light absorption characteristics of each functional layer, simulation results of the overall tandem organic solar cell's light absorption are generated for subsequent short-circuit current density calculations, current matching analysis, and global parameter co-optimization processes.

[0041] The proposed simulation method for the light field distribution of tandem organic solar cells constructs interface and propagation matrices based on the electromagnetic field boundary conditions and complex refractive index optical parameters at the interfaces of each functional layer, thereby achieving accurate modeling of the light propagation process in multilayer optical film systems. By cascading the interface and propagation matrices, a total transmission matrix characterizing the overall optical response is established, improving the overall integrity and consistency of the light propagation analysis of the tandem structure. By combining the relationship between functional layer thickness and coherence length, coherent and incoherent propagation processes are distinguished, achieving high-precision solutions for the local light field distribution within the complex tandem structure. Furthermore, by performing integral analysis on the photon absorption power density within each functional layer, the photon absorption characteristics of each functional layer and the overall light absorption simulation results are obtained, significantly improving the accuracy of the light field distribution analysis, the optical response characterization capability, and the reliability of subsequent device parameter optimization for tandem organic solar cells.

[0042] In some embodiments, the method for obtaining the theoretically calculated value of the short-circuit current density of each sub-cell is as follows: The incident spectral data within the target wavelength range is acquired and photon flux conversion processing is performed to obtain the corresponding photon flux distribution. Based on the light absorption characteristics of each functional layer, the photon flux distribution is calculated by light absorption coupling to obtain the effective absorbed photon distribution of each functional layer. Based on the photoelectric conversion characteristics, the carrier conversion of the effectively absorbed photon distribution is calculated to obtain the short-circuit current density corresponding to each functional layer. The short-circuit current density of each functional layer is summarized and calculated to obtain the theoretical short-circuit current density of each sub-cell. The theoretical short-circuit current density is corrected based on the parasitic absorption loss of the non-photosensitive layer to obtain the theoretical calculated value of the short-circuit current density of each sub-cell.

[0043] For example, using the AM1.5G standard solar spectrum as the incident spectral data, the corresponding photon flux distribution in the wavelength range of 300nm to 900nm is calculated; combining the light absorption characteristics of the active layer PM6:Y6 of the front sub-cell and the active layer D18:L8-BO of the rear sub-cell, the number of effectively absorbed photons under each wavelength condition is calculated, and the absorbed photons are converted into corresponding carrier currents based on the photoelectric conversion efficiency; for example, the short-circuit current density corresponding to the front sub-cell is 15.8mA / cm. 2 The short-circuit current density corresponding to the secondary battery is 15.3 mA / cm². 2 Furthermore, the parasitic absorption losses generated by the PEDOT:PSS layer and electrode layer can be further corrected, ultimately yielding theoretical short-circuit current densities of 15.2 mA / cm² for the front and rear sub-cells. 2 With 14.9 mA / cm 2 This is used for subsequent current matching analysis and parameter co-optimization.

[0044] The theoretical calculation method for the short-circuit current density of each sub-cell described in this application achieves a quantitative characterization of the solar spectral energy distribution by performing photon flux conversion processing on the incident spectral data within the target wavelength range; it obtains the effective absorbed photon distribution corresponding to each functional layer by combining the light absorption characteristics of each functional layer for light absorption coupling calculation, thereby improving the accuracy of photon utilization process analysis; it achieves accurate solution of the short-circuit current density of each functional layer by performing carrier conversion calculation on the effectively absorbed photons based on photoelectric conversion characteristics; on this basis, by summarizing the short-circuit current densities of each functional layer and correcting for parasitic absorption losses, it obtains theoretically calculated values ​​of the sub-cell short-circuit current density that are closer to the actual device operating state, thereby significantly improving the accuracy of current matching analysis, photoelectric performance prediction capability, and reliability of subsequent parameter optimization results of tandem organic solar cells.

[0045] In some embodiments, the method for global collaborative optimization of the parameter space formed by the thicknesses of each functional layer is as follows: Parameter initialization sampling is performed within the parameter space composed of the thickness parameters of each functional layer to generate an initial parameter population for multi-objective optimization. Based on the functional layer thickness combination corresponding to each initial parameter population, optical and electrical response simulation calculations are performed to obtain the objective function evaluation results corresponding to each thickness combination. The initial parameter population is subjected to non-dominated sorting, and selection, crossover and mutation operations are performed sequentially on the sorted initial parameter population to generate a new generation of parameter population; During the population iteration process, the current matching constraint is used to determine the constraint of each thickness combination, and the fitness of the corresponding individual is adjusted according to the constraint satisfaction. The population iteration process is terminated based on preset convergence conditions, and after termination, the combination of thickness parameters with the best comprehensive performance is selected from the Pareto optimal solution set to obtain the global collaborative optimization result.

[0046] For example, the thicknesses of the PM6:Y6 active layer of the front sub-cell, the D18:L8-BO active layer of the rear sub-cell, and the PEDOT:PSS layer are used as variables to be optimized. The PM6:Y6 thickness range is set to 80nm–140nm, the D18:L8-BO thickness range is set to 90nm–150nm, and the PEDOT:PSS thickness range is set to 20nm–40nm. First, 100 initial thickness parameter combinations are randomly generated in the corresponding parameter space as the initial parameter population. Then, optical field distribution simulation and short-circuit current density calculation are performed on each parameter combination to obtain the corresponding objective function evaluation results.

[0047] The objective function is set to maximize the sum of the short-circuit current densities of the front and rear sub-cells, while constraining the current matching error between the front and rear sub-cells to not exceed 0.5 mA / cm. 2 Subsequently, the NSGA-II multi-objective optimization algorithm was used to perform non-dominated sorting on the initial parameter population, and selection, crossover and mutation operations were performed in sequence to generate a new generation of parameter population.

[0048] During the iteration process, when the short-circuit current density of the front and rear sub-cells corresponding to a certain thickness combination is 15.6 mA / cm, 2 With 15.1 mA / cm 2 When the current matching error meets the constraint, the fitness of the corresponding individual is increased; however, for parameter combinations where the current matching error exceeds a preset threshold, their fitness evaluation is reduced. After approximately 120 iterations, the optimization process terminates when the objective function's variation amplitude remains below a preset convergence threshold for multiple consecutive iterations. The optimal thickness parameter combination with the best overall performance is then selected from the Pareto optimal solution set. Examples include PM6:Y6 layer thickness of 102nm, D18:L8-BO layer thickness of 118nm, and PEDOT:PSS layer thickness of 28nm, which serve as target parameter combinations for subsequent device fabrication and process optimization. See also... Figure 2 The specific steps and population iteration logic of the above-mentioned global collaborative optimization process are as follows: Figure 2 As shown in the diagram on the left.

[0049] The global collaborative optimization method described in this application establishes a parameter population for multi-objective optimization by initial sampling within the parameter space composed of the thickness parameters of each functional layer, thus achieving a systematic search of the parameter space of complex stacked structures. By combining optical and electrical response simulations, the device performance corresponding to different thickness combinations is comprehensively evaluated, improving the ability of the parameter optimization process to characterize the overall photoelectric performance. Iterative optimization of thickness parameter combinations under multi-objective conditions is achieved by performing non-dominated sorting and selection, crossover, and mutation operations on the parameter population. On this basis, by introducing current matching constraints and fitness correction mechanisms, the influence of local optima on the optimization results is reduced, and the synergy and performance balance among sub-cells are improved. Finally, by comprehensively screening the Pareto optimal solution set, a global collaborative optimization result that takes into account both short-circuit current density and current matching characteristics is obtained, thereby significantly improving the parameter optimization accuracy, overall photoelectric conversion performance, and optimization result stability of stacked organic solar cells.

[0050] In some embodiments, the inter-cell current matching error is expressed by a current matching error function. A quantitative evaluation is conducted, and the current matching error function is defined as the ratio of the standard deviation of the short-circuit current density of each sub-cell to the mean of the short-circuit current density of all sub-cells. The preset matching threshold is dynamically set based on the number of sub-cells M of the tandem organic solar cell and the target power conversion efficiency level. When M equals 2, the default value of the preset matching threshold is set to 0.03, and when M equals 3, the default value is set to 0.04. When global collaborative optimization cannot find a parameter combination that makes the sum of short-circuit current densities meet the preset minimum performance index under the constraint of satisfying the current preset matching threshold, the preset matching threshold is relaxed in a step of 0.005 and the search is restarted. At the same time, an alarm prompt is output and a suggestion is given to evaluate whether to adjust the active layer material ratio.

[0051] In some embodiments, the method for determining the disturbance amount is as follows: Obtain the thickness control parameters of the thin film deposition process corresponding to each functional layer, and determine the thickness perturbation range of each functional layer based on the thickness fluctuation characteristics corresponding to different deposition processes. Based on the thickness perturbation range corresponding to each functional layer, the perturbation level of the thickness parameters of each functional layer is set, and the corresponding combination of thickness perturbation parameters is generated by experimental design method. Based on the number of functional layers, the combination of thickness perturbation parameters is optimized to obtain a representative combination of thickness parameters for performance analysis. The representative thickness parameters are combined and input into the preset optical and electrical response simulation model for calculation to obtain the corresponding current matching error and short-circuit current density. Based on the current matching error and short-circuit current density corresponding to each parameter combination, the robustness of the thickness parameters of each functional layer is evaluated, and the target thickness parameter combination that meets the preset performance requirements is selected.

[0052] For example, taking PM6:Y6 active layer, D18:L8-BO active layer and PEDOT:PSS layer as examples, the rotation speed fluctuation range, solution concentration error and evaporation rate fluctuation parameters in the corresponding spin coating process are obtained respectively, and the thickness perturbation range corresponding to each functional layer is determined by combining historical preparation data; for example, the thickness perturbation range of PM6:Y6 layer is set to ±5nm, the thickness perturbation range of D18:L8-BO layer is set to ±6nm, and the thickness perturbation range of PEDOT:PSS layer is set to ±2nm.

[0053] Subsequently, based on the aforementioned thickness perturbation range, an orthogonal experimental design method was used to generate multiple sets of thickness perturbation parameter combinations, and the parameter combinations were screened to reduce the computational complexity under high-dimensional combination conditions; for example, 60 representative parameter combinations were selected from the initially generated 300 sets of perturbation parameter combinations as performance analysis samples.

[0054] Furthermore, the representative thickness parameters are combined and input into the optical field distribution simulation model and the short-circuit current density calculation model for joint simulation to obtain the corresponding current matching error and short-circuit current density results. For example, when the thickness perturbation of the PM6:Y6 layer is +3nm and the thickness perturbation of the D18:L8-BO layer is -4nm, the corresponding short-circuit current densities of the front and rear sub-cells are 15.3mA / cm². 2 With 14.8 mA / cm 2 The current matching error is 0.5 mA / cm. 2 .

[0055] Finally, statistical analysis was performed on the performance fluctuation results under different disturbance conditions, and the combination of thickness parameters that can still maintain a high short-circuit current density and a small current matching error under thickness fluctuation conditions were selected as the target parameter combination to meet the preset robust performance requirements.

[0056] The method for determining the perturbation amount described in this application obtains the thickness control parameters of the corresponding thin film deposition process for each functional layer, and determines the corresponding thickness perturbation range by combining the thickness fluctuation characteristics of different deposition processes, thereby achieving a quantitative characterization of the fluctuation characteristics of the actual fabrication process. By setting the perturbation level of the thickness parameters of each functional layer and generating thickness perturbation parameter combinations using experimental design methods, the sample coverage capability under complex parameter fluctuation conditions is improved. By performing combination optimization processing on the thickness perturbation parameter combinations, representative performance analysis parameter combinations are obtained, reducing the computational complexity in the high-dimensional parameter analysis process. On this basis, by combining optical and electrical response simulation, the current matching error and short-circuit current density corresponding to each parameter combination are analyzed, thereby achieving a comprehensive evaluation of the stability of the thickness parameters of each functional layer, thus significantly improving the adaptability of the parameter optimization results of the tandem organic solar cell to process fluctuations, the robustness of device performance, and the consistency of actual fabrication.

[0057] In some embodiments, the method for selecting robust parameter combinations that satisfy the preset robustness threshold constraint is as follows: For each combination of thickness disturbance parameters, the current matching error and short-circuit current density are calculated, and the calculation results are compared with the performance results corresponding to the initial optimal parameter combination. Based on the current matching error and short-circuit current density results corresponding to various thickness disturbance parameter combinations, current matching robustness index and current performance robustness index are constructed respectively. Statistical analysis of performance fluctuations is performed on candidate parameter combinations that meet the preset robustness conditions to obtain statistical characteristic parameters that characterize performance stability. A comprehensive evaluation is conducted based on the average performance, performance stability, and worst-case performance of each candidate parameter combination, and robust parameter combinations that meet the preset robustness threshold constraints are selected.

[0058] For example, the PM6:Y6 layer thickness of 102nm, D18:L8-BO layer thickness of 118nm, and PEDOT:PSS layer thickness of 28nm obtained by global collaborative optimization are used as the initial optimal parameter combination, and multiple sets of perturbation parameter combinations are generated based on the corresponding thickness perturbation range. Then, the short-circuit current density and current matching error corresponding to each perturbation combination are calculated respectively, and the performance results are compared and analyzed with those corresponding to the initial optimal parameter combination.

[0059] For example, when the PM6:Y6 layer thickness perturbation is +4nm and the D18:L8-BO layer thickness perturbation is -3nm, the corresponding short-circuit current densities of the front and rear sub-cells are 15.1mA / cm². 2 With 14.7 mA / cm 2 The current matching error is 0.4 mA / cm. 2When the PM6:Y6 layer thickness perturbation exceeds +8nm, the current matching error between the front and rear sub-cells increases to 0.9mA / cm. 2 The match exceeds the preset matching threshold.

[0060] Furthermore, based on the performance results corresponding to each combination of disturbance parameters, current matching robustness index and current performance robustness index are constructed respectively. The current matching robustness index characterizes the fluctuation degree of current matching error under different disturbance conditions, while the current performance robustness index characterizes the ability to maintain short-circuit current density. Subsequently, statistical analysis is performed on candidate parameter combinations that meet the preset robustness conditions to obtain corresponding statistical characteristic parameters such as average short-circuit current density, standard deviation, and performance degradation under worst-case conditions.

[0061] For example, when a certain set of parameters maintains an average short-circuit current density of 14.9 mA / cm under all disturbance conditions. 2 The above conditions apply, and the standard deviation of the current matching error is less than 0.15 mA / cm. 2 If the parameter combination is deemed to have good robustness, then a comprehensive evaluation is conducted based on average performance, performance stability, and worst-case performance to select robust parameter combinations that meet the preset robustness threshold constraints for use in subsequent device fabrication and process parameter mapping processes. See also Figure 2 The selection logic for the above robust parameter combinations and the relationship between each indicator are as follows: Figure 2 As shown in the diagram on the right.

[0062] The robust parameter combination screening method described in this application calculates the current matching error and short-circuit current density corresponding to each thickness disturbance parameter combination and compares them with the performance results of the initial optimal parameter combination to achieve quantitative analysis of the device performance variation law under different process fluctuation conditions. By constructing current matching robustness index and current performance robustness index respectively, the comprehensive characterization ability of the parameter stability and performance retention capability of tandem organic solar cells is improved. By performing performance fluctuation statistical analysis on candidate parameter combinations that meet the preset robustness conditions, statistical characteristic parameters characterizing performance stability are obtained, realizing a refined evaluation of the device performance fluctuation characteristics. On this basis, by combining average performance, performance stability and worst-case performance for comprehensive evaluation, robust parameter combinations that meet the preset robustness threshold constraints are screened, thereby significantly improving the resistance to process fluctuations, performance stability and actual mass production adaptability of the optimized parameter results of tandem organic solar cells.

[0063] In some embodiments, the method for establishing the mapping relationship between robust parameter combinations and actual thin film deposition process parameters is as follows: For different types of thin film deposition processes, a thickness mapping model between process parameters and thin film deposition thickness is constructed, and the model parameters of the thickness mapping model are determined based on experimental calibration data. Based on the target thickness parameters corresponding to each functional layer, the thickness mapping model of the corresponding deposition process is called to solve the corresponding combination of process parameters; For situations where multiple candidate deposition processes correspond to the same functional layer, the candidate deposition processes are matched and screened based on the target thickness requirements, thickness uniformity requirements, and deposition efficiency requirements to determine the target deposition process. Based on the target deposition process, process window constraint information is introduced into the corresponding thickness mapping model, and the adjustable range and parameter sensitivity characteristics of each process parameter are determined. Based on the target deposition process and the corresponding combination of process parameters, a mapping relationship is established between the robust parameter combination and the actual thin film deposition process parameters.

[0064] For example, taking the PM6:Y6 active layer using spin coating, the PEDOT:PSS layer using slot coating, and the Ag electrode layer using thermal evaporation as examples, corresponding thickness mapping models are established respectively; among them, for the PM6:Y6 active layer, a spin coating thickness mapping model can be established based on the relationship between spin coating speed, solution concentration, spin coating time and film thickness, and the model parameters are determined through experimental calibration.

[0065] For example, in the PM6:Y6 material system, when the solution concentration is 16 mg / mL, the spin coating speed is 3000 rpm, and the spin coating time is 40 s, an active layer film with a thickness of about 102 nm is formed. Furthermore, based on multiple sets of experimental data, a fitting relationship between spin coating parameters and film thickness is established to achieve the reverse solution of process parameters corresponding to the target thickness.

[0066] For the PEDOT:PSS layer, when the target thickness is 28nm, the thickness uniformity and deposition efficiency of spin coating and slot coating processes can be evaluated separately. If the slot coating process has higher film thickness uniformity and is suitable for continuous preparation, then the slot coating process is determined as the target deposition process, and the corresponding coating speed, flow rate parameters and drying temperature range are further determined.

[0067] Furthermore, process window constraint information is introduced into the corresponding thickness mapping model. For example, the adjustable range of the spin coating speed of PM6:Y6 active layer is set to 2500rpm~3500rpm, and the fluctuation range of solution concentration is set to ±0.5mg / mL. The sensitivity characteristics of different process parameters to film thickness changes are analyzed in combination with historical process data. Among them, when the spin coating speed changes by 100rpm, the corresponding film thickness change is about 2nm, so the spin coating speed is determined to be a highly sensitive process parameter.

[0068] Finally, based on the target deposition process and the corresponding combination of process parameters, a mapping relationship between the robust parameter combination and the actual thin film deposition process parameters is established. For example, "PM6:Y6 layer thickness 102nm" is mapped to the process parameter combination of "solution concentration 16mg / mL, spin coating speed 3000rpm, spin coating time 40s" for subsequent device fabrication and online process control.

[0069] The method for establishing the mapping relationship between the robust parameter combination and the actual thin film deposition process parameters described in this application constructs a thickness mapping model between process parameters and thin film deposition thickness for different types of thin film deposition processes, thereby achieving a correlation characterization between the target thickness of the functional layer and the actual process parameters. By combining experimental calibration data to determine the parameters of the thickness mapping model, the accuracy and feasibility of the process parameter solution results are improved. Multiple candidate deposition processes are matched and screened based on target thickness requirements, thickness uniformity requirements, and deposition efficiency requirements, achieving synergistic adaptation of different functional layer deposition processes. Furthermore, by introducing process window constraint information into the thickness mapping model and determining the adjustable range and parameter sensitivity characteristics of each process parameter, the adaptability of the process parameter configuration to actual fabrication fluctuations is improved. This achieves an effective mapping between the robust parameter combination and the actual thin film deposition process parameters, significantly improving the process feasibility, fabrication consistency, and mass production stability of the optimized parameters for tandem organic solar cells.

[0070] In some embodiments, the method for equivalently updating the thickness parameter based on the thickness deviation is as follows: After the deposition of each functional layer is completed, the actual thickness data of the corresponding functional layer output by the online thickness monitoring system is obtained; Based on the actual thickness data, the theoretical thickness parameters of the corresponding functional layers in the simulation model are replaced and updated, and the propagation matrix parameters of the corresponding functional layers are updated synchronously. The thickness parameters of the functional layers that have been deposited are locked, and the thickness parameters of the functional layers that have not been deposited are retained as variables to be optimized, so as to construct a hybrid parameter model that includes the measured thickness parameters and the thickness parameters to be optimized. Based on the hybrid parameter model, the optical and electrical response simulation calculations were re-executed to obtain the updated light field distribution, short-circuit current density of each sub-cell, and current matching error results. Based on the updated current matching error results, the thickness parameters of subsequent functional layers are adjusted to compensate for the errors, and the equivalent update process and corresponding calculation results are recorded to complete the dynamic equivalent update of the thickness parameters.

[0071] For example, after the active layer of the PM6:Y6 pre-cell battery is deposited, the actual film thickness is detected by an online ellipticity monitoring system as 106 nm, while the corresponding theoretically optimized thickness is 102 nm. Subsequently, the PM6:Y6 layer thickness parameter in the simulation model is updated from 102 nm to 106 nm, and the phase propagation parameter in the corresponding propagation matrix is ​​updated simultaneously.

[0072] Furthermore, the thickness parameter of the already deposited PM6:Y6 layer was locked at 106 nm, while the thickness of the undeposited D18:L8-BO active layer and PDINN layer was retained as variables to be optimized, thus constructing a hybrid parameter model that includes the measured thickness parameter and the parameter to be optimized; then, the optical field distribution simulation and short-circuit current density calculation were re-executed based on the updated model.

[0073] For example, the updated simulation results show that the short-circuit current density of the front cell has increased to 15.5 mA / cm². 2 Then the short-circuit current density of the sub-cell dropped to 14.7 mA / cm². 2 This resulted in an increase in the current matching error between the front and rear sub-cells to 0.8 mA / cm. 2 Based on this result, the target thickness of the subsequent D18:L8-BO layer was adjusted to compensate for the loss. For example, the original target thickness was adjusted from 118nm to 123nm to improve the light absorption capability of the corresponding wavelength band of the subsequent sub-cell.

[0074] During subsequent functional layer deposition, the equivalent update process described above is repeated, and the thickness update values, current matching error changes, and compensation adjustment parameters for each update are recorded and stored to achieve dynamic parameter correction and synergistic optimization in the fabrication process of tandem organic solar cells.

[0075] The thickness parameter equivalent update method based on thickness deviation described in this application achieves real-time perception of the actual deposition state of tandem organic solar cells by acquiring the actual thickness data of the functional layers output by an online thickness monitoring system. It improves the consistency between the light field simulation results and the actual fabrication state by synchronously updating the theoretical thickness parameters and corresponding propagation matrix parameters in the simulation model based on the actual thickness data. Furthermore, it constructs a hybrid parameter model containing both measured and optimized parameters by locking the thickness parameters of the completed functional layers and retaining the thickness parameters of the incomplete functional layers as variables to be optimized, thus achieving dynamic collaborative optimization during the fabrication process. Based on this, it re-executes optical and electrical response simulation calculations and adjusts the subsequent functional layer thickness parameters by combining the updated current matching error results, achieving dynamic correction of the impact of process deviations. This significantly improves the parameter control accuracy, photoelectric performance stability, and process adaptive optimization capability during the fabrication process of tandem organic solar cells.

[0076] In some embodiments, the method for setting performance optimization criteria is as follows: Set a lower limit for the short-circuit current density and construct short-circuit current performance judgment conditions based on the real-time predicted short-circuit current density results; Set an acceptable upper limit for the current matching error between each sub-cell, and construct current matching performance judgment conditions based on the current matching error; Based on the correlation between short-circuit current density, open-circuit voltage, and fill factor, the expected energy conversion efficiency of the device is estimated, and efficiency performance judgment conditions are constructed. Based on the current matching performance judgment condition, short-circuit current performance judgment condition, and efficiency performance judgment condition, a hierarchical performance judgment rule and a comprehensive evaluation rule are constructed, and each performance index is normalized and weighted. Based on the grading and comprehensive evaluation results, performance optimization standards are determined, and the performance indicators and grading results are recorded and stored.

[0077] For example, taking a dual-junction tandem organic solar cell as an example, the acceptable lower limit of the short-circuit current density of each sub-cell can be set to 14.5 mA / cm². 2 The upper limit of the current matching error between the front and rear sub-cells is set to 0.5mA / cm. 2 The real-time prediction results showed that the short-circuit current density of the front battery was 15.2 mA / cm². 2 The short-circuit current density of the secondary battery is 14.9 mA / cm². 2 If the short-circuit current performance meets the preset requirements, then it is determined that the short-circuit current performance meets the preset requirements.

[0078] Furthermore, based on the correlation between short-circuit current density, open-circuit voltage, and fill factor, the expected energy conversion efficiency of the device is estimated. For example, when the device open-circuit voltage is 1.85V and the fill factor is 0.78, the corresponding predicted energy conversion efficiency is approximately 21.5%, and the efficiency qualification threshold is set to above 20%.

[0079] Subsequently, a comprehensive evaluation rule is constructed based on short-circuit current performance, current matching performance, and efficiency performance. For example, the short-circuit current density, current matching error, and prediction efficiency are assigned weight coefficients of 0.4, 0.3, and 0.3, respectively, and the comprehensive evaluation score is calculated after normalizing the corresponding performance indicators.

[0080] When the comprehensive evaluation score is higher than the preset threshold, the current preparation state is determined to meet the performance optimization standard, and the subsequent functional layer deposition is continued; otherwise, a local adaptive update and incremental global collaborative optimization process is triggered to dynamically compensate and adjust the parameters of the subsequent functional layers.

[0081] The performance optimization standard setting method described in this application achieves constraint control over key photoelectric performance indicators of tandem organic solar cells by setting a lower acceptable limit for short-circuit current density and an upper acceptable limit for current matching error. By combining real-time predicted short-circuit current density and current matching error results, short-circuit current performance judgment conditions and current matching performance judgment conditions are constructed respectively, improving the real-time performance and accuracy of device operation status evaluation. The expected energy conversion efficiency of the device is estimated based on the correlation between short-circuit current density, open-circuit voltage, and fill factor, achieving quantitative analysis of the overall device performance. Furthermore, by constructing hierarchical performance judgment rules and comprehensive evaluation rules, and by normalizing and weighting each performance indicator, synergistic evaluation and unified judgment among multi-dimensional performance indicators are achieved, thereby significantly improving the comprehensiveness of tandem organic solar cell performance evaluation, the accuracy of optimization judgment, and the performance control capability during the fabrication process.

[0082] In some embodiments, the local adaptive update strategy specifically includes: Based on sensitivity analysis, the undeposited functional layer that has the greatest impact on the target performance is identified. Sensitivity is characterized by the change in short-circuit current density or current matching error caused by a unit change in the thickness of the functional layer. For the identified key functional layers, a local search space is constructed based on their initial optimization range, and the search boundary is set to ±10% to ±20% of the current thickness, while the constraint does not exceed the preset physical feasible range, wherein the minimum thickness is not less than 5nm and the maximum thickness does not exceed 3 times the material feature diffusion length; Gradient-based optimization algorithms are used to perform local optimization calculations within the constructed local search space, with the gradient information of the objective function serving as the basis for the search direction. During the local optimization process, the thickness parameters of the deposited functional layers are kept fixed, and only the thickness parameters of the undeposited functional layers are adjusted and optimized. Set the upper limit of local optimization iterations to 30-50 times, and terminate the optimization process early when the improvement of the objective function is less than 0.1% for several consecutive iterations; If the local optimization yields a parameter combination that meets the performance judgment criteria, then update the functional layer thickness setting parameters in the corresponding process guidance scheme. If a feasible solution is not obtained within the iteration limit of the local optimization, the constraints are evaluated, and the constraints are relaxed or the parameters of the previously fixed functional layers are adjusted back according to the evaluation results. If necessary, manual decision prompts are output. The entire process of local adaptive update is recorded, including at least the triggering conditions, parameter adjustment range, optimization iteration trajectory, and final decision result.

[0083] For example, taking a dual-junction tandem organic solar cell as an example, after the PM6:Y6 layer of the front cell is deposited, sensitivity analysis revealed that the thickness of the active layer D18:L8-BO of the back cell has the greatest impact on the current matching error, with a sensitivity coefficient of 0.18 mA·cm. -2 / nm, higher than other undeposited functional layers.

[0084] Accordingly, a local search space was constructed based on the initial optimized thickness of 118 nm for the D18:L8-BO layer, limiting the search range to 106 nm to 130 nm, while satisfying the physical constraints that the minimum thickness is not less than 5 nm and the maximum thickness does not exceed three times the material diffusion length. Subsequently, an optimization algorithm based on gradient descent was used for iterative search within this local space, with the thickness parameters updated guided by minimizing the current matching error.

[0085] For example, in the 12th iteration, the improvement in the objective function decreased to 0.08%, satisfying the early termination condition; finally, a parameter combination with an optimized thickness of 122nm was obtained, which increased the short-circuit current density of the front and rear sub-cells to 15.4mA / cm. 2 With 15.2 mA / cm 2 The current matching error was reduced to 0.2 mA / cm. 2 It meets the performance judgment criteria and updates the D18:L8-BO layer thickness setting parameters in the process guidance scheme simultaneously.

[0086] If local optimization fails to converge within 50 iterations under certain extreme conditions, the current matching constraint tension is assessed, and the constraint is appropriately relaxed or the previously fixed functional layer thickness parameters are adjusted retrospectively. Manual decision prompts are triggered when necessary, and the triggering conditions, iteration trajectory, and parameter adjustment range of the entire local optimization process are fully recorded and stored.

[0087] The aforementioned local adaptive update strategy of this application achieves rapid location of key influencing parameters by identifying the critical undeposited functional layers that have the greatest impact on short-circuit current density and current matching error through sensitivity analysis. It improves the adaptability of the local optimization process to the constraints of the actual device structure by constructing a local search space constrained by the current thickness parameters within a physically feasible range. Furthermore, it achieves rapid convergence optimization of the undeposited functional layer thickness parameters by employing gradient-based optimization algorithms within the local search space and dynamically adjusting the search direction based on the gradient information of the objective function. On this basis, it reduces the disturbance to the existing fabrication state by keeping the parameters of the deposited functional layers fixed and adjusting only the undeposited functional layers. Simultaneously, it improves the stability and feasibility of the local optimization results by combining iteration termination conditions, constraint evaluation mechanisms, and anomaly backtracking strategies. Finally, it achieves traceable management of the optimization decision-making process by recording the entire local adaptive update process, thereby significantly improving the dynamic parameter compensation capability, process adaptive optimization capability, and device performance stability in the fabrication process of tandem organic solar cells.

[0088] In some embodiments, the method for performing incremental global collaborative optimization is as follows: Taking the current preparation stage as a new starting point for optimization, the thickness of the already deposited functional layer is used as a fixed parameter input, and the thickness of the undeposited functional layer is used as a variable to be optimized, so as to construct an incremental optimization problem; The optimization model is reconstructed based on the current preparation state, where the objective function is to maximize the sum of the short-circuit current densities of each sub-cell, and the constraint is that the current matching error does not exceed a preset threshold. Based on the reduction of the dimension of the variable to be optimized, the population size is adaptively adjusted and set to 50-100 individuals. A hot-start strategy is adopted to select the solution that best matches the currently fixed parameters from the Pareto optimal solution set obtained from the initial global optimization, and extract its unfixed layer thickness parameter as the initial population center for initialization; A compensation weight term is introduced into the objective function to guide the optimization process to search for parameter combinations that can offset the effects of previous preparation deviations. When the current of a certain sub-cell decreases due to the deviation of the previous layer, the weight of the corresponding sub-cell current in the objective function is increased. During the optimization process, the current matching constraint and robustness constraint remain unchanged to ensure the consistency between the incremental optimization results and the global optimization objective. The upper limit of incremental optimization iteration is set to 100-150 generations, and the optimization process is terminated early when the improvement of the objective function is less than 0.2% for 20 consecutive generations. After optimization, the updated global collaborative parameter scheme is output, and the process parameters of the undeposited functional layers in the process guidance scheme are revised simultaneously. The incremental global collaborative optimization process can be repeatedly triggered after the deposition of each functional layer to achieve dynamic adaptive optimization based on the current preparation state.

[0089] For example, in the fabrication process of a dual-junction tandem organic solar cell, the deposition of the PM6:Y6 front cell active layer and the PEDOT:PSS layer has been completed, and the actual thickness of PM6:Y6 has been detected as 106 nm (a deviation of +4 nm from the target of 102 nm). Based on this fabrication state, the thickness of the deposited layers is fixed as input, and the thicknesses of the D18:L8-BO layer and subsequent functional layers are used as variables to be optimized to construct an incremental optimization problem.

[0090] Under the hot start strategy, the solution closest to the current fixed parameters is selected from the Pareto solution set obtained from the initial global optimization, and its corresponding undeposited layer thickness (e.g., 118 nm for D18:L8-BO) is extracted as the initial population center. Based on this, 50 to 100 individuals are generated to form the initial population.

[0091] Meanwhile, since the short-circuit current of the front sub-cell decreases due to the thickness deviation, the weighting coefficient of the front sub-cell current is increased in the objective function to guide the optimization process to prioritize compensating for the impact of this deviation. After approximately 120 iterations of optimization, the optimized thickness of the D18:L8-BO layer is approximately 123 nm, which increases the short-circuit current density of the rear sub-cell to 15.3 mA / cm². 2 The current matching error is controlled within 0.3 mA / cm. 2 Within.

[0092] Finally, the updated incremental global collaborative parameter scheme is output, and the process parameters of subsequent undeposited functional layers are revised simultaneously, so that the subsequent deposition process can continuously and adaptively optimize based on the current actual fabrication state, thereby realizing dynamic closed-loop collaborative control of the stacked device fabrication process.

[0093] The incremental global collaborative optimization method described in this application uses the current fabrication stage as a new optimization starting point, with the thickness of the deposited functional layer as a fixed parameter and the thickness of the undeposited functional layer as the variable to be optimized. This constructs an incremental optimization model that matches the actual fabrication state, achieving dynamic collaborative optimization during the fabrication process of tandem organic solar cells. By reconstructing an optimization model based on the current fabrication state, with the maximization of short-circuit current density and current matching error constraints as its core, the method improves the adaptability of the optimization results to the current device state. Furthermore, by combining variable dimensionality reduction with adaptive adjustment of the optimization population size, and employing a hot-start strategy to extract matching parameters from the initial global optimization results, the method achieves this. Using a number as the initial population center improves the search efficiency and convergence speed of the incremental optimization process. Based on this, by introducing a compensation weight term into the objective function, the optimization process is guided to dynamically compensate for previous fabrication deviations, while maintaining consistency in current matching constraints and robustness constraints, ensuring coordination and unity between the incremental optimization results and the overall global optimization objective. Finally, by synchronously revising the process parameters of the undeposited functional layers and repeatedly triggering incremental optimization after each functional layer is deposited, dynamic adaptive collaborative optimization based on real-time fabrication status is achieved, thereby significantly improving the parameter compensation capability, global optimization consistency, and device performance stability during the fabrication of tandem organic solar cells.

[0094] In some embodiments, the method for maintaining consistency of global collaborative constraints is as follows: After the deposition of each functional layer is completed and the corresponding thickness parameters are locked, the current fixed parameter combination is obtained and together with the parameter space to be optimized for the remaining undeposited functional layers to construct the current global parameter state. Based on the fixed parameter combination and current matching constraints, the constraint propagation algorithm is used to reverse derive the feasible domain of the thickness parameters of the remaining undeposited functional layers in order to obtain the parameter boundary range that satisfies the constraints. The feasible domain of the thickness parameters of the remaining undeposited functional layers is updated based on the parameter boundary range, and infeasible parameter intervals that violate the constraints are removed. The feasibility of the updated parameter feasible region is determined. When there is a non-empty feasible region in the remaining parameter space, the global collaborative constraint consistency is determined to be valid, and the next functional layer deposition is carried out. When no feasible solution exists in the remaining parameter space, a constraint conflict determination is triggered, and a constraint conflict warning message is output. Under conflict constraints, the recently deposited and locked functional layers are fine-tuned or equivalently modified to restore the feasible domain of parameters. When fine-tuning fails to restore feasibility, the current matching constraint threshold or the material and structural parameters of subsequent functional layers are adjusted to restore global constraint consistency. A constraint violation index is constructed based on the normalized distance between the current parameter combination and the constraint boundary, and the constraint violation index is monitored and warned in real time. Before each functional layer is deposited, a feedforward evaluation mechanism is used to analyze the impact of the current layer thickness variation on the subsequent constraint space, and priority constraint management is carried out on the thickness control accuracy of key sensitive functional layers.

[0095] For example, in the fabrication process of a dual-junction tandem organic solar cell, the PM6:Y6 front cell and PEDOT:PSS layer have been deposited, and their thickness parameters are locked at 106 nm and 28 nm, respectively. At this point, the fixed parameters are combined with the undeposited D18:L8-BO layer to construct the current global parameter state, based on current matching constraints (error ≤ 0.5 mA / cm²). 2 The feasible region of the remaining parameters is derived in reverse.

[0096] For example, through constraint propagation calculations, it was found that the feasible range of the D18:L8-BO layer thickness was reduced from the original 118nm±12nm to 112nm~126nm to meet the current matching requirements of the preceding and following sub-cells. After eliminating infeasible intervals that exceed this range, if the remaining parameter space is still a non-empty set, it is determined that the global cooperative constraints remain consistent, and subsequent layer deposition continues.

[0097] If a preceding bias is introduced under certain circumstances (such as a PM6:Y6 thickness bias of +6nm), causing the feasible region to shrink to an empty set after constraint propagation, a constraint conflict warning will be triggered. In this case, the locked PM6:Y6 layer will be fine-tuned first, for example, the thickness will be adjusted from 106nm to 104nm, in order to restore the parameter feasible region.

[0098] If fine-tuning still fails to restore feasibility, the current matching error threshold is further relaxed or the material system of subsequent functional layers is adjusted to ensure that the overall constraint system regains a feasible solution space. At the same time, the degree of constraint violation is evaluated in real time by calculating the normalized distance between the current parameter combination and the boundary of the feasible region, and the thickness control accuracy requirements for sensitive functional layers near the boundary (such as the D18:L8-BO layer) are increased.

[0099] Before each functional layer is deposited, a feedforward evaluation mechanism is used to predict the impact of the current layer thickness perturbation on the subsequent constraint space, thereby identifying potential conflict risks in advance and implementing a more stringent thickness control strategy for key sensitive layers to ensure that the global collaborative constraint remains consistent throughout the entire fabrication process.

[0100] The method described in this application for maintaining global collaborative constraint consistency achieves global constraint collaborative management during the fabrication process of tandem organic solar cells by obtaining fixed parameter combinations after the deposition of each functional layer and constructing the current global parameter state in conjunction with the parameter space to be optimized for the remaining undeposited functional layers. It also achieves dynamic analysis and constraint updates for the feasibility of subsequent parameter spaces by using a constraint propagation algorithm based on current matching constraints to deduce the feasible domain of the remaining functional layer thickness parameters. Furthermore, it improves the ability of the global optimization process to maintain constraints on complex parameter coupling relationships by eliminating infeasible parameter intervals that violate constraints and performing real-time feasibility judgments on the updated parameter feasible domains. Based on this, it achieves adaptive recovery and coordinated control of global constraint conflict problems through constraint conflict early warning, functional layer fine-tuning correction, and dynamic adjustment of constraint thresholds. Simultaneously, it prioritizes the management of the thickness control accuracy of key sensitive functional layers by combining constraint violation index monitoring and feedforward evaluation mechanisms, thereby significantly improving the global collaborative optimization consistency, parameter constraint stability, and process dynamic adaptive control capabilities during the fabrication process of tandem organic solar cells.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the parameters of organic solar cells based on numerical simulation of light fields, characterized in that, Includes the following steps: Using the thickness of each functional layer as a global collaborative optimization variable, and based on a complex refractive index database, the light field distribution of the stacked organic solar cell is simulated to obtain the photon absorptivity distribution of each functional layer. Based on the photon absorptivity distribution, the short-circuit current density distribution corresponding to each functional layer and the overall stacked structure is calculated to obtain the theoretical calculated value of the short-circuit current density of each sub-cell. Global collaborative optimization is performed on the parameter space formed by the thickness of each functional layer to obtain the initial optimal parameter combination; The perturbation amount is determined based on the thickness fluctuation range of each functional layer in the actual fabrication process, and the perturbation amount is applied to the initial optimal parameter combination to generate a perturbation parameter combination; Based on the current matching error and short-circuit current density attenuation amplitude under various perturbation parameter combinations, robust parameter combinations that meet the preset robust threshold constraints are selected. A mapping relationship between robust parameter combinations and actual thin film deposition process parameters is established to generate a global collaborative process guidance scheme for the fabrication of tandem organic solar cells.

2. The method for optimizing organic solar cell parameters according to claim 1, characterized in that, The method for simulating the light field distribution of tandem organic solar cells is as follows: Based on the electromagnetic field boundary conditions at the interface of each functional layer and the complex refractive index optical parameters of each functional layer, the interface matrix and the propagation matrix are constructed respectively. The interface matrix and the propagation matrix are concatenated according to the light propagation order to construct the total transmission matrix that characterizes the overall optical response relationship. Based on the relationship between the thickness of the functional layer and the coherence length, coherent and incoherent propagation are distinguished for each functional layer, and the local optical field distribution along the thickness direction inside each functional layer is solved based on the total transmission matrix. Calculate the photon absorption power density at discrete positions along the thickness direction of each functional layer, and integrate the photon absorption characteristics of each functional layer along the thickness direction. Based on the photon absorption characteristics of each functional layer, light absorption simulation results are generated to characterize the optical response capability of tandem organic solar cells.

3. The method for optimizing organic solar cell parameters according to claim 1, characterized in that, The method for obtaining the theoretical calculation value of the short-circuit current density of each sub-cell is as follows: The incident spectral data within the target wavelength range is acquired and photon flux conversion processing is performed to obtain the corresponding photon flux distribution. Based on the light absorption characteristics of each functional layer, the photon flux distribution is calculated by light absorption coupling to obtain the effective absorbed photon distribution of each functional layer. Based on the photoelectric conversion characteristics, the carrier conversion of the effectively absorbed photon distribution is calculated to obtain the short-circuit current density corresponding to each functional layer. The short-circuit current density of each functional layer is summarized and calculated to obtain the theoretical short-circuit current density of each sub-cell. The theoretical short-circuit current density is corrected based on the parasitic absorption loss of the non-photosensitive layer to obtain the theoretical calculated value of the short-circuit current density of each sub-cell.

4. The method for optimizing organic solar cell parameters according to claim 1, characterized in that, The method for global collaborative optimization of the parameter space formed by the thickness of each functional layer is as follows: Parameter initialization sampling is performed within the parameter space composed of the thickness parameters of each functional layer to generate an initial parameter population for multi-objective optimization. Based on the functional layer thickness combination corresponding to each initial parameter population, optical and electrical response simulation calculations are performed to obtain the objective function evaluation results corresponding to each thickness combination. The initial parameter population is subjected to non-dominated sorting, and selection, crossover and mutation operations are performed sequentially on the sorted initial parameter population to generate a new generation of parameter population; During the population iteration process, the current matching constraint is used to determine the constraint of each thickness combination, and the fitness of the corresponding individual is adjusted according to the constraint satisfaction. The population iteration process is terminated based on preset convergence conditions, and after termination, the combination of thickness parameters with the best comprehensive performance is selected from the Pareto optimal solution set to obtain the global collaborative optimization result.

5. The method for optimizing organic solar cell parameters according to claim 1, characterized in that, The method for determining the disturbance is as follows: Obtain the thickness control parameters of the thin film deposition process corresponding to each functional layer, and determine the thickness perturbation range of each functional layer based on the thickness fluctuation characteristics corresponding to different deposition processes. Based on the thickness perturbation range corresponding to each functional layer, the perturbation level of the thickness parameters of each functional layer is set, and the corresponding combination of thickness perturbation parameters is generated by experimental design method. Based on the number of functional layers, the combination of thickness perturbation parameters is optimized to obtain a representative combination of thickness parameters for performance analysis. The representative thickness parameters are combined and input into the preset optical and electrical response simulation model for calculation to obtain the corresponding current matching error and short-circuit current density. Based on the current matching error and short-circuit current density corresponding to each parameter combination, the robustness of the thickness parameters of each functional layer is evaluated, and the target thickness parameter combination that meets the preset performance requirements is selected.

6. The method for optimizing organic solar cell parameters according to claim 1, characterized in that, The method for selecting robust parameter combinations that satisfy the preset robustness threshold constraint is as follows: For each combination of thickness disturbance parameters, the current matching error and short-circuit current density are calculated, and the calculation results are compared with the performance results corresponding to the initial optimal parameter combination. Based on the current matching error and short-circuit current density results corresponding to various thickness disturbance parameter combinations, current matching robustness index and current performance robustness index are constructed respectively. Statistical analysis of performance fluctuations is performed on candidate parameter combinations that meet the preset robustness conditions to obtain statistical characteristic parameters that characterize performance stability. A comprehensive evaluation is conducted based on the average performance, performance stability, and worst-case performance of each candidate parameter combination, and robust parameter combinations that meet the preset robustness threshold constraints are selected.

7. The method for optimizing organic solar cell parameters according to claim 1, characterized in that, The method for establishing the mapping relationship between robust parameter combinations and actual thin film deposition process parameters is as follows: For different types of thin film deposition processes, a thickness mapping model between process parameters and thin film deposition thickness is constructed, and the model parameters of the thickness mapping model are determined based on experimental calibration data. Based on the target thickness parameters corresponding to each functional layer, the thickness mapping model of the corresponding deposition process is called to solve the corresponding combination of process parameters; For situations where multiple candidate deposition processes correspond to the same functional layer, the candidate deposition processes are matched and screened based on the target thickness requirements, thickness uniformity requirements, and deposition efficiency requirements to determine the target deposition process. Based on the target deposition process, process window constraint information is introduced into the corresponding thickness mapping model, and the adjustable range and parameter sensitivity characteristics of each process parameter are determined. Based on the target deposition process and the corresponding combination of process parameters, a mapping relationship is established between the robust parameter combination and the actual thin film deposition process parameters.

8. The method for optimizing the parameters of organic solar cells according to any one of claims 1-7, characterized in that, The method for optimizing organic solar cell parameters also includes: During the execution of the process guidance scheme, online thickness monitoring data of each functional layer is collected simultaneously, and the thickness parameters are updated accordingly based on the thickness deviation. At the same time, the corresponding optical field simulation results and short-circuit current density prediction values ​​are calculated in real time.

9. The method for optimizing the parameters of an organic solar cell according to claim 8, characterized in that, The method for equivalently updating thickness parameters based on thickness deviation is as follows: After the deposition of each functional layer is completed, the actual thickness data of the corresponding functional layer output by the online thickness monitoring system is obtained; Based on the actual thickness data, the theoretical thickness parameters of the corresponding functional layers in the simulation model are replaced and updated, and the propagation matrix parameters of the corresponding functional layers are updated synchronously. The thickness parameters of the functional layers that have been deposited are locked, and the thickness parameters of the functional layers that have not been deposited are retained as variables to be optimized, so as to construct a hybrid parameter model that includes the measured thickness parameters and the thickness parameters to be optimized. Based on the hybrid parameter model, the optical and electrical response simulation calculations were re-executed to obtain the updated light field distribution, short-circuit current density of each sub-cell, and current matching error results. Based on the updated current matching error results, the thickness parameters of subsequent functional layers are adjusted to compensate for the errors, and the equivalent update process and corresponding calculation results are recorded to complete the dynamic equivalent update of the thickness parameters.

10. The method for optimizing the parameters of an organic solar cell according to claim 8, characterized in that, The method for optimizing organic solar cell parameters also includes: Based on the optical field simulation results and the predicted short-circuit current density, it is determined whether the current fabrication state meets the preset performance optimization standard. If the conditions are not met, the initial optimal parameter combination will be locally adaptively updated based on the current online film thickness deviation data, and incremental global collaborative optimization will be performed. If the conditions are met, the parameters of the current functional layer are locked, and the process guidance scheme is continued to be executed while maintaining the consistency of global collaborative constraints.

Images