Analysis device, analysis system, analysis method, and program

Abstract

This invention provides an analytical device that expands the scope of analysis based on data measured using organic EL elements. [Solution] The analysis device comprises an acquisition unit that acquires data on the optical properties of an organic electroluminescent element into which a subject has been introduced, and an analysis unit that analyzes the metal contained in the subject using machine learning based on the data acquired by the acquisition unit.

Description

Technical Field

[0001] The present invention relates to an analysis device, an analysis system, an analysis method, and a program.

Background Art

[0002] In an organic electroluminescence device (organic EL device), when moisture, oxygen, etc. are contained as foreign substances, dark spots are partially generated. By utilizing the generation of these dark spots, the organic EL device can be used as a sensor for detecting the presence or absence of moisture, gas, etc. (Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Since emission spectrum data is a large amount of complex data, high-precision analysis can be expected by using such data. However, conventional sensing technologies using organic EL devices have limited analysis targets, and further expansion of analysis targets is required.

[0005] The problem to be solved by the present invention is to provide an analysis device etc. that expands the analysis targets based on data measured using an organic EL device.

Means for Solving the Problems

[0006] To solve the above problems, the analysis device of the present invention an acquisition unit that acquires data on the optical characteristics of an organic electroluminescence device into which the subject is introduced, The system includes an analysis unit that analyzes the metals contained in the subject based on the data acquired by the acquisition unit using machine learning.

[0007] The invention described in claim 2 is, in the invention described in claim 1, The aforementioned metal is contained in the subject in the form of a metal ion with a valence of 2 or higher.

[0008] The invention described in claim 3 is, in the invention described in claim 2, The aforementioned metal ion is a transition metal ion.

[0009] The invention described in claim 4 is, in the invention described in claim 2, The aforementioned metal ions are alkaline earth metal ions.

[0010] The invention described in claim 5 is, in the invention described in claim 1, The above data is emission spectrum data when a voltage is applied to the organic electroluminescent element.

[0011] The invention described in claim 6 is, in the invention described in claim 1, The aforementioned machine learning prediction algorithm includes an algorithm that reduces the number of dimensions of multiple features contained in the data and compresses the data.

[0012] The invention described in claim 7 is, in the invention described in claim 1, The aforementioned machine learning prediction algorithm includes linear discriminant analysis.

[0013] The invention described in claim 8 is, in the invention described in claim 1, The aforementioned machine learning prediction algorithm includes canonical discriminant analysis.

[0014] The invention described in claim 9 is, in the invention described in claim 1, The item analyzed by the aforementioned analysis unit is the type of metal.

[0015] The invention according to claim 10 is the invention according to claim 1, wherein the item to be analyzed by the analysis unit is the content of the metal.

[0016] The invention according to claim 11 is the invention according to claim 1, wherein the item to be analyzed by the analysis unit is the ionic state of the metal.

[0017] The invention according to claim 12 is the invention according to claim 1, wherein the analysis unit analyzes two or more metals contained in the specimen.

[0018] The analysis system of the present invention according to claim 13 comprises the analysis device according to claim 1, and a measurement device for measuring the data, and the acquisition unit acquires the data measured by the measurement device.

[0019] The analysis method of the present invention according to claim 14 is an analysis method executed by an analysis device, and includes an acquisition step of acquiring data on the light characteristics of an organic electroluminescence element into which a specimen is introduced, and an analysis step of analyzing the metal contained in the specimen by machine learning based on the data acquired in the acquisition step.

[0020] The program of the present invention according to claim 15 causes a computer of an analysis device to function as an acquisition unit that acquires data on the light characteristics of an organic electroluminescence element into which a specimen is introduced, and an analysis unit that analyzes the metal contained in the specimen by machine learning based on the data acquired by the acquisition unit.

Advantages of the Invention

[0021] According to the present invention, the scope of analysis can be expanded in an analysis device that performs analysis based on data measured using an organic EL element. [Brief explanation of the drawing]

[0022] [Figure 1] This is a block diagram showing the schematic configuration of the analysis system. [Figure 2] These are emission spectrum data from organic EL elements treated with pure water and magnesium ion aqueous solution, respectively. [Figure 3] This is a schematic cross-sectional view of an organic EL element. [Figure 4] This is a schematic cross-sectional view of the laminate. [Figure 5] This is a schematic cross-sectional view of a laminate coated with the subject material. [Figure 6] This is a schematic cross-sectional view of a laminate coated with a light-emitting dopant. [Figure 7] This is a schematic cross-sectional view of an organic EL element. [Figure 8] This is a flowchart of the analysis method. [Figure 9] This graph shows the results of principal component analysis on the acquired spectral data. [Figure 10] This graph shows the results of canonical discriminant analysis on the acquired spectral data. [Figure 11] This graph shows the results of canonical discriminant analysis on the acquired spectral data. [Figure 12] This graph shows the relationship between predicted and measured values ​​of concentration in the common logarithm. [Modes for carrying out the invention]

[0023] Hereinafter, one or more embodiments of the present invention will be described with reference to the drawings. However, the scope of the present invention is not limited to the disclosed embodiments.

[0024] [Configuration of the analysis system] Figure 1 is a block diagram showing the schematic configuration of the analysis system 100 of this embodiment. The analysis system 100 comprises a measuring device 10 and an analysis device 20. The measuring device 10 and the analysis device 20 may communicate with each other. Alternatively, the measuring device 10 and the analysis device 20 may be integrated. The measuring device 10 comprises a measuring unit 11, a communication unit 12, and a control unit 13. The analysis device 20 comprises a display unit 21, an operation unit 22, a communication unit 23, and a control unit 24.

[0025] It has been found that when a different type of metal ion M2 is introduced into an organic electroluminescent device (organic EL device) that contains metal ions M1 as an electron injection material, the luminescence characteristics (emission wavelength, emission intensity, etc.) change to a very small extent. The mechanism of this action is not yet clear, but it is speculated as follows: When metal ions M1 and M2 are of the same type, the electron injection function does not decrease even when metal ions M2 are introduced. However, when metal ions M1 and M2 are of different types, it is thought that the luminescence characteristics of the organic EL device change because metal ions M2 inhibit the electron injection function of metal ions M1.

[0026] While the amount of metals in water (drinking water, wastewater, etc.) is generally very small, it has been found that even trace amounts of metal ions can alter the luminescence characteristics of organic EL elements.

[0027] Figure 2 shows the emission spectrum data of organic EL elements treated with pure water and magnesium ion aqueous solution (concentration: 100 ppm by mass), respectively. As shown in Figure 2, it can be seen that the peak waveform originating from the light-emitting dopant changes slightly depending on the presence or absence of metal ions.

[0028] In the analysis system 100 of this embodiment, an organic EL element is used as the sensor of the measuring device 10, and the acquired data is analyzed by machine learning. This allows for the analysis of metals contained in the sample, and expands the scope of analysis in an analysis device that analyzes based on data measured using an organic EL element.

[0029] Generally, mass spectrometry and ion chromatography are known instruments for analyzing metals contained in a sample. These instruments require a relatively long time for analysis. Furthermore, since each metal species is analyzed separately, analyzing multiple metal species takes even longer. In addition, these instruments have difficulty analyzing the ionic state of metals contained in the sample.

[0030] On the other hand, the analysis system 100 of this embodiment can shorten the time required for analysis by using an organic EL element as a sensor. Furthermore, since multiple target metal species can be analyzed simultaneously, the time required for analysis is almost the same as when analyzing only one type of metal species. In addition, the analysis system 100 of this embodiment can also analyze the ionic state of the metal contained in the sample.

[0031] Any metal contained in the sample, other than the metal species contained in the electron injection layer, can be analyzed by the analysis system 100 of this embodiment. Examples of alkaline earth metals that can be analyzed include beryllium (Be), magnesium (Mg), and calcium (Ca). Examples of transition metals that can be analyzed include scandium (Sc), titanium (Ti), vanadium (V), chromium (Cr), manganese (Mn), iron (Fe), cobalt (Co), nickel (Ni), copper (Cu), and zinc (Zn).

[0032] The details of each device are described below.

[0033] [Measuring device] (Measurement part) In this embodiment, the measurement unit 11 acquires data on the optical properties of the organic EL element into which the subject has been introduced, specifically the spectral data of the obtained light. By acquiring spectral data, a large amount of complex data can be obtained, improving the prediction accuracy of the analysis. The measurement unit 11 is not particularly limited as long as it is a device that measures the optical properties of the organic EL element.

[0034] In this embodiment, the organic EL element has a light-emitting layer between electrodes, and the light-emitting layer is a detection region. The light-emitting layer contains a light-emitting dopant. Various data can be obtained by acquiring the light emission information and current density of the light-emitting dopant. In this embodiment, "light-emitting dopant" refers to a substance that emits light when returning from an excited state to a ground state.

[0035] There are no particular limitations on the method for introducing a sample into an organic EL element, but two methods can be considered. The first is to introduce the sample as a material for the light-emitting layer when forming the light-emitting layer of the organic EL element. The second is to introduce the sample by allowing it to penetrate into the interior of the organic EL element. These methods will be explained in order below.

[0036] First, we will explain the method of introducing the subject as the material for the light-emitting layer when forming the light-emitting layer of an organic EL element.

[0037] Figure 3 is a schematic cross-sectional view of the organic EL element 110. The organic EL element 110 includes a transparent substrate 111, a transparent electrode 112, a hole transport layer 113, a receiving layer 114, a test subject layer 121, a light-emitting layer 122, an electron injection layer 116, and a counter electrode layer 115. The organic EL element 110 may further include a hole injection layer, an electron blocking layer, an electron transport layer, etc., as needed. Furthermore, if the light-emitting layer 122 contains a compound having an electron injection function, the electron injection layer 116 is not necessarily required.

[0038] A method for acquiring data using organic EL elements will be explained based on Figures 3 to 6.

[0039] In this method, first, a laminate comprising a transparent substrate 111, a transparent electrode 112 (e.g., an ITO film), a hole transport layer 113 (e.g., a polyaromatic diamine), and a receiving layer 114 (e.g., polystyrene) is prepared (Figure 4). The transparent substrate 111, transparent electrode 112, hole transport layer 113, etc., can be the same as those used in known organic EL devices. The receiving layer 114 only needs to be able to receive the test subject and the light-emitting dopant. The receiving layer 114 may also serve as the host layer.

[0040] Next, the sample layer 121 is applied to a desired area of ​​the receiving layer 114 of the laminate using any method (Figure 5).

[0041] Next, the luminescent dopant is applied in a pattern to desired locations on the receptor layer 114 coated with the subject layer 121 using an inkjet method or the like (Figure 6). As a result, the areas coated with the luminescent dopant become the luminescent layer 122. For example, by applying a component that dissolves the receptor layer 114 together with the luminescent dopant, the applied luminescent dopant, a part of the subject layer 121, and a part of the receptor layer 114 are mixed, and an integrated luminescent layer 122 is formed. Although not shown in the figure, the luminescent dopant may also be applied to areas where the subject layer 121 is not coated to obtain luminescence information from the luminescent dopant alone.

[0042] Next, an electron injection layer 116 is placed on the light-emitting layer 122. Note that if the light-emitting layer 122 contains a compound with electron injection functionality, the electron injection layer 116 does not need to be placed. Finally, a counter electrode layer 115, which is paired with the transparent electrode 112, is placed on the electron injection layer 116 to fabricate an organic EL element (Figure 3).

[0043] In the fabricated organic EL device, the light-emitting dopant in the light-emitting layer 122 is excited by a conventional method, and emission spectral data is acquired.

[0044] When acquiring light spectra using organic EL elements, i.e., when using organic EL elements as sensing devices, fluorescent compounds, delayed-fluorescence compounds, and phosphorescent compounds can be used as light-emitting dopants. Furthermore, different phosphorescent compounds may be used in combination as light-emitting dopants, or phosphorescent compounds and fluorescent compounds may be used in combination. This allows for the acquisition of any desired emission color. Additionally, multiple light-emitting compounds with different emission colors may be combined to produce white light.

[0045] In this specification, "fluorescent compound" refers to a compound that emits fluorescence other than delayed fluorescence. "Fluorescence" refers to the light emitted when returning from a singlet excited state to the ground state. "Fluorescence other than delayed fluorescence" refers to fluorescence excluding the "delayed fluorescence" exemplified below. An example of "delayed fluorescence" is "thermally activated delayed fluorescence (TADF)". Another example of "delayed fluorescence" is "triplet-triplet annihilation (TTA) delayed fluorescence".

[0046] In other words, in this specification, "fluorescent compound" does not include "delayed fluorescent compounds" such as "thermally activated delayed fluorescent compounds" and "triplet-triplet annihilation delayed fluorescent compounds." "Fluorescent compound" refers to a compound that does not undergo upconversion by reverse intersystem crossing from the lowest excited triplet energy level to the lowest excited singlet energy level.

[0047] Fluorescent compounds do not necessarily need to be heavy metal complexes like phosphorescent compounds. So-called organic compounds, composed of common combinations of elements such as carbon, oxygen, nitrogen, and hydrogen, can be used as fluorescent compounds. Other nonmetallic elements such as phosphorus, sulfur, and silicon may also be used as fluorescent compounds. Complexes of typical metals such as aluminum and zinc may also be used as fluorescent compounds. Known fluorescent compounds used in the light-emitting layers of organic EL devices may also be used as fluorescent compounds.

[0048] In this specification, "phosphorescent compound" refers to a compound that emits phosphorescence. Specifically, a "phosphorescent compound" refers to a compound that emits phosphorescence at room temperature (25°C) and has a phosphorescence quantum yield of 0.01 or higher at 25°C. A phosphorescence quantum yield of 0.1 or higher is preferred.

[0049] "Phosphorescence" refers to the light emitted when a substance returns from a triplet excited state to its ground state. Known phosphorescent compounds used in the light-emitting layers of organic EL devices can be used as the phosphorescent compound.

[0050] In this specification, "delayed fluorescence compound" refers to a compound that emits delayed fluorescence. "Delayed fluorescence" refers to the light emitted when a singlet excited state returns to the ground state as a result of upconversion by reverse intersystem crossing from the lowest excited triplet energy level to the lowest excited singlet energy level. Known delayed fluorescence compounds used in the light-emitting layer of organic EL elements can be used as the delayed fluorescence compound.

[0051] Next, we will explain a method of introducing the test subject by allowing it to penetrate the interior of the organic EL element.

[0052] Figure 7 is a schematic cross-sectional view of the organic EL element 110. The organic EL element 110 has a transparent substrate 111, a transparent electrode 112, a hole transport layer 113, a light-emitting layer 122, an electron injection layer 116, and a counter electrode layer 115 in this order. The organic EL element 110 may further have a hole injection layer, an electron blocking layer, a hole blocking layer, an electron transport layer, etc., as needed. Also, if the light-emitting layer 122 contains a compound having an electron injection function, the electron injection layer 116 is not necessarily required. The material of each layer may be a known material. The manufacturing method of each layer may be a known manufacturing method. The light-emitting dopant contained in the light-emitting layer 122 is the same as described above.

[0053] The organic EL element may be subjected to interaction with the test material by allowing the entire element to be permeated, or it may be subjected to interaction with the test material by allowing it to be partially permeated. When the entire element is permeated, the entire organic EL element is considered the permeated region. When the element is subjected to partial permeation, the organic EL element may be considered as either a partially permeated region or a partially non-permeated region.

[0054] One method for creating a non-penetrating region is to have a passivation film or encapsulant on the side of the counter electrode layer 115 opposite to the transparent substrate 111. Both a passivation film and an encapsulant may be present. By protecting the area around the organic EL element with the transparent substrate 111 and the passivation film or encapsulant, penetration by the subject can be suppressed.

[0055] Conversely, one method for creating a penetration region is to omit the passivation film or sealing material on the side of the counter electrode layer 115 opposite the transparent substrate 111, that is, to expose the counter electrode layer 115. Alternatively, the counter electrode layer 115 may have a layer that allows molecules of a certain size or smaller to pass through on the side of the counter electrode layer 115 opposite the transparent substrate 111. If the subject passes through this layer, it can be considered a penetration region.

[0056] Furthermore, the counter electrode layer 115 may have a passivation film or sealing material on the side opposite to the transparent substrate 111, and may also have partial through-holes. Having through-holes allows the counter electrode layer 115 to be exposed and become a penetration region. The through-holes may penetrate the counter electrode layer 115. This allows the electron injection layer 116 to be exposed and become a penetration region.

[0057] In an organic EL element impregnated with the test sample, the light-emitting dopant in the light-emitting layer 122 is excited by a conventional method, and emission spectral data is acquired.

[0058] The number of data points measured by the measurement unit 11 and acquired by the control unit 24 of the analysis device 20 is not particularly limited, but the more data points there are, the better the prediction accuracy of the analysis. The number of data points is appropriately selected depending on the measurement method, etc. Preferably, the number of data points is equal to or greater than the number of explanatory variables. If a large number of data points can be acquired, preferably the number of data points is 10 times or more the number of explanatory variables, and more preferably 100 times or more.

[0059] When using an organic EL element as a sensing device, multiple data points may be taken for each of the different emission colors—blue, green, and red—at intervals of a few nanometers. Alternatively, the current density may be measured at 1V intervals. This allows for the acquisition of a large number of data points.

[0060] (Communications Department) The communication unit 12 transmits the data measured by the measurement unit 11 to the communication unit 23 of the analysis device 20. The communication unit 12 may use either wireless communication or wired communication.

[0061] (Control Unit) The control unit 13 is a processor that provides overall control over the operation of the measuring device 10. The control unit 13 includes a CPU (Central Processing Unit) that performs various calculations, and RAM (Random Access Memory) that provides the CPU with working memory space and stores temporary data.

[0062] [Analysis equipment] (Display) The display unit 21 displays various information on its screen based on display control signals received from the control unit 24. The display unit 21 is equipped with a display device. Examples of display devices include displays and projectors. The display unit 21 displays the results of analysis performed by the control unit 24 based on data measured by the measurement unit 11 of the measuring device 10 and notifies the user.

[0063] (Operation unit) The control unit 22 accepts various inputs based on user operations. The control unit 22 is equipped with input devices. Examples of input devices include keyboards, mice, various switches, touchscreens, touchpads, etc.

[0064] (Communications Department) The communication unit 23 receives data transmitted from the communication unit 12 of the measuring device 10. The communication unit 23 may use either wireless communication or wired communication.

[0065] (Control Unit) The control unit 24 is a processor that provides overall control over the operation of the analysis device 20. The control unit 24 includes a CPU (Central Processing Unit) that performs various calculations, and RAM (Random Access Memory) that provides the CPU with working memory space and stores temporary data.

[0066] The control unit 24 acquires data measured by the measurement unit 11 of the measuring device 10, that is, data on the optical properties of the organic EL element 110 in which the subject is included in the light-emitting layer. At this time, the control unit 24 functions as an acquisition unit. Based on the data acquired by the acquisition unit, the control unit 24 analyzes the metals contained in the subject using machine learning. At this time, the control unit 24 functions as an analysis unit. Specifically, the CPU reads the stored program and loads it into RAM, and performs data acquisition and analysis in cooperation with the program loaded into RAM.

[0067] (Analysis using machine learning) In machine learning, multiple predictive models are created based on the data acquired by the acquisition unit. By combining the results of these multiple predictive models, a trained model capable of analyzing the metals contained in the sample is constructed.

[0068] The items analyzed regarding the metals contained in the sample include the type of metal, the amount (concentration), and the ionic state. Here, "ionic state" refers to whether or not the metal exists in an ionic state, and if the metal exists in an ionic state, the valence of the ion is also included.

[0069] If the type, content, and ionic state of the metal contained in the sample are known in advance, a predictive model can be created by performing machine learning with the characteristics of the analytical data as explanatory variables and the type, content, and ionic state of the metal as the dependent variable.

[0070] As explanatory variables, numerical values ​​representing the characteristics of the data acquired by the acquisition unit, and numerical values ​​calculated from them, can be used. If the data acquired by the acquisition unit is a spectral distribution, the intensity of light at each wavelength can be used as explanatory variables. As the dependent variable, the type of metal, content, ionic state, etc., can be selected according to the purpose of the analysis.

[0071] Machine learning can be either supervised or unsupervised. Supervised learning is a learning method that learns the relationship between input and output from training data that has correct labels. Unsupervised learning is a learning method that learns the structure of a data set from training data that does not have correct labels.

[0072] Machine learning may be reinforcement learning, deep learning, or deep reinforcement learning. "Reinforcement learning" refers to a learning method that learns the "optimal sequence of actions" through trial and error. "Deep learning" refers to a learning method that learns the features contained in a large amount of data in a stepwise, deeper way. "Deep reinforcement learning" refers to a learning method that combines reinforcement learning and deep learning.

[0073] Analysis in machine learning can be performed using statistical analysis software "JMP16.2" or "JMPpro16.2" manufactured by SAS Institute Japan Co., Ltd. Examples of prediction algorithms used in machine learning include principal component analysis (PCA), cluster analysis (hierarchical method, k-means method, normal mixture method), linear discriminant analysis (LDA), canonical discriminant analysis, and partial least squares regression (PLS regression). These algorithms may also be used in combination.

[0074] Other prediction algorithms used in machine learning include decision trees, random forests, bootstrap forests, neural networks, K-nearest neighbors, simple Bayesian algorithms, support vector machines (SVMs), nominal logistic regression (multiple logits), and generalized regression (Ridge, Lasso).

[0075] When the sample size is approximately 50 or less, the explanatory variables are 100 or more, and the variables are spectral data, the algorithm is preferably linear regression from the viewpoint of suppressing overfitting and ensuring robustness (reproducibility) of the results. In particular, the algorithm is preferably linear discriminant analysis or canonical discriminant analysis, and canonical discriminant analysis is preferred from the viewpoint of simultaneously analyzing multiple types of metals.

[0076] From the standpoint of suppressing overfitting and ensuring robust results, it is preferable to use a dimensionality reduction technique in the algorithm. Dimensionality reduction techniques reduce the number of dimensions of multiple features contained in the data, thereby compressing the data. This makes the data easier to handle and speeds up the entire analysis process.

[0077] For evaluating discrimination performance, for example, the entire sample data is divided into training data and validation data in an arbitrary proportion, discriminant analysis is performed, and the discrimination accuracy (correctness), misclassification rate, and entropy R-squared are calculated. When calculating validation results using validation data, it is preferable to use the hold-out method, which uses an arbitrary or random set of validation data. Alternatively, when calculating validation results using validation data, it is preferable to use k-fold cross-validation, which divides the entire sample data into k sets and performs cross-validation on combinations of these k sets.

[0078] In calculating validation results using k-fold cross-validation, it is preferable to display the k mean values ​​and the one with the best statistical performance among the k discriminant models. Here, "best statistical performance" means that the error from the mean is small. In addition, in calculating validation results using k-fold cross-validation, it is preferable to display the one with the smallest discriminant accuracy between training and validation, or the one with the smallest difference in entropy R-squared.

[0079] [Subject] The type of subject in this embodiment is not particularly limited as long as it contains a metal, and may be a substance with a known structure or a substance with an unknown structure. The subject may be a mixture of various compounds, etc. The subject may belong to any field, such as the medical field, the industrial field, or the food field.

[0080] Examples of subjects belonging to the medical field include sweat and blood. Examples of subjects belonging to the industrial field include rainwater, rivers, small ponds, tropical fish breeding water, soil, and wastewater. In other words, the analysis system 100 of this embodiment can be applied to water quality management. Examples of subjects belonging to the food field include food products in general.

[0081] [Analysis method] Figure 8 shows a flowchart of the analysis method of this embodiment. The analysis method of this embodiment has an acquisition step (step S1) and an analysis step (step S2). Details of each step are as described in the Control Unit 24 (Acquisition Unit and Analysis Unit).

[0082] In this embodiment, the analysis device 20 comprises an acquisition unit (control unit 24) and an analysis unit (control unit 24). The acquisition unit acquires data on the optical properties of the organic EL element 110 into which the subject has been introduced. The analysis unit analyzes the metals contained in the subject using machine learning based on the data acquired by the acquisition unit. As a result, the luminescence properties of the organic EL element 110 into which the subject has been introduced change, and the metals contained in the subject can be analyzed. Consequently, the scope of analysis can be expanded in an analysis device that performs analysis based on data measured using an organic EL element.

[0083] In this embodiment, it is preferable that the metal is contained in the sample in the form of metal ions with a valence of 2 or higher. This changes the luminescence characteristics of the organic EL element 110 into which the sample is introduced, allowing for analysis of the metal contained in the sample.

[0084] In this embodiment, the metal ion is preferably a transition metal ion. This changes the luminescence characteristics of the organic EL element 110 into which the sample is introduced, allowing for analysis of the metal contained in the sample.

[0085] In this embodiment, the metal ions are preferably alkaline earth metal ions. This changes the luminescence characteristics of the organic EL element 110 into which the sample is introduced, allowing for analysis of the metals contained in the sample.

[0086] In this embodiment, the data is preferably emission spectrum data when a voltage is applied to the organic EL element 110. This allows for the acquisition of a large amount of complex data, improving the prediction accuracy of the analysis.

[0087] In this embodiment, the machine learning prediction algorithm preferably includes an algorithm that reduces the number of dimensions of multiple features contained in the data and compresses the data. This speeds up the series of analysis steps and reduces overfitting.

[0088] In this embodiment, the machine learning prediction algorithm preferably includes linear discriminant analysis. This improves the robustness of the analysis.

[0089] In this embodiment, the machine learning prediction algorithm preferably includes canonical discriminant analysis. This allows for the simultaneous analysis of multiple types of metals.

[0090] In this embodiment, the item analyzed by the analysis unit is preferably the type of metal. The analysis device 20 of this embodiment can analyze the type of metal contained in the sample.

[0091] In this embodiment, the item analyzed by the analysis unit is preferably the metal content. The analysis device 20 of this embodiment can analyze the metal content contained in the sample.

[0092] In this embodiment, the item analyzed by the analysis unit is preferably the ionic state of the metal. The analysis device 20 of this embodiment can analyze the ionic state of the metal contained in the sample.

[0093] In this embodiment, it is preferable that the analysis unit analyzes two or more metals contained in the sample. The analysis device 20 of this embodiment can analyze two or more metals contained in the sample.

[0094] In this embodiment, the analysis system 100 comprises an analysis device 20 and a measuring device 10 for measuring data. The acquisition unit acquires the data measured by the measuring device 10.

[0095] In this embodiment, the analysis method performed by the analysis device 20 includes an acquisition step S1 and an analysis step S2. The acquisition step S1 acquires data on the optical properties of the organic EL element 110 into which the subject has been introduced. The analysis step S2 analyzes the metals contained in the subject using machine learning based on the data acquired in the acquisition step S1. As a result, the luminescence properties of the organic EL element 110 into which the subject has been introduced change, and the metals contained in the subject can be analyzed. Consequently, the scope of analysis can be expanded in an analysis device that performs analysis based on data measured using an organic EL element.

[0096] In this embodiment, the program causes the computer of the analysis device 20 to function as an acquisition unit and an analysis unit. The acquisition unit acquires data on the optical properties of the organic EL element 110 into which the subject has been introduced. The analysis unit analyzes the metals contained in the subject using machine learning based on the data acquired by the acquisition unit. As a result, the luminescence properties of the organic EL element 110 into which the subject has been introduced change, and the metals contained in the subject can be analyzed. Consequently, the scope of analysis can be expanded in an analysis device that performs analysis based on data measured using an organic EL element.

[0097] Furthermore, the detailed configuration and operation of each device constituting the analysis apparatus can also be modified as appropriate without departing from the spirit of the present invention. [Examples]

[0098] The present invention will be described in detail below with reference to examples, but it is not limited to these examples.

[0099] <Data Acquisition> For the analysis of metal species, we prepared metal ion aqueous solutions with different metal species and concentrations. The metal species were sodium (Na), magnesium (Mg), or calcium (Ca), for a total of three types. The concentrations were 10 ppm by mass, 100 ppm by mass, or 1000 ppm by mass, for a total of three types. For comparison, we also prepared water containing no metal species, i.e., with a concentration of 0 ppm by mass. In total, we prepared 10 types of metal ion aqueous solutions. For concentration analysis, magnesium ion aqueous solutions with concentrations of 0.01 ppm, 0.1 ppm, 1 ppm, 10 ppm, 100 ppm, and 1000 ppm were prepared. For comparison, water containing no metal species, i.e., with a concentration of 0 ppm, was also prepared.

[0100] [Data acquisition using OLED sensors] Data was acquired using an OLED sensor into which the test subject was introduced as the material for the light-emitting layer. The details are shown below. Note that an OLED sensor was fabricated using the same procedure except that the test subject was not introduced into the light-emitting layer, and the above-mentioned metal ion aqueous solution was impregnated into the fabricated OLED sensor as the test subject. The same data was obtained when the following data was acquired from this OLED sensor impregnated with metal ions.

[0101] (1) Fabrication of a blue OLED sensor A substrate was prepared by depositing a 100 nm thick indium tin oxide (ITO) film onto a 30 mm × 30 mm × 0.7 mm glass substrate. After patterning the substrate, it was ultrasonically cleaned with isopropyl alcohol and dried with dry nitrogen gas. The substrate was then cleaned with UV ozone for 5 minutes to obtain a transparent support substrate with an ITO transparent electrode (anode).

[0102] An insulating polymer solution was obtained by diluting polystyrene (manufactured by ACROS ORGANICS Co., Ltd., molecular weight = 260,000) as the solute and n-propyl acetate as the solvent to 1.0% by mass. The insulating polymer solution was applied to the anode by spin coating at 500 rpm for 30 seconds. The coating film was then dried at 120°C for 30 minutes to obtain a laminate with a receiving layer 50 nm thick.

[0103] The above-mentioned metal ion aqueous solutions were prepared as test subjects. The laminates obtained above were immersed in petri dishes containing each metal ion aqueous solution for 1 minute, then the liquid was removed with an air gun and the laminates were dried.

[0104] n-propyl acetate was used as the solvent, and a luminescent dopant (blue phosphorescent compound Ep1, described below) was mixed with the solvent at a concentration of 10 mg / mL. The resulting mixture was heated with ultrasound for 30 minutes, and then filtered through a 0.2 μm filter to remove aggregated components, thereby preparing a luminescent ink.

[0105] [ka]

[0106] The obtained luminescent ink was dropped onto the receiving layer of the laminate using an inkjet method under the following conditions to form a luminescent layer (detection region). The laminate was dried at 120°C for 30 minutes to evaporate the solvent.

[0107] The conditions for the inkjet method were as follows: The inkjet device used was the "IJCS-1" (manufactured by Konica Minolta, Inc.), and the inkjet head used was the "KM512" (manufactured by Konica Minolta, Inc.). The number of ejection shots was 2, the distance between ejection nozzles from the head was 140 μm pitch, and printing was performed at a head scan speed of 90 mm / sec.

[0108] The laminated body with the detection region formed was attached to the vacuum deposition apparatus. The vacuum chamber was 4 × 10 -4The pressure was reduced to Pa, and an electron injection layer and a cathode were formed on the detection region of the laminate under the following conditions to obtain a blue OLED sensor. The electron injection layer was formed by depositing potassium fluoride at a deposition rate of 0.1 Å / sec to a thickness of 2.0 nm. The cathode was formed by depositing aluminum at a deposition rate of 4 Å / sec to a thickness of 100 nm.

[0109] The resulting blue OLED sensor had four detection areas (2 x 2 mm) within a 30 x 30 mm area, with the detection area (inkjet area) being the same size as the electrode. The polystyrene receiving layer is insoluble in the metal ion aqueous solution, which is the test subject. Therefore, when the laminate was immersed in the test subject and dried, trace amounts of solid components (metal components) from the aqueous solution were present on the surface of the polystyrene receiving layer. Subsequently, when the luminescent ink was applied to the receiving layer using an inkjet method, the solvent of the luminescent ink dissolved the polystyrene. The dissolved polystyrene, luminescent dopant, and test subject reached the lower electrode while forming a dispersed state. As a result, a detection area was formed in which the luminescent dopant and test subject were dispersed at the molecular level.

[0110] In this embodiment, 24 blue OLED sensors were fabricated for each type of sample, resulting in 4 × 24 = 96 measurement points for each sample. The anode and cathode were configured to allow voltage application via wiring.

[0111] (2) Fabrication of green and red OLED sensors The green and red OLED sensors were fabricated using the same procedure as the blue OLED sensor, except that the light-emitting dopant was changed to compound 1 (Green) or compound 2 (RED).

[0112] [ka]

[0113] (3) Data acquisition For the three types of OLED sensors fabricated (blue, green, and red), the spectral radiance spectra [W·sr] were analyzed for each. -1 ·m -2 ·nm -1 ] and the current density [mA / cm²] for each drive voltage 2 The following measurements were taken: The drive voltage was set in 1V increments from 6V to 12V, and the current at each voltage was measured. A spectroradiometer "CS-2000" (manufactured by Konica Minolta, Inc.) was used to measure brightness. A "6243 DC VOLTAGE CURRENT SOURCE / MONITOR" (manufactured by ADCMT Corporation) was used to measure current.

[0114] Using OLED sensors containing B (blue dopant), G (green dopant), and R (red dopant), 57 spectral data points were recorded at 5nm intervals from 420 to 700nm. The spectral data was normalized with the value at the wavelength with the highest radiance set to 1.

[0115] <Data Analysis> (1) Analysis of metal species The spectral data of all aqueous solutions (for metal species analysis) obtained as described above were analyzed. The software used for the analysis was "JMPpro" (manufactured by SAS Institute Japan Co., Ltd.). Figure 9 is a graph showing the results of principal component analysis for the spectral data obtained above. While the graph suggests that the data is roughly clustered by metal species, it is clear that distinguishing between metal species is difficult using principal component analysis alone.

[0116] Figure 10 is a graph showing the results of canonical discriminant analysis on the spectral data obtained above. Note that in the circles corresponding to each metal species, the inner circle represents the 95% confidence interval. The graph shows that clustering occurs for each metal species. The circle corresponding to water and the inner circle corresponding to sodium overlap, indicating that sodium does not have a significant difference compared to water. On the other hand, the circle corresponding to water does not overlap with the inner circles corresponding to magnesium and calcium. Therefore, magnesium and calcium have significant differences compared to water. Furthermore, the inner circle corresponding to magnesium and the inner circle corresponding to calcium do not overlap. Therefore, magnesium and calcium have significant differences.

[0117] In other words, canonical discriminant analysis can distinguish between magnesium and calcium when the metal species are magnesium or calcium.

[0118] (2) Analysis of concentration Principal component analysis was performed on the spectral data of the magnesium ion aqueous solution (for concentration analysis) obtained above. As a result, a correlation was observed between the first principal component and the concentration of magnesium ions.

[0119] Figure 11 is a graph showing the results of canonical discriminant analysis on the spectral data obtained above. The number after "Mg" represents the concentration; for example, "Mg0.01" means that the magnesium ion concentration is 0.01 ppm by mass. The graph shows a correlation between the first canonical variable and the magnesium ion concentration. In other words, canonical discriminant analysis can distinguish between metal ion concentrations. Furthermore, it can be seen that canonical discriminant analysis can distinguish between concentrations over a wide range from 0.01 to 1000 ppm by mass, and is particularly effective at distinguishing even extremely low concentrations (0.01 ppm by mass).

[0120] Next, a correlation analysis was performed between the concentration of the magnesium ion aqueous solution and the data obtained by the OLED sensor. Table I below shows the correlation coefficients obtained from the correlation analysis.

[0121] [Table 1]

[0122] In Analysis 1, principal component analysis was performed on spectral data (luminescence intensity) from an OLED sensor. Table I shows the correlation coefficients between the first principal component and concentration, the second principal component and concentration, the first principal component and the common logarithm of concentration, and the second principal component and the common logarithm of concentration. It can be seen that the first principal component shows a higher correlation with both concentration and the common logarithm of concentration compared to the second principal component.

[0123] In Analysis 2, principal component analysis was performed on the current value data from the OLED sensor. Table I shows the correlation coefficients for the first principal component and concentration, the second principal component and concentration, the first principal component and the common logarithm of concentration, and the second principal component and the common logarithm of concentration. It can be seen that the first principal component shows a higher correlation with both concentration and the common logarithm of concentration compared to the second principal component.

[0124] In Analysis 3, principal component analysis was performed on both the spectral data and current value data from the OLED sensor. Table I shows the correlation coefficients for the first principal component and concentration, the second principal component and concentration, the first principal component and the common logarithm of concentration, and the second principal component and the common logarithm of concentration. It can be seen that the first principal component shows a higher correlation with both concentration and the common logarithm of concentration compared to the second principal component.

[0125] In analyses 1-3, comparing the correlation coefficients between each principal component and its concentration, and between each principal component and its common logarithm, it can be seen that the common logarithm of the concentration shows a higher correlation with each principal component compared to the concentration itself.

[0126] Next, regression analysis was performed on spectral data and current values ​​acquired by the OLED sensor, along with the common logarithm of the concentration, to create a predictive model. The created predictive model was then cross-validated with k=5 folds.

[0127] Figure 12 shows the relationship between predicted and measured values ​​in the common logarithm of concentration. Furthermore, the R-squared value and the root mean squared error (RMSE) of the developed prediction model were 0.62 and 7.4 [mass ppm]. From the graph and various evaluation indices (R-squared value and root mean squared error) shown in Figure 12, it can be seen that the developed prediction model is sufficient for analyzing metal concentrations. [Explanation of Symbols]

[0128] 10 Measuring device 11 Measuring part 12 Communications Department 13 Control Unit 20 Analyzer 21 Display section 22 Control section 23 Communications Department 24 Control Unit 100 Analysis Systems 110 Organic EL elements 111 Transparent substrate 112 Transparent electrode 113 Hole transport layer 114 Receptor layer 115 Counter electrode layer 116 Electron injection layer 121 Subject layer 122 Emitting layer

Claims

1. A data acquisition unit that acquires data on the optical properties of an organic electroluminescent element into which a subject has been introduced, An analysis device comprising: an analysis unit that analyzes the metals contained in the subject by machine learning based on the data acquired by the acquisition unit.

2. The analytical apparatus according to claim 1, wherein the metal is contained in the sample in the form of a metal ion with a valence of 2 or higher.

3. The analytical apparatus according to claim 2, wherein the metal ion is a transition metal ion.

4. The analytical apparatus according to claim 2, wherein the metal ion is an alkaline earth metal ion.

5. The analysis apparatus according to claim 1, wherein the data is emission spectrum data when a voltage is applied to the organic electroluminescent element.

6. The analysis apparatus according to claim 1, wherein the machine learning prediction algorithm includes an algorithm for reducing the number of dimensions of multiple features contained in the data and for reducing the data.

7. The analysis apparatus according to claim 1, wherein the machine learning prediction algorithm includes linear discriminant analysis.

8. The analysis apparatus according to claim 1, wherein the machine learning prediction algorithm includes canonical discriminant analysis.

9. The analysis apparatus according to claim 1, wherein the item to be analyzed by the analysis unit is the type of metal.

10. The analysis apparatus according to claim 1, wherein the item analyzed by the analysis unit is the content of the metal.

11. The analysis apparatus according to claim 1, wherein the item analyzed by the analysis unit is the ionic state of the metal.

12. The analysis apparatus according to claim 1, wherein the analysis unit analyzes two or more metals contained in the sample.

13. The analysis apparatus described in claim 1, The device comprises a measuring device for measuring the aforementioned data, The acquisition unit is an analysis system that acquires the data measured by the measuring device.

14. In the analysis method performed by the analysis device, A data acquisition process in which a subject obtains data on the optical properties of an organic electroluminescent element introduced inside, An analysis method comprising: an analysis step of analyzing the metals contained in the subject by machine learning based on the data acquired in the acquisition step.

15. The computer of the analysis device, A data acquisition unit that acquires data on the optical properties of an organic electroluminescent element into which a subject has been introduced, A program that functions as an analysis unit that analyzes the metals contained in the subject based on the data acquired by the acquisition unit using machine learning.

Images