Composition evaluation method, sensor, and evaluation system

Abstract

The purpose of the present invention is to provide an evaluation method of a composition in which, during research and development, or on a production line, it is possible to perform overall evaluation of the state of the target composition without affecting the composition. This evaluation method of said composition involves: a step in which a light-emitting layer is prepared that contains a light-emitting substance which changes light emission behavior in response to the state of the composition; a step in which, in a state in which the light-emitting layer and the composition are brought into contact, the light-emitting layer is irradiated with excitation light and light emission information about the light emission substance is acquired; and a step in which the light emission information is analyzed on the basis of fluorescence fingerprint information acquired in advance, and the state of the composition is evaluated.

Description

[Technical Field]

[0001] The present invention relates to a method for evaluating a composition, a sensor and an evaluation system for doing so. [Background technology]

[0002] Traditionally, data science has been used downstream in the manufacturing supply chain, in areas such as sales, inventory management, accounting, and quality assurance. However, in recent years, there has been a desire to utilize data science in more upstream areas such as product prototyping, development, research, and manufacturing. However, in order to adapt data science to upstream areas of the manufacturing industry, it is essential to understand the concepts and use it appropriately.

[0003] For example, if we take all of the academic fields and behaviors in question and analyze them from the perspective of "regularities" and "correlations," the areas that have benefited greatly from the development of AI (Artificial Intelligence) in recent years are not theoretical, rule-based academic fields like physics and electricity, but areas like images and purchasing behavior that have no rules and where solutions can be found solely through correlations. In these areas, the use of AI has led to dramatic advances in business expansion and problem-solving.

[0004] Meanwhile, in the fields of technological development and research and development, it has traditionally been important for researchers to start with data obtained from experiments, "deductively" derive patterns and correlations from them, and then formulate and verify (abduction) the resulting "hypothesis." In fields where a theory can be derived with fewer hypothesis tests, there are fewer benefits to using AI. Examples of such fields include physics and electrical engineering. Meanwhile, in fields dealing with low-molecular-weight compounds, polymeric materials, and biotechnology, hypotheses can be formulated, but the effects are often diverse, making it easy for hypotheses to be inaccurate or for hypothesis verification to take a long time.

[0005] Here, the opposite of "deductive" hypothesis formulation is the "inductive" interpretation method. In Industry 4.0, an inductive approach is dominant over deductive hypothesis formulation. Materials informatics applies this inductive approach to chemistry, materials, and compounding, while process informatics extends it to manufacturing processes. To put it very simply, minimizing abduction is key to reaping the benefits of AI in these fields. Furthermore, it is not AI but data generation that will drive digital transformation (DX) in the industrial sector. For example, it is crucial to generate binary data that captures subtle differences and characteristics of target objects, fluids, and gases.

[0006] Here, data used in inductive interpretation techniques refers to data that is difficult to handle deductively in human thought, such as scientific data that is not sufficiently grounded or that is not sufficiently assigned meaning, etc. Furthermore, even if scientific data has a sufficiently grounded or assigned meaning, scientific data that is difficult to process deductively in human thought due to its complexity or large amount of information is included in inductive data.

[0007] One type of inductive data is fluorescence fingerprinting. Fluorescence fingerprinting is a measurement method that has been known since the 1970s, in which fluorescence spectra are acquired while changing the excitation wavelength. Because foods contain a relatively large number of components that emit fluorescence when exposed to excitation light, fluorescence fingerprinting is used in the food industry for quality control and other purposes.

[0008] On the other hand, Patent Documents 1 and 2 propose a sensor having a tip carrying a fluorophore or reagent whose fluorescence activity changes depending on the concentration of a specific substance, and a system including such a sensor. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] International Publication No. 1999 / 023476 [Patent Document 2] Japanese Patent Application Publication No. 02-183145 Summary of the Invention [Problem to be solved by the invention]

[0010] As mentioned above, digital transformation is also required in manufacturing industries that handle low-molecular-weight compounds, polymer materials, biotechnology, and other fields, and fluorescent fingerprinting could be used to accelerate this transformation. However, fluorescent fingerprinting has not yet become widespread outside of the food industry. One reason for this is that the objects being measured in these industries do not contain fluorescent components.

[0011] Therefore, in product research and development and on a production line, it is conceivable to mix a luminescent substance into a raw material composition, an intermediate product, a final product, etc. However, adding such a substance is not appropriate because it may impair the quality of the raw material composition or the product.

[0012] The present invention aims to provide an evaluation method, a sensor used therefor, and an evaluation system that can comprehensively evaluate the state of a composition in product research and development or on a production line without affecting the composition that is the subject of production or research. [Means for solving the problem]

[0013] As one embodiment of the present invention, there is provided a method for evaluating a composition, comprising the steps of: preparing a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition; irradiating the light-emitting layer with excitation light while the light-emitting layer and the composition are in contact with each other, thereby acquiring light-emitting information of the light-emitting substance; and analyzing the light-emitting information based on previously acquired fluorescence fingerprint information, thereby evaluating the state of the composition.

[0014] As one embodiment of the present invention, there is provided a sensor comprising: a light-emitting section having a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition; an excitation light source that emits excitation light to excite the light-emitting substance; and a detection section that acquires information about the light emitted by the light-emitting substance, wherein the wavelength of the excitation light is determined based on previously acquired fluorescence fingerprint information.

[0015] In another embodiment, the present invention also provides a sensor comprising: a light-emitting unit having a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition; an excitation light source that emits excitation light to excite the light-emitting substance; and a detection unit that acquires information about the light emitted by the light-emitting substance, wherein the light-emitting substance is a substance selected based on previously acquired fluorescence fingerprint information.

[0016] As one embodiment of the present invention, there is provided an evaluation system having the sensor and an information processing unit that analyzes light information acquired by the detection unit using previously acquired fluorescence fingerprint information and evaluates the state of the composition. [Effects of the Invention]

[0017] The composition evaluation method described above makes it possible to comprehensively evaluate the state of a composition in various production lines, prototypes, and research without affecting the composition itself. Furthermore, the sensor and evaluation system described above enable the evaluation method to be carried out efficiently. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 is a diagram showing the flow of one embodiment of the composition evaluation method of the present invention. [Figure 2] FIG. 2 is a diagram showing the flow of a modified example of one embodiment of the composition evaluation method of the present invention. [Figure 3] FIG. 3 is a diagram showing the flow of the luminescent material selection process. [Figure 4] FIG. 4 is a schematic diagram showing an example of the structure of the sensor of the present invention. [Figure 5]FIG. 5A is a fluorescence fingerprint of a composition containing a recyclable cycloolefin polymer used in an example of the present invention, and FIG. 5B is a fluorescence fingerprint of a composition containing a non-recyclable cycloolefin polymer. DETAILED DESCRIPTION OF THE INVENTION

[0019] The present invention will be described in detail below by way of an embodiment, but the present invention is not limited to this embodiment.

[0020] 1. Method for evaluating the composition The composition evaluation method of this embodiment is a method for evaluating various performance and properties of a composition, such as determining whether the composition is good or bad, specifying the degree of deterioration, and distinguishing the place of origin and lot. Generally, when the structure and properties of a substance to be measured are known, it is easy to select test conditions and reagents, etc., tailored to the substance as a target. In contrast, when a composition contains multiple substances that interact in a complex manner, or when multiple components contribute to the properties or performance, evaluation is difficult. According to the method of this embodiment, it is possible to comprehensively evaluate the performance and properties of compositions that are difficult to evaluate using general methods.

[0021] The composition evaluated by the composition evaluation method of this embodiment may contain two or more components, and may be, for example, a food product or various industrial products. As will be described in detail later, the composition must be sufficiently interacted with the luminescent material in the luminescent layer, so the composition is preferably a fluid, more preferably a gas or liquid, and particularly preferably a liquid. The composition evaluation method of this embodiment may be performed during the trial production, development, or research process of a product, during the process of testing a product production line (in a test plant) or the process of producing a prototype (in a pilot plant), or even during the product manufacturing process or the product quality inspection process. It may also be performed during the manufacturing, storage, or quality control process of agricultural products or processed agricultural products.

[0022] As shown in the flow diagram of FIG. 1 , the composition evaluation method of the present embodiment includes the steps of: step S11 of preparing a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition (hereinafter also referred to as the "light-emitting layer preparation step"); step S12 of bringing the light-emitting layer and the composition into contact, irradiating the light-emitting layer with excitation light, and acquiring light-emitting information of the light-emitting substance (hereinafter also referred to as the "light-emitting information acquisition step"); and step S13 of analyzing the light-emitting information using previously acquired fluorescence fingerprint information to evaluate the state of the composition (hereinafter also referred to as the "evaluation step").

[0023] The composition evaluation method of this embodiment may include other steps as needed. For example, as shown in the flow diagram of Figure 2, before the step S11 of preparing the light-emitting layer, a step S10 of selecting a light-emitting material to be used in the light-emitting layer (hereinafter also referred to as a "light-emitting material selection step") may be included. Hereinafter, an embodiment in which the light-emitting material selection step S10 is performed before the light-emitting layer preparation step S11 will be described. However, this embodiment is not limited to this embodiment.

[0024] (Luminescent material selection process) In the luminescent material selection step S10, a luminescent material to be used in the luminescent layer to be prepared in the luminescent layer preparation step S11 (described later) is selected. Specifically, as shown in the flow diagram of FIG. 3, multiple samples containing multiple compositions in different states and multiple types of candidate luminescent materials are prepared (S101). Then, sample fluorescence fingerprints are obtained for the multiple samples (S102). Furthermore, the luminescent material to be used in the luminescent layer is determined based on the sample fluorescence fingerprint information (S103).

[0025] In the sample preparation step S101, first, multiple compositions in different states are prepared. Here, "different states" refers to different properties, performance, etc., and are appropriately selected depending on the properties or performance to be evaluated in the evaluation method of this embodiment. For example, when evaluating the taste or flavor of food, an ideal composition (good product) and a composition with poor taste or flavor (defective product) are prepared. Furthermore, when evaluating product lots or origins, multiple compositions with different degrees of polymerization are prepared. Furthermore, when determining the polymerization state of monomers during the manufacturing process, multiple compositions with different degrees of polymerization are prepared. Note that the performance or property evaluated in the evaluation step S13 described below does not have to be one, and multiple properties may be evaluated. When evaluating multiple performances or properties, multiple compositions may be prepared for each performance or property.

[0026] Meanwhile, in the sample preparation step S101, multiple candidate luminescent substances are prepared that may be used as luminescent substances in the luminescence information acquisition step S12 described below. In this specification, the term "luminescent substance" refers to any substance that can emit light when irradiated with light in the luminescence information acquisition step described below and whose luminescence behavior changes depending on the state of the composition. Note that "changes in luminescence behavior" refers to a change in the peak wavelength of the light emitted by the substance, a change in the intensity of the emitted light, or a change in the spectrum of the emitted light. Typically, the light emitted by a luminescent substance is fluorescence or phosphorescence generated by excitation of the luminescent substance.

[0027] Examples of luminescent substances (candidate luminescent substances) whose luminescence behavior changes depending on the state of the composition include substances that hydrogen bond with components in the composition and substances that interact with π-π interactions. Such luminescent substances (candidate luminescent substances) may be inorganic, organic, or metal complexes. Examples of inorganic substances that can be used as luminescent substances (candidate luminescent substances) include YAG phosphors, quantum dot phosphors, thermochromic materials such as VO2, and Ag nanoparticles. Examples of organic substances that can be used as luminescent substances (candidate luminescent substances) include fluorescent materials such as coumarin and perylene. Examples of metal complexes that can be used as luminescent substances (candidate luminescent substances) include phosphorescent materials such as Eu (europium) complexes and Ir (iridium) complexes. Compounds that interact with water and change their luminescence intensity depending on its concentration, and compounds that exhibit optical anisotropy, becoming cis-conformer when irradiated with ultraviolet light and becoming trans-conformer when irradiated with visible light, exhibiting switching properties, can also be used as luminescent substances (candidate luminescent substances). These luminescent substances (candidate luminescent substances) may have a group (e.g., Si—OR) for covalently bonding with a binder in the luminescent layer described below, or may be luminescent substances obtained by silane coupling with the inorganic substance, organic substance, metal complex, or the like.

[0028] Then, in the sample preparation step S101, the above-mentioned multiple compositions and multiple candidate luminescent substances are mixed or brought into contact with each other to prepare the sample. The ratio of the compositions and the candidate luminescent substances in the sample to be prepared is appropriately selected. If necessary, they may be mixed using a solvent or the like. Furthermore, as long as the compositions and the candidate luminescent substances can be sufficiently contacted, they do not necessarily have to be mixed. When preparing the sample, it is also preferable to prepare a sample that does not contain the candidate luminescent substance, i.e., a sample consisting of only the composition or only the composition and solvent.

[0029] The number of samples prepared in the sample preparation step S101 is not particularly limited and is appropriately selected depending on the number of different compositions and the number of candidate luminescent substances. Multiple samples can be prepared by, but are not limited to, the following method. For example, 95 types of candidate luminescent substances are placed in each well of a 96-well plate. The remaining well is left empty. A first composition (e.g., a non-defective product) is then added to each of the 96 wells. Similarly, another plate is prepared in which 95 types of candidate luminescent substances are placed in each well, and a second composition (e.g., a defective product) is added to each of the 96 wells. This allows 192 samples to be prepared, but further samples may be prepared using a third composition, a fourth composition, etc. Note that each candidate luminescent substance may consist of only one compound, or may be a mixture of two or more compounds.

[0030] After preparing the multiple samples, a step of acquiring a fluorescence fingerprint (also referred to herein as a "sample fluorescence fingerprint") for each sample is performed (step S102). The method of acquiring a sample fluorescence fingerprint is not particularly limited, and can be performed, for example, as follows. First, each sample is irradiated with excitation light of a specific wavelength from an excitation light source. Then, the wavelength and intensity of light (fluorescence or phosphorescence) emitted by the sample when irradiated with the excitation light are measured. Next, the wavelength of the excitation light is shifted by a desired width (e.g., 10 nm), and the wavelength and intensity of the light are similarly measured. These steps are then repeated to acquire data on the wavelength of the excitation light and the wavelength and intensity of light emitted by the sample (light-emitting substance). These are then converted into three-dimensional data to acquire a sample fluorescence fingerprint. Note that the sample fluorescence fingerprint referred to herein may be created based on the wavelength and intensity of light emitted by the sample (light-emitting substance), and may be created based on, for example, the wavelength and intensity of phosphorescence emitted by the sample (light-emitting substance).

[0031] The wavelength of the excitation light used to obtain the sample fluorescence fingerprint is appropriately selected depending on the type of candidate luminescent substance, the type of composition, etc. For example, when a substance that can be excited by visible light is used as the candidate luminescent substance, visible light is used as the excitation light. On the other hand, when a substance that can be excited by ultraviolet light is used as the candidate luminescent substance, ultraviolet light is used as the excitation light.

[0032] The light source for the excitation light is not particularly limited, but can be a supercontinuum light source (a broadband pulsed light source that utilizes the nonlinear effect of optical fiber to emit strong, phase-coherent light over an extremely wide wavelength range, also known as an "SC light source") or an LED. These light sources can increase the amount of light, making it easier to obtain a clear sample fluorescence fingerprint. Note that a sample fluorescence fingerprint can also be obtained by combining multiple light sources.

[0033] On the other hand, the wavelength and intensity of the fluorescence emitted by each sample can be measured using a spectrofluorometer, etc. The measurement may be performed using a plurality of spectrofluorometers.

[0034] Furthermore, the device that creates a sample fluorescence fingerprint from the wavelength of the excitation light and the wavelength and intensity of the light emitted by the candidate luminescent substance, i.e., the device that converts this data into three-dimensional data, can be a general information processing device, such as a personal computer.

[0035] After obtaining a sample fluorescence fingerprint for each sample, information on these sample fluorescence fingerprints (collectively referred to herein as "sample fluorescence fingerprint information") is collected, and a luminescent substance suitable for the desired evaluation of the composition is determined from among multiple candidate luminescent substances based on the sample fluorescence fingerprint information (step S103). Specifically, the sample fluorescence fingerprints of each sample are compared, and candidate luminescent substances for which the sample fluorescence fingerprints show significant differences due to differences in the state of the composition, as well as the wavelengths of their excitation light and the wavelengths of light emitted by the candidate luminescent substances, are identified. The identified candidate luminescent substance is then selected as the luminescent substance for the luminescent layer to be used in the luminescence information acquisition step S12, which will be described later. Note that in this step, only one luminescent substance may be selected, but two or more luminescent substances may also be selected. Furthermore, multiple luminescent substances may be selected according to the evaluation items of the composition.

[0036] The method for comparing the fluorescence fingerprints of multiple samples in this step is not particularly limited, and the fluorescence fingerprints of multiple samples may simply be superimposed and compared. Alternatively, the fluorescence fingerprint information of the samples may be reduced in dimension through statistical analysis processing, and parameterized to succinctly represent the characteristics of each composition state, and then compared. Examples of statistical analysis processing methods include multivariate analysis and data mining. Specific examples include data structure analysis, discriminant analysis, pattern classification, multidimensional data analysis, regression analysis, and machine learning.

[0037] The data structure analysis includes principal component analysis, factor analysis, correspondence analysis, and independent component analysis. The discriminant analysis includes linear discriminant analysis or nonlinear discriminant analysis. The linear discriminant analysis includes canonical discriminant analysis, and the nonlinear discriminant analysis includes decision trees.

[0038] Examples of the pattern classification include cluster analysis and multidimensional scaling. Examples of the regression analysis include linear regression and nonlinear regression. Here, examples of linear discriminant analysis include partial least squares (PLS) regression, simple regression analysis, multiple regression analysis, and principal component regression, and examples of nonlinear discriminant analysis include logistic regression and regression tree.

[0039] Examples of the machine learning include neural networks, self-organizing maps, ensemble learning, and genetic algorithms. The statistical analysis process may be performed using any analytical method that can accurately analyze the composition.

[0040] (Emitting layer preparation process) In the light-emitting layer preparation step, a light-emitting layer containing a light-emitting substance is prepared. The light-emitting layer may have any shape or structure as long as it can emit light when in contact with the composition in the light-emitting information acquisition step described below. For example, the light-emitting layer may be disposed at the tip of an optical fiber, or may be disposed on one surface of a light-guiding planar member such as glass.

[0041] In a probe or the like in which a light-emitting layer is disposed at the tip of an optical fiber, light is guided to the light-emitting layer via the optical fiber, and light emitted by the light-emitting substance in the light-emitting layer can be transmitted to a detection device or the like via the optical fiber. Therefore, detection can be performed efficiently. Hereinafter, an example in which the light-emitting layer is disposed at the tip of an optical fiber (probe) will be described, but the present embodiment is not limited to this example.

[0042] An example of the structure of a sensor including a probe is shown in Figure 4. The sensor 200 has a probe 21, a light source 22, a detection unit 23, and cables 210a and 210b connecting these.

[0043] The probe 21 may have a light-guiding member 21 a and a light-emitting layer 21 b containing a light-emitting substance disposed at the tip of the light-guiding member 21 a. The light-emitting substance contained in the light-emitting layer 21 b is, for example, the light-emitting substance selected in the light-emitting substance selection step described above.

[0044] Here, light-guiding member 21a may be any member capable of guiding light (fluorescence or phosphorescence) emitted by the luminescent material in light-emitting layer 21b toward detection unit 23. It is preferable that light-guiding member 21a is a member capable of guiding excitation light emitted by light source 22 for exciting the luminescent material in light-emitting layer 21b toward light-emitting layer 21b and capable of guiding light (fluorescence or phosphorescence) emitted by the luminescent material in light-emitting layer 21b toward detection unit 23, because the excitation light can be reliably irradiated onto the luminescent material.

[0045] The light-guiding member 21a may be an optical fiber, and may be composed of, for example, one type of optical fiber. In this case, the optical fiber serves to guide the excitation light from the light source 22 to the light-emitting layer 21b side, and to guide the light (fluorescence or phosphorescence) emitted by the light-emitting substance from the light-emitting layer 21b side to the detection unit 23 side. On the other hand, the light-guiding member 21a may be composed of multiple types of optical fibers. In this case, the light-guiding member 21a includes an optical fiber for guiding the excitation light from the light source 22 to the light-emitting layer 21b side, and an optical fiber for guiding the light (fluorescence or phosphorescence) emitted by the light-emitting substance from the light-emitting layer 21b side to the detection unit 23 side. However, from the viewpoint of space saving, it is preferable that the light-guiding member 21a be composed of one type of optical fiber.

[0046] It should be noted that cables 210a, 210b, etc. may be arranged between the light guide member 21a of the probe 21 and the light source 22 or the detector 23, as necessary.

[0047] On the other hand, the light-emitting layer 21b only needs to be disposed at least at the tip of the light-guiding member 21a (probe 21), and may be disposed so as to cover the entire light-guiding member 21a, for example. The light-emitting layer 21b may be a layer containing only a light-emitting substance. Examples of layers containing only a light-emitting substance include porous films and monolayers such as self-assembled monolayers. Such monolayers may cover the entire tip of the probe, or may cover the tip of the probe in an island-like manner, i.e., partially. However, from the viewpoint of strength, it is preferable that the light-emitting substance be a layer bound to the light-guiding member 21a by a binder.

[0048] Furthermore, the amount of the luminescent substance in the luminescent layer 21b is not particularly limited as long as it is an amount that can sufficiently come into contact with the composition and can sufficiently emit light (fluorescence or phosphorescence) when exposed to excitation light. The luminescent layer 21b may contain only one type of luminescent substance, or may contain two or more types. For example, when it is desired to evaluate multiple performances and properties of the composition, one luminescent layer 21b may contain multiple luminescent substances for evaluating each performance or property. Alternatively, multiple probes 21 may be prepared for each luminescent substance, and these may be used to perform the luminescence information acquisition step S12 described below. Preparing multiple probes 21 prevents the luminescence information from each luminescent substance from mixing, making evaluation easier in the evaluation step S13 described below.

[0049] On the other hand, the type of binder contained in the light-emitting layer 21b is not particularly limited, but a material that can transmit excitation light for exciting the light-emitting substance and light emitted by the light-emitting substance is preferable. For example, it may be an inorganic material such as a silicone resin, or an organic resin such as an epoxy resin. The light-emitting layer 21b may contain only one type of binder, or may contain two or more types.

[0050] Here, the method for forming the light-emitting layer 21b at the tip of the light-guiding member 21a is not particularly limited. For example, a light-emitting substance and a binder precursor (e.g., triethoxysilane) may be mixed, applied to the periphery of the light-guiding member 21a, and cured to form the light-emitting layer 21b. In this case, the mixture may be cured by polymerizing only the binder precursor, or the light-emitting substance and the binder precursor may be copolymerized and cured. When the binder precursor is triethoxysilane, a sol-gel reaction may be used for polymerization. Alternatively, for example, an epoxy resin (polymer), a light-emitting substance, and a solvent may be mixed, applied to the periphery of the light-guiding member 21a, and then the solvent may be removed to form the light-emitting layer 21b. Furthermore, when forming a light-emitting layer consisting only of a light-emitting substance, a light-emitting substance having a group (—Si(OCH3)3) derived from a silane coupling agent may be applied to the tip of the optical fiber by an inkjet method or the like, and then polycondensed and cured by a sol-gel reaction or the like. Alternatively, for example, the tip of an optical fiber may be immersed in a light-emitting substance having a group (-Si(OCH3)3) derived from a silane coupling agent, followed by polycondensation and hardening. Note that the light-emitting layer can also be formed in a similar manner when it is formed on a light-guiding planar member such as glass, rather than on the tip of an optical fiber.

[0051] Here, the light source 22 that can be used in the sensor 200 is not particularly limited as long as it can irradiate light of a predetermined wavelength, and can be a supercontinuum light source, an LED, a white light source, etc. Furthermore, the light source 22 may be equipped with a filter for irradiating only light of a specific wavelength, a mechanism for adjusting the intensity of light, etc.

[0052] On the other hand, the detection unit 23 may be any device capable of measuring the wavelength and intensity of light (fluorescence or phosphorescence) emitted by the luminescent substance, and may be, for example, a spectrofluorometer. The detection unit 23 may be connected to an information processing device (not shown) for analyzing luminescence information of the luminescent substance.

[0053] (Light emission information acquisition process) In the luminescence information acquisition step S12, the luminescent layer prepared in the above-mentioned luminescent layer preparation step S11, for example, the luminescent layer 21b of the probe 21 of the above-mentioned sensor 200, is brought into contact with the composition, and in this state, excitation light from the light source 22 is irradiated onto the luminescent substance in the luminescent layer 21b.

[0054] The method for bringing the light-emitting layer 21b into contact with the composition is not particularly limited, and for example, the light-emitting layer 21b may be immersed in the composition, or the composition may be applied to the surface of the light-emitting layer 21b.

[0055] Here, when the luminescence information acquisition step S12 is performed during the production of the composition, the composition may be sampled and the composition may be brought into contact with the light-emitting layer 21b. That is, the luminescence information acquisition step S12 may be performed offline. Alternatively, the light-emitting layer may be immersed in the composition in the production line to bring them into contact. That is, the luminescence information acquisition step S12 may be performed inline. In this embodiment, since the light-emitting substance is fixed to the probe surface, there is very little possibility that the composition will be contaminated by the light-emitting substance, and the luminescence information acquisition step can also be performed inline.

[0056] Furthermore, the wavelength of the excitation light irradiated onto the light-emitting layer 21b (light-emitting substance) in the luminescence information acquisition step S12 may be a specific wavelength that produced a large difference in the fluorescence fingerprint in the above-mentioned luminescent substance selection step S10. On the other hand, since irradiating light with a wide range of wavelengths as excitation light allows for the acquisition of a large amount of information, light from a wide wavelength range may be irradiated sequentially while shifting the wavelength. However, when the luminescence information acquisition step is performed in-line, irradiating light with a specific wavelength is preferable from the viewpoint of efficiency.

[0057] In the luminescence information acquisition step S12, the light emitted by the luminescent substance may be detected by a detection device or the like. For example, when the sensor 200 is used, the light is detected by the detection unit 23 via the light-guiding member 21a or the cable 210b. The luminescence information acquired by the detection unit 23 may be information that allows the composition to be evaluated in the evaluation step S13 described below. For example, the luminescence information may be only the intensity of a specific wavelength emitted by the luminescent substance, or may be the spectrum of light emitted by the luminescent substance. Furthermore, the luminescence information acquired in this step may be information regarding fluorescence emitted by the luminescent substance, or information regarding phosphorescence emitted by the luminescent substance.

[0058] When using the sensor 200, the luminescence information may be obtained using only one probe 21, or multiple probes 21 with different types of luminescent substances may be prepared to obtain multiple pieces of luminescence information.

[0059] Alternatively, a probe (not shown) not having the light-emitting layer 21b may be prepared separately, and the difference in luminescence information may be obtained. Obtaining the difference means canceling out the influence of various factors other than the target item and obtaining only the influence of the target item. This increases the signal-to-noise ratio and provides more effective information. Specifically, a probe 21 having a light-emitting layer 21b and a probe whose light-emitting layer does not contain a luminescent substance are prepared. Then, each of these is brought into contact with a composition, and in this state, light (excitation light) is irradiated onto the probe 21 from the light source 22. The light from each probe is then detected by the detection unit 23. The data detected from the probe whose light-emitting layer does not contain a luminescent substance is then subtracted from the data detected from the probe 21 having the light-emitting layer 21b. This removes noise, allowing for more accurate acquisition of luminescence information derived from the luminescent substance.

[0060] Furthermore, in the light emission information acquisition step S12, a plurality of data may be acquired at regular intervals. By acquiring a plurality of data at regular intervals, for example, the reaction rate, maturation rate, fermentation rate, drying rate, stability over time, etc., when manufacturing a specific product can be grasped over time.

[0061] (Evaluation process) In the evaluation step S13, the luminescence information acquired in the luminescence information acquisition step S12 is analyzed based on the previously acquired fluorescence fingerprint information to evaluate the state of the composition. The fluorescence fingerprint information used in the evaluation step S13 may be the sample fluorescence fingerprint information acquired in the above-mentioned luminescent material selection step S10, or may be fluorescence fingerprint information acquired separately.

[0062] For example, if the luminescent material selection step S10 is not performed, the luminescent material is mixed with a plurality of compositions in different states to prepare a plurality of samples, and a fluorescence fingerprint is obtained for each sample. Information regarding these fluorescence fingerprints (fluorescence fingerprint information) may then be used to evaluate the compositions. The method for obtaining the fluorescence fingerprint information is the same as the method for obtaining sample fluorescence fingerprint information in the luminescent material selection step S10 described above.

[0063] In the evaluation step S13, the previously acquired fluorescence fingerprint information is compared with the luminescence information acquired in the luminescence information acquisition step S12 to determine the state of the composition. As described above, only one performance or property of the composition may be evaluated, or multiple properties or performances may be evaluated. These comparisons can be performed using a general information processing device, such as a personal computer.

[0064] Alternatively, a state estimation model may be created in advance and the state of the composition may be identified using the estimation model. When there are many evaluation items for the composition, the use of the state estimation model allows for more appropriate evaluation.

[0065] The state estimation model can be created by machine learning a previously acquired fluorescence fingerprint and the corresponding state of the composition. Known methods can be used as the machine learning method. For example, for a composition in a specific state, a multivariate analysis is performed using the fluorescence fingerprint as an explanatory variable and the state as a target variable to determine a similarity index. The similarity index can be selected from cosine similarity, Pearson's correlation coefficient, deviation pattern similarity, Euclidean distance similarity, Morishita's similarity index, standard Euclidean distance similarity, Mahalanobis distance similarity, Manhattan distance similarity, Chebyshev distance similarity, Minkowski distance similarity, Jaccard coefficient similarity, Dice coefficient similarity, Simpson coefficient similarity, etc.

[0066] In the similarity index calculation step, the fluorescence fingerprint similarity index is calculated for a specific composition under the assumption that there is a correlation between the state of the composition (the objective variable) and the fluorescence fingerprint similarity index. This is then performed for other compositions in different states. Then, by repeatedly executing and optimizing the calculation so as to minimize the error between the calculated similarity index and the actual state of the composition, an estimation model can be created.

[0067] Many analytical methods have been developed for two-dimensional data. The aforementioned fluorescence fingerprint, on the other hand, is three-dimensional data consisting of the wavelength of excitation light, the wavelength of light emitted by the luminescent substance, and the intensity of that light. Therefore, the three-dimensional data may be expanded into two dimensions for multivariate analysis. For example, the fluorescence fingerprint may be expanded into two-dimensional data consisting of wavelength conditions (combination of excitation wavelength and fluorescence wavelength) and fluorescence intensity, and multivariate analysis may be performed. Furthermore, the two-dimensionally expanded fluorescence fingerprint may be subjected to processes such as mean centering, normalization, autoscaling, second derivative, baseline correction, and smoothing. This allows for emphasizing the information contained in each data point and aligning the scale of data from different samples. Alternatively, multivariate analysis may be performed on the three-dimensional data as is.

[0068] Furthermore, if necessary, marker signals (e.g., combinations of excitation wavelengths and fluorescence wavelengths) important for estimation (i.e., those that change significantly depending on the state of the composition) can be detected from the obtained similarity index. In this case, marker signals may be detected by performing multivariate analysis such as principal component regression, cluster analysis, discriminant analysis, SIMCA, multiple regression analysis, PLS regression analysis, PLS discrimination, SVM regression, SVM discrimination, RF regression, and / or RF discrimination. Marker signals may also be detected based on one or more indices indicating the contribution rate to regression and discrimination, including regression coefficients, factor loadings, loadings, selectivity ratios, variable importance inprojection, variable importance, and out-of-bag error, obtained from multivariate analysis.

[0069] Then, the numerical values ​​of the marker signals are estimated for a composition different from the composition for which the similarity index was calculated. After that, the estimated data is compared with the actual data, and optimization is performed repeatedly to reduce the error between them, thereby obtaining a state estimation model that can estimate the state of the composition based on a specific marker signal.

[0070] When such a state estimation model is created, the state of the composition can be evaluated by applying the light emission information obtained in the light emission information acquisition step to the state estimation model.

[0071] (others) In the above explanation, an embodiment in which a luminescent material selection process, a luminescent layer preparation process, a luminescence information acquisition process, and an evaluation process are performed is described. However, if a suitable luminescent material is already known, the luminescent material selection process does not have to be performed.

[0072] (effect) As described above, in the evaluation method of this embodiment, the state of a composition is evaluated based on previously acquired fluorescence fingerprint information. This method allows for comprehensive evaluation of the properties and performance of a composition, even if the individual components of the composition are not identified or if multiple substances in the composition interact in a complex manner. In other words, it is not necessary to perform complex component analysis of the composition, and it is possible to evaluate the performance and characteristics of a composition even when the correlation between each component and the performance or characteristics is unclear.

[0073] Furthermore, the use of fluorescence fingerprint information makes it possible to detect even slight changes, and it is possible to comprehensively evaluate the state of a composition at the appropriate time during product research and development or on the production line without affecting the composition.

[0074] 2. Sensors and evaluation systems As one embodiment, the present invention also provides a sensor that can be used in the above-mentioned evaluation method. The sensor simply needs to include a light-emitting unit having a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition, an excitation light source that emits excitation light to excite the light-emitting substance, and a detection unit that acquires information about the light emitted by the light-emitting substance. For example, the light-emitting unit, excitation light source, and detection unit do not need to be connected to each other. For example, the light-emitting unit may be a structure (probe) having an optical fiber and a light-emitting layer disposed at the tip of the optical fiber, or may be a structure having a light-guiding planar member such as glass and a light-emitting layer disposed on the light-guiding planar member.

[0075] However, if the light-emitting unit is the above-mentioned probe, it is preferable in that it can be easily connected to an excitation light source and a detection unit, the sensor is easy to handle, and it can be made smaller. Hereinafter, a sensor having such a probe will be described as an example, but the sensor of this embodiment is not limited to this structure.

[0076] 4, the sensor includes a probe 21 having a light-guiding member 21a and a light-emitting layer 21b disposed at the tip of the light-guiding member 21a and containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition, a light source 22 that emits excitation light for exciting the light-emitting substance, and a detection unit 23 that acquires information about the light emitted by the light-emitting substance. The configurations of these components are the same as those of the sensor prepared in the light-emitting layer preparation step S12 of the evaluation method described above.

[0077] However, the wavelength of the excitation light emitted by the light source 22 in the sensor 20 is a wavelength determined based on previously acquired fluorescence fingerprint information (Aspect 1), or the luminescent substance is a substance determined based on previously acquired fluorescence fingerprint information (Aspect 2).

[0078] In the first embodiment, the previously acquired fluorescence fingerprint information is information obtained when a plurality of samples are prepared by mixing a specific luminescent substance (the luminescent substance contained in the luminescent layer 21b) with a plurality of compositions in different states, and a fluorescence fingerprint is obtained for each of these samples. By measuring the fluorescence fingerprints of each of these samples and comparing the fluorescence fingerprints, it becomes clear what wavelength of excitation light irradiation is necessary to easily identify a change in state of the composition. The wavelength at which the change in state is easily identified is then determined as the wavelength of the excitation light emitted by the light source 22. Therefore, with the sensor 20 of the first embodiment, the state of the composition can be ascertained simply by irradiating the composition with light of a specific wavelength as excitation light, without using excitation light over a wide wavelength range.

[0079] On the other hand, the previously acquired fluorescence fingerprint information in embodiment 2 is information obtained when multiple samples are prepared by mixing multiple candidate luminescent substances with multiple compositions in different states, and a fluorescence fingerprint is obtained for each of these samples. By obtaining a fluorescence fingerprint for each of these samples and comparing the fluorescence fingerprints, it becomes clear which candidate luminescent substance (light-emitting substance) is used to easily determine a change in the state of the composition. The candidate luminescent substance whose change in state is easily determined is then determined as the luminescent substance contained in light-emitting layer 21b. Therefore, according to sensor 20 of embodiment 2, differences in the state of the composition are easily reflected in the luminescence information from the light-emitting substance, making it easy to grasp the state of the composition.

[0080] In either embodiment, the method for obtaining the fluorescence fingerprint is the same as the method described in the luminescent material selection step S10.

[0081] As one embodiment, the present invention also provides an evaluation system having any one of the above sensors and an information processing unit. The information processing unit of the evaluation system analyzes the luminescence information acquired by the detection unit of the sensor using previously acquired fluorescence fingerprint information to evaluate the state of the composition. Here, the type of information processing unit is not particularly limited and can be a general information processing device, such as a personal computer.

[0082] The fluorescence fingerprint information used by the evaluation system may be fluorescence fingerprint information acquired to determine the wavelength of the excitation light of the sensor, or may be fluorescence fingerprint information acquired to determine the luminescent material of the sensor, or may be fluorescence fingerprint information acquired separately.

[0083] Furthermore, in the evaluation system, the information processing unit may be trained in advance by machine learning to create a state estimation model, and the composition may be evaluated using the state estimation model. The state estimation model can be created in the same way as the state estimation model used in the above-mentioned composition evaluation method. [Example]

[0084] Specific examples of the present invention will be described below along with comparative examples, but the present invention is not limited to these.

[0085] (1) Selection of luminescent material Resin composition A (concentration of cycloolefin polymer a: 30 g / L) was prepared by dissolving recyclable cycloolefin polymer a in methylene chloride, and resin composition B (concentration of cycloolefin polymer b: 30 g / L) was prepared by dissolving non-recyclable cycloolefin polymer b (cycloolefin polymer b is a resin obtained by recycling cycloolefin polymer a multiple times) in methylene chloride. Resin composition A and resin composition B were then mixed with 24 types of luminescent substances, respectively. The number of standard samples was 50, i.e., 2 (resin compositions A and B) × 25 (24 types (number of luminescent substances) + 1 (no luminescent substance)). Fluorescence fingerprints were measured for each of these standard samples. The fluorescence fingerprints were measured using a commercially available spectrofluorometer (Hitachi High-Tech Science Corporation, F-7000). The fluorescence fingerprints were measured by measuring the wavelength and intensity of the fluorescence emitted by the standard samples over an excitation wavelength range of 250 nm to 700 nm, shifting the wavelength by 10 nm. These were then converted into three-dimensional data to obtain a fluorescent fingerprint.

[0086] The obtained fluorescence fingerprints were analyzed, and the fluorescence fingerprints of resin compositions A and B were compared to identify the luminescent substance (BASF's IRGANOX 1076) that showed the greatest change, as well as the characteristic excitation and emission wavelengths (excitation wavelength: 320 nm / emission wavelength: 410 nm). There was no difference in the data without the luminescent substance. Figure 5A shows the fluorescence fingerprint obtained for resin composition A using the luminescent substance (BASF's IRGANOX 1076), and Figure 5B shows the fluorescence fingerprint obtained for resin composition B using the luminescent substance (BASF's IRGANOX 1076).

[0087] (2) Sensor fabrication A composition made by mixing an epoxy polymer and a light-emitting material (IRGANOX 1076 manufactured by BASF) was applied to the end of an optical fiber to form a 1 mm-thick light-emitting layer, thereby producing a sensor. The optical fiber was equipped with an excitation light fiber and a detection light fiber. The excitation light fiber was connected to a light source capable of emitting light with a wavelength of 320 nm. Meanwhile, the detection light fiber was connected to a detection unit (spectrofluorophotometer, F-7000 manufactured by Hitachi High-Tech Science Corporation).

[0088] (3) Evaluation of the composition A resin composition C (concentration of cycloolefin polymer c: 30 g / L) was prepared by dissolving cycloolefin polymer c, the number of times of recycling of which was unknown, in methylene chloride. The end of the sensor was immersed in the resin composition C, and light with a wavelength of 320 nm was irradiated using an excitation light fiber. The fluorescence emitted by the light-emitting layer was then detected by a spectrofluorometer via a detection light fiber. The obtained data was compared with the fluorescence fingerprint data by an information processing unit to determine whether cycloolefin polymer c could be recycled.

[0089] This application claims priority from Japanese Patent Application No. 2021-143165, filed September 2, 2021. The contents of the specification and drawings of that application are incorporated herein by reference in their entirety. [Industrial Applicability]

[0090] The composition evaluation method of the present invention makes it possible to comprehensively evaluate the state of a composition in product research and development or on a production line without affecting the composition itself, and is therefore useful for the testing, research, production, etc. of various compositions. [Explanation of symbols]

[0091] 21 Probe 21a Light guide member 21b Light-emitting layer 22 Light source 23 Detector 210a, 210b cables

Claims

1. A method for manufacturing a light-emitting device, comprising: preparing a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition; a step of irradiating the light-emitting layer with excitation light while the light-emitting layer and the composition are in contact with each other, and acquiring luminescence information of the light-emitting substance; analyzing the luminescence information based on previously acquired fluorescence fingerprint information to evaluate the state of the composition; and The light-emitting layer is disposed at the tip of an optical fiber or on a light-guiding planar member. Methods for evaluating compositions.

2. A method of manufacturing a light-emitting device, comprising: preparing a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition; a step of irradiating the light-emitting layer with excitation light while the light-emitting layer and the composition are in contact with each other, and acquiring luminescence information of the light-emitting substance; analyzing the luminescence information based on previously acquired fluorescence fingerprint information to evaluate the state of the composition; and The step of acquiring the light emission information is carried out within a production line for the composition. Methods for evaluating compositions.

3. Before the step of preparing the light-emitting layer, preparing a plurality of samples containing a plurality of the compositions in different states and a plurality of types of candidate luminescent substances; acquiring sample fluorescence fingerprint information for the plurality of samples; selecting the luminescent material based on the sample fluorescence fingerprint information; The method for evaluating a composition according to claim 1 or 2, further comprising:

4. determining a wavelength of the excitation light based on the sample fluorescence fingerprint information; A method for evaluating the composition according to claim 3.

5. a light-emitting portion having a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition; an excitation light source that emits excitation light for exciting the luminescent material; a detection unit that acquires information about the light emitted by the luminescent substance; and the wavelength of the excitation light is determined based on previously acquired fluorescence fingerprint information, In the light-emitting section, the light-emitting layer is disposed at a tip of an optical fiber or on a light-guiding planar member. sensor.

6. a light-emitting portion having a light-emitting layer containing a light-emitting substance whose light-emitting behavior changes depending on the state of the composition; an excitation light source that emits excitation light for exciting the luminescent material; a detection unit that acquires information about the light emitted by the luminescent substance; and the luminescent substance is a substance selected based on previously acquired fluorescence fingerprint information, In the light-emitting section, the light-emitting layer is disposed at a tip of an optical fiber or on a light-guiding planar member. sensor.

7. A sensor according to claim 5 or 6; an information processing unit that analyzes the light information acquired by the detection unit using previously acquired fluorescence fingerprint information to evaluate the state of the composition; and A rating system having:

Images