Carbon nanotube length measurement system

Abstract

Provided is a carbon nanotube length measurement system comprising: an image data acquisition unit for acquiring image data of a carbon nanotube; a physical property data acquisition unit for acquiring a plurality of pieces of physical property data of the carbon nanotube; at least one processor for executing a command for measuring the length of the carbon nanotube by using the data acquired by the acquisition units; and a memory storing at least one command executed through the at least one processor, wherein the at least one command comprises: a command for directly extracting the length from the image data of the carbon nanotube; a command for generating an indirect length for correcting an error occurring in the direct length by combining the plurality of pieces of physical property data of the carbon nanotube; a command for generating a correction model for correcting a length measurement error on the basis of the direct length and the indirect length; and a command for determining the length of the carbon nanotube of which the length measurement error has been corrected by using the correction model. The representativeness and reliability of measurement results may be improved through the carbon nanotube length measurement system including the above configuration and commands.

Description

Carbon nanotube length measurement system

[0001] The present application claims the benefit of priority based on Korean Patent Application No. 10-2025-0021384 filed February 19, 2025 and Korean Patent Application No. 10-2026-0021436 filed February 3, 2026, and includes all contents disclosed in the documents of said Korean patent applications as part of the specification.

[0002] The present invention relates to a carbon nanotube length measurement system. Specifically, it relates to a carbon nanotube length measurement system using a correction model. More specifically, the invention relates to a carbon nanotube length measurement system comprising: an image data acquisition unit for acquiring image data of a carbon nanotube; a material property data acquisition unit for acquiring a plurality of material property data for a carbon nanotube; at least one processor for executing a command to measure the length of a carbon nanotube using data acquired by the acquisition unit; and a memory for storing at least one command executed through the at least one processor. The at least one command comprises: a command to extract a direct length from the image data of the carbon nanotube; a command to generate an indirect length to correct an error occurring in the direct length by combining a plurality of material property data for the carbon nanotube; a command to generate a correction model to correct a length measurement error based on the direct length and the indirect length; and a command to determine the length of the carbon nanotube with the length measurement error corrected using the correction model.

[0003] Carbon nanotubes (CNTs) are tube-shaped carbon allotropic materials with a diameter of several to tens of nanometers and a length ranging from tens of micrometers to tens of centimeters. They can be broadly classified into single-walled carbon nanotubes (SWCNTs) and multi-walled carbon nanotubes (MWCNTs) depending on the number of graphene layers. These carbon nanotubes are well known to possess excellent properties, such as mechanical strength, electrical conductivity, thermal conductivity, and chemical resistance, along with their unique structure.

[0004] Due to their excellent inherent properties, carbon nanotubes as described above are the subject of extensive research in various fields, such as reinforcing agents for composite materials and the manufacture of transparent conductive films. In particular, as interest in energy storage technology continues to grow and lithium-ion batteries—which are capable of charging and discharging, have high operating voltages, and possess significantly higher energy densities among various electrochemical devices—are gaining prominence, carbon nanotubes are being actively utilized as conductive materials with excellent electrical conductivity and strength during electrode manufacturing.

[0005] As such, carbon nanotubes are used as additives in various fields, and the primary purpose of their use is to improve properties such as mechanical strength or electrical conductivity. When carbon nanotubes are used as additives, their utilization and content are determined based on the numerical values ​​of their physical properties, such as mechanical strength or electrical conductivity.

[0006] However, there are physical limitations to individually measuring properties such as mechanical strength or electrical conductivity for all available carbon nanotubes to determine their suitability; consequently, the length of carbon nanotubes, which is closely related to these characteristics, plays a crucial role in their selection. Accordingly, methods primarily used to determine carbon nanotube length have involved measuring aspect ratios using Atomic Force Microscopy (AFM) or Scanning Electron Microscopy (SEM), or indirectly deriving length by measuring viscosity in a liquid state. However, while aspect ratio measurements via microscope allow for accurate readings of observed nanotubes, they are difficult to apply to long nanotubes and are limited to localized measurements. Furthermore, the method of indirectly deriving length through viscosity is susceptible to variations caused by various variables such as temperature due to limitations in the dynamic properties of the dispersion; additionally, significant discrepancies may arise from the actual values ​​if the carbon nanotubes are not completely dispersed.

[0007] Therefore, there is a need to develop a method for measuring the length of carbon nanotubes that enables accurate measurement of the carbon nanotubes present in the sample to be measured, while also allowing for an overall length evaluation of the sample.

[0008] [Prior Art Literature]

[0009] [Non-patent literature]

[0010] (Non-patent literature 1) Macromolecules 2016, 49, 2, 681-689

[0011] The objective of the present invention is to provide a carbon nanotube length measurement system comprising: an image data acquisition unit for acquiring image data of a carbon nanotube; a material property data acquisition unit for acquiring a plurality of material property data for a carbon nanotube; at least one processor for executing a command to measure the length of a carbon nanotube using data acquired by the acquisition unit; and a memory for storing at least one command executed through the at least one processor, wherein the at least one command comprises: a command to extract a direct length from the image data of the carbon nanotube; a command to generate an indirect length to correct an error occurring in the direct length by combining a plurality of material property data for the carbon nanotube; a command to generate a correction model to correct a length measurement error based on the direct length and the indirect length; and a command to determine the length of the carbon nanotube with the length measurement error corrected using the correction model.

[0012] A first aspect of the present invention provides a method for measuring the length of a carbon nanotube, comprising: a generation step of generating a correction model for correcting a length measurement error by inputting an indirect length and a direct length of the carbon nanotube; and a determination step of determining the length of the carbon nanotube with the length measurement error corrected through the correction model, wherein the indirect length is generated by combining a plurality of different carbon nanotube physical property data, and the direct length is extracted from carbon nanotube image data.

[0013] In one embodiment of the present invention, the correction model is a Bayesian learning model.

[0014] In one embodiment of the present invention, the correction model is a model capable of learning the probability distribution between random variables using input data as a random variable.

[0015] In one embodiment of the present invention, the generation step includes a process of learning the probability distribution of an indirect length using input indirect length data as a probability variable.

[0016] In one embodiment of the present invention, the generation step includes a process of learning the probability distribution of the indirect length using input indirect length data as a probability variable, and the determination step includes a process of adjusting the probability distribution of the indirect length by reflecting the direct length.

[0017] In one embodiment of the present invention, the determination step includes the step of deriving a posterior probability of a carbon nanotube length through a prior probability and a likelihood, wherein the prior probability is generated from an indirect length and the likelihood is generated from a direct length.

[0018] In one embodiment of the present invention, the plurality of physical property data are selected from the group consisting of thickness, aspect ratio, extensional viscosity, and combinations thereof.

[0019] In one embodiment of the present invention, the indirect length is generated from a combination of the thickness and aspect ratio of the carbon nanotube, and the aspect ratio is generated from the elongation viscosity of the carbon nanotube.

[0020] In one embodiment of the present invention, the thickness of the carbon nanotube is measured using an atomic force microscope (AFM).

[0021] In one embodiment of the present invention, the extraction of direct lengths includes the step of extracting each carbon nanotube backbone from the image data.

[0022] In one embodiment of the present invention, the extraction of the direct length includes the step of calculating the circumference of each skeleton from the extracted carbon nanotube skeleton.

[0023] In one embodiment of the present invention, the perimeter of the skeleton is calculated through polygon approximation.

[0024] In one embodiment of the present invention, image data of the carbon nanotube is captured from a scanning electron microscope (SEM).

[0025] In one embodiment of the present invention, the carbon nanotube comprises a single-walled carbon nanotube (SWCNT).

[0026] A second aspect of the present invention provides a carbon nanotube length measurement system comprising: an image data acquisition unit for acquiring image data of a carbon nanotube; a material property data acquisition unit for acquiring a plurality of material property data for a carbon nanotube; at least one processor for executing a command to measure the length of a carbon nanotube using data acquired by the acquisition unit; and a memory for storing at least one command executed through the at least one processor. The at least one command comprises: a command to extract a direct length from the image data of the carbon nanotube; a command to generate an indirect length to correct an error occurring in the direct length by combining a plurality of material property data for the carbon nanotube; a command to generate a correction model to correct a length measurement error based on the direct length and the indirect length; and a command to determine the length of the carbon nanotube with the length measurement error corrected using the correction model.

[0027] In one embodiment of the present invention, the command for generating the correction model includes a command for learning the probability distribution of the indirect length using input indirect length data as a probability variable.

[0028] In one embodiment of the present invention, a command for generating the correction model includes a command for learning the probability distribution of the indirect length using input indirect length data as a probability variable, and a command for determining the length of the carbon nanotube includes a command for adjusting the probability distribution of the indirect length by reflecting the direct length.

[0029] In one embodiment of the present invention, a command for determining the length of the carbon nanotube includes a command for deriving a posterior probability of the carbon nanotube length through a prior probability and a likelihood, wherein the prior probability is generated from an indirect length and the likelihood is generated from a direct length.

[0030] In one embodiment of the present invention, the material property data acquisition unit includes an atomic force microscope, and the thickness of the carbon nanotube is acquired from the atomic force microscope.

[0031] In one embodiment of the present invention, the direct length extraction command includes a command to extract each carbon nanotube backbone from the image data.

[0032] In one embodiment of the present invention, the direct length extraction command includes a command to calculate the circumference of each skeleton from the extracted carbon nanotube skeleton.

[0033] In one embodiment of the present invention, the image data acquisition unit includes a scanning electron microscope, and the image data of the carbon nanotube is acquired from the scanning electron microscope.

[0034] A carbon nanotube length measurement system according to the present invention is a system for measuring the length of a carbon nanotube, comprising: an image data acquisition unit for acquiring image data of a carbon nanotube; a material property data acquisition unit for acquiring a plurality of material property data for a carbon nanotube; at least one processor for executing a command to measure the length of a carbon nanotube using data acquired by the acquisition unit; and a memory for storing at least one command executed through the at least one processor. The at least one command includes a command to extract a direct length from the image data of the carbon nanotube; a command to generate an indirect length to correct an error occurring in the direct length by combining a plurality of material property data for the carbon nanotube; a command to generate a correction model to correct a length measurement error based on the direct length and the indirect length; and a command to determine the length of the carbon nanotube with the length measurement error corrected using the correction model. By combining information on the overall length of the carbon nanotube included in the sample to be measured with information on the length of a specific carbon nanotube in the sample, the system has the characteristic of improving the representativeness and reliability of the measurement result.

[0035] Furthermore, if the measurement system of the present invention is utilized in processes in related fields where carbon nanotubes are used, the predictability of physical properties that change directly or indirectly depending on the length of the carbon nanotubes can be improved, thereby enabling the securing of optimal process conditions. In particular, if the measurement system of the present invention is utilized in a battery manufacturing process, the predictability of electrical conductivity affected by the length of the carbon nanotubes can be improved, thereby enabling the manufacture of batteries with improved electrochemical properties.

[0036] [Fig. 1] is an image obtained after performing carbon nanotube skeleton extraction through a skeletonize algorithm for an embodiment of the present invention.

[0037] [Fig. 2] is a flowchart of the operation sequence of a carbon nanotube length measurement system according to an embodiment of the present invention.

[0038] [Fig. 3] is a flowchart of the operation sequence of a carbon nanotube direct length extraction command according to an embodiment of the present invention.

[0039] [Fig. 4] is a block diagram of a carbon nanotube length measurement system according to an embodiment of the present invention.

[0040] The present invention will be described in more detail below.

[0041] Prior to this, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention. Accordingly, the configurations described in the embodiments described in this specification are merely one preferred embodiment of the invention and do not represent all of the technical spirit of the invention; therefore, it should be understood that various equivalents and modifications capable of replacing them may exist at the time of filing this application.

[0042] In this specification, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0043]

[0044] The present invention will be described in detail below.

[0045]

[0046] Carbon nanotube length measurement method

[0047] The present invention provides a method for measuring the length of carbon nanotubes.

[0048] The above carbon nanotube is a secondary structure formed by assembling carbon nanotube units to form a bundle shape, either wholly or partially, wherein the carbon nanotube units have a graphite sheet in the shape of a cylinder with a nano-sized diameter, and sp 2 They possess a bonded structure. Carbon nanotubes exhibit unique mechanical, electrical, and thermal properties due to this distinctive structure, and are utilized in various fields such as electrochemistry, biomedicine, and mechanical engineering because of these unique properties.

[0049] Carbon nanotubes are primarily used in the aforementioned fields to complement mechanical, electrical, or thermal properties. To achieve this objective, it is necessary to identify the properties of the carbon nanotubes to be used and then appropriately adjust their content. However, since there are physical limitations to measuring all existing carbon nanotubes for the desired properties and selecting the most suitable one, a method is currently employed to select the nanotubes intended for priority use by measuring factors such as length, which influence intrinsic properties like electrical conductivity.

[0050] Conventionally, to measure the length of such carbon nanotubes, it was common practice to use an Atomic Force Microscope (AFM) to capture images and derive the length by calculating the thickness of the carbon nanotubes. However, there was a problem in that the accuracy and reliability were reduced due to errors in the measurement results caused by self-aggregation resulting from the Van der Waals action of the carbon nanotubes and impurities in the carbon nanotube samples to be measured.

[0051] Although methods utilizing dynamic light scattering or multiagngle light scattering are used to address these issues, they have the disadvantage of only being able to measure the average length of liquid samples. Furthermore, problems regarding reduced accuracy due to self-aggregation and impurities still remain.

[0052] Accordingly, the present invention provides a method for measuring the length of a carbon nanotube in which the representativeness and reliability of the derived carbon nanotube length are enhanced by determining the length of the carbon nanotube with the length measurement error corrected through a method of inputting the indirect and direct lengths of the carbon nanotube into a correction model to solve the above-mentioned problems.

[0053] In one embodiment of the present invention, the carbon nanotube length measurement method comprises a generation step of generating a correction model for correcting a length measurement error by inputting the indirect length and direct length of the carbon nanotube, and a determination step of determining the length of the carbon nanotube with the length measurement error corrected through the correction model.

[0054] Below, each component of the above-mentioned carbon nanotube length measurement method will be explained in detail.

[0055]

[0056] Indirect length

[0057] The above indirect length is the total length of carbon nanotubes within the sample to be measured, generated by combining multiple different carbon nanotube property data.

[0058] The aforementioned multiple different carbon nanotube property data refers to multiple data sets measuring different types of properties; it refers to data that does not directly measure the length of the carbon nanotube but can generate the length of the carbon nanotube through combination with other types of properties via mathematical relationships, etc. For example, crystallinity and I of the Raman spectrum, which can be directly or indirectly related to the length of the carbon nanotube. G / I D Ratio, elongation viscosity, thickness, aspect ratio, etc., may correspond to such material property data, but are not limited thereto if the length of the carbon nanotube can be generated through a combination with other material property-related data. Multiple different data refer to the selection of data measuring different materials among these materials.

[0059] In one embodiment of the present invention, the carbon nanotube material property data comprises being selected from the group consisting of thickness, aspect ratio, extensional viscosity, and combinations thereof. When an indirect length is generated by combining the selected from the group, a length with a small error from the actual carbon nanotube length can be generated.

[0060] The thickness, aspect ratio, and elongation viscosity of the carbon nanotubes can be measured and generated through measurement methods for the physical properties generally known in the relevant technical field. For example, they can be measured using an atomic force microscope (AFM), a rheometer, a transmission electron microscope (TEM), etc., but are not limited thereto.

[0061] In one embodiment of the present invention, the indirect length is generated from a combination of the thickness and aspect ratio of the carbon nanotube, and the aspect ratio is derived from the elongation viscosity of the carbon nanotube. Specifically, the derivation of the indirect length comprises the steps of measuring the elongation viscosity of the carbon nanotube, calculating the aspect ratio of the carbon nanotube from the elongation viscosity, and generating the indirect length by combining the aspect ratio with the thickness of the carbon nanotube.

[0062] In one embodiment of the present invention, the thickness of the carbon nanotube is measured using an atomic force microscope (AFM). When the thickness of the carbon nanotube is measured using the atomic force microscope, the occurrence of error in the measurement result can be minimized.

[0063] In one embodiment of the present invention, the elongation viscosity of the carbon nanotube is measured using a rheometer. Measuring the elongation viscosity of the carbon nanotube using the rheometer can improve the speed and accuracy of the measurement.

[0064] The derivation of the aspect ratio from the elongation viscosity of the carbon nanotubes can be achieved through formulas or theoretical models representing the relationship between elongation viscosity and the aspect ratio. Specifically, it can be derived from a relationship representing the theoretical relationship between elongation viscosity and the aspect ratio in a rigid rod solution, but is not limited thereto as long as a reliable aspect ratio can be derived, and can be performed by hardware, software, or a combination thereof.

[0065] The combination of the above-mentioned multiple different carbon nanotube property data can be achieved through simple mathematical relationships or theoretical modeling, but is not limited thereto, and can be performed by hardware, software, or a combination thereof. For example, the crystallinity of carbon nanotubes or I of the Raman spectrum G / I DThe length of the carbon nanotube can be generated through a mathematical calculation process involving other physical properties that can derive the length of the carbon nanotube by measuring ratios, etc., or through theoretical modeling; alternatively, the length of the carbon nanotube can be generated by measuring the thickness and aspect ratio of the carbon nanotube and multiplying the aforementioned values.

[0066] In one embodiment of the present invention, the indirect length generation from the combination of the thickness and aspect ratio of the carbon nanotube is to generate the indirect length through the product of the thickness and aspect ratio of the carbon nanotube.

[0067]

[0068] Direct length

[0069] The above direct length was obtained by selecting a specific carbon nanotube within the sample to be measured and directly measuring its length, and was extracted from image data of the carbon nanotube.

[0070] The image data of the carbon nanotubes described above refers to image data captured at a scale sufficient to confirm the skeleton of the carbon nanotubes contained within the measurement sample, and said image data may be captured using methods generally known in the relevant technical field. For example, it may be captured using an Atomic Force Microscope (AFM), a Transmission Electron Microscope (TEM), a Scanning Electron Microscope (SEM), etc., but is not limited thereto.

[0071] In one embodiment of the present invention, the image data of the carbon nanotube is captured using a scanning electron microscope (SEM). When an image of the carbon nanotube is captured using the scanning electron microscope (SEM), an image suitable for measuring the length of the carbon nanotube can be obtained, while also demonstrating ease of measurement and suitability for application in actual processes.

[0072] Extracting the length directly from the image data of the carbon nanotube above may involve manually identifying the skeleton of the carbon nanotube on the image data and comparing it with the scale of the image data to directly extract the length of the carbon nanotube, or it may involve separating the skeleton of the carbon nanotube to be identified and extracting its length through a separate processing step on the captured image data, but is not limited thereto and may be performed by hardware, software, or a combination thereof.

[0073] In one embodiment of the present invention, the extraction of direct lengths includes the step of extracting each carbon nanotube backbone from the image data. The extraction of the carbon nanotube backbones may be performed manually or through a separate image processing process, but is not limited thereto and may be performed by hardware, software, or a combination thereof.

[0074] In one embodiment of the present invention, the extraction of the carbon nanotube skeleton is performed through image data processing. The processing of image data for skeleton extraction refers to extracting the skeleton of a specific object through relationships such as shading differences and color tone changes between each pixel within the image. This process can be performed using a relevant program or by methods commonly known in the art. For example, algorithm libraries that can be used for image analysis, such as OpenCV or Scikit Image, or in particular algorithms such as skeletonization algorithms within said libraries, can be executed through programs such as Python, but are not limited thereto.

[0075] In one embodiment of the present invention, the extraction of the carbon nanotube skeleton is performed using a skeletonize algorithm.

[0076] In one embodiment of the present invention, the derivation of the direct length includes the step of calculating the circumference of each skeleton from the extracted carbon nanotube skeleton. By including the step of calculating the circumference of the skeleton in the derivation of the direct length, the accuracy of the length measurement can be improved by compensating for problems such as the selection of reference points or the uneven progression path of the skeleton that may occur when attempting to directly extract the length from the extracted carbon nanotube skeleton. The extraction of the carbon nanotube skeleton may be performed manually or through a separate image data processing process, but is not limited thereto and may be performed by hardware, software, or a combination thereof.

[0077] In one embodiment of the present invention, the calculation of the perimeter of the skeleton is performed through image data processing. The processing of image data for the perimeter calculation refers to predicting the three-dimensional structure of the carbon nanotube from the extracted carbon nanotube skeleton. This process can be performed using a relevant program or by methods commonly known in the art. For example, algorithm libraries that can be used for image analysis, such as OpenCV or Scikit Image, or in particular algorithms such as polygon approximation within said libraries, can be performed through programs such as Python, but are not limited thereto.

[0078] In one embodiment of the present invention, the perimeter of the skeleton is calculated through polygon approximation.

[0079] In one embodiment of the present invention, the direct length is calculated as half the value of the circumference of the skeleton calculated as above.

[0080] In one embodiment of the present invention, the extraction of direct lengths includes a step of recognizing each carbon nanotube from the image data prior to the step of extracting the carbon nanotube skeleton. The step of recognizing carbon nanotubes is a step of separating information about the background and information about the carbon nanotubes within the image data and individually recognizing a plurality of carbon nanotubes included in the image data. The recognition of carbon nanotubes may be performed manually or through a separate image data processing process, but is not limited thereto and may be performed by hardware, software, or a combination thereof. Specifically, the derivation of direct lengths includes (1) a step of recognizing each carbon nanotube from the image data and (2) a step of extracting each carbon nanotube skeleton from the image data.

[0081] In one embodiment of the present invention, the recognition of carbon nanotubes is performed through image data processing. The processing of image data for carbon nanotube recognition refers to distinguishing similar objects and separating each object individually through relationships such as shading differences, color tone changes, and the presence or absence of patterns between each pixel within the image data. This process can be performed using a relevant program or by methods commonly known in the relevant technical field. For example, algorithm libraries that can be used for image analysis, such as the Segment Anything model, OpenCV, or Scikit image, and particularly algorithms capable of segmentation within said libraries, can be executed through programs such as Python, but are not limited thereto.

[0082] In one embodiment of the present invention, the recognition of the carbon nanotube is performed using a Segment Anything model.

[0083]

[0084] Generation and Decision Steps

[0085] The above generation step is a step in which the indirect and direct lengths of carbon nanotubes are input as parameters usable in a correction model for correcting length measurement errors, and the correction model generates them in accordance with the theory upon which the model is based. Furthermore, the above determination step is a step in which, based on the results of the generation, the input parameters are utilized to reach the length of the carbon nanotube with the length measurement error corrected, and a determined value is output.

[0086] While the aforementioned indirect and direct lengths have improved accuracy in their derivation processes, the indirect length is not a value generated by directly measuring the length of the carbon nanotubes contained within the sample, but rather by combining multiple different data points related to the physical properties of carbon nanotubes. Therefore, in this regard, although the indirect length possesses a certain representativeness by representing the overall length of carbon nanotubes within the sample, it lacks some reliability as errors may occur depending on the degree of aggregation and impurities of the carbon nanotubes within the sample. Conversely, the direct length is a value extracted by directly measuring the length from image data of the carbon nanotubes. Accordingly, while the reliability of the results is very high, the direct length lacks representativeness of the overall carbon nanotubes within the sample, as it represents only a tiny fraction of the countless carbon nanotubes contained therein.

[0087] Accordingly, the carbon nanotube length measurement method of the present invention aims to provide a carbon nanotube length measurement method that includes the above-mentioned generation and determination steps to derive the length of the carbon nanotube with the length measurement error corrected from the indirect length and the direct length, thereby compensating for the issues of representativeness and reliability inherent in each variable and determining the length of the carbon nanotube in the sample to approximate the actual value.

[0088] The aforementioned correction model is a model capable of deriving a value with corrected length measurement errors from the theory on which the model is based by learning input variables. In other words, it refers to a model that is based on a theory capable of complementing the mutual representativeness and reliability of indirect and direct lengths, and can determine the length of a carbon nanotube with corrected length measurement errors by learning the input indirect and direct lengths as variables according to the theory on which the model is based.

[0089] As such correction models, generative models generally known in the relevant technical field may be used; for example, models such as Bayesian learning models, non-linear regression models, and decision tree models may be used, but are not limited thereto.

[0090] In one embodiment of the present invention, the indirect length of carbon nanotubes input to the correction model is 10 to 500. Specifically, it may be 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, or 70 or more, 500 or less, 450 or less, 400 or less, 350 or less, 300 or less, 250 or less, 200 or less, 150 or less, 100 or less, 90 or less, or 80 or less, and may be 10 to 500, 20 to 400, or 50 to 200. When the number of indirect lengths satisfies the above range, it is possible to generate a reliable correction model while ensuring the economic feasibility and efficiency of carbon nanotube length measurement suitable for actual application.

[0091] In one embodiment of the present invention, the correction model is a model capable of learning the probability distribution among probability variables using input data as a probability variable. Learning the probability distribution means deriving an overall probability distribution for a given variable based on input homogeneous variables. For example, if the correction model uses indirect length as a probability variable, multiple indirect lengths are input to the correction model, and the correction model derives the probability distribution of the indirect length for a single measurement sample based on the multiple indirect lengths.

[0092] By using a model capable of learning the probability distribution between the above random variables as the correction model, the length of the entire population can be considered without bias toward the results of the input sample group, and the influence of errors existing within the input variables on the results can also be reduced to some extent.

[0093] In one embodiment of the present invention, the probability distribution represents a Gaussian distribution.

[0094] In one embodiment of the present invention, the generation step includes a process of learning the probability distribution of the indirect length using the input indirect length data as a probability variable. The learning of the probability distribution is identical in content to the method described above, and a plurality of indirect measurement data are input to utilize the indirect length data as a probability variable. As described above, the indirect length is a variable that possesses a certain representativeness as it represents the overall length of the carbon nanotube, although its reliability is somewhat lacking; therefore, when the probability distribution is learned using such indirect length as a probability variable, a probability distribution for the carbon nanotube length with a certain representativeness can be derived.

[0095] In one embodiment of the present invention, the generation step includes a process of learning the probability distribution of the indirect length using input indirect length data as a random variable, and the determination step includes a process of adjusting the probability distribution of the indirect length by reflecting the direct length. The process of adjusting by reflecting the direct length refers to supplementing the probability distribution by inputting reliable data to correct errors in the learned probability distribution. That is, the probability distribution for the indirect length generated in the generation step possesses a certain representativeness but lacks reliability; therefore, the determination step performs a process of supplementing the probability distribution to approximate the probability distribution of the length of the carbon nanotube in the actual sample by utilizing the reliable direct length input into the correction model.

[0096] That is, the measurement method according to one embodiment of the present invention secures representativeness of the carbon nanotube length by including the generation step as described above, and at the same time secures reliability by including the determination step as described above, thereby enabling the determination of the carbon nanotube length with the length measurement error corrected.

[0097] In one embodiment of the present invention, the correction model is a Bayesian learning model.

[0098] The aforementioned Bayesian learning model is a learning model that performs probabilistic inference by updating the probability of a hypothesis based on Bayes' theorem using new information. By utilizing this Bayesian learning model, heterogeneous variables input into the model can be linked to derive results that complement each other. The aforementioned Bayesian learning model can be applied in a form commonly used in the relevant technical field to derive the carbon nanotube length corrected for measurement errors.

[0099] In one embodiment of the present invention, the Bayesian learning model derives the posterior probability of a probabilistic inference target through a prior probability and additional information regarding it. The prior probability is derived as a learning result based on the minimum data input to the learning model, and the additional information refers to the likelihood. The Bayesian learning model can derive an error-corrected posterior probability by updating the prior probability through the combination of the likelihood with the prior probability. When using such a Bayesian learning model, the final decision is made in a manner that complements the overall probability of the inference target; thus, even if outliers exist in the input data, the change in the final decision result is not significant. Furthermore, because it is based on prior probability, it is possible to quantify the uncertainty that occurs when using other machine learning models, thereby improving the accuracy of the decision result.

[0100] In one embodiment of the present invention, the determination step includes a step of deriving a posterior probability of the carbon nanotube length through a prior probability and a likelihood, wherein the prior probability is generated from an indirect length and the likelihood is generated from a direct length. As described above, by deriving a prior probability using data for an indirect length with certain representativeness and adjusting it by reflecting a likelihood generated from a direct length with guaranteed reliability, the representativeness and reliability of the finally determined carbon nanotube length can be simultaneously improved.

[0101]

[0102] carbon nanotubes

[0103] The above carbon nanotube monomers can be classified into single-walled carbon nanotubes (SWCNT), double-walled carbon nanotubes (DWCNT), and multi-walled carbon nanotubes (MWCNT) depending on the number of bonds forming the walls.

[0104] In one embodiment of the present invention, the carbon nanotube comprises one or more of single-walled, double-walled, and multi-walled carbon nanotubes.

[0105] In one embodiment of the present invention, the carbon nanotube comprises a single-walled carbon nanotube (SWCNT).

[0106] In the case of the aforementioned single-walled carbon nanotubes, since there is only one wall forming the unit, the length of the carbon nanotube has a more direct influence on the thickness, diameter, etc., and accordingly, the accuracy of carbon nanotube length measurement is considered more important.

[0107]

[0108] Carbon nanotube length measurement system

[0109] The present invention also provides a carbon nanotube length measuring system (100).

[0110] In one embodiment of the present invention, the carbon nanotube length measuring system (100) comprises an image data acquisition unit (110) for acquiring image data of a carbon nanotube, a physical property data acquisition unit (120) for acquiring a plurality of physical property data for a carbon nanotube, at least one processor (130) for executing a command to measure the length of a carbon nanotube using data acquired from the acquisition unit (110, 120), and a memory (140) for storing at least one command executed through the at least one processor (130).

[0111] The image data acquisition unit (110) is capable of obtaining image data (S111) of carbon nanotubes, meaning it can directly capture an image of a carbon nanotube or provide an image of a captured carbon nanotube. Specifically, the image data (S111) of carbon nanotubes refers to image data (S111) captured at a scale sufficient to confirm the skeleton of carbon nanotubes contained within a measurement sample. The image data (S111) can be acquired by a method generally known in the relevant technical field and can be captured from an atomic force microscope, a transmission electron microscope, a scanning electron microscope, etc., but is not limited thereto.

[0112] In one embodiment of the present invention, the image data acquisition unit (110) includes a scanning electron microscope.

[0113] In one embodiment of the present invention, the image data (S111) of the carbon nanotube is obtained from the scanning electron microscope.

[0114] The above-mentioned physical property data acquisition unit (120) is capable of obtaining physical property data (S121) of carbon nanotubes, meaning it can directly measure the physical properties of carbon nanotubes or provide the measured physical property data (S121) of carbon nanotubes. Specifically, the physical property data (S121) of carbon nanotubes refers to data that can generate the length of carbon nanotubes through a combination with other types of physical properties, such as mathematical relationships, even though the length of the carbon nanotubes has not been directly measured. The above-mentioned physical property data (S121) can be acquired by methods generally known in the relevant technical field, and can be acquired through an atomic force microscope, a rheometer, a transmission electron microscope, etc., but is not limited thereto. If multiple physical property data (S121) can be acquired in a single configuration, the above-mentioned physical property data acquisition unit (120) may include only one configuration.

[0115] In one embodiment of the present invention, the plurality of data are selected from the group consisting of thickness, aspect ratio, elongation viscosity, and combinations thereof.

[0116] In one embodiment of the present invention, the physical property data acquisition unit (120) includes an atomic force microscope.

[0117] In one embodiment of the present invention, the thickness of the carbon nanotube is obtained from the atomic force microscope.

[0118] In one embodiment of the present invention, the physical property data acquisition unit (120) includes a rheometer.

[0119] In one embodiment of the present invention, the elongation viscosity of the carbon nanotube is measured from a rheometer.

[0120] In one embodiment of the present invention, the system includes a memory (140), a processor (130), a transmission and reception system (150), an input interface system (160), an output interface system (170), and a storage system.

[0121] The components within the above system can be connected to each other to communicate with one another. Specifically, each component can be connected by a bus (190) to communicate with one another.

[0122] The memory (140) and storage system may be composed of at least one of a volatile storage medium and a non-volatile storage medium. For example, the memory (140) and storage device (180) may be composed of at least one of a read-only memory (ROM) and a random access memory (RAM).

[0123] The memory (140) includes at least one instruction executed by the processor (130).

[0124] The above processor (130) may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor (130) on which methods according to embodiments of the present invention are performed.

[0125] As previously described, the processor (130) can execute at least one program command stored in memory (140).

[0126] In one embodiment of the present invention, the at least one command comprises: a command (S110) for extracting a direct length from image data of the carbon nanotube; a command (S120) for generating an indirect length to correct an error occurring in the direct length by combining a plurality of physical property data for the carbon nanotube; a command (S130) for generating a correction model to correct a length measurement error based on the direct length and the indirect length; and a command (S140) for determining the length of the carbon nanotube with the length measurement error corrected using the correction model.

[0127] The above correction model is the same as the correction model described above, and the command (S110) for extracting the direct length is the same as the extraction of the direct length described above.

[0128] In one embodiment of the present invention, the direct length extraction command (S110) includes a command (S113) for extracting each carbon nanotube backbone from the image data (S111).

[0129] In one embodiment of the present invention, the extraction (S113) of the carbon nanotube framework is performed through the processing of image data (S111).

[0130] In one embodiment of the present invention, the extraction (S113) of the carbon nanotube backbone is performed using a backbone algorithm.

[0131] Specifically, the command (S110) for extracting the direct length includes a command (S113) for extracting each carbon nanotube skeleton from the image data (S111) of the carbon nanotube, and the extraction of the carbon nanotube skeleton (S113) can be performed through a skeletonization algorithm.

[0132] In one embodiment of the present invention, the direct length extraction command (S110) includes a command (S114) for calculating the circumference of each skeleton from the extracted carbon nanotube skeleton.

[0133] In one embodiment of the present invention, the calculation of the circumference of the skeleton is performed through the processing of image data (S111).

[0134] In one embodiment of the present invention, the calculation of the perimeter of the skeleton (S114) is performed through polygonal approximation.

[0135] Specifically, the command (S110) for extracting the direct length includes a command (S114) for calculating the circumference of each skeleton from the extracted carbon nanotube skeleton, and the calculation of the skeleton's circumference (S114) can be performed through polygonal approximation.

[0136] In one embodiment of the present invention, in one embodiment of the present invention, the command (S110) for extracting the direct length includes a command (S112) for recognizing each carbon nanotube from the image data (S111) of the carbon nanotube.

[0137] In one embodiment of the present invention, the recognition (S112) of the carbon nanotube is performed through the processing of image data (S111).

[0138] In one embodiment of the present invention, the recognition (S112) of the carbon nanotube is performed using a Segment Anything model.

[0139] Specifically, the command (S110) for extracting the direct length includes a command (S112) for recognizing each carbon nanotube from the image data (S111), and the recognition of the carbon nanotube (S112) can be performed using a Segment Anything model.

[0140] The command (S120) for generating the above indirect length is the same as the generation of the indirect length described above.

[0141] In one embodiment of the present invention, the command (S120) for generating the indirect length includes a command for generating the indirect length by combining the thickness and aspect ratio of the carbon nanotube.

[0142] In one embodiment of the present invention, the command (S120) for generating the indirect length includes a command for deriving the aspect ratio from the elongation viscosity of the carbon nanotube. The derivation of the aspect ratio may be achieved through a formula or theoretical model representing the relationship between the elongation viscosity and the aspect ratio. Specifically, the derivation of the aspect ratio from the carbon nanotube elongation viscosity may be performed using a relational equation representing the theoretical relationship between the elongation viscosity and the aspect ratio in the rigid rod solution.

[0143] The command for extracting the direct length (S110) and the command for generating the indirect length (S120) can be executed in parallel.

[0144] The command (S130) for generating the above correction model is the same as the generation step described above, and the command (S140) for determining the length of the carbon nanotube is the same as the determination step described above.

[0145] In one embodiment of the present invention, the command (S130) for generating the correction model includes a command to learn the probability distribution of the indirect length using the input indirect length data as a probability variable.

[0146] In one embodiment of the present invention, the command (S130) for generating the correction model includes a command to learn the probability distribution of the indirect length using the input indirect length data as a probability variable.

[0147] In one embodiment of the present invention, a command (S140) for determining the length of the carbon nanotube includes a command for adjusting the probability distribution of the indirect length to reflect the direct length.

[0148] In one embodiment of the present invention, a command (S140) for determining the length of the carbon nanotube includes a command for deriving a posterior probability of the carbon nanotube length through a prior probability and a likelihood, wherein the prior probability is generated from an indirect length and the likelihood is generated from a direct length.

[0149] Specific embodiments of the present invention are presented below. However, the embodiments described below are merely for the purpose of specifically illustrating or explaining the present invention and do not limit the present invention. Furthermore, details not described herein can be sufficiently technically inferred by a person skilled in the art, so their description is omitted.

[0150]

[0151] Examples

[0152]

[0153] <Example 1: Indirect Length of Carbon Nanotubes>

[0154] (1) Aspect ratio of carbon nanotubes

[0155] Different single-walled carbon nanotube dispersions of samples A to C were prepared, and the dispersions were deformed at 25°C using a capillary breakage expansion rheometer (CaBER1, Thermo Fischer) to a state where the Hencky Strain was 0.76. The degree of diameter reduction over time was observed to measure the extensional viscosity, and the aspect ratio of the carbon nanotubes was calculated from the extensional viscosity. The calculation of the carbon nanotube aspect ratio was performed by referring to the method of non-patent literature 1 (Macromolecules 2016, 49, 2, 681-689).

[0156] (2) Indirect length of carbon nanotubes

[0157] A few drops of the above carbon nanotube dispersion were dropped onto a silicon wafer (Si wafer) and dried to prepare a carbon nanotube dispersion for measurement, and the thickness of the carbon nanotube was measured using an atomic force microscope (AFM; NX10, manufactured by Park System), and the indirect length of the carbon nanotube was derived by multiplying the aspect ratio and thickness of the carbon nanotube calculated above.

[0158]

[0159] <Example 2: Direct Length of Carbon Nanotubes>

[0160] Images of the above carbon nanotube dispersion samples were captured using a scanning electron microscope (SEM; FESEM-09-JSM7610F, manufactured by JEOL). Each carbon nanotube skeleton was extracted from the images, and the circumference of each skeleton was calculated from the extracted skeletons. Subsequently, the direct length of the carbon nanotube was derived as half the value of the circumference.

[0161] The extraction of the carbon nanotube skeleton was performed using a skeletonize algorithm, and the extraction results are shown in [Fig. 1]. The numbers 0 to 13 shown in [Fig. 1] represent the numbering of each carbon nanotube extracted within the image. The perimeter of the extracted skeleton was calculated using polygon approximation.

[0162]

[0163] <Example 3: Determination of Carbon Nanotube Length>

[0164] For each carbon nanotube sample, 16 aspect ratios and 5 thicknesses were measured, and 80 indirect lengths were derived from this. Then, the indirect length values ​​were input into a Bayesian learning model to learn the probability distribution of the indirect lengths. Subsequently, the probability distribution of the indirect lengths of the learning model was adjusted to reflect the direct lengths, using the indirect lengths as the prior probability and the direct lengths of each sample as the likelihood. Through this, the posterior probability of the carbon nanotube length was calculated, and the length of the carbon nanotube corrected for the length measurement error was determined.

[0165] The average value for the above indirect length, the direct length set as the likelihood value, and the length of the carbon nanotube corresponding to the median value of the posterior probability corrected for the length measurement error through this are shown in [Table 1] below.

[0166]

[0167] Indirect Length (㎛) Direct Length (㎛) Median Posterior Probability (㎛) Sample A 2.336 2.28 2.344 Sample B 2.643 1.56 2.218 Sample C 2.203 1.39 1.867

[0168]

[0169] Although preferred embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept of the present invention as defined in the following claims also fall within the scope of the present invention.

[0170]

[0171] [Explanation of the symbol]

[0172] 100: Carbon nanotube length measurement system

[0173] 110: Image data acquisition unit

[0174] 120: Material Property Data Acquisition Unit

[0175] 130: Processor

[0176] 140: Memory

[0177] 150: Transmitting and receiving system

[0178] 160: Input Interface System

[0179] 170: Output Interface System

[0180] 180: Storage device

[0181] 190: Bus

[0182] S100: Instructions executed through the processor of the carbon nanotube length measurement system

[0183] S110: Command to directly extract length from image data of carbon nanotubes

[0184] S111: Image data of carbon nanotubes

[0185] S112: Command to recognize each carbon nanotube from image data of carbon nanotubes

[0186] S113: Command to extract each carbon nanotube backbone from carbon nanotube image data

[0187] S114: Command to calculate the circumference of each skeleton from the extracted carbon nanotube skeleton

[0188] S120: Command to generate an indirect length to correct an error occurring in the direct length by combining multiple material property data for carbon nanotubes.

[0189] S121: Multiple physical property data for carbon nanotubes

[0190] S130: Command to generate a correction model for correcting length measurement errors based on direct length and indirect length

[0191] S140: Command to determine the length of a carbon nanotube with the length measurement error corrected using a correction model

Claims

1. As a system for measuring the length of carbon nanotubes, Image data acquisition unit for acquiring image data of carbon nanotubes; A property data acquisition unit for acquiring multiple property data for carbon nanotubes; At least one processor for executing a command to measure the length of a carbon nanotube using data acquired from the above-mentioned acquisition unit; and It includes a memory that stores at least one instruction executed through the above-mentioned at least one processor, and The above at least one command is, A command to directly extract the length from the image data of the carbon nanotube above; A command to generate an indirect length for correcting an error occurring in the direct length by combining multiple physical property data for the carbon nanotube; A command to generate a correction model for correcting length measurement errors based on the above direct length and indirect length; and Instruction to determine the length of a carbon nanotube with the length measurement error corrected using the above correction model; including, Carbon nanotube length measurement system.

2. In Claim 1, The above correction model is a Bayesian learning model, Carbon nanotube length measurement system.

3. In Claim 1, The above correction model is a model capable of learning the probability distribution between random variables using input data as a random variable. Carbon nanotube length measurement system.

4. In Claim 1, The command for generating the above correction model includes a command for learning the probability distribution of the indirect length using the input indirect length data as a random variable, Carbon nanotube length measurement system.

5. In Claim 1, The command for generating the above correction model includes a command for learning the probability distribution of the indirect length using the input indirect length data as a random variable, and The command for determining the length of the carbon nanotube includes a command for adjusting the probability distribution of the indirect length to reflect the direct length. Carbon nanotube length measurement system.

6. In Claim 1, The command for determining the length of the carbon nanotube includes a command for deriving the posterior probability of the carbon nanotube length through prior probability and likelihood, and The above prior probability is generated from the indirect length, and The above possibility is generated from the direct length, Carbon nanotube length measurement system.

7. In Claim 1, The above plurality of physical property data includes those selected from a group consisting of thickness, aspect ratio, extensional viscosity, and combinations thereof. Carbon nanotube length measurement system.

8. In Claim 1, The above indirect length is generated from a combination of the thickness and aspect ratio of the carbon nanotube, and The above aspect ratio is derived from the elongation viscosity of carbon nanotubes, Carbon nanotube length measurement system.

9. In Claim 8, The above material property data acquisition unit includes an atomic force microscope (AFM), and The thickness of the carbon nanotube above is obtained from the atomic force microscope, Carbon nanotube length measurement system.

10. In Claim 1, The above direct length extraction command includes a command to extract each carbon nanotube backbone from the image data, Carbon nanotube length measurement system.

11. In Claim 10, The above direct length extraction command includes a command to calculate the circumference of each skeleton from the extracted carbon nanotube skeleton, Carbon nanotube length measurement system.

12. In Claim 11, The calculation of the perimeter of the above skeleton is performed through polygon approximation, Carbon nanotube length measurement system.

13. In Claim 1, The above image data acquisition unit includes a scanning electron microscope (SEM), and The image data of the carbon nanotube above is obtained from the scanning electron microscope, Carbon nanotube length measurement system.

14. In Claim 1, The above carbon nanotubes include single-walled carbon nanotubes (SWCNT). Carbon nanotube length measurement system.

Images