JP7209287B2 - How to identify nanoparticles - Google Patents

Description

TECHNICAL FIELD The present invention relates to a method for discriminating nanoparticles, and more particularly to a method for specifically discriminating nanoparticles under dark field.

When gold nanoparticles are observed under a dark field microscope, they can be observed as bright bright spots due to light scattering due to strong surface plasmon resonance. A number of molecular detection methods have been proposed that utilize this strong light scattering under the dark field.

For example, Non-Patent Document 1 discloses that a DNA measurement system was constructed using DNA-modified gold nanoparticles and magnetic microparticles. In this measurement system, DNA to be measured is first added to the presence of DNA-modified gold colloids, streptavidin-modified magnetic particles, and biotin-modified probes to form complexes of magnetic particles and gold nanoparticles via the DNA to be measured. Let Next, magnetism is collected using a magnetic force, and non-adsorbed components are washed. After that, by dissociating the DNA association, the complexes of the magnetic fine particles and the gold nanoparticles are separated, and the gold nanoparticles are obtained in the supernatant. Since the number of gold nanoparticles present in the supernatant correlates with the target DNA concentration, the target DNA concentration is measured by counting the bright spots of the particles under dark-field microscope observation.

In addition, Non-Patent Document 2 shows detection of amyloid fibrils using anti-amyloid peptide antibody-modified gold nanoparticles. In this measurement system, gold nanoparticles bound to amyloid monomers (present as monomers) and gold nanoparticles bound to amyloid fibrils (present as aggregates) are imaged by dark field microscopy. Amyloid fibrils are detected with high sensitivity by calculating the average pixel value of the spots in the obtained observation image as brightness, creating a histogram, and evaluating nanoparticle aggregation.

These molecular detection methods are based on the strong scattered light of gold nanoparticles. However, when impurities are present in the field of view of the dark-field microscope, they may be observed as bright spots similar to gold nanoparticles, which may affect the measurement. In contrast, Patent Literature 1 discloses a technique for distinguishing between impurities and gold nanoparticles by utilizing the fact that negatively charged gold nanoparticles do not adhere to a glass substrate and undergo irregular Brownian motion in liquid. It is Specifically, when the exposure time during image data acquisition is lengthened, the brightness increases in proportion to the exposure time because the impurities are immobilized on the glass substrate, but the gold nanoparticles are not immobilized. , the image data is obtained as the trajectory of the Brownian motion.

JP 2009-229103 A

Anal. Chem. , 2016, 88, 4188 Analytical. Sciences. , 2016, 32, 307

However, although the method described in Patent Document 1 is a useful method for distinguishing between nanoparticles and impurities under a dark field, it is difficult to apply to observation of nanoparticles immobilized on a substrate. For example, in the methods shown in Non-Patent Documents 1 and 2, amino group-modified glass is used to immobilize nanoparticles in order to increase the brightness of the particles and obtain clear image data. Therefore, the method described in Patent Document 1 cannot be used for these methods.

Accordingly, the present invention has been made in view of the above problems, and an object of the present invention is to provide a novel nanoparticle discrimination method that can also be applied to observation of immobilized nanoparticles.

In order to solve the above problems, the method for distinguishing nanoparticles according to the present invention is a method for distinguishing nanoparticles that can be observed as bright spots under dark-field observation, wherein a sample containing nanoparticles is observed under dark-field observation. observing to obtain a color observation image; color-separating the observation image to obtain color information of luminescent spots including information on each color component; and obtaining luminescent spots in the observation image from the color information. It includes a step of determining whether or not it is a nanoparticle bright spot.

Further, in one aspect of the method for distinguishing nanoparticles according to the present invention, the sample contains two or more types of nanoparticles that produce bright spots of different colors.

In one aspect of the method for distinguishing nanoparticles according to the present invention, in the step of obtaining the observation image, the nanoparticles in the sample are immobilized on a substrate and observed.

In one aspect of the method for distinguishing nanoparticles according to the present invention, in the step of obtaining the observation image, the nanoparticles in the sample are observed without being immobilized on a substrate.

In one aspect of the method for determining nanoparticles according to the present invention, the sample contains a gold nanoparticle monomer as the nanoparticles.

In one aspect of the method for identifying nanoparticles according to the present invention, in the identifying step, it is determined whether the bright spots in the observed image are the nanoparticles, aggregates of the nanoparticles, or impurities. discriminate.

In one aspect of the method for distinguishing nanoparticles according to the present invention, the color separation of the observation image is performed by color-separating the observation image into RGB.

In one aspect of the method for distinguishing nanoparticles according to the present invention, in the distinguishing step, whether or not the bright spots in the observed image are bright spots of nanoparticles is determined based on pre-obtained discrimination information. discriminate.

Further, in one aspect of the method for discriminating nanoparticles according to the present invention, the discriminating information is the bright spot classification model obtained by machine-learning the correlation between the color information and the attribution of the bright spots.

According to the method for distinguishing nanoparticles of the present invention, since the attribution of bright spots is judged based on color information, it can also be applied to a sample in which nanoparticles are immobilized on a substrate.

It is the figure which showed the scheme until it creates a three-dimensional plot in the discrimination|determination method which concerns on embodiment of this invention. FIG. 11 is a diagram showing RGB three-dimensional plots in Example 4; FIG. 11 is a diagram showing RGB three-dimensional plots in Example 7; FIG. 10 is a diagram showing the particle size distribution by DLS in (3) of Example 8; FIG. 22 is a diagram showing RGB three-dimensional plots in (5) of Example 8; FIG. 20 is a diagram showing RGB three-dimensional plots in (6) of Example 8;

The method for distinguishing nanoparticles according to the present invention includes the steps of observing a sample containing nanoparticles under dark field observation, obtaining a color observation image, color-separating the obtained observation image, and obtaining color information of bright spots. It includes a step of obtaining, and a step of determining whether or not the bright spots are the bright spots of the nanoparticles from the obtained color information. An embodiment of the method for distinguishing nanoparticles according to the present invention will be described below.

[Nanoparticles]
The nanoparticles to be identified in this embodiment are particles that can be observed as bright spots under dark-field observation. The nanoparticles applied to the present embodiment are not particularly limited as long as they can be observed as bright spots under dark field observation. Examples include gold nanoparticles, silver nanoparticles, gold nanorods, gold nanoartine, and gold nanoparticle aggregates. Aggregates of homogeneous particles such as silver nanoparticles-gold nanoarchine dimers, aggregates of heterogeneous particles such as silver nanoparticles-gold nanoarchine dimers, and quantum dots.

Also, the particle size of the nanoparticles is not particularly limited, but the particle size is preferably 10 nm to 1 μm, more preferably 20 nm to 100 nm. In the present specification, the particle diameter is the diameter in the case of spherical particles, and the length in the long axis direction in the case of rod-like particles such as nanorods.

The material of the nanoparticles is not particularly limited as long as the nanoparticles can be observed as characteristic bright spots under dark field. However, from the viewpoint of brightness, metals having plasmon resonance characteristics in the visible to near-infrared wavelength band are preferred.

Further, in the method described in Patent Document 1 above, when a plurality of nanoparticles such as gold nanoparticle monomers and gold nanoparticle aggregates, or gold nanoparticles and silver nanoparticles are present in the same solution, the discrimination is Have difficulty. However, according to the determination method of the present embodiment, by using particles with different colors of generated bright spots, two or more different particles can be distinguished and determined under dark-field observation. That is, when distinguishing two or more different particles under dark-field observation, it is preferable to use particles that produce bright spots of different colors. For example, a gold nanoparticle monomer (particle size 40 nm, green bright spots under dark field observation) as the first nanoparticles and a gold nanoparticle aggregate (yellow bright spots under dark field observation) as the second nanoparticles. The combination used, or the gold nanoparticle monomer as the first nanoparticle (particle size 40 nm, green bright spot under dark field observation) and the silver nanoparticle monomer as the second nanoparticle (particle size 60 nm, under dark field observation A combination using a blue bright spot) can be exemplified.

The discrimination method of this embodiment distinguishes bright spots of nanoparticles from bright spots of impurities. The term "impurity" as used herein refers to a substance observed during dark-field observation for purposes other than the purpose, and is not particularly limited. Impurities include, for example, glass powder and dust that exist on the glass substrate and cause bright spots under dark field observation. In addition, in the case where a specimen sample containing nanoparticles is to be measured, biological substances that produce bright spots during dark-field observation are examples of impurities.

[Dark field observation]
The dark-field observation in the present embodiment may be a conventionally known dark-field observation method in which light is obliquely applied to the sample and only scattered light and reflected light from the sample are observed. The illumination method may be either a transmissive type or an epi-illumination type, and may be appropriately selected by the user.

In order to image-analyze the bright spots generated under dark-field observation, the observed image is obtained as a color image. There are no particular restrictions on the imaging device for acquiring the observed image as long as it can capture a color image. When observing nanoparticles in a sample without immobilizing them, it is preferable to use an imaging device with a high shutter speed.

For dark-field observation of nanoparticles, a liquid sample containing nanoparticles is spread on a substrate. Examples of substrates include slide glass and glass petri dishes. The material of the substrate is not particularly limited as long as the liquid sample can be developed, and examples thereof include soda glass, quartz glass, acrylic resin and polycarbonate.

In dark-field observation in this embodiment, nanoparticles may be immobilized on a substrate. By immobilizing the nanoparticles on the substrate, the sharpness of the acquired observation image can be improved. As a result, it is possible to improve the accuracy of the discrimination method of this embodiment using the observation image. A method for immobilizing nanoparticles on a substrate includes, but is not limited to, a technique utilizing interaction between nanoparticles and the substrate surface. For example, when gold nanoparticles are used as the nanoparticles, the gold nanoparticles are immobilized on the substrate by electrostatic interaction with the negatively charged gold nanoparticles by modifying the substrate surface with amino groups. be able to. Alternatively, the nanoparticles can be immobilized on the substrate by increasing the salt concentration in the sample containing the nanoparticles and using a slide glass as the substrate.

As described above, according to the discrimination method of the present embodiment, nanoparticles can be discriminated even in a sample in which nanoparticles are immobilized on a substrate. However, immobilization of nanoparticles is not an essential step. When the nanoparticles are not immobilized on the substrate, it is preferable to obtain observation images using a camera (imaging device) with a high shutter speed in order to suppress the influence of the Brownian motion of the nanoparticles. . In this specification, "fast shutter speed" refers to a case where an image is output quickly due to a short exposure time, as well as a process of acquiring multiple images and synthesizing them (for example, pixel shift multiplexing). It is also intended that the image can be output more quickly because it does not perform shooting). Alternatively, Brownian motion of the nanoparticles may be physically suppressed by adding a thickening agent to the sample containing the nanoparticles. Alternatively, a combination of using a camera with a fast shutter speed and adding a thickening agent to the sample may be used.

A thickener that does not interfere with observation may be appropriately selected according to the target sample. Examples of thickening agents include carboxymethylcellulose, chitosan and pectin.

[Analysis of bright spots]
Next, a method for analyzing bright spots using an acquired observation image will be described. FIG. 1 is a diagram showing a scheme in one aspect of the bright spot analysis method described below.

(color separation of bright spots)
In the determination method of the present embodiment, a color observation image obtained by dark-field observation is color-separated into a plurality of color components, color information of bright spots including information of each color component is obtained, and based on the color information, do the analysis. In the following, an example will be described in which a color observation image obtained by dark-field observation is color-separated into RGB, and analysis is performed based on the RGB values of each bright spot.

First, a color observation image obtained by dark-field observation is image-converted into an 8-bit grayscale image. Bright spots are identified in the grayscale image and the regions identified as bright spots are determined as regions of interest (ROI). For identification of bright spots, for example, when image analysis software (ImageJ) shown in Examples below is used, the conditions of threshold value 30-255, size 30-10000, and circularity 0.3-1.0 are used. What satisfies the condition may be identified as a bright spot at the stage of image processing. Next, the observed color image of interest is divided into R, G, and B component images, and intensity measurements are performed on all ROIs for each color component image. The intensity of each color component (R value, G value and B value) in the ROI is used as color information of the bright spot. That is, the color information of the bright spots in this embodiment is trivariate data consisting of RGB values.

(Determination of bright spots)
As one mode for discriminating the bright spots based on the color information, the obtained color information of the bright spots is three-dimensionally plotted as trivariate data of RGB values. By three-dimensional plotting, it becomes possible to visually recognize the difference in bright spots due to different attributions. In the present specification, the attribution means whether the bright spot is derived from any of nanoparticles, nanoparticle aggregates, and impurities, and when there are multiple types of nanoparticles, which nanoparticle it is derived from. It indicates whether or not

For example, when the nanoparticles are gold nanoparticle monomers, the color of the bright spots is green, and the G value is the largest when color-separated into RGB. That is, bright spots plotted at positions where the value of G is larger than that of R and B can be recognized as bright spots of the gold nanoparticle monomer. On the other hand, bright spots derived from impurities are white, and when color-separated into RGB, the values of RGB are almost the same. That is, it can be recognized that the bright spots plotted at positions where the values of RGB are almost the same are impurities. By performing three-dimensional plotting of RGB in this way, the nanoparticles can be visually distinguished.

Three-dimensional plotting makes it possible to distinguish nanoparticles visually. It is preferable to discriminate Here, the discriminant information is a reference, model, algorithm, program, etc. derived by collecting color information of bright spots with clear attribution and using the collected color information, and is obtained by observation. By applying the discriminant information to the color information of the bright spot, the attribution of the bright spot is concluded.

(machine learning method)
Even if the nanoparticles are of the same kind, there are variations in the RGB values of the bright spots of each nanoparticle. In addition, even if the nanoparticles or impurities are of different types, the desired nanoparticles and bright spots may have similar colors. Therefore, there is a limit to human visual discrimination in order to perform discrimination with high accuracy. For this reason, it is preferable to use software (computer program) that outputs determination results of attribution using RGB values as input data as the above-described discrimination information in determining attribution. In particular, from the viewpoint of determination accuracy, it is desirable to use software (classification model) that automatically determines attribution based on machine learning results as the above-described determination information. The software (algorithm) that can be used for discrimination is not particularly limited, but for example, a neural network that has undergone supervised learning, or a support vector machine that has undergone supervised learning can also be used.

For example, it is possible to cause the software to output the result of discrimination between gold nanoparticle monomers, gold nanoparticle aggregates, and impurities from input data that are RGB values. In this case, the color information of the gold nanoparticle monomer is obtained in advance using the gold nanoparticle monomer, the color information of the gold nanoparticle aggregate is obtained using the gold nanoparticle aggregate, and impurities (for example, biological samples and (soil sample, etc.) to obtain the color information of impurities. Then, machine learning of the software is performed using the color information with known attribution obtained in this way as learning data, and the accuracy of machine learning is confirmed using the color information with known attribution as test data. This process is repeated. This enables the software to perform determination with desired determination accuracy.

An example of such software is the statistical analysis free software R (R DEVELOPMENT CORE TEAM (2005). R: A LANGUAGE
AND environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www. R-project. org. ).

By adopting a discrimination method using machine learning, it becomes possible to discriminate the attribution of nanoparticles with higher accuracy.

As described above, when nanoparticles are observed under a dark field, the presence of coexisting impurities may make it difficult to identify the nanoparticles. According to the determination method of the present embodiment, by adopting the approach of color-separating a color observation image, it is possible to apply the method to an observation target, which has conventionally been difficult to apply. In addition, a technique using a spectroscope is known for analyzing nanoparticles in dark field observation. Nanoparticles can be easily distinguished without using expensive equipment such as In addition, since the nanoparticles can be easily distinguished, it is possible to focus only on the nanoparticles by distinguishing the bright spots of the nanoparticles from the bright spots of the impurities. can be raised. Furthermore, in the discrimination method of this embodiment, it is possible to automate all the analysis operations. Therefore, it is possible to greatly shorten the required time compared to manual observation and a method using a spectroscope.

Furthermore, according to the discrimination method of the present embodiment, the first nanoparticles and the second nanoparticles, which are different from the first nanoparticles, can be discriminated by using the difference in color of the bright spots observed under the dark field. becomes possible. In addition, an automatic detection system for nanoparticles by allowing predetermined software to perform machine learning of the correlation between the RGB value of each bright point and the attribution of the bright point (e.g., gold nanoparticle monomer, gold nanoparticle aggregate, and impurity) can also be designed. This makes it possible to more quickly and accurately detect nanoparticles even in samples containing impurities under dark-field observation.

EXAMPLES The embodiments of the present invention will be described in more detail below with reference to examples. Of course, the present invention is not limited to the following examples, and it goes without saying that various aspects are possible for the details. Furthermore, the present invention is not limited to the above-described embodiments, and can be modified in various ways within the scope of the claims. It is included in the technical scope of the invention. In addition, all documents described in this specification are incorporated by reference.

[Example 1. Preparation of Amino Group-Introduced Slide Glass]
A slide glass (manufactured by Matsunami Glass Co., Ltd., S1111) was washed with ultrapure water and 70% ethanol and dried. Next, 1% APTES ((3-Aminopropyl)triethoxysilane, diluted with ultrapure water) was applied to the slide glass, allowed to stand at room temperature for 30 minutes, and then washed with ultrapure water. This was dried to obtain an amino group-introduced slide glass.

[Example 2. Dark-field imaging of gold nanoparticle monomers, gold nanoparticle aggregates, and impurities]
(1) Imaging of Gold Nanoparticle Monomer A sample was prepared by diluting 40 nm gold nanoparticles (manufactured by BBI, 150 pM) with ultrapure water to 5 pM. 3.5 μL of the sample was dropped onto the amino group-introduced slide glass prepared in Example 1, covered with a cover glass, and sealed by applying a top coat to the four sides of the cover glass to prepare a slide.

The prepared slide was observed under dark field by oil immersion. A microscope (BX53, manufactured by OLYMPUS), a CCD camera (color, DP73, manufactured by OLYMPUS), and an objective lens (Uplan FLN60×, manufactured by OLYMPUS) were used as observation devices. Observation conditions were an exposure time of 200 ms, ISO 200, and no optical filter. Focusing was performed under the above conditions, and six observation images were taken at different observation positions (4800×3600, pixel multi-photographing).

(2) Imaging of Aggregate of Gold Nanoparticles Gold nanoparticles of 40 nm (manufactured by BBI, 150 pM) were diluted to 10 pM with ultrapure water. A preparation was prepared using the method described in (1) above, except that the same amount of phosphate-buffered saline (PBS) was added to this and allowed to stand for 10 minutes, and the observation image was taken. did

(3) Imaging of impurities Except for using a soil sample or serum (100-fold dilution, unfiltered) treated with a filter (MILLEX-GV 0.22 μm Filter Unit, Merck Millipore) as a sample, the above (1) The preparation was observed by the method described in 1., and an observation image was taken.

[Example 3. Image processing〕
Using image analysis software ImageJ (Biophotonics International, vol 11, issue 7, pp. 36-42, 2004), image processing was performed by the following method.

(1) Grayscale Conversion, Bright Point Analysis, and ROI Storage The target image was converted to grayscale, and bright point analysis and bright point area information (ROI) acquisition were performed. Bright spot analysis parameters were threshold 30-255, size 30-10000, and roundness 0.3-1.0.

(2) Division into RGB and Bright Point Analysis in ROI in Each Image The target image composed of RGB was divided into three images, an R component image, a G component image and a B component image. . Bright spot analysis was performed on each image using the ROI obtained in (1) above, and the intensity of the ROI in each of the R, G, and B images was obtained. The obtained intensity of each ROI was used as the intensity of each of R, G, and B of each bright spot. This was used as subsequent analysis data.

[Example 4. Data processing using R]
The analysis data obtained in Example 3 was processed using statistical analysis free software R (R DEVELOPMENT CORE TEAM (2005). R: A LANGUAGE AND
environment for statistical computing.
R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www. R-project. org. ) was used.

Specifically, the analysis data was output as a text file, and a table in which each bright point and its RGB value were associated was created. By three-dimensionally plotting the RGB values of each bright spot, the distribution of the RGB values of the bright spots was visualized. The analytical schemes of Examples 3 and 4 are as shown in FIG. The results of observing the three-dimensional plot of the visualized RGB values from different angles are shown in (a) to (c) of FIG. In FIG. 2, gold nanoparticle monomers, gold nanoparticle aggregates, and impurities are distinguished by different shades of plots. The plot with the lightest gradation is the plot of the bright spots of the gold nanoparticle monomer, and the plot with the darkest gradation is the plot of the bright spots of the impurities (especially see (c) in FIG. 2). In the three-dimensional plot of FIG. 2, the axes corresponding to the R value, G value and B value are indicated by arrows with signs in the figure.

As is clear from (a) to (c) of FIG. 2, bright spots of gold nanoparticle monomers, gold nanoparticle aggregates, and impurities showed different distributions in the three-dimensional plot of RGB values. Therefore, it was confirmed that gold nanoparticle monomers, gold nanoparticle aggregates, and impurities can be distinguished by color separation of bright spots of gold nanoparticle monomers, gold nanoparticle aggregates, and impurities, respectively.

[Example 5. Attribution of bright spots using machine learning]
Assignment is facilitated by providing criteria for determining which nanoparticle or impurity an observed bright spot corresponds to. However, when the plots of the bright spots cannot be linearly separable as shown in the three-dimensional plot of FIG. 2, it is difficult to manually define the criteria. Therefore, a method for discriminating the attribution of bright spots using machine learning was adopted.

The RGB value information of the bright spots obtained in Examples 2 and 3, which is clearly attributed to gold nanoparticle monomers, gold nanoparticle aggregates, and impurities, is combined with the attribution to provide a training data set for machine learning. and This data set was divided into two by random numbers, one of which was used as learning data and the other as test data. A classification model for determining the attribution of bright spots by machine learning was created using R, a statistical analysis free software. In machine learning using any divided learning data, it was possible to classify the test data into gold nanoparticle monomers, gold nanoparticle aggregates, or impurities, and the correct answer rate was 98.8%. .

[Example 6. Evaluation of error between measurements]
In order to evaluate the influence of deviations and variations in RGB values between measurements on the attribution results of machine learning, the same sample was measured multiple times to confirm measurement errors.

Using a gold nanoparticle monomer sample, a preparation was prepared according to the method described in Example 2, and an image was taken under a dark field microscope. At this time, while the stage position (XY direction) was fixed, only the objective lens and the condenser lens (Z direction) were moved, and the process of performing photographing was repeated six times. Specifically, after taking an image, the objective lens and the condenser lens were moved to remove the focus, the focus was adjusted again, and the image was taken again. Determination of the assignment of nanoparticles was performed by R, a machine-learned statistical analysis free software, using all the data of the training set used in Example 5 as learning data. Table 1 shows the analysis results of the six image data.

The standard deviation of the aggregation rate when measured 6 times was 0.000437. These results suggest that the errors between measurements are small, and that deviations and variations between measurements have little effect on the results of machine learning.

From the above results, it was confirmed that nanoparticles can be distinguished by color separation according to the method of the present invention.

[Example 7. Observation and assignment of suspended gold nanoparticles]
In order to observe floating gold nanoparticles, carboxymethyl cellulose (CMC) was used as a thickener to suppress the movement of the gold nanoparticles and the images were taken at a high shutter speed.

(1) Imaging of gold nanoparticle monomer 40 nm gold nanoparticles (manufactured by BBI, 150 pM) were diluted with ultrapure water to 10 pM, and the same amount of 1.0% CMC was added thereto to prepare a sample. 3.5 μL of the sample is dropped onto a slide glass (manufactured by Matsunami Glass Co., Ltd., S1111, amino group not introduced), covered with a cover glass from above, and sealed by applying a top coat to the four sides of the cover glass to form a preparation of the monomer sample. was made.

The prepared slide was observed under dark field by oil immersion. A microscope (BX53, manufactured by OLYMPUS), a CCD camera (color, DP73, manufactured by OLYMPUS), and an objective lens (UplanFLN60×, manufactured by OLYMPUS) were used as observation equipment, and the observation conditions were an exposure time of 200 ms, ISO 200, and no optical filter. I went with Focusing was performed under the above conditions, and six observation images were photographed (1600×1200, normal photography).

(2) Imaging of Aggregate of Gold Nanoparticles Gold nanoparticles of 40 nm (manufactured by BBI, 150 pM) were diluted with ultrapure water to 20 pM, and the same amount of PBS was added thereto and allowed to stand for 10 minutes. Samples were then prepared by adding 1.0% CMC in a 1:1 volume. Drop 3.5 μL of the sample onto a slide glass (manufactured by Matsunami Glass Co., Ltd., S1111, amino group not introduced), cover it with a cover glass, and apply a top coat to the four sides of the cover glass to seal it. A slide was prepared.

Observation of the prepared preparation was performed by the method described in (1) above.

(3) Photographing Impurity Samples Samples were prepared by mixing soil samples (filtered) 1:1 by volume with 1.0% CMC. 3.5 μL of the sample was dropped onto a slide glass (manufactured by Matsunami Glass Co., Ltd., S1111, amino group not introduced), covered with a cover glass from above, and sealed by applying a top coat to the four sides of the cover glass to form a preparation of the impurity sample. was made.

Observation of the prepared preparation was performed by the method described in (1) above.

(4) Bright spot analysis The images obtained by the above (1) to (3) were subjected to data processing using the methods described in Examples 3 and 4, and three-dimensional plotting of RGB values was performed. . The parameters for bright spot analysis by ImageJ were threshold 30-255, size 5-200, and circularity 0.3-1.0. Also, the results of observing the three-dimensional plot of the visualized RGB values from different angles are shown in (a) to (c) of FIG. In FIG. 3 , gold nanoparticle monomers, gold nanoparticle aggregates, and impurities are shown separately as plots with different shades similar to those in Example 4. As is clear from (a) to (c) of FIG. 3, even when imaging was performed without immobilizing the gold nanoparticles, the gold nanoparticle monomer, the gold It was confirmed that the nanoparticle aggregates and bright spots of the impurities showed different distributions in the three-dimensional plot of RGB values.

For determination of bright spot attribution by machine learning, the RGB information of the bright spots whose attribution is clear in (1) to (3) was combined with the attribution and used as learning data. As a result of evaluating the accuracy of the learning data using leave-one-out cross-validation, the average correct answer rate was 99.59%. As described above, it was confirmed that nanoparticles can be identified by one aspect of the method of the present invention even in observation of floating gold nanoparticles that are not immobilized on a substrate.

[Example 8. Discrimination of Silver Nanoparticles, Gold Nanoartin and Their Dimers]
Silver nanoparticles, gold nanoartines, and silver nanoparticle-gold nanoartine dimers were produced by the following method, and their discrimination was performed.

(1) Preparation of anti-BSA rabbit antibody-modified silver nanoparticles 50 µL of 40 nm silver nanoparticles (manufactured by BBI) and 1 µL of standard BSA (manufactured by Thermo Fisher) adjusted to 0.125 mg/mL are mixed and incubated at room temperature for 30 minutes. Let stand for 1 minute. An anti-BSA rabbit polyclonal antibody (manufactured by Molecular Probes, 20 µg/mL, 50 µL) was added to this mixed solution, and the mixture was allowed to stand at room temperature for another 2 hours. After that, 1% BSA solution (100 μL) and ultrapure water (800 μL) were added and mixed. The supernatant and nanoparticles are separated by centrifugation at 8000 rpm for 10 minutes, and the collected nanoparticles are dispersed in 50 μL of 0.1×PBS buffer to obtain an anti-BSA antibody-modified silver nanoparticle solution. rice field.

(2) Preparation of anti-rabbit antibody-modified gold nanoartin 2 μL of anti-rabbit polyclonal antibody (manufactured by R&D system) was diluted with 18 μL of 50 mM sodium phosphate buffer (pH 8.0) and added to 200 μL of 80 nm gold nanoartin (manufactured by Cytodiagnostics). and left at room temperature for 30 minutes. After that, 80 μL of 1% BSA solution was added, stirred, and allowed to stand for 15 minutes. This mixed solution was centrifuged (8000 rpm for 10 minutes) to separate the supernatant from nanoartin to obtain anti-rabbit antibody-modified gold nanoartin. The collected modified gold nanoartin was dispersed in 500 μL of ultrapure water, and the same centrifugation operation was performed twice to remove excess anti-rabbit polyclonal antibody. An anti-rabbit antibody-modified gold nanoartin solution was obtained by dispersing the obtained nanoartin in 200 μL of ultrapure water.

(3) Confirmation of prepared nanoparticles The modified nanoparticles prepared in (1) and (2) above are evaluated by particle size distribution measurement using dynamic light scattering (DLS), and the modification of the target protein is confirmed. bottom. The results are shown in FIG. FIG. 4A shows the results before and after modification of BSA-modified silver nanoparticles with anti-BSA rabbit polyclonal antibodies, and FIG. 4B shows the results before and after modification of gold nanoartins with anti-rabbit polyclonal antibodies. In both cases, modification of the antibody shifted the particle size distribution in the positive direction, suggesting that the target protein was modified.

(4) Dark-field Observation of Nanoparticles The modified nanoparticles prepared in (1) and (2) above and a mixed sample thereof were observed under a dark-field. As the mixed sample, 5 μL of the solution (1) and 5 μL of the solution (2) were mixed and left to stand for 3 hours. Preparation of the preparation for observation was performed by the method described in Example 2. The prepared slide was observed under dark field by oil immersion. A microscope (BX53, manufactured by OLYMPUS), a CCD camera (color, DP73, manufactured by OLYMPUS), and an objective lens (UplanFLN60×, manufactured by OLYMPUS) were used as observation equipment, and the observation conditions were an exposure time of 10 sec and ISO 100 and 4800. Photographed at ×3600 pixel shift. In the mixed sample, bright spots of different colors (pink to purple) were observed for both silver nanoparticles and gold nanoartins.

(5) Prediction of silver nanoparticle-gold nanoarchine dimer image The modified nanoparticles prepared in (1) and (2) above are each observed in the dark field, and the observed 10 bright spots are each observed at 30 × 30 pixels. Extracted. Next, a calculated dimer image was obtained by superimposing the bright spots of the modified nanoparticles of (1) and the bright spots of the modified nanoparticles of (2). Specifically, all combinations (10 × 10) of the 10 bright spots of the modified nanoparticles of (1) and the 10 bright spots of the modified nanoparticles of (2) are superimposed, A total of 100 calculated dimer images were obtained. Calculated dimer is intended to be an image presumed to be observed when the modified nanoparticles of (1) and modified nanoparticles of (2) form a dimer. Each of the 10 extracted bright spots and the calculated dimer image were visualized by three-dimensional plotting of RGB values (Fig. 5). As a result, as shown in (a) of FIG. 5, it was confirmed that the clusters were divided into three clusters: clusters of silver nanoparticles, clusters of gold nanoartins, and clusters of calculated dimer images. Also, it was presumed that three-dimensional plots of RGB values of bright spots observed as dimers were at the positions shown in (a) to (c) of FIG.

(6) RGB three-dimensional plot of mixed sample Each dark field image observed in (4) above was processed by the method shown in Examples 3 and 4, and the RGB three-dimensional plot including the calculated dimer in (5) above was plotted. It is shown in FIG. As shown in (a) to (c) of FIG. 6, a plot clearly different from the plot derived from silver nanoparticles and gold nanoartin was confirmed in the mixed sample, and the plot was sufficiently close to the calculated dimer plot. Met. These results suggested the presence of dimers of silver nanoparticles and gold nanoartines in the mixed sample. In addition, as shown in (a) to (c) of FIG. 6, it was confirmed that silver nanoparticles, gold nanoartine, and silver nanoparticle-gold nanoartine dimer can be distinguished by color separation. In the plot shown in FIG. 6, clusters of silver nanoparticles also include bright spots of modified silver nanoparticles present as monomers in the mixed sample. Similarly, in the plot shown in FIG. 6, the clusters of gold nanoartins also contain bright spots of modified gold nanoartins present as monomers in the mixed sample.

INDUSTRIAL APPLICABILITY The present invention can be used in fields using techniques for observing nanoparticles under a dark field.

Claims (8)

A method for distinguishing nanoparticles observable as bright spots under dark field observation,
Observing a sample containing nanoparticles under dark field observation to obtain a color observation image;
a step of color-separating the observation image to obtain color information of the luminescent spots including information of each color component; including a step of determining whether it is a bright spot of
A method for distinguishing nanoparticles, wherein the material of the nanoparticles is a metal having plasmon resonance characteristics . 2. The method for distinguishing nanoparticles according to claim 1, wherein the sample contains two or more kinds of nanoparticles that produce bright spots of different colors. 3. The method of discriminating nanoparticles according to claim 1, wherein in the step of obtaining the observation image, the nanoparticles in the sample are immobilized on a substrate and observed. 3. The method of discriminating nanoparticles according to claim 1, wherein in the step of obtaining the observation image, the nanoparticles in the sample are observed without being immobilized on a substrate. 5. The method for determining nanoparticles according to claim 1, wherein the sample contains a gold nanoparticle monomer as the nanoparticles. 6. Any one of claims 1 to 5, wherein in the discriminating step, it is discriminated whether the bright spots in the observed image are the nanoparticles, aggregates of the nanoparticles, or impurities. The method for distinguishing nanoparticles according to . 7. The method for distinguishing nanoparticles according to claim 1, wherein the color separation of the observation image is performed by color-separating the observation image into RGB. 2. The method for discriminating nanoparticles according to claim 1 , wherein the discriminating information is a bright spot classification model obtained by machine-learning a correlation between the color information and the attribution of the bright spots.

Images