JP7653198B2 - Sample Measurement Device - Google Patents

Description

Application of Article 30, Paragraph 2 of the Patent Act. Presented at the Tokyo Institute of Technology/Tokyo Medical and Dental University Matching Fund Research Results Presentation and Research Exchange Meeting held at Tokyo Medical and Dental University on April 10, 2024.

Patent Law Article 30, Paragraph 2 applied Poster presentation at the Tokyo Institute of Technology/Tokyo Medical and Dental University Matching Fund Research Results Presentation and Research Exchange Meeting held at Tokyo Medical and Dental University on April 10, 2024

Application of Article 30, Paragraph 2 of the Patent Act Announced at Transducers 2023 held at the National Kyoto International Conference Center on June 29, 2023

Article 30, paragraph 2 of the Patent Act applied. "Evaluation of frequency characteristics of single nanoparticles by AC nanopore method using flow-type nanopore" was announced in the proceedings of the 13th Micro-Nano Engineering Symposium, published on November 7, 2022 at https://shunkosha1.sakura.ne.jp/mnm2022/.

Article 30, paragraph 2 of the Patent Act applied. November 15, 2022: At the 13th Micro-Nano Engineering Symposium held at Asti Tokushima, a presentation was made on "Evaluation of frequency characteristics of single nanoparticles by AC nanopore method using flow-type nanopores."

Article 30, paragraph 2 of the Patent Act applies. A paper titled "Study on nanopore structure for broadband AC nanopore method" was published in the proceedings of the 13th Micro-Nano Engineering Symposium, published on November 7, 2022 at https://shunkosha1.sakura.ne.jp/mnm2022/.

Article 30, paragraph 2 of the Patent Act applied. November 15, 2022: At the 13th Micro-Nano Engineering Symposium held at Asti Tokushima, a presentation was made on "Study of nanopore structure for broadband AC nanopore method."

The present invention relates to a sample measurement device, and in particular to a sample measurement device capable of measuring minute measurement objects.

As detection methods for viruses, bacteria, microorganisms, or viroids (hereinafter simply referred to as viruses, etc.), the polymerase chain reaction (PCR) method, immunochromatography, etc. are known. Meanwhile, as another detection method, a method of sensing viruses, etc. using electrical measurement is being considered. One example of the electrical measurement method is a technique in which the virus, etc. to be detected is dispersed in water and the virus, etc. is electrically detected as particles. From the viewpoint of preventing infectious diseases and the like and preventing their spread, there is a demand for detection of viruses, etc. in environments such as indoors, livestock barns, and outdoors. The above-mentioned electrical measurement method is suitable for sensing in such environments (for example, Patent Documents 1 and 2).

JP 2020-202824 A JP 2020-098211 A

As described above, there is a need for technology that can quickly and accurately identify minute measurement objects such as viruses, bacteria, microorganisms, or viroids in the environment from the perspective of preventing and spreading infectious diseases. In view of this problem, the object of the present invention is to provide a sample measurement device that can quickly and accurately identify minute measurement objects.

A sample measurement device according to one aspect of the present invention includes a flow path device including a microchannel through which a sample solution containing a measurement object dispersed in a solvent flows, an inlet portion disposed on one side of the microchannel for introducing the sample solution into the microchannel, a outlet portion disposed on the other side of the microchannel for discharging the sample solution from the microchannel, a first electrode disposed at a position corresponding to the inlet portion when viewed in a plan view, and a second electrode disposed at a position corresponding to the outlet portion when viewed in a plan view, and a measurement portion that applies an AC voltage to the first electrode and the second electrode and identifies the measurement object based on the AC characteristics of the measurement object when the measurement object passes through the microchannel.

The present invention provides a sample measurement device that can quickly and accurately identify minute measurement objects.

FIG. 1 is a cross-sectional view for explaining a sample measuring device according to an embodiment. FIG. 1 is a cross-sectional view for explaining a sample measuring device according to an embodiment. FIG. 2 is an exploded perspective view for explaining a flow path device of the sample measurement device according to the embodiment. 1 is a graph showing AC characteristics measured by a sample measuring device according to an embodiment. 4 is a graph showing an example of a measurement result of the sample measurement device according to the embodiment. FIG. 1 is a diagram for explaining the effect of the present invention. FIG. 4 is a diagram for explaining the state of an electric field in a microchannel. FIG. 1 is a diagram for explaining the effect of the present invention. FIG. 13 is a diagram for explaining machine learning in the sample measurement device according to the embodiment. FIG. 13 is a diagram for explaining machine learning in the sample measurement device according to the embodiment. FIG. 13 is a diagram showing the measurement results of a sample measuring device. FIG. 13 is a diagram showing the measurement results of a sample measuring device. FIG. 13 is a diagram showing the measurement results of a sample measuring device. FIG. 13 is a diagram showing the measurement results of a sample measuring device. FIG. 13 is a diagram showing the measurement results of a sample measuring device.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
1 and 2 are cross-sectional views for explaining a sample measurement device according to an embodiment. As shown in Fig. 1 and Fig. 2, the sample measurement device 1 according to the present embodiment includes a flow path device 10 and a measurement unit 30. The flow path device 10 includes an inlet unit 13, a minute flow path 14, an outlet unit 15, a first electrode 23, and a second electrode 24. Fig. 2 shows a state in which a measurement target 42 has been introduced into the sample measurement device 1 shown in Fig. 1.

2, the flow channel device 10 includes a microchannel 14 through which a sample solution 40 containing a measurement target 42 dispersed in a solvent 41 flows, an inlet 13 arranged on one side (upstream side) of the microchannel 14 to introduce the sample solution 40 into the microchannel 14, and an outlet 15 arranged on the other side (downstream side) of the microchannel 14 to discharge the sample solution 40 from the microchannel 14. An inlet 16 is provided on the upper surface of the inlet 13, and the sample solution 40 is introduced into the inlet 13 from the inlet 16. An outlet 17 is provided on the upper surface of the outlet 15, and the sample solution 40 is discharged to the outside from the outlet 17. A first electrode 23 is arranged at a position corresponding to the inlet 13 when viewed in a plane. A second electrode 24 is arranged at a position corresponding to the outlet 15 when viewed in a plane.

The sample measurement device 1 according to this embodiment is a device used to identify a measurement target 42 such as a virus, bacteria, or microorganism. The sample measurement device 1 according to this embodiment can also identify targets other than viruses, bacteria, and microorganisms, as long as the objects are dispersible in the solvent 41. In this embodiment, an ionic liquid can be used as the solvent 41. Also, in this embodiment, a liquid other than an ionic liquid (e.g., an aqueous solution) can be used as the solvent 41.

Figure 3 is an exploded perspective view for explaining the flow path device of the sample measurement apparatus according to this embodiment. As shown in Figure 3, the flow path device 10 includes a flow path chip 11 in which an inlet 13, a micro flow path 14, and an outlet 15 are formed, and an electrode chip 21 in which a first electrode 23 and a second electrode 24 are formed on a substrate 22. In this embodiment, the flow path device 10 is constructed by bonding the electrode chip 21 and the flow path chip 11 together. Note that Figure 3 shows the flow path chip 11 upside down.

The flow channel chip 11 can be formed using silicone rubber 12. Specifically, the flow channel chip 11 can be produced by micromolding the silicone rubber 12. In producing a mold for micromolding, the micro flow channel 14 can be produced using a focused ion beam chemical vapor deposition (FIB-CVD) method, and the inlet section 13 and outlet section 15 can be produced using photoresist. In addition, the mold for micromolding can be produced using a back exposure method in which a photoresist matrix pattern is directly formed on a photomask.

The electrode tip 21 can be produced by forming a first electrode 23 and a second electrode 24 on a substrate 22. For example, a PCB substrate can be used for the substrate 22. Furthermore, Au, Cu, etc. can be used as the material for the first electrode 23 and the second electrode 24.

The flow path chip 11 and the electrode chip 21 may be bonded together using an adhesive. In addition, since the flow path chip 11 is formed using silicone rubber 12, the adhesive properties of the silicone rubber 12 may be utilized to bond the flow path chip 11 and the electrode chip 21 together.

As an example, the cross-sectional area of the microchannel 14 shown in Fig. 2 perpendicular to the direction in which the sample solution 40 flows is 0.05 ^µm2 or more and 100 ^µm2 or less. The length of the microchannel 14 is 0.1 µm or more and 100 µm or less. These sizes can be changed appropriately depending on the size of the measurement target 42.

The measurement unit 30 is connected to the first electrode 23 and the second electrode 24. The measurement unit 30 applies an AC voltage to the first electrode 23 and the second electrode 24, and identifies the object to be measured 42 based on the AC characteristics of the object to be measured 42 when the object to be measured 42 passes through the microchannel 14.

The measurement unit 30 may be any circuit capable of measuring the AC characteristics of the measurement object 42. The measurement unit 30 is configured to be capable of applying an AC voltage having a frequency of several kHz to several GHz, preferably 1 kHz to 100 MHz, and more preferably 1 MHz to 10 MHz, to the first electrode 23 and the second electrode 24. For example, a lock-in amplifier may be used for the measurement unit 30.

The measurement unit 30 is also configured to determine the composite impedance and phase using the measured AC characteristics, and to identify the object using the determined composite impedance and phase. In this embodiment, the composite impedance and phase may be used to determine parameters corresponding to the resistance component, zeta potential, and dielectric constant of the object 42, respectively, and the object 42 may be identified using the determined parameters corresponding to the resistance component, zeta potential, and dielectric constant.

Next, we will explain the operation of identifying the measurement target 42 using the sample measurement device 1 according to this embodiment.

When identifying the object to be measured 42, first, the object to be measured 42 dispersed in the solvent 41 is caused to flow through the microchannel 14. Next, an AC voltage is applied to the first electrode 23 and the second electrode 24, and the AC characteristics of the object to be measured 42 are measured when the object to be measured 42 passes through the microchannel 14. For example, the measurement unit 30 applies an AC voltage having a frequency of 1 kHz to 100 MHz to the first electrode 23 and the second electrode 24. Note that in this embodiment, an AC voltage may be applied to the first electrode 23 and the second electrode 24 in advance, and then the object to be measured 42 dispersed in the solvent 41 may be caused to flow through the microchannel 14.

In this embodiment, the measurement unit 30 may apply a DC voltage between the first electrode 23 and the second electrode 24 to introduce the positively or negatively charged measurement object 42 from the introduction unit 13 into the microchannel 14. Specifically, when the measurement object 42 is positively charged, the first electrode 23 is set to a positive electrode, the second electrode 24 is set to a negative electrode, and a DC voltage for electrophoresis is applied between the first electrode 23 and the second electrode 24, so that the positively charged measurement object 42 can be introduced from the introduction unit 13 into the microchannel 14. When the measurement object 42 is negatively charged, the first electrode 23 is set to a negative electrode, the second electrode 24 is set to a positive electrode, and a DC voltage for electrophoresis is applied between the first electrode 23 and the second electrode 24, so that the negatively charged measurement object 42 can be introduced from the introduction unit 13 into the microchannel 14. At this time, the measurement unit 30 may apply an AC voltage for measurement and a DC voltage for electrophoresis in a superimposed manner to the first electrode 23 and the second electrode 24.

At this time, the speed at which the object to be measured 42 passes through the microchannel 14 can be adjusted by adjusting the magnitude and time of the DC voltage applied to the first electrode 23 and the second electrode 24. In this embodiment, the number of objects to be measured 42 that pass through the microchannel 14 at the same time is, in principle, one. When the object to be measured 42 is introduced into the microchannel 14 using the DC voltage applied to the first electrode 23 and the second electrode 24 in this manner, the accuracy of introducing the object to be measured 42 into the microchannel 14 can be improved. Therefore, the detection accuracy of the object to be measured 42 can be improved.

In other words, when the introduction section 13 is pressurized to move the sample solution 40 into the microchannel 14 and move the measurement object 42 into the microchannel 14, if the number of measurement objects 42 passing through the microchannel 14 is small, such as when the concentration of the measurement object 42 in the sample solution 40 is low, the movement speed cannot be controlled, and there is a problem in that the measurement accuracy does not improve. In contrast, in this embodiment, the measurement object 42 is introduced into the microchannel 14 using a DC voltage applied to the first electrode 23 and the second electrode 24, so that the accuracy when introducing the measurement object 42 into the microchannel 14 can be improved.

In addition, in this embodiment, the polarity of the DC voltage applied to the first electrode 23 and the second electrode 24 can be switched to move the measurement object 42 flowing through the microchannel 14 back and forth. This can further improve the detection accuracy of the measurement object 42.

In addition, in this embodiment, equipment (external pump, etc.) for introducing the measurement object 42 into the microchannel 14 is not required, so the device configuration can be made smaller. In addition, in this embodiment, concentration polarization can be induced with a DC voltage and the state can be measured with an AC voltage, so the surface potential of the measurement object 42 can be measured more accurately.

After measuring the AC characteristics of the measurement object 42 as described above, the measurement object 42 is identified using the measured AC characteristics. Specifically, the measured AC characteristics are used to determine the composite impedance and phase, and the measurement object 42 is identified using the determined composite impedance and phase.

In this embodiment, the synthetic impedance and phase are used to determine parameters corresponding to the resistance component, zeta potential, and dielectric constant of the object to be measured 42, respectively, and the object to be measured 42 may be identified using the parameters corresponding to the determined resistance component, zeta potential, and dielectric constant. For example, the object to be measured 42 may be identified by mapping the parameters corresponding to the resistance component, zeta potential, and dielectric constant of the object to be measured 42 on a three-dimensional coordinate system with the resistance component, zeta potential, and dielectric constant as its axes. Here, the parameter corresponding to the resistance component is a parameter corresponding to the size of the object to be measured 42. The parameter corresponding to the zeta potential is a parameter corresponding to the surface potential of the object to be measured 42. The parameter corresponding to the dielectric constant is a parameter corresponding to the structure of the object to be measured 42, such as a spherical shell structure or membrane capacitance, or a parameter corresponding to the material of the object to be measured 42.

Since the size, surface potential, and structure of each measurement object 42 are different, the parameters corresponding to these, namely, the resistance component, zeta potential, and dielectric constant, are calculated, and the parameters corresponding to the resistance component, zeta potential, and dielectric constant of the measurement object 42 are mapped onto three-dimensional coordinates, thereby classifying the measurement object 42. For example, by mapping the measurement results of multiple types of measurement objects 42 onto three-dimensional coordinates and accumulating the data, the accuracy of identifying the measurement object 42 can be improved.

In addition, in this embodiment, a lock-in amplifier may be used to extract the in-phase component, which is an AC characteristic, and the phase component that is out of phase with the in-phase component, and the extracted in-phase component and phase component may be used to determine the composite impedance and phase, and the object may be identified using the time change in the composite impedance and the time change in the phase.

Figure 4 is a graph showing AC characteristics measured by the sample measurement device according to this embodiment, and shows AC characteristics measured using a lock-in amplifier. In other words, Figure 4 shows the time response waveforms of the current and phase change. Looking at the enlarged view in Figure 4, it can be seen that the current and phase changes are synchronized and are waveforms originating from the same measurement object 42.

Figure 5 is a graph showing an example of the measurement results of the sample measurement device according to this embodiment. Figure 5 shows histograms of current change values for polystyrene beads with a diameter of 1 μm, polystyrene beads with a diameter of 2 μm, and E. coli. In the measurement in Figure 5, the width, height, and length of the microchannel 14 of the flow channel device 10 were set to 3 μm, 4 μm, and 30 μm, respectively. Au plating was used for the first electrode 23 and the second electrode 24. A lock-in amplifier was used for the measurement.

As shown in FIG. 5, in this embodiment, a unimodal distribution according to particle size was obtained. Here, the horizontal axis in FIG. 5 is the current value, and the larger the particle diameter of the object to be measured 42, the larger the amount of change in current is, suggesting that a change in current value according to the particle diameter of the object to be measured 42 is obtained.

As described above, in the invention according to this embodiment, the measurement object 42 is measured using a flow path device 10 including a microchannel 14, an inlet 13 for introducing a sample solution 40 into the microchannel 14, an outlet 15 for discharging the sample solution 40 from the microchannel 14, a first electrode 23 arranged at a position corresponding to the inlet 13, and a second electrode 24 arranged at a position corresponding to the outlet 15. Thus, the present invention can provide a sample measurement device capable of quickly and accurately identifying a minute measurement object.

Figure 6 is a diagram for explaining the effect of the present invention. As shown in the left diagram of Figure 6, in the conventional nanopore configuration, that is, in the configuration in which electrodes are provided on each of the upper and lower surfaces of a substrate having a hole through which the measurement object passes, the substrate, which is an insulator, is sandwiched between two electrodes, so the capacitance of the flow path device tends to be large. In other words, if the capacitance of the flow path device is C, the dielectric constant of the substrate is ε, the thickness of the substrate is d, and the area of the substrate is S, the capacitance of the device is expressed as C = εS/d. Therefore, in the conventional nanopore configuration, as shown in the left diagram of Figure 6, d is small and S is large, so the capacitance of the flow path device tends to be large. In contrast, in the flow path device (straw-type nanopore) of this embodiment, as shown in the right diagram of Figure 6, d can be made large while S can be made small, so the capacitance of the flow path device can be reduced.

Figure 7 is a top view for explaining the state of the electric field in the microchannel. The left diagram of Figure 7 shows the state of the electric field in a conventional nanopore, and the right diagram shows the state of the electric field in the straw-shaped nanopore of this embodiment. As shown in the left diagram of Figure 7, in a conventional nanopore, the electric field was concentrated only in the hole through which the object to be measured passed. In contrast, in this embodiment, when viewed in a plane, the first electrode 23 and the second electrode 24 sandwich the microchannel 14. Therefore, as shown in Figure 7, when an AC voltage is applied to the first electrode 23 and the second electrode 24, the electric field can be concentrated near the microchannel 14, thereby improving the detection accuracy of the object to be measured 42.

In this embodiment, the capacitance of the flow path device can be reduced as described above, so that the cutoff frequency can be increased by 400 times (2 MHz) compared to the conventional frequency, as shown in FIG. 8. This improves the detection accuracy of the object to be measured.

In addition, in this embodiment, the measurement unit 30 may further include a prediction model generation unit that uses the measured AC characteristics and information on the measurement object 42 as teacher data and generates a prediction model by machine learning that predicts the measurement object 42 based on the measured AC characteristics.

The prediction model generation unit may extract an in-phase component, which is an AC characteristic, and a phase component that is out of phase with the in-phase component, calculate a composite impedance and a phase using the extracted in-phase component and phase component, and generate a prediction model using a feature of the time change of the composite impedance and a feature of the time change of the phase.

The prediction model generation unit may generate the prediction model by machine learning using a neural network. For example, the prediction model generation unit may generate the prediction model by machine learning using a convolution neural network (CNN) or a long short term memory (LSTM).

Machine learning will be explained in detail below.
The present inventors have been working on the development of a measurement unit 30 (classifier) using a neural network to obtain higher accuracy in classifying bacterial species from individual measurement waveforms derived from bacteria. As a result, feature quantities as shown in FIG. 9 were extracted, and a classification algorithm using a neural network as shown in FIG. 10 was constructed. That is, in the spike (waveform) shown in FIG. 9, the spike height h and the spike width t _n (n=0, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90) were extracted as feature quantities, and a total of 12 feature quantities were extracted per spike. Here, the value of n corresponds to the depth (%) when the baseline is 0% and the bottom of the spike is 100%, and the spike width t _n at each depth is extracted as a feature quantity.

The extracted features were then used to generate a prediction model. A neural network like that shown in FIG. 10 was used for this purpose. By combining the sample measurement device according to the present embodiment described above with machine learning, it was possible to build a sample measurement device with a maximum accuracy of 97.6%.

Examples will be described below.
<Experimental Method>
The following bacterial samples were prepared:
Phosphate-buffered saline (PBS; 17-516Q, Lonza, Morrisville, NC, USA) was used as the measurement solvent, and 0.01% artificial phospholipid surfactant (Lipidoid BL206, NOF Corporation) was used as the anti-adsorption agent. The following six human infectious bacteria were used:
(1) S.enterica (Se, NBRC 3313)
(2) V. parahaemolyticus (Vp, NBRC12711)
(3) P. vulgaris (Pv, NBRC 3045)
(4) K. pneumoniae (Kp,NBRC 13277)
(5) M. intracellulare (Mi, distributed by Kobe City Institute of Environmental Health)
(6) M. mesophilicum (Mm, NBRC 15688)

M. intracellulare was cultured on "7H11 agar medium," and the rest were cultured on "standard agar medium." One colony from each bacterial culture was dispersed in 300 pL of the measurement solution and used for the measurement immediately after vortexing for about 10 seconds.

(AC Nanopore Measurement Method)
The microchannel in the conventional nanopore in the left diagram of FIG. 6 and the microchannel in the present invention are also referred to as nanopores below. In this experiment, measurements were performed using these nanopores. Here, the method of measuring the AC characteristics of a sample using the nanopores described above is referred to as AC nanopore measurement. A lock-in amplifier was used for the measurements. The lock-in amplifier extracts only the signal components with the same frequency as the reference signal from the noise background and measures the magnitude and phase of the current. In the AC nanopore measurement method, a spike-like descending current and an ascending phase change are observed synchronously for a single particle. In other words, a waveform similar to that in FIG. 4 is obtained. The magnitude of the current is proportional to the volume of the particle and was used to determine the size of the bacteria. Furthermore, a set of these current and phase spike waveforms was used as a machine learning feature for classifying bacteria.

(Nanopores)
In the AC nanopore measurement method according to this experiment, a commercially available nanopore (NP2000, Izon Science Ltd., hereinafter referred to as Izon nanopore) and a straw-shaped nanopore according to the present invention (see FIG. 1) were used. The cross-sectional shape of the microchannel of the straw-shaped nanopore was 3×3 μm, and the length was 30 μm.

(Measurement method)
For the Izon nanopore, the bacterial suspension was added to the upper cell with a pressure head of 3 mm (30 Pa) in all experiments, and the nanopore was driven by hydrostatic pressure. For the straw-shaped nanopore, 50 pl was injected into the inlet and 20 pl into the outlet, and the nanopore was driven by the pressure head generated by the difference between them. The applied voltage was 0.1 V for the Izon nanopore and 0.7 V for the straw-shaped nanopore. A minimum of 500 counts were measured for all measurements.

<Experimental Results>
The change in the magnitude of the current accompanying the passage of a particle (measurement object) is proportional to the volume of the particle. FIG. 11 is a histogram of the size distribution of bacterial particles measured by AC nanopore measurement using a straw-shaped nanopore. In FIG. 11, the horizontal axis indicates the magnitude of the current, and the vertical axis indicates the frequency. The histogram was created by counting at least 500 particles for each particle size, but due to the variation in counts, the vertical axis was normalized so that the total count for each particle was 100%. In the histogram shown in FIG. 11, each bacterial size (current value) is distributed overlappingly, making it difficult to classify the bacteria when evaluated based only on size.

Therefore, the inventors of the present application classified bacteria using machine learning.
First, the pulse waveforms of current magnitude and phase obtained by AC nanopore measurement were used for learning. In machine learning, the number of waveforms in the waveform set consisting of current magnitude and phase was first uniformed. Then, the 12 combinations with the highest learning effect for the peak height of each waveform when the peak height was set to 100% and the width of the waveform for every 10% increase were used as training data (see Figure 9). For CNN-based classification, the model was built with three layers (one input layer, one hidden layer, and one output layer).

Figures 12 and 13 show matrix evaluations of particle classification accuracy when the Random Forest method and the CNN method are applied to the measurement results using the Izon nanopore. In Figures 12 and 13, (a) shows the classification results for current magnitude only, (b) shows the classification results for phase only, and (c) shows the classification results for current magnitude and phase. In Figures 12 and 13, the darker the color of the squares on the diagonal line sloping downward to the right, the higher the classification accuracy. The numbers in the matrix indicate the number of classified particles.

Comparing the random forest method and the CNN method, as shown in (a), the classification accuracy when only the magnitude of the current was used was 61.4% for the random forest method and 66.6% for the CNN method. As shown in (b), the classification accuracy when only the phase was used was 57.2% for the random forest method and 70.7% for the CNN method. On the other hand, as shown in (c), when both the magnitude and phase of the current were used, the classification accuracy was 66.5% for the random forest method and 78.3% for the CNN method, showing that the CNN method showed higher classification accuracy.

Next, we will consider the classification accuracy of particles when the CNN method is applied to the measurement results using the straw-shaped nanopore. Figure 14 shows a matrix evaluation of the classification accuracy of particles when the CNN method is applied to the measurement results using the straw-shaped nanopore. As shown in Figure 14, the measurement using the straw-shaped nanopore showed higher classification accuracy in all sample sets compared to the measurement using the Izon nanopore (Figure 13). When comparing the case (c) using both the magnitude and phase of the current, the classification accuracy of the Izon nanopore (Figure 13) was 78.3%, while the classification accuracy of the straw-shaped nanopore (Figure 14) was 97.8%, showing a high classification accuracy of more than 19% for the straw-shaped nanopore. These results show that the combination of the CNN method and the straw-shaped nanopore can significantly improve the classification accuracy of bacteria.

In the above, we have described an example in which bacteria were used as the particles to be measured, but in order to confirm the classification accuracy for smaller particles, we will explain the experimental results when viruses were used as the measurement objects in the AC nanopore measurement method using a straw-shaped nanopore. Here, the following four viruses were used as the measurement objects.
(A) Influenza Virus type A (H1N1)
(B) Influenza Virus type A (H3N2)
(C) Influenza Virus type B (Tokio/53/99 strain)
(D) Respiratory Syncytial Virus

For these viruses, the classification accuracy was evaluated when the CNN method was applied to the measurement results of the magnitude and phase of the current to classify the particles. Figure 15 shows a matrix evaluation of the classification accuracy of particles when the CNN method was applied to the measurement results of the magnitude and phase of the current for viruses in a straw-shaped nanopore. The numbers in the matrix in Figure 15 indicate the accuracy rate of particle classification. As shown in Figure 15, high classification accuracy was shown, with 0.94 for Influenza Virus A (H1N1), 1.00 for Influenza Virus A (H3N2), 0.95 for Influenza Virus B (Tokio/53/99 strain), and 0.86 for Respiratory Syncytial Virus. The overall classification accuracy was 96.52%, and it was confirmed that the classification accuracy was roughly equivalent to that of bacteria (97.8%, Figure 14).

Therefore, by using the straw-shaped nanopore of the present invention to perform measurements using the AC nanopore measurement method, it is possible to classify microscopic particles such as bacteria and viruses with high accuracy.

The present invention has been described above in accordance with the above embodiment, but the present invention is not limited to the configuration of the above embodiment, and of course includes various modifications, alterations, and combinations that a person skilled in the art could make within the scope of the invention of the claims of this patent application.

1. Sample measurement device
REFERENCE SIGNS LIST 10 Flow path device 11 Flow path chip 12 Silicone rubber 13 Inlet portion 14 Micro flow path 15 Outlet portion 16 Inlet port 17 Outlet port 21 Electrode chip 22 Substrate 23 First electrode 24 Second electrode 30 Measurement portion 40 Sample solution 41 Solvent 42 Measurement target

Claims (15)

A flow path device;
A measuring unit,
the flow channel device includes a flow channel chip in which an inlet portion, a microflow channel, and a discharge portion are formed, and an electrode chip in which a first electrode and a second electrode are formed on a substrate;
the microchannel is formed in the channel chip so as to extend in a direction parallel to a main surface of the channel chip, and a sample solution containing a measurement target dispersed in a solvent flows through the microchannel ;
The introduction unit is disposed on one side of the microchannel, and introduces the sample solution into the microchannel;
the discharge unit is disposed on the other side of the microchannel, and the sample solution is discharged from the microchannel ;
the first electrode is disposed at a position corresponding to the introduction portion when viewed in a plan view from a direction perpendicular to a main surface of the electrode tip ,
the second electrode is disposed at a position corresponding to the discharge portion when viewed in a plan view from the direction perpendicular to the main surface of the electrode tip ,
the measurement unit applies an AC voltage to the first electrode and the second electrode, and identifies the object to be measured based on AC characteristics of the object to be measured when the object to be measured passes through the microchannel;
The flow channel device is configured by bonding the electrode chip and the flow channel chip together.
Sample measurement device. The sample solution is
the first electrode is introduced into the introduction section through an introduction port, which is an opening provided in the channel chip at a position corresponding to the first electrode;
The sample measurement device according to claim 1 , wherein the sample is discharged from the discharge section through a discharge port that is an opening provided in the channel chip at a position corresponding to the second electrode. 3. The sample measurement device according to claim 1, wherein the minute channel is a channel whose longitudinal direction is parallel to the main surface of the channel chip. 4. The sample measurement device according to claim 3 , wherein the area of a cross section of the microchannel perpendicular to a direction in which the sample solution flows is 0.05 μm ² or more and 100 μm ² or less. 4. The sample measurement device according to claim 3 , wherein the length of the microchannel is 0.1 μm or more and 100 μm or less. The sample measurement device according to claim 1 or 2, wherein the measurement unit is configured to be able to introduce the positively or negatively charged measurement object from the introduction unit into the microchannel by applying a DC voltage between the first electrode and the second electrode. 7. The sample measurement device according to claim 6 , wherein the measurement section is configured to be able to apply the AC voltage and the DC voltage in a superimposed manner to the first electrode and the second electrode. The sample measurement device according to claim 1 or 2, wherein the measurement unit includes a lock-in amplifier. The sample measurement device according to claim 1 or 2, wherein the measurement unit is configured to be capable of applying an AC voltage having a frequency of 1 kHz to 100 MHz to the first electrode and the second electrode. 3. The sample measurement device according to claim 1, wherein the measurement section determines a composite impedance and a phase using AC characteristics of the object to be measured, and identifies the object to be measured using the determined composite impedance and phase. Using the synthetic impedance and the phase, parameters corresponding to a resistance component, a zeta potential, and a dielectric constant of the object to be measured are obtained, respectively;
The sample measurement device according to claim 10 , wherein the object to be measured is identified using parameters corresponding to the determined resistance component, zeta potential, and dielectric constant. 3. The sample measurement device according to claim 1, wherein the measurement unit includes a prediction model generation unit that uses the AC characteristics of the object to be measured and information about the object to be measured as teacher data and generates a prediction model by machine learning to predict the object to be measured based on the AC characteristics of the object to be measured. The prediction model generation unit,
extracting an in-phase component and a phase component out of phase with the in-phase component, which are the AC characteristics;
determining a synthetic impedance and a phase using the extracted in-phase component and the extracted phase component;
The sample measurement device according to claim 12 , wherein the prediction model is generated using a feature amount of the time change of the synthetic impedance and a feature amount of the time change of the phase. The sample measurement device according to claim 12 , wherein the prediction model generation unit generates the prediction model by machine learning using a neural network. The sample measurement device according to claim 1 or 2, wherein the measurement object is a bacterium or a virus.