JP7184974B2 - RESIN COMPOSITION FOR FORMING ODOR SUBSTANCE RECEIVING LAYER, SENSOR ELEMENT, ODOR SENSOR, AND ODOR MEASURING DEVICE USING SAME - Google Patents

Description

The present invention relates to a resin composition for forming an odorant-receiving layer, a sensor element using the same, an odor sensor, and an odor measuring device.

With the development of information processing technology in recent years, if it is possible to quantify the sense of smell, which has not been sufficiently mechanically measured among the five human senses, in a wide range of fields such as medicine, environment and safety, and marketing. It is expected that it can be used in the industrial field. Semiconductor gas sensors have hitherto been used to detect specific gaseous substances (gases) with high accuracy and high sensitivity.

The invention described in Patent Document 1 proposes a mechanism in which the semiconductor of a semiconductor gas sensor is replaced with a conductive polymer to detect adsorption of odor components to the surface of the conductive polymer. Patent Document 1 reports that it is possible to detect an odor component that easily decomposes thermally and a substance that does not cause an oxidation-reduction reaction on the detection surface of the sensor.

Further, Patent Document 2 focuses on the property that the electrical resistance of a mixture of an organic polymer and a conductive substance changes when exposed to an organic gas. In Patent Document 2, a plurality of combinations of organic polymers/conductive substances with different organic polymer compositions are prepared from the above mixture, and when these are used as an electric resistance array in a sensor, the electric resistance when exposed to the same organic gas It is described that the resistance changes are different. Patent document 2 reports that odors can be identified by attributing the pattern of electrical resistance change and the type of odor (=a mixture of organic gases) using this.

Furthermore, Patent Document 3 reports that the response speed of the sensor is improved by adding a plasticizer to the above organic polymer.

JP-A-11-23508 Japanese Patent Publication No. 11-503231 Japanese Patent Publication No. 2002-519633

However, the conventional techniques described above have room for improvement from the standpoint of odor identification performance.

An object of one aspect of the present invention is to provide a resin composition for forming an odorant-receiving layer with improved odor discrimination performance, a sensor element, an odor sensor, and an odor measuring device using the resin composition.

The present inventors arrived at the present invention as a result of conducting studies to achieve the above object.

That is, one aspect of the present invention is a resin composition for forming an odorant-receiving layer, the resin composition containing a fluororesin (A), a surfactant (B), and a conductive carbon material (C). , a sensor element, an odor sensor, and an odor measuring device using the same.

According to one aspect of the present invention, it is possible to provide a resin composition for forming an odorant-receiving layer with improved odor discrimination performance, a sensor element, an odor sensor, and an odor measuring device using the resin composition.

1 is a block diagram showing an example of the configuration of an odor measurement device according to one embodiment of the present invention; FIG. FIG. 4 is a top view showing an example of the configuration of a sensor element; 3 is a cross-sectional view showing an example of the configuration of the sensor element shown in FIG. 2; FIG. It is a functional block diagram showing an example of a configuration of an odor measuring device. 10 is a flow chart showing an example of the flow of processing in which an estimation device generates an estimation model; It is a functional block diagram showing an example of a configuration of an odor measuring device. 4 is a flow chart showing an example of the flow of processing in which an estimating device estimates an odorant. FIG. 4 is a block diagram showing an example of the configuration of an odor measuring device according to another embodiment of the present invention;

An embodiment of the invention will be described below, but the invention is not limited thereto. In addition, unless otherwise specified in this specification, "A to B" representing a numerical range intends "A or more and B or less".

[1. Resin composition]
A resin composition according to one embodiment of the present invention is a resin composition for forming an odorant-receiving layer, comprising a fluororesin (A), a surfactant (B) and a conductive carbon material (C). include.

As used herein, the term “odorant” broadly means a substance that can be adsorbed to the odorant-receiving layer. Therefore, it also includes substances that are not generally regarded as odor-causing substances. "Smell" often includes multiple odorants, and there are substances that are not recognized as odorants or unknown odorants. One embodiment of the present invention focuses on the fact that the amount of odorant adsorbed to the odorant-receiving layer varies depending on the type of odorant.

In this specification, even when the term "odorant" is used simply, it may mean not an individual odorant but an "aggregate of odorants" that may contain a plurality of odorants.

The "odorant" is not particularly limited, but examples include hexane, ethyl acetate, methanol, diethyl carbonate, toluene, d-limonene, bornan-2-one, cis-3-hexenol, β-phenylethyl alcohol, citral, L- carvone, γ-undecalactone, eugenol, linalyl acetate, menthol, benzaldehyde, vanillin, hexanal, ethanol, pentyl valerate, linalool, 2-propanol and the like.

Further, in this specification, the term "odorant-receiving layer" means a layer that adsorbs an odorant to be identified. The odorant-receiving layer is formed from the resin composition described above. The odorant-receiving layer may be provided as part of the sensor element described below.

It is considered that the sensor described in Cited Document 1 can detect an odor composed of a single compound. On the other hand, many odors are mixtures of multiple substances. The sensor described in Cited Document 1 does not have a function of identifying odor components in the detection section, so the odor identification performance for mixtures is not sufficient. In Cited Document 2, by utilizing the difference in the chemical structure of the conductive polymer used in the detection part, the response to various compounds shown by the detection part through each conductive polymer is different, so that the mixture It has been shown that it is possible to recognize the odor as However, the chemical structures of conductive polymers are limited, and it is difficult to sensitively separate the response of the detection unit to arbitrary odor components, and it is difficult to distinguish between odors composed of similar components. Cited Document 3 proposes a method in which a mixture consisting of an organic polymer, a plasticizer, and a conductive substance is used as a detection material in a detection section, and the permeation of an odor component into the organic polymer is detected as a change in electrical resistance of the mixture. there is By utilizing the fact that different odor components are used to permeate organic polymers having different compositions, a plurality of detection units composed of the above-described detection materials containing organic polymers with different compositions are arranged in an array to detect odors as a mixture. can be recognized. However, in the organic polymer containing the above organic polymer and plasticizer, even if a plurality of combinations of the organic polymer/conductive substance are prepared, the difference in chemical properties between the organic polymers is small. not enough. These conventional techniques cannot accurately detect, for example, an actual odor pattern in which a plurality of substances interact with each other or an actual odor pattern due to a substance whose composition is unknown.

The present inventors have found that the electrical conductivity of the resin composition varies depending on the amount of the odorant adsorbed on the resin composition, and that the process of adsorption to the resin composition differs for each odorant. The present inventors have focused on the fact that there is a problem, and have invented a resin composition, a sensor element, and the like according to an embodiment of the present invention. By using such a resin composition, it is possible to improve the odor identification performance. For example, real-world odor patterns with multiple substances interacting or with substances of unknown composition can also be identified.

<Fluororesin (A)>
Examples of the fluororesin (A) include polyvinylidene fluoride (PVDF), vinylidene fluoride-propylene hexafluoride copolymer (PVDF-HFP), tetrafluoroethylene-hexafluoropropylene copolymer (FEP), poly Examples include chlorotrifluoroethylene (PCTFE) and chlorotrifluoroethylene-ethylene copolymer (ECTFE).

Among the above-described fluororesins, PVDF, PVDF-HFP, PCTFE and ECTFE are preferred, and PVDF, PVDF-HFP and FEP are more preferred from the viewpoint of detection accuracy of odorants.

The dielectric constant (60 Hz) of the fluororesin (A) is preferably 2.0 to 9.0, more preferably 2.3 to 9.0, most preferably 2.3 to 9.0, from the viewpoint of odor discrimination performance. 6.8 to 9.0.

<Measurement conditions of relative permittivity>
In this specification, the dielectric constant of the fluororesin (A) can be measured with reference to the method described in JIS C2138 of Japanese Industrial Standards (JIS).

Specifically, it is measured by the following method. E4908A (manufactured by Agilent Technologies Inc.) can be used as the LCR meter. Measurement conditions can be, for example, a frequency of 1 MHz and 25°C.
1. A test piece with a diameter of 100 mm and a thickness of 1.5 mm is produced.
2. Aluminum foil having a diameter of 36 mm with a tab is pasted on the top and bottom of the test piece using a small amount of silicone grease to form electrodes.
3. Connect the electrode tabs to an LCR meter to measure the dielectric constant.

<Surfactant (B)>
Surfactant (B) is not particularly limited, but preferably has an HLB value of 8-18, more preferably 9-17, and particularly preferably 10-16. By using the surfactant (B) having such an HLB value, good odor discrimination performance can be obtained.

The “HLB value” here is an index showing the balance between hydrophilicity and lipophilicity, and is described, for example, in “Introduction to Surfactants” [published by Sanyo Chemical Industries, Ltd., 2007, by Takehiko Fujimoto], page 212. It is known as the calculated value by the Oda method, and not the calculated value by the Griffin method.

The HLB value can be calculated from the ratio of the organic value and the inorganic value of the organic compound.

HLB = 10 x inorganic/organic Here, the inorganic and organic values in the above formula represent the index values expressing the organic and inorganic properties proposed by Fujita et al. It can be calculated using the values in the table described on page 213 of "Introduction".

Surfactants (B) include, for example, anionic surfactants, cationic surfactants, amphoteric surfactants and nonionic surfactants.

Examples of anionic surfactants include alkali metal salts of carboxylic acids having 10 to 24 carbon atoms and alkali metal salts of alkylsulfonic acids having 14 to 24 carbon atoms.

Examples of the carboxylic acid having 10 to 24 carbon atoms include decanoic acid, undecanoic acid, dodecanoic acid, tridecanoic acid, tetradecanoic acid, hexadecanoic acid, heptadecanoic acid, octadecanoic acid, pentadecanoic acid, nonadecanoic acid, icosanoic acid, henicosanoic acid, docosanoic acid, tricosanoic acid, tetracosanoic acid and the like.

Examples of the alkyl group of the alkylsulfonic acid having 14 to 24 carbon atoms include tetradecyl group, pentadecyl group, hexadecyl group, heptadecyl group, octadecyl group, nonadecyl group, icosyl group, henicosyl group, docosyl group, tricosyl group and tetracosyl. and the like.

Examples of the alkali metal contained in the alkali metal salt include sodium and potassium.

Examples of cationic surfactants include halide salts of quaternary ammonium having an alkyl group of 12 to 24 carbon atoms.

Examples of the quaternary ammonium having an alkyl group having 12 to 24 carbon atoms include tetrapropylammonium, tetrabutylammonium, tetrapentylammonium, tetrahexylammonium, dimethyldioctylammonium, didecyldimethylammonium, decyltrimethylammonium and dodecyltrimethyl. ammonium, tridecyltrimethylammonium, hexadecyltrimethylammonium, methyltrioctylammonium, octyltrimethylammonium, tributylmethylammonium, octadecyltrimethylammonium, tetradecyltrimethylammonium, nonadecyltrimethylammonium, icosyltrimethylammonium, henicosyltrimethylammonium, heptadecyltrimethylammonium, pentadecyltrimethylammonium and the like.

Examples of the halide salts include fluoride salts, chloride salts, bromide salts and iodide salts.

Amphoteric surfactants include, for example, dimethyl(3-sulfopropyl)ammonium inner salt having an alkyl group having 10 to 22 carbon atoms, N-alkyl-N,N-dimethyl having an alkyl group having 10 to 22 carbon atoms. glycine and the like.

Examples of the dimethyl(3-sulfopropyl)ammonium hydroxide inner salt having an alkyl group having 10 to 22 carbon atoms include decyldimethyl(3-sulfopropyl)ammonium hydroxide inner salt, undecyldimethyl(3-sulfopropyl ) ammonium hydroxide inner salt, dodecyldimethyl(3-sulfopropyl) ammonium hydroxide inner salt, tridecyldimethyl(3-sulfopropyl) ammonium hydroxide inner salt, tetradecyldimethyl(3-sulfopropyl) ammonium hydroxide demolecular salt, pentadecyldimethyl(3-sulfopropyl)ammonium hydroxide inner salt, hexadecyldimethyl(3-sulfopropyl)ammonium hydroxide inner salt, heptadecyldimethyl(3-sulfopropyl)ammonium hydroxide molecule inner salt, octadecyldimethyl(3-sulfopropyl)ammonium hydroxide inner salt, nonadecyldimethyl(3-sulfopropyl)ammonium hydroxide inner salt, icosyldimethyl(3-sulfopropyl)ammonium hydroxide inner salt, henicosyldimethyl(3-sulfopropyl)ammonium hydroxide inner salt and docosyldimethyl(3-sulfopropyl)ammonium hydroxide inner salt, and the like.

Examples of N-alkyl-N,N-dimethylglycine having an alkyl group having 10 to 22 carbon atoms include N-dodecyl-N,N-dimethylglycine and N-octadecyl-N,N-dimethylglycine.

Examples of nonionic surfactants include higher alcohol ethylene oxide adducts.

Higher alcohols include 1-hexyl alcohol, 1-heptyl alcohol, 1-octyl alcohol, 1-nonyl alcohol, 1-decyl alcohol, 1-undecyl alcohol, 1-dodecyl alcohol, 1-tridecyl alcohol, 1-tetra decyl alcohol, 1-pentadecyl alcohol, 1-hexadecyl alcohol, 1-heptadecyl alcohol, 1-octadecyl alcohol and the like.

The number of moles of ethylene oxide added is preferably 5 to 50, more preferably 5 to 40, still more preferably 5 to 30 from the viewpoint of odor discrimination performance.

Surfactant (B) is preferably an anionic surfactant, a cationic surfactant or a nonionic surfactant, more preferably a nonionic surfactant or a cationic surfactant, from the viewpoint of odor discrimination performance. agent, most preferably a cationic surfactant.

The weight ratio [(A)/(B)] of the fluororesin (A) and the surfactant (B) is preferably 1.0 to 4.0, more preferably 1.0 to 4.0, from the viewpoint of odor discrimination performance. is between 1.0 and 2.3, most preferably between 1.0 and 1.5.

The fluororesin (A) and the surfactant (B) may or may not be compatible with each other.

<Conductive carbon material (C)>
In this specification, the conductive carbon material (C) is a carbon material having a volume resistivity of 0.1 Ω·cm or less. The resin composition described above is in a state in which the conductive carbon material (C) is dispersed in the mixture of the fluororesin (A) and the surfactant (B). The conductive carbon material (C) is in contact with each other to form a conductive path, so that the resin composition has conductivity.

Examples of the conductive carbon material (C) include carbon black, carbon nanotubes and graphene. Examples of carbon black include furnace black and acetylene black.

Commercially available furnace blacks include Ketjenblack EC (trade name manufactured by Akzo, the Netherlands), Ketjenblack EC-300J (trade name manufactured by Lion Specialty Chemicals Co., Ltd.), and Ketjenblack EC-600JD (Lion Specialty Chemicals). (manufactured by Tokai Carbon Co., Ltd.), SEAST G116, 116 (manufactured by Tokai Carbon Co., Ltd.), and Niteron #10 (manufactured by Nippon Steel Chemical Co., Ltd.).

Commercially available acetylene black products include DENKA BLACK (trade name manufactured by Denki Kagaku Kogyo KK) and SUPER C-65 (trade name manufactured by MTI Corporation, USA).

Commercially available carbon nanotubes include VGCF-H (trade name of Showa Denko KK).

Graphene is commercially available from Sigma-Aldrich.

The shape of the conductive carbon material (C) is preferably fibrous or spherical.

When fibrous, the fiber diameter is preferably 0.1 to 10 μm, more preferably 0.1 to 5 μm. The fiber length is preferably 0.1-10 μm, more preferably 1-10 μm.

When spherical, the primary particle size is preferably 10 nm to 200 nm, more preferably 20 nm to 150 nm.

The content of the conductive carbon material (C) is preferably 100% by weight in total of the fluororesin (A), the surfactant (B) and the conductive carbon material (C) from the viewpoint of odorant receptivity. is 25-75% by weight, more preferably 30-65% by weight, most preferably 35-55% by weight. Alternatively, the content of the conductive carbon material (C) is 5 to 30% by weight with respect to the total 100% by weight of the fluororesin (A), the surfactant (B) and the conductive carbon material (C). may be from 5 to 20% by weight, or from 5 to 10% by weight.

<Method for producing resin composition>
A specific example of the method for producing a resin composition according to one embodiment of the present invention is as follows.

The resin composition is obtained as a slurry by mixing the fluororesin (A), the surfactant (B), the conductive carbon material (C) and the solvent (D) and uniformly kneading the mixture with a stirrer. The odorant-receiving layer is obtained by applying this to the gaps between a set of two metal wires, followed by drying by heating.

The solvent (D) is not particularly limited as long as it can be removed by drying, but is preferably N-methylpyrrolidone, N,N-dimethylformamide, N,N-dimethylacetamide, tetrahydrofuran, ethyl acetate, water, toluene. and xylene.

[2. sensor element 31]
The resin composition described above shows different changes in electrical conductivity with time when the odorant A is adsorbed on the resin composition and when an odorant B different from the odorant A is adsorbed on the resin composition. By utilizing this property, it is possible to realize the sensor element 31 capable of detecting and identifying odorants.

The resin composition described above shows different changes in electrical conductivity with time when the odorant A is adsorbed on the resin composition and when an odorant B different from the odorant A is adsorbed on the resin composition. By utilizing this property, it is possible to realize the sensor element 31 capable of detecting and identifying odorants.

Below, the outline and effects of the sensor element 31 to which the resin composition according to one embodiment of the present invention is applied will be described.

The sensor element 31 includes an odorant-receiving layer 315 containing the resin composition described above, first metal wiring 313A, and second metal wiring 313B. Note that, hereinafter, the first metal wiring 313A and the second metal wiring 313B may be referred to as the metal wiring 313 when not distinguished from each other.

Here, first metal wiring 313A and second metal wiring 313B will be described with reference to FIGS. 2 and 3. FIG. 2 is a top view showing an example of the configuration of the sensor element 31, and FIG. 3 is a cross-sectional view showing an example of the configuration of the sensor element 31 shown in FIG.

The first metal wiring 313A and the second metal wiring 313B are metal wirings that function as electrodes for measuring changes in the electrical conductivity of the odorant-receiving layer 315 (that is, the resin composition). That is, the first metal wiring 313A and the second metal wiring 313B are separated from each other, and the odorant-receiving layer 315 is in contact with at least part of the first metal wiring and at least part of the second metal wiring. . In one example, the first metal wiring 313A and the second metal wiring 313B are metal wirings that are not in direct contact with each other, and as shown in FIG. 2, may be metal wirings that are substantially parallel to each other.

Metal wiring 313 including first metal wiring 313 A and second metal wiring 313 B as shown in FIG. 2 may be arranged on substrate 311 . Substrate 311 may be a substrate such as glass epoxy commonly used in electronic circuits. Metal wiring 313 may be a metal wiring such as copper or gold. The thickness of each of the first metal wiring 313A and the second metal wiring 313B viewed from the direction perpendicular to the surface of the substrate is preferably 10 μm to 2 mm, more preferably 10 μm to 1 mm. The height of each of the first metal wiring 313A and the second metal wiring 313B when viewed in a direction parallel to the surface of the substrate, that is, the thickness is preferably 1 μm to 100 μm, more preferably 10 μm to 50 μm. The interval between the first metal wiring 313A and the second metal wiring 313B is preferably 1 μm to 1 mm, more preferably 1 μm to 100 μm. The length of the metal wiring 313 is preferably 10 μm to 50 mm, more preferably 10 μm to 30 mm.

The metal wiring 313 may be arranged on the seal substrate 312 . FIG. 3 shows the AA section of FIG. As shown in FIG. 3, a seal substrate 312 may be arranged on a substrate 311 made of glass epoxy or the like, and metal wiring 313 may be arranged on the seal substrate 312 . A vinyl tape 314 may be used to fix the seal substrate 312 on the substrate 311 . Vinyl tape 314 can also be used to adjust the length of the exposed portion of metal wire 313 by masking the excess portion of metal wire 313 . Here, the exposed portion of the metal wiring 313 is the portion where the metal wiring 313 and the odorant-receiving layer 315 are in contact with each other. The vinyl tape 314 can also be an insulator for adjusting the length of the portion where the metal wiring 313 and the odorant-receiving layer 315 are in contact.

The odorant-receiving layer 315 may be in contact with at least part of the first metal wiring 313A and at least part of the second metal wiring 313B. For example, as shown in FIGS. 2 and 3, the odorant-receiving layer 315 may be arranged to fill the area sandwiched between the first metal wiring 313A and the second metal wiring 313B.

When the electrical conductivity of the odorant-receiving layer 315 (that is, the electrical conductivity of the sensor element 31) is low, the distance between the first metal wiring 313A and the second metal wiring 313B is a predetermined distance (eg, 500 μm) or less. is desirable.

The sensor element 31 uses a resin composition that exhibits different electrical conductivity changes over time depending on whether the odorant A is adsorbed or the odorant B different from the odorant A is adsorbed. It is possible to detect and identify various odorants.

[3. Odor sensor 30]
The outline and effects of the odor sensor 30 to which the sensor element 31 is applied will be described below with reference to FIG. FIG. 1 is a block diagram showing an example of the configuration of an odor measurement device 100 having an odor sensor 30 to which a sensor element 31 is applied. Incidentally, in the sensor element 31 shown in FIG. 1, the illustration of the vinyl tape 314 is omitted for the sake of simplification.

The odor sensor 30 includes a sensor element 31 that detects an odorant, a constant current source 32 (power source), and a voltmeter 33 (measuring device).

A lead wire W connects the first metal wiring 313A and the second metal wiring 313B of the sensor element 31 . FIG. 1 shows an example in which the lead wire W is provided with a constant current source 32 and a voltmeter 33 .

A constant current source 32 is a power supply for supplying power to the sensor element 31 . A constant current source 32 supplies a constant current (for example, a direct current of 1 mA) to the sensor element 31 through a lead wire.

The voltmeter 33 measures the potential difference generated between the first metal wiring 313A and the second metal wiring 313B when the constant current supplied from the constant current source 32 is supplied to the odorant-receiving layer 315 .

The odor sensor 30 may further include a housing 34, although this is not an essential component. The housing 34 is a container capable of enclosing air containing an odorant. If housing 34 is provided, sensor element 31 is installed in housing 34 .

The housing 34 has an inlet 341 for introducing odorants and an outlet 342 for discharging air containing odorants. The introduction of the odorant may be performed by inserting a filter paper P or the like soaked with the odorant into the housing 34 through the introduction port 341, or by introducing air containing the odorant into the housing 34 through the introduction port 341. It may be done by inserting The housing 34 is a container for containing air containing a predetermined concentration (for example, 200 ppm) or more of odorants.

An airflow generating fan 35 may be arranged at the outlet 342 of the housing 34, although it is not essential. The airflow generating fan 35 is for generating an airflow inside the housing 34 and discharging the gas inside the housing 34 to the outside of the housing 34 through the exhaust port 342 .

Note that the odor sensor 30 may include a constant voltage source (power source) (not shown) as an alternative to the constant current source 32 and an ammeter (measuring device) (not shown) as an alternative to the voltmeter 33 . In this case, the constant voltage source functions as a power source for supplying power to the sensor element 31, and applies a constant voltage to the sensor element 31 through the lead wire. On the other hand, the ammeter measures the current value flowing between the first metal wiring 313A and the second metal wiring 313B when a constant voltage is applied to the odorant-receiving layer 315 .

The odor sensor 30 outputs a measured value indicating a temporal change in electrical conductivity of the sensor element 31 before and after the odorant is adsorbed on the sensor element 31 . This makes it possible to detect and identify various odorants.

[4. Odor measuring device 100]
When various odorants are adsorbed on the sensor element 31, the odor sensor 30 described above can output changes in electrical conductivity of the sensor element 31 over time for each odorant. If this odor sensor 30 is applied, the electrical conductivity of the sensor element 31 changes with time when the odorant A is adsorbed on the sensor element 31, and the sensor element 31 when the odorant B is adsorbed. A comparison can be made with the change in the electrical conductivity of the element 31 over time. Based on such a comparison result, the odor measuring device 100 capable of estimating the odorant adsorbed on the sensor element 31 can be realized.

Furthermore, the odor measuring device 100 can estimate the odorant with high accuracy by using the estimation model 22 generated by machine learning. The estimation model 22 is learning including a combination of a measurement value measured when each of a plurality of odorants is adsorbed to at least one sensor element 31 and identification information unique to the odorant that gave the measurement value. can be generated using data for

The outline and effect of the odor measuring device 100 to which the odor sensor 30 is applied will be described below. The odor measuring device 100 is a device for estimating an odorant adsorbed on the sensor element 31 from a change in electrical conductivity occurring in the sensor element 31 to which the resin composition described above is applied.

First, the configuration of an odor measurement device 100 according to an embodiment of the present invention will be described using FIG. FIG. 1 is a block diagram showing an example of the configuration of an odor measurement device 100. As shown in FIG.

As shown in FIG. 1, the odor measuring device 100 includes an estimating device 10 and an odor sensor 30. As shown in FIG.

(Estimation device 10)
The estimating device 10 is a device that estimates an odorant detected by the odor sensor 30 . The estimating device 10 is, for example, a computer, and includes a CPU and memory (not shown). The estimating device 10 is communicably connected to the odor sensor 30 . Specifically, the estimation device 10 estimates the odorants by analyzing the measured values obtained from the odor sensor 30 . The configuration of the estimation device 10 will be described later.

<Generation of estimation model 22>
Next, the configuration of the odor measurement device 100 that performs the process of generating the estimation model 22 used for estimating the odorant and the process of generating the estimation model 22 will be described with reference to FIGS. 4 and 5. FIG.

The estimation model 22 is a combination of measured values measured by the voltmeter 33 when each of a plurality of odorants is adsorbed on at least one sensor element, and identification information specific to the odorants that give the measured values. It is generated by machine learning using learning data including Here, the identification information unique to the odorant may be, for example, the name of the odorant, the CAS number, the chemical formula, or the like.

(Configuration of estimation device 10 (generation of estimation model 22))
FIG. 4 is a functional block diagram showing an example of the configuration of the odor measuring device 100. As shown in FIG. For convenience of explanation, members having the same functions as the members explained with reference to FIG.

As shown in FIG. 4 , the estimation device 10 includes an input section 15 , a control section 1 and a storage section 2 .

The input unit 15 is for receiving various input operations from the user, and may be a keyboard, mouse, touch panel, or the like, for example.

The control unit 1 includes a measured value acquisition unit 11 (acquisition unit), a change pattern analysis unit 12 (analysis unit), a learning control unit 13 and an estimation model generation unit 14 .

The measured value acquisition unit 11 acquires measured values from the voltmeter 33 . The measured value acquiring unit 11 also uses the acquired measured values to calculate a value (for example, a resistance value, an impedance, etc.) indicating the electrical conductivity of the sensor element 31 . The measured value acquiring unit 11 may acquire measured values from the voltmeter 33 at predetermined time intervals (for example, 0.1 second intervals).

The change pattern analysis unit 12 analyzes changes over time in the electrical conductivity of at least one sensor element 31 . The change pattern analysis unit 12 uses the resistance value calculated by the measurement value acquisition unit 11 to calculate a value indicating the amount of change in electrical conductivity of the sensor element 31 due to the adsorption of the odorant. The change pattern analysis unit 12 generates data indicating a change pattern indicating a temporal change in the calculated amount of change in electrical conductivity. When the generated change pattern is a known odorant, the change pattern analysis unit 12 associates the generated change pattern with identification information specific to the known odorant, and stores the generated change pattern in the change pattern database 21 (learning data). may be stored.

The learning control unit 13 reads the change pattern database 21 from the storage unit 2 and controls generation of the estimation model 22 by machine learning. Here, the change pattern database 21 includes a combination of measured values measured when a plurality of odorants are adsorbed on the sensor element 31 and identification information specific to known odorants that gave the measured values. database. The learning control unit 13 inputs the change pattern read from the change pattern database 21 to the estimation model generation unit 14 . In addition, the learning control unit 13 compares the identification information of the odorant corresponding to the change pattern input to the estimation model generation unit 14 and the estimation result output from the estimation model generation unit 14, and performs correction according to the comparison result. The instruction is output to the estimation model generation unit 14 .

The estimated model generation unit 14 generates the estimated model 22 by a machine learning algorithm using the variation patterns stored in the variation pattern database 21 . The estimation model generation unit 14 may be configured to generate the estimation model 22 using a known supervised machine learning algorithm. Machine learning algorithms applicable to the estimation model generator 14 include, for example, the k-nearest neighbor method, logistic regression, support vector machine, random forest, and neural network.

(Process for generating estimated model 22)
Specific processing performed by each unit of the control unit 1 will be described below with reference to FIG. FIG. 5 is a flowchart showing an example of the flow of processing in which the estimation device 10 generates the estimation model 22. As shown in FIG.

First, the measured value acquisition unit 11 acquires the voltage value V0 measured by the odor sensor 30 before the filter paper P soaked with the odorant is inserted into the housing 34, and calculates the resistance value R0 (step S11). .

On the other hand, the input unit 15 receives input such as the name of a known odorant immersed in the filter paper P inserted into the housing 34 (step S12). The process of step S12 may be performed before step S11.

Next, the measured value acquiring unit 11 acquires the voltage value V measured by the odor sensor 30 immediately after the filter paper P soaked with the known odorant is inserted into the housing 34, and calculates the resistance value R. (step S13).

Subsequently, the change pattern analysis unit 12 uses the resistance value R0 and the resistance value R to calculate R/R0 (step S14). R/R0 is a value indicating the amount of change in electrical conductivity of the sensor element 31 due to adsorption of a known odorant. Note that the change pattern analysis unit 12 may calculate R−R0 instead of R/R0. The change pattern analysis unit 12 stores the change pattern of R/R0 over time in the change pattern database 21 in association with the input name of the known odorant (step S15).

If the change pattern is not stored for the predetermined type of existing odorant (NO in step S16), that is, if the data used for machine learning is still insufficient, the process returns to step S11.

When the change pattern is stored for the predetermined type of existing odorant (YES in step S16), the learning control unit 13 reads out the change pattern for the known odorant stored in the change pattern database 21. , is input to the estimation model generation unit 14 . The estimated model generating unit 14 generates the estimated model 22 by a machine learning algorithm using the change patterns stored in the change pattern database 21 (step S17).

The estimated model generating unit 14 stores the estimated model 22 generated by predetermined machine learning in the storage unit 2 (step S18).

Although the estimation device 10 generates the estimation model 22 in the examples shown in FIGS. 4 and 5, the present invention is not limited to this. For example, an external computer different from the estimation device 10 and having the same functions as the learning control unit 13 and the estimation model generation unit 14 is provided with the same data as the change pattern database 21 to create the estimation model 22. may

<Estimation of odorants>
Next, the configuration of the odor measuring device 100a for estimating the odorant using the estimation model 22 and the estimation process will be described with reference to FIGS. 6 and 7. FIG.

(Configuration of estimation device 10a (execution of estimation processing))
FIG. 6 is a functional block diagram showing an example of the configuration of the odor measuring device 100a. For convenience of explanation, members having the same functions as the members explained with reference to FIGS. 1 and 4 are denoted by the same reference numerals, and the explanation thereof will not be repeated.

As shown in FIG. 6, the estimation device 10a includes a control unit 1a, a storage unit 2a, and an output unit 18. Here, FIG. 6 shows a configuration example when the estimation device 10 shown in FIG. 4 is used for odor substance estimation processing. That is, the estimation device 10 shown in FIG. 4 and the estimation device 10a shown in FIG. 6 may be computers having the same hardware configuration.

The output unit 18 is for presenting the estimation results to the user, and may be, for example, a display, a speaker, a lamp, or the like.

The control unit 1 a includes a measured value acquisition unit 11 (acquisition unit), a change pattern analysis unit 12 (analysis unit), an estimation unit 16 and an output control unit 17 .

The estimating unit 16 uses the estimating model 22 to estimate the odorant from the analysis results obtained by analyzing the measured values obtained from the odor sensor 30 .

The output control unit 17 controls the output unit 18 to output the estimation result.

(estimation process)
Specific processing performed by each unit of the control unit 1a will be described below with reference to FIG. FIG. 7 is a flowchart showing an example of the flow of processing for estimating an odorant by the estimating device 10a.

First, the measured value acquiring unit 11 acquires the voltage value V0 measured by the odor sensor 30 before the filter paper P soaked with the odorant is inserted into the housing 34, and calculates the resistance value R0 (step S1). . The resistance value R0 is preferably 200-1000Ω, more preferably 250-900Ω, and most preferably 300-800Ω.

Next, the measured value acquisition unit 11 acquires the voltage value V measured by the odor sensor 30 immediately after the filter paper P soaked with the unknown (that is, the target of estimation) odorant is inserted into the housing 34. and the resistance value R is calculated (step S2).

Subsequently, the change pattern analysis unit 12 uses the resistance value R0 and the resistance value R to calculate R/R0 (step S3).

Next, the estimation unit 16 estimates an unknown odorant from the change pattern of R/R0 over time based on the estimation model 22 (step S4).

The output control unit 17 controls the output unit to output the estimation result (step S5).

<Embodiment 2>
Although the odor sensor 30 having one sensor element 31 has been described in the above embodiment, the odor sensor 30 may have two or more sensor elements 31 . For example, the odor sensor 30b may include sensor elements 31 and 31b in which the resin composition used for the odorant-receiving layer 315 is different from each other. This will be described with reference to FIG. FIG. 8 is a block diagram showing an example configuration of an odor measurement device 100b according to another embodiment of the present invention. For convenience of explanation, members having the same functions as the members explained with reference to FIG.

For example, an odor measuring device 100b shown in FIG. 8 includes odor sensors 30, 30b and an estimating device 10b. The odor sensor 30b includes a sensor element 31 and a sensor element 31b. The odorant-receiving layer 315 of the sensor element 31 and the odorant-receiving layer 315b of the sensor element 31b use different resin compositions. may

The estimating device 10b may be a computer having the same configuration as the estimating devices 10 and 10a. The estimating device 10b obtains a first measurement value measured by the voltmeter 33 when a constant current is supplied from the constant current source 32 to the sensor element 31, and A second measured value measured by the voltmeter 33b is obtained and analyzed.

By providing a plurality of sensor elements whose odorant-receiving layers are made of resin compositions having different odorant-adsorbing properties, the odor measurement device 100b can simultaneously estimate a plurality of odorants. In addition to the sensor element according to one embodiment of the present invention, a sensor element containing no surfactant (B) in the odorant-receiving layer may be used in combination.

Further, if the odor measuring device 100b is used, for each known odorant, a first change pattern indicating a change in electrical conductivity of the sensor element 31 and a second change pattern indicating a change in the electrical conductivity of the sensor element 31b can be obtained. It is possible to obtain The estimated model 22 may be generated by machine learning using both the first change pattern and the second change pattern. Since the odor measurement device 100b estimates odorants using the estimation model 22 generated in this way, each odorant can be identified more precisely.

<Example of realization by software>
The control blocks (especially the control unit 1) of the estimation devices 10, 10a, and 10b may be realized by logic circuits (hardware) formed in an integrated circuit (IC chip) or the like, or may be realized by software. .

In the latter case, the estimating devices 10, 10a, 10b comprise computers that execute instructions of programs, which are software that implements the respective functions. This computer includes, for example, one or more processors, and a computer-readable recording medium storing the above program. In the computer, the processor reads the program from the recording medium and executes it, thereby achieving the object of the present invention. As the processor, for example, a CPU (Central Processing Unit) can be used. As the recording medium, a "non-temporary tangible medium" such as a ROM (Read Only Memory), a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. In addition, a RAM (Random Access Memory) or the like for developing the above program may be further provided. Also, the program may be supplied to the computer via any transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. Note that one aspect of the present invention can also be implemented in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

The present invention is not limited to the above-described embodiments, but can be modified in various ways within the scope of the claims, and can be obtained by appropriately combining technical means disclosed in different embodiments. is also included in the technical scope of the present invention.

EXAMPLES The present invention will be further described below with reference to Examples and Comparative Examples, but the present invention is not limited to these. Hereinafter, unless otherwise specified, % means % by weight, and part means part by weight.

<Preparation of higher alcohol ethylene oxide adduct> (Examples 1 to 13)
A reaction vessel equipped with a stirring blade, a stirrer, a nitrogen inlet, an outlet and an ethylene oxide inlet was equipped with a heating/cooling device, a thermometer, a pressure gauge, a temperature controller and a nitrogen inlet pipe. This reaction vessel has a sealing structure that allows the inside of the reaction vessel to be decompressed or pressurized. Under a nitrogen atmosphere, the amount of higher alcohol (dried with 30 parts of molecular sieves 3A for 2 hours) shown in Table 1 and 3 parts of potassium hydroxide as a catalyst were charged, followed by nitrogen replacement under reduced pressure. The temperature was raised to 160° C., and the amount of ethylene oxide shown in Table 1 was added dropwise for reaction while adjusting the flow rate so that the pressure inside the reaction vessel was 0.5 MPa (G). Stirring was continued for 1 hour after completion of dropping, and then the temperature was lowered to room temperature to obtain higher alcohol ethylene oxide adducts (NS-1) to (NS-13).

<Preparation of slurry>
Fluorine resin (A), surfactant (B), conductive carbon (C) and N-methylpyrrolidone as solvent (D) are weighed into a polypropylene container in the amounts shown in Tables 2 and 3, and the mixture is got The mixture was stirred at 2000 rpm for 60 minutes using a rotation/revolution mixer (ARE-310 manufactured by THINKY Co., Ltd.) to obtain a slurry. The slurry was used as a resin composition for forming an odorant-receiving layer. In Tables 2 and 3 below, the "(A)/(B) ratio" means the weight ratio of the fluororesin (A) and the surfactant (B).

Among the fluororesins (A) shown in Tables 2 and 3, polyvinylidene fluoride (PVDF) is manufactured by Sigma-Aldrich, vinylidene fluoride-propylene hexafluoride copolymer (PVDF-HFP) and chlorotrifluoroethylene- Ethylene copolymer (ECTFE) manufactured by Solvay, tetrafluoroethylene-hexafluoropropylene copolymer (FEP) manufactured by DuPont, and polychlorotrifluoroethylene (PCTFE) manufactured by Daikin were used.

Among surfactants (B) shown in Table 2, nonionic surfactants (NS-1) to (NS-13) prepared by the above method were used. Other than that, those commercially available from Tokyo Kasei Kogyo Co., Ltd. were used.

The conductive carbon materials (C) shown in Tables 2 and 3 are SUPER-C65 manufactured by MTI Corporation, VGCF-H manufactured by Showa Denko Co., Ltd., Denka Black manufactured by Denka Co., Ltd., and Ketjen Black. EC-300J and KETJEN BLACK EC-600JD manufactured by Lion Specialty Chemical Co., Ltd. were used.

<Production of sensor element> (Examples 14 to 158, Comparative Examples 1 to 5)
A seal substrate containing a set of two metal wires was cut out from a seal substrate (ICB-073, manufactured by Sunhayato Co., Ltd.) having a plurality of metal wires with a gap width of 500 μm. The cut out seal substrate was further cut so that the metal wiring had a length of 3.5 cm.

The cut seal substrate was pasted on a glass plate with double-sided tape so that the metal wiring was up. In addition, the excess portion of the metal wiring was masked with vinyl tape so that the length of the exposed portion of the metal wiring was 3.0 cm. Then, each slurry prepared by the above method was applied to the exposed portion of the metal wiring using a bar coater (No. 4). After the coating, it was dried for 3 hours in a smooth wind dryer heated to 100°C. After drying and cooling to room temperature, the metal wiring provided with the odorant-receiving layer was peeled off from the glass plate, and the sensor elements (E-1) to (E-158) and the comparative sensor element (E'-1) were obtained. ) ~ (E'-5) were obtained.

<<Evaluation of Resin Composition and Sensor Element>> (Examples 159 to 293, Comparative Examples 6 to 10)
The resin composition can be evaluated by comparing the data obtained from the sensor element (E) and the comparative sensor element (E').

<Measurement method>
A housing equipped with an inlet for introducing a specimen (odorant) and a fan for creating an airflow so that the specimen spreads uniformly was prepared. Among the sensor elements (E-1) to (E-135) and the comparative sensor elements (E'-1) to (E'-5) with soldered lead wires for extracting the terminals to the outside, the evaluation target A sensor element was installed in the housing.

A 1 mA constant current power source and a voltmeter for measuring the voltage applied to both terminals of the lead wire were attached to the end of the lead wire taken out from the housing, and the measured value of the voltmeter was recorded by a computer.

A filter paper was immersed in the sample so that the concentration of the sample in the housing was 200 ppm, and the filter paper was inserted through the introduction port. Measurement was started immediately after inserting the filter paper. After 60 seconds from the start of the measurement, the fan was rotated again for 60 seconds, and the measurement was performed while discharging the steam inside the housing to the outside. The voltage measurement was performed at intervals of 0.1 seconds. Moreover, the measurement was repeated 100 times under the same conditions.

Hexane, ethyl acetate, methanol, diethyl carbonate, or toluene was used as the specimen.

<Evaluation method>
Using the voltage measured at each time and the current value of 1 mA supplied from the constant current power source, the electrical resistance R was calculated from Ohm's law. The resistance R0 was measured in advance before the sample was introduced, and R/R0 was calculated. The measured R0 values are listed in Table 4. Table 4 also shows the value of Rmax/R0 (%), where Rmax is the maximum value of R during measurement.

The responsiveness of the sensor element to each analyte with time changes of R/R0 at 0.1 second intervals was analyzed using the k-nearest neighbor method. For the sensor elements (E-1) to (E-135) or the comparative sensor elements (E'-1) to (E'-5), the above measurements are performed, and 100 corresponding to each example and each comparative example The measurement data for a total of 500 times (times×5 samples) was randomly divided so that the number of learning data:the number of test data=80:20, and a classifier (learning model) based on the k-nearest neighbor method was created for the learning data. The accuracy rate when classifying the test data by the classifiers of each example and comparative example is used as the performance index of the sensor element, and it can be determined that the higher the accuracy rate, the higher the performance as the sensor element. For the creation of the classifier and the calculation of the accuracy rate in each example, the data obtained by the corresponding sensor element (E) and the comparative sensor element ( E′) were used for both.

<Evaluation results>
The evaluation results are shown below.

Comparing Examples 159 to 293 with Comparative Examples 6 to 10, the surfactant (E'-1) to (E'-5), which do not contain the surfactant (B), are more surfactant The data obtained using the sensor elements (E-1) to (E-135) containing the agent (B) are compared to the data obtained using the same fluororesin (A) (E'-1) to (E'- The accuracy rate was improved by combining with the data obtained using 5). It can be said that the use of the surfactant (B) improved the odor discrimination performance of the odorant-receiving layer and the sensor element.

Also, the value of Rmax/R0 tended to decrease as the content of the conductive carbon material (C) increased. Since the smaller the value of Rmax/R0, the larger the measurement error, the larger the value of Rmax/R0, the better.

<<Evaluation of the content of the conductive carbon material (C)>> (Examples 297 to 306)
By changing the content of the conductive carbon material (C), the value of R/R0 immediately after the start of measurement can be controlled. R/R0 was measured using each sensor element (E-136) to (E-145) in the same manner as described above. After the end of the measurement, R/R0 was extracted one second after the start of the measurement and listed in Table 6. Table 6 also shows the value of Rmax/R0 (%), where Rmax is the maximum value of R during measurement.

From Table 6, as the conductive carbon material (C) increases, the value of R/R0 increases one second after the start of measurement, and this is remarkable when the content of the conductive carbon material (C) is 30% by weight or more. It can be said that the content of the conductive carbon material (C) is preferably 30% by weight or more because the larger R/R0 one second after the start of measurement leads to a shorter measurement time and a reduction in power consumption. When the content of the conductive carbon material (C) was 75% by weight or less, the adhesion of the conductive carbon material (C) to the seal substrate was better, and it was more suitable for the sensor element (E).

On the other hand, as described above, as the content of the conductive carbon material (C) increases, the value of Rmax/R0 decreases and the measurement error increases. Therefore, from the viewpoint of keeping the accuracy rate by keeping the measurement time and measurement error small, the content of the conductive carbon material (C) is preferably 25 to 75% by weight, more preferably 30 to 65% by weight. %, most preferably 35-55% by weight.

<<Evaluation of odor sensor and odor measuring device>>
The evaluation of the odor sensor and the odor measuring device is based on the accuracy rate and the comparison sensor element (E'- 1) to (E'-5) can be used to compare the accuracy rate when odors are identified.

<Measurement method>
A housing equipped with an inlet for introducing the sample (odor) and a fan for creating an airflow so that the sample spreads uniformly was prepared. Sensor elements (E-1) to (E-135) with soldered lead wires for extracting terminals to the outside are stored in a housing (F), a sensor element for comparison (E'-1) to A comparative odor sensor (F') in which (E'-5) was housed in a housing was used.

A 1 mA constant current power source and a voltmeter for measuring the voltage applied to both terminals of the lead wire were attached to the end of the lead wire taken out from the housing, and the measured value of the voltmeter was recorded by a computer.

A filter paper was immersed in the sample so that the concentration of the sample in the housing was 200 ppm, and the filter paper was inserted through the introduction port. Measurement was started immediately after inserting the filter paper. After 60 seconds from the start of the measurement, the fan was rotated again for 60 seconds, and the measurement was performed while discharging the steam inside the housing to the outside. The voltage measurement was performed at intervals of 0.1 seconds. Moreover, the measurement was repeated 100 times under the same conditions.

As samples, d-limonene, bornan-2-one, cis-3-hexenol, β-phenylethyl alcohol, citral, L-carvone, γ-undecalactone, eugenol, and linalyl acetate were used. All specimens were manufactured by Tokyo Kasei Kogyo Co., Ltd.

<Evaluation method>
Using the voltage measured at each time and the current value of 1 mA supplied from the constant current power supply, the electrical resistance R was calculated from Ohm's law. The resistance R0 was measured in advance before the sample was introduced, and R/R0 was calculated.

The responsiveness of the sensor element to each analyte with time changes of R/R0 at 0.1 second intervals was analyzed using the k-nearest neighbor method. The odor sensor (F) and the comparative odor sensor (F') were subjected to the above measurements, and the measurement data for each sensor was measured 100 times x 9 samples = 900 times in total. Number of learning data: Number of test data = 80:20 , and a classifier (learning model) based on the k-nearest neighbor method was created for the learning data. The accuracy rate when classifying the test data with the classifiers of each example and comparative example is used as the performance index of the odor sensor, and it can be determined that the higher the accuracy rate, the higher the performance as a sensor.

<Evaluation results>
As a result of the above measurement, the accuracy rate of the odor sensor (F) was 92%, and the accuracy rate of the comparative odor sensor (F') was 41%. It can be said that the odor sensor using the sensor element has good odor discrimination performance.

<<Evaluation when using a specimen that is a mixture>> (Examples 304 to 438, Comparative Examples 11 to 15)
<Method for preparing a specimen that is a mixture>
In addition to the single sample, a mixture sample was prepared by the following method. Menthol, benzaldehyde, ethyl acetate, vanillin, hexanal, ethanol, pentyl valerate, linalool, and 2-propanol were each prepared in a desiccator so as to have a gas concentration of 200 ppm, and used as a mixture raw material. Then, a 500 mL two-neck eggplant flask equipped with a three-way cock and septum rubber was vacuumed and then sealed. The following volumes were taken from each mixture raw material with a syringe and injected through the rubber septum of a sealed two-neck eggplant flask.
Mixture sample 1:
Menthol (150 mL)
Benzaldehyde (150 mL)
Ethyl acetate (150 mL)
Mixture sample 2:
Vanillin (150 mL)
Hexanal (150 mL)
Ethanol (150 mL)
Mixture sample 3:
Pentyl valerate (150 mL)
Linalool (150 mL)
2-propanol (150 mL).

<Method of measurement involving a specimen that is a mixture>
Similarly to Examples 159 to 293 and Comparative Examples 6 to 10, among sensor elements (E-1) to (E-135) and comparative sensor elements (E'-1) to (E'-5), evaluation The target sensor element was installed in the housing. A constant current power supply, a voltmeter, and a computer were similarly arranged.

20 mL of the mixture sample prepared by the above method was collected with a syringe, and the mixture sample was injected from the introduction port. Measurements were started immediately after injection of the mixture sample. After 60 seconds from the start of the measurement, the fan was rotated again for 60 seconds, and the measurement was performed while discharging the steam inside the housing to the outside. The voltage measurement was performed at intervals of 0.1 seconds. Moreover, the measurement was repeated 100 times under the same conditions.

<Evaluation method when using a specimen that is a mixture>
Using the data obtained by the above method and the data obtained from the measurement performed on the above-mentioned single specimen, evaluate by the same method as the evaluation method performed on the above-mentioned single specimen, and determine the correct answer. rate was calculated. That is, evaluation was performed by the following methods.

Using the voltage measured at each time and the current value of 1 mA supplied from the constant current power source, the electrical resistance R was calculated from Ohm's law. The resistance R0 was measured in advance before the sample was introduced, and R/R0 was calculated.

The responsiveness of the sensor element to each analyte with time changes of R/R0 at 0.1 second intervals was analyzed using the k-nearest neighbor method. For the sensor elements (E-1) to (E-135) or the comparative sensor elements (E'-1) to (E'-5), the above measurements are performed, and 100 corresponding to each example and each comparative example The measurement data for a total of 800 times (times×8 samples) was randomly divided so that the number of learning data:the number of test data=80:20, and a classifier (learning model) based on the k nearest neighbor method was created for the learning data. The accuracy rate when classifying the test data by the classifier of each example and comparative example is used as the performance index of the sensor element, and it can be determined that the higher the accuracy rate, the higher the performance as a sensor. For the creation of the classifier and the calculation of the accuracy rate in each example, the data obtained by the corresponding sensor element (E) and the comparative sensor element ( E′) were used for both.

<Evaluation results when using a specimen that is a mixture>
Comparing Examples 304 to 438 with Comparative Examples 11 to 15, the surfactant (E'-1) to (E'-5), which do not contain the surfactant (B), are more surfactant The data obtained using the sensor elements (E-1) to (E-135) containing the agent (B) were compared with the comparative sensor elements (E'-1) to (E'- The accuracy rate was improved by combining with the data obtained using 5). It can be said that this effect is due to the change in the odorant adsorption characteristics of the sensor element due to the use of the surfactant (B).

INDUSTRIAL APPLICABILITY The present invention is useful as an odor identification sensor for medical use, gas detection, agriculture, and other industries and daily life. For example, farmers can use smell sensors to determine the maturity of fragrant crops and manage the optimal harvest timing. Moreover, it is also possible to convert the odors of products such as foods and cosmetics into data using an odor sensor, thereby supporting the improvement of efficiency in product development and the stabilization of quality.

10, 10a, 10b Estimation device 11 Measurement value acquisition unit (acquisition unit)
12 Change pattern analysis unit (analysis unit)
16 estimation units 30, 30b odor sensors 31, 31b sensor elements 32, 32b constant current source (power source)
33, 33b voltmeter (measuring instrument)
100, 100a, 100b odor measuring device 313A first metal wiring 313B second metal wiring 315, 315b odorant receiving layer

Claims (8)

A resin composition for forming an odorant-receiving layer,
Containing fluororesin (A), surfactant (B) and conductive carbon material (C),
The content of the conductive carbon material (C) is 25 to 75% by weight with respect to the total 100% by weight of the fluororesin (A), the surfactant (B) and the conductive carbon material (C). There is a resin composition. The resin composition according to claim 1, wherein the surfactant (B) has an HLB value of 8-18. 3. The resin composition according to claim 1, wherein the weight ratio [(A)/(B)] of the fluororesin (A) and the surfactant (B) is 1.0 to 4.0. A sensor element comprising an odorant-receiving layer containing the resin composition according to any one of claims 1 to 3 , a first metal wiring, and a second metal wiring,
the first metal wiring and the second metal wiring are spaced apart,
The sensor element, wherein the odorant-receiving layer is in contact with at least part of the first metal wiring and at least part of the second metal wiring. at least one sensor element according to claim 4 ;
a power source for powering the sensor element;
and a measuring device that outputs a measured value indicative of the electrical conductivity of the odorant-receiving layer of the sensor element powered by the power supply. 6. The odor sensor according to claim 5 , wherein said power supply supplies a constant current or applies a constant voltage to said at least one sensor element. An odor measuring device comprising the odor sensor according to claim 5 or 6 and an estimating device,
The estimation device is
an acquisition unit that acquires the measured value from the measuring device;
an analysis unit that analyzes changes over time in electrical conductivity of the at least one sensor element;
an estimation unit that estimates an odorant based on the estimation model,
The estimation model includes a measurement value measured by the measuring device when each of a plurality of odorants is adsorbed on the at least one sensor element, and identification information unique to the odorant that gave the measurement value. An odor measuring device generated by machine learning using learning data including combinations. 8. A control program for causing a computer to function as the odor measuring device according to claim 7 , the control program for causing the computer to function as the acquiring section, the analyzing section, and the estimating section.

Images