Electron nose containing laser-inducing graphene with embedded gas-activated particles and method for manufacturing the same

Abstract

This invention provides an electronic nose. [Solution] An electronic nose is provided, comprising a substrate containing a carbon precursor and a gas-active substance precursor, and an olfactory sensor array including a plurality of olfactory sensors formed by irradiating the substrate with lasers of different parameters, wherein each olfactory sensor includes a laser-induced graphene layer formed from the carbon precursor by laser irradiation of one parameter, and gas-active particles formed from the gas-active substance precursor and embedded in the laser-induced graphene layer, and the plurality of olfactory sensors react to different types of gaseous substances.

Description

[Technical Field]

[0001] The present invention relates to an electronic nose and a method for manufacturing the same, and more particularly to an electronic nose and a method for manufacturing the same, which includes laser-guided graphene containing gas-active particles formed by a simple method of irradiating with lasers having different parameters, and which includes multiple olfactory sensors capable of reacting with various types of gaseous substances. [Background technology]

[0002] The digitalization of human senses has spurred innovation across various technologies and dramatically changed daily life. However, the digitalization of the sense of smell remains a challenging area.

[0003] An electronic nose is a device that utilizes the principle of detecting and identifying odors using sensors, based on the human nose's ability to smell. Such electronic noses hold immense potential in various fields, including environmental monitoring, food safety, medicine, and the perfume industry.

[0004] Specifically, it is known that olfactory receptors in the human body do not bind to olfactory molecules in a one-to-one manner. Instead, olfactory receptors transmit information about odors through a combination coding process in which odor molecules bind to various types of olfactory receptors. Through this mechanism, humans are known to have over 400 types of olfactory nerves and be able to distinguish between approximately 1 trillion different odors. Therefore, the development of electronic nose technology attempts to achieve the digitalization of olfaction by mimicking the same function and form as human olfaction. Conventionally, to manufacture electronic noses that mimic human olfaction, technologies such as electronic nose sensor array manufacturing technology using individual sensors, electronic nose manufacturing technology using metal oxide deposition, electronic nose manufacturing technology using carbon compound substrates, and single gas sensor manufacturing technology using laser-guided graphene substrates have been utilized.

[0005] In particular, the electronic nose sensor array manufacturing technology using individual sensors is known to involve configuring individual gas sensors, each with different reactivity, in parallel to mimic the mechanism of olfaction, and predicting the type of gas by analyzing the reaction patterns obtained from each sensor.

[0006] Furthermore, the electron nose manufacturing technology using metal oxide deposition is known to be advantageous for miniaturization and sensor diversification by integrating the electron nose, which was manufactured using individual gas sensors, onto a substrate on which a metal oxide sensor array was created using a semiconductor patterning (exposure-deposition) process.

[0007] Furthermore, carbon compound-based electronic nose manufacturing technology is known as a technique for manufacturing gas sensor arrays based on highly reactive carbon compounds (such as rGO (Reduced Graphene Oxide)) to enable the sensors to operate at room temperature.

[0008] In particular, the single-gas sensor manufacturing technology for laser-induced graphene substrates using the carbon compound-based technology described above is widely used in research and development of single-gas sensors targeting single gas species, thanks to the large surface area and abundant functional groups of laser-induced graphene. The laser-induced graphene-based technology is known to primarily utilize hybrid materials with functional catalysts (such as VOx, MOF, and MoS2) rather than using laser-induced graphene alone.

[0009] However, the electronic nasal sensor array manufacturing technology using individual sensors consists of a collection of individual gas sensors, each with a different reactivity, in order to mimic the olfactory mechanism. The individual gas sensors used here are large in size, which limits the integration and diversification of sensors.

[0010] In particular, the electron nose manufacturing technology using metal oxide deposition involves complex exposure and vacuum deposition processes for patterning, which not only makes manufacturing time-consuming and costly, but is also known to require a high-temperature (>150°C) operating environment due to the operating mechanism of the metal oxide gas sensor.

[0011] On the other hand, carbon compound-based electronic nose manufacturing technology can operate at room temperature, and for electronic nose manufacturing using gas sensors based on carbon compounds (such as rGO) that allow for relatively free reactivity tuning, methods are employed to introduce particles synthesized by solvent thermal synthesis or other methods using an ex-situ method such as drop casting. However, the aforementioned ex-situ method is known to have limitations in mass production due to large deviations (variations) between sensors caused by aggregation development and manual manufacturing problems, as well as the long time required for manufacturing.

[0012] Furthermore, laser-guided graphene-based single-gas sensor manufacturing technology has only been studied primarily for single-channel applications, and the process of manufacturing laser-guided graphene arrays for application in the production of electronic noses for the classification (identification) of various gases has not yet been achieved. [Prior art documents] [Patent Documents]

[0013] [Patent Document 1] Korean Patent No. 10-1852074 (Publication Date: 2018.04.25) [Overview of the Initiative] [Problems that the invention aims to solve]

[0014] The present invention aims to provide an electronic nose that includes an olfactory sensor array capable of reacting to various types of gaseous substances by utilizing laser-guided graphene-based sensor manufacturing technology.

[0015] Furthermore, the present invention aims to provide a method for manufacturing an electronic nose that allows substrate manufacturing and sensor array manufacturing to be performed in-situ, and that enables the production of an olfactory sensor array in a short time. [Means for solving the problem]

[0016] An electronic nose according to an embodiment includes a substrate containing a carbon precursor and a gas active substance precursor, and an olfactory sensor array including a plurality of olfactory sensors formed by irradiating the substrate with lasers having different mediation variables respectively. Each of the olfactory sensors includes a laser-induced graphene layer formed from the carbon precursor by laser irradiation with one mediation variable, and gas active particles formed from the gas active substance precursor and incorporated in the laser-induced graphene layer. The plurality of olfactory sensors are characterized by reacting to different types of gas substances.

[0017] A method for manufacturing an electronic nose according to an embodiment includes the steps of preparing a mixed solution containing a gas active substance precursor and a carbon precursor, forming a substrate using the mixed solution, and irradiating the substrate with lasers having different mediation variables respectively to generate an olfactory sensor array including a plurality of olfactory sensors. Each of the olfactory sensors includes a laser-induced graphene layer formed from the carbon precursor by laser irradiation with one mediation variable, and gas active particles formed from the gas active substance precursor and incorporated in the laser-induced graphene layer. The plurality of olfactory sensors are characterized by reacting to different types of gas substances.

Brief Description of the Drawings

[0018] [Figure 1] It is a configuration diagram showing an electronic nose including an olfactory sensor array in which gas active particles are incorporated in a porous laser-induced graphene layer according to an embodiment. [Figure 2] It is a flowchart showing a method for manufacturing an electronic nose according to an embodiment. [Figure 3] It is a diagram showing the results of analyzing the characteristics of a CeLIG sensor formed by the method according to the example. [Figure 4] It is a diagram showing the results of analyzing the chemical and physical diversity of a CeLIG sensor array formed by the method according to the example. [Figure 5] It is a diagram showing the response data of a CeLIG sensor array formed by the method according to the example to different odor substances. [Figure 6]This figure shows the odor classification learning process of the machine learning algorithm-based CeLIG sensor array according to the embodiment. [Figure 7] This figure shows the olfactory prediction data and performance comparison results of the machine learning algorithm-based CeLIG sensor array according to the embodiment. [Figure 8] This figure shows the stress analysis results and flexibility test results of the CeLIG sensor array according to the example. [Modes for carrying out the invention]

[0019] The specific structural or functional descriptions of embodiments of the concept of the present invention disclosed herein are merely illustrative for the purpose of illustrating embodiments of the concept of the present invention, and embodiments of the concept of the present invention can be carried out in various forms and are not limited to those described herein.

[0020] Embodiments of the concept of the present invention can be modified in various ways and may take on various forms; therefore, embodiments are illustrated in the drawings and described in detail herein. However, this is not intended to limit embodiments of the concept of the present invention to any particular disclosure, but rather to include modifications, equivalents, or substitutions that fall within the spirit and technical scope of the present invention.

[0021] Terms such as "first" or "second" may be used to describe multiple components, but such terms should be interpreted solely for the purpose of distinguishing one component from others. For example, the first component may be named the second component, and similarly, the second component may also be named the first component.

[0022] When it is mentioned that one component is “linked” or “connected” to another component, it must be understood that it is directly linked or connected to the other component, but that other components may exist in between. On the other hand, when it is mentioned that one component is “directly linked” or “directly connected” to another component, it should be understood that there are no other components in between. Expressions describing the relationship between components, such as “between,” “immediately between,” or “directly adjacent to,” must be interpreted in the same way.

[0023] A singular expression includes plural expressions unless the context clearly indicates otherwise. In this specification, terms such as “includes” or “has” indicate the presence of features, figures, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood not to presuppose the existence or addition of one or more other features, figures, steps, actions, components, parts, or combinations thereof.

[0024] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as those generally understood by a person of ordinary skill in the art to which this embodiment belongs. Commonly used, predefined terms should be interpreted as having the meaning consistent with their meaning in the context of the relevant art, and not as ideal or overly formal unless expressly defined herein.

[0025] The embodiments will be described in detail below with reference to the attached drawings. However, the scope of the patent application is not limited or restricted by these embodiments. The same reference numerals shown in each drawing indicate the same component.

[0026] Figure 1 is a diagram showing an electronic nose 1 that includes an olfactory sensor array 20 in which gas-activated particles 23 are embedded within a porous laser-induced graphene layer 21, according to an embodiment.

[0027] Referring to Figure 1, the electronic nose 1 according to this embodiment includes a substrate 10, an olfactory sensor array 20, and a contact pad 30. Here, the olfactory sensor array 20 may include a plurality of olfactory sensors 20-1 to 20-n.

[0028] The substrate 10 may contain a carbon precursor and a gas-activated substance precursor.

[0029] Specifically, the substrate 10 is formed using a mixture containing the carbon precursor and the gas-activated substance precursor.

[0030] The carbon precursor, as a carbon-containing polymer, can form a laser-induced graphene layer 21 by laser irradiation, as described later. The laser-induced graphene layer 21 includes multiple graphene layers and can have a porous structure by irradiating the carbon precursor with a laser. The laser-induced graphene layer 21 has a large surface area, is rich in functional groups on its surface, can bind to gas-active substances, and can form an olfactory sensor with high gas detection sensitivity. The odor substance reacts with the laser-induced graphene layer 21, thereby generating an electrical signal.

[0031] The carbon precursors include polydimethylsiloxane (PDMS), polyimide (PI), polyurethane (PU), cellulose, cellulose acetate, acetate butyrate, cellulose acetate propionate, polymethyl methacrylate (PMMA), polymethyl acrylate (PMA), polyacrylic copolymer, polyvinyl acetate copolymer, polyvinyl acetate (PVAc), polyvinylpyrrolidone (PVP), and polyvinyl alcohol (polymethyl The material may also contain alcohol (PVA), polyfurfuryl alcohol (PPFA), polystyrene (PS), polystyrene copolymer, polyethylene oxide (PEO), polypropylene oxide (PPO), polyethylene oxide copolymer, polypropylene oxide copolymer, polycarbonate (PC), polyvinyl chloride (PVC), polycaprolactone, polyvinylidene fluoride, polyvinylidene fluoride copolymer, polyamide, polyethylene terephthalate (PET), polyethylene (PE), or mixtures thereof. In particular, the carbon precursor may be polyimide.

[0032] The gas-active substance precursor can form gas-active particles by laser irradiation, as described later. The gas-active particles refer to metal oxides that act as oxidation-reduction catalysts to promote reactions with gaseous substances, i.e., odorous substances. The gas-active particle precursor contains components for generating the gas-active particles by laser irradiation.

[0033] The gas-activated particles may include cerium oxide (CeO2), zinc oxide (ZnO), tin oxide (SnO2), indium oxide (In2O3), zirconium oxide (ZrO2), tungsten oxide (WO3), iron oxide (Fe2O3), copper oxide (CuO), or a mixture thereof.

[0034] The olfactory sensor array 20 includes a plurality of olfactory sensors 20-1 to 20-n formed on the substrate 10. The olfactory sensors 20-1 to 20-n play a role in generating and transmitting signals when exposed to gas.

[0035] The olfactory sensors 20-1 to 20-n can be formed by irradiating a carbon precursor substrate, which is mixed with gas-active substance precursors, with lasers having different parameters so that they react to different gaseous substances (odor substances). Each of the olfactory sensors 20-1 to 20-n can be formed by irradiating specific regions of the substrate that are separated from each other with a laser. When the substrate is irradiated with a laser, a laser-induced graphene layer 21 can be formed because the substrate contains carbon precursors. The gas-active particles may be embedded in the laser-induced graphene layer 21. The olfactory sensors 20-1 to 20-n, produced by irradiating with lasers having different parameters, have different physical and chemical properties, and thus can be formed to react to different odor substances. This enables olfactory combination coding for classifying odor molecules.

[0036] In the olfactory sensor constituting the electronic nose according to this embodiment, a heterojunction is formed in which the gas-activated particles 23 act as an n-type semiconductor and the laser-induced graphene layer 21 acts as a p-type semiconductor. Here, the gas-activated particles 23 may be embedded on the laser-induced graphene layer 21 in the form of nanometer-sized particles.

[0037] The contact pad 30 constituting the electronic nose according to this embodiment includes wiring that electrically connects the olfactory sensors 20-1 to 20-n on the upper part of the substrate 10 so that it can output the resistance change due to the oxidation-reduction reaction of the olfactory sensors 20-1 to 20-n by gaseous substances.

[0038] The wiring includes a signal line 31 that outputs the resistance change due to the oxidation-reduction reaction of gaseous substances in the olfactory sensors 20-1 to 20-n, and a grounding line 33 that connects the olfactory sensors 20-1 to 20-n to a common ground.

[0039] The electronic nose 1 may further include an odor prediction model learned by an artificial intelligence algorithm to analyze the signal output patterns of each of the plurality of olfactory sensors 20-1 to 20-n and predict the gas that each of the plurality of olfactory sensors reacts to, i.e., the type of odorous substance and the concentration of the odorous substance.

[0040] On the other hand, Figure 2 is a flowchart showing the manufacturing method of the electronic nose according to the embodiment.

[0041] Referring to Figure 2, the method for manufacturing an electronic nose according to the embodiment includes the steps of: preparing a mixed solution S10; forming a substrate 10 using the mixed solution S20; generating an olfactory sensor array equipped with a plurality of olfactory sensors S30; and forming a contact pad S40.

[0042] First, step S10, which involves preparing a mixed solution, includes the step of mixing a gas-active substance precursor with a solvent to form a gas-active substance precursor solution.

[0043] The solvent can be any of the common types of solvents used to dissolve metal precursors. For example, the solvent may include at least one of methylpyrrolidone (1-Methyl-2-pyrrolidinone, NMP), gasoline, benzene, toluene, hexane, thinner, methylene chloride, ether, acetone, methyl ethyl ketone (MEK), methanol, ethanol, n-propanol, isopropanol, n-butanol, isobutanol, t-butanol, and pentanol.

[0044] This step may include a step of dissolving the carbon precursor to form a liquid-phase carbon precursor solution. In step S10, which prepares the mixed solution of the gas-active substance precursor and the carbon precursor, the gas-active substance precursor solution and the carbon precursor solution may be mixed and stirred to form the mixed solution of the gas-active substance precursor and the carbon precursor.

[0045] In step S20, where the substrate is formed using the mixed solution, the mixed solution containing the gas-activated substance precursor and the carbon precursor is applied to a substrate layer or mold, and then cured to form the substrate. The substrate layer may be a glass substrate or a silicon substrate, etc.

[0046] As an example, in this step, the mixed solution may be spin-coated onto a substrate layer, followed by soft baking, to form a substrate containing a gas-activated substance precursor and a carbon precursor.

[0047] In particular, this step may be configured to form a substrate 10 having a predetermined thickness on the substrate layer by performing a unit cycle at least once, which includes the step of spin-coating the mixed solution onto the upper part of the substrate layer to form a mixed layer, and the step of curing the mixed layer.

[0048] Step S30, which involves forming the olfactory sensor array, involves irradiating the mixed layer of the gas-active substance precursor and carbon precursor with lasers having different output characteristics to generate an olfactory sensor array 20 formed from olfactory sensors 20-1 to 20-n, each containing a laser-guided graphene layer 21 with gas-active substance nanoparticles 23 that react with different gaseous substances.

[0049] In this step, the parameters, including at least one of the laser's specific power, pulses per inch (PPI), and laser scanning velocity, can be adjusted to differ in order to form the plurality of olfactory sensors, so that each of the plurality of olfactory sensors reacts to a different type of gaseous substance. When olfactory sensors are formed on a substrate by irradiating it with lasers that have different parameters as described above, each olfactory sensor can react to gaseous substances with different odors because it has different physicochemical properties from each other.

[0050] Step S40, which involves forming the contact pad, is the step of forming a contact pad 30 that includes a signal line 31 and a ground line 33 for the signal output of the olfactory sensors 20-1 to 20-n. The contact pad 30 may be coated with a conductive paste to ensure stable connection with the probe tip.

[0051] The present invention will be described in more detail below with reference to examples.

[0052] The examples presented are merely specific illustrations of the present invention and are not intended to limit the technical scope of the present invention.

[0053] <Examples> (1) Fabrication of laser-induced graphene layers (LIG) containing CeOx nanoparticles A cerium precursor solution was prepared by mixing 720 mg of Ce(NO3)·6H2O with 8.3 ml of NMP (1-Methyl-2-pyrrolidinone) and dissolving it through sonication. The cerium precursor solution was mixed with 41.7 ml of carbon precursor solution and prepared a mixed solution by shaking and alternating stirring at room temperature. Here, the carbon precursor solution was prepared using liquid-phase PI (Polyimid) to produce a total of 50 ml of mixed solution. The mixed solution was spin-coated onto a glass substrate covered with Kapton tape at 2000 rpm for 30 seconds, and then soft-baked at 100°C. The unit cycle including the spin-coating and soft-baking was repeated 5 times to form a substrate with a predetermined thickness, and the final curing was performed at 200°C.

[0054] Next, in the laser process for manufacturing laser-guided graphene (CeLIG) with embedded cerium nanoparticles, the laser scanning speed was fixed at 50% of the maximum scanning speed, and the process was performed by adjusting the specific power and pulses per inch (PPI). In this process, a 25W laser cartridge was used, and the focus was adjusted to the substrate surface.

[0055] Next, to manufacture the olfactory sensor, vector scanning was applied to pattern 1 mm long filament-type channels.

[0056] The contact pads were patterned using rasterization and coated with Ag paste to ensure stable contact with the probe tip.

[0057] As a result, a total of 10 different CeLIG olfactory sensors (hereinafter referred to as "channels (Ch1~Ch10)") were generated by varying the laser process parameters (Ch.1: 5% relative power and 200 PPI, Ch.2: 5% relative power and 500 PPI, Ch.3: 10% relative power and 200 PPI, Ch.4: 10% relative power and 500 PPI, Ch.5: 15% relative power and 100 PPI, Ch.6: 15% relative power and 200 PPI, Ch.7: 15% relative power and 500 PPI, Ch.8: 20% relative power and 100 PPI, Ch.9: 20% relative power and 200 PPI, Ch.10: 20% relative power and 500 PPI) to form an olfactory sensor array.

[0058] <Example Test> (1) Characteristic analysis The properties of CeLIG produced by the method described in the examples were analyzed, and the results are shown in Figure 3. CeLIG characterization was performed using field emission scanning electron microscopy (FE-SEM), ultra-high resolution TEM, high-resolution Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), and ultraviolet photoelectron spectroscopy (UPS).

[0059] Figure 3 shows the results of analyzing the characteristics of a CeLIG sensor formed by the method described in the embodiment.

[0060] Specifically, the SEM image of CeLIG shown in Figure 3(a) exhibits large porous properties. Figure 3(b) shows a TEM image of CeLIG. Figures 3(c) to 3(f) show EDS-TEM images of CeLIG, respectively, and are maps of electrons (ZC), cerium (Ce), carbon (C), and oxygen (O). Figure 3(g) is a high-magnification TEM image of CeLIG showing the CeO2 lattice structure. Figure 3(h) shows the uniform size distribution of CeOx nanoparticles contained in LIG, Figure 3(i) is the Raman spectrum, and Figure 3(j) is the XPS O1s spectrum of CeLIG. Figure 3(k) is an energy band diagram of CeLIG showing typical pn junction formation between CeOx and LIG.

[0061] As can be seen from the scanning electron microscope (SEM) image shown in Figure 3(a), the CeLIG produced by the method described in the example exhibited a typical alternating porous structure of LIG. The formation of many large pores (macropores) was predicted to be due to localized explosions and release of gas species during the photothermal decomposition of the cerium (Ce)-doped polyimide film under laser irradiation. Such pores are important for increasing the surface area, which can be conveniently applied in the field of gas detection. As can be seen from the transmission electron microscope (TEM) image shown in Figure 3(b) and the energy-dispersive X-ray spectroscopy (EDS) images shown in Figures 3(c) to 3(f), the CeLIG described in the example showed a uniform distribution of CeO2 nanoparticles throughout the laser-induced graphene layer. As can be seen from the high-resolution TEM image shown in Figure 3(g), characteristic lattice spacings of 0.32 nm and 0.27 nm corresponding to the (111) plane and (200) plane of CeO2, respectively, were clearly observed. In particular, the average size (Diameter (nm)) of CeO2 nanoparticles was 6.47 ± 0.79 nm (Figure 3(h), n=150), indicating an extremely narrow size distribution and no signs of aggregation.

[0062] Furthermore, the homogeneity of CeO2 nanoparticles within the CeLIG was presumed to be due to the rapid heating / cooling process during nucleation and the influence of the laser-induced graphene layer (LIG) acting as a functional backbone. It was confirmed that the rapid heating and cooling due to the photothermal effect of the laser provides optimal conditions for nanoparticle synthesis. It was confirmed that the photothermal effect induced by laser irradiation provides optimal conditions for nanoparticle synthesis. Here, it was determined that rapid heating causes simultaneous nucleation and growth within the irradiated area, resulting in a uniform nanoparticle size distribution (particle size distribution).

[0063] Furthermore, it was confirmed that minimizing the movement of nanoclusters during the laser heat treatment process promotes nanoparticle growth, thereby forming a backbone with defects for strong surface bonding with the nanoclusters.

[0064] As can be seen in Figure 3(i), the laser-induced graphene layer (LIG) is shown as a strong D peak in the Raman spectrum of the graphite material, confirming that it has abundant defect sites and functional groups.

[0065] Furthermore, as confirmed by the X-ray photoelectron spectroscopy (XPS) results shown in Figure 3(j), the O1s spectrum was found to have a strong CO peak (532.1 eV) and a C=O peak (533.5 eV) along with a Ce-O peak (529.9 eV). These peaks suggest the presence of hydroxyl and carboxyl groups on the carbon surface of LIG. Therefore, it was determined that the surface-functionalized LIG carbon backbone helps stabilize and immobilize cerium nanoclusters and prevents aggregation. CeO2 nanoparticles produced individually by the in-situ process bonded more strongly to the backbone compared to the ex-situ introduction method, and improved stability was confirmed in application areas such as gas detection, catalytic reactions, and mechanical motion. Therefore, it was confirmed that small-sized CeO2 nanoparticles can be uniformly doped into LIG using a laser process.

[0066] The energy band diagram shown in Figure 3(k) confirms the gas detection performance of CeOx-doped CeLIG. The band gap energy is 3.1 eV, and n-type CeO2 with a work function Φ2 of 3.3 eV forms a junction with p-type narrow-bandgap LIG with a work function Φ1 of 4.5 eV. The built-in potential creates a depletion layer, and this pn heterojunction at the interface improves the detection performance of gas sensors based on metal oxides and various carbon derivatives. The heterogeneous structure with uniformly dispersed nanoparticles maximizes the number of depletion layers and amplifies the ability to convert chemical changes into electrical signals. Figure 3(k) shows the changes in the depletion layer and work function (Φ), electron depletion, and hole depletion after exposure to an odorous substance following the formation of the pn heterojunction. In other words, it was confirmed that when the adsorbed target gas is a reducing gas, the electron concentration of CeOx increases, strengthening the pn junction, which in turn expands the depletion layer and consequently increases resistance. Conversely, it was confirmed that when an oxidizing gas is adsorbed, the energy band shifts in the opposite direction, decreasing resistance. Therefore, it was confirmed that even a slight change in charge carrier concentration can be effectively converted into a large sensor reaction, thereby improving the odor detection performance.

[0067] (2) Chemical and structural diversification of olfactory sensors by changing laser-mediated parameters Olfactory sensor arrays with different chemical and physical properties can be combined using a coding scheme to generate identifiable reaction patterns for various odor substances, thereby improving classification accuracy.

[0068] To achieve this, the specific power output and PPI, which are parameters of the laser, were adjusted. Here, the PPI parameter represents the pulse stacking density along the laser movement direction.

[0069] Figure 4 shows the results of analyzing the chemical and physical diversity of CeLIG sensor arrays formed by the method described in the example.

[0070] Specifically, Figure 4(a) shows a heat map of the specific power output and electrical resistance against PPI during laser irradiation. Figure 4(b) shows the Raman spectra (fixed 100 PPI) of CeLIGs with different specific power outputs. Figure 4(c) shows I D / I G The ratio is shown. Also, Figure 4(d) shows I as a function of the laser irradiation parameter. 2D / I G The ratios are shown. Figure 4(e) shows the normalized atomic percentages of carbon, oxygen, and cerium due to the change in specific power and PPI during laser irradiation. Figure 4(f) shows the XPS Ce 3d spectrum (fixed 200 PPI) with increasing specific power. Figure 4(g) shows the Ce spectrum with changing specific power (fixed 200 PPI). 3+ and Ce 4+ The area percentage of the separated u''' peaks, which indicate changes in the composition, is shown. Figures 4(h) to 4(m) show SEM images of CeLIG processed at various specific powers at a fixed 200 PPI (Figure 4(h) is (k): 5%, Figure 4(i) is (l): 10%, and Figure 4(j) is (m): 15%). The scale bar is 200 μm for Figures 4(h) to 4(j) and 50 μm for Figures 4(k) to 4(m).

[0071] As shown in Figure 4(a), the change in electrical resistance as a response to various laser parameters was evaluated, and it was confirmed that as the laser specific power (%) and PPI increased, the laser fluence increased, the graphitization of the PI substrate was promoted, and the resistance (Ω) decreased.

[0072] Figures 4(b) to 4(d) show the results of Raman spectroscopic analysis for characterizing the carbon derivatives. The G peak and the D peak indicate that the polymer precursor has been converted into a material rich in sp2 carbon. The presence of the 2D peak, which is the second overtone of the D peak, functions as a fingerprint for graphene materials. As confirmed from Figures 4(b) to 4(d), by increasing the specific output (Power) of the laser from 5% to 20% and increasing the PPI from 100 to 500, the ratio of I D / I G decreased from a maximum of 1.51 to a minimum of 0.88. Conversely, the ratio of I 2D / I G indicating a thinner sp2 carbon layer and the quality of graphene increased from a minimum of 0.54 to a maximum of 0.77. Thus, it was predicted that the increased laser fluence induces local high-temperature and high-pressure conditions, promotes carbonization, graphitization, and exfoliation, and generates thinner and higher-quality sp2 carbon layers. The above results are consistent with the results of the resistance heat map (see Figure 4(a)).

[0073] It was confirmed by examples that various sensors with different reaction characteristics according to the degree of graphene exfoliation, surface oxidation, and defect density can be fabricated by changing the laser irradiation conditions (i.e., mediation variables) without performing complex wet chemical treatments.

[0074] Figure 4(e) is a three-dimensional mapping graph of the atomic percentages of carbon, oxygen, and cerium calculated using XPS surface analysis. Each element is normalized for visualization (Normalized atomic percentage shown in Figure 4(e)). It was confirmed that for the CeLIG according to the example, when irradiated with a laser at room temperature, the output rate increases, and as a result, the carbon surface is oxidized. Consequently, the carbon content decreased while the oxygen content increased. This change was significant under the 200 PPI condition. Also, it was confirmed that the atomic ratio of cerium varied from 0.19% to 0.89% depending on the laser mediation variable.

[0075] It was confirmed that changes in laser-mediated parameters affect not only the carbon backbone of CeLIG but also the diversity of CeOx nanoparticles. Therefore, the chemical environment and atomic state of CeOx catalyst nanoparticles in CeLIG were analyzed by XPS Ce3d scanning, and the results are shown in Figure 4(f).

[0076] Specifically, the oxidation state of cerium significantly influences its energy structure, band gap, and the activity of redox catalysts, thereby diversifying the gas reaction pattern.

[0077] Specifically, Ce 3+ The amount of can directly improve the catalytic activity of cerium oxide. As shown in Figure 4(f), the Ce 3d core-level spectrum is Ce 3d 3 / 2 and Ce 3D 5 / 2 The components were divided into the corresponding "u" and "v" groups, with a spin-orbit separation of 18.6 eV. Each component was further divided into multiple peaks. The peaks of V (882.8 eV), v'' (887.8 eV), v''' (898.6 eV), u (901.4 eV), u'' (907.4 eV), and u''' (916.9 eV) were Ce 4+ While the peaks v0 (880.5eV), v' (885.4eV), u0 (900.1eV), and u' (904.2eV) are due to Ce 3+ This indicates the state in which mixed atomic states of Ce(III) and Ce(IV) coexist within CeOx. This signifies high redox activity of Ce. 3+ The quantification of Ce 3+ This is achieved by calculating the relative area percentage of the u''' peak, which is unrelated to the quantity of (the area of ​​the u''' peak is Ce 3+ (It is inversely proportional to the amount of ). Therefore, Ce 3+ The ratio was found to increase with increasing laser power (Figure 4(g)). This development is due to the generation of oxygen vacancies in the metal oxide at high temperatures as the laser process power increases, and Ce is used to maintain charge neutrality. 3+This is thought to be because it facilitated the formation of [something]. Therefore, changing the laser parameter allows Ce to [something] to form. 3+ A gradient was observed in the amount of [the substance], and as a result, it was determined that the catalytic activity of CeOx could be adjusted (controlled). Also, Ce 3+ As the amount of [substance] increased, the Ce 4f orbital was partially filled, narrowing the band gap and altering the depletion region (see Figure 3(k)). This change diversified the gas reaction pattern, confirming that a sensor array could be designed.

[0078] Furthermore, it was confirmed that laser processing can impart morphological diversity to LIGs.

[0079] As can be seen from the SEM images in Figures 4(h) to 4(m), the formation of a large amount of anisotropic fibers was further promoted as the laser power rate increased. However, at low power conditions, an isotropic porous structure was generated. It was confirmed that the rapid generation of gas by laser irradiation rapidly increased the local pressure and simultaneously caused carbonization expansion and delamination, converting the LIG from a porous structure to a fibrous structure.

[0080] (3) Classification of olfactory substances using a CeLIG-based electronic nose Figure 5 shows the response data of a CeLIG sensor array formed by the method described in the embodiment to different odor substances.

[0081] Specifically, Figure 5(a) shows the structural formulas of nine types of odor molecule molecules used in the perfume industry (1-heptanol, 2,3,5-trimethylpyrazine, 2-ethylphenol, d-limonene, eugenol, geraniol, octanal, cis-3-hexenol, and ethanol). Figure 5(b) shows the dynamic detection curves of 10 CeLIG olfactory sensors for D-limonene under a 1000 sccm gas flow (10 minutes exposure, 50 minutes recovery). Figure 5(c) is a polar coordinate plot showing the various reaction patterns of 10 olfactory sensors for various odor substances. All measurements were performed at room temperature without the use of heaters.

[0082] By varying the laser-mediated parameters (specific power and PPI), we utilized the chemical and physical diversity introduced into CeLIG to fabricate an artificial olfactory sensor array. The fabricated olfactory sensor array was then analyzed using nine representative olfactory substances (geraniol (Floral: GRL), d-limonene (Citrus: LIM), octanal (Waxy: OCT), eugenol (Spicy: EUG), 1-heptanol (Fruity: HEP), 2-ethylfencol (Earthy: 2EF), cis-3-hexenol (Green: CHX), 2,3,5-trimethylpyrazine). The sensors were tested against nuts (TP) and ethanol (EtOH) (see Figure 5(a)). These scents are nontoxic and widely used in perfumes, food additives, and fragrances due to their distinctive aromas. As shown in Figure 5(b), repeated exposure to these scents changes the resistance (Ω) of each sensor, allowing for the calculation of the reaction value. Therefore, 10 types of characteristics for each scent could be extracted using the sensor array.

[0083] The response value of the olfactory sensor was derived by applying the following formula (1).

[0084] Reactivity = (R - R0) / R0X 100(%) (1)

[0085] Here, R is the resistance of the olfactory sensor after exposure to the target gas, and R0 is the initial resistance value of the olfactory sensor.

[0086] As shown in Figure 5(c), the average reaction value could be confirmed through the radar graph. As a result, a unique reaction pattern that clearly distinguishes the aroma was confirmed. Furthermore, unlike conventional metal oxide semiconductor gas sensors that generally require high temperatures (>250°C) or a light source for activation, the LIG-based gas sensor can operate at room temperature, resulting in high energy efficiency, simplified device configuration, and the use of flexible substrates.

[0087] Odor molecules often possess a variety of functional groups and complex structures. For example, such molecules are composed of various combinations of alkyl groups (methyl, ethyl, heptyl, etc.) of different lengths, which mostly act as nonpolar electron donors. Alcohol groups act as strong electron donors, while phenol and ether groups act as weak donors. Aldehydes act as weak acceptors, and pyrazine rings act as intermediate acceptors. Consequently, such molecules have rarely been used in conventional gas sensor research to avoid complexity, and their reaction mechanisms remain largely unknown. In practical applications for classifying various odor molecules, a one-to-one correspondence between gas sensor arrays and reaction mechanisms is inefficient. Therefore, the electronic nose described in the example provides a method for identifying gas-active substances, i.e., odor molecules, using the reaction patterns of multiple olfactory sensors, rather than a one-to-one reaction mechanism between olfactory sensors and odor molecules.

[0088] (4) Data processing of machine learning infrastructure Figure 6 shows the odor classification learning process based on the machine learning algorithm of the CeLIG sensor array according to the embodiment.

[0089] Specifically, referring to Figure 6, the response values ​​of each olfactory sensor were used as features for data analysis. t-SNE (t-distributed stochastic neighbor embedding) analysis was performed using the Python scikit-learn library to reduce multidimensional parameters to 2D. Regression for odor substance type classification and concentration prediction using SVM (Support Vector Machine) was performed using the Python scikit-learn library with the RBF kernel. The training and test sets were split in a 7:3 ratio. Tanimoto coefficients for comparing molecular similarity were calculated using the Python Rdkit library after converting each compound to a simplified molecular input line input system (SMILES) notation.

[0090] Figure 6(a) shows the data processing and learning process for odor substance prediction using a machine learning platform.

[0091] As shown in Figures 6(a) and 5, responses from 10 integrated sensors were collected for various types and conditions of gas species (Measurement). Next, as data preprocessing, the class and concentration of odor molecules were assigned to the dataset (Data Labeling). For machine learning characteristic data, the response values ​​from the 10 sensors were extracted, and dimensionality reduction was performed using t-SNE to transform the data into a 2D space. This unsupervised learning-based data preparation was useful for computation efficiency and visualization because it allowed high-dimensional data with many characteristics to be represented in a reduced dimensional space while preserving the distance between data points. Finally, an SVM algorithm was trained using the labeled data (response data) and characteristic data (Machine Learning) to generate an odor prediction model. The trained odor prediction model was used to classify the type of odorous substance and predict its concentration through regression.

[0092] The scatter plot shown in Figure 6(b) illustrates the t-SNE results derived from response values ​​obtained from the CeLIG integrated olfactory sensor array. The T-SNE algorithm identified the 2D representation that best preserved the distances between data points. Through manifold learning, 10-dimensional Euclidean distances were transformed into 2D coordinates, allowing for clear clustering of each odor molecule class without overlap. This access method provides significant visualization of large datasets that would be difficult to analyze using intrinsic morphological or radial plots (see Figure 5(c)).

[0093] Machine learning was performed using t-SNE data obtained to identify and predict nine distinct ordered molecules as feature data. SVM is a widely used supervised learning method for classification problems, which searches for a hyperplane for classifying data and learns an effective shunt in the feature space using a kernel method. In this example, the SVM model was implemented with a radial basis function kernel to improve its ability to handle nonlinear relationships in the data. The data set was split into 70% for training and 30% for testing. The SVM model was trained on the training set, and its performance was evaluated on the test set. For accurate generalization, the test set was not used for parameter selection or model training.

[0094] Figure 6(b) is a graph showing the result of applying the SVM model to display the decision boundary map of 2D t-SNE coordinates. As can be seen from Figure 6(b), the trained model was able to provide discriminable boundaries for nine types of olfactory substances (GRL, LIM, OCT, EUG, HEP, 2EF, CHX, TP, EtOH) in a dimensionally reduced space, providing clear visual information.

[0095] Figure 6(c) is a graph showing the confusion matrix between the actual and predicted classes for nine types of olfactory substances, based on the learned SVM classification results. As can be seen from Figure 6(c), the trained model demonstrated high predictive accuracy (Accuracy (%)) of 97.46% on the training set and 97.06% on the test set, confirming that successful generalization is possible without overfitting or underfitting.

[0096] Figure 6(d) shows estimates of shunt uncertainty in a probabilistic format. In Figure 6(d), the subplots represent specific olfactory substance classes, and the color gradient of the heatmap indicates the confidence level of the model predicting each class. Separated, narrow areas on the map indicate high probabilities and high classification confidence for the shunt. Conversely, broad areas indicate low probabilities and regions with uncertainty or low classification confidence.

[0097] Therefore, as can be seen from Figure 6(d), the trained shunts exhibited concentrated, high-probability regions with little overlap, confirming a low risk of false positives or false negatives.

[0098] Figure 7 shows the olfactory prediction data and performance comparison results based on the machine learning algorithm of the CeLIG sensor array according to the example. In Figure 7, the concentration of odor substances was predicted using support vector regression from the scikit-learn machine learning library. Odor concentration plays an important role in determining the depth, persistence, and overall perception of the odor.

[0099] Figure 7(a) shows the actual concentration data on the x-axis (True, Gas flow (sccm)) and the predicted data on the y-axis (Prediction, Gas flow (sccm)) as the ratio of carrier gas to base gas increases to 1:3, 1:1, and 3:1. The solid line shows the ideal prediction curve (y=x), and the dotted line shows the linear approximation of the predicted data using regression. The R value was calculated using the residual variance for the test set. 2 The accuracy also showed high values. For 2,3,5-trimethylpyrazine (TP), d-limonene (LIM), and 1-heptanol (HEP), the accuracy was confirmed as 96.34%, 94.44%, and 99.92%, respectively (compared to 99.34%, 99.99%, and 100% for the training set). These results suggest that t-SNE-based data processing provides clear and intuitive visual information, and that machine learning-based models can accurately predict all types and concentrations of odorous substances.

[0100] To quantitatively evaluate the selectivity of the fabricated olfactory sensor array, it was compared with previously reported chemical resistance gas sensors operating at room temperature. Of various calculation methods, the Tanimoto coefficient was selected to measure molecular similarity. Figure 7(b) shows the number of gases and molecular similarity (including studies) used in the selectivity evaluation of reported chemical resistance gas sensors. The molecular similarity used here is the maximum value derived by calculating the Tanimoto coefficient for all possible gas combinations. The results indicate that the fabricated CeLIG olfactory sensor array was able to efficiently distinguish between various types of similar gases (odors). This comparison includes not only electronic noses designed for odor identification, but also single sensors targeting specific gases.

[0101] As can be seen in Figure 7(b), the artificial olfactory platform, which can identify a large number of odor molecules at room temperature, combined with its structural simplicity, low energy consumption, and stability, has been shown to be applicable to extend the artificial olfactory platform to a wide range of application areas, including complex odor molecules, beyond existing targets such as compound 1, NH3, and specific VOCs. [ka]

[0102] (5) Application to flexible devices Figure 8 shows the stress analysis and flexibility test results for CeLIG.

[0103] For reference, flexibility is an important characteristic for applying sensor platforms in various fields, such as patch sensors, smart packaging, or outdoor environments. When designing this sensor platform, even if the mechanical stability is excellent, unstable electrical characteristics will interfere with sensor operation. To ensure reliable performance, it is important to set an operating range in which all mechanical and electrical stability is maintained.

[0104] Referring to Figure 8(a), the operating range for measuring changes in mechanical stress (MPa) and electrical resistance (Ω) during the tensile process must be limited to a deformation (strain) rate of 3% (shown in the blue shaded area), which ensures all mechanical and electrical stability.

[0105] Furthermore, in the case of the superelastic flexible substrate used in the examples, it was confirmed that applying a load lower than the previously applied maximum load resulted in nonlinear and irreversible elastic behavior due to stress softening known as the Mullins effect. To more accurately predict and analyze the mechanical behavior, it was necessary to observe and consider the material behavior by performing load and load removal cycle tests within a specific deformation range. For this purpose, as shown in Figure 8(b), the repeated stress-deformation (SS) characteristics were measured within a preset deformation range of 3%. Next, to improve the accuracy of prediction, conditions and results that minimized changes in SS characteristics were applied to the mechanical behavior analysis.

[0106] Based on the data shown in Figure 8(b), the spatial distribution of internal stresses in the bending was visualized using FEA (Finite Element Analysis) results, and the location where the maximum stress occurred was identified (see Figure 8(c)). Through this analysis, as shown in Figure 8(d), the smallest achievable radius of curvature could be predicted. The maximum stress and maximum deformation rate (maximum strain rate) occurred at the outermost surface of the flexible structure, and the bending occurred at the smallest radius of curvature within the bending. As shown in Figure 8(d), the flexible structure could be designed to operate within an electrically stable deformation range (radius of curvature > 1.7 mm).

[0107] However, the results obtained from this FEA analysis did not take into account fatigue failure that could arise from invisible microcracks due to cyclic loading. Therefore, direct verification by repeated loading tests was necessary. Based on the FEA results integrating the Mullins effect, fatigue tests were performed using small radii of curvature of 2.0 mm and 2.5 mm, which belong to the electrically and mechanically stable range (see Figure 8(e)). The 2.0 mm radius of curvature was within the stable range, and in repeated fatigue tests, the structure maintained operational stability for a limited number of bending cycles. However, beyond 1000 cycles, invisible damage accumulated and propagated, compromising structural stability. Conversely, with a radius of curvature of 2.5 mm, which provided a safer margin within the electrically stable range, the initial resistance value was maintained even after 30,000 cycles of repeated bending. This result indicates that by optimizing flexibility, the device can operate stably without fatigue failure even under repeated bending conditions. Therefore, it was confirmed that the CeLIG platform can be applied to future application areas requiring flexible electronic noses.

[0108] Although embodiments of the present invention have been described in detail above with reference to the drawings, the present invention is not limited to the embodiments described above, and various modifications and variations can be made from the above description by those who are ordinary skill in the art. For example, the described techniques may be performed in a different order than described, and / or the described systems, structures, devices, circuits, and other components may be combined or assembled in a different manner than described, or substituted or replaced by other components or equivalents, and still achieve the desired results.

[0109] Therefore, other embodiments, other embodiments, and those equivalent to the claims described below also fall within the scope of the claims.

Claims

1. A substrate containing a carbon precursor and a gas-active substance precursor, The olfactory sensor array includes a plurality of olfactory sensors formed by irradiating the substrate with lasers of different parameters, respectively. Each of the olfactory sensors comprises a laser-induced graphene layer formed from the carbon precursor by laser irradiation of a single parameter, and gas-active particles formed from the gas-active substance precursor and embedded in the laser-induced graphene layer. The aforementioned multiple olfactory sensors react to different types of gaseous substances, forming an electronic nose.

2. The electronic nose according to claim 1, wherein the parameter includes at least one of the specific power of the laser, the pulses per inch (PPI) of the laser, and the laser scanning velocity.

3. In the olfactory sensor, The electron nose according to claim 1, having a heterojunction structure in which the gas-activated particles act as n-type semiconductors and the laser-induced graphene layer acts as a p-type semiconductor.

4. The aforementioned gas-activated particles are Cerium oxide (CeO ２ ), zinc oxide (ZnO), tin oxide (SnO ２ ), indium oxide (In ２ O ３ ), zirconium oxide (ZrO ２ ), tungsten oxide (WO ３ ), iron oxide (Fe ２ O ３ The electron nose according to claim 1, comprising one or more selected from the group consisting of ), or copper oxide (CuO).

5. The aforementioned carbon precursor is Polydimethylsiloxane (PDMS), polyimide (PI), polyurethane (PU), cellulose, cellulose acetate, acetate butyrate, cellulose acetate propionate, polymethyl methacrylate (PMMA), polymethyl acrylate Acrylate (PMA), polyacrylic copolymer, polyvinyl acetate copolymer, polyvinyl acetate (PVAc), polyvinylpyrrolidone (PVP), polyvinyl alcohol (PVA), polyfurfuryl alcohol (PPFA), polystyrene (PS), polystyrene copolymer, polyethylene oxide (PEO), polypropylene oxide (PPO), polyethylene oxide copolymer, polypropylene oxide copolymer, polycarbonate (PC), polyvinyl chloride (PVC), polycaprolactone, polyvinylidene fluoride, polyvinylidene fluoride copolymer, polyamide, polyethylene terephthalate (polyethylene terephthalate) The electronic nose according to claim 1, comprising one or more selected from the group consisting of terrephthalate (PET) and polyethylene (polyethylene, PE).

6. The electronic nose according to claim 1, further comprising a contact pad including wiring connected to each of the plurality of olfactory sensors to output resistance due to the gas activation of the olfactory sensor array.

7. The electronic nose according to claim 1, further comprising an odor prediction model learned by an artificial intelligence algorithm to analyze the signal output patterns of each of the plurality of olfactory sensors and predict the gaseous substance to which each of the plurality of olfactory sensors reacts and the concentration of the gaseous substance.

8. A step of preparing a mixed solution containing a gas-active substance precursor and a carbon precursor, The steps include forming a substrate using the mixed solution, The process includes the step of irradiating the substrate with lasers of different parameter sizes to generate an olfactory sensor array including a plurality of olfactory sensors, Each of the olfactory sensors comprises a laser-induced graphene layer formed from the carbon precursor by laser irradiation of a single parameter, and gas-active particles formed from the gas-active substance precursor and embedded in the laser-induced graphene layer. The aforementioned multiple olfactory sensors react to different types of gaseous substances. A method for manufacturing an electronic nose.

9. The method for manufacturing an electronic nose according to claim 8, wherein the step of forming a substrate using the mixed solution includes a step of spin-coating the mixed solution onto the upper part of a substrate layer to form a mixed layer, and a step of curing the mixed layer, and the unit cycle is performed at least once to form a substrate having a predetermined thickness.

10. The step of generating the olfactory sensor array includes at least one of the following: the specific power of the laser, the pulses per inch (PPI) of the laser, and the laser scanning velocity, such that each of the plurality of olfactory sensors reacts to a different type of gaseous substance. The method for manufacturing an electronic nose according to claim 8, wherein the mediating variables are adjusted to be different in order to form the plurality of olfactory sensors.

11. The method for manufacturing an electronic nose according to claim 8, wherein the step of generating the olfactory sensor array further includes the step of generating an olfactory sensor array having the plurality of olfactory sensors, and then forming contact pads for signal output of the olfactory sensors.

Images