# Fileset

[biosensors-13-00368 (2).pdf](https://mdr.nims.go.jp/filesets/2ebe1d45-b27c-48e5-a160-ac9073f793a5/download)

## Creator

[Gaku Imamura](https://orcid.org/0000-0002-3130-7190), [Kosuke Minami](https://orcid.org/0000-0003-4145-1118), [Genki Yoshikawa](https://orcid.org/0000-0002-9136-8964)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Repetitive Direct Comparison Method for Odor Sensing](https://mdr.nims.go.jp/datasets/5ca2d2e8-da83-4012-80a3-20372102d8b7)

## Fulltext

Repetitive Direct Comparison Method for Odor SensingCitation: Imamura, G.; Minami, K.;Yoshikawa, G. Repetitive DirectComparison Method for OdorSensing. Biosensors 2023, 13, 368.https://doi.org/10.3390/bios13030368Received: 31 January 2023Revised: 2 March 2023Accepted: 8 March 2023Published: 10 March 2023Copyright: © 2023 by the authors.Licensee MDPI, Basel, Switzerland.This article is an open access articledistributed under the terms andconditions of the Creative CommonsAttribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).biosensorsArticleRepetitive Direct Comparison Method for Odor SensingGaku Imamura 1,2,* , Kosuke Minami 3 and Genki Yoshikawa 3,41 International Center for Materials Nanoarchitectonics (MANA), National Institute for Materials Science (NIMS),1-1 Namiki, Tsukuba 305-0044, Japan2 Graduate School of Information Science and Technology, Osaka University, 1-2 Yamadaoka,Suita 565-0871, Japan3 Research Center for Functional Materials, National Institute for Materials Science (NIMS), 1-1 Namiki,Tsukuba 305-0044, Japan4 Materials Science and Engineering, Graduate School of Pure and Applied Science, University of Tsukuba,1-1-1 Tennodai, Tsukuba 305-8571, Japan* Correspondence: imamura.gaku@nims.go.jp; Tel.: +81-(0)-29-860-4988Abstract: Olfactory sensors are one of the most anticipated applications of gas sensors. To distinguishodors—complex mixtures of gas species, it is necessary to extract sensor responses originating fromthe target odors. However, the responses of gas sensors tend to be affected by interfering gases withmuch higher concentrations than target odor molecules. To realize practical applications of olfactorysensors, extracting minute sensor responses of odors from major interfering gases is required. In thisstudy, we propose a repetitive direct comparison (rDC) method, which can highlight the differencein odors by alternately injecting the two target odors into a gas sensor. We verified the feasibilityof the rDC method on chocolates with two different flavors by using a sensor system based onmembrane-type surface stress sensors (MSS). The odors of the chocolates were measured by therDC method, and the signal-to-noise ratios (S/N) of the measurements were evaluated. The resultsshowed that the rDC method achieved improved S/N compared to a typical measurement. Theresult also indicates that sensing signals could be enhanced for a specific combination of receptormaterials of MSS and target odors.Keywords: gas sensor; olfactory sensor; signal processing; membrane-type surface stress sensor1. IntroductionOlfaction plays an important role in judging the quality of food; for example, onecan detect spoilage of foods and estimate the ripeness of fruit with his/her nose. Inthe food industry, the flavor is one of the most important factors of products [1]. Foodcompanies make great efforts to improve the flavor of their products, while logisticscompanies and retailers maintain and control the food quality, including flavors, duringtransportation and storage. To automate the quality control of flavors, olfactory sensors—also known as electronic noses (e-noses), artificial olfaction, or machine olfaction—havebeen attracting much attention [2]. Numerous studies utilizing olfactory sensors on foodproducts have been reported, such as the identification of flavors and the detection ofripening, spoilage, and adulteration [3,4]. Despite such studies, the number of practicalapplications of olfactory sensors in the food industry is still limited. This is due to technicalproblems of olfactory sensors stemming from the complex nature of odor.As an olfactory sensor is a gas sensor system that mimics human odor perception,an olfactory sensor should consist of two main parts: detection of odor molecules andidentification of an odor [5]. In the mammalian olfactory mechanism, odor moleculesinhaled through the nose are detected by olfactory receptors in the nasal epithelium. Thereare approximately 400 types of olfactory receptors in humans [6]. Each olfactory receptorhas a different affinity for odor molecules and transduces the chemical interactions intoneural signals. The signals are then transmitted to the olfactory bulb, followed by odorBiosensors 2023, 13, 368. https://doi.org/10.3390/bios13030368 https://www.mdpi.com/journal/biosensorshttps://doi.org/10.3390/bios13030368https://doi.org/10.3390/bios13030368https://creativecommons.org/https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://www.mdpi.com/journal/biosensorshttps://www.mdpi.comhttps://orcid.org/0000-0002-3130-7190https://orcid.org/0000-0003-4145-1118https://orcid.org/0000-0002-9136-8964https://doi.org/10.3390/bios13030368https://www.mdpi.com/journal/biosensorshttps://www.mdpi.com/article/10.3390/bios13030368?type=check_update&version=1Biosensors 2023, 13, 368 2 of 11identification at the limbic system. To realize practical olfactory sensors, therefore, it isnecessary to improve both odor detection and identification [7,8]. Recent dramatic progressin artificial intelligence has led to various studies utilizing data science approaches inolfactory sensors [9]. Although such approaches are effective in identifying odors fromsensor data, the data should be collected with an appropriate measurement protocol sothat the data contain information about the odors. In many cases, the concentration oftarget odor molecules (mostly volatile organic compounds, VOCs) is quite low, on the orderof parts per million (ppm; 10−6) or less. Sensing signals for such low-concentration odormolecules are hindered by interfering gases that exist at high concentrations. Figure 1 showsan example of gas sensor measurements of water and chocolates with different flavors. Thesample vapors and carrier gas (cleaned air) were alternately injected into a gas sensor (amembrane-type surface stress sensor (MSS) coated with poly(2,6-diphenyl-p-phenyleneoxide) (PPPO)). The details of the measurements are explained in Section 2. As observedin this example, the sensing output for water vapor, which is a typical interfering gas, ismuch higher than those for the target odors in many cases; odor molecules are present atvery low concentrations and contribute little to the sensor response, while background orinterfering gases strongly affect the sensor response. It should be noted that the sensorresponses reflect the composition of the gases, but the responses are not necessarily relatedto human olfaction. It is inferred from Figure 1 that both eliminating the effect of interferinggases and detecting low-concentration gases with a high signal-to-noise ratio (S/N) areessential in odor sensing.Biosensors 2023, 13, x FOR PEER REVIEW 2 of 11  signals. The signals are then transmitted to the olfactory bulb, followed by odor identifi-cation at the limbic system. To realize practical olfactory sensors, therefore, it is necessary to improve both odor detection and identification [7,8]. Recent dramatic progress in arti-ficial intelligence has led to various studies utilizing data science approaches in olfactory sensors [9]. Although such approaches are effective in identifying odors from sensor data, the data should be collected with an appropriate measurement protocol so that the data contain information about the odors. In many cases, the concentration of target odor mol-ecules (mostly volatile organic compounds, VOCs) is quite low, on the order of parts per million (ppm; 10−6) or less. Sensing signals for such low-concentration odor molecules are hindered by interfering gases that exist at high concentrations. Figure 1 shows an example of gas sensor measurements of water and chocolates with different flavors. The sample vapors and carrier gas (cleaned air) were alternately injected into a gas sensor (a mem-brane-type surface stress sensor (MSS) coated with poly(2,6-diphenyl-p-phenylene oxide) (PPPO)). The details of the measurements are explained in Section 2. As observed in this example, the sensing output for water vapor, which is a typical interfering gas, is much higher than those for the target odors in many cases; odor molecules are present at very low concentrations and contribute little to the sensor response, while background or in-terfering gases strongly affect the sensor response. It should be noted that the sensor re-sponses reflect the composition of the gases, but the responses are not necessarily related to human olfaction. It is inferred from Figure 1 that both eliminating the effect of interfer-ing gases and detecting low-concentration gases with a high signal-to-noise ratio (S/N) are essential in odor sensing.  Figure 1. Sensing outputs of the measurements with the PPPO-coated MSS. Vapors of ultrapure water and two chocolates (“Manjari” and “Fraise” flavors) were used as samples. The sample vapors and the carrier gas (deodorized and dehumidified air) were alternately injected into the MSS with an interval of 5 s. The baselines are corrected so that the signal output at 5 s is 0 mV. The measure-ment protocol follows the normal measurement, which is explained in Section 2. To make reliable olfactory sensors that are applicable in the food industry, it is nec-essary to design a sensing system that is robust to interfering gases in order to detect the slight difference between odors. One such approach is the elimination of interfering gases by filters or traps [10]. Although this approach works well in some cases [11], such filters or traps may eliminate target molecules as well as the interfering gases. The use of a pre-concentrator and micro-gas chromatography are other approaches for detecting low-con-centration VOCs from interfering gases [12–14]. This approach, however, often requires a temperature control unit in the measurement system and a long measurement time, re-sulting in a complicated measurement protocol. Another issue in realizing reliable olfactory sensors is increasing the S/N of the meas-urement system. One of the most common methods for improving S/N is signal averaging, Figure 1. Sensing outputs of the measurements with the PPPO-coated MSS. Vapors of ultrapurewater and two chocolates (“Manjari” and “Fraise” flavors) were used as samples. The sample vaporsand the carrier gas (deodorized and dehumidified air) were alternately injected into the MSS with aninterval of 5 s. The baselines are corrected so that the signal output at 5 s is 0 mV. The measurementprotocol follows the normal measurement, which is explained in Section 2.To make reliable olfactory sensors that are applicable in the food industry, it is nec-essary to design a sensing system that is robust to interfering gases in order to detectthe slight difference between odors. One such approach is the elimination of interferinggases by filters or traps [10]. Although this approach works well in some cases [11], suchfilters or traps may eliminate target molecules as well as the interfering gases. The useof a preconcentrator and micro-gas chromatography are other approaches for detectinglow-concentration VOCs from interfering gases [12–14]. This approach, however, oftenrequires a temperature control unit in the measurement system and a long measurementtime, resulting in a complicated measurement protocol.Another issue in realizing reliable olfactory sensors is increasing the S/N of themeasurement system. One of the most common methods for improving S/N is signalaveraging, which is widely used in many types of sensor applications. However, suchan approach has rarely been applied to gas sensors because it has been difficult to obtainBiosensors 2023, 13, 368 3 of 11reproducible sensor responses from repeated measurements. Gas sensors, including metaloxide sensors, quartz crystal microbalance sensors, and chemiresistive sensors, absorb oradsorb target gases on their sensing units. These types of sensors often suffer from theeffect of residual gases because it takes some time for the absorbed/adsorbed gases todesorb from the sensing units. To obtain reproducible responses with such sensors, theresidual gases must be completely eliminated by purging with carrier gas at each repetition,which takes too long for practical use. One exception to this is optical gas sensors, whichdetect and analyze light passing through a cell filled with a sample gas [15–19]. As theprobe light does not usually have a chemical interaction with gas samples, the signalaveraging can be performed with a repetitive measurement over the light exposure time.In addition, as the cell can be easily cleaned by carrier gas purging, the sensing response isunlikely to be affected by residual gas. In contrast to this exceptional case in optical gassensors, improving S/N by signal averaging remains a challenging task in chemical gassensor measurements.In this study, we have developed a repetitive direct comparison (rDC) method that canhighlight the difference between odors by alternately injecting target odors into a gas sensor.The key feature of this measurement method is that the effects of common interfering gasesare canceled, and the sensor responses based on the difference in VOCs of each samplecan be obtained. This is the most distinctive feature of the rDC method from typical odormeasurements where sensing responses are separately obtained for each sample. As such,sensor responses are weak due to the low concentration of VOCs—typically less thanppm order. We repeated the cycles of odor injections and averaged the sensing outputs toimprove the S/N. Since it has already been demonstrated with MSS that the reproduciblesensing outputs can be obtained after some repetitions even though the receptor layercontains residual gases [20,21], S/N can be improved by signal averaging, which has rarelybeen applied to gas sensor measurements.2. Materials and Methods2.1. Measurement SetupFigure 2a shows the measurement setup for the rDC method. Two samples wereset in the vials, and their headspace gases were injected into the sensors. Unlike typicalodor measurement methods where an odor of a sample and a carrier gas are alternatelydelivered to a gas sensor [22,23], the odors of the two samples are alternately injectedin this method. The rDC method is effective for samples mainly composed of the samebase components but have slightly different odors because the sensing responses highlightthe difference in odors. In this study, two kinds of chocolates were used as samples. Forcomparison, we also measured the odors of the chocolates with a typical measurementprotocol in which the odor and the cleaned air were alternately injected (hereafter, denotedas “normal measurements”). Such measurements were performed by setting the chocolatesin the Sample 1 vial and emptying the Sample 2 vial. The normal measurement of ultrapurewater was performed with the same configuration. The outer air was deodorized anddehumidified for the measurements by passing through the glass vials filled with activatedcarbon pellets and silica gel beads, as shown in Figure 2a. The cleaned air was flown tothe vials for 5 min before the measurements to fill the vials with the cleaned air. By usingthe cleaned air as a carrier gas, the headspace gases of the two vials containing sampleswere delivered to a sensor chamber in which MSS (purchased from NanoWorld AG) coatedwith two different polymers (i.e., PPPO and polystyrene (PS)) were set. The flow rate wascontrolled by a pump and set at 30 mL/min. All the measurements were conducted at 25 ◦C.The repetitive measurements were carried out by switching the flow paths with a valveplaced in front of the sensor. Odors of Sample 1 and Sample 2 were alternately injectedwith an interval of 5 s. The repetitions of the sample odor injections started with Sample 2.The total measurement duration was set at 120 s. To stabilize the device, measurementswithout samples were performed for several hours prior to the experiments.Biosensors 2023, 13, 368 4 of 11Biosensors 2023, 13, x FOR PEER REVIEW 4 of 11  started with Sample 2. The total measurement duration was set at 120 s. To stabilize the device, measurements without samples were performed for several hours prior to the ex-periments.  Figure 2. (a) A schematic illustration of the measurement setup for the rDC method. In this study, MSS was used as a gas sensor. (b,c) Optical microscope images of MSS coated with (b) PPPO and (c) PS. The molecular structures of PPPO and PS are shown on the right of the microscope images. We also conducted normal measurements on Manjari and Fraise chocolates. A meas-urement with the rDC method was performed by setting Manjari and Fraise chocolates in Sample 1 and Sample 2 vials, respectively. This set of measurements was performed 6 times, and the performance was evaluated statistically. As the sensor chamber contains both the PPPO- and the PS-coated MSS, the sensing outputs for the two MSS could be obtained simultaneously. 2.2. Odor Samples Chocolates with different flavors, i.e., “Fraise” and “Manjari” flavors (purchased from Valrhona), were used as odor samples. Although the flavor of chocolate consists of several hundred VOCs [24], the compositions of the vapors from the two kinds of choco-lates are considered to be similar. VOCs from the base material—cacao, milk, and butter—are dominant, while the flavors are additional components. Fraise and Manjari chocolates have strawberry and vanilla flavors, of which representative components are furaneol and vanillin, respectively [25]. Thus, it is expected that only a slight difference originating from the flavors could be seen in sensing responses with the rDC method. 2.3. Membrane-Type Surface Stress Sensors (MSS) Nanomechanical sensors are sensors that detect changes in mechanical properties such as volume, strain, and stress. Since first demonstrated with micro-cantilevers in 1994 [26,27], nanomechanical sensors have been shown as a great sensing platform [28–30]. MSS, which was developed in 2011 by Yoshikawa et al., is one of the nanomechanical sensors [31]. An MSS is a small gas sensor fabricated through a micro-electro-mechanical systems (MEMS) process [31]. With its distinctive structure (shown in Figure 2a), MSS efficiently detects surface stress caused by mechanical deformation of a receptor layer, resulting in high sensitivity (i.e., sub-ppm level gas detection) [32,33]. The receptor layer Figure 2. (a) A schematic illustration of the measurement setup for the rDC method. In this study,MSS was used as a gas sensor. (b,c) Optical microscope images of MSS coated with (b) PPPO and(c) PS. The molecular structures of PPPO and PS are shown on the right of the microscope images.We also conducted normal measurements on Manjari and Fraise chocolates. A mea-surement with the rDC method was performed by setting Manjari and Fraise chocolatesin Sample 1 and Sample 2 vials, respectively. This set of measurements was performed6 times, and the performance was evaluated statistically. As the sensor chamber containsboth the PPPO- and the PS-coated MSS, the sensing outputs for the two MSS could beobtained simultaneously.2.2. Odor SamplesChocolates with different flavors, i.e., “Fraise” and “Manjari” flavors (purchased fromValrhona), were used as odor samples. Although the flavor of chocolate consists of severalhundred VOCs [24], the compositions of the vapors from the two kinds of chocolatesare considered to be similar. VOCs from the base material—cacao, milk, and butter—aredominant, while the flavors are additional components. Fraise and Manjari chocolateshave strawberry and vanilla flavors, of which representative components are furaneol andvanillin, respectively [25]. Thus, it is expected that only a slight difference originating fromthe flavors could be seen in sensing responses with the rDC method.2.3. Membrane-Type Surface Stress Sensors (MSS)Nanomechanical sensors are sensors that detect changes in mechanical propertiessuch as volume, strain, and stress. Since first demonstrated with micro-cantilevers in1994 [26,27], nanomechanical sensors have been shown as a great sensing platform [28–30].MSS, which was developed in 2011 by Yoshikawa et al., is one of the nanomechanicalsensors [31]. An MSS is a small gas sensor fabricated through a micro-electro-mechanicalsystems (MEMS) process [31]. With its distinctive structure (shown in Figure 2a), MSSefficiently detects surface stress caused by mechanical deformation of a receptor layer,resulting in high sensitivity (i.e., sub-ppm level gas detection) [32,33]. The receptor layercoated on the sensing membrane absorbs gaseous molecules to expand. The expansionresults in the deformation of the sensing membrane and induces surface stress, which isBiosensors 2023, 13, 368 5 of 11electrically read by the four piezoresistors. As the piezoresistors compose the Wheatstonebridge circuit, the sensing output (Vout) is obtained via the following formula [31]:Vout =VB4(∆R1R1− ∆R2R2+∆R3R3− ∆R4R4), (1)where VB is the bridge voltage applied to the circuit. R1, R2, R3, and R4 are the resistanceof the piezoresistors, and ∆R1, ∆R2, ∆R3, and ∆R4 are the changes in the resistance of thepiezoresistors. As receptor layers, PPPO and PS were used in this study. PPPO, whichis also known as Tenax®, is typically used as a sorbent for low-concentration VOCs. Thesuperior absorption ability of PPPO is also preferable as a receptor material of nanome-chanical sensors, which we confirmed in our previous study [21]. PS has been widelyused in nanomechanical sensors, and its characteristics as a receptor material are well-studied [20,34]. The molecular structures of PPPO and PS are shown in Figure 2b,c. Thesetwo polymers are hydrophobic properties, which are thought to be preferable for detectingodor molecules [22]. These polymers were dissolved in 1,1,2,2-tetrachloroethane and coatedon MSS with an inkjet spotter. Optical microscope images of the PPPO- and PS-coated MSSare shown in Figure 2b,c. Even though the receptor materials are not uniformly coated,the sensing outputs of MSS are robust to the topography of the receptor layer comparedto cantilever-type sensors [35]. Sensing signals of MSS were read with VB = −1.0 [V] at asampling frequency of 100 Hz. More details of MSS are described in Refs. [31,36].2.4. Signal ProcessingTime-series data of sensing outputs were obtained from the measurements. Thesensing outputs between 55 s and 115 s were used for analysis (yellow shaded area inFigure 3). Averaging of the signal outputs was performed by the following process: thesensing outputs were divided into 6 parts according to the repetitions. The baseline wascorrected for each part so that the sensing output at the start of Sample 1 injection becomes0. After the baseline correction, the sensing outputs of the 6 parts were averaged.Biosensors 2023, 13, x FOR PEER REVIEW 5 of 11  coated on the sensing membrane absorbs gaseous molecules to expand. The expansion results in the deformation of the sensing membrane and induces surface stress, which is electrically read by the four piezoresistors. As the piezoresistors compose the Wheatstone bridge circuit, the sensing output (𝑉 ) is obtained via the following formula [31]: 𝑉 = 𝑉4 Δ𝑅𝑅 − Δ𝑅𝑅 + Δ𝑅𝑅 − Δ𝑅𝑅 , (1) where 𝑉  is the bridge voltage applied to the circuit. 𝑅 , 𝑅 , 𝑅 , and 𝑅  are the re-sistance of the piezoresistors, and Δ𝑅 , Δ𝑅 , Δ𝑅 , and Δ𝑅  are the changes in the re-sistance of the piezoresistors. As receptor layers, PPPO and PS were used in this study. PPPO, which is also known as Tenax®, is typically used as a sorbent for low-concentration VOCs. The superior absorption ability of PPPO is also preferable as a receptor material of nanomechanical sensors, which we confirmed in our previous study [21]. PS has been widely used in nanomechanical sensors, and its characteristics as a receptor material are well-studied [20,34]. The molecular structures of PPPO and PS are shown in Figure 2b,c. These two polymers are hydrophobic properties, which are thought to be preferable for detecting odor molecules [22]. These polymers were dissolved in 1,1,2,2-tetrachloroethane and coated on MSS with an inkjet spotter. Optical microscope images of the PPPO- and PS-coated MSS are shown in Figure 2b,c. Even though the receptor materials are not uni-formly coated, the sensing outputs of MSS are robust to the topography of the receptor layer compared to cantilever-type sensors [35]. Sensing signals of MSS were read with 𝑉 = – 1.0  [V] at a sampling frequency of 100 Hz. More details of MSS are described in Refs. [31,36]. 2.4. Signal Processing Time-series data of sensing outputs were obtained from the measurements. The sens-ing outputs between 55 s and 115 s were used for analysis (yellow shaded area in Figure 3). Averaging of the signal outputs was performed by the following process: the sensing outputs were divided into 6 parts according to the repetitions. The baseline was corrected for each part so that the sensing output at the start of Sample 1 injection becomes 0. After the baseline correction, the sensing outputs of the 6 parts were averaged.  Figure 3. Whole sensing outputs of the measurements with (a) the PPPO-coated MSS and (b) the PS-coated MSS. Baselines are corrected at 0 s. Yellow-shaded areas are used for averaging. The noise levels were evaluated with the sensing outputs of the last 1 s of the repeti-tion. As the sensing outputs can be considered linear in this period, the sensing outputs were fitted with straight lines. The noise levels were calculated from the residuals; the noise level was defined by the root mean square (RMS) of the residuals. Figure 3. Whole sensing outputs of the measurements with (a) the PPPO-coated MSS and (b) thePS-coated MSS. Baselines are corrected at 0 s. Yellow-shaded areas are used for averaging.The noise levels were evaluated with the sensing outputs of the last 1 s of the repetition.As the sensing outputs can be considered linear in this period, the sensing outputs werefitted with straight lines. The noise levels were calculated from the residuals; the noiselevel was defined by the root mean square (RMS) of the residuals.Signal intensity is defined as the difference between the initial and peak sensingoutputs in a cycle, the difference between the signal outputs at 0 s and 5 s.Biosensors 2023, 13, 368 6 of 113. ResultsFigure 3 shows the whole sensing outputs from one set of the six measurements: thenormal measurements of Manjari and Fraise chocolates, the difference between Manjariand Fraise chocolates, and the measurements by the rDC method of Manjari and Fraisechocolates. The baselines were corrected so that the sensing outputs become 0 µV at 0 s.The difference between Manjari and Fraise was calculated by subtracting the Manjarisensing outputs from the Fraise sensing outputs. Figure 3a,b show the results of theMSS coated with PPPO and PS, respectively. The results of normal measurements ofManjari and Fraise shown in Figure 3a are also shown in Figure 1. Clear signal peakscan be seen for Manjari and Fraise chocolates; signal peaks and valleys correspond tothe injection of the odors and the carrier gas (i.e., cleaned air), respectively. As explainedabove, these signals are basically induced by the difference between chocolates and carriergases, reflecting almost all components in each chocolate sample, including major basecomponents. In contrast, lower signal peaks compared to the normal measurements ofManjari and Fraise are observed for the rDC method reflecting the slight difference in odors.For any measurement, the signal shapes for a single cycle of odor injections graduallychange for the first few repetitions because of the sensing mechanism of MSS; sensingoutputs reflect the gas diffusion process in the receptor layer and the changes in viscoelasticproperties of the receptor materials [20,37,38]. After some repetitions, the sensing outputsreach a steady state; for each repetition, almost the same signal appears. The baselineshifted slightly even after the sensing outputs reached the steady state, which may be dueto the temperature shift associated with the thermal heating of the circuit. Such sensingoutputs can be averaged to improve S/N.The effect of the averaging is shown in Figure 4. Blue and red lines represent theraw and averaged sensing outputs, respectively. The results of the MSS coated withPPPO and PS correspond to Figure 4a,b, respectively. Raw sensing outputs shown inFigure 4a,b were trimmed from the last cycle in the yellow shaded in Figure 3. It is clearfrom Figure 4a,b that the noise levels were reduced after the averaging. Figure 4c,d are theresidual plots for the PPPO- and PS-coated MSS, respectively, representing the differencebetween the experimental values and the expected values from the linear fit. This analysiswas performed in the range between 9.0 s and 10.0 s (green shaded area in Figure 4a,b). Itis also clear from Figure 4c,d that the noise level was reduced by the signal averaging foreach case. As can be seen from Figure 4c,d, the noise level of the difference between Fraiseand Manjari is higher than that of other measurements due to the propagation of errors bysignal processing of two data with noise.Biosensors 2023, 13, x FOR PEER REVIEW 6 of 11  Signal intensity is defined as the difference between the initial and peak sensing out-puts in a cycle, the difference between the signal outputs at 0 s and 5 s. 3. Results Figure 3 shows the whole sensing outputs from one set of the six measurements: the normal measurements of Manjari and Fraise chocolates, the difference between Manjari and Fraise chocolates, and the measurements by the rDC method of Manjari and Fraise chocolates. The baselines were corrected so that the sensing outputs become 0 μV at 0 s. The difference between Manjari and Fraise was calculated by subtracting the Manjari sensing outputs from the Fraise sensing outputs. Figure 3a,b show the results of the MSS coated with PPPO and PS, respectively. The results of normal measurements of Manjari and Fraise shown in Figure 3a are also shown in Figure 1. Clear signal peaks can be seen for Manjari and Fraise chocolates; signal peaks and valleys correspond to the injection of the odors and the carrier gas (i.e., cleaned air), respectively. As explained above, these signals are basically induced by the difference between chocolates and carrier gases, re-flecting almost all components in each chocolate sample, including major base compo-nents. In contrast, lower signal peaks compared to the normal measurements of Manjari and Fraise are observed for the rDC method reflecting the slight difference in odors. For any measurement, the signal shapes for a single cycle of odor injections gradually change for the first few repetitions because of the sensing mechanism of MSS; sensing outputs reflect the gas diffusion process in the receptor layer and the changes in viscoelastic prop-erties of the receptor materials [20,37,38]. After some repetitions, the sensing outputs reach a steady state; for each repetition, almost the same signal appears. The baseline shifted slightly even after the sensing outputs reached the steady state, which may be due to the temperature shift associated with the thermal heating of the circuit. Such sensing outputs can be averaged to improve S/N. The effect of the averaging is shown in Figure 4. Blue and red lines represent the raw and averaged sensing outputs, respectively. The results of the MSS coated with PPPO and PS correspond to Figure 4a,b, respectively. Raw sensing outputs shown in Figure 4a,b were trimmed from the last cycle in the yellow shaded in Figure 3. It is clear from Figure 4a,b that the noise levels were reduced after the averaging. Figure 4c,d are the residual plots for the PPPO- and PS-coated MSS, respectively, representing the difference between the experimental values and the expected values from the linear fit. This analysis was performed in the range between 9.0 s and 10.0 s (green shaded area in Figure 4a,b). It is also clear from Figure 4c,d that the noise level was reduced by the signal averaging for each case. As can be seen from Figure 4c,d, the noise level of the difference between Fraise and Manjari is higher than that of other measurements due to the propagation of errors by signal processing of two data with noise.  Figure 4. Raw (blue) and averaged (red) sensing outputs from (a) the PPPO-coated MSS and (b) the PS-coated MSS. Residuals of sensing outputs for the linear fit over the time period of 9.0 to 10.0 s Figure 4. Raw (blue) and averaged (red) sensing outputs from (a) the PPPO-coated MSS and (b) thePS-coated MSS. Residuals of sensing outputs for the linear fit over the time period of 9.0 to 10.0 s(green-shaded area in (a) and (b)) for (c) the PPPO-coated MSS and (d) the PS-coated MSS. Blue andred lines indicate raw and averaged sensing outputs, respectively.Biosensors 2023, 13, 368 7 of 11The noise levels for the PPPO-coated MSS and the PS-coated MSS are summarizedin Figure 5a,b, respectively. The bars and error bars represent the mean and standarddeviation of the noise level evaluated from the six measurements. Noise reduction wasconfirmed for all the measurements after averaging. Theoretical values are also shown inFigure 5 as references. If the noise is considered to have a Gaussian distribution, the noiselevel is inversely proportional to the square root of N. Thus, theoretical values for normalmeasurements on Manjari and Fraise chocolates and the rDC method are calculated bydividing the noise levels of the raw sensing outputs with√N, where N is the number ofrepetitions. In this study, N = 6 corresponds to the number of repetitions in the yellow-shaded areas in Figure 3. The theoretical values for the difference between Manjari andFraise were calculated by the following form:√σ2th,Manjari + σ2th,Fraise, (2)where σ2th,Manjari and σ2th,Fraise are the theoretical noise levels (i.e., RMS) for the normalmeasurements of Manjari and Fraise chocolates, respectively. Although noise levels werereduced by averaging, deviations from the theoretical values were observed in some cases:the normal measurement of Fraise and the rDC method with the PS-coated MSS. Moreover,undulations in noises can be seen for some measurements in Figure 4a,b. The deviationfrom the theoretical expectation might be due to such noise that does not follow a Gaussiandistribution. Such noise components, which can originate from electronic noises from thecircuits, including pumps and valves, could be canceled or enhanced by averaging.Biosensors 2023, 13, x FOR PEER REVIEW 7 of 11  (green-shaded area in (a) and (b)) for (c) the PPPO-coated MSS and (d) the PS-coated MSS. Blue and red lines indicate raw and averaged sensing outputs, respectively. The noise levels for the PPPO-coated MSS and the PS-coated MSS are summarized in Figure 5a,b, respectively. The bars and error bars represent the mean and standard devi-ation of the noise level evaluated from the six measurements. Noise reduction was con-firmed for all the measurements after averaging. Theoretical values are also shown in Fig-ure 5 as references. If the noise is considered to have a Gaussian distribution, the noise level is inversely proportional to the square root of 𝑁. Thus, theoretical values for normal measurements on Manjari and Fraise chocolates and the rDC method are calculated by dividing the noise levels of the raw sensing outputs with √𝑁, where 𝑁 is the number of repetitions. In this study, 𝑁 = 6 corresponds to the number of repetitions in the yellow-shaded areas in Figure 3. The theoretical values for the difference between Manjari and Fraise were calculated by the following form: 𝜎 , + 𝜎 , , (2) where 𝜎 ,  and 𝜎 ,  are the theoretical noise levels (i.e., RMS) for the normal measurements of Manjari and Fraise chocolates, respectively. Although noise levels were reduced by averaging, deviations from the theoretical values were observed in some cases: the normal measurement of Fraise and the rDC method with the PS-coated MSS. Moreover, undulations in noises can be seen for some measurements in Figure 4a,b. The deviation from the theoretical expectation might be due to such noise that does not follow a Gaussian dis-tribution. Such noise components, which can originate from electronic noises from the cir-cuits, including pumps and valves, could be canceled or enhanced by averaging.  Figure 5. Summary of the noise level of each measurement with (a) the PPPO-coated MSS and (b) the PS-coated MSS. Error bars represent the standard deviations. 4. Discussion The signal intensities, noise levels after averaging (from Figure 5), and S/N are sum-marized in Figure 6 and Table 1. S/N was calculated for each measurement, and the mean and standard deviation were evaluated from the S/N. In the case of the PPPO-coated MSS, the signal intensity is slightly higher for the rDC method than the difference of the normal measurements, while the noise level is almost the same for both methods. A comparison of the S/N shows that the rDC method achieved a slightly higher S/N. An interesting fea-ture can be seen for the PS-coated MSS. The rDC method achieved significantly higher signal intensity than the normal measurement. This enhancement effect might be due to the absorption/desorption dynamics of VOCs at the receptor layer. As the receptor layer is not purged with a carrier gas for the rDC method, common VOCs between the two samples remain in the receptor layer during the measurement. Such remaining VOCs may interact with the different components of the odors and cause nonlinear expansion of the Figure 5. Summary of the noise level of each measurement with (a) the PPPO-coated MSS and (b) thePS-coated MSS. Error bars represent the standard deviations.4. DiscussionThe signal intensities, noise levels after averaging (from Figure 5), and S/N aresummarized in Figure 6 and Table 1. S/N was calculated for each measurement, andthe mean and standard deviation were evaluated from the S/N. In the case of the PPPO-coated MSS, the signal intensity is slightly higher for the rDC method than the differenceof the normal measurements, while the noise level is almost the same for both methods.A comparison of the S/N shows that the rDC method achieved a slightly higher S/N.An interesting feature can be seen for the PS-coated MSS. The rDC method achievedsignificantly higher signal intensity than the normal measurement. This enhancement effectmight be due to the absorption/desorption dynamics of VOCs at the receptor layer. Asthe receptor layer is not purged with a carrier gas for the rDC method, common VOCsbetween the two samples remain in the receptor layer during the measurement. Suchremaining VOCs may interact with the different components of the odors and causenonlinear expansion of the receptor layer, leading to enhanced signal intensity. The resultimplies that a specific selection of receptor material can heighten the sensing signalsoriginating from the difference in odors. However, further investigation is needed to clarifythe mechanism of signal enhancement.Biosensors 2023, 13, 368 8 of 11Biosensors 2023, 13, x FOR PEER REVIEW 8 of 11  receptor layer, leading to enhanced signal intensity. The result implies that a specific se-lection of receptor material can heighten the sensing signals originating from the differ-ence in odors. However, further investigation is needed to clarify the mechanism of signal enhancement.  Figure 6. Summary of (a) the intensity and (b) S/N. Error bars represent the standard deviations. Table 1. Summary of the signal intensities, noise levels, and S/N. Receptor Material Measurement Type Signal  Intensity (μV) Noise Level after Averaging (μV) S/N PPPO rDC 10.6 ± 2.4 0.87 ± 0.45 14.6 ± 6.4 PPPO Difference of the Normal Measurements 4.7 ± 3.3 1.09 ± 0.31 4.5 ± 2.7 PS rDC 98 ± 26 1.48 ± 0.15 67 ± 19 PS Difference of the Normal Measurements 29 ± 23 1.58 ± 0.28 21 ± 20 Although signal averaging is a common technique in many types of sensor measure-ments, it has rarely been applied to gas sensor measurements, except for optical gas sen-sors [15–19]. The reason is that the same signal outputs could hardly be obtained with gas sensors because of residual gases in their sensing units. In this study, we approached this problem by allowing the sensing responses to reach a steady state by repeating the cycle of sample gas injection and carrier gas purging. This approach has made it possible to obtain reproducible sensing outputs for each repetition without fully cleaning the sensing units. The theoretical background for the sensing responses of MSS in repetitive measure-ments has been investigated in Refs. [20,21]. This approach is not limited to MSS and can be applied to other types of gas sensors. Besides averaging the sensing outputs after repetitions of odor injections, noise re-duction techniques such as low-pass filters and moving average filters are typically used for reducing the noise. While such methods can efficiently reduce high-frequency noise, the methods may also attenuate the rapid sensor response, which contains information on the interaction between the sensing element and the gas molecules. The responses of MSS, for example, reflect the dynamics of gas sorption and viscoelastic stress relaxation of the receptor layer, with time constants varying typically between 0.1 to 100 s [20,39]. As the rapid sensor response with a short time constant is important in odor discrimination and identification from the viewpoint of feature extraction for machine learning [22,40], noise reduction methods that attenuate high-frequency components should be applied to the sensing outputs after careful parameter setting (e.g., cutoff frequency and window width). Figure 6. Summary of (a) the intensity and (b) S/N. Error bars represent the standard deviations.Table 1. Summary of the signal intensities, noise levels, and S/N.Receptor Material Measurement Type Signal Intensity (µV) Noise Level afterAveraging (µV) S/NPPPO rDC 10.6 ± 2.4 0.87 ± 0.45 14.6 ± 6.4PPPO Difference of the NormalMeasurements 4.7 ± 3.3 1.09 ± 0.31 4.5 ± 2.7PS rDC 98 ± 26 1.48 ± 0.15 67 ± 19PS Difference of the NormalMeasurements 29 ± 23 1.58 ± 0.28 21 ± 20Although signal averaging is a common technique in many types of sensor mea-surements, it has rarely been applied to gas sensor measurements, except for optical gassensors [15–19]. The reason is that the same signal outputs could hardly be obtained withgas sensors because of residual gases in their sensing units. In this study, we approachedthis problem by allowing the sensing responses to reach a steady state by repeating thecycle of sample gas injection and carrier gas purging. This approach has made it possible toobtain reproducible sensing outputs for each repetition without fully cleaning the sensingunits. The theoretical background for the sensing responses of MSS in repetitive measure-ments has been investigated in Refs. [20,21]. This approach is not limited to MSS and canbe applied to other types of gas sensors.Besides averaging the sensing outputs after repetitions of odor injections, noise re-duction techniques such as low-pass filters and moving average filters are typically usedfor reducing the noise. While such methods can efficiently reduce high-frequency noise,the methods may also attenuate the rapid sensor response, which contains information onthe interaction between the sensing element and the gas molecules. The responses of MSS,for example, reflect the dynamics of gas sorption and viscoelastic stress relaxation of thereceptor layer, with time constants varying typically between 0.1 to 100 s [20,39]. As therapid sensor response with a short time constant is important in odor discrimination andidentification from the viewpoint of feature extraction for machine learning [22,40], noisereduction methods that attenuate high-frequency components should be applied to thesensing outputs after careful parameter setting (e.g., cutoff frequency and window width).5. ConclusionsIn conclusion, we propose a measurement method in which the difference in odorsof samples is highlighted. In this repetitive direct comparison (rDC) method, the odorsof samples are alternately injected into a sensor without carrier gas purging. The cycleof the odor injections is repeated to obtain multiple sensing responses for averaging theoutputs. As a demonstration, we measured the odors of chocolates by the rDC methodand detected the sensor responses originating from the difference in flavors with higherBiosensors 2023, 13, 368 9 of 11S/N than a typical measurement protocol. By utilizing the rDC method, not only thecommon components from the base material but also the changes in the experimentalconditions (e.g., temperature, humidity, and VOCs in the air) could be canceled, leadingto applications such as odor quality control and the detection of defective products. If astandard product is available, the quality of each product can be examined by the rDCmethod without being affected by the disturbance of the experimental conditions. Sincethis measurement method is advantageous for highlighting the slight difference in odor, itcan be used for various purposes such as quality control of food, monitoring of indoor airquality, and classification of cosmetics, especially for measuring samples consisting of largeamounts of interfering components with different odors (e.g., fruit-flavored carbonateddrinks and room fragrances). In a product inspection, for example, one can examine theodor of a product by comparing the product to a reference product with the rDC method;the product passes the inspection if the signal intensity is within a permission range. Webelieve this method will become a key technology for odor sensing and contribute to thewidespread use of odor sensors.6. PatentsJapanese patent application No. 2022-074098.Author Contributions: Conceptualization, G.I. and G.Y.; methodology, G.I., K.M., and G.Y.; software,G.I.; validation, K.M. and G.Y.; investigation, G.I., K.M., and G.Y.; writing—original draft preparation,G.I.; writing—review and editing, K.M. and G.Y.; supervision, G.I. and G.Y.; project administration,G.I. and G.Y.; funding acquisition, G.I., K.M., and G.Y.; All authors have read and agreed to thepublished version of the manuscript.Funding: This research was funded by the Leading Initiative for Excellent Young Researchers,Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan; a Grant-in-Aidfor Scientific Research (C) [20K05345, MEXT, Japan]; a Grant-in-Aid for Scientific Research (C)[22K05324, MEXT, Japan]; a Grant-in-Aid for Scientific Research (A) [18H04168, MEXT, Japan]; Leavea Nest Grant, Yoshinoya Award, Leave a Nest Co., ltd. and Yoshinoya Co., Ltd.; the Public/PrivateR&D Investment Strategic Expansion Program (PRISM), Cabinet Office, Japan; the World PremierInternational Research Center Initiative (WPI) on Materials Nanoarchitectonics (MANA), NIMS; theCenter for Functional Sensor & Actuator (CFSN), NIMS.Institutional Review Board Statement: Not applicable.Informed Consent Statement: Not applicable.Data Availability Statement: Not applicable.Acknowledgments: We would like to thank Dissanayake Santha Kumara, Masaaki Matoba, andFusako Hidaka for their technical support and Takahiro Nemoto for the fruitful discussion. We wouldalso like to appreciate MSS Alliance/Forum/Partnership.Conflicts of Interest: The authors declare no conflict of interest.References1. Kader, A.A. Flavor quality of fruits and vegetables. J. Sci. Food Agric. 2008, 88, 1863–1868. [CrossRef]2. Shi, H.; Zhang, M.; Adhikari, B. Advances of electronic nose and its application in fresh foods: A review. Crit. Rev. Food Sci. Nutr.2018, 58, 2700–2710. [CrossRef]3. Jia, W.; Liang, G.; Jiang, Z.; Wang, J. Advances in Electronic Nose Development for Application to Agricultural Products. FoodAnal. Methods 2019, 12, 2226–2240. [CrossRef]4. Tan, J.; Xu, J. Applications of electronic nose (e-nose) and electronic tongue (e-tongue) in food quality-related propertiesdetermination: A review. Artif. Intell. Agric. 2020, 4, 104–115. [CrossRef]5. Su, C.-Y.; Menuz, K.; Carlson, J.R. Olfactory Perception: Receptors, Cells, and Circuits. Cell 2009, 139, 45–59. [CrossRef]6. Niimura, Y.; Nei, M. Evolution of olfactory receptor genes in the human genome. Proc. Natl. Acad. Sci. USA 2003, 100, 12235–12240.[CrossRef]7. Persaud, K.; Dodd, G. Analysis of discrimination mechanisms in the mammalian olfactory system using a model nose. Nature1982, 299, 352–355. [CrossRef]8. Persaud, K.C. Biomimetic Olfactory Sensors. IEEE Sens. J. 2012, 12, 3108–3112. [CrossRef]http://doi.org/10.1002/jsfa.3293http://doi.org/10.1080/10408398.2017.1327419http://doi.org/10.1007/s12161-019-01552-1http://doi.org/10.1016/j.aiia.2020.06.003http://doi.org/10.1016/j.cell.2009.09.015http://doi.org/10.1073/pnas.1635157100http://doi.org/10.1038/299352a0http://doi.org/10.1109/JSEN.2012.2214118Biosensors 2023, 13, 368 10 of 119. Ye, Z.; Liu, Y.; Li, Q. Recent Progress in Smart Electronic Nose Technologies Enabled with Machine Learning Methods. Sensors2021, 21, 7620. [CrossRef] [PubMed]10. van den Broek, J.; Weber, I.C.; Güntner, A.T.; Pratsinis, S.E. Highly selective gas sensing enabled by filters. Mater. Horiz. 2021, 8,661–684. [CrossRef] [PubMed]11. Jirayupat, C.; Nagashima, K.; Hosomi, T.; Takahashi, T.; Samransuksamer, B.; Hanai, Y.; Nakao, A.; Nakatani, M.; Liu, J.; Zhang,G.; et al. Breath odor-based individual authentication by an artificial olfactory sensor system and machine learning. Chem.Commun. 2022, 58, 6465. [CrossRef]12. Reddy, K.; Guo, Y.; Liu, J.; Lee, W.; Khaing Oo, M.K.; Fan, X. Rapid, sensitive, and multiplexed on-chip optical sensors formicro-gas chromatography. Lab A Chip 2012, 12, 901–905. [CrossRef]13. James, F.; Breuil, P.; Pijolat, C.; Camara, M.; Briand, D.; Bart, A.; Cozic, R. Development of a MEMS Preconcentrator for Micro-gasChromatography Analyses. Procedia Eng. 2014, 87, 500–503. [CrossRef]14. Biswas, P.; Zhang, C.; Chen, Y.; Liu, Z.; Vaziri, S.; Zhou, W.; Sun, Y. A Portable Micro-Gas Chromatography with IntegratedPhotonic Crystal Slab Sensors on Chip. Biosensors 2021, 11, 326. [CrossRef]15. Werle, P.; Mücke, R.; Slemr, F. The limits of signal averaging in atmospheric trace-gas monitoring by tunable diode-laser absorptionspectroscopy (TDLAS). Appl. Phys. B 1993, 57, 131–139. [CrossRef]16. Bauer, R.; Stewart, G.; Johnstone, W.; Boyd, E.; Lengden, M. 3D-printed miniature gas cell for photoacoustic spectroscopy of tracegases. Opt. Lett. 2014, 39, 4796–4799. [CrossRef] [PubMed]17. Li, J.; Deng, H.; Sun, J.; Yu, B.; Fischer, H. Simultaneous atmospheric CO, N2O and H2O detection using a single quantum cascadelaser sensor based on dual-spectroscopy techniques. Sens. Actuators B Chem. 2016, 231, 723–732. [CrossRef]18. Khan, A.; Schaefer, D.; Tao, L.; Miller, D.J.; Sun, K.; Zondlo, M.A.; Harrison, W.A.; Roscoe, B.; Lary, D.J. Low Power GreenhouseGas Sensors for Unmanned Aerial Vehicles. Remote Sens. 2012, 4, 1355–1368. [CrossRef]19. Langridge, J.M.; Ball, S.M.; Shillings, A.J.L.; Jones, R.L. A broadband absorption spectrometer using light emitting diodes forultrasensitive, in situ trace gas detection. Rev. Sci. Instrum. 2008, 79, 123110. [CrossRef] [PubMed]20. Minami, K.; Shiba, K.; Yoshikawa, G. Sorption-induced static mode nanomechanical sensing with viscoelastic receptor layers formultistep injection-purge cycles. J. Appl. Phys. 2021, 129, 124503. [CrossRef]21. Inada, K.; Kojima, H.; Cho-Isoda, Y.; Tamura, R.; Imamura, G.; Minami, K.; Nemoto, T.; Yoshikawa, G. Statistical Evaluation ofTotal Expiratory Breath Samples Collected throughout a Year: Reproducibility and Applicability toward Olfactory Sensor-BasedBreath Diagnostics. Sensors 2021, 21, 4742. [CrossRef] [PubMed]22. Shiba, K.; Tamura, R.; Imamura, G.; Yoshikawa, G. Data-driven nanomechanical sensing: Specific information extraction from acomplex system. Sci. Rep. 2017, 7, 3661. [CrossRef]23. Ngo, H.T.; Minami, K.; Imamura, G.; Shiba, K.; Yoshikawa, G. Effects of Center Metals in Porphines on Nanomechanical GasSensing. Sensors 2018, 18, 1640. [CrossRef]24. Aprotosoaie, A.C.; Luca, S.V.; Miron, A. Flavor Chemistry of Cocoa and Cocoa Products—An Overview. Compr. Rev. Food Sci.Food Saf. 2016, 15, 73–91. [CrossRef]25. Sen, A.; Schieberle, P.; Grosch, W. Quantitative-determination of 2,5-dimethyl-4-hydroxy-3(2H)-furanone and its methyl-etherusing a stable isotope-dilution assay. Food Sci. Technol.-Lebensm.-Wiss. Technol. 1991, 24, 364–369.26. Gimzewski, J.K.; Gerber, C.; Meyer, E.; Schlittler, R.R. Observation of a chemical reaction using a micromechanical sensor. Chem.Phys. Lett. 1994, 217, 589–594. [CrossRef]27. Thundat, T.; Warmack, R.J.; Chen, G.Y.; Allison, D.P. Thermal and ambient-induced deflections of scanning force microscopecantilevers. Appl. Phys. Lett. 1994, 64, 2894–2896. [CrossRef]28. Lang, H.P.; Hegner, M.; Gerber, C. Cantilever array sensors. Mater. Today 2005, 8, 30–36. [CrossRef]29. Sader, J.E. Surface stress induced deflections of cantilever plates with applications to the atomic force microscope: Rectangularplates. J. Appl. Phys. 2001, 89, 2911–2921. [CrossRef]30. Watari, M.; Galbraith, J.; Lang, H.-P.; Sousa, M.; Hegner, M.; Gerber, C.; Horton, M.A.; McKendry, R.A. Investigating the MolecularMechanisms of In-Plane Mechanochemistry on Cantilever Arrays. J. Am. Chem. Soc. 2007, 129, 601–609. [CrossRef]31. Yoshikawa, G.; Akiyama, T.; Gautsch, S.; Vettiger, P.; Rohrer, H. Nanomechanical Membrane-type Surface Stress Sensor. Nano Lett.2011, 11, 1044–1048. [CrossRef] [PubMed]32. Osica, I.; Imamura, G.; Shiba, K.; Ji, Q.; Shrestha, L.K.; Hill, J.P.; Kurzydłowski, K.J.; Yoshikawa, G.; Ariga, K. Highly NetworkedCapsular Silica–Porphyrin Hybrid Nanostructures as Efficient Materials for Acetone Vapor Sensing. ACS Appl. Mater. Interfaces2017, 9, 9945–9954. [CrossRef] [PubMed]33. Imamura, G.; Minami, K.; Shiba, K.; Mistry, K.; Musselman, K.P.; Yavuz, M.; Yoshikawa, G.; Saiki, K.; Obata, S. Graphene Oxide asa Sensing Material for Gas Detection Based on Nanomechanical Sensors in the Static Mode. Chemosensors 2020, 8, 82. [CrossRef]34. Baller, M.K.; Lang, H.P.; Fritz, J.; Gerber, C.; Gimzewski, J.K.; Drechsler, U.; Rothuizen, H.; Despont, M.; Vettiger, P.; Battiston,F.M.; et al. A cantilever array-based artificial nose. Ultramicroscopy 2000, 82, 1–9. [CrossRef]35. Loizeau, F.; Akiyama, T.; Gautsch, S.; Vettiger, P.; Yoshikawa, G.; de Rooij, N.F. Comparing membrane- and cantilever-basedsurface stress sensors for reproducibility. Sens. Actuator A Phys. 2015, 228, 9–15. [CrossRef]36. Minami, K.; Imamura, G.; Tamura, R.; Shiba, K.; Yoshikawa, G. Recent Advances in Nanomechanical Membrane-Type SurfaceStress Sensors towards Artificial Olfaction. Biosensors 2022, 12, 762. [CrossRef]http://doi.org/10.3390/s21227620http://www.ncbi.nlm.nih.gov/pubmed/34833693http://doi.org/10.1039/D0MH01453Bhttp://www.ncbi.nlm.nih.gov/pubmed/34821311http://doi.org/10.1039/D2CC90179Jhttp://doi.org/10.1039/c2lc20922ehttp://doi.org/10.1016/j.proeng.2014.11.413http://doi.org/10.3390/bios11090326http://doi.org/10.1007/BF00425997http://doi.org/10.1364/OL.39.004796http://www.ncbi.nlm.nih.gov/pubmed/25121877http://doi.org/10.1016/j.snb.2016.03.089http://doi.org/10.3390/rs4051355http://doi.org/10.1063/1.3046282http://www.ncbi.nlm.nih.gov/pubmed/19123548http://doi.org/10.1063/5.0039045http://doi.org/10.3390/s21144742http://www.ncbi.nlm.nih.gov/pubmed/34300482http://doi.org/10.1038/s41598-017-03875-7http://doi.org/10.3390/s18051640http://doi.org/10.1111/1541-4337.12180http://doi.org/10.1016/0009-2614(93)E1419-Hhttp://doi.org/10.1063/1.111407http://doi.org/10.1016/S1369-7021(05)00792-3http://doi.org/10.1063/1.1342018http://doi.org/10.1021/ja065222xhttp://doi.org/10.1021/nl103901ahttp://www.ncbi.nlm.nih.gov/pubmed/21314159http://doi.org/10.1021/acsami.6b15680http://www.ncbi.nlm.nih.gov/pubmed/28234457http://doi.org/10.3390/chemosensors8030082http://doi.org/10.1016/S0304-3991(99)00123-0http://doi.org/10.1016/j.sna.2015.02.039http://doi.org/10.3390/bios12090762Biosensors 2023, 13, 368 11 of 1137. Wenzel, M.J.; Josse, F.; Heinrich, S.M.; Yaz, E.; Datskos, P.G. Sorption-induced static bending of microcantilevers coated withviscoelastic material. J. Appl. Phys. 2008, 103, 064913. [CrossRef]38. Heinrich, S.M.; Wenzel, M.J.; Josse, F.; Dufour, I. An analytical model for transient deformation of viscoelastically coated beams:Applications to static-mode microcantilever chemical sensors. J. Appl. Phys. 2009, 105, 124903. [CrossRef]39. Imamura, G.; Shiba, K.; Yoshikawa, G.; Washio, T. Analysis of nanomechanical sensing signals; physical parameter estimation forgas identification. AIP Adv. 2018, 8, 075007. [CrossRef]40. Imamura, G.; Shiba, K.; Yoshikawa, G.; Washio, T. Free-hand gas identification based on transfer function ratios without gas flowcontrol. Sci. Rep. 2019, 9, 9768. [CrossRef]Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individualauthor(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury topeople or property resulting from any ideas, methods, instructions or products referred to in the content.http://doi.org/10.1063/1.2902500http://doi.org/10.1063/1.3148291http://doi.org/10.1063/1.5036686http://doi.org/10.1038/s41598-019-46164-1 Introduction  Materials and Methods  Measurement Setup  Odor Samples  Membrane-Type Surface Stress Sensors (MSS)  Signal Processing  Results  Discussion  Conclusions  Patents  References