JP6710826B2 - Combustible gas analysis method - Google Patents

Description

The present invention relates to a combustible gas analysis method .

There are various odorants in the living environment. Most of these odor components are trace amounts of organic molecules and are called volatile organic compounds (VOC). VOCs have various properties. Various low-concentration VOCs equivalent to several tens to hundreds of tens of ppb are present in the room due to radiation from building materials, interiors, furniture, etc., and use of cosmetics and air fresheners by residents. Therefore, in an atmosphere where a specific VOC is to be detected, water vapor having a constant concentration and volatile impurities having a low concentration usually exist.

For example, in the medical field, it has been pointed out that VOC contained in exhaled breath is associated with bad breath, metabolism and disease, and it is said that there is a correlation between the VOC in exhaled breath and the health condition. In the field of food, it is said that there is a correlation between freshness and quality and VOC corresponding to the odor emitted from the food. Therefore, the monitoring of these specific VOCs is effective in managing the health condition in the former case and as food control in the latter case. However, this measurement must be carried out in indoor air, and since volatile impurities are present in the air, a method for detecting a specific VOC without affecting volatile impurities is necessary. is there.

On the other hand, in order to detect a combustible gas represented by VOC, a semiconductor sensor using a metal oxide n-type semiconductor has been conventionally used. The semiconductor type sensor is characterized by low price and running cost. A typical semiconductor type sensor is tin oxide, and oxygen in the atmosphere is adsorbed on the surface of the tin oxide particles that compose the sensor element sensitive area, and the carrier electrons are captured by the adsorbed oxygen, and the interface between particles is That is, a potential barrier is formed at the grain boundary, increasing the electrical resistance of the grain boundary. When the flammable gas reaches the surface of the tin oxide particles and consumes the adsorbed oxygen on the surface to oxidize the particles, the potential barrier at the grain boundary is reduced and the electrical resistance is reduced. Tin oxide is classified as a so-called grain boundary responsive semiconductor gas sensor.

However, it is difficult for the grain boundary responsive semiconductor gas sensor to contribute to the gas selectivity of each sensor element due to its nature.

Under such circumstances, it is desired to develop a gas sensor group capable of detecting a specific combustible gas in the air containing volatile impurities.

In order to detect a specific VOC with a simple and inexpensive gas sensor group, a gas sensor group having a plurality of gas sensors typified by a semiconductor sensor is simultaneously detected, and statistical analysis such as principal component analysis is performed using the obtained signals. Are being studied.

It is difficult for each gas sensor to detect only a specific combustible gas by its nature, but a certain tendency can be obtained by statistically analyzing various signals obtained from a plurality of gas sensors.

Patent Document 1 discloses a composite odor sensor that performs determination by combining a sensor array configured by combining a plurality of ceramic semiconductor type sensors and a sensor array configured by combining a plurality of crystal oscillator type sensors. There is.

Patent Document 2 discloses a gas molecule detection element in which a gas molecule selection material in which an ionic liquid is held in a plasma organic thin film is formed on the surface of a quartz oscillator, and a plurality of gas molecule detections in which the concentration of the ionic liquid in the gas molecule selection material is different. A gas sensor array in which elements are arranged is disclosed.

Patent Document 3 discloses a gas identification and detection method in which a plurality of types of gases are identified and detected by a plurality of semiconductor gas sensors.

JP-A-10-170422 JP, 2006-53059, A JP, 2004-12193, A

However, the combustible gas in the humidified air containing volatile impurities cannot be sufficiently separated and detected, and it has been an urgent task to solve this problem.

An aspect of the present invention is to provide a method for analyzing a combustible gas capable of sufficiently separating and detecting the combustible gas in humidified air containing volatile impurities.

One embodiment of the present invention is a method of analyzing a combustible gas, the step of detecting a combustible gas using a gas sensor group, and a step of performing a principal component analysis of a response value obtained by detecting the combustible gas. The gas sensor group includes a plurality of types of grain boundary response type semiconductor gas sensors and one or more types of bulk response type semiconductor gas sensors, and the bulk response type semiconductor gas sensor is cerium oxide or cerium oxide. - look containing zirconium-based composite oxide, semiconductor type gas sensor of the one or more bulk-responsive, the cerium - on the membrane of zirconium complex oxide, through the alumina film, the platinum particles are supported It includes a sensor in which a carrier of the cerium-zirconium-based composite oxide is formed .

According to one embodiment of the present invention, a method for analyzing a flammable gas capable of sufficiently separating and detecting the flammable gas in humidified air containing volatile impurities can be provided.

It is a gas flow diagram of the measurement system used in the example. FIG. 5 is a diagram showing principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component of the first embodiment. FIG. 9 is a diagram showing a principal component score and an eigenvector by the first principal component and the second principal component, the first principal component and the third principal component of the second embodiment. FIG. 10 is a diagram showing a principal component score and an eigenvector by the first principal component and the second principal component, the first principal component and the third principal component of the third embodiment. FIG. 16 is a diagram showing a principal component score and an eigenvector based on the first principal component, the second principal component, the first principal component, and the third principal component of the fourth embodiment. FIG. 16 is a diagram showing a principal component score and an eigenvector by the first principal component and the second principal component, and the first principal component and the third principal component of the fifth embodiment. It is a figure which shows the principal component score and eigenvector by a 1st principal component and a 2nd principal component, a 1st principal component, and a 3rd principal component of Example 6. It is a figure which shows the principal component score by a 1st main component and a 2nd main component, a 1st main component, and a 3rd main component, and an eigenvector when the humidity of the comparative example 1 corresponds to 0%RH at 20 degreeC conversion. It is a figure which shows the principal component score by a 1st main component and a 2nd main component, a 1st main component, and a 3rd main component, and an eigenvector, when the humidity of the comparative example 1 is equivalent to 60%RH at 20 degreeC conversion. 11 is a diagram showing a principal component score and an eigenvector by the first principal component and the second principal component, and the first principal component and the third principal component of Comparative Example 2. FIG. 11 is a diagram showing a principal component score and an eigenvector by the first principal component and the second principal component, the first principal component and the third principal component of Comparative Example 3. FIG.

Next, a mode for carrying out the present invention will be described.

[Gas sensor group]
The gas sensor group of the present embodiment is used for detection of combustible gas, and has a plurality of types of grain boundary responsive semiconductor gas sensors and one or more types of bulk responsive semiconductor gas sensors. By using a sensor having a different detection principle, the accuracy of the principal component analysis described later can be improved.

In the present specification and claims, a combustible gas is defined as a gas that is a substance that can exist as a gas and that can be detected by a semiconductor gas sensor. That is, the flammable gas is a substance that can be oxidized on the surface of the metal oxide (particles) constituting the semiconductor gas sensor described later.

Examples of the flammable gas include organic compounds such as VOC and inorganic compounds such as hydrogen sulfide, hydrogen, carbon monoxide and the like.

In the present specification and claims, VOC is defined as all volatile organic compounds.

Examples of VOCs include aliphatic hydrocarbons, aromatic hydrocarbons, esters, alcohols, aldehydes, ketones, terpenes and halogenated hydrocarbons. Of these, nonane, decane, acetone, methyl isobutyl ketone, acetoin, and other components contained in the breath of the patient are preferable.

In the grain boundary responsive semiconductor type gas sensor of the present embodiment, the sensor element sensitive portion is composed of n-type semiconductor metal oxide particles. Therefore, oxygen in the atmosphere is adsorbed on the surface of the metal oxide particles, and the carrier electrons are captured by the adsorbed oxygen, so that a potential barrier is formed at the particle-particle interface, that is, at the grain boundary. Resistance increases. Here, when the flammable gas reaches the surface of the metal oxide particles, the adsorbed oxygen on the surface is consumed and oxidized, whereby the potential barrier at the grain boundary is reduced and the electrical resistance is also reduced. Therefore, the electric resistance varies depending on the type and concentration of the combustible gas. Therefore, when the humidity changes, the amount of water adsorbed on the surface of the metal oxide particles changes, which affects the amount of oxygen adsorbed on the surface of the metal oxide particles. As a result, in the grain boundary response type semiconductor gas sensor, the sensor response is easily affected by humidity.

The bulk response type semiconductor gas sensor according to the present embodiment is an n-type semiconductor metal oxide that easily supplies oxygen necessary for the oxidation reaction of the combustible gas from oxygen atoms in the crystal lattice according to the concentration of the combustible gas. The sensor element sensitive section is constituted by. Therefore, when the flammable gas reaches the surface of the metal oxide, the oxygen atoms in the lattice are consumed by the oxidation reaction. The diffusion rate of the generated oxygen vacancies is very high, and the oxygen vacancies accumulate on the surface of the metal oxide and do not rate-control the entire bulk. Depends on. When the concentration of oxygen vacancies increases, the concentration of carrier electrons also increases, so that the electrical resistance of the metal oxide also decreases. Therefore, the electrical resistance of the metal oxide varies depending on the type and concentration of the combustible gas. Therefore, the bulk response type semiconductor gas sensor is unlikely to be affected by changes in humidity.

The sensor element sensitive part of the bulk response type semiconductor gas sensor of the present embodiment is not particularly limited, but a metal oxide having an oxygen storage capacity may be mentioned, and two or more kinds may be used in combination. Among them, cerium oxide and cerium-zirconium-based composite oxides are preferable, and the composition formula Ce _1-x Zr _x O _{2 is used.}
(In the formula, x is in the range of 0 to 0.4.)
Compounds represented by are more preferable.

Further, the sensor element sensitive part of the bulk response type semiconductor gas sensor of the present embodiment is provided with an insulating layer on the upper part thereof, and in a state where it is electrically insulated from the sensor element sensitive part, noble metal catalyst particles are further provided on the upper part thereof. A supported metal oxide catalyst carrier can be provided.

The insulating layer is required to have not only electrical insulation but also gas permeability that allows VOC to reach the lower sensor element sensitive portion, and is composed of insulating fine particles.

The insulating fine particles are not particularly limited, but examples thereof include insulating metal oxides and silica.

The insulating metal oxide is not particularly limited, but examples thereof include alumina, zirconia, magnesia, hafnia, yttria, and calcia. Of these, alumina is preferable.

The material constituting the noble metal catalyst particles is not particularly limited, but platinum, palladium, gold, silver, rhodium, ruthenium and the like can be mentioned. Of these, platinum is preferable.

The catalyst carrier for the metal oxide is not particularly limited, and examples thereof include metal oxides having an oxygen storage capacity, alumina, zirconia, magnesia, hafnia, yttria, and calcia. Among them, cerium oxide and cerium-zirconium-based composite oxides are preferable, and the composition formula Ce _1-x Zr _x O _{2 is used.}
(In the formula, x is in the range of 0 to 0.4.)
Compounds represented by are more preferable.

[Combustible gas analysis method]
The flammable gas analysis method of the present embodiment includes a step of detecting the flammable gas using the gas sensor group of the present embodiment.

For example, by performing a principal component analysis on the gas sensor response value obtained by detecting the combustible gas, it is possible to sufficiently separate and detect the combustible gas in the humidified air containing volatile impurities.

The flammable gas analysis method of the present embodiment can be applied to, for example, a system that manages a health condition by detecting exhaled breath in the medical field, and a system that manages food in the food field.

Next, the present invention will be specifically described based on examples, but the present invention is not limited to the following examples.

[Example 1]
In this embodiment, six types of grain boundary responsive semiconductor type gas sensors (TGS2600, TGS2602, TGS2610, TGS2620 (above, manufactured by Figaro Giken Co., Ltd.), #31b, #33b) and one type of bulk response type semiconductor gas sensor ( No. 71) was used.

The TGS2600, TGS2602, TGS2610, and TGS2620 are commercially available gas sensors and are claimed to be effective for organic compound gases. In this example, the heater voltage recommended by the manufacturer was set to 5 V, and the current was applied. Response characteristics of these gas sensors are adjusted by adding an additive to tin oxide. Therefore, these gas sensors are grain boundary responsive semiconductor type sensors.

# 31b and # 33b were prepared by AIST is Pt on SnO _2, gas sensors Pd and Au were added each 1 wt% (Pt, Pd, Au / SnO 2) (Sens. Actuators B 187 (2013) 135-141). Here, in order to show constant sensitivity to various VOCs, Pt, Pd, and Au are added to SnO ₂ to enhance the performance. Therefore, these gas sensors are grain boundary responsive semiconductor type sensors. Since the film thicknesses of #31b and #33b are 4.6 μm and 2.8 μm, respectively, the gas sensor response characteristics are different. All of these gas sensors were used at an operating temperature of 250°C. #31b and #33b were produced according to the production method shown in Sens. Actuators B 187 (2013) 135-141.

No. Reference numeral 71 is a gas sensor manufactured by the National Institute of Advanced Industrial Science and Technology (see Sensors 15 (2015) 9427-9437). The material forming the sensor element sensitive portion is Ce _0.9 Zr _0.1 O ₂ (CeZr10) obtained by adding 10 mol% of zirconium to cerium oxide. No. Reference numeral 71 is a sensor (CeZr10) in which a thick film of alumina (Al ₂ O ₃ ) is placed on the CeZr10 thick film, and CeZr10 (3 wt% Pt-CeZr10) containing 3% by mass of Pt is further mounted thereon. / _Al 2 _O 3/3 wt% Pt-CeZr10) is. Therefore, the present gas sensor is a bulk response type semiconductor sensor. This gas sensor was used at an operating temperature of 400°C.

FIG. 1 shows a gas flow diagram of the measurement system used in this example.

Seven kinds of gas sensors were put in the sensor chamber 1 capable of controlling the gas atmosphere and sealed.

Next, the gas flowing into the sensor chamber 1 will be specifically described.

First, the total gas flow rate to the sensor chamber 1 during measurement was always 500 mL/min. Next, the concentration ratio of nitrogen and oxygen to the sensor chamber 1 during measurement was always set to 4:1. That is, the flow rate of nitrogen is 400 mL/min and the flow rate of oxygen is 100 mL/min. Of these, 200 mL/min for nitrogen and 100 mL/min for oxygen were passed through a water bubbler 2 containing distilled water. Here, the water bubbler 2 was controlled so that the water temperature was always 20°C. That is, 200 mL/min of nitrogen and 100 mL/min of oxygen are saturated with water vapor, and are mixed with 200 mL/min of nitrogen in a dry state to reach the sensor chamber 1. Humidity corresponds to 60% RH in terms of 20° C. (hereinafter referred to as synthetic air).

The target VOCs used in this example were five kinds of acetoin, nonanal, decane, acetone, and methyl isobutyl ketone, and were adjusted so that the concentration when reaching the sensor chamber 1 would be 1 ppm. Acetoin (G. Songa, et al., Lung Cancer 67 (2010) 227-231), nonanal (P. Fuchs et al., Int. J. Cancer 126 (2010) 2663-2670.), decane (D .Poli et al., Respiratory Research 6 (2005) 71-80.) describes that acetone and methyl isobutyl ketone (C. Wang, et al., IEEE Sens. J. 10 (2010) 54- 63.) is reported to be associated with the exhaled breath of diabetic patients.

In addition, in this example, in order to reproduce the volatile impurities contained in the room air, 31-type mixed gas was used as TVOC. 31 kinds of mixed gas is a component defined as a volatile organic compound (VOC) in the standard issued by the International Organization for Standardization, from the viewpoint of the effect on the human body issued by the Ministry of Health, Labor and Welfare and the European Joint Research Center. It is a mixed gas filled with 31 kinds of VOCs that can be filled in a gas cylinder described in the list of VOCs to be monitored. The main components of TVOC are classified into aromatic hydrocarbons (30%) and aliphatic hydrocarbons (37%). TVOC concentrations when reaching the sensor chamber 1 in Europe, especially in reference to the target concentration 300 [mu] g / ^{m 3} of the room indicated Germany Environment Agency, the 0-fold (0μg ^{/ m} 3), 1-fold (300 [mu] g / ^{m 3} ), 2 times (600 μg/m ³ ) and 3 times (900 μg/m ³ ). When TVOC concentration is expressed in ppm, it corresponds to 0 times (0 ppm), 1 times (0.060 ppm), 2 times (0.12 ppm), and 3 times (0.18 ppm) in order.

As these target VOCs and TVOCs, a gas produced from a nitrogen-based gas cylinder, or a gas produced by passing a nitrogen gas through a permeator (made by Gastec Co., Ltd.) as a gas generator 3 producing a gas from a solvent was used. Specifically, among the target VOCs, acetone, methyl isobutyl ketone, and TVOC correspond to the former, and among the target VOCs, acetoin, nonanal, and decane correspond to the latter. When these gases are flowed, the flow rate of dry nitrogen is reduced by an amount corresponding to the flow rate of the flowed gas so that the total flow rate of gas is 500 mL/min and the concentration ratio of nitrogen and oxygen is 4:1. did.

After putting the seven kinds of gas sensors in the sensor chamber 1, the TGS2600, TGS2602, TGS2610 and TGS2620 apply a heater voltage of 5V recommended by the manufacturer, and #31b and #33b apply voltage to the heater so that the heaters have a predetermined operating temperature. Was applied to raise the temperature of the gas sensor to the driving temperature. Next, 500 mL/min of synthetic air was flowed for 2 hours to stabilize the resistance value of the gas sensor and then start recording the resistance value. Subsequently, synthetic air was flowed for 30 minutes, synthetic air containing 1 ppm of target VOC was flowed for 20 minutes, and the resistance value was recorded while flowing synthetic air again for 30 minutes. This operation was carried out on all five target gases of acetoin, nonanal, decane, acetone, and methyl isobutyl ketone.

Further, in the same manner as above, after flowing synthetic air containing a TVOC concentration of 1 time (300 μg/m ³ ) for 30 minutes, a synthetic air containing a TVOC concentration of 1 time and 1 ppm of target TVOC was flowed for 20 minutes, and the TVOC concentration was again adjusted. The resistance value was recorded while flowing synthetic air containing 1 time for 30 minutes. In the same manner, the TVOC concentration was doubled and tripled. These operations were also performed on all five target gases of acetoin, nonanal, decane, acetone, and methyl isobutyl ketone.

Principal component analysis was performed using the response value S of the gas sensor. The response value S of the gas sensor is expressed by the formula S=Ra/Rg
(In the formula, Ra is a resistance value of the gas sensor when flowing synthetic air (or synthetic air containing TVOC), and Rg is a synthetic air containing 1 ppm target TVOC (or synthetic air containing TVOC). The resistance value of the gas sensor when
Is defined by

FIG. 2 shows the principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component. The contribution rate of each principal component is shown on each axis.

From FIG. 2, it can be seen that the cumulative contribution rate up to the third principal component is 93.9%. This indicates that out of the seven gas sensors, the value of 6.573 is collected. In FIG. 2, the size of the vector is tripled in order to show the eigenvector in an easy-to-understand manner. The plot of the principal component score varied depending on the concentration of TVOC, which is a volatile impurity. The arrow in FIG. 2 indicates the direction in which the concentration of TVOC increases. Therefore, the concentration of TVOC is 0 times (0 ppm) on the starting point side of the arrow and 3 times (0.18 ppm) the ending point (all the same from FIG. 3 to FIG. 11). As shown in Fig. 2, acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, show overlapping main component scores, but the main component scores between acetoin, nonanal, decane, and acetone-methyl isobutyl ketone are It is understood that they are separated and can be sufficiently separated and detected.

The principal component score of the target VOC varied in the axial direction of the second principal component. This is caused by changing the concentration of TVOC to 0 to 3 times the standard. The main components of TVOC are aromatic hydrocarbons and aliphatic hydrocarbons. Therefore, it can be said that the second main component tends to be derived from the total number of carbon atoms in VOC. On the other hand, the first main component is in the order of acetoin having a hydroxyl group, nonanal having an aldehyde group, ketone-based acetone and methyl isobutyl ketone, and decane of an aliphatic hydrocarbon in descending order of the main component score. , VOCs, especially in the presence or absence of oxidative functional groups.

In the principal component scores of the first principal component and the second principal component, TGS2602, #31b, #33b, No. 2 which shows eigenvectors substantially parallel to the axis of the first principal component. 71 will be in charge. On the other hand, in Comparative Example 1 which will be described later, not only acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, but also decane, which is not similar in structure, overlap in the main component scores. From this, the bulk response type No. It is considered that 71 particularly contributes to the separation of the ketone system and the aliphatic hydrocarbon.

Also, regarding the main component scores of the first main component and the third main component, in Comparative Example 1 described later, the main component scores of acetone/methylisobutylketone and decane were overlapped, but in Example 1, The main component scores of acetone/methyl isobutyl ketone and decane are separated. No. The eigenvector of 71 is parallel to the direction of its separation, and is considered to particularly contribute to the separation of the ketone system and the aliphatic hydrocarbon.

[Example 2]
In this embodiment, six kinds of grain boundary response type semiconductor gas sensors (TGS2600, TGS2602, TGS2610, TGS2620 (above, manufactured by Figaro Giken Co., Ltd.), #31b, #33b) and two types of bulk response type semiconductor gas sensors ( No. 9 and No. 71) were used. In addition, Example 2 corresponds to Example 1 with one type of bulk response type semiconductor gas sensor added.

No. 9 was fabricated by the Industrial Technology Research Institute, which is a zirconium oxide, cerium added 10 _mol%, and a Ce _{0.9 Zr} 0.1 _{O 2} gas sensor (CeZr10) (Sens. Actuators B 108 (2005) 238 -243). Therefore, the present gas sensor is a bulk response type semiconductor sensor. This gas sensor was used at an operating temperature of 400°C.

In this example, as in Example 1, the response values S of the eight types of gas sensors were measured, and then the principal component analysis was performed using the response values S of the eight types of gas sensors.

FIG. 3 shows the principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component. The contribution rate of each principal component is shown on each axis.

From FIG. 3, it can be seen that the cumulative contribution rate up to the third principal component is 93.2%. This indicates that the value of 7.456 types out of the 8 types of gas sensors is collected. In FIG. 3, the size of the vector is tripled in order to make the eigenvector easy to understand. As shown in Fig. 3, acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, show overlapping main component scores, but the main component scores between acetoin, nonanal, decane, and acetone-methyl isobutyl ketone are separated. Therefore, it can be seen that the separation detection can be sufficiently performed.

The principal component score of the target VOC varied in the axial direction of the second principal component. As in Example 1, it can be said that the second main component tends to be derived from the total number of carbon atoms in VOC. On the other hand, the first main component is in the order of acetoin, nonanal, acetone and methyl isobutyl ketone, and decane from the one having the highest main component score, as in the case of Example 1. It can be said that the tendency tends to be derived from the presence or absence of a functional group.

In the principal component scores of the first principal component and the second principal component, TGS2602, #31b, #33b, No. 2 which shows eigenvectors substantially parallel to the axis of the first principal component. 9, No. 71 will be in charge. On the other hand, in Comparative Example 1 which will be described later, not only acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, but also decane, which is not similar in structure, overlap in the main component scores. From this, the bulk response type No. 9, No. It is considered that 71 particularly contributes to the separation of the ketone system and the aliphatic hydrocarbon.

Also, regarding the main component scores of the first component and the third main component, in Comparative Example 1 described later, an overlap of the main component scores of acetone/methylisobutylketone and decane was observed, but in Example 2, acetone was used. Not only the main component scores of methyl isobutyl ketone and decane are separated, but the difference in the main component scores is larger than that in Example 1. No. The eigenvector of 71 is parallel to the direction of its separation, and is considered to particularly contribute to the separation of the ketone system and the aliphatic hydrocarbon. Here, No. The eigenvector of No. 9 is also No. Since it is almost parallel to the eigenvector of 71, it is considered that the separability of the principal component scores is increased by adding one type of bulk response type gas sensor.

[Example 3]
In this embodiment, four kinds of grain boundary responsive semiconductor type gas sensors (TGS2602, TGS2610 (above, manufactured by Figaro Giken Co., Ltd.), #31b, #33b) and one kind of bulk response type semiconductor gas sensor (No. 71). Was used. The third embodiment corresponds to the first embodiment with the TGS2600 and TGS2620 removed.

In this example, as in Example 1, the response values S of the five gas sensors were measured, and then the principal component analysis was performed using the response values S of the five gas sensors.

FIG. 4 shows the principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component. The contribution rate of each principal component is shown on each axis.

It can be seen from FIG. 4 that the cumulative contribution rate up to the third principal component is 97.6%. This indicates that the value of 4.88 types among the 5 types of gas sensors is collected. In FIG. 4, the size of the vector is tripled in order to make the eigenvector easy to understand. As shown in Fig. 4, acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, have overlapping main component scores, but the main component scores between acetoin, nonanal, decane, and acetone/methyl isobutyl ketone are separated. Therefore, it can be seen that the separation detection can be sufficiently performed.

In addition, the variation in the axial direction of the second main component depending on the TVOC concentration was reduced as compared with the first embodiment. This is because TGS2600 and TGS2620, which show an eigenvector nearly parallel to the axis of the second principal component and strongly show the influence of the concentration of TVOC, were removed and an appropriate sensor was selected according to the purpose.

In the principal component scores of the first principal component and the second principal component, TGS2602, #31b, #33b, No. 2 which shows eigenvectors substantially parallel to the axis of the first principal component. 71 will be in charge. On the other hand, in Comparative Example 2 to be described later, not only acetone and methylisobutylketone, which are ketone-based and very similar in structure, but also decane, which is not similar in structure, overlap in the main component scores. From this, the bulk response type No. It is considered that 71 particularly contributes to the separation of the ketone system and the aliphatic hydrocarbon.

Also, regarding the main component scores of the first main component and the third main component, in Comparative Example 2 described later, the main component scores of acetone/methylisobutylketone and decane were found to overlap, but in Example 3, The main component scores of acetone/methyl isobutyl ketone and decane are separated. Here, No. The eigenvector of 71 is parallel to the direction of its separation, and is considered to particularly contribute to the separation of the ketone system and the aliphatic hydrocarbon.

[Example 4]
In this embodiment, four kinds of grain boundary responsive semiconductor type gas sensors (TGS2602, TGS2610 (above, manufactured by Figaro Giken Co., Ltd.), #31b, #33b) and two kinds of bulk response type semiconductor gas sensors (No. 9, No. 71) was used. In addition, Example 4 corresponds to Example 2 from which TGS 2600 and TGS 2620 are removed. In addition, the fourth embodiment is the same as the third embodiment. It is also equivalent to the addition of 9.

In this example, as in Example 1, the response values S of the six types of gas sensors were measured, and then the principal component analysis was performed using the response values S of the six types of gas sensors.

FIG. 5 shows the principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component. The contribution rate of each principal component is shown on each axis.

From FIG. 5, it can be seen that the cumulative contribution rate up to the third principal component is 95.6%. This indicates that out of the 6 types of gas sensors, the value of 5.736 types is collected. In FIG. 5, the size of the vector is tripled in order to make the eigenvector easy to understand. As shown in Fig. 5, although acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, have overlapping main component scores, the main component scores between acetoin, nonanal, decane, and acetone-methyl isobutyl ketone are separated. Therefore, it can be seen that the separation detection can be sufficiently performed.

In addition, the variation in the axial direction of the second main component depending on the TVOC concentration was smaller than that in the second embodiment. This is because the information of TGS2600 and TGS2620 showing an eigenvector nearly parallel to the axis of the second principal component and strongly showing the influence of the concentration of TVOC is removed, and an appropriate sensor is selected according to the purpose. ..

In the principal component scores of the first principal component and the second principal component, TGS2602, #31b, #33b, No. 2 which shows eigenvectors substantially parallel to the axis of the first principal component. 9, No. 71 will be in charge. On the other hand, in Comparative Example 2 to be described later, not only acetone and methylisobutylketone, which are ketone-based and very similar in structure, but also decane, which is not similar in structure, overlap in the main component scores. From this, the bulk response type No. 9, No. It is considered that 71 particularly contributes to the separation of the ketone system and the aliphatic hydrocarbon.

Also, regarding the main component scores of the first main component and the third main component, in Comparative Example 2 described later, the main component scores of acetone/methylisobutylketone and decane were overlapped, but in Example 4, Not only are the main component scores of acetone/methylisobutylketone and decane separated, but the difference in the main component scores is larger than in Example 3. Here, No. The eigenvector of 71 is parallel to the direction of its separation, and is considered to particularly contribute to the separation of the ketone system and the aliphatic hydrocarbon. In addition, No. The eigenvector of No. 9 is also No. Since it is almost parallel to the eigenvector of 71, it is considered that the separability of the principal component scores is increased by adding one type of bulk response type gas sensor.

[Example 5]
In this example, three types of grain boundary responsive semiconductor type gas sensors (TGS2602 (manufactured by Figaro Giken Co., Ltd., #31b, #33b) and one type of bulk response type semiconductor gas sensor (No. 71) were used. Example 5 corresponds to Example 3 with the TGS 2610 removed.

In this example, as in Example 1, the response values S of the four gas sensors were measured, and then the principal component analysis was performed using the response values S of the four gas sensors.

FIG. 6 shows principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component. The contribution rate of each principal component is shown on each axis.

From FIG. 6, it can be seen that the cumulative contribution rate up to the third principal component is 99.9%. This indicates that the value of 3.996 types among the 4 types of gas sensors is collected. In FIG. 6, the size of the vector is tripled in order to make the eigenvector easy to understand. As shown in Fig. 6, acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, show overlapping main component scores, but the main component scores between acetoin, nonanal, decane, and acetone-methyl isobutyl ketone are separated. Therefore, it can be seen that the separation detection can be sufficiently performed.

In addition, the variation in the axial direction of the second main component depending on the TVOC concentration was reduced as compared with the third embodiment. This is because TGS2610, which shows an eigenvector nearly parallel to the axis of the second principal component and strongly shows the influence of the concentration of TVOC, is removed, and an appropriate sensor is selected according to the purpose.

In the principal component scores of the first principal component and the second principal component, TGS2602, #31b, #33b, No. 2 which shows eigenvectors substantially parallel to the axis of the first principal component. 71 will be in charge. On the other hand, in Comparative Example 3 to be described later, not only acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure to each other, but also decane having a dissimilar structure overlaps in the main component scores. From this, the bulk response type No. It is considered that 71 particularly contributes to the separation of the ketone system and the aliphatic hydrocarbon.

Also, regarding the main component scores of the first main component and the third main component, in Comparative Example 3 described later, the overlap of the main component scores of acetone/methylisobutylketone and decane was observed, but in Example 5, The main component scores of acetone/methyl isobutyl ketone and decane are separated. Here, No. The eigenvector of 71 is parallel to the direction of its separation, and is considered to particularly contribute to the separation of the ketone system and the aliphatic hydrocarbon.

[Example 6]
In this embodiment, three types of grain boundary responsive semiconductor type gas sensors (TGS2602 (manufactured by Figaro Giken Co., Ltd., #31b, #33b) and two types of bulk response type semiconductor gas sensors (No. 9, No. 71). Was used. Example 6 corresponds to Example 4 with the TGS 2610 removed. In addition, the sixth embodiment is the same as the fifth embodiment. It is also equivalent to the addition of 9.

In this example, as in Example 1, the response values S of the five gas sensors were measured, and then the principal component analysis was performed using the response values S of the five gas sensors.

FIG. 7 shows the principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component. The contribution rate of each principal component is shown on each axis.

From FIG. 7, it can be seen that the cumulative contribution rate up to the third principal component is 96.1%. This indicates that the value of 4.805 types among the 5 types of gas sensors is collected. In FIG. 7, the size of the vector is tripled in order to make the eigenvector easy to understand. As shown in Fig. 7, although acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, have overlapping main component scores, the main component scores between acetoin, nonanal, decane, and acetone-methyl isobutyl ketone are separated. Therefore, it can be seen that the separation detection can be sufficiently performed.

In addition, the variation in the axial direction of the second main component depending on the TVOC concentration was reduced as compared with the fourth embodiment. This is because the information of TGS2610 which shows the eigenvector nearly parallel to the axis of the second principal component and which strongly shows the influence of the concentration of TVOC is removed, and an appropriate sensor is selected according to the purpose.

In the principal component scores of the first principal component and the second principal component, TGS2602, #31b, #33b, No. 2 which shows eigenvectors substantially parallel to the axis of the first principal component. 9, No. 71 will be in charge. On the other hand, in Comparative Example 3 to be described later, not only acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure to each other, but also decane having a dissimilar structure overlaps in the main component scores. From this, the bulk response type No. 9, No. It is considered that 71 particularly contributes to the separation of the ketone system and the aliphatic hydrocarbon.

Also, regarding the main component scores of the first main component and the third main component, in Comparative Example 3 described later, the main component scores of acetone/methylisobutylketone and decane were overlapped, but in Example 6, Not only the main component scores of acetone/methylisobutylketone and decane were separated, but the difference in the main component score was larger than in Example 5. Here, No. The eigenvector of 71 is parallel to the direction of its separation, and is considered to particularly contribute to the separation of the ketone system and the aliphatic hydrocarbon. In addition, No. The eigenvector of No. 9 is also No. Since it is almost parallel to the eigenvector of 71, it is considered that the separability of the principal component scores is increased by adding one type of bulk response type gas sensor.

[Comparative Example 1]
In this comparative example, six kinds of grain boundary responsive semiconductor gas sensors (TGS2600, TGS2602, TGS2610, TGS2620 (above, manufactured by Figaro Giken Co., Ltd.), #31b, #33b) were used. The comparative example 1 corresponds to the one obtained by removing the bulk response type semiconductor gas sensor from the example 1 or the example 2.

In this comparative example, first, the response values S of 6 types of gas sensors were measured in the same manner as in Example 1 except that the humidity was changed to correspond to 0% RH in terms of 20° C., and then 6 types of Principal component analysis was performed using the response value S of the gas sensor. At this time, all of nitrogen and oxygen did not pass through the water bubbler 2. Next, in the same manner as in Example 1, the response values S of the six types of gas sensors were measured, and then the principal component analysis was performed using the response values S of the six types of gas sensors.

FIG. 8 and FIG. 9 show the principal component scores of the first principal component and the second principal component, the first principal component and the third principal component when the humidity corresponds to 0% RH and 60% RH at 20° C., respectively. And eigenvectors are shown. The contribution rate of each principal component is shown on each axis.

From FIGS. 8 and 9, it can be seen that the cumulative contribution rates up to the third principal component are 94.2% and 98.3%, respectively. This indicates that the value of 5.652 types and the value of 5.898 types of 6 types of gas sensors are aggregated. In FIGS. 8 and 9, the size of the vector is tripled for easy understanding of the eigenvector.

As shown in FIG. 8, in the main component scores by the first main component and the second main component, although the main component scores of methyl isobutyl ketone and acetone and acetone and decane are close to each other, the overlap of the main component scores is not found. It can be seen that separation detection can be performed without being carried out. Also, regarding the principal component scores of the first principal component and the third principal component, although the principal component scores of nonanal and acetone are close to each other, no overlapping of the principal component scores is seen, and it can be seen that separation detection is possible.

However, as shown in FIG. 9, when the humidity corresponds to 60% RH at 20° C., the separation of the main component scores is not sufficient. When the humidity corresponds to 60% RH in terms of 20° C., the eigenvector and the principal component score greatly change. As a result, in the principal component scores of the first and second principal components, not only acetone and methyl isobutyl ketone, which are ketone-based and very similar in structure, but also decane having a dissimilar structure, overlaps in the principal component scores. Was given. Also, in the main component scores of the first and third main components, overlapping of the main component scores of acetone/methylisobutylketone/decane/acetoin was observed. This is because the mechanism of response in the grain boundary responsive semiconductor type gas sensor is that oxygen adsorbed on the surface is consumed by the oxidation of VOC, but when the humidity corresponds to 60% RH at 20° C., the adsorption of oxygen It is considered that this is because the sensitivity decreases due to the adsorption of water molecules on a part of the sites, which reduces the difference in characteristics between the grain boundary responsive semiconductor gas sensors and reduces the amount of information. That is, it can be seen that the six kinds of grain boundary responsive semiconductor gas sensors used in this comparative example cannot sufficiently separate and detect under the environment where the humidity, which is the same as in nature, is 60% RH in terms of 20°C.

[Comparative example 2]
In this comparative example, four kinds of grain boundary responsive semiconductor gas sensors (TGS2602, TGS2610 (above, manufactured by Figaro Giken Co., Ltd., #31b, #33b) were used. In addition, Comparative Example 2 corresponds to the one obtained by removing the bulk response type semiconductor gas sensor from Example 3 or Example 4. Comparative Example 2 also corresponds to Comparative Example 1 with the TGS 2600 and TGS 2620 removed.

In the same manner as in Example 1, the response values S of the four gas sensors were measured, and then the principal component analysis was performed using the response values S of the four gas sensors.

FIG. 10 shows principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component. The contribution rate of each principal component is shown on each axis.

It can be seen from FIG. 10 that the cumulative contribution rate up to the third principal component is 100%. This shows that virtually all of the values of the four gas sensors are aggregated. In FIG. 10, the size of the vector is tripled in order to make the eigenvector easy to understand.

As shown in FIG. 10, in the main component score by the first main component and the second main component, the difference between the main component scores of nonanal and acetoin is larger than that in Comparative Example 1, and the first main component and the third main component The difference in the main component scores of acetoin and acetone/methylisobutylketone/decane was larger than that of Comparative Example 1 in the main component score. However, not only acetone and methylisobutylketone, which have very similar structures in the ketone system, but also decane, which has a dissimilar structure, showed overlapping main component scores. That is, it can be seen that even in this comparative example, separation and detection cannot be sufficiently performed.

[Comparative Example 3]
In this comparative example, three kinds of grain boundary responsive semiconductor type gas sensors (TGS2602 (manufactured by Figaro Giken Co., Ltd., #31b, #33b) were used. In addition, Comparative Example 3 corresponds to the one obtained by removing the bulk response type semiconductor gas sensor from Example 5 or Example 6. In addition, Comparative Example 3 corresponds to Comparative Example 2 with the TGS 2610 removed.

In the same manner as in Example 1, the response values S of the three types of gas sensors were measured, and then the principal component analysis was performed using the response values S of the three types of gas sensors.

FIG. 11 shows the principal component scores and eigenvectors of the first principal component and the second principal component, and the first principal component and the third principal component. The contribution rate of each principal component is shown on each axis.

As is clear from FIG. 11, the principal component analysis is performed by three types of grain boundary response type semiconductor gas sensors, and since the third principal component is included, the cumulative contribution rate up to the third principal component is 100%. Note that in FIG. 11, the size of the vector is tripled in order to make the eigenvector easy to understand.

As shown in FIG. 11, although the variation in the axial direction of the second main component depending on the TVOC concentration was smaller than that in Comparative Example 2, the main component scores of acetone/methylisobutylketone/decane were overlapped. That is, it can be seen that even in this comparative example, separation and detection cannot be sufficiently performed.

1 Sensor Room 2 Water Bubbler 3 Gas Generator 4 Valve 5 Flow Meter with Needle Valve 6 Three-way Valve 7 Crossover Four-way Valve

Claims (4)

A step of detecting a combustible gas using a gas sensor group,
Comprising a step of principal component analysis of the response value obtained by detecting the flammable gas,
The gas sensor group includes a plurality of types of grain boundary response type semiconductor gas sensors and one or more types of bulk response type semiconductor gas sensors,
The semiconductor gas sensor of the bulk response type, cerium oxide or cerium - see containing zirconium-based composite oxide,
The one or more bulk response type semiconductor gas sensors include a carrier of the cerium-zirconium-based composite oxide in which platinum particles are supported on the cerium-zirconium-based composite oxide film via an alumina film. A method for analyzing flammable gas, comprising: The flammable gas analysis method according to claim 1, wherein the plurality of types of grain boundary responsive semiconductor gas sensors include a sensor including tin oxide particles. The cerium oxide or cerium - zirconium based compound oxide represented by the composition formula Ce _{1-x Zr} x _{O 2}
(In the formula, x is in the range of 0 to 0.4.)
The analysis method of the combustible gas according to claim 1 or 2, characterized in that in a compound represented. The flammable gas analysis method according to any one of claims 1 to 3, wherein the flammable gas contained in the breath of the patient is detected.

Images