Metal oxide semiconductor gas sensor

Abstract

This metal oxide semiconductor gas sensor has: a first electrode and a second electrode; and a sensitive layer in contact with both the first electrode and the second electrode. The sensitive layer includes an n-type semiconductor and a p-type semiconductor. When the conduction band lowest energy level of the p-type semiconductor at 300K is pCB, the conduction band lowest energy level of the n-type semiconductor at 300K is nCB, and the volume fraction of the p-type semiconductor to the total volume of the p-type and n-type semiconductors is α, the value of ΔCB=(pCB×α)−(nCB×(1−α)) is at least 1.8 eV.　

Description

Metal oxide semiconductor gas sensor

[0001] The present invention relates to a metal oxide semiconductor gas sensor.

[0002] Carbon monoxide (CO) and nitrogen dioxide (NO 2 Since carbon monoxide (CO) is a product of incomplete combustion and is highly toxic, it can poison catalysts as an impurity in hydrogen fuel, so a high-performance sensor is needed to monitor its concentration.

[0003] In Patent Document 1 shown below, 2-x Ce X BO 4 The publication describes a gas sensor material for detecting CO using the compound. However, the sensor described in this publication has insufficient selectivity and sensitivity to CO gas, and further improvements are required.

[0004] There is also a demand for a metal oxide semiconductor gas sensor that can effectively detect VOC (Volatile Organic Compounds) gases such as ethanol gas and acetone gas.

[0005] Patent No. 6586694

[0006] A first object of the exemplary embodiment of the present invention is to provide a metal oxide semiconductor gas sensor capable of effectively detecting VOC (Volatile Organic Compounds) gases such as ethanol gas, acetone gas, etc. A second object is to provide a metal oxide semiconductor gas sensor capable of detecting carbon monoxide (CO gas) contained in low concentrations in a measurement gas.

[0007] In order to achieve the first object, a metal oxide semiconductor gas sensor according to an exemplary embodiment of the present invention includes a first electrode, a second electrode, and a sensitive layer in contact with both the first electrode and the second electrode, wherein the sensitive layer includes an n-type semiconductor and a p-type semiconductor, and the lowest energy level of the conduction band of the p-type semiconductor at 300 K is p CB The lowest energy level of the conduction band of the n-type semiconductor at 300 K is nCB When the volume ratio of the p-type semiconductor to the total volume of the p-type semiconductor and the n-type semiconductor is α, ΔCB=(p CB × α)-(n CB ×(1−α)) is 1.8 eV or more.

[0008] The n-type semiconductor may have a band gap smaller than 3.4 eV at 300K, and the p-type semiconductor may have a band gap larger than 1.75 eV at 300K.

[0009] The band gap of the n-type semiconductor at 300K may be 2.4 eV or more and 5.0 eV or less, and the band gap of the p-type semiconductor at 300K may be 1.0 eV or more and 5.0 eV or less.

[0010] In order to achieve the second object, a metal oxide semiconductor gas sensor according to another exemplary embodiment of the present invention comprises a first electrode, a second electrode, and a sensitive layer in contact with both the first electrode and the second electrode, wherein the sensitive layer includes an n-type semiconductor and a p-type semiconductor, the n-type semiconductor having a band gap of 2.4 eV or more and 5.0 eV or less at 300 K, and the p-type semiconductor having a band gap of 1.0 eV or more and 5.0 eV or less at 300 K.

[0011] In the sensitive layer, the surface area ratio occupied by the n-type semiconductor may be 47.9% or more and 99.5% or less.

[0012] In a cross section of the sensitive layer, the coefficient of variation of element distribution of the n-type semiconductor may be 0.9 or less.

[0013] In a cross section of the sensitive layer, the coefficient of variation of element distribution of the p-type semiconductor may be 0.9 or less.

[0014] The sensitive layer may have an average porosity of 15.6% or more and 22.0% or less in cross section.

[0015] The sensitive layer may have an average void size in a cross section of 52 nm or more and 145 nm or less.

[0016] When the sensitive layer is divided into three equal parts in the thickness direction, the average porosity in the region closest to the first electrode and the second electrode may be φn (%), and the average porosity in the region farthest from the first electrode and the second electrode may be φf (%), where φf-φn≧0.4.

[0017] The thickness of the sensitive layer may be 0.4 μm or more.

[0018] Fig. 1 is a plan view of a metal oxide semiconductor gas sensor according to an embodiment of the present invention, with the sensitive layer omitted. Fig. 2 is a plan view of a metal oxide semiconductor gas sensor according to an embodiment of the present invention. Fig. 3 is a cross-sectional view of the metal oxide semiconductor gas sensor shown in Fig. 2, taken along line III-III. Fig. 4 is a schematic diagram showing an EDS mapping image of a cross section of the sensitive layer of a metal oxide semiconductor gas sensor according to an example of the present invention. Fig. 5 is a schematic diagram showing the energy levels of the conduction band minimums of n-type and p-type semiconductors at 300 K.

[0019] First Embodiment A metal oxide semiconductor gas sensor according to an embodiment of the present invention will be described below. The metal oxide semiconductor gas sensor of this embodiment is a gas sensor that can mainly detect VOC gases such as ethanol gas and acetone gas.

[0020] As shown in Figures 1 and 3, the metal oxide semiconductor gas sensor 10 has a substrate 1 and a first electrode 2 and a second electrode 3 formed on the substrate 1. In this embodiment, the first electrode 2 and the second electrode 3 are shaped like comb teeth and face each other at a predetermined distance, but the shape is not particularly limited as long as they face each other at a predetermined distance. As shown in Figures 1 to 3, the sensitive layer 4 is in contact with both the first electrode 2 and the second electrode 3 so as to cover the portions of the electrodes 2 and 3 other than the lead-out electrode portions. The sensitive layer 4 is formed on the surface of the substrate 1 on which the first electrode 2 and the second electrode 3 are formed.

[0021] The sensitive layer 4 is made of a metal oxide material and contains a plurality of pores, the average porosity of which is preferably 15.6% to 22.0%, more preferably 15.9% to 22.0%, and even more preferably 16.0% to 21.5%, although these are not shown in Fig. 3, which shows the cross section of the sensitive layer 4. There are no particular limitations on the method for measuring the average porosity.

[0022] For example, the ratio of the total area of ​​multiple voids to the total area of ​​at least one cross section of the sensitive layer 4 may be considered the average porosity. The lower limit of the size recognized as a void is, for example, 3 nm or more. In order to calculate the average, it is preferable that the area within a predetermined range of the cross section of the sensitive layer 4 is a cross section of the sensitive layer in which at least 100 voids are observed. There are no particular restrictions on the method for observing the cross section, as long as it can be observed in a manner that allows the metal oxide material and voids to be distinguished. For example, the average porosity can be calculated by observing the cross section of the sensitive layer 4 using an SEM, optical microscope, or the like. In the cross section of the sensitive layer 4 in Figure 4, the white areas represent voids 5.

[0023] There are no particular limitations on the method for setting at least one cross section of the sensitive layer 4, as long as the average porosity of the sensitive layer 4 can be measured appropriately. For example, the total area ratio of voids in a cross section parallel to the thickness direction of the sensitive layer 4 can be calculated. The average porosity is preferably determined by averaging the area of ​​the sensitive layer 4 in the thickness direction, in the region in contact with the substrate 1 and the region on the surface side exposed to the outside.

[0024] By keeping the average porosity of the sensitive layer 4 and the proportions of each component within the specified ranges, a metal oxide semiconductor gas sensor with high sensor sensitivity to VOC gases and a short response time can be obtained. Furthermore, by appropriately controlling the average porosity of the sensitive layer 4, VOC gases can easily penetrate into the pores, making it easier for oxygen adsorbed on the surface of the oxide to react with the gas, making it easier to lower the height of the spatial charge barrier, reducing resistance, further improving sensitivity, and contributing to a shorter response time.

[0025] The average pore size of the sensitive layer 4 is preferably 52 nm or more and 145 nm or less, more preferably 55 nm or more and 75 nm or less, and particularly preferably 55 nm or more and 70 nm or less. There are no particular limitations on the method for measuring the average pore size. For example, the equivalent circle diameter of each pore in a cross section parallel to the thickness direction of the sensitive layer 4 may be calculated and averaged. The average pore size of the sensitive layer 4 can be defined as the average size of all pores observed in a cross section of the sensitive layer 4 within a specified area range. The lower limit of the size recognized as a pore is the same as that for the average porosity.

[0026] In order to calculate the average void size, the area within a predetermined range of the cross section of the sensitive layer 4 is preferably the same as that for the average porosity. That is, the area within a predetermined range of the cross section of the sensitive layer 4 is preferably a cross section of the sensitive layer in which at least 100 voids are observed. In addition, in order to calculate the average void size, it is preferable to obtain an average along the thickness direction of the sensitive layer 4 in the region in contact with the substrate 1 and the region on the surface side exposed to the outside. When calculating the average void size, voids of 52 nm or less or 122 nm or more may also be counted, but it is preferable that voids outside these ranges account for 10% or less of the total number of voids observed.

[0027] The average pore size is preferably determined based on the mean free path (55 to 75 nm) of VOC gases such as ethanol gas and acetone gas. By setting the average pore size within the above range, it becomes easier to further increase the sensor sensitivity to VOC gases and further shorten the response time.

[0028] As shown in Fig. 4, the sensitive layer 4 includes a first oxide 4a of an n-type semiconductor and a second oxide 4b of a p-type semiconductor. In Fig. 4, the first oxide 4a and the second oxide 4b are depicted as connected, but in reality, they exist separately as particles. Here, as shown in Fig. 5, the lowest energy level of the conduction band of a p-type semiconductor at 300 K is defined as p CB The lowest energy level of the conduction band of an n-type semiconductor at 300 K is n CBIn addition, when the volume ratio of the p-type semiconductor to the total volume of the p-type semiconductor and the n-type semiconductor is α, in this embodiment, ΔCB=(p CB × α)-(n CB × (1-α)) is 1.8 eV or more, preferably 2 or more, 3 or more, or 4 or more. Within such a range, the detection sensitivity of VOC gases such as ethanol gas and acetone gas is improved.

[0029] In this embodiment, the band gap n of the n-type semiconductor at 300K is BG (see FIG. 5) may be 2.4 eV or more and 5.0 eV or less, and the band gap n BG (See FIG. 5) may be 1.0 eV or more and 5.0 eV or less. By being in such a range, a metal oxide semiconductor gas sensor having high sensor sensitivity to VOC gases, a short response time, and excellent gas selectivity can be obtained.

[0030] Alternatively, in this embodiment, it is preferable that the band gap of the n-type semiconductor is smaller than 3.4 eV at 300 K, and the band gap of the p-type semiconductor is larger than 1.75 eV at 300 K. When these values ​​are in this range, the selective detection sensitivity of acetone gas over ethanol gas, among VOC gases, is likely to be improved.

[0031] Examples of n-type semiconductor oxides having the above band gap include SnO 2 , W.O. 3 , V 2 O 5 , Sb 2 O 3 , TiO 2 (both rutile and anatase), Cr 2 O 3 , ZnO, Nb 2 O 5 , BaO, In 2 O 3 , Ta 2 O 5 , Ga 2 O 3 At least one of these oxides can be used alone or in combination. For VOC gas detection, the n-type semiconductor oxide is preferably SnO. 2 , W.O. 3 , V 2 O 5 , ZnO, etc., and at least one of these oxides can be used alone or in combination.

[0032] Furthermore, examples of oxides of p-type semiconductors having the above band gap include CuO and MnO 2 , Co 3 O 4 , Cu 2 O, Mn 2 O 3 , U.O. 2 , CoO, PbO, Bi 2 O 3 , NiO, MnO, etc., and at least one of these oxides can be used alone or in combination. For VOC gas detection, the oxide of the p-type semiconductor is preferably CuO, Co 3 O 4 , NiO, etc., and at least one of these oxides can be used alone or in combination.

[0033] In the sensitive layer 4, the surface area ratio of the n-type semiconductor first oxide 4a in the sensitive layer 4 is preferably 47.9% to 99.5%, more preferably 50% to 99%. Also, the surface area ratio of the p-type semiconductor second oxide 4b in the sensitive layer 4 is preferably 0.5% to 52.1%, more preferably 1% to 50%. When the ratios are within these ranges, the response time can be shortened, and particularly the sensor sensitivity and gas selectivity can be improved.

[0034] The surface area ratio of the first oxide in the sensitive layer 4 is, in other words, the ratio of n-type semiconductors in the metal oxide material that constitutes the sensitive layer 4. The surface area ratio of the second oxide in the sensitive layer 4 is, in other words, the ratio of p-type semiconductors in the metal oxide material that constitutes the sensitive layer 4. Note that the sensitive layer 4 does not include the first electrode 2 or the second electrode 3.

[0035] There are no particular limitations on the method for measuring the surface area ratio of the n-type semiconductor and the surface area ratio of the p-type semiconductor. For example, the sensitive layer 4 can be cut out from the metal oxide semiconductor gas sensor 10, and a mapping image of each element (oxide) such as that shown in FIG. 4 can be created using STEM-EDS or the like, and the surface area ratio of each can be determined based on the mapping image. Specifically, the perimeters of the p-type and n-type semiconductors can be calculated from the mapping image, and the squares of these can be expressed as the surface area ratio. In this case, a visual field containing at least 50 or more p-type and n-type particles can be selected, and the total surface area ratio can be calculated by observing at least 10 visual fields.

[0036] Regarding the method for measuring the volume fraction of the n-type semiconductor and the volume fraction of the p-type semiconductor, the measurement results of the surface area fraction of the n-type semiconductor and the surface area fraction of the p-type semiconductor may be converted into volumes, or alternatively, the weight fractions at the time of weighing may be used to convert into volumes.

[0037] The coefficient of variation of the element distribution of the n-type semiconductor in the cross section of the sensitive layer 4 is preferably 0.9 or less. The coefficient of variation of the element distribution of the p-type semiconductor in the cross section of the sensitive layer 4 is preferably 0.9 or less. The coefficient of variation of the element distribution can be evaluated, for example, by the method shown below.

[0038] An STEM image with a field of view of 200 nm × 200 nm or more is obtained in a cross section parallel to the thickness direction of the sensitive layer 4. The obtained STEM image is divided into 1 nm × 1 nm square regions and subjected to EDS analysis, and data is obtained on the intensities of elements derived from the n-type semiconductor and elements derived from the p-type semiconductor in each region. The coefficient of variation of the element distribution can be calculated based on the data obtained for all regions (40,000 points).

[0039] There is no particular limit to the thickness of the sensitive layer 4. For example, it may be 0.4 μm or more, or 1.0 μm or more. When the sensitive layer 4 is thicker than a predetermined thickness, the sensor sensitivity to VOC gases is likely to be improved. There is no particular upper limit to the thickness of the sensitive layer 4. For example, it may be 20 μm or less. Note that, when the first electrode 2 and the second electrode 3 are embedded in the sensitive layer 4 as shown in FIG. 3, the thickness of the sensitive layer 4 is the thickness of the portion that does not include the first electrode 2 and the second electrode 3. Specifically, the thickness of the portion indicated as d in FIG. 3 is the thickness of the sensitive layer 4.

[0040] 3, when the sensitive layer 4 is divided into three equal parts in the thickness d direction, where φn (%) is the average porosity in the region d1 closest to the first electrode 2 and the second electrode 3, and φf (%) is the average porosity in the region d2 farthest from the first electrode 2 and the second electrode 3, it is preferable that φf - φn ≧ 0.4, and it is even more preferable that φf - φn ≧ 1.0. That is, it is preferable that the average porosity is high in the portion of the sensitive layer 4 that mainly comes into contact with the gas to be measured (the outer surface side of the sensitive layer 4) and in the vicinity thereof. By setting φf - φn to a predetermined value or more, it becomes easier to further increase the sensor sensitivity to VOC gases and further shorten the response time.

[0041] There are no particular limitations on the method for measuring φn and φf. For example, an image of a cross section parallel to the thickness direction of the sensitive layer 4 and not including the first electrode 2 and the second electrode 3, observed with an SEM or the like, may be divided into three equal parts in the thickness direction. φn may be the total area ratio of voids in the image corresponding to the region closest to the first electrode 2 and the second electrode 3 (the region closest to the substrate 1). φf may be the total area ratio of voids in the region farthest from the first electrode 2 and the second electrode 3.

[0042] The sensitive layer 4 may contain components other than the above-mentioned metal oxides as long as they do not significantly affect the sensor sensitivity, response time, and gas selectivity to VOC gases. The total area ratio of the components other than the metal oxides in the sensitive layer 4 may be 5% or less.

[0043] The structure and manufacturing method of the metal oxide semiconductor gas sensor 10 according to this embodiment will be described in more detail below.

[0044] 1 shows a metal oxide semiconductor gas sensor 10 according to this embodiment in a state before a sensitive layer 4 (VOC gas sensitive layer) is formed. A first electrode 2 and a second electrode 3 are formed on a substrate 1.

[0045] The type of substrate 1 is not particularly limited. For example, it may be a substrate made of an insulating material that is heat-resistant to the operating temperature of the metal oxide semiconductor gas sensor 10. Specific examples of the substrate 1 that can be used include an electrically insulating ceramic substrate and a substrate with a thermal oxide film. Examples of the electrically insulating ceramic substrate include an alumina substrate and a zirconia substrate. Examples of the substrate with a thermal oxide film include a silicon substrate with a thermal oxide film.

[0046] The substrate 1 is not an essential component of the metal oxide semiconductor gas sensor 10. The metal oxide semiconductor gas sensor 10 only needs to have at least the first electrode 2, the second electrode 3, and the sensitive layer 4 in contact with both the first electrode 2 and the second electrode 3 in a manner that allows the sensor to function as a gas sensor.

[0047] There are no particular limitations on the materials of the first electrode 2 and the second electrode 3. It is sufficient that the first electrode 2 and the second electrode 3 are made of an electrically conductive material. For example, Pt, Au, etc. may be used as the material of the first electrode 2 and the second electrode 3.

[0048] There are no particular limitations on the shapes of the first electrode 2 and the second electrode 3. To increase the contact area with the sensitive layer 4, the first electrode 2 and the second electrode 3 may have, for example, a comb-like shape as shown in FIG. 1 . This is preferable because it is easier to increase the contact area between the first electrode 2 and the second electrode 3 and the sensitive layer 4. As shown in FIG. 1 , the first electrode 2 and the second electrode 3 are preferably arranged to face each other. There are no particular limitations on the distance between the first electrode 2 and the second electrode 3. Considering the need to miniaturize the metal oxide semiconductor gas sensor 10, it is preferable that the distance between the first electrode 2 and the second electrode 3 is small. For example, it may be 50 μm or less.

[0049] There are no particular limitations on the method for forming the first electrode 2 and the second electrode 3. For example, they can be formed by sputtering, vacuum deposition, screen printing, or the like.

[0050] After the first electrode 2 and the second electrode 3 are formed, the sensitive layer 4 may be formed on the first electrode 2 and the second electrode 3 as shown in Fig. 2. There are no particular restrictions on the shape, size, etc. of the sensitive layer 4. It is sufficient that the sensitive layer 4 is in contact with at least both the first electrode 2 and the second electrode 3. For example, as shown in Fig. 2, the sensitive layer 4 may be formed so as to cover the first electrode 2 and the second electrode 3.

[0051] Fig. 3 shows a cross-sectional view taken along line III-III in Fig. 2. As shown in Fig. 3, the sensitive layer 4 may be formed on the first electrode 2 and the second electrode 3 so as to fill the area between the first electrode 2 and the second electrode 3, in order to ensure a sufficient contact area with the first electrode 2 and the second electrode 3.

[0052] There are no particular limitations on the method for forming the sensitive layer 4. For example, the sensitive layer 4 can be formed by forming a paste of n-type and p-type semiconductors together with a binder, etc., applying the paste to the substrate 1 on which the first electrode 2 and the second electrode 3 are formed, and firing the paste. The firing may be performed so that the binder, etc., contained in the paste evaporates and the sensitive layer 4 contains both n-type and p-type semiconductors. However, the firing temperature is preferably such that rapid grain growth and necking between particles are not promoted. Rapid grain growth and necking between particles would reduce the surface area of ​​the sensitive layer 4 that comes into contact with the gas. Therefore, specifically, the firing temperature (binder removal temperature) is preferably 350°C to 600°C. Note that the higher the firing temperature, the smaller the φf-φn tends to be.

[0053] There are no particular limitations on the method for preparing the paste for forming the sensitive layer 4. For example, it can be prepared by mixing and stirring an n-type semiconductor and a p-type semiconductor in a vehicle prepared using a binder, a dispersant, an organic solvent, etc. As the binder, for example, a binder made of a polymer compound such as cellulose, ethyl cellulose, or hydroxyethyl cellulose can be used. As the dispersant, a polymer-type dispersant or a surfactant-type dispersant can be used. As the organic solvent, toluene, xylene, terpineol, ethylene glycol, etc. can be used. The higher the binder content, the higher the porosity of the final sensitive layer 4 tends to be. Furthermore, the longer the mixing and stirring time, the smaller the average pore size tends to be.

[0054] The metal oxide semiconductor gas sensor 10 of this embodiment has a predetermined sensitive layer 4, which improves the sensitivity, responsiveness, and gas selectivity of VOC gases. Specifically, the metal oxide semiconductor gas sensor 10 of this embodiment has high responsiveness and can detect VOC gases contained in the measurement gas at low concentrations of 30 ppm or less or 20 ppm or less. There is no particular lower limit for the detectable VOC gas concentration, but the sensor can generally detect VOC gases at concentrations of 50 ppb or more.

[0055] When the semiconductor sensitive layer 4 is exposed to a measurement gas containing VOC gases, the sensor resistance, i.e., the resistance between the first electrode 2 and the second electrode 3, changes. VOC gases can be detected from the change in the sensor resistance.

[0056] The principle behind the change in sensor resistance is explained below. In the case of many metal oxide semiconductor gas sensors, the sensor resistance when exposed to air is used as the reference. When exposed to air, oxygen, which has electron-withdrawing properties, is adsorbed to the surface of the semiconductor (sensitive layer) of a metal oxide semiconductor gas sensor. When the semiconductor is solely an n-type semiconductor, oxygen adsorbed to the semiconductor surface forms a space charge layer near the semiconductor surface. The formation of the space charge layer creates a potential barrier between the semiconductors, preventing electrons from moving between the semiconductors.

[0057] If an oxidizing gas is introduced onto the semiconductor surface when oxygen is adsorbed onto the surface, the space charge layer becomes thicker. This causes the sensor resistance to increase. If a reducing gas is introduced onto the semiconductor surface when oxygen is adsorbed onto the surface, the adsorbed oxygen is consumed, causing the space charge layer to become thinner. This causes the sensor resistance to decrease. As a result, the concentration of the gas to be measured can be detected from the change in sensor resistance.

[0058] When the semiconductor is solely a p-type semiconductor, the opposite reaction occurs to that when the semiconductor is an n-type semiconductor. When an oxidizing gas is introduced onto the semiconductor surface, the sensor resistance value decreases. When a reducing gas is introduced onto the semiconductor surface, the sensor resistance value increases.

[0059] In this embodiment, it is considered that the potential barrier is neither between n-type semiconductors nor between p-type semiconductors. That is, in this embodiment, it is considered that a pn junction is formed at the interface between the n-type oxide semiconductor and the p-type oxide semiconductor inside the sensitive layer 4. The formation of the pn junction causes holes and electrons to combine, resulting in a higher barrier and a wider electron depletion layer at the junction interface than in the past, increasing resistance in air. Furthermore, in a reducing gas or flammable gas, some of the oxygen adsorbed on the oxide surface is consumed by reaction with the gas, returning the captured electrons to the semiconductor, lowering the height of the space charge barrier and reducing resistance. As a result, it is considered that the sensor sensitivity is improved.

[0060] The sensor resistance value when the metal oxide semiconductor gas sensor 10 is exposed to air containing no VOC gas is defined as Ra, and the sensor resistance value when the metal oxide semiconductor gas sensor 10 is exposed to a measurement gas containing a specific VOC gas concentration is defined as Rg. The sensor sensitivity to a VOC gas having a specific concentration is defined as S. VOC as Rg / Ra or S' VOC The sensor sensitivity is expressed as the larger of Ra / Rg. The sensor sensitivity varies depending on the concentration of VOC gases.

[0061] When the metal oxide semiconductor gas sensor 10 using the sensitive layer 4 according to this embodiment is exposed to VOC gases, Ra>Rg is often satisfied.

[0062] After switching the gas supplied to the metal oxide semiconductor gas sensor 10, it takes some time for the sensor resistance value to reach a value corresponding to the gas supplied. Therefore, when measuring Ra, the sensor resistance value is continuously measured after starting to supply air to the metal oxide semiconductor gas sensor 10, and it is confirmed that the sensor resistance value has stabilized. The sensor resistance value when the sensor resistance value has stabilized is Ra. When measuring Rg, the sensor resistance value is continuously measured after starting to supply the measurement gas to the metal oxide semiconductor gas sensor 10, and it is confirmed that the sensor resistance value has stabilized. The sensor resistance value when the sensor resistance value has stabilized is Rg.

[0063] The absolute value of the change in sensor resistance value |(Ra - Rg)| from the sensor resistance value Ra in air to the sensor resistance value Rg in the measured gas is set to 100, and the 90% response time (T90) is the time from when the gas supplied to the metal oxide semiconductor gas sensor 10 is switched from air to the measured gas until the absolute value of the change in sensor resistance value reaches 90. The sensor resistance value at the time of the 90% response time is R, which can be expressed as R = {(Ra - Rg) × 0.1 + Rg} (Equation (1)). In other words, T90 is the time until the sensor resistance value changes from Ra to R.

[0064] When the difference between Ra and Rg is large and a graph is created by plotting time on the horizontal axis and sensor resistance on the vertical axis, with the vertical axis displayed on a logarithmic scale, the sensor resistance at the time of 90% response time may be represented as R', and the result may be expressed as logR' = {(logRa - logRg) x 0.1 + logRg} (Equation (2)). In this case, the time it takes for the sensor resistance value to change from Ra to R' is T90.

[0065] When the metal oxide semiconductor gas sensor 10 is exposed to a measurement gas containing VOC gases and then exposed to air containing no VOC gases, the sensor resistance changes again from Rg to Ra. It is preferable that the time required for the sensor resistance to change from Rg to Ra be short. Furthermore, if the sensor resistance in air does not stabilize and continues to increase or decrease during measurement, the sensor resistance when exposed to a measurement gas containing VOC gases will be higher or lower than it should be.

[0066] Therefore, when Ra>Rg, the metal oxide semiconductor gas sensor 10 of this embodiment preferably changes its sensor resistance value Ra to 80% to 100% of the sensor resistance value Ra before exposure to the measurement gas within 5 minutes after exposing the metal oxide semiconductor gas sensor 10 to air containing no VOC gases after being exposed to a measurement gas containing VOC gases.When Ra<Rg, the metal oxide semiconductor gas sensor 10 of this embodiment preferably changes its sensor resistance value Ra to 100% to 120% of the sensor resistance value Ra before exposure to the measurement gas within 5 minutes after exposing the metal oxide semiconductor gas sensor 10 to air containing no VOC gases after being exposed to a measurement gas containing VOC gases.

[0067] The selective detection sensitivity of acetone gas relative to ethanol gas, among VOC gases, can be expressed as, for example, Ace / EtOH. The selective detection sensitivity, Ace / EtOH, is preferably greater than 1, more preferably 1.2 or greater, 1.5 or greater, or even 2 or greater, and such a combination of p-type and n-type semiconductors and / or ΔCB value is preferred.

[0068] The operating temperature of the metal oxide semiconductor gas sensor 10 of this embodiment is not particularly limited. For example, it is preferably 150°C or higher and 400°C or lower. This is because heating to 150°C or higher can improve responsiveness to VOC gases, which are the gases to be measured. Furthermore, heating to temperatures higher than 400°C can cause grain growth of metal oxide particles, which can result in deterioration of the metal oxide semiconductor gas sensor 10. In order to suppress grain growth, prevent deterioration of the metal oxide semiconductor gas sensor 10, and ensure long-term use, it is preferable to set the operating temperature to 400°C or lower, as described above.

[0069] There are no particular limitations on the method for maintaining the metal oxide semiconductor gas sensor 10 at the above-mentioned operating temperature. For example, various heating methods can be selected, such as external heating using an electric furnace or resistance heating, in which a heater such as a Pt heater is formed on the back surface of the substrate and electricity is passed through the heater.

[0070] The metal oxide semiconductor gas sensor of this embodiment described above is small and inexpensive and can detect VOC gases contained in the measurement gas at concentrations of about 30 ppm or 20 ppm.

[0071] The type of gas to be measured is not particularly limited, and any gas containing VOC gases can be measured. In particular, VOC gases in combustion engine exhaust gases can be measured.

[0072] In particular, the metal oxide semiconductor gas sensor of this embodiment can detect VOC gases contained at a concentration of about 20 ppm with high sensitivity, and therefore can be suitably used in home medical care applications, for example.

[0073] Second Embodiment A metal oxide semiconductor gas sensor according to another embodiment of the present invention will be described below. This embodiment has many common features with the first embodiment, so some of the common features will be omitted and the following description will mainly focus on the differences. The metal oxide semiconductor gas sensor of this embodiment is a gas sensor that can mainly detect CO gas.

[0074] In this embodiment, as shown in FIG. 4 , the sensitive layer 4 also includes a first oxide 4a of an n-type semiconductor and a second oxide 4b of a p-type semiconductor. Note that, although the first oxide 4a and the second oxide 4b are depicted as connected in FIG. 4 , they are actually separate particles. Here, the band gap of the n-type semiconductor at 300 K is 2.4 eV to 5.0 eV, preferably 2.6 to 4.8 eV, and the band gap of the p-type semiconductor at 300 K is 1.0 eV to 5.0 eV, preferably 1.2 to 4.3 eV. By having the n-type and p-type semiconductors with band gaps within the specified ranges, a metal oxide semiconductor gas sensor with high sensitivity to CO gas, a short response time, and excellent gas selectivity can be obtained.

[0075] Examples of n-type semiconductor oxides having the above band gap include SnO 2 , W.O. 3 , V 2 O 5 , Sb 2 O 3 , TiO 2 (both rutile and anatase), Cr 2 O 3 , ZnO, Nb 2 O 5 , BaO, In 2 O 3 , Ta 2 O 5 , Ga 2 O 3 At least one of these oxides can be used alone or in combination. For CO gas detection, the n-type semiconductor oxide is particularly preferably SnO. 2 , WO3, ZnO, etc., and at least one of these oxides can be used alone or in combination.

[0076] Furthermore, examples of oxides of p-type semiconductors having the above band gap include CuO and MnO 2 , Co 3 O 4 , Cu 2 O, Mn 2 O 3 , U.O. 2 , CoO, PbO, Bi 2 O 3 , NiO, MnO, etc., and at least one of these oxides can be used alone or in combination. For CO gas detection, particularly preferred examples of p-type semiconductor oxides include CuO, CO3O4, NiO, etc., and at least one of these oxides can be used alone or in combination.

[0077] In the sensitive layer 4, the surface area ratio of the n-type semiconductor first oxide 4a in the sensitive layer 4 is preferably 47.9% to 99.5%, more preferably 50% to 99%. Also, the surface area ratio of the p-type semiconductor second oxide 4b in the sensitive layer 4 is preferably 0.5% to 52.1%, more preferably 1% to 50%. When the ratios are within these ranges, the response time can be shortened, and particularly the sensor sensitivity and gas selectivity can be improved.

[0078] The surface area ratio of the first oxide in the sensitive layer 4 is, in other words, the ratio of n-type semiconductors in the metal oxide material that constitutes the sensitive layer 4. The surface area ratio of the second oxide in the sensitive layer 4 is, in other words, the ratio of p-type semiconductors in the metal oxide material that constitutes the sensitive layer 4. Note that the sensitive layer 4 does not include the first electrode 2 or the second electrode 3.

[0079] There are no particular limitations on the method for measuring the surface area ratio of the n-type semiconductor and the surface area ratio of the p-type semiconductor. For example, the sensitive layer 4 can be cut out from the metal oxide semiconductor gas sensor 10, and a mapping image of each element (oxide) such as that shown in FIG. 4 can be created using STEM-EDS or the like, and the surface area ratio of each can be determined based on the mapping image. Specifically, the perimeters of the p-type and n-type semiconductors can be calculated from the mapping image, and the squares of these can be expressed as the surface area ratio. In this case, a visual field containing at least 50 or more p-type and n-type particles can be selected, and the total surface area ratio can be calculated by observing at least 10 visual fields.

[0080] The sensitive layer 4 may contain components other than the above-mentioned metal oxides as long as they do not significantly affect the sensor sensitivity, response time, and gas selectivity for CO gas. The total area ratio of the components other than the metal oxides in the sensitive layer 4 may be 5% or less.

[0081] The average pore size of the sensitive layer 4 is preferably 52 nm or more and 145 nm or less, more preferably 80 nm or more and 122 nm or less. There are no particular limitations on the method for measuring the average pore size. For example, the equivalent circle diameter of each pore in a cross section parallel to the thickness direction of the sensitive layer 4 may be calculated and averaged. The average pore size of the sensitive layer 4 can be defined as the average size of all pores observed in a cross section of the sensitive layer 4 within a specified area range. The lower limit of the size recognized as a pore is the same as that for the average porosity.

[0082] The average pore size is preferably determined based on the mean free path of CO (90 to 130 nm), and by setting the average pore size within the above range, it becomes easier to further increase the sensor sensitivity to CO gas and further shorten the response time.

[0083] The structure and manufacturing method of the metal oxide semiconductor gas sensor 10 according to this embodiment are the same as those of the previously described embodiment, and therefore a common description will be omitted.

[0084] The metal oxide semiconductor gas sensor 10 of this embodiment has a predetermined sensitive layer 4, which improves the sensitivity, responsiveness, and gas selectivity to CO gas. Specifically, the metal oxide semiconductor gas sensor 10 of this embodiment has high responsiveness and can detect CO gas even at low concentrations of 20 ppm or less contained in the measured gas. There is no particular lower limit for the detectable CO gas concentration, but it can generally detect CO gas at concentrations of 50 ppb or more.

[0085] When the semiconductor sensitive layer 4 is exposed to a measurement gas containing CO gas, the sensor resistance, i.e., the resistance between the first electrode 2 and the second electrode 3, changes. CO gas can be detected from the change in the sensor resistance.

[0086] The principle behind the change in sensor resistance is explained below. In the case of many metal oxide semiconductor gas sensors, the sensor resistance when exposed to air is used as the reference. When exposed to air, oxygen, which has electron-withdrawing properties, is adsorbed to the surface of the semiconductor (sensitive layer) of a metal oxide semiconductor gas sensor. When the semiconductor is solely an n-type semiconductor, oxygen adsorbed to the semiconductor surface forms a space charge layer near the semiconductor surface. The formation of the space charge layer creates a potential barrier between the semiconductors, preventing electrons from moving between the semiconductors.

[0087] If an oxidizing gas is introduced onto the semiconductor surface when oxygen is adsorbed onto the surface, the space charge layer becomes thicker. This causes the sensor resistance to increase. If a reducing gas is introduced onto the semiconductor surface when oxygen is adsorbed onto the surface, the adsorbed oxygen is consumed, causing the space charge layer to become thinner. This causes the sensor resistance to decrease. As a result, the concentration of the gas to be measured can be detected from the change in sensor resistance.

[0088] When the semiconductor is solely a p-type semiconductor, the opposite reaction occurs to that when the semiconductor is an n-type semiconductor. When an oxidizing gas is introduced onto the semiconductor surface, the sensor resistance value decreases. When a reducing gas is introduced onto the semiconductor surface, the sensor resistance value increases.

[0089] In this embodiment, it is considered that the potential barrier is neither between n-type semiconductors nor between p-type semiconductors. That is, in this embodiment, it is considered that a pn junction is formed at the interface between the n-type oxide semiconductor and the p-type oxide semiconductor inside the sensitive layer 4. The formation of the pn junction causes holes and electrons to combine, resulting in a higher barrier and a wider electron depletion layer at the junction interface than in the past, increasing resistance in air. Furthermore, in a reducing gas or flammable gas, some of the oxygen adsorbed on the oxide surface is consumed by reaction with the gas, returning the captured electrons to the semiconductor, lowering the height of the space charge barrier and reducing resistance. As a result, it is considered that the sensor sensitivity is improved.

[0090] The sensor resistance value when the metal oxide semiconductor gas sensor 10 is exposed to air containing no CO gas is Ra, and the sensor resistance value when the metal oxide semiconductor gas sensor 10 is exposed to a measurement gas containing a specific concentration of CO gas is Rg. The sensor sensitivity to the specific concentration of CO gas is S. CO as Rg / Ra or S' CO The sensitivity of the sensor varies depending on the concentration of CO gas.

[0091] When the metal oxide semiconductor gas sensor 10 using the sensitive layer 4 according to this embodiment is exposed to CO gas, Ra>Rg is often satisfied.

[0092] After switching the gas supplied to the metal oxide semiconductor gas sensor 10, it takes some time for the sensor resistance value to reach a value corresponding to the gas supplied. Therefore, when measuring Ra, the sensor resistance value is continuously measured after starting to supply air to the metal oxide semiconductor gas sensor 10, and it is confirmed that the sensor resistance value has stabilized. The sensor resistance value when the sensor resistance value has stabilized is Ra. When measuring Rg, the sensor resistance value is continuously measured after starting to supply the measurement gas to the metal oxide semiconductor gas sensor 10, and it is confirmed that the sensor resistance value has stabilized. The sensor resistance value when the sensor resistance value has stabilized is Rg.

[0093] The absolute value of the change in sensor resistance value |(Ra - Rg)| from the sensor resistance value Ra in air to the sensor resistance value Rg in the measured gas is set to 100, and the 90% response time (T90) is the time from when the gas supplied to the metal oxide semiconductor gas sensor 10 is switched from air to the measured gas until the absolute value of the change in sensor resistance value reaches 90. The sensor resistance value at the time of the 90% response time is R, which can be expressed as R = {(Ra - Rg) × 0.1 + Rg} (Equation (1)). In other words, T90 is the time until the sensor resistance value changes from Ra to R.

[0094] When the difference between Ra and Rg is large and a graph is created by plotting time on the horizontal axis and sensor resistance on the vertical axis, with the vertical axis displayed on a logarithmic scale, the sensor resistance at the time of 90% response time may be represented as R', and the result may be expressed as logR' = {(logRa - logRg) x 0.1 + logRg} (Equation (2)). In this case, the time it takes for the sensor resistance value to change from Ra to R' is T90.

[0095] When the metal oxide semiconductor gas sensor 10 is exposed to a measurement gas containing CO gas and then exposed to air containing no CO gas, the sensor resistance changes again from Rg to Ra. In this case, it is preferable that the time required for the sensor resistance to change from Rg to Ra is short. Furthermore, if the sensor resistance in air does not stabilize during measurement and continues to increase or decrease, the sensor resistance when exposed to a measurement gas containing CO gas will be higher or lower than it should be.

[0096] Therefore, when Ra>Rg, it is preferable that the metal oxide semiconductor gas sensor 10 of this embodiment, after being exposed to a measurement gas containing CO gas, changes in sensor resistance to 80% to 100% of the sensor resistance Ra before exposure to the measurement gas within 5 minutes after being exposed to air containing no CO gas. When Ra<Rg, it is preferable that the metal oxide semiconductor gas sensor 10 of this embodiment, after being exposed to a measurement gas containing CO gas, changes in sensor resistance to 100% to 120% of the sensor resistance Ra before exposure to the measurement gas within 5 minutes after being exposed to air containing no CO gas.

[0097] Similar to the sensor sensitivity to CO gas, the sensor sensitivity to CH4 gas is CH4 Or S' CH4 Then, S CO / S CH4 Or S' CO / S' CH4 The larger of these can be calculated as the CO gas selectivity of the semiconductor gas sensor.

[0098] The operating temperature of the metal oxide semiconductor gas sensor 10 of this embodiment is not particularly limited. For example, it is preferably 150°C or higher and 400°C or lower. This is because heating to 150°C or higher can improve responsiveness to CO gas, which is the gas to be measured. Furthermore, heating to a high temperature above 400°C can cause grain growth of metal oxide particles. As a result, the metal oxide semiconductor gas sensor 10 may deteriorate. In order to suppress grain growth, prevent deterioration of the metal oxide semiconductor gas sensor 10, and ensure long-term use, it is preferable to set the operating temperature to 400°C or lower, as described above.

[0099] There are no particular limitations on the method for maintaining the metal oxide semiconductor gas sensor 10 at the above-mentioned operating temperature. For example, various heating methods can be selected, such as external heating using an electric furnace or resistance heating, in which a heater such as a Pt heater is formed on the back surface of the substrate and electricity is passed through the heater.

[0100] The metal oxide semiconductor gas sensor of this embodiment described above is small and inexpensive and can detect CO gas contained in a measurement gas at a concentration of about 20 ppm.

[0101] There are no particular limitations on the type of gas to be measured, and any gas containing CO gas can be used as a measurement target. In particular, CO gas in combustion engine exhaust gas can be suitably measured.

[0102] In particular, the metal oxide semiconductor gas sensor of this embodiment can detect CO gas at a concentration of about 20 ppm with high sensitivity, and therefore can be suitably used in home medical care applications, for example.

[0103] The present invention is not limited to the above-described embodiment and can be modified in various ways. For example, the arrangement of the electrodes 2 and 3 of the sensor 10 is not particularly limited, and any structure may be used as long as both the first electrode 2 and the second electrode 3 are in contact with and face the sensitive layer 4. Furthermore, an embodiment may be a combination of the first and second embodiments described above.

[0104] The present invention will be described in detail below with reference to examples, but the present invention is not limited to the examples shown below.

[0105] (Sample No. 1) According to the following procedure, the metal oxide semiconductor gas sensor 10 shown in FIGS. 1 to 4 was fabricated and evaluated.

[0106] Substrate 1: 9.5 x 5.0 mm 2 A Si substrate molded into the shape shown in FIG.

[0107] Electrodes 2 and 3 having the comb-tooth shape shown in FIG. 1 were formed on a substrate 1. Specifically, the dimensions were 2.5×4.0 mm. 2 Pt electrodes in a comb shape with an electrode width of 15 μm and an inter-electrode distance of 15 μm were formed by sputtering.

[0108] A paste prepared by the method described below was applied to electrodes 2 and 3. An inkjet printer was used to apply the paste. The paste was then fired in a firing furnace at a firing temperature of 430°C in air to form a sensitive layer 4 on electrodes 2 and 3, as shown in Figure 2, and a metal oxide semiconductor gas sensor 10 was thus fabricated.

[0109] The method for preparing the paste will be described below.

[0110] Powders made of the materials listed in Table 1A for the n-type and p-type semiconductors were mixed so that the volume ratios of the n-type and p-type semiconductors were as listed in Table 1A to obtain a mixed powder. Separately, a binder was prepared by mixing polyvinyl butyral resin (BM-S manufactured by Sekisui Chemical Co., Ltd.), a dispersant, and BDG (diethylene glycol monobutyl ether). The resin was weighed out so that the mass ratio relative to the mixed powder was 5 php, and the mixture was mixed and stirred for 1 hour in a room-temperature grinder (mixer mill) manufactured by SPEX to prepare a paste. ΔCB was calculated from the characteristic values ​​of each material and the volume ratios listed above. The results are shown in Table 1A.

[0111] A paste for the sensitive layer 4 was applied to the substrate 1 on which the electrodes 2 and 3 were formed, and heat treatment (400°C to 600°C) was performed to form a single-layer sensitive layer 4 with a thickness of 1.2 μm, thereby obtaining a sample of a metal oxide semiconductor gas sensor.

[0112] (Sample Nos. 2 to 35) Metal oxide semiconductor gas sensors were fabricated in the same manner as sample No. 1, except that the types and volume ratios of the n-type semiconductor and / or p-type semiconductor were changed to the types and ratios shown in Table 1A.

[0113] Next, a method for evaluating the fabricated metal oxide semiconductor gas sensor will be described.

[0114] The fabricated metal oxide semiconductor gas sensor was placed in a sample chamber equipped with a heater, and the sample holder equipped with a heater was heated to 150 to 400°C.

[0115] Synthetic air was prepared by mixing nitrogen gas and oxygen gas at a flow rate ratio of 4:1. The synthetic air was flowed into the sample chamber at a flow rate of 500 mL / min, and the sensor resistance was measured at a temperature of 300°C.

[0116] The sensor resistance value was measured by the two-terminal method using a 2700-type multi-channel DMM manufactured by Keithley Instruments, Inc. The measurement interval for the sensor resistance value was 10 seconds. After confirming that the sensor resistance value had stabilized, the sensor resistance value was measured at 10-second intervals for 200 seconds, and the average of the obtained sensor resistance values ​​was taken as the sensor resistance value (Ra) in synthetic air.

[0117] After measuring Ra, a constant flow rate of EtOH gas was introduced from an EtOH standard gas cylinder into the air gas to form an ethanol-containing gas. The ethanol concentration in the ethanol-containing gas was set to 30 ppm or less. The ethanol-containing gas was then supplied to the metal oxide semiconductor gas sensor by flowing it into the sample chamber at a flow rate of 500 mL / min, and the change in sensor resistance due to the supply of the ethanol-containing gas was examined. Rg was the sensor resistance value 15 minutes after the start of the supply of the ethanol-containing gas. Ra / Rg and Rg / Ra were calculated as the sensor sensitivity. The results are shown in Table 1B.

[0118] Further, Ra / Rg and Rg / Ra were determined as sensor sensitivities in the same manner, except that the ethanol-containing gas was replaced with an acetone-containing gas. The results are shown in Table 1B. Further, Ra / Rg and Rg / Ra were determined as sensor sensitivities in the same manner, except that the ethanol-containing gas was replaced with a CO-containing gas. The results are shown in Table 1B. Then, the ratio of Ra / Rg when acetone was used as the detection gas to Ra / Rg when ethanol was used as the detection gas, i.e., the gas selectivity Ace / EtOH for acetone over ethanol, was determined. The results are shown in Table 1B.

[0119] The average porosity, porosity difference (φf-φn), and average pore size of the sensitive layer 4 were also measured. The average thickness d of the sensitive layer 4 was also determined from the cross-sectional photograph of the STEM image. Furthermore, a sample was prepared by cutting out only the sensitive layer from a metal oxide semiconductor gas sensor sample, and a mapping image of each element (oxide) was created as shown in Figure 4. Based on the mapping image, the respective surface area ratios were calculated. The specific surface area in the sensitive layer was calculated using the above formula, and the surface area ratios of the n-type and p-type semiconductors were determined. Specifically, the perimeters of the p-type and n-type particles were calculated from the mapping image, and their squares were used to determine the surface area ratio. A visual field containing at least 50 p-type and n-type particles was selected, and the total surface area was calculated from at least 10 visual fields. As a result, in each example, the average porosity was 15.6% or more and 22.0% or less, the porosity difference (φf-φn) was 0.4 or more, and the average pore size was 52 nm or more and 145 nm or less.

[0120] Specifically, the average void size and average porosity were measured according to the following procedure.

[0121] A backscattered electron image of a cross section parallel to the thickness direction of the sensitive layer 4 taken with a scanning electron microscope (SU5000, manufactured by Hitachi, Ltd.) was processed to calculate the average porosity in the cross section of the sensitive layer 4. Furthermore, the average void size in the cross section of the sensitive layer 4 was calculated. The size of the backscattered electron image was set to a size that allowed the entire sensitive layer 4 to be observed in the thickness direction. The photographed area was a location where the first and second electrodes were not included in the backscattered electron image.

[0122] Furthermore, the backscattered electron image was divided into three equal parts in the thickness direction of the sensitive layer 4. The average porosity calculated by image processing the backscattered electron image of the region closest to the first and second electrodes was defined as φn (%), and the average porosity calculated by image processing the backscattered electron image of the region farthest from the first and second electrodes was defined as φf (%), and φf-φn was calculated.

[0123] Furthermore, the element distributions of the n-type and p-type semiconductors of these samples were evaluated by the following procedure.

[0124] A cross section parallel to the thickness direction of the sensitive layer 4 was cut using an FIB, and an STEM image with a field of view of 200 nm x 200 nm or more was obtained. The obtained STEM image was divided into 1 nm x 1 nm square regions and subjected to EDS analysis, and data was obtained on the intensity of Sn derived from the n-type semiconductor and Cu derived from the p-type semiconductor in each section. The coefficient of variation of the element distribution was calculated based on the data obtained for all sections (40,000 points). As a result, the coefficient of variation of the element distribution of the n-type semiconductor was 0.9 or less, and the coefficient of variation of the element distribution of the p-type semiconductor was 0.9 or less.

[0125]

[0126]

[0127] The results shown in Tables 1A and 1B show that in each of the Examples in which the n-type semiconductor and the p-type semiconductor are configured so that ΔCB is 1.8 eV, preferably 2 or more, 3 or more, or 4 or more, the sensitivity to VOC gases such as ethanol and acetone was better than that of the Comparative Examples. In contrast, the sensitivity to ethanol gas and acetone was not good in any of the Comparative Examples.

[0128] Furthermore, from the results shown in Table 1B, the band gap n of the n-type semiconductor at 300K BG (See FIG. 5) is 2.4 eV or more and 5.0 eV or less, and the band gap n BG It was found that when the voltage (see FIG. 5) is 1.0 eV or more and 5.0 eV or less, a metal oxide semiconductor gas sensor having high sensitivity to VOC gases such as ethanol and acetone can be obtained.

[0129] Furthermore, the results shown in Table 1B indicate that gas sensors with high selectivity for acetone gas can be realized in examples where the band gap of the n-type semiconductor at 300 K is smaller than 3.4 eV and the band gap of the p-type semiconductor at 300 K is larger than 1.75 eV (sample numbers 13, 14, 21, and 22). Also, gas sensors with high selectivity for ethanol gas can be realized in examples where the band gap of the n-type semiconductor at 300 K is larger than 3.4 eV (sample numbers 25, 26, 32, and 33).

[0130] (Sample No. 53) According to the following procedure, the metal oxide semiconductor gas sensor 10 shown in FIGS. 1 to 4 was fabricated and evaluated.

[0131] Substrate 1: 9.5 x 5.0 mm 2 A Si substrate molded into the shape shown in FIG.

[0132] Electrodes 2 and 3 having the comb-tooth shape shown in FIG. 1 were formed on a substrate 1. Specifically, the dimensions were 2.5×4.0 mm. 2 Pt electrodes in a comb shape with an electrode width of 15 μm and an inter-electrode distance of 15 μm were formed by sputtering.

[0133] A paste prepared by the method described below was applied to electrodes 2 and 3. An inkjet printer was used to apply the paste. The paste was then fired in a firing furnace at a firing temperature of 430°C in air to form a sensitive layer 4 on electrodes 2 and 3, as shown in Figure 2, and a metal oxide semiconductor gas sensor 10 was thus fabricated.

[0134] The method for preparing the paste will be described below.

[0135] SnO as an n-type semiconductor 2 Powder (BET specific surface area 75m 2 / g) and CuO powder (BET specific surface area 38 m) as a p-type semiconductor. 2 / g) were mixed so that the volume ratio was 70 vol% n-type semiconductor and 30 vol% p-type semiconductor (surface area ratio: 83.3 vol% n-type semiconductor and 16.7 vol% p-type semiconductor), to obtain a mixed powder. Separately, polyvinyl butyral resin (BM-S manufactured by Sekisui Chemical Co., Ltd.), a dispersant, and BDG (diethylene glycol monobutyl ether) were mixed to prepare a binder. The resin was weighed out so that the mass ratio relative to the mixed powder was 5 php, and the mixture was mixed and stirred for 1 hour in a room temperature grinder (mixer mill) manufactured by SPEX Corporation to prepare a paste.

[0136] A paste for the sensitive layer 4 was applied to the substrate 1 on which the electrodes 2 and 3 were formed, and heat treatment (400°C to 600°C) was performed to form a single-layer sensitive layer 4 with a thickness of 1.2 μm, thereby obtaining a sample of the metal oxide semiconductor gas sensor according to sample number 53.

[0137] Next, a method for evaluating the fabricated metal oxide semiconductor gas sensor will be described.

[0138] The fabricated metal oxide semiconductor gas sensor was placed in a sample chamber equipped with a heater, and the sample holder equipped with a heater was heated to 150 to 400°C.

[0139] Synthetic air was prepared by mixing nitrogen gas and oxygen gas at a flow rate ratio of 4: 1. The synthetic air was flowed into the sample chamber at a flow rate of 500 mL / min, and the sensor resistance was measured.

[0140] The sensor resistance value was measured by the two-terminal method using a 2700-type multi-channel DMM manufactured by Keithley Instruments, Inc. The measurement interval for the sensor resistance value was 10 seconds. After confirming that the sensor resistance value had stabilized, the sensor resistance value was measured at 10-second intervals for 200 seconds, and the average of the obtained sensor resistance values ​​was taken as the sensor resistance value (Ra) in synthetic air.

[0141] After measuring Ra, a constant flow rate of CO gas was introduced from a CO standard gas cylinder into the air gas to form a CO-containing gas. The CO concentration in the CO-containing gas was set to 20 ppm or less. The CO-containing gas was then supplied to the metal oxide semiconductor gas sensor by flowing it into the sample chamber at a flow rate of 500 mL / min, and the change in sensor resistance due to the supply of CO-containing gas was investigated. Rg was the sensor resistance value 15 minutes after the start of supplying the CO-containing gas. Furthermore, S was used to measure the sensor sensitivity to CO gas. CO = Ra / Rg and S' CO The results are shown in Table 2. CO Or S' CO When the value was 2.00 or more, it was judged as good, and when it was 2.50 or more, it was judged as even better.

[0142] The response time T90 was determined by the above method. T90 was the time from when the supply of the CO-containing gas started until the sensor resistance value reached R' shown in the above formula (2). A T90 of 120 seconds or less was considered good, and a T90 of 100 seconds or less was considered even better.

[0143] Furthermore, the sensor sensitivities SCH4 and S' for CH4 gas were measured in the same manner except that the CO gas was changed to CH4 gas and the CH4 concentration in the CH4-containing gas was set to 300 ppm. CH4 Then, S CO / S CH4 The gas selectivity was calculated as follows. The results are shown in Table 2. CO / S CH4 When the value was 1.5 or more, it was judged as good, and when it was 1.7 or more, it was judged as even better.

[0144] For sample number 53, the average porosity, porosity difference (φf-φn), and average pore size of the sensitive layer 4 were measured. The results are shown in Table 3. The average thickness d of the sensitive layer 4 was also determined from the cross-sectional STEM image. Furthermore, a sample was prepared by cutting out only the sensitive layer from a metal oxide semiconductor gas sensor sample, and a mapping image of each element (oxide) was created as shown in Figure 4. Based on the mapping image, the surface area ratio of each element was calculated. The specific surface area in the sensitive layer was calculated using the above formula, and the surface area ratios of the n-type and p-type semiconductors were determined. Specifically, the perimeters of the p-type and n-type semiconductors were calculated from the mapping image, and the squares of these were used to determine the surface area ratio. A visual field containing at least 50 p-type and n-type particles was selected, and the total surface area ratio was calculated for at least 10 visual fields. The results are shown in Table 3.

[0145] Specifically, the average void size and average porosity were measured according to the following procedure.

[0146] A backscattered electron image of a cross section parallel to the thickness direction of the sensitive layer 4 taken with a scanning electron microscope (SU5000, manufactured by Hitachi, Ltd.) was processed to calculate the average porosity in the cross section of the sensitive layer 4. Furthermore, the average void size in the cross section of the sensitive layer 4 was calculated. The size of the backscattered electron image was set to a size that allowed the entire sensitive layer 4 to be observed in the thickness direction. The photographed area was a location where the first and second electrodes were not included in the backscattered electron image.

[0147] Furthermore, the backscattered electron image was divided into three equal parts in the thickness direction of the sensitive layer 4. The average porosity calculated by image processing the backscattered electron image of the region closest to the first and second electrodes was defined as φn (%), and the average porosity calculated by image processing the backscattered electron image of the region farthest from the first and second electrodes was defined as φf (%), and φf-φn was calculated.

[0148] (Sample Nos. 51, 52, 54, and 55) Metal oxide semiconductor gas sensors were fabricated in the same manner as in Example 1, except that the type of n-type semiconductor and / or p-type semiconductor was changed to the type shown in Table 2. The obtained metal oxide semiconductor gas sensors were evaluated in the same manner as in Sample No. 53. The results are shown in Table 2.

[0149]

[0150] The results shown in Table 2 show that the example (sample number 53) in which the n-type semiconductor and p-type semiconductor had predetermined band gaps had better CO sensor sensitivity, response time, and gas selectivity than the reference examples (sample numbers 51, 52, 54, and 55).

[0151] (Sample Nos. 56 to 61) Compared with sample No. 53, SnO as an n-type semiconductor 2 Metal oxide semiconductor gas sensors were fabricated in the same manner as sample number 53, except that the mixing ratio of Cu and CuO as a p-type semiconductor was varied to adjust the surface area ratios of each to the values ​​shown in Table 3. The obtained metal oxide semiconductor gas sensors were evaluated in the same manner as sample number 53. The results are shown in Table 3. In sample numbers 60 and 61, Rg was greater than Ra, so S' was used as the CO sensor sensitivity. CO = Rg / Ra as gas selectivity, and S' CO / S' CH4 The following is stated.

[0152]

[0153] The results shown in Table 3 show that even when the surface area ratio of the n-type semiconductor and the surface area ratio of the p-type semiconductor were changed, the sensor sensitivity, response time, and gas selectivity were good in each example in which the n-type semiconductor and the p-type semiconductor had predetermined bandgaps in the sensitive layer 4. It was also found that the response time and gas selectivity were further improved when the surface area ratio of the n-type semiconductor and the surface area ratio of the p-type semiconductor fell within the predetermined ranges.

[0154] Furthermore, from the results shown in Table 3, the surface area ratio of the n-type semiconductor is preferably 47.9% or more and 99.5% or less, and more preferably 50% or more and 99.5% or less, but from the viewpoint of improving the sensitivity and gas responsiveness of the CO sensor, it is more preferably 68% or more and 99.5% or less. Furthermore, the specific surface area is preferably 49.7 m 2 / g or more, more preferably 50.5m 2 / g or more, or 52.2m 2 / g or more.

[0155] (Sample Nos. 62 and 63) Metal oxide semiconductor gas sensors were fabricated in the same manner as sample No. 53, except that the type and amount of dispersant were adjusted to achieve the desired degree of dispersion compared to sample No. 58. The obtained metal oxide semiconductor gas sensors were evaluated in the same manner as sample No. 53. The results are shown in Table 4.

[0156] Furthermore, the element distributions of the n-type and p-type semiconductors of these samples were evaluated by the following procedure.

[0157] A cross section parallel to the thickness direction of the sensitive layer 4 was cut using an FIB, and an STEM image with a field of view of 200 nm x 200 nm or more was obtained. The obtained STEM image was divided into 1 nm x 1 nm square regions and subjected to EDS analysis, and data was obtained on the intensity of Sn derived from the n-type semiconductor and Cu derived from the p-type semiconductor in each region. The coefficient of variation of the element distribution was calculated based on the data obtained for all regions (40,000 points).

[0158]

[0159] The results shown in Table 4 show that when the coefficient of variation of the n-type semiconductor was 0.9 or less and / or when the coefficient of variation of the p-type semiconductor was 0.9 or less, the sensor sensitivity, response time, and gas selectivity were good, and the sensor sensitivity was particularly improved.

[0160] (Sample Nos. 64 to 66) Sample Nos. 64 to 66 were carried out in the same manner as sample No. 53, except that the binder resin content was changed from sample No. 53 so that the average porosity of the sensitive layer 4 became the value shown in Table 5. The binder resin content increased in the order of sample No. 64, sample No. 53, sample No. 65, and sample No. 66. The higher the resin content in the binder, the higher the average porosity. The results are shown in Table 5.

[0161]

[0162] The results shown in Table 5 show that even when the average porosity was different, each example in which the n-type semiconductor and p-type semiconductor in the sensitive layer 4 had a predetermined band gap exhibited good sensor sensitivity, response time, and gas selectivity. Furthermore, sample numbers 53 and 65, whose average porosity was within the predetermined range, exhibited good response times compared to sample number 64, whose average porosity was less than 15.8%, and sample number 66, whose average porosity was more than 21%.

[0163] (Sample Nos. 67 to 69) These samples were prepared under the same conditions as sample No. 53, except that the mixing and stirring time during paste preparation was changed. The mixing and stirring time was shorter for sample Nos. 67, 53, 68, and 69 in that order. The shorter the mixing and stirring time, the larger the average void size. The final average void size was adjusted to the value shown in Table 6. The results are shown in Table 6.

[0164]

[0165] Even when the average void size was changed, the sensor sensitivity, response time, and gas selectivity were good in each example in which the n-type semiconductor and p-type semiconductor in the sensitive layer 4 had a predetermined band gap. Samples 53 and 68, in which the average void size was 80 nm or more and 130 nm or less, had better sensor sensitivity and response time than sample 67, in which the average void size was less than 80 nm, and sample 69, in which the average void size was more than 130 nm.

[0166] (Sample Nos. 70-71) Two types of paste with different binder content ratios were prepared. The paste with the lower binder content ratio was applied first, and the paste with the higher binder content ratio was applied second. The final φf-φn value was set to the value shown in Table 7. All other points were the same as in Example 1. The results are shown in Table 7. Note that sample No. 70 had a smaller difference in the amount of binder resin between the first and second layer pastes.

[0167]

[0168] Even when φf-φn was changed, the sensor sensitivity, response time, and gas selectivity were good in each example in which the n-type and p-type semiconductors in the sensitive layer 4 had predetermined band gaps.

[0169] (Sample No. 72) The same conditions as in Example 1 were used, except that the heat treatment temperature was increased when forming the sensitive layer 4. The final value of φf-φn was set to the value shown in Table 7. The results are shown in Table 7.

[0170] Even when φf-φn was changed, the sensor sensitivity, response time, and gas selectivity were good in each example in which the n-type and p-type semiconductors in the sensitive layer 4 had predetermined band gaps. Furthermore, it was found that sample numbers 70, 71, and 53, in which φf-φn was 1.1% or more, had a superior response time compared to sample number 72.

[0171] (Sample Nos. 73 to 75) The test was carried out under the same conditions as sample No. 53, except that the thickness of the sensitive layer 4 was changed by changing the thickness of the applied paste as shown in Table 8. The results are shown in Table 8.

[0172]

[0173] Even when the thickness of the sensitive layer 4 was changed, the sensor sensitivity, response time, and gas selectivity were good in each example in which the n-type and p-type semiconductors in the sensitive layer 4 had predetermined bandgaps. Note that sample numbers 53, 74, and 75, in which the thickness of the sensitive layer 4 was 1.0 μm or more, had better sensor sensitivity and response time than sample number 73, in which the thickness of the sensitive layer was less than 1.0 μm.

Claims

1. A semiconductor device comprising a first electrode, a second electrode, and a sensitive layer in contact with both the first electrode and the second electrode, wherein the sensitive layer includes an n-type semiconductor and a p-type semiconductor, and the lowest energy level of the conduction band of the p-type semiconductor at 300 K is p CB The lowest energy level of the conduction band of the n-type semiconductor at 300 K is n CB When the volume ratio of the p-type semiconductor to the total volume of the p-type semiconductor and the n-type semiconductor is α, ΔCB=(p CB × α)-(n CB ×(1-α)) is 1.8 eV or more.

2. The metal oxide semiconductor gas sensor according to claim 1, wherein the band gap of said n-type semiconductor at 300K is smaller than 3.4 eV, and the band gap of said p-type semiconductor at 300K is larger than 1.75 eV.

3. The metal oxide semiconductor gas sensor according to claim 1, wherein the band gap of said n-type semiconductor at 300K is 2.4 eV or more and 5.0 eV or less, and the band gap of said p-type semiconductor at 300K is 1.0 eV or more and 5.0 eV or less.

4. A metal oxide semiconductor gas sensor having a first electrode, a second electrode, and a sensitive layer in contact with both the first electrode and the second electrode, wherein the sensitive layer includes an n-type semiconductor and a p-type semiconductor, wherein the band gap of the n-type semiconductor at 300K is 2.4 eV or more and 5.0 eV or less, and the band gap of the p-type semiconductor at 300K is 1.0 eV or more and 5.0 eV or less.

5. The metal oxide semiconductor gas sensor according to claim 1 or 4, wherein the surface area ratio of the n-type semiconductor in the sensitive layer is 47.9% or more and 99.5% or less.

6. The metal oxide semiconductor gas sensor according to claim 1 or 4, wherein the coefficient of variation of the element distribution of the n-type semiconductor in a cross section of the sensitive layer is 0.9 or less.

7. The metal oxide semiconductor gas sensor according to claim 1 or 4, wherein the coefficient of variation of the element distribution of the p-type semiconductor in the cross section of the sensitive layer is 0.9 or less.

8. The metal oxide semiconductor gas sensor according to claim 1 or 4, wherein the average porosity in the cross section of the sensitive layer is 15.6% or more and 22.0% or less.

9. The metal oxide semiconductor gas sensor according to claim 1 or 4, wherein the average void size in the cross section of the sensitive layer is 52 nm or more and 145 nm or less.

10. A metal oxide semiconductor gas sensor as described in claim 1 or 4, wherein when the sensitive layer is divided into three equal parts in the thickness direction, the average porosity in the region closest to the first electrode and the second electrode is φn (%), and the average porosity in the region farthest from the first electrode and the second electrode is φf (%), and φf-φn≧0.

4.

11. The metal oxide semiconductor gas sensor according to claim 1 or 4, wherein the thickness of the sensitive layer is 0.4 μm or more.

Images