The development of high performance, portable and miniaturized gas sensors is gaining increasing attention in the fields of environmental monitoring, security, medical diagnostics and agriculture. Among various detection tools, metal-oxide-semiconductor (MOS) chemo-resistive gas sensors are the most popular choice for commercial applications due to their high stability, low cost, and high sensitivity. One of the most important approaches to further improve the performance of the sensor is the creation of nanosized MOS-based heterojunctions (hetero-nanostructured MOS) from MOS nanomaterials. However, the sensing mechanism of a heteronanostructured MOS sensor is different from that of a single MOS gas sensor, as it is quite complex. Sensor performance is affected by various parameters, including the physical and chemical properties of the sensitive material (such as grain size, defect density, and material oxygen vacancies), operating temperature, and device structure. This review presents several concepts for designing high performance gas sensors by analyzing the sensing mechanism of heterogeneous nanostructured MOS sensors. In addition, the influence of the geometric structure of the device, determined by the relationship between the sensitive material and the working electrode, is discussed. To study sensor behavior systematically, this article introduces and discusses the general mechanism of perception of three typical geometric structures of devices based on various heteronanostructured materials. This overview will serve as a guide for future readers who study the sensitive mechanisms of gas sensors and develop high performance gas sensors.
Air pollution is an increasingly serious problem and a serious global environmental problem that threatens the well-being of people and living beings. Inhalation of gaseous pollutants can cause many health problems such as respiratory disease, lung cancer, leukemia and even premature death1,2,3,4. From 2012 to 2016, millions of people were reported to have died from air pollution, and each year, billions of people were exposed to poor air quality5. Therefore, it is important to develop portable and miniaturized gas sensors that can provide real-time feedback and high detection performance (eg, sensitivity, selectivity, stability, and response and recovery times). In addition to environmental monitoring, gas sensors play a vital role in safety6,7,8, medical diagnostics9,10, aquaculture11 and other fields12.
To date, several portable gas sensors based on different sensing mechanisms have been introduced, such as optical13,14,15,16,17,18, electrochemical19,20,21,22 and chemical resistive sensors23,24. Among them, metal-oxide-semiconductor (MOS) chemical resistive sensors are the most popular in commercial applications due to their high stability and low cost25,26. The contaminant concentration can be determined simply by detecting the change in MOS resistance. In the early 1960s, the first chemo-resistive gas sensors based on ZnO thin films were reported, generating great interest in the field of gas detection27,28. Today, many different MOS are used as gas sensitive materials, and they can be divided into two categories based on their physical properties: n-type MOS with electrons as the majority charge carriers and p-type MOS with holes as the majority charge carriers. charge carriers. In general, the p-type MOS is less popular than the n-type MOS because the inductive response of the p-type MOS (Sp) is proportional to the square root of the n-type MOS (\(S_p = \sqrt {S_n}\ ) ) at the same assumptions (for example, the same morphological structure and the same change in the bending of the bands in the air) 29,30. However, single-base MOS sensors still face problems such as insufficient detection limit, low sensitivity and selectivity in practical applications. Selectivity issues can be addressed to some extent by creating arrays of sensors (called “electronic noses”) and incorporating computational analysis algorithms such as training vector quantization (LVQ), principal component analysis (PCA), and partial least squares (PLS) analysis31 , 32, 33, 34, 35. In addition, the production of low-dimensional MOS32,36,37,38,39 (e.g. one-dimensional (1D), 0D and 2D nanomaterials), as well as the use of other nanomaterials (e.g. MOS40,41,42 , noble metal nanoparticles (NPs))43,44, carbon nanomaterials45,46 and conductive polymers47,48) to create nanoscale heterojunctions (i.e., heteronanostructured MOS) are other preferred approaches to solve the above problems. Compared with traditional thick MOS films, low-dimensional MOS with high specific surface area can provide more active sites for gas adsorption and facilitate gas diffusion36,37,49. In addition, the design of MOS-based heteronanostructures can further tune carrier transport at the heterointerface, resulting in large changes in resistance due to different operating functions50,51,52. In addition, some of the chemical effects (e.g., catalytic activity and synergistic surface reactions) that occur in the design of MOS heteronanostructures can also improve sensor performance.50,53,54 Although designing and fabricating MOS heteronanostructures would be a promising approach to improve sensor performance, modern chemo-resistive sensors typically use trial and error, which is time-consuming and inefficient. Therefore, it is important to understand the sensing mechanism of MOS based gas sensors as it can guide the design of high performance directional sensors.
In recent years, MOS gas sensors have developed rapidly and some reports have been published on MOS nanostructures55,56,57, room temperature gas sensors58,59, special MOS sensor materials60,61,62 and specialty gas sensors63. A review paper in Other Reviews focuses on elucidating the sensing mechanism of gas sensors based on the intrinsic physical and chemical properties of MOS, including the role of oxygen vacancies 64 , the role of heteronanostructures 55, 65 and charge transfer at heterointerfaces 66. In addition, many other parameters affect sensor performance, including heterostructure, grain size, operating temperature, defect density, oxygen vacancies, and even open crystal planes of the sensitive material25,67,68,69,70,71. 72, 73. However, the (rarely mentioned) geometric structure of the device, determined by the relationship between the sensing material and the working electrode, also significantly affects the sensitivity of the sensor74,75,76 (see section 3 for more details). For example, Kumar et al. 77 reported two gas sensors based on the same material (eg, two-layer gas sensors based on TiO2@NiO and NiO@TiO2) and observed different changes in NH3 gas resistance due to different device geometries. Therefore, when analyzing a gas-sensing mechanism, it is important to take into account the structure of the device. In this review, the authors focus on MOS-based detection mechanisms for various heterogeneous nanostructures and device structures. We believe that this review can serve as a guide for readers wishing to understand and analyze gas detection mechanisms and can contribute to the development of future high performance gas sensors.
On fig. 1a shows the basic model of a gas sensing mechanism based on a single MOS. As the temperature rises, the adsorption of oxygen (O2) molecules on the MOS surface will attract electrons from the MOS and form anionic species (such as O2- and O-). Then, an electron depletion layer (EDL) for an n-type MOS or a hole accumulation layer (HAL) for a p-type MOS is then formed on the surface of the MOS 15, 23, 78. The interaction between O2 and the MOS causes the conduction band of the surface MOS to bend upward and form a potential barrier. Subsequently, when the sensor is exposed to the target gas, the gas adsorbed on the surface of the MOS reacts with ionic oxygen species, either attracting electrons (oxidizing gas) or donating electrons (reducing gas). Electron transfer between the target gas and the MOS can adjust the width of the EDL or HAL30,81 resulting in a change in the overall resistance of the MOS sensor. For example, for a reducing gas, electrons will be transferred from the reducing gas to an n-type MOS, resulting in a lower EDL and lower resistance, which is referred to as n-type sensor behavior. In contrast, when a p-type MOS is exposed to a reducing gas that determines the p-type sensitivity behavior, the HAL shrinks and the resistance increases due to electron donation. For oxidizing gases, the sensor response is opposite to that for reducing gases.
Basic detection mechanisms for n-type and p-type MOS for reducing and oxidizing gases b Key factors and physico-chemical or material properties involved in semiconductor gas sensors 89
Apart from the basic detection mechanism, the gas detection mechanisms used in practical gas sensors are quite complex. For example, the actual use of a gas sensor must meet many requirements (such as sensitivity, selectivity, and stability) depending on the needs of the user. These requirements are closely related to the physical and chemical properties of the sensitive material. For example, Xu et al.71 demonstrated that SnO2 based sensors achieve the highest sensitivity when the crystal diameter (d) is equal to or less than twice the Debye length (λD) of SnO271. When d ≤ 2λD, SnO2 is completely depleted after the adsorption of O2 molecules, and the response of the sensor to the reducing gas is maximum. In addition, various other parameters can affect sensor performance, including operating temperature, crystal defects, and even exposed crystal planes of the sensing material. In particular, the influence of the operating temperature is explained by the possible competition between the rates of adsorption and desorption of the target gas, as well as the surface reactivity between adsorbed gas molecules and oxygen particles4,82. The effect of crystal defects is strongly related to the content of oxygen vacancies [83, 84]. The operation of the sensor can also be affected by different reactivity of open crystal faces67,85,86,87. Open crystal planes with lower density reveal more uncoordinated metal cations with higher energies, which promote surface adsorption and reactivity88. Table 1 lists several key factors and their associated improved perceptual mechanisms. Therefore, by adjusting these material parameters, detection performance can be improved, and it is critical to determine the key factors affecting sensor performance.
Yamazoe89 and Shimanoe et al.68,71 performed a number of studies on the theoretical mechanism of sensor perception and proposed three independent key factors influencing sensor performance, specifically receptor function, transducer function, and utility (Fig. 1b). . Receptor function refers to the ability of the MOS surface to interact with gas molecules. This function is closely related to the chemical properties of MOS and can be significantly improved by introducing foreign acceptors (for example, metal NPs and other MOS). The transducer function refers to the ability to convert the reaction between the gas and the MOS surface into an electrical signal dominated by the grain boundaries of the MOS. Thus, sensory function is significantly affected by MOC particle size and density of foreign receptors. Katoch et al.90 reported that grain size reduction of ZnO-SnO2 nanofibrils resulted in the formation of numerous heterojunctions and increased sensor sensitivity, consistent with transducer functionality. Wang et al.91 compared various grain sizes of Zn2GeO4 and demonstrated a 6.5-fold increase in sensor sensitivity after introducing grain boundaries. Utility is another key sensor performance factor that describes the availability of gas to the internal MOS structure. If gas molecules cannot penetrate and react with the internal MOS, the sensor’s sensitivity will be reduced. The usefulness is closely related to the diffusion depth of a particular gas, which depends on the pore size of the sensing material. Sakai et al. 92 modeled the sensitivity of the sensor to flue gases and found that both the molecular weight of the gas and the pore radius of the sensor membrane affect the sensitivity of the sensor at different gas diffusion depths in the sensor membrane. The discussion above shows that high performance gas sensors can be developed by balancing and optimizing receptor function, transducer function, and utility.
The above work clarifies the basic perception mechanism of a single MOS and discusses several factors that affect the performance of a MOS. In addition to these factors, gas sensors based on heterostructures can further improve sensor performance by significantly improving sensor and receptor functions. In addition, heteronanostructures can further improve sensor performance by enhancing catalytic reactions, regulating charge transfer, and creating more adsorption sites. To date, many gas sensors based on MOS heteronanostructures have been studied to discuss mechanisms for enhanced sensing95,96,97. Miller et al. 55 summarized several mechanisms that are likely to improve the sensitivity of heteronanostructures, including surface-dependent, interface-dependent, and structure-dependent. Among them, the interface-dependent amplification mechanism is too complicated to cover all interface interactions in one theory, since various sensors based on heteronanostructured materials (for example, nn-heterojunction, pn-heterojunction, pp-heterojunction, etc.) can be used. Schottky knot). Typically, MOS-based heteronanostructured sensors always include two or more advanced sensor mechanisms98,99,100. The synergistic effect of these amplification mechanisms can enhance the reception and processing of sensor signals. Thus, understanding the mechanism of perception of sensors based on heterogeneous nanostructured materials is crucial in order to help researchers develop bottom-up gas sensors in accordance with their needs. In addition, the geometric structure of the device can also significantly affect the sensitivity of the sensor 74, 75, 76. In order to systematically analyze the behavior of the sensor, the sensing mechanisms of three device structures based on different heteronanostructured materials will be presented and discussed below.
With the rapid development of MOS based gas sensors, various hetero-nanostructured MOS have been proposed. The charge transfer at the heterointerface depends on the different Fermi levels (Ef) of the components. At the heterointerface, electrons move from one side with a larger Ef to the other side with a smaller Ef until their Fermi levels reach equilibrium, and holes, vice versa. Then the carriers at the heterointerface are depleted and form a depleted layer. Once the sensor is exposed to the target gas, the heteronanostructured MOS carrier concentration changes, as does the barrier height, thereby enhancing the detection signal. In addition, different methods of fabricating heteronanostructures lead to different relationships between materials and electrodes, which leads to different device geometries and different sensing mechanisms. In this review, we propose three geometric device structures and discuss the sensing mechanism for each structure.
Although heterojunctions play a very important role in gas detection performance, the device geometry of the entire sensor can also significantly influence the detection behavior, since the location of the sensor conduction channel is highly dependent on the device geometry. Three typical geometries of heterojunction MOS devices are discussed here, as shown in Figure 2. In the first type, two MOS connections are randomly distributed between two electrodes, and the location of the conductive channel is determined by the main MOS, the second is the formation of heterogeneous nanostructures from different MOS, while only one MOS is connected to the electrode. electrode is connected, then the conductive channel is usually located inside the MOS and is directly connected to the electrode. In the third type, two materials are attached to two electrodes separately, guiding the device through a heterojunction formed between the two materials.
A hyphen between compounds (eg “SnO2-NiO”) indicates that the two components are simply mixed (type I). An “@” sign between two connections (e.g. “SnO2@NiO”) indicates that the scaffold material (NiO) is decorated with SnO2 for a type II sensor structure. A slash (e.g. “NiO/SnO2”) indicates a type III sensor design .
For gas sensors based on MOS composites, two MOS elements are randomly distributed between the electrodes. Numerous fabrication methods have been developed to prepare MOS composites, including sol-gel, coprecipitation, hydrothermal, electrospinning, and mechanical mixing methods98,102,103,104. Recently, metal-organic frameworks (MOFs), a class of porous crystalline structured materials composed of metal centers and organic linkers, have been used as templates for the fabrication of porous MOS composites105,106,107,108. It is worth noting that although the percentage of MOS composites is the same, the sensitivity characteristics can vary greatly when using different manufacturing processes.109,110 For example, Gao et al.109 fabricated two sensors based on MoO3±SnO2 composites with the same atomic ratio ( Mo:Sn = 1:1.9) and found that different fabrication methods lead to different sensitivities. Shaposhnik et al. 110 reported that the reaction of co-precipitated SnO2-TiO2 to gaseous H2 differed from that of mechanically mixed materials, even at the same Sn/Ti ratio. This difference arises because the relationship between MOP and MOP crystallite size varies with different synthesis methods109,110. When the grain size and shape are consistent in terms of donor density and semiconductor type, the response should remain the same if the contact geometry does not change 110 . Staerz et al. 111 reported that the detection characteristics of SnO2-Cr2O3 core-sheath (CSN) nanofibers and ground SnO2-Cr2O3 CSNs were nearly identical, suggesting that the nanofiber morphology does not offer any advantage.
In addition to the different fabrication methods, the semiconductor types of the two different MOSFETs also affect the sensitivity of the sensor. It can be further divided into two categories depending on whether the two MOSFETs are of the same type of semiconductor (nn or pp junction) or different types (pn junction). When gas sensors are based on MOS composites of the same type, by changing the molar ratio of the two MOS, the sensitivity response characteristic remains unchanged, and the sensor sensitivity varies depending on the number of nn- or pp-heterojunctions. When one component predominates in the composite (eg 0.9 ZnO-0.1 SnO2 or 0.1 ZnO-0.9 SnO2), the conduction channel is determined by the dominant MOS, called the homojunction conduction channel 92 . When the ratios of the two components are comparable, it is assumed that the conduction channel is dominated by the heterojunction98,102. Yamazoe et al. 112,113 reported that the heterocontact region of the two components can greatly improve the sensitivity of the sensor because the heterojunction barrier formed due to the different operating functions of the components can effectively control the drift mobility of the sensor exposed to electrons. Various ambient gases 112,113. On fig. Figure 3a shows that sensors based on SnO2-ZnO fibrous hierarchical structures with different ZnO contents (from 0 to 10 mol % Zn) can selectively detect ethanol. Among them, a sensor based on SnO2-ZnO fibers (7 mol.% Zn) showed the highest sensitivity due to the formation of a large number of heterojunctions and an increase in the specific surface area, which increased the function of the converter and improved sensitivity 90 However, with a further increase in the ZnO content to 10 mol.%, the microstructure SnO2-ZnO composite can wrap surface activation areas and reduce sensor sensitivity85. A similar trend is also observed for sensors based on NiO-NiFe2O4 pp heterojunction composites with different Fe/Ni ratios (Fig. 3b)114.
SEM images of SnO2-ZnO fibers (7 mol.% Zn) and sensor response to various gases with a concentration of 100 ppm at 260 °C; 54b Responses of sensors based on pure NiO and NiO-NiFe2O4 composites at 50 ppm of various gases, 260 °C; 114 ( c) Schematic diagram of the number of nodes in the xSnO2-(1-x)Co3O4 composition and the corresponding resistance and sensitivity reactions of the xSnO2-(1-x)Co3O4 composition per 10 ppm CO, acetone, C6H6 and SO2 gas at 350 °C by changing the molar ratio of Sn/Co 98
The pn-MOS composites show different sensitivity behavior depending on the atomic ratio of MOS115. In general, the sensory behavior of MOS composites is highly dependent on which MOS acts as the primary conduction channel for the sensor. Therefore, it is very important to characterize the percentage composition and nanostructure of composites. Kim et al.98 confirmed this conclusion by synthesizing a series of xSnO2 ± (1-x)Co3O4 composite nanofibers by electrospinning and studying their sensor properties. They observed that the behavior of the SnO2-Co3O4 composite sensor switched from n-type to p-type by reducing the percentage of SnO2 (Fig. 3c)98. In addition, heterojunction-dominated sensors (based on 0.5 SnO2-0.5 Co3O4) showed the highest transmission rates for C6H6 compared to homojunction-dominant sensors (eg, high SnO2 or Co3O4 sensors). The inherent high resistance of the 0.5 SnO2-0.5 Co3O4 based sensor and its greater ability to modulate the overall sensor resistance contribute to its highest sensitivity to C6H6. In addition, lattice mismatch defects originating from SnO2-Co3O4 heterointerfaces can create preferential adsorption sites for gas molecules, thereby enhancing sensor response109,116.
In addition to semiconductor-type MOS, the touch behavior of MOS composites can also be customized using the chemistry of MOS-117. Huo et al.117 used a simple soak-bake method to prepare Co3O4-SnO2 composites and found that at a Co/Sn molar ratio of 10%, the sensor exhibited a p-type detection response to H2 and an n-type sensitivity to H2. response. Sensor responses to CO, H2S and NH3 gases are shown in Figure 4a117. At low Co/Sn ratios, many homojunctions form at the SnO2±SnO2 nanograin boundaries and exhibit n-type sensor responses to H2 (Figs. 4b,c)115. With an increase in the Co/Sn ratio up to 10 mol. %, instead of SnO2-SnO2 homojunctions, many Co3O4-SnO2 heterojunctions were simultaneously formed (Fig. 4d). Since Co3O4 is inactive with respect to H2, and SnO2 reacts strongly with H2, the reaction of H2 with ionic oxygen species mainly occurs on the surface of SnO2117. Therefore, electrons move to SnO2 and Ef SnO2 shifts to the conduction band, while Ef Co3O4 remains unchanged. As a result, the resistance of the sensor increases, indicating that materials with a high Co/Sn ratio exhibit p-type sensing behavior (Fig. 4e). In contrast, CO, H2S, and NH3 gases react with ionic oxygen species on the SnO2 and Co3O4 surfaces, and electrons move from the gas to the sensor, resulting in a decrease in barrier height and n-type sensitivity (Fig. 4f). . This different sensor behavior is due to the different reactivity of Co3O4 with different gases, which was further confirmed by Yin et al. 118 . Similarly, Katoch et al. 119 demonstrated that SnO2-ZnO composites have good selectivity and high sensitivity to H2. This behavior occurs because H atoms can be easily adsorbed to the O positions of ZnO due to strong hybridization between the s-orbital of H and the p-orbital of O, which leads to metallization of ZnO120,121.
a Co/Sn-10% dynamic resistance curves for typical reducing gases such as H2, CO, NH3 and H2S, b, c Co3O4/SnO2 composite sensing mechanism diagram for H2 at low % m. Co/Sn, df Co3O4 Mechanism detection of H2 and CO, H2S and NH3 with a high Co/Sn/SnO2 composite
Therefore, we can improve the sensitivity of the I-type sensor by choosing appropriate fabrication methods, reducing the grain size of the composites, and optimizing the molar ratio of the MOS composites. In addition, a deep understanding of the chemistry of the sensitive material can further enhance the selectivity of the sensor.
Type II sensor structures are another popular sensor structure that can use a variety of heterogeneous nanostructured materials, including one “master” nanomaterial and a second or even third nanomaterial. For example, one-dimensional or two-dimensional materials decorated with nanoparticles, core-shell (CS) and multilayer heteronanostructured materials are commonly used in type II sensor structures and will be discussed in detail below.
For the first heteronanostructure material (decorated heteronanostructure), as shown in Fig. 2b(1), the conductive channels of the sensor are connected by a base material. Due to the formation of heterojunctions, modified nanoparticles can provide more reactive sites for gas adsorption or desorption, and can also act as catalysts to improve sensing performance109,122,123,124. Yuan et al.41 noted that decorating WO3 nanowires with CeO2 nanodots can provide more adsorption sites at the CeO2@WO3 heterointerface and the CeO2 surface and generate more chemisorbed oxygen species for reaction with acetone. Gunawan et al. 125. An ultra-high sensitivity acetone sensor based on one-dimensional Au@α-Fe2O3 has been proposed and it has been observed that the sensitivity of the sensor is controlled by the activation of O2 molecules as an oxygen source. The presence of Au NPs can act as a catalyst promoting the dissociation of oxygen molecules into lattice oxygen for the oxidation of acetone. Similar results were obtained by Choi et al. 9 where a Pt catalyst was used to dissociate adsorbed oxygen molecules into ionized oxygen species and enhance the sensitive response to acetone. In 2017, the same research team demonstrated that bimetallic nanoparticles are much more efficient in catalysis than single noble metal nanoparticles, as shown in Figure 5126. 5a is a schematic of the manufacturing process for platinum-based bimetallic (PtM) NPs using apoferritin cells with an average size of less than 3 nm. Then, using the electrospinning method, PtM@WO3 nanofibers were obtained to increase the sensitivity and selectivity to acetone or H2S (Fig. 5b–g). Recently, single atom catalysts (SACs) have shown excellent catalytic performance in the field of catalysis and gas analysis due to the maximum efficiency of the use of atoms and tuned electronic structures127,128. Shin et al. 129 used Pt-SA anchored carbon nitride (MCN), SnCl2 and PVP nanosheets as chemical sources to prepare Pt@MCN@SnO2 inline fibers for gas detection. Despite the very low content of Pt@MCN (from 0.13 wt.% to 0.68 wt.%), the detection performance of gaseous formaldehyde Pt@MCN@SnO2 is superior to other reference samples (pure SnO2, MCN@SnO2 and Pt NPs@SnO2 ). . This excellent detection performance can be attributed to the maximum atomic efficiency of the Pt SA catalyst and the minimum coverage of SnO2129 active sites.
Apoferritin-loaded encapsulation method to obtain PtM-apo (PtPd, PtRh, PtNi) nanoparticles; dynamic gas sensitive properties of bd pristine WO3, PtPd@WO3, PtRn@WO3, and Pt-NiO@WO3 nanofibers; based, for example, on the selectivity properties of PtPd@WO3, PtRn@WO3 and Pt-NiO@WO3 nanofiber sensors to 1 ppm of interfering gas 126
In addition, heterojunctions formed between scaffold materials and nanoparticles can also effectively modulate conduction channels through a radial modulation mechanism to improve sensor performance130,131,132. On fig. Figure 6a shows the sensor characteristics of pure SnO2 and Cr2O3@SnO2 nanowires for reducing and oxidizing gases and the corresponding sensor mechanisms131. Compared to pure SnO2 nanowires, the response of Cr2O3@SnO2 nanowires to reducing gases is greatly enhanced, while the response to oxidizing gases is worsened. These phenomena are closely related to the local deceleration of the conduction channels of the SnO2 nanowires in the radial direction of the formed pn heterojunction. The sensor resistance can be simply tuned by changing the EDL width on the surface of pure SnO2 nanowires after exposure to reducing and oxidizing gases. However, for Cr2O3@SnO2 nanowires, the initial DEL of SnO2 nanowires in air is increased compared to pure SnO2 nanowires, and the conduction channel is suppressed due to the formation of a heterojunction. Therefore, when the sensor is exposed to a reducing gas, the trapped electrons are released into the SnO2 nanowires and the EDL is drastically reduced, resulting in higher sensitivity than pure SnO2 nanowires. Conversely, when switching to an oxidizing gas, DEL expansion is limited, resulting in low sensitivity. Similar sensory response results were observed by Choi et al., 133 in which SnO2 nanowires decorated with p-type WO3 nanoparticles showed significantly improved sensory response to reducing gases, while n-decorated SnO2 sensors had improved sensitivity to oxidizing gases. TiO2 nanoparticles (Fig. 6b) 133. This result is mainly due to the different work functions of SnO2 and MOS (TiO2 or WO3) nanoparticles. In p-type (n-type) nanoparticles, the conduction channel of the framework material (SnO2) expands (or contracts) in the radial direction, and then, under the action of reduction (or oxidation), further expansion (or shortening) of the conduction channel of SnO2 – rib ) of the gas ( Fig. 6b).
Radial modulation mechanism induced by modified LF MOS. a Summary of gas responses to 10 ppm reducing and oxidizing gases based on pure SnO2 and Cr2O3@SnO2 nanowires and corresponding sensing mechanism schematic diagrams; and corresponding schemes of WO3@SnO2 nanorods and detection mechanism133
In bilayer and multilayer heterostructure devices, the conduction channel of the device is dominated by the layer (usually the bottom layer) in direct contact with the electrodes, and the heterojunction formed at the interface of the two layers can control the conductivity of the bottom layer. Therefore, when gases interact with the top layer, they can significantly affect the conduction channels of the bottom layer and the resistance 134 of the device. For example, Kumar et al. 77 reported the opposite behavior of TiO2@NiO and NiO@TiO2 double layers for NH3. This difference arises because the conduction channels of the two sensors dominate in layers of different materials (NiO and TiO2, respectively), and then the variations in the underlying conduction channels are different77.
Bilayer or multilayer heteronanostructures are commonly produced by sputtering, atomic layer deposition (ALD) and centrifugation56,70,134,135,136. The film thickness and the contact area of the two materials can be well controlled. Figures 7a and b show NiO@SnO2 and Ga2O3@WO3 nanofilms obtained by sputtering for ethanol detection135,137. However, these methods generally produce flat films, and these flat films are less sensitive than 3D nanostructured materials due to their low specific surface area and gas permeability. Therefore, a liquid-phase strategy for fabricating bilayer films with different hierarchies has also been proposed to improve perceptual performance by increasing the specific surface area41,52,138. Zhu et al139 combined sputtering and hydrothermal techniques to produce highly ordered ZnO nanowires over SnO2 nanowires (ZnO@SnO2 nanowires) for H2S detection (Fig. 7c). Its response to 1 ppm H2S is 1.6 times higher than that of a sensor based on sputtered ZnO@SnO2 nanofilms. Liu et al. 52 reported a high performance H2S sensor using a two-step in situ chemical deposition method to fabricate hierarchical SnO2@NiO nanostructures followed by thermal annealing (Fig. 10d). Compared to conventional sputtered SnO2@NiO bilayer films, the sensitivity performance of the SnO2@NiO hierarchical bilayer structure is significantly improved due to the increase in specific surface area52,137.
Double layer gas sensor based on MOS. NiO@SnO2 nanofilm for ethanol detection; 137b Ga2O3@WO3 nanofilm for ethanol detection; 135c highly ordered SnO2@ZnO bilayer hierarchical structure for H2S detection; 139d SnO2@NiO bilayer hierarchical structure for detecting H2S52.
In type II devices based on core-shell heteronanostructures (CSHNs), the sensing mechanism is more complex, since the conduction channels are not limited to the inner shell. Both the manufacturing route and the thickness (hs) of the package can determine the location of the conductive channels. For example, when using bottom-up synthesis methods, conduction channels are usually limited to the inner core, which is similar in structure to two-layer or multilayer device structures (Fig. 2b(3)) 123, 140, 141, 142, 143. Xu et al. 144 reported a bottom-up approach to obtaining CSHN NiO@α-Fe2O3 and CuO@α-Fe2O3 by depositing a layer of NiO or CuO NPs on α-Fe2O3 nanorods in which the conduction channel was limited by the central part. (nanorods α-Fe2O3). Liu et al. 142 also succeeded in restricting the conduction channel to the main part of CSHN TiO2 @ Si by depositing TiO2 on prepared arrays of silicon nanowires. Therefore, its sensing behavior (p-type or n-type) depends only on the semiconductor type of the silicon nanowire.
However, most reported CSHN-based sensors (Fig. 2b(4)) were fabricated by transferring powders of the synthesized CS material onto chips. In this case, the conduction path of the sensor is affected by the housing thickness (hs). Kim’s group investigated the effect of hs on gas detection performance and proposed a possible detection mechanism100,112,145,146,147,148. It is believed that two factors contribute to the sensing mechanism of this structure: (1) the radial modulation of the EDL of the shell and (2) the electric field smearing effect (Fig. 8) 145. The researchers mentioned that the conduction channel of the carriers is mostly confined to the shell layer when hs > λD of the shell layer145. It is believed that two factors contribute to the sensing mechanism of this structure: (1) the radial modulation of the EDL of the shell and (2) the electric field smearing effect (Fig. 8) 145. The researchers mentioned that the conduction channel of the carriers is mostly confined to the shell layer when hs > λD of the shell layer145. Считается, что в механизме восприятия этой структуры участвуют два фактора: (1) радиальная модуляция ДЭС оболочки и (2) эффект размытия электрического поля (рис. 8) 145. Исследователи отметили, что канал проводимости носителей в основном приурочено к оболочке, когда hs > λD оболочки145. It is believed that two factors are involved in the mechanism of perception of this structure: (1) radial modulation of the EDL of the shell and (2) the effect of blurring the electric field (Fig. 8) 145. The researchers noted that the carrier conduction channel is mainly confined to the shell when hs > λD shells145. It is believed that two factors contribute to the detection mechanism of this structure: (1) the radial modulation of the DEL of the shell and (2) the effect of electric field smearing (Fig. 8) 145.研究人员提到传导通道当壳层的hs > λD145 时，载流子的数量主要局限于壳层。 > λD145 时，载流子的数量主要局限于壳层。 Исследователи отметили, что канал проводимости Когда hs > λD145 оболочки, количество носителей в основном ограничено оболочкой. The researchers noted that the conduction channel When hs > λD145 of the shell, the number of carriers is mainly limited by the shell. Therefore, in the resistive modulation of the sensor based on CSHN, the radial modulation of the cladding DEL prevails (Fig. 8a). However, at hs ≤ λD of the shell, the oxygen particles adsorbed by the shell and the heterojunction formed at the CS heterojunction are completely depleted of electrons. Therefore, the conduction channel is not only located inside the shell layer but also partially in the core part, especially when hs < λD of the shell layer. Therefore, the conduction channel is not only located inside the shell layer but also partially in the core part, especially when hs < λD of the shell layer. Поэтому канал проводимости располагается не только внутри оболочечного слоя, но и частично в сердцевинной части, особенно при hs < λD оболочечного слоя. Therefore, the conduction channel is located not only inside the shell layer, but also partly in the core part, especially at hs < λD of the shell layer.因此，传导通道不仅位于壳层内部，而且部分位于芯部，尤其是当壳层的hs < λD 时。 hs < λD 时。 Поэтому канал проводимости располагается не только внутри оболочки, но и частично в сердцевине, особенно при hs < λD оболочки. Therefore, the conduction channel is located not only inside the shell, but also partly in the core, especially at hs < λD of the shell. In this case, both the fully depleted electron shell and the partially depleted core layer help modulate the resistance of the entire CSHN, resulting in an electric field tail effect (Fig. 8b). Some other studies have used the EDL volume fraction concept instead of an electric field tail to analyze the hs effect100,148. Taking these two contributions into account, the total modulation of the CSHN resistance reaches its greatest value when hs is comparable to the sheath λD, as shown in Fig. 8c. Therefore, the optimal hs for CSHN can be close to the shell λD, which is consistent with experimental observations99,144,145,146,149. Several studies have shown that hs can also affect the sensitivity of CSHN-based pn-heterojunction sensors40,148. Li et al. 148 and Bai et al. 40 systematically investigated the effect of hs on the performance of pn-heterojunction CSHN sensors, such as TiO2@CuO and ZnO@NiO, by changing the cladding ALD cycle. As a result, sensory behavior changed from p-type to n-type with increasing hs40,148. This behavior is due to the fact that at first (with a limited number of ALD cycles) heterostructures can be considered as modified heteronanostructures. Thus, the conduction channel is limited by the core layer (p-type MOSFET), and the sensor exhibits p-type detection behavior. As the number of ALD cycles increases, the cladding layer (n-type MOSFET) becomes quasi-continuous and acts as a conduction channel, resulting in n-type sensitivity. Similar sensory transition behavior has been reported for pn branched heteronanostructures 150,151 . Zhou et al. 150 investigated the sensitivity of Zn2SnO4@Mn3O4 branched heteronanostructures by controlling the Zn2SnO4 content on the surface of Mn3O4 nanowires. When Zn2SnO4 nuclei formed on the Mn3O4 surface, a p-type sensitivity was observed. With a further increase in the Zn2SnO4 content, the sensor based on branched Zn2SnO4@Mn3O4 heteronanostructures switches to the n-type sensor behavior.
A conceptual description of the two-functional sensor mechanism of CS nanowires is shown. a Resistance modulation due to radial modulation of electron-depleted shells, b Negative effect of smearing on resistance modulation, and c Total resistance modulation of CS nanowires due to a combination of both effects 40
In conclusion, type II sensors include many different hierarchical nanostructures, and sensor performance is highly dependent on the arrangement of the conductive channels. Therefore, it is critical to control the position of the conduction channel of the sensor and use a suitable heteronanostructured MOS model to study the extended sensing mechanism of type II sensors.
Type III sensor structures are not very common, and the conduction channel is based on a heterojunction formed between two semiconductors connected to two electrodes, respectively. Unique device structures are usually obtained through micromachining techniques and their sensing mechanisms are very different from the previous two sensor structures. The IV curve of a Type III sensor typically exhibits typical rectification characteristics due to heterojunction formation48,152,153. The I–V characteristic curve of an ideal heterojunction can be described by the thermionic mechanism of electron emission over the height of the heterojunction barrier152,154,155.
where Va is the bias voltage, A is the device area, k is the Boltzmann constant, T is the absolute temperature, q is the carrier charge, Jn and Jp are the hole and electron diffusion current densities, respectively. IS represents the reverse saturation current, defined as: 152,154,155
Therefore, the total current of the pn heterojunction depends on the change in the concentration of charge carriers and the change in the height of the barrier of the heterojunction, as shown in equations (3) and (4) 156
where nn0 and pp0 are the concentration of electrons (holes) in an n-type (p-type) MOS, \(V_{bi}^0\) is the built-in potential, Dp (Dn) is the diffusion coefficient of electrons (holes), Ln (Lp ) is the diffusion length of electrons (holes), ΔEv (ΔEc) is the energy shift of the valence band (conduction band) at the heterojunction. Although the current density is proportional to the carrier density, it is exponentially inversely proportional to \(V_{bi}^0\). Therefore, the overall change in current density strongly depends on the modulation of the height of the heterojunction barrier.
As mentioned above, the creation of hetero-nanostructured MOSFETs (for example, type I and type II devices) can significantly improve the performance of the sensor, rather than individual components. And for type III devices, the heteronanostructure response can be higher than two components48,153 or higher than one component76, depending on the chemical composition of the material. Several reports have shown that the response of heteronanostructures is much higher than that of a single component when one of the components is insensitive to the target gas48,75,76,153. In this case, the target gas will interact only with the sensitive layer and cause a shift Ef of the sensitive layer and a change in the height of the heterojunction barrier. Then the total current of the device will change significantly, since it is inversely related to the height of the heterojunction barrier according to the equation. (3) and (4) 48,76,153. However, when both n-type and p-type components are sensitive to the target gas, detection performance can be somewhere in between. José et al.76 produced a porous NiO/SnO2 film NO2 sensor by sputtering and found that the sensor sensitivity was only higher than that of the NiO based sensor, but lower than that of the SnO2 based sensor. sensor. This phenomenon is due to the fact that SnO2 and NiO exhibit opposite reactions to NO276. Also, because the two components have different gas sensitivities, they may have the same tendency to detect oxidizing and reducing gases. For example, Kwon et al. 157 proposed a NiO/SnO2 p-n-heterojunction gas sensor by oblique sputtering, as shown in Fig. 9a. Interestingly, the NiO/SnO2 pn-heterojunction sensor showed the same sensitivity trend for H2 and NO2 (Fig. 9a). To solve this result, Kwon et al. 157 systematically investigated how NO2 and H2 change carrier concentrations and tuned \(V_{bi}^0\) of both materials using IV-characteristics and computer simulations (Fig. 9bd). Figures 9b and c demonstrate the ability of H2 and NO2 to change the carrier density of sensors based on p-NiO (pp0) and n-SnO2 (nn0), respectively. They showed that pp0 of p-type NiO slightly changed in the NO2 environment, while it changed dramatically in the H2 environment (Fig. 9b). However, for n-type SnO2, nn0 behaves in the opposite way (Fig. 9c). Based on these results, the authors concluded that when H2 was applied to the sensor based on the NiO/SnO2 pn heterojunction, an increase in nn0 led to an increase in Jn, and \(V_{bi}^0\) led to a decrease in the response (Fig. 9d ). After exposure to NO2, both a large decrease in nn0 in SnO2 and a small increase in pp0 in NiO lead to a large decrease in \(V_{bi}^0\), which ensures an increase in the sensory response (Fig. 9d) 157 In conclusion, changes in the concentration of carriers and \(V_{bi}^0\) lead to changes in the total current, which further affects the detection ability.
The sensing mechanism of the gas sensor is based on the structure of the Type III device. Scanning electron microscopy (SEM) cross-sectional images, p-NiO/n-SnO2 nanocoil device and sensor properties of p-NiO/n-SnO2 nanocoil heterojunction sensor at 200°C for H2 and NO2; b , cross-sectional SEM of a c-device, and simulation results of a device with a p-NiO b-layer and an n-SnO2 c-layer. The b p-NiO sensor and the c n-SnO2 sensor measure and match the I–V characteristics in dry air and after exposure to H2 and NO2. A two-dimensional map of the b-hole density in p-NiO and a map of c-electrons in the n-SnO2 layer with a color scale were modeled using the Sentaurus TCAD software. d Simulation results showing a 3D map of p-NiO/n-SnO2 in dry air, H2 and NO2157 in the environment.
In addition to the chemical properties of the material itself, the structure of the Type III device demonstrates the possibility of creating self-powered gas sensors, which is not possible with Type I and Type II devices. Because of their inherent electric field (BEF), pn heterojunction diode structures are commonly used to build photovoltaic devices and show potential for making self-powered photoelectric gas sensors at room temperature under illumination74,158,159,160,161. BEF at the heterointerface, caused by the difference in the Fermi levels of the materials, also contributes to the separation of electron-hole pairs. The advantage of a self-powered photovoltaic gas sensor is its low power consumption as it can absorb the energy of the illuminating light and then control itself or other miniature devices without the need for an external power source. For example, Tanuma and Sugiyama162 have fabricated NiO/ZnO pn heterojunctions as solar cells to activate SnO2-based polycrystalline CO2 sensors. Gad et al. 74 reported a self-powered photovoltaic gas sensor based on a Si/ZnO@CdS pn heterojunction, as shown in Fig. 10a. Vertically oriented ZnO nanowires were grown directly on p-type silicon substrates to form Si/ZnO pn heterojunctions. Then CdS nanoparticles were modified on the surface of ZnO nanowires by chemical surface modification. On fig. 10a shows off-line Si/ZnO@CdS sensor response results for O2 and ethanol. Under illumination, the open-circuit voltage (Voc) due to the separation of electron-hole pairs during BEP at the Si/ZnO heterointerface increases linearly with the number of connected diodes74,161. Voc can be represented by an equation. (5) 156,
where ND, NA, and Ni are the concentrations of donors, acceptors, and intrinsic carriers, respectively, and k, T, and q are the same parameters as in the previous equation. When exposed to oxidizing gases, they extract electrons from ZnO nanowires, which leads to a decrease in \(N_D^{ZnO}\) and Voc. Conversely, gas reduction resulted in an increase in Voc (Fig. 10a). When decorating ZnO with CdS nanoparticles, photoexcited electrons in CdS nanoparticles are injected into the conduction band of ZnO and interact with the adsorbed gas, thereby increasing the perception efficiency74,160. A similar self-powered photovoltaic gas sensor based on Si/ZnO was reported by Hoffmann et al. 160, 161 (Fig. 10b). This sensor can be prepared using a line of amine-functionalized ZnO nanoparticles ([3-(2-aminoethylamino)propyl]trimethoxysilane) (amino-functionalized-SAM) and thiol ((3-mercaptopropyl)-functionalized, to adjust the work function of the target gas for selective detection of NO2 (trimethoxysilane) (thiol-functionalized-SAM)) (Fig. 10b) 74,161.
A self-powered photoelectric gas sensor based on the structure of a type III device. a Self-powered photovoltaic gas sensor based on Si/ZnO@CdS, self-powered sensing mechanism and sensor response to oxidized (O2) and reduced (1000 ppm ethanol) gases under sunlight; 74b Self-powered photovoltaic gas sensor based on Si ZnO/ZnO sensors and sensor responses to various gases after functionalization of ZnO SAM with terminal amines and thiols 161
Therefore, when discussing the sensitive mechanism of type III sensors, it is important to determine the change in the height of the heterojunction barrier and the ability of the gas to influence the carrier concentration. In addition, illumination can generate photogenerated carriers that react with gases, which is promising for self-powered gas detection.
As discussed in this literature review, many different MOS heteronanostructures have been fabricated to improve sensor performance. The Web of Science database was searched for various keywords (metal oxide composites, core-sheath metal oxides, layered metal oxides, and self-powered gas analyzers) as well as distinctive characteristics (abundance, sensitivity/selectivity, power generation potential, manufacturing). Method The characteristics of three of these three devices are shown in Table 2. The overall design concept for high performance gas sensors is discussed by analyzing the three key factors proposed by Yamazoe. Mechanisms for MOS Heterostructure Sensors To understand the factors influencing gas sensors, various MOS parameters (eg, grain size, operating temperature, defect and oxygen vacancy density, open crystal planes) have been carefully studied. Device structure, which is also critical to the sensor’s sensing behavior, has been neglected and rarely discussed. This review discusses the underlying mechanisms for detecting three typical types of device structure.
The grain size structure, manufacturing method, and number of heterojunctions of the sensing material in a Type I sensor can greatly affect the sensitivity of the sensor. In addition, the behavior of the sensor is also affected by the molar ratio of the components. Type II device structures (decorative heteronanostructures, bilayer or multilayer films, HSSNs) are the most popular device structures consisting of two or more components, and only one component is connected to the electrode. For this device structure, determining the location of the conduction channels and their relative changes is critical in studying the mechanism of perception. Because type II devices include many different hierarchical heteronanostructures, many different sensing mechanisms have been proposed. In a type III sensory structure, the conduction channel is dominated by a heterojunction formed at the heterojunction, and the perception mechanism is completely different. Therefore, it is important to determine the change in the height of the heterojunction barrier after exposure of the target gas to the type III sensor. With this design, self-powered photovoltaic gas sensors can be made to reduce power consumption. However, since the current fabrication process is rather complicated and the sensitivity is much lower than traditional MOS-based chemo-resistive gas sensors, there is still a lot of progress in the research of self-powered gas sensors.
The main advantages of gas MOS sensors with hierarchical heteronanostructures are the speed and higher sensitivity. However, some key problems of MOS gas sensors (e.g., high operating temperature, long-term stability, poor selectivity and reproducibility, humidity effects, etc.) still exist and need to be addressed before they can be used in practical applications. Modern MOS gas sensors typically operate at high temperatures and consume a lot of power, which affects the long-term stability of the sensor. There are two common approaches to solving this problem: (1) development of low power sensor chips; (2) development of new sensitive materials that can operate at low temperature or even at room temperature. One approach to the development of low-power sensor chips is to minimize the size of the sensor by fabricating microheating plates based on ceramics and silicon163. Ceramic based micro heating plates consume approximately 50–70 mV per sensor, while optimized silicon based micro heating plates can consume as little as 2 mW per sensor when operating continuously at 300 °C163,164. The development of new sensing materials is an effective way to reduce power consumption by lowering the operating temperature, and can also improve sensor stability. As the size of the MOS continues to be reduced to increase the sensitivity of the sensor, the thermal stability of the MOS becomes more of a challenge, which can lead to drift in the sensor signal165. In addition, high temperature promotes the diffusion of materials at the heterointerface and the formation of mixed phases, which affects the electronic properties of the sensor. The researchers report that the optimum operating temperature of the sensor can be reduced by selecting suitable sensing materials and developing MOS heteronanostructures. The search for a low-temperature method for fabricating highly crystalline MOS heteronanostructures is another promising approach to improve stability.
The selectivity of MOS sensors is another practical issue as different gases coexist with the target gas, while MOS sensors are often sensitive to more than one gas and often exhibit cross sensitivity. Therefore, increasing the selectivity of the sensor to the target gas as well as to other gases is critical for practical applications. Over the past few decades, the choice has been partly addressed by building arrays of gas sensors called “electronic noses (E-nose)” in combination with computational analysis algorithms such as training vector quantization (LVQ), principal component analysis (PCA), etc. e. Sexual problems. Partial Least Squares (PLS), etc. 31, 32, 33, 34. Two main factors (the number of sensors, which are closely related to the type of sensing material, and computational analysis) are critical to improving the ability of electronic noses to identify gases169. However, increasing the number of sensors usually requires many complex manufacturing processes, so it is critical to find a simple method to improve the performance of electronic noses. In addition, modifying the MOS with other materials can also increase the selectivity of the sensor. For example, selective detection of H2 can be achieved due to the good catalytic activity of MOS modified with NP Pd. In recent years, some researchers have coated the MOS MOF surface to improve sensor selectivity through size exclusion171,172. Inspired by this work, material functionalization may somehow solve the problem of selectivity. However, there is still a lot of work to be done in choosing the right material.
The repeatability of the characteristics of sensors manufactured under the same conditions and methods is another important requirement for large-scale production and practical applications. Typically, centrifugation and dipping methods are low cost methods for fabricating high throughput gas sensors. However, during these processes, the sensitive material tends to aggregate and the relationship between the sensitive material and the substrate becomes weak68, 138, 168. As a result, the sensitivity and stability of the sensor deteriorate significantly, and the performance becomes reproducible. Other fabrication methods such as sputtering, ALD, pulsed laser deposition (PLD), and physical vapor deposition (PVD) allow the production of bilayer or multilayer MOS films directly on patterned silicon or alumina substrates. These techniques avoid buildup of sensitive materials, ensure sensor reproducibility, and demonstrate the feasibility of large-scale production of planar thin-film sensors. However, the sensitivity of these flat films is generally much lower than that of 3D nanostructured materials due to their small specific surface area and low gas permeability41,174. New strategies for growing MOS heteronanostructures at specific locations on structured microarrays and precisely controlling the size, thickness, and morphology of sensitive materials are critical for low-cost fabrication of wafer-level sensors with high reproducibility and sensitivity. For example, Liu et al. 174 proposed a combined top-down and bottom-up strategy for fabricating high-throughput crystallites by growing in situ Ni(OH)2 nanowalls at specific locations. . Wafers for microburners.
In addition, it is also important to consider the effect of humidity on the sensor in practical applications. Water molecules can compete with oxygen molecules for adsorption sites in sensor materials and affect the sensor’s responsibility for the target gas. Like oxygen, water acts as a molecule through physical sorption, and can also exist in the form of hydroxyl radicals or hydroxyl groups at a variety of oxidation stations through chemisorption. In addition, due to the high level and variable humidity of the environment, a reliable response of the sensor to the target gas is a big problem. Several strategies have been developed to address this problem, such as gas preconcentration177, moisture compensation and cross-reactive lattice methods178, as well as drying methods179,180. However, these methods are expensive, complex, and reduce the sensitivity of the sensor. Several inexpensive strategies have been proposed to suppress the effects of humidity. For example, decorating SnO2 with Pd nanoparticles can promote the conversion of adsorbed oxygen into anionic particles, while functionalizing SnO2 with materials with high affinity for water molecules, such as NiO and CuO, are two ways to prevent moisture dependence on water molecules. . Sensors 181, 182, 183. In addition, the effect of humidity can also be reduced by using hydrophobic materials to form hydrophobic surfaces36,138,184,185. However, the development of moisture-resistant gas sensors is still at an early stage, and more advanced strategies are required to address these issues.
In conclusion, improvements in detection performance (eg, sensitivity, selectivity, low optimum operating temperature) have been achieved by creating MOS heteronanostructures, and various improved detection mechanisms have been proposed. When studying the sensing mechanism of a particular sensor, the geometric structure of the device must also be taken into account. Research into new sensing materials and research into advanced fabrication strategies will be required to further improve the performance of gas sensors and address remaining challenges in the future. For controlled tuning of sensor characteristics, it is necessary to systematically build the relationship between the synthetic method of sensor materials and the function of heteronanostructures. In addition, the study of surface reactions and changes in heterointerfaces using modern characterization methods can help elucidate the mechanisms of their perception and provide recommendations for the development of sensors based on heteronanostructured materials. Finally, the study of modern sensor fabrication strategies may allow the fabrication of miniature gas sensors at the wafer level for their industrial applications.
Genzel, N.N. et al. A longitudinal study of indoor nitrogen dioxide levels and respiratory symptoms in children with asthma in urban areas. neighborhood. Health perspective. 116, 1428–1432 (2008).
Post time: Nov-04-2022
