The issue of air quality is still a major concern in many countries. A clean air supply is essential to our health and the environment. The human nose serves as a highly advanced sensing system which may differentiate between hundreds of smells but fails if absolute gas concentrations or odorless gases need to be detected. The demand for detecting toxic and deleterious gases is accordingly urgent to support or replace human nose. Although a large number of gas detecting systems have currently been used in process control and laboratory analytics [1–4], high performance gas sensors with high sensitivity, high selectivity and rapid response speed are also needed to improve the levels of gas detection.

Metal oxide gas sensors have been widely used in portable gas detection systems because of their advantages such as low cost, easy production, compact size and simple measuring electronics [5,6]. However, the performance of such sensors is significantly influenced by the morphology and structure of sensing materials, resulting in a great obstacle for gas sensors based on bulk materials or dense films to achieve highly-sensitive properties. Gas sensors based on nanomaterials are a greatly developing direction to improve gas sensing properties in sensitivity, selectivity and response speed. Although there are already some reviews on metal oxide gas sensor [7–9], it is still necessary to systematically summarize the features of metal oxides from the perspective of nanoscience and nanotechnology. In this review, we provide a brief summary on metal oxide nanostructures and their gas sensing properties from the aspects of particle size, morphology and doping. Most of the examples are given based on n-type metal oxides which are more extensively investigated and applied among the metal oxide gas sensors.

It is necessary to reveal the sensing mechanism of metal oxide gas sensors which is helpful for designing and fabricating novel gas sensing materials with excellent performance. Although the exact fundamental mechanisms that cause a gas response are still controversial, it is essentially responsible for a change in conductivity that trapping of electrons at adsorbed molecules and band bending induced by these charged molecules. Herein, a brief introduction to the sensing mechanism of n-type metal oxides in air is given based on the example of SnO₂. Typically, oxygen gases are adsorbed on the surface of the SnO₂ sensing material in air. The adsorbed oxygen species can capture electrons from the inner of the SnO₂ film. The negative charge trapped in these oxygen species causes a depletion layer and thus a reduced conductivity. When the sensor is exposed to reducing gases, the electrons trapped by the oxygen adsorbate will return to the SnO₂ film, leading to a decrease in the potential barrier height and thus an increase in conductivity. There are different oxygen species including molecular (O₂⁻) and atomic (O⁻, O²⁻) ions on the surface depending on working temperature. Generally, below 150 °C the molecular form dominates while above this temperature the atomic species are found [9,10].

The overall surface stoichiometry has a decisive influence on the surface conductivity for the metal oxides. Oxygen vacancies act as donors, increasing the surface conductivity, whereas adsorbed oxygen ions act as surface acceptors, binding elections and diminishing the surface conductivity. Figure 1 shows the energy diagram of various oxygen species in the gas phase, adsorbed at the surface and bound within the lattice of SnO₂ [11,12]. On SnO₂ films the reaction O₂⁻_ads + e⁻ = 2O⁻_ads takes place as the temperature increases. The desorption temperatures from the SnO₂ surface are around 550 °C for O⁻_ads ions and around 150 °C for O₂⁻_ads ions. At constant oxygen coverage, the transition causes an increase in surface charge density with corresponding variations of band bending and surface conductivity. From conductance measurements, it is concluded that the transition takes place slowly. Therefore, a rapid temperature change on the part of the sensors is usually followed by a gradual and continuous change in the conductance. The oxygen coverage adjusts to a new equilibrium and the adsorbed oxygen is converted into another species which may be used in measurement method of dynamic modulated temperature as reported previously [13–19].

Gas sensors based on metal oxide nanostructures generally consist of three parts, i.e., sensing film, electrodes and heater. Metal oxide nanostructures react in the form of a film which will change in resistance upon exposure to target gases. A pair of electrodes is used to measure the resistance of the sensing film. Usually the gas sensors are furnished with a heater so that they are heated externally to reach an optimum working temperature. Currently, metal oxide nanostructures sensors have been characterized in three ways: conductometric, field effect transistor (FET) and impedometric ones [20]. FET type is usually exploited to fabricate sensors based on single or arrays of one-dimensional (1D) semiconducting nanomaterials, which have a complex fabrication process. Impedometric type sensors are based on impedance changes and are operated under alternating voltage upon exposure to target species, which has not yet attracted much attention. The conductometric type is the most common gas sensor which is suitable for most nanomaterials. There are two types of device structures in conductometric sensors: directly heated and indirectly heated. A directly heated type structure means the heater is contacted with the sensing material, which may lack stability and anti-interference ability, so most of the current nanostructure-based gas sensors are indirectly heated type structures which can be divided into two types, i.e., cylindrical and planar layouts, as shown in Figures 2 and 3. Alumina ceramics (wafers or tubes) are generally used as substrates to support sensing films. In the ceramic tube-based device, a piece of heating wire is placed in the interior of the ceramic tube, while, in the ceramic wafer-based device, heating paste is placed on the backside of the ceramic wafer. Some silica wafers can also be used as the substrate, which is advantageous in manufacturing small sized gas sensor because of its compatibility with integrated circuits.

The “small size effect” of metal oxides has been reported by many publications [21–27]. As shown in Figure 4, a sensor is considered to be composed of partially sintered crystallites that are connected to their neighbors by necks. Those interconnected grains form larger aggregates that are connected to their neighbors by grain boundaries [28]. On the surface of the grains, adsorbed oxygen molecules extract electrons from the conduction band and trap the electrons at the surface in the form of ions, which produces a band bending and an electron depleted region called the space-charge layer. When the particle size of the sensing film is close to or less than double the thickness of the space-charge layer, the sensitivity of the sensor will increase remarkably. Xu et al. explained the phenomena by a semiquantitative model [29]. Three different cases can be distinguished according to the relationship between the particle size (D) and the width of the space-charge layer (L) that is produced around the surface of the crystallites due to chemisorbed ions and the size of L is about 3 nm for pure SnO₂ material in literatures [30–34]. When D >> 2L, the conductivity of the whole structure depends on the inner mobile charge carriers and the electrical conductivity depends exponentially on the barrier height. It is not so sensitive to the charges acquired from surface reactions. When D ≥ 2L, the space-charge layer region around each neck forms a constricted conduction channel within each aggregate. Consequently, the conductivity not only depends on the particle boundaries barriers, but also on the cross section area of those channels and so it is sensitive to reaction charges. Therefore, the particles are sensitive to the ambient gas composition. When D < 2L, the space-charge layer region dominates the whole particle and the crystallites are almost fully depleted of mobile charge carriers. The energy bands are nearly flat throughout the whole structure of the interconnected grains and there are no significant barriers for intercrystallite charge transport and then the conductivity is essentially controlled by the intercrystallite conductivity. Few charges acquired from surface reactions will cause large changes of conductivity of the whole structure, so the crystalline SnO₂ becomes highly sensitive to ambient gas molecules when its particle size is small enough.

Based on Xu’s model, many new sensing materials are developed to achieve high gas sensing properties [35–37]. Typically, the nanocomposite of SnO₂ and multiwall carbon nanotube (MWCNT) was exploited to detect persistent organic pollutants (POPs) which possess stable chemical properties and are ordinarily difficult to detected with metal oxides [38]. The preparation of materials with size and porosity in the nanometer range is of technological importance for a wide range of sensing applications. The ultrasensitive detection of aldrin and dichlorodiphenyltrichloroethane (DDT), has been carried out using the nanocomposite of small SnO₂ particles and MWCNTs. The nanocomposite shows a very attractive improved sensitivity compared with a conventional SnO₂ sensor. A sharp response of low limiting concentration about 1 ng was observed in both aldrin and DDT, suggesting potential applications as a new analytical approach. One major advantage of this sensing material is its stable attachment between sub-10 nm SnO₂ nanoparticles and carbon nanotubes shown is Figure 5. Besides, the SnO₂/MWCNT nanocomposite synthesized by a wet chemical method may control the size of SnO₂ particles under 10 nm and form highly porous three dimensional (3D) structures. Among the highly porous 3D structures, MWCNTs can be regarded as the framework and the SnO₂ particles uniformly packed on them, which may enhance the ability of gas diffusion into and out of the sensing film. The high sensitivity can also be attributed to an effect of p-n junction formed between p-type carbon nanotubes and n-type SnO₂ nanoparticles. The investigation results make SnO₂/MWCNT nanocomposites attractive for the purpose of POPs detection.

Commonly, metal oxide sensing films are divided into dense and porous [10]. In dense films, the gas interaction takes place only at the surface of the film since the analyte cannot penetrate into the sensing film. In porous films, the gas can penetrate into the film and interact with the inner grains. In fact, metal oxide films are usually produced with a certain overall porosity through several processes, which is yet insufficient for gas sensing.

Apart from large surface-to-volume ratios, well-defined and uniform pore structures are particularly desired for metal oxides to improve sensing performance. Porous materials are classified into several kinds according to their size. According to the definition of the International Union of Pure and Applied Chemistry (IUPAC) [39], microporous materials have pore diameters of less than 2 nm and macroporous materials have pore diameters of greater than 50 nm; the mesoporous category thus lies in the middle. Mesoporous materials have lots of application in the fields of drug delivery, catalysts, energy storage and detection of gas pollution. Mesoporous oxide structures with well-aligned pore structures are fascinating for gas sensing investigation. For example, mesoporous SnO₂ has attracted more interests because of their high sensitive and rapid gas response, which can facilitate gas diffusion and mass transport due to mesopores providing regions for exchanging gases. Mesoporous materials can be prepared via many methods such as template synthesis [40–43], hydrothermal/solvothermal approaches [44–46], self-assembly reaction [47–51], the Kirkendall effect [52–54], Ostwald ripening [55,56] and so on. Among those methods, the mesoporous SnO₂ (Figure 6) prepared by the method of MWCNT templates exhibited an excellent gas sensing properties [57].

Compared to traditional SnO₂, the mesoporous SnO₂ has better permeability because mesoporous SnO₂ provides more space for gas molecules to diffuse in and out of the film. The gas sensing comparison indicates SnO₂ mesoporous materials have much better response to ethanol and benzene, especially benzene. The key parameters to determine the gas sensing characteristics are thickness, permeability and surface morphology, while mesoporous structure has better permeability. During the response and recovery process, target gas molecules diffuse in and out of SnO₂ film. A diffusion equation assuming a first-order reaction of target gas is inducted by Sakai and co-workers to explain gas diffusion dynamics in the response process. In the mesopores, gas diffusion constant (D_k) is determined by temperature (T), pore radius (r), molecular weight (M) of the diffusion gas as following equation [58]: (1) D k = 4 r 3   2 RT π Mwhere R is gas constant. The molecules of sample gas diffuse into the surface of the mesoporous SnO₂ film and react with the surface oxygen of SnO₂ chains subsequently [58,59]. The reaction of molecules occurs only on the out surface region of the traditional SnO₂ film. Since the SnO₂ mesoporous structure can increase response region and the inner parts become active, the mesoporous SnO₂ materials are more sensitive.

In the past few years, many efforts have been devoted to improve the sensitivity of gas sensors. Sakai et al. found that the porous structure of the sensing film played a critical role in the performance of the sensor because it decided the rate of gas diffusion [58]. Xu et al. found that the particle size heavily affected the sensitivity of sensor [29]. Although many methods have been reported to synthesize monodisperse nanoparticles of metal oxides [60–64], small size of nanoparticles are not stable which may easily congregate and grow up under heating conditions [31]. Besides, small sized nanoparticles will be compactly sintered together during the film coating process which is a disadvantage for gas diffusion in them. If porous nanostructures are used as gas sensing materials, the gas sensing properties will be much improved. On the basis of those reasons, nanoparticles-assembled nanostructures with many kinds of shapes such as porous nanowires, porous nanotubes, porous nanospheres and so on are reviewed in this chapter, which exhibited excellent gas sensing properties because they not only possessed large surface area and relatively mass reactive sites, but also formed relatively loose film structures.

Doping of metal oxide sensing film is a traditional technology for gas sensors. The traditional concept of doping is to enhance catalytic activity and adjust electrical resistance of the intrinsic metal oxide [103–105]. The dopant is usually high active, which make it react preferentially with adsorbed molecules. As shown in Figure 11, the dopant is generally dispersed on the metal oxide matrix so that they are available near all the intergranular contacts. In air, the oxygen molecules react preferentially with the dopant forming oxygen anions and then spill over to the metal oxide matrix. When the target gases are adsorbed on to the surface of the dopant and then migrate to the oxide surface to react with surface oxygen species thereby increasing the surface conductivity [106].

However, as the development of nanotechnology, doping is given many novel meanings. A typical doping phenomenon concerns the fact that the particle size of the doped metal oxide becomes smaller than the pure one [27,28] which can be explained by Nae-Lih Wu’s theory [29], i.e., because of the interaction on the boundaries between host and dopant crystallites, the motion of crystallites is restricted [107–109]. In other words, the advancing of grain boundaries which is required for crystal growth is stunted. As a result, the size of crystallites is decreased by the doping of impurities.

Gong et al. have investigated the role of the Cu doping in enhancing the capability to adsorb CO molecules [110]. According to their results, the Cu site in ZnO film plays an important role to adsorb CO molecules at both low and high temperatures. When CO molecules are adsorbed on the film, they are preferably adsorbed on the Cu sites to form bonds between Cu and CO. The interacting bonding between Cu and CO consists of the donation of CO 5σ electrons to the metal and the back donation of π electrons from d-orbitals of Cu to CO. That adsorption results in the enhancement of the reactivity to CO. The CO adsorption mainly takes place at the Cu sites but not at the Zn sites, and then CO molecules migrate from the Cu to the Zn sites [111], by which the Cu sites enhance the CO adsorption and thus the reaction of CO with oxygen species.

Another important role of doping in enhancing gas sensing properties is to form p-n junctions which may increase the depletion barrier height due to the electron transfer from n-type materials to p-type ones [112]. When the sensor was exposed to reducing target gases, the electrons trapped by absorbed oxygen species and p-type materials are feed back to n-type materials through surface interactions, resulting in a significantly decreased sensor resistance. Therefore, the sensor response was improved remarkably.

If the doping is integrated into a high-sensitive nanostructure, the sensitivity will be further improved. Xue et al. have prepared n-type SnO₂ nanorods uniformly coated with p-type CuO nanoparticles via a hydrothermal method which exhibited super-high sensitivity to H₂S [113]. Besides, both the gold- or Pt-doped In₂O₃ nanowires have revealed higher sensitivity than the bare ones [114,115]. He et al. further improved the sensing properties to H₂S of CuO-doped SnO₂ material by replacing SnO₂ nanorods with SnO₂ hollow spheres. The CuO-doped SnO₂ hollow spheres as shown in Figure 12 exhibited a ppb-leveled detection limit at a relatively low working temperature of 35 °C [116]. Besides, high selectivity was also acquired as shown in Figure 13, from which it can be seen that CuO-doped SnO₂ hollow spheres could distinguish a small amount of (10 ppm) H₂S among large amount of other gases including 1000 ppm of H₂, NH₃, ethanol and benzene.

Recently, metal oxide nanosturctures has been doped by many physical or chemical methods, such as thermal evaporation [117], sputter deposition [118], spin coating [119] and wet chemical methods [116]. However, a new technology for uniform and dense doping is highly desired. Liu et al. have developed a plasma-assisted strategy for highly dense doping of metal oxide nanostructures [120]. Figure 14 has schematically illustrated the plasma-assisted strategy for preparing highly dense In-doped SnO₂ coral-like nanostructures.

Firstly, coral-like SnO₂/carbonaceous nanocomposites were synthesized via a hydrothermal route. Then, the nanocomposites were functionalized by plasma treatment. The densities of some functional groups, such as hydroxyl and carboxyl, can be greatly increased on the surface of nanocomposites, which is significant for further adsorbing In³⁺ ions to achieve dense doping. The plasma-treated SnO₂/carbonaceous nanocomposites were ultrasonically dispersed in In³⁺ ion solution and left static for a long time and subsequently washed and centrifugated. Finally, the In-doped SnO₂ coral-like nanostructures combined with porous and hollow structures were prepared by following an annealing process to remove the sacrificed carbonaceous templates. In gas-sensing measurements, the In-doped SnO₂ coral-like nanostructures with plasma treatment exhibited highly sensitive to chlorobenzene with a high response and short response and recovery times.

Although metal oxide gas sensors are predominantly solid-state gas detecting devices with many advantages such low cost, easy production, and compact size, and thus have been widely-used in many fields such as public safety, pollutant monitoring and so on, there is still room to improve the gas sensing performance of such sensors by controlling the morphology and structure of sensing materials. Here, gas sensing mechanisms have been reviewed first for better understanding their working principles. Then, the influences of size effect, porous nanostructure and doping on nanoscale levels have been described. By considering those influencing factors on nanoscale, novel metal oxide nanostructures will be developed and then gas sensing properties of metal oxides will be much improved.

On the basis of current progress in the field of metal oxide gas sensors, it is anticipated that the following aspects would be promising directions for developing in the future: (1) novel nanostuctures or nanocomposites which may achieve super-sensitive detection; (2) combining porous nanostructures which possess fast responses and recovery characteristics to a chromatographic technique; (3) exploiting first principles to further investigate the gas sensing mechanisms. The research on gas sensors is related to many fields such as physics, chemistry, electronics and mathematics. Addressing those problems will be one of the great challenges and it is important to enhance interdisciplinary collaboration.