The mechanisms of sensory phenomena in binary metal-oxide nanocomposites

The mechanisms of sensory phenomena in binary metal-oxide nanocomposites

Sensors and Actuators B 240 (2017) 613–624 Contents lists available at ScienceDirect Sensors and Actuators B: Chemical journal homepage: www.elsevie...

2MB Sizes 0 Downloads 12 Views

Sensors and Actuators B 240 (2017) 613–624

Contents lists available at ScienceDirect

Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb

The mechanisms of sensory phenomena in binary metal-oxide nanocomposites G.N. Gerasimov a , V.F. Gromov a , O.J. Ilegbusi b,∗ , L.I. Trakhtenberg a,c a b c

Semenov Institute of Chemical Physics, Russian Academia of Sciences, 4, Kosygina Str., Moscow 119991, Russia University of Central Florida, 4000 Central Florida Boulevard, Orlando, FL 32816-2450, USA Moscow Institute of Physics and Technology (State University), 9, Institutskii per., Dolgoprudny, Moscow Region 141700 Russia

a r t i c l e

i n f o

Article history: Received 13 June 2016 Received in revised form 23 August 2016 Accepted 3 September 2016 Available online 4 September 2016 Keywords: Nanostructured semiconductors Core-shell type nanofibers Electronic and chemical sensitization Percolation transition Reducing gases Study in operando

a b s t r a c t This study reviews the structure and properties of nanostructured composite conductometric sensors based on semiconducting metal oxides, and the physico-chemical processes occurring when applied for detection of ambient reducing gases. It discusses the mechanisms of electronic and chemical sensitization in composites comprised of metal oxides of different electronic and chemical properties. In particular, the relationship between the conductivity mechanisms and sensory effect is examined, considering the transfer of electrons between the oxide components of the composite semiconductor sensor. A separate section is devoted to new systems consisting of composite nanofibers of the core-shell type, the sensory characteristics of which depend on the transfer of electrons between the core and the shell in the nanofiber. It is demonstrated that by changing the nature of the components and their relative locations in such nanofibers, the sensitivity and selectivity of the sensor system can be tailored to various chemical compounds. Thus, the use of composite metal oxide systems can lead to improved efficiency and selectivity of conductometric sensors, and enable the development of sensor systems with the desired operating properties. © 2016 Published by Elsevier B.V.

Contents 1. 2. 3. 4.

5. 6.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613 Electronic structure of semiconductor nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614 Models of sensory phenomena in nanostructured semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615 Sensors based on metal-oxide composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 616 4.1. Binary composites of type 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 616 4.2. Binary composites of type 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 617 Sensor properties of core-shell type nanofibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 620 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 621 Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 622 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 622 Biography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 622

1. Introduction Conductometric chemical sensors based on nanocrystalline semiconducting metal oxides are prospective sensitive elements for the detection of harmful environmental substances. As a result, there is considerable interest in the design and investigation of such

∗ Corresponding author. E-mail address: [email protected] (O.J. Ilegbusi). http://dx.doi.org/10.1016/j.snb.2016.09.007 0925-4005/© 2016 Published by Elsevier B.V.

sensors. The observed sensory effect when applied for detection of reducing gases results from the reaction of adsorbed molecules of these gases with oxygen anions chemisorbed on the sensor surface. The electrons released in this reaction are subsequently returned to the conduction band of the semiconductor nanoparticles. In the presence of reducing gases the conductivity of n-type semiconductors increases, while it decreases for p-type semiconductors. A typical example is the following sensory reaction occurring in the detection of ambient CO: COad + O(−)ad → CO2 + e .

(I)

614

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

The CO2 molecules produced are loosely coupled with the sensor surface and are easily removed, such that the CO adsorption sites on the sensor surface, and its functionality are retained. The sensor response is defined as the ratio Ra /Rg , where Ra and Rg are respectively the sensor resistances in air, and air containing the detected gas. One of the major drawbacks of one-component semiconductor sensors, such as SnO2 or In2 O3 sensors, is their low selectivity to compounds of the same type, e.g., various reducing gases. The use of mixed systems containing metal oxides with different electronic and chemical properties represents a promising direction to enhance the sensor selectivity and sensitivity. By varying the type and composition of the components in binary metal oxide systems as well as the temperature, the response rate and selectivity of the sensor response can be adjusted for gas detection, for example, H2 and CO [1,2]. The addition of small amounts of ZnO to SnO2 , doped with 1 mol% CuO was found to shift the maximal sensory efficiency in the detection of H2 in the presence of CO towards higher temperatures [1]. At a relatively low temperature (160 ◦ C), the value of SCO /SH2 is ∼5 (SCO and SH2 being the sensor efficiency in the detection of CO and hydrogen, respectively) while an increase in the temperature to 310 ◦ C leads to selective detection of hydrogen (SH2 /SCO ∼2). This review focuses primarily on the study of the influence of charge transfer between the components of binary sensor systems on sensor properties. It also clarifies the relationship between the conductivity mechanisms and sensory characteristics of binary nanocomposite sensors of different compositions in the detection of reducing gases.

2. Electronic structure of semiconductor nanoparticles The conductive and sensory properties of composite sensors depend primarily on the electronic structure of the semiconductor nanoparticles. Qualitatively, the state of electrons and positive charges in nanostructured materials can be described as follows. There is a significant concentration of electrons in the semiconductor nanoparticles conduction band, which is provided by the thermal ionization of donor impurities. The electronic levels of such donors are located near the bottom of the conduction band [3]. In air ambience, the oxygen molecules adsorbed on the nanoparticle surface dissociate at 200–400 ◦ C. The atoms produced capture electrons from the nanoparticle volume [4], and the number of free electrons in the conduction band of the nanoparticles is markedly reduced. The conduction electron density (nc ) and the number of charged surface sites (O− ), depend on temperature, depth and concentration of the traps of donor (nd ), the energy of the electrons in O− and diameter (d) of the nanoparticle [5]. In order to form a surface layer of negative O− it is necessary that the amount of conduction electrons (/6)d3 nc be quite large (»1). For example, in a nanocrystalline film of SnO2 at 200 ◦ C the nc value measured by the Hall effect is 1016 –1018 cm−3 [6,7]. In this case the nanoparticle with a diameter of 100 nm at nc = 1016 cm−3 must contain only 5 electrons, and considering that only a fraction of the electrons comes to the surface, the formation of the surface layer is essentially eliminated [8]. For nanostructured systems such as In2 O3 where nc reaches 1019 cm−3 [9,10], the number of conduction electrons in the nanoparticle is sufficient to form a negatively charged surface layer. The adsorbates i.e. electron traps, on the surface are approximately uniformly distributed, and when there are a large number of trapped electrons, the surface of the spherical particles can be assumed to be uniformly charged. According to the Gauss theorem the electric field generated by this charge inside the nanoparticles is zero. The field outside the nanoparticles, due to its electrical neu-

trality, is also absent. Therefore, in the case of high concentrations of conduction electrons, the charged surface layer does not directly affect the behavior of the charges inside the nanoparticles. The main consequence of the electron capture by surface traps is inequality of the number of positive and negative charges within the nanoparticles. This inequality leads to highly inhomogeneous distribution of conduction electrons inside the nanoparticles such that the nanoparticles in the near-surface region form a layer of positive charge, which compensates for the negative charge of oxygen ions on the particle surface [5]. The electron distribution and sensory characteristics have been calculated for the plane semiconductor layer [11,12], which is valid only for nanoparticles having sufficiently large radius. The distribution of the electron density inside the spherical nanoparticles containing a negative charge on the surface has also been calculated by solving the Poisson equation [13,14]. In these studies, the charge on the surface was set arbitrarily, based only on the density of surface levels, and the distribution of positive charges inside the particle was assumed to be homogeneous. The distribution of positive and negative charges in nanoparticles having low and high electron density in the conduction band has been thoroughly analyzed [5,8,15]. If at the low electron density in the conduction band the negatively charged layer on the nanoparticle surface is absent [8], such a layer exists at high electron density and must be taken into account. The problem of spatial and energy distribution of charge in a spherical nanoparticle has been solved [5]. The equations, from which these distributions are determined, were obtained by minimizing the total free energy of the charges in the nanoparticle system. This approach allowed determination of the concentrations of electrons at the surface, the conduction electrons and positively charged ionized donors, as well as the electrical potential within the nanoparticles. The values of all these properties are functions of temperature and distance from the center as well as the radius of the nanoparticles and the depth of traps on the surface. It should be emphasized that in contrast to the studies in [13,14], all the characteristics of nanoparticles, including surface charge, distribution of positive charges and the dependence of the conduction electron concentration on temperature, were determined without any additional assumptions and simplifications [5]. The calculated characteristics correspond to complete statistical equilibrium in the particle, including the interaction of all charges. In addition, both the surface charge density, and the degree of ionization of the donor impurities are functions of temperature. At incomplete ionization of donors, as a rule, the conduction electrons as well as the positive charges exhibit non-uniform distribution. This significantly changes the charge distribution and electric field inside the nanoparticles. In order to demonstrate the critical effect of inhomogeneity of the positive charge distribution, the spatial distribution of conduction electrons was calculated assuming that the positively charged ionized donors were homogeneously distributed throughout the nanoparticle volume [13,14], and also for the model [5]. A typical example was considered of nanoparticles with radius R = 25 nm at T = 500 K. It was found that the difference in the relative concentrations of the conduction electrons in these two models could be up to five orders of magnitude [5]. The inhomogeneity of positive charge distribution inside the particle results in fundamental changes in the characteristics of the nanoparticles considered that should significantly influence the sensory properties of nanostructured semiconductor. The temperature dependence of the electron density at the surface of the nanoparticles, in contrast to the dependence accepted in most previous studies, in particular, the models in [11,12] is markedly different from the Boltzmann’s distribution [5].

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

3. Models of sensory phenomena in nanostructured semiconductors

[20]. The sensory process in this case proceeds in the following manner:

The total resistance of nanostructured systems with d  LD (LD – Debye length) is determined by the contacts between the nanocrystals, because the resistance between intergranular contacts is much higher than the resistance inside the crystals [11,16]. The conductivity of such systems is proportional to nc in the surface layers of nanoparticles [8,12]. The existing schemes of the sensory process are based on a variety of mechanisms of interaction between the reducing gas and the surface of the sensor nanoparticles. In accordance with the ionosorption model, the oxygen chemisorption occurs as a result of electron capture from the conduction band of a semiconductor. This leads to the dissociation of the oxygen molecule and O− formation. According to the scheme of sensory processes in nanostructured semiconductors [12,17] developed on the basis of this model [18], all the oxygen vacancies are ionized completely and all electrons from these vacancies have moved to the conduction band. The barrier of electron transfer between the particles, which determines the stationary conductivity of the nanostructured sensor, depends on the surface concentration of the O− anion-radicals. This value in the presence of reducing agents is determined by the processes of O− formation upon the dissociation of adsorbed oxygen, as a result of capture of electrons from the conduction band, and O− death due to their recombination, as well as the reaction between O− and reducing agents thus, k1

k2

k−1

A

O2 + 2e  2O− →2e + reaction products.

(II)

Here k1 is the rate constant of adsorbed oxygen dissociation as a consequence of the conduction electrons captured from the semiconductor surface, and k-1 is the rate constant of O− recombination and return of the electrons to the conduction band, A is the analyzed reducing gas and k2 is the rate constant of the reaction of this gas with O− . The value of the sensory effect, S, in this case is proportional to the square root of the partial pressure of analyzed gas, which agrees well with the experimental data [19]. At high sensor sensitivity when the rate of reaction of O− with the reducing gas on the nanoparticle surface far exceeds the O− death rate, the sensor sensitivity does not depend on the rate of O− formation on the semiconductor nanoparticle surface. The sensitivity of the sensor is determined solely by the concentration of adsorbed reducing gas on the surface and the ratio of the constant rate of gas reaction with chemisorbed oxygen anions to the rate constant of their spontaneous death. The same conclusion can be drawn from the model presented in [11], in which the spontaneous death of oxygen anions is considered as a monomolecular process, and steady state concentration of O− anion radicals in the presence of a reducing agent A is defined by the following scheme, k1

k2

k−1

A

1/2O2 + e  O− →e + reaction products.

615

(III)

The expression for the sensory effect corresponding to scheme (III) was determined for nanostructured semiconductors having both d ≤ LD , and d  LD . It was shown that in both cases as well as in [12], the sensory effect essentially does not depend on the rate of formation of O− on the nanoparticle surface. In some models [8,15] unlike others [11,12], the first stage is the dissociation of oxygen molecules chemisorbed on neutral atoms, which then capture electrons from the conduction band. In SnO2 such a process can occur with the participation of oxygen vacancies

k1

O2 ad  2Oad k−1 ke

Oad + e− O(−)ad

(IV)

k−e k2

H2 ad + O(−)ad →H2 Odes + e− Here k1 is the rate constant of adsorbed oxygen dissociation, k-1 is the rate constant of oxygen atoms recombination, ke and k-e are respectively the rate constants of electron capture by O atoms and electron removal from O− , codes «ad» and «des» refer, respectively, to molecules adsorbed on nanoparticle surface and desorbed from the surface. The constant k2 is the same as in the previous Scheme II. In addition, the transfer of electrons from the oxygen vacancies in the conduction band is considered a thermally activated process. Therefore, in contrast to the models in [11,12] the calculation of sensory effect takes into account the fact that the proportion of ionized vacancies and the ratio of the concentration of O to O− depend on the temperature and the surface concentration of adsorbed gas. For nanoparticles with low nc (SnO2 ), the theoretical dependence of the sensitivity on the temperature and size of the nanoparticles are in good agreement with the experimental data [8,15]. It should be noted that although ionosorption model describes the functionality of metal oxide sensors, at present there is no convincing direct evidence of formation of anion-radicals O2 − and O− [21] on the nanoparticle surface. An alternative reductionreoxidation model of sensory effect is proposed in [22,23]. According to this model, under the influence of a reducing gas, there occurs a reduction of O2 − anions of metal oxide lattice on the nanoparticle surface. As a result, electron centers are formed, that donate electrons to the metal oxide conduction band. This leads to an increase in sensor conductivity. After removal of the reducing gas these centers are completely oxidized by oxygen, and sensor conductivity drops to its original value. Characteristically, the sensory effect in reducing gases detected by metal oxide sensors is also observed in inert N2 atmosphere in the absence of oxygen. In this case, the return of the conductivity to the original value after removal of the reducing gas from the environment is much slower than in the presence of oxygen. The mechanism of this process has not yet been elucidated [23]. In order to determine the mechanism of sensory effect at different concentrations of oxygen and reducing gas in the environment, some studies have considered simultaneous measurements of spectral and electrical characteristics of the sensor in gas detection (an operando) [24,25]. The IR spectra of the diffuse scattering (IRDS) of nanocrystalline SnO2 films and sensor responses of these films to CO and H2 in N2 -O2 gas mixtures with different concentrations of the components have been investigated in the operando mode [26–28]. In nitrogen atmosphere the intensity of Sn-O absorption bands in IRDS spectra is reduced under the influence of CO. This indicates partial removal of oxygen atoms from the surface layers of SnO2 nanocrystals. Simultaneously the film conductivity increases significantly due to the formation of donor centers, i.e. oxygen vacancies. Additives of oxygen lead to oxidation of the vacancies and corresponding increase in the intensity of Sn-O absorption bands in the IRDS spectra, and a decrease in film conductivity [26]. When the oxygen concentration is less than 250 ppm, the sensory effect is due to a balance between the processes of formation and death (oxidation) of the vacancies [29]. At such oxygen concentrations, a similar reduction-oxidation process also defines the sensory effect for H2 . The donor centers in this case are hydroxyl groups formed upon reaction of H2 with

616

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

oxygen ions on SnO2 nanoparticle surface and then are oxidized by oxygen [30]. When the sensor is operated in air the sensory mechanism changes. An operando study of conductivity and IRDS spectra of hydroxyl groups of SnO2 in air showed that the increase in conductivity of the sensor under the influence of hydrogen is accompanied by a decrease in the concentration of hydroxyl groups on the SnO2 surface [27]. On the other hand, the reaction of hydrogen with SnO2 particle surface should lead to an increase in the concentration of the hydroxyl groups [30]. Thus sensory effect in air is a result of the reaction of reducing gases with the oxygen, adsorbed on the SnO2 surface and captured conduction electrons of SnO2 [26,27]. The reason for the change in the sensory mechanism is that the O− anion radicals generated upon the dissociative adsorption of O2 on the metal oxide surface have a higher reactivity than the oxygen anions of the metal-oxide lattice. Thus the studies in operando mode show that in the inert gaseous environment without oxygen, the increase in conductivity of the metal oxide sensor in the presence of reducing agents is due to the reaction of these agents with the metal oxide surface, resulting in formation of electron donor centers, which supply electrons to the conduction band of the metal oxide. These centers are being depleted under the influence of oxygen whereupon the sensor conductivity returns to its original level. In air the metal oxide surface is covered with chemisorbed oxygen, which captures electrons from the conduction band of the metal oxide. The studies in operando confirm the ionosorption model in which sensor response in air results from the concurrence of two processes − the reaction of the reducing gas with chemisorbed oxygen anions that removes these anions and return electrons to the conduction band, and the chemisorption of oxygen to form anions. 4. Sensors based on metal-oxide composites Binary semiconductor composites can be divided into two broad categories depending on the path of current flow in the sensor system. In the composites of the first type the current flows through the nanocrystals of one component, while the second component is a modifier that changes the chemical and electronic characteristics of conductive nanocrystals. In the second-type composites the current can flow through the various metal oxides which determine the conductivity of the sensor system. 4.1. Binary composites of type 1 The composites in this category consist of the primary metal oxide component and a modifier. Modifiers may be added within the nanoparticle volume, especially in the surface layers of the particles, forming a solid solution with the main oxide. As a result, the nanoparticles of the main oxide acquire new chemical properties [31,32], resulting in formation of new adsorption sites, which interact specifically with different molecules on the nanoparticle surface [32–34]. The coordinated unsaturated metal ions [31] in the oxygen vacancies of the metal oxide lattice are active centers of gas adsorption on the metal oxide surface. The role of such vacancies during the adsorption and chemical reactions of gases on the surface of the metal oxide sensors has been shown, in particular, for the conductometric detection of ethanol [35], hydrogen [36] and CO [37]. Fig. 1 illustrates the model of dissociative adsorption of H2 on oxygen vacancy of TiO2 [38]. The processes of adsorption and desorption of detected gas largely determine the temperature dependence of the sensor response. In the vast majority of semiconducting conductometric sensors, this dependence has the form of a curve with a maximum Smax at a certain temperature Tmax . The presence of temperature

Fig. 1. Model of dissociative adsorption of H2 -molecule on the oxygen vacancy at TiO2 (110) surface [38].

maximum of the sensitivity is due to the competition of two processes – the reaction of the detected gas adsorbed on the surface of the sensitive layer with oxygen centers and desorption of the detected gas [39]. Raising the temperature causes an increase in the rate constant of the sensory reaction, but also lowers the equilibrium concentration of the adsorbed gas by increasing its rate of desorption. Typical examples are sensors based on ZnO-In2 O3 nanofibers [40]. Doping of In2 O3 nanofibers by low concentrations of ZnO produces highly crystalline In2 O3 composite nanofibers, containing in the lattice about 2 at% dissolved Zn+2 ions. This leads to the formation of positively charged holes in the nanofibers and consequently, a sharp decrease in electronic conductivity. The sensor response to CO and H2 increases by approximately the same extent (8–10 folds). As a result of the zinc ions introduced in the In2 O3 nanofibers the sensitivity of the nanofibers to ethanol increases rather significantly (about 20 folds). The increase in sensory effect due to dissolution of Zn+2 ions in the In2 O3 lattice is due to a significant increase in the number of oxygen vacancies that are centers of chemisorption of the detected gases [40]. It can be assumed that the oxygen vacancies of In2 O3 lattice contacting with Zn+2 , which replaced the In+3 , are more favorable for the adsorption of polar molecules such as ethanol molecules. However, in the most cases a modifier forms a separate phase and typically as small nanoparticles or clusters, and included in the composite with the nanocrystals of the primary oxide. In such systems the best sensory characteristics are achieved not in the randomly packed mixture of nanoparticles but when small modifier particles are located on the surface of the primary nanocrystals [41]. The most suitable method for producing such composites is the impregnation of semiconductor crystals with solutions of the respective metal salts. Subsequent heat treatment of the impregnated crystals leads to the formation of small metal or metal oxide particles on the surface of the crystals [42,43]. Fig. 2 shows PdO clusters on the surface of SnO2 nanocrystals obtained by the impregnation method [44]. The influence of modifier nanoparticles on the conductivity and sensory properties of a composite depends on electron transfer between these nanoparticles and the conductive nanocrystals of the composite. Such a transfer is caused by difference of the potentials between the contacting particles of the semiconductor or metal with different electron work function. As such, the levels of potential energy of electrons (such as the conduction electrons and the valence electrons) are reduced in particles with a high work function and rise in particles with a low work function. Accord-

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

617

Fig. 2. TEM images of PdO nanoclusters deposited on the surface of SnO2 crystallites. Pd-concentration: (a) – 3 wt%, (b) – 5 wt% [44].

ingly, there is a redistribution of electrons and mutual charging of contacting nanoparticles. The charges induced by the contact are distributed on both sides of the contact plane to a depth of about the Debye length of charge screening LD = (εkT/q2 nc )1/2 , where ε is dielectric constant of the crystal, and q is electron charge [45]. The sensors consisting of metal oxide crystals, with the surface modified by clusters of noble metals and/or their oxides have been investigated in great detail. A study has considered the nanocrystalline SnO2 films containing semiconducting Ag2 O and PdO clusters the size range 5–10 nm, obtained by the impregnation method. These clusters cause a sharp increase in the resistance of nanocrystalline SnO2 film due to the transition of electrons from SnO2 to Ag2 O or PdO, having substantially higher the work function than SnO2 . The electron transfer under the influence of the contact potential at the boundary between PdO clusters and SnO2 nanoparticles was detected by XPS [44]. In the presence of H2 and CO, the clusters quickly reduce to the metal, such that the work function of the cluster is sharply reduced and the electron outflow from SnO2 to the clusters stops, producing a significant increase in sensory effect. Such sensitization of the sensory effect due to the change of electron transfer between the components of the composite has been defined as electronic sensitization [44]. The addition of Pt practically has no influence on nanocrystalline SnO2 film resistance [46]. This implies that in nanocomposites with Pt clusters located on the surface of SnO2 nanocrystals, electronic interactions between the clusters and SnO2 are virtually non-existent, since Pt does not form stable oxides in air [44]. The enhanced sensory effect under the influence of Pt clusters observed in the detection of hydrogen and methane, is due to purely chemical processes [44,47]. The Pt clusters catalyze the dissociation of the analyte molecules adsorbed on the surface of the clusters. This is followed by a spillover of hydrogen atoms and hydrocarbon radicals onto the SnO2 surface, where they react with the O− anion radicals. As a result, chemical sensitization of sensory effects occurs [44,47]. Similar processes appear to occur also on PdO clusters at high temperatures (400–500 ◦ C) [44] when the PdO begins to break down and acts as a metal catalyst. The sensory characteristics of nanocrystalline SnO2 films doped with palladium or platinum have been studied in detail, including determination of optimal compositions of the composites for detectionof CO and H2 in air ambience [48,49]. By varying the nature of small metal oxide clusters located on the surface of nanocrystals, which form the sensitive layer of the sensor, it is possible to promote chemical sensitization of the sensory effects for certain types of gases and, thereby, increase the detection selectivity of these gases. A typical example is the modification of nanostructured SnO2 film with La2 O3 clusters having basic nature. The chemisorption of ethanol on the surface of such clusters results in dehydrogenation of the ethanol and formation of acetone. Acetone molecules migrate towards the surface of the SnO2 nanocrystals, resulting in conductivity of the composite. The

acetone molecules react with O− anion-radicals, releasing trapped conduction electrons of SnO2 and increasing the sensor conductivity [44] that provides a high sensory response. In addition, acidic MnO2 clusters essentially reduce the sensitivity of SnO2 sensor to ethanol to zero. These clusters absorb ethanol and catalyze its dehydration to ethylene, which has low reactivity toward O− [44]. The sensitivity of the nanocrystalline In2 O3 sensor to CO and the selectivity of CO detection in the presence of hydrogen increase sharply after addition of small amounts of Rb2 O and Co3 O4 to the sensitive layer. In particular, addition of Co3 O4 strongly influences the sensory properties of In2 O3 . Thus, at 200 ◦ C the addition of 0.5 mol% of Co3 O4 to impregnated In2 O3 samples increases SCO by about 128-folds, and the ratio SCO /SH2 about 35-folds [50]. It should be noted that Rb2 O and Co3 O4 are p-semiconductors, such that p-n-contacts are formed at the boundaries of In2 O3 nanocrystal with the particles of the added oxides and there is a transfer of electrons from In2 O3 to these oxides. Perhaps these contacts act as centers of CO and H2 chemisorption, and the rate constant of chemisorption of the CO polar molecules, containing lone electron pairs on the p-n-centers with separated charges, is higher than for H2 molecules. 4.2. Binary composites of type 2 The binary metal oxide composites of the second type include nanocrystalline systems in which there can be different current paths, depending on which component determines the conductivity of the composite. Such randomly packed composites are typically produced by the screen-printing method [51], the most widely used for synthesizing composite sensor films from ground mixtures of metal oxide nanocrystals. The sensory properties of the nanocomposite also change in accordance with a change in the current paths. The conductivity of randomly packed composites, consisting of particles having different electronic characteristics is described by the percolation theory [52]. Detailed analysis of the conductivity on the basis of this theory is given in [16] for a nanocomposite consisting of a mixture of n- and p-type TiO2 nanocrystals. In this system, there are four types of electrical contacts between the crystals: n–n, p–n in the direction of field from the p- to n-crystal, n–p in reverse direction of the field in the contact region and p–p. The redistribution of the electrons and the resulting mutual charging of the contacting nanocrystals lead to a reduction in the barrier to electron transport in the p-n contacts, but this also increase the barrier for n–p contacts. The paths of current flow in the composites are aggregates of the contacting particles, provided that these aggregates permeate the entire space and represent endless clusters [52]. The current paths through the nanocrystalline aggregates with heterogeneous intercrystalline contacts inevitably include contacts with high resistance, which determine the total resistance of the system

618

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

Fig. 3. Temperature dependence of the resistance of SnO2 -In2 O3 nanostructured films. In2 O3 concentration (wt%): (ɑ)  – 0;  – 6; 䊉 – 12; (b) 䊏 – 25; 䊉 – 37;  – 50; (c) 䊏 – 100.

[16]. Therefore, current will bypass the heterogeneous contacts and flow through homogeneous aggregates, consisting of nanocrystals of one type (n-n or p–p). The formation of endless clusters from particles of one type (e.g., n- or p-) in the composite containing particles of another type (p- or n-) depends on the ratio between the particle size and concentration of particles of the type from which the cluster is formed. In the randomly packed composite consisting of n-TiO2 nanocrystals with 150 nm size and p-TiO2 crystals with 1 ␮m size, the percolation threshold of formation of n-TiO2 endless clusters with electronic conductivity is 5 wt% of n-TiO2 . Similar clusters of p-TiO2 crystals of 1 ␮m size, which have hole conductivity, are formed at p-TiO2 concentration (Xp ) above 75 wt% [16]. Thus, in the composite with Xp less than 75 wt% only electronic conductivity is realized, and at Xp larger than 95 wt% only hole conduction occurs. In the range of Xp from 75 to 95 wt% both hole and electron conductivity coexist, and the composite conductivity has mixed characteristics. The direction of sensory response in the detection of reducing gases similarly changes according to a change in the type of conductivity. The reduction in composite conductivity under the influence of CO and CH4 , which is characteristic of composites with hole conductivity at Xp > 95 wt%, is replaced by an increase in conductivity in the electronic type composite at Xp < 75 wt% [16]. In p-n nanocomposites it is necessary to consider the interaction between the components and the electron transfer from the n- to p-metal oxide. This is because the work function W for metal oxide crystals of n-type is less than W for metal oxide crystals of p-type [53]. Such a transfer plays a crucial role in the conductive and sensing behavior of composites. In n-SnO2 –p-Co3 O4 composite having electronic conductivity, when current flows through the n-SnO2 nanocrystals, the electron transfer from n-SnO2 to p-Co3 O4 through heterogeneous contacts leads to a reduction in the concentration of conduction electrons in the current paths and a sharp increase in the composite resistance with an increase in p-Co3 O4 concentration (XCo ) up to 10 wt% [2]. With further increase in XCo percolation transition occurs from electronic to hole conductivity through pCo3 O4 nanocrystals, and the composite resistance begins to drop sharply [2]. The direction of sensor response in CO and H2 detection changes [2] in accordance with the variation of conductivity type as well as in the system [16]. Specifically, these gases cause a decrease in resistance of the composite with electronic conductivity, and an increase in the composite having p-type conductivity [2]. The peculiarities of conductivity and sensory effects associated with the interaction between the nanocrystals having different work functions are also typical for composites consisting of similar metal oxide semiconductors of the n-type. An example of such composites are SnO2 -In2 O3 nanocrystalline films [54]. The temperature dependence of conductivity depends on which component forms the current path in the composite. The resistance of these films at

In2 O3 concentrations (XIn ) less than 20%, as well as the resistance of SnO2 nanocrystalline films, decreases with increasing temperature in the range from 200 to 550 ◦ C (Fig. 3a). Thus, the current in the composite films of this composition flows through the aggregates of SnO2 nanocrystals. Since the work function from SnO2 crystal is 4.9 eV [55], and from In2 O3 crystals is 4.3 eV [56], additives of In2 O3 nanocrystals increase the conductivity of the films due to the electron transfer from In2 O3 to SnO2 particles [54,57]. Characteristically, the increase of conductivity in this case is accompanied by an increase in sensor response to CO and H2 . This behavior can be attributed to the electronic sensitization of the sensor effect due to increase in the number of electrons which pass from the In2 O3 additives to the conducting nanocrystalline SnO2 clusters under the influence of CO and H2 . Such influence of these reducing gases on the transfer of electrons in SnO2 -In2 O3 heterocontacts has been confirmed by data on the conductivity of the sensors. It was shown that an increase in XIn from 0 to 12% in pure air increases the conductivity of the composite only by 15–20%, but in air containing 0.5% CO, it is 250% [57]. Temperature dependence of conductivity changes principally in composites with XIn larger than 20 wt%. An unusual increase in resistance with increasing temperature is observed in these composites in the range 300–400 ◦ C (Fig. 3b). This effect is characteristic of nanocrystalline In2 O3 (Fig. 3c) and, is most probably, caused by the change in the nature of chemisorbed oxygen centers on the surface of the In2 O3 nanocrystals [54,58]. Experimental data show that increasing XIn above 20 wt% results in a percolation transition from conductivity through SnO2 -nanocrystals to conductivity through In2 O3 -nanocrystals. Percolation transition in conductivity is accompanied by a change in the sensor behavior of the composite. Indeed, there is a maximum in the area of percolation transition at XIn ≈ 20 wt% on the curve of dependence of maximal sensory effect of SnO2 In2 O3 system on composition (Fig. 4, curve 1) [59]. When current passes through In2 O3 nanocrystals (at XIn above 20 wt%) the electron transfer from In2 O3 to SnO2 results in a decrease in the sensor response of the composite to reducing gases. The degree of this electron transfer decreases with diminishing SnO2 concentration and the corresponding increase of XIn . This explains the slow rise of the response to CO and H2 with XIn for XIn larger than 50 wt%, when the effect of the electron transfer from In2 O3 to SnO2 on the sensor response begins to gradually decrease due to reduction in the number of In2 O3 –SnO2 intercrystallite contacts. Correspondingly, the sensor response approaches that of nanocrystalline In2 O3 [57]. It should be noted that the effect of the electron transfer from In2 O3 to SnO2 on the sensor response of composites with conductivity through In2 O3 -nanocrystals (XIn > 20 wt%) is much less than that of the composites with conductivity through SnO2 -nanocrystals (XIn < 20 wt%). This is because the concentration of electrons in In2 O3 is about 2–3 orders of magnitude greater than in SnO2 .

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

Fig. 4. Maximal response of composite films to 2% H2 : 1-SnO2 -In2 O3 and 2–In2 O3 ZnO at 450 ◦ C [59].

The increase in conductivity and sensor response due to small additives of In2 O3 nanocrystals (up to 20 wt%) also occurs in ZnOIn2 O3 nanocomposites [59] for which conduction occurs through the ZnO nanocrystals. These effects are due to electron transfer from In2 O3 to ZnO [60]. Unlike the SnO2 -In2 O3 composite, two maxima are observed in the ZnO-In2 O3 system on the curve of maximal sensor response as a function of composite composition (Fig. 4, curve 2) [59]. The first maximum, just as for SnO2 -In2 O3 , is due to the appearance of a percolation cluster from the In2 O3 particles, which are rich in electrons. With further increase in XIn up to 80 wt% there are two types of percolation clusters in the system: one type consists of In2 O3 nanoparticles and the second, ZnO. Herewith the second maximum at XIn = 80 wt% appears on the curve of sensor response dependence on In2 O3 composition in the composite. At XIn > 80 wt% the percolation clusters of ZnO nanoparticles disappear. Meanwhile, in the nanocomposites with conduction through In2 O3 nanocrystals at high In2 O3 concentration (over 80 wt%), additives of ZnO nanocrystals produce a sharp decrease in conductivity of the composite, accompanied by an increase of sensor response to H2 [59]. It is assumed that the ZnO nanocrystals containing active oxygen vacancies [38] catalyze the dissociation of adsorbed oxygen on O atoms. These atoms diffuse to In2 O3 nanocrystals and capture the conduction electrons of the In2 O3 [58]. The ZnO nanocrystals also catalyze the dissociation of adsorbed H2 [38]. Therefore these nanoctystals are chemical sensitizers of sensory response in ZnOIn2 O3 composite. The effect of ZnO on sensor response in ZnO-In2 O3 nanocomposite can be explained by the spillover of H atoms from ZnO to In2 O3 nanocrystals where they react with O− , returning electrons to In2 O3 conduction band [59]. Similar processes of dissociation of hydrogen molecules at the catalytically active particles and transfer of the atoms produced to metal oxide nanocrystals have been noted in many studies (see for example [61]). The combination of electron-donor semiconductors with high electron concentration and electron-acceptor semiconductors containing chemically active sites allows both electronic and chemical sensitization of sensory effects in these composites. This in turn provides opportunities to increase the efficiency of the sensor. The role of contacts between nanocrystals of the metal oxides with different electronic structure was also observed in the detection of ethanol by nanostructured ZnO-In2 O3 composite containing a layer of In2 O3 nanofibers deposited on a layer of ZnO nanofibers [62]. As a result of the formation of contacts between ZnO and In2 O3 nanocrystals, the sensitivity to ethanol of such two-layer ZnOIn2 O3 films was found to be 5–10 times higher than the sensitivity of one-component films from ZnO nanofibers, prepared by the same method. In addition, the maximum sensitivity of the two-layer film

619

compared to a similar ZnO film was shifted by 100 ◦ towards lower temperatures, specifically from 300 to 200 ◦ C [38]. An increase in sensor response to ethanol was also observed at contacts of In2 O3 and Fe2 O3 nanocrystalline layers [63]. In the two-layer ZnO-In2 O3 composite with parallel connection of the layered resistors the conductivity is determined by the conductivity of the In2 O3 layer, which is increased as a result of the reaction of C2 H5 OH with O− on the surface of the In2 O3 nanocrystals. It can be assumed that this reaction is accelerated because ZnO nanocrystals catalyze C2 H5 OH decomposition, leading to formation of active intermediate species [64]. Because of contacts between ZnO and In2 O3 nanocrystals these intermediates pass through the In2 O3 nanocrystals, where they react with O− . In addition, the transfer of electrons from In2 O3 to ZnO should facilitate chemisorption of polar molecules with lone electron pairs, such as ethanol. The same factors are likely to cause a high sensory response to trimethylamine and ethanol of composite nanofibers, consisting of In2 O3 and ZnO nanocrystals contacting with one another [65]. The sensitivity of such a sensor to these compounds is much more than the sensitivity of one-component sensors consisting of nanocrystalline ZnO or In2 O3 fibers. At certain concentration ratios between the components of the sensor there is a significant increase in the selectivity in the detection of trimethylamine. The causes of this effect have not yet been clarified. It should be noted that in a complex system of intersecting composite fibers composed of nanocrystals with different electronic characteristics, there is a lot of nanocrystal aggregates with different sensor responses to detected gases. The concentration ratio between the various aggregates and, correspondingly, the sensory characteristics of the system depend on the composition of the composite fibers. Numerous studies have considered the binary nanocrystal systems containing CeO2 [66–70]. Ceria and CeO2 -containing materials have generated considerable interest and have come under intense scrutiny as catalysts and as structural and electronic promoters of heterogeneous catalytic reactions. The catalytic properties of ceria nanocrystals are due to the relative lability of the lattice oxygen on the nanocrystal surface that leads to formation of a large number of oxygen vacancies on the surface [71]. The sensory characteristics of binary composites containing ceria depend on the structure and properties of the second component of the composite, and the nature of the detected gas. In the nanocrystalline composites synthesized by co-precipitation of CeO2 and SnO2 from a solution at the addition of 10 wt% CeO2 , the sensor response to polar organic compounds (butanone, ethanol, acetone) significantly increases [68,69], but the response to CO and CH4 [67], as well as aromatic hydrocarbons and H2 [66] is reduced. By the same token, the addition of ceria to ZnO based nanocomposites increases response to aromatics [72]. Nanotubes from In2 O3 and CeO2 nanocrystals, exhibit much higher response to acetone than individual nanocrystalline oxides [73]. The CeO2 -In2 O3 nanocomposite system has also been investigated [74,75]. Introduction of small amounts of CeO2 to nanocrystalline In2 O3 film leads to a significant increase in sensor response to H2 and CO, reaching a maximum at 3–5 wt% CeO2 . It is remarkable that the increase in response is accompanied by only a slight increase in the resistance of the composite sensor. This indicates that CeO2 is not dissolved in In2 O3 and the observed increase in sensor response is the result of chemical sensitization of the sensory process by CeO2 nanoparticles in contact with the In2 O3 nanocrystals that constitute the path of current flow in the composite. Sensitization of the CeO2 -In2 O3 composite to hydrogen is associated with the promotion of H2 dissociation by CeO2 nanoparticles having high concentration of oxygen vacancies. The creation of vacancies is accompanied by the formation of chemically active Ce+3 ions, localized in these vacancies [76].

620

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

Fig. 5. Proposed scheme of the sensory process in In2 O3 –based nanocomposite with deposited CeO2 clusters; 1–oxygen vacancy; (a) – H2 detection; (b) – CO detection.

The CeO2 vacancies containing Ce+3 ions are centers of dissociative chemisorption of H2 molecules with the formation of H atoms, the part of which are relatively weakly connected with the surface of CeO2 clusters [77,78] and at 300–500 ◦ C can be passed onto the surface of In2 O3 crystal. The reaction of hydrogen atoms with O− anions adsorbed on the surface of In2 O3 occurs, most likely, at the interface between the CeO2 nanoparticles and the substrate of In2 O3 nanocrystals. This reaction improves sensor response of CeO2 -In2 O3 sensor to H2 compared with the response of In2 O3 alone [74]. The scheme of this process is illustrated in Fig. 5a. The addition of CeO2 to In2 O3 film also improves the sensory effect in the detection of CO [74]. Chemisorption of CO in this composite occurs mainly on the CeO2 nanoparticles, as they have a much higher concentration of oxygen vacancies than In2 O3 nanocrystals. At the interface between the In2 O3 and CeO2 clusters there are particularly favorable conditions for the reaction of CO with chemisorbed anions O− located on the surface of the In2 O3 crystals near this boundary. In the complex between the CO and O− anion-radical, formed in the oxygen vacancy of CeO2 , the reactivity of CO in the reaction with O− is enhanced, since the bond between the CO and the vacancy leads to the formation of unpaired electron in CO molecule (Fig. 5b). The highest sensitivity has been observed for the composite film with 3–10 wt% CeO2 . A further enrichment of the composite with ceria significantly decreases the sensitivity. It was shown by XPS method that increasing XCe from 3 to 40 wt% leads to a sharp reduction in the concentration of Ce+3 ions in ceria nanoparticles [74]. This implies that the number of H2 and CO chemisorption centers in these nanoparticles decreases with increasing XCe most likely due to the increase in nanoparticle size [79]. At CeO2 concentrations of more than 40 wt% the resistance of the composite film increases dramatically due to the percolation transition. There is a change of current flow paths − from one through In2 O3 nanocrystals with a high concentration of conduction electrons to one through CeO2 nanocrystals with a low concentration of conduction electrons. Accordingly, the composite sensor sensitivity decreases, approaching the sensitivity of nanocrystalline CeO2 .

5. Sensor properties of core-shell type nanofibers The attention of specialists in the field of semiconducting sensors has recently been attracted to composite nanofibers of core-shell type, in which the nanofiber of one oxide is surrounded by a layer of another oxide (see, e.g., SnO2 -ZnO nanofiber, Fig. 6). A variety of studies on nanoheterogeneous core-shell sensors including SnO2 -ZnO [80,81], ZnO-SnO2 [82,83], In2 O3 -ZnO [84] and ␣-Fe2 O3 -ZnO [85–87] (here the first oxide is the core, and the second oxide is the shell) have exhibited sensitivity in the detection of reducing gases that is much greater than the sensitivity of nanofiber sensors based on individual metal oxides,

Fig. 6. Schematic image of sensor based on SnO2 -ZnO core-shell nanofibers.

Fig. 7. Dependence of SnO2 -ZnO core-shell nanofiber sensor resistance on shell thickness.

The effect of shell thickness on the resistance of SnO2 -ZnO nanofibers and their sensitivity in the detection of CO and aromatic hydrocarbons has been studied [81]. ZnO was deposited on monocrystalline SnO2 fiber by atomic layer deposition method, such that the resulting shell fully shielded the internal nanofiber. First the resistance and sensitivity of such nanoheterogeneous systems increase with increasing thickness of the shell to certain maximum values, and then decrease (Figs. 7 and 8). The maximum values of resistance and sensitivity are achieved when the thickness of the shell is close to the Debye length of charge screening (LD )ZnO in ZnO that is estimated to be 22–35 nm [81]. System resistance is determined by contacts between the composite nanofibers, and increases with a decrease in the concentration of electrons in the ZnO shell [39].

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

621

Fig. 8. Sensor responses of SnO2 -ZnO core-shell nanofibers to 10 ppm of different gases. (ɑ) – CO, (b) – C6 H6 and (c) – C7 H8 .

In SnO2 -ZnO composite fiber the conduction electron concentration in the ZnO shell is reduced, compared with the concentration of electrons in pure ZnO due to two factors. The first factor is the transition of conduction electrons from the shell to the core, because the work function of ZnO is less than that of SnO2 . Furthermore, there occurs dissociation of adsorbed oxygen molecules on the surface of ZnO shell in the temperature range 200–400 ◦ C, and the oxygen atoms produced capture the electrons from the ZnO conduction band. The combined effect of these factors is that in the shell with a thickness dZnO , which is substantially less than (LD )ZnO , the density of conduction electrons in ZnO is close to zero [81]. In this case there comes into play the reverse process of electron transfer from SnO2 to ZnO, specifically, thin ZnO shell can be filled with electrons from SnO2 core, just as happens with thin dielectric layers on metal or semiconductor surface [45,81]. Electron transfer in heterocontacts, including core-shell heterocontacts in composite nanofibers, proceeds with the participation of surface states formed due to the difference in lattice structures of contacting crystals, in particular, crystalline core and shell [60]. With increasing thickness of the ZnO shell in the range of 0 to (LD )ZnO the influence of electron transfer from SnO2 to ZnO on the resistance of the SnO2 -ZnO core-shell nanowires decreases rapidly. Correspondingly, the resistance of SnO2 -ZnO nanofibers is increased up to a maximum value at dZnO of about (LD )ZnO . With further increase in dZnO the effect of electron outflow from the ZnO shell to the SnO2 core is reduced. This leads to an increase in the concentration of electrons at the surface of the shell and, correspondingly, a reduction in resistance of the ZnO-SnO2 nanofibers. When dZnO is significantly greater than (LD )ZnO , the resistance approaches the value of the resistance of pure ZnO nanofibers [81]. The reasons for the observed dependence of sensitivity of the SnO2 -ZnO nanofiber composite system on the shell thickness have not been completely clarified. A possible explanation of this relationship has been provided in a model [4,12]. According to this model, the sensitivity and resistance of the nanostructured sensor are increased with decrease in concentration of electrons in the volume of the nanoparticles. In the system of core-shell nanofibers the nc value is the concentration of electrons in the shell, and a decrease in nc under the influence of electron-acceptor core according to the model, must increase the sensitivity of the SnO2 -ZnO system to the

reducing gases. It should be noted that the application of SnO2 film for detection of hydrogen resulted in an increase in sensitivity of the nanostructured film sensor with increase in resistance [88]. A similar result has been obtained with the modification of ␣MoO3 nanofibers by iron oxide. For this process small particles of Fe2 (MoO4 )3 were formed on the nanofiber surface [89]. The sensory response to 500 ppm ethanol at 220 ◦ C was observed to be six times larger than the response of the pure ␣-MoO3 , due to the high catalytic activity of the Fe2 (MoO4 )3 mixed oxide. It is noteworthy that in the ZnO-SnO2 nanowire system where the ZnO core can be considered as an electron donor for the SnO2 shell, the sensitivity to reducing gases CO and H2 is quite low at the temperature range of 200–300 ◦ C. Meanwhile, this system at the same temperature range exhibits high sensitivity to NO2 oxidant that is much higher than the sensitivity of individual SnO2 and ZnO nanofibers [80]. As temperature increases to 400 ◦ C, the sensory characteristics of ZnO-SnO2 nanofibers are changed radically. In particular, the sensitivity to NO2 sharply reduces and there is apparently a significant sensitivity to reducing compounds, especially ethanol, which is 7 times larger than the sensitivity of ZnO nanofibers to ethanol [82]. It should be noted that ZnO addition to SnO2 results in an increase in sensor sensitivity of the net-like SnO2 /ZnO hetero-nanostructures in the detection of ultra-small H2 S concentrations [90]. The results of the above studies collectively demonstrate that changing the nature of the components and their relative position in core-shell type composite nanofibers, can greatly affect the sensitivity and selectivity of the sensor to various chemical compounds. Under certain conditions, these sensor systems are very sensitive to traces of organic compounds in the air (see, e.g., [80,81]). This characteristic makes such nanofibers very promising in the development of new and improved sensors for detection of toxic and explosive gases.

6. Conclusion This review examined various models of the sensory phenomena occurring in the application of nanostructured semiconducting metal oxides for detection of reducing gases. The simultaneous investigations of spectral and electrical characteristics of metal

622

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

oxide sensors in the gas detection (an operando) show that the ionosorption model is valid in air in real conditions of sensors exploitation. According to this model the sensory effect is due to the reaction of reducing gases with O− anion-radicals produced during the dissociative chemisorption of oxygen adsorbed on the surface of the metal oxide sensor particles, which then capture the electrons from the conduction band of the sensor. This reaction leads to a release of the trapped electrons and their return to the conduction band. This reduces the resistance of the sensor, which is the physical manifestation of the sensory effect. The conductivity and sensory characteristics of nanostructured sensors depend on the electronic structure of the nanoparticles. In nanostructured metal oxide composites the electronic structure of the nanoparticles is changed as a result of interaction between the components of the composite having different electronic and physico-chemical characteristics. This interaction largely determines the behavior of metal oxide nanocomposite sensors and allows significant change in their operating properties. The modification of nanoparticles of initial metal oxide components during the formation of the composite is due mainly to two factors: a) the dissolution of ions of one metal oxide component in the lattice of the other oxide and b) the transfer of charges between the contacting nanocrystals having different electron work function. The redistribution of ions in the composite results in changes in the electronic structure and lattice of the nanoparticles, in particular, the formation of new centers of adsorption of the detected gas, which in turn increases sensor response and selectivity. The extent of this modification depends on the nanoparticle type and metal oxide structure, and the method used to produce the composite. The second factor responsible for modification of the nanoparticles is the redistribution of electrons between the contacting nanoparticles of different work functions. As a result, mutual charging these nanoparticles is observed in these systems. This creates conditions favoring the adsorption of polar molecules with electron-donor and electron-acceptor groups. The difference between the values of the work function of the contacting metal oxide crystals and, correspondingly, the degree of electron transfer between these crystals are changed on exposure of the sensitive sensor layer to a reducing gas. Electronic sensitization of the sensory response occurs if redistribution of electrons caused by the reaction leads to an influx of conduction electrons in the nanocrystals, which form a path of current flow in the composite. The mechanism of chemical sensitization of sensory response can be realized in the nanocomposites as well. A characteristic example is a composite of nanocrystalline metal oxide semiconductor with chemically-active metal or metal oxide nanoclusters, deposited on the surface of semiconductor nanocrystals. The deposited nanoclusters contain a high concentration of the centers to chemisorption of the detected gases. The molecules chemisorbed on the nanoclusters (or active intermediates formed in reactions on the surface of nanoclusters) diffuse into the semiconductor nanocrystals, through which current flows in the composite, and react with oxygen anions at the surface of the nanocrystals (spillover effect) thereby increasing the conductivity. Thus the nanocluster additives enhance the sensor response. Binary metal oxide nanocomposites are particularly promising for the detection of complex organic compounds with electrondonor and electron-acceptor groups. Sensor response in the detection of this type of compounds is improved due to the presence in the composite sensor of acidic and basic centers, which favor adsorption of such molecules onto the sensor surface. An example is the combination of positively and negatively charged nanoparticles in contact with each other, obtained as a result of the electron transfer between the metal oxides. Optimal conditions for the detection are achieved in a composite sensor in which the centers, providing increased adsorption capacity of the sensor,

are combined with the reactive centers. These in turn activate the adsorbed molecules in the reaction with oxygen anions at the surface of nanocrystals. Such centers, for example, are active oxygen vacancies and the ions in the intermediate depend on the extent of oxidation in the CeO2 nanoparticles. Of particular interest are the new composite systems of the core-shell type nanofibers, whose sensory characteristics depend on the transfer of electrons between the core and the shell in the nanofiber. The sensitivity and selectivity of this sensor system to a variety of chemical compounds depend on the nature and the mutual arrangement of the metal oxide components. In closure, binary nanostructured metal oxides provide is an effective means to improve the efficiency and selectivity of semiconductor sensors. They enable development of sensor systems to meet desired complexity of operational properties. Acknowledgement This study is supported by the Russian Scientific Foundation under grant number 14-19-00781. References [1] W.J. Moon, J.H. Yu, G.M. Choi, The CO and H2 gas selectivity of CuO-doped SnO2 -ZnO composite gas sensor, Sens. Actuators B 87 (2002) 464–470. [2] U.S. Choi, G. Sakai, K. Shimanoe, N. Yamazoe, Sensing properties of SnO2 -Co3 O4 composites to CO and H2 , Sens. Actuators B 98 (2004) 166–173. [3] D.F. Cox, T.B. Fryberger, S. Semancik, Oxygen vacancies and defect electronic states on the SnO 2 (110)-1 × 1 surface, Phys. Rev. B 38 (1988) 2072–2078. [4] N. Yamazoe, K. Shimanoe, Receptor function and response of semiconductor gas sensor, J. Sens. 138 (2009) 1–21. [5] M.A. Kozhushner, B.V. Lidskii, I.I. Oleynik, V.S. Posvyanskii, L.I. Trakhtenberg, Inhomogeneous charge distribution in semiconductor nanoparticles, J. Phys. Chem. C 119 (2015) 16286–16292. [6] H. Ogawa, M. Nishikawa, A. Abe, Hall measurement studies and an electrical conduction model of tin oxide ultrafine particle films, J. Appl. Phys. 53 (1982) 4448–4454. [7] S. Rembeza, E. Rembeza, T. Svistova, Electrophysical properties of gas sensitive films SnO2 doped with palladium, Sens. Transducers Mag. 40 (2004) 145–151. [8] M.A. Kozhushner, L.I. Trakhtenberg, A.C. Landerville, I.I. Oleynik, Theory of sensing response of nanostructured tin-dioxide thin films to reducing hydrogen gas, J. Phys. Chem. C 117 (2013) 11562–11568. [9] L.C. Jimenez, H.A. Mendez, B.A. Paez, M.E. Ramirez, H. Rodriguez, Production and characterization of indium oxide and indium nitride, Braz. J. Phys. 36 (2006) 1017–1024. [10] P. Prathap, G.G. Devi, Y.P.V. Subbaiah, K.T.R. Reddy, V. Ganesan, Growth and characterization of indium oxide films, Curr. Appl. Phys. 8 (2008) 120–127. [11] N. Barsan, U. Weimar, Conduction model of metal oxide gas sensors, J. Electroceram. 7 (2001) 143–167. [12] N. Yamazoe, K. Shimanoe, Theory of power laws for semiconductor gas sensors, Sens. Actuators B 128 (2008) 566–573. [13] C. Malagu, V. Guidi, M. Stefancich, M.C. Carotta, G. Martinelli, Model for Schottky barrier and surface states in nanostructured n-type semiconductors, J. Appl. Phys. 91 (2002) 808–814. [14] N.P. Zaretskiy, L.I. Menshikov, A.A. Vasiliev, On the origin of sensing properties of the nanostructured layers of semiconducting metal oxide materials, Sens. Actuators B 170 (2012) 148–157. [15] M.A. Kozhushner, L.I. Trakhtenberg, V.L. Bodneva, T.V. Belysheva, A.C. Landerville, I.I. Oleynik, Effect of temperature and nanoparticle size on sensor properties of nanostructured tin dioxide films, J. Phys. Chem. C 118 (2014) 11440–11444. [16] N. Savage, B. Chwieroth, A. Ginwalla, B.R. Patton, Sh A. Akbar, P.K. Dutta, Composite n-p-semiconducting titanium oxide as gas sensors, Sens. Actuators B 79 (2001) 17–27. [17] N. Yamazoe, K. Shimanoe, Explicit formulation for the response of neat oxide semiconductor gas sensor to reducing gas, Sens. Actuators B 158 (2011) 28–34. [18] S.R. Morrison, Surface barrier effects in adsorption, illustrated by zinc oxide, Adv. Catal. 7 (1955) 259–301. [19] V.F. Gromov, G.N. Gerasimov, T.V. Belysheva, L.I. Trakhtenberg, Sensing mechanisms at reductive gases detecting by conductometric sensors based on SnO2 , Russ. Chem. J. 52 (2008) 80–87. [20] B. Slater, C.R.A. Catlow, D.E. Williams, A.M. Stoneham, Dissociation of O2 on the reduced SnO2 (110) surface, Chem. Commun. 123 (2000) 5–1236. [21] A. Gurlo, Interplay between O2 and SnO2 : Oxygen ionosorption and spectroscopic evidence for adsorbed oxygen, Eur. J. Chem. Phys. Phys. Chem. 7 (2006) 2041–2052. [22] J.N. Zemel, Theoretical description of gas-film interaction on SnOx , Thin Solid Films 163 (1988) 189–202.

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624 [23] A. Gurlo, R. Riedel, In situ and operando spectroscopy for assessing mechanisms of gas sensing, Angew. Chem. Int. Ed. 46 (2007) 3826–3848. [24] O.V. Safonova, T. Neisius, A. Ryzhikov, B. Chenevier, A.M. Gaskov, M. Labeaub, Characterization of the H2 sensing mechanism of Pd-promoted SnO2 by XAS in operando conditions, Chem. Commun. (2005) 5202–5204. [25] D. Koziej, K. Thomas, N. Barsan, F. Thibault-Starzyk, U. Weimar, Influence of annealing temperature on the CO sensing mechanism for tin dioxide based sensors – Operando studies, Catal. Today 126 (2007) 211–218. [26] D. Degler, S. Wicker, U. Weimar, N. Barsan, Identifying the active oxygen species in SnO2 based gas sensing materials: an operando IR spectrsocopy study, J. Phys. Chem. C 119 (2015) 11792–11799. [27] K. Grossmann, R.G. Pavelko, N. Barsan, U. Weimar, Interplay of H2 water vapor and oxygen at the surface of SnO2 based gas sensors −An operando investigation utilizing deuterated gases, Sens. Actuators B 166–167 (2012) 787–793. [28] Y. Chen, X. Wang, Ch. Shi, L. Li, H. Qin, J. Hu, Sensing mechanism of SnO2 (110) surface to H2 : Density functional theory calculations, Sens. Actuators B 220 (2015) 279–287. [29] N. Bârsan, M. Hübner, U. Weimar, Conduction mechanisms in SnO2 based poly-crystalline thick film gas sensors exposed to CO and H2 in different oxygen backgrounds, Sens. Actuators B 157 (2011) 510–517. [30] M. Hübner, R.G. Pavelko, N. Barsan, U. Weimar, Influence of oxygen backgrounds on hydrogen sensing with SnO2 nanomaterials, Sens. Actuators B 154 (2011) 264–269. [31] V.V. Krivetsky, M.N. Rumyantseva, A.M. Gas’kov, Chemical modification of nanocrystalline tin dioxide for selective gas sensors, Russ. Chem. Rev. 82 (2013) 917–941. [32] A.V. Marikutsa, M.N. Rumyantseva, E.A. Konstantinova, T.B. Shatalova, A.M. Gaskov, Active sites on nanocrystalline tin dioxide surface: effect of palladium and ruthenium oxides clusters, J. Phys. Chem. C 118 (2014) 21541–21549. [33] M.N. Rumyantseva, A.M. Gas‘kov, Chemical modification of nanocrystalline metal oxides: influence of real structure and surface chemistry on sensing properties, Russ. Chem. Bull. 57 (2008) 1094–1115. [34] M. Rumyantseva, V. Kovalenko, A. Gaskov, E. Makshina, V. Yuschenko, I. Ivanova, A. Ponzoni, G. Faglia, E. Comini, Nanocomposites SnO2 /Fe2 O3 : sensor and catalytic properties, Sens. Actuators B 118 (2006) 208–214. [35] G. Kwak, K. Yong, Adsorption and reaction of ethanol on ZnO, J. Phys. Chem. C 112 (2008) 3041–3046. [36] M. Ahmad, R. Din, C. Pan, J. Zhu, Investigation of hydrogen storage capabilities of ZnO-based nanostructures, J. Phys. Chem. C 114 (2010) 2560–2565. [37] A. Linsebigler, G. Lu, T. Yates, CO chemisorption on TiO2 (110): Oxygen vacancy site influence on CO adsorption, J. Phys. Chem. 103 (1995) 9438–9445. [38] W. Göpel, G. Rocker, R. Feierabend, Intrinsic defects of TiO2 : interaction with chemisorbed O2 , H2 , CO and CO2 , Phys. Rev. B 28 (1983) 3427–3438. [39] S. Ahlers, G. Muller, T. Doll, A rate equation approach to the gas sensitivity of thin film metal oxide materials, Sens. Actuators B 107 (2005) 587–599. [40] N. Singh, A. Ponzoni, E. Comini, P.S. Lee, Chemical sensing investigations on Zn-In2 O3 nanowires, Sens. Actuators B 171–172 (2012) 244–248. [41] S.D. Bakrania, M.S. Wooldridge, The effects of the location of Au additives on combustion-generated SnO2 nanopowders for CO gas sensing, Sensors 10 (2010) 7002–7017. [42] G.N. Gerasimov, M.I. Ikim, P.S. Timashev, V.F. Gromov, T.V. Belysheva, E.Y. Spiridonova, V.N. Bagratashvili, L.I. Trakhtenberg, Small CeO2 clusters on the surface of semiconductor particles, Russ. J. Phys. Chem. A 89 (2015) 1002–1007. [43] M.M. Natile, A. Glisenti, Nanostructured oxide-based powders: investigation of the growth mode of the CeO2 clusters on the YSZ surface, J. Phys. Chem. B 110 (2006) 2515–2521. [44] N. Yamazoe, New approaches for improving semiconductor gas sensors, Sens. Actuators B 5 (1991) 7–19. [45] V.L. Bonch-Bruevich, S.G. Kalashnikov, Physics of Semiconductors, Science, Moscow, 1977, 672 pp. [46] D.-D. Lee, W.-Y. Chung, Gas-sensing characteristics of SnO2-x thin film with added Pt fabricated by the dipping method, Sens. Actuators B 20 (1989) 301–305. [47] Sh. Matsushima, J. Teraoka, N. Miura, N. Yamazoe, Electronic interaction between metal – additives and tin dioxide-based sensors, Jpn. J. Appl. Phys. 27 (1988) 1798–1802. [48] V.V. Malyshev, A.V. Pislyakov, Investigation of gas-sensitivity of sensor structures to hydrogen in a wide range of temperature, concentration and humidity of gas medium, Sens. Actuators B 134 (2008) 913–921. [49] V.V. Malyshev, A.V. Pislyakov, Investigation of gas-sensitivity of sensor structures to carbon monoxide in a wide range of temperature, concentration and humidity of gas medium, Sens. Actuators B 123 (2007) 71–81. [50] H. Yamaura, K. Moriya, N. Miura, N. Yamazoe, Mechanism of sensitivity promotion in CO sensor using indium oxide and cobalt oxide, Sens. Actuators B 65 (2000) 39–41. [51] G.F. Fine, L.M. Cavanagh, A. Afonja, R. Binions, Metal oxide semi-conductor gas sensors in environmental monitoring, Sensors 10 (2010) 5469–5502. [52] A.L. Efros, Physics and Geometry of Disorder: Percolation Theory, Mir, Moscow, 1987. [53] H. Gleiter, J. Weissmuller, O. Wollerscheim, R. Wurschum, Nanocrystalline materials: a way to solids with tunable electronic structures and properties, Acta Mater. 49 (2001) 737–745.

623

[54] T.V. Belysheva, G.N. Gerasimov, V.F. Gromov, E.Yu. Spiridonova, L.I. Trakhtenberg, Conductivity of SnO2 -In2 O3 nanocrystaline composite films, Russ. J. Phys. Chem. A 84 (2010) 1554–1559. [55] M.N. Islam, M.O. Hakim, Electron affinity and work function of polycrystalline SnO2 thin film, J. Mat. Sci. Lett. 5 (1986) 63–65. [56] O. Lang, C. Pettenkofer, J.F. Sa´ınchez-Royo, A. Segura, A. Klein, W. Jaegermann, Thin film growth and band lineup of In2 O3 on the layered semiconductor InSe, J. Appl. Phys. 86 (1999) 5687–5693. [57] L.I. Trakhtenberg, G.N. Gerasimov, V.F. Gromov, T.V. Belysheva, O.J. Ilegbusy, Effect of composition on sensing properties of SnO2 -In2 O3 mixed nanostructured films, Sens. Actuators В 169 (2012) 32–38. [58] G. Korotcenkov, V. Brinzari, A. Cerneavschi, M. Ivanova, A. Cornet, J. Morante, A. Cabot, J. Arbiol, In2 O3 films deposited by spray pyrolysis: gas response to reducing (CO H2 ) gases, Sens. Actuators B 98 (2004) 122–129. [59] L.I. Trakhtenberg, G.N. Gerasimov, V.F. Gromov, T.V. Belysheva, O.J. Ilegbusy, Conductivity and sensing properties of In2 O3 + ZnO mixed nanostructured films: effect of composition and temperature, Sens. Actuators B 187 (2013) 514–521. [60] S. Park, H. Ko, S. Kim, C. Lee, Role of the interfaces in multiple networked one-dimensional core-shell nanostructured gas sensors, ACS Appl. Mater. Interfaces 6 (2014) 9595–9600. [61] D.A. Panayotov, S.P. Burrows, J.T. Yates, J.R. Morris, Mechanistic studies of hydrogen dissociation and spillover on Au/TiO2 : IR spectroscopy of coadsorbed CO and H-donated electrons, J. Phys. Chem. C 115 (2011) 22400–22408. [62] X.-J. Zhang, G.-J. Qiao, High performance ethanol sensing films fabricated from ZnO and In2 O3 nanofibers with a double-layer structure, Appl. Surf. Sci. 258 (2012) 6643–6649. [63] M. Ivanovskaya, D. Kotsikau, G. Faglia, P. Nelli, Influence of chemical composition and structural factors of Fe2 O3 /In2 O3 sensors on their selectivity and sensitivity to ethanol, Sens. Actuators B 96 (2003) 498–503. [64] K.-W. Kim, P.-S. Cho, S.-J. Kim, J.-H. Lee, C.-Y. Kang, J.-S. Kim, S.-J. Yoon, The selective detection of C2 H5 OH using SnO2 -ZnO thin film gas sensors prepared by combinatorial solution deposition, Sens. Actuators B 123 (2007) 318–324. [65] C.-S. Lee, I.-D. Kim, J.-H. Lee, Selective and sensitive detection of trimethylamine using ZnO-In2 O3 composite nanofibers, Sens. Actuators B 181 (2013) 463–470. [66] C. Li, Z. Yu, S. Fang, S. Wu, Y. Gui, R. Chen, Synthesis and gas-sensing propertiesof Ce-doped SnO2 materials, J. Phys. Conf. Ser. 152 (2009) 012033. [67] A. Khodadadi, S.S. Mohajerzadeh, Y. Mourtazavi, A.M. Miri, Cerium oxide/SnO2 -based semiconductor gas sensor with improved sensitivity to CO, Sens. Actuators B 80 (2001) 267–271. [68] Z. Jiang, Z. Guo, B. Sun, Y. Jia, M. Li, J. Liu, Highly sensitive and selective butanone sensors based on cerium-doped SnO2 thin films, Sens. Actuators B 145 (2010) 667–673. [69] M.E.M. Hassouna, A.M. El-Sayed, F.M. Ismail, M.H. Khder, A.A. Farghali, S.M. Yakout, Investigation on the structure, electrical conductivity and ethanol gas sensitive properties of Ce-doped SnO2 nanoparticles sensors, Int. J. Nanomater. Biostruct. 2 (2012) 44–49. [70] C. Ge, C. Xie, S. Cai, Preparation and gas-sensing properties of Ce-doped ZnO thin-film sensors by dip-coating, Mater. Sci. Eng. B 137 (2007) 53–58. [71] F. Esch, S. Fabris, L. Zhou, L. Montini, C. Africh, P. Fornasiero, G. Comelli, R. Rosie, Electron localization determines defect formation on ceria substrates, Science 309 (2005) 752. [72] C. Ge, C. Xie, S. Cai, Preparation and gas-sensing properties of Ce-doped ZnO thin-film sensors by dip-coating, Mater. Sci. Eng. B 137 (2007) 53–58. [73] L. Xu, H. Song, B. Dong, Y. Wang, J. Chen, X. Bai, Preparation and bifunctional gas sensing properties of porous In2 O3 -CeO2 binary oxide nanotubes, Inorg. Chem. 49 (2010) 10590–10597. [74] L.I. Trakhtenberg, G.N. Gerasimov, V.F. Gromov, T.V. Belysheva, O.J. Ilegbusi, Effect of composition and temperature on conductive and sensing properties of CeO2 + In2 O3 nanocomposite films, Sens. Actuators B 209 (2015) 562–569. [75] G.N. Gerasimov, V.F. Gromov, L.I. Trakhtenberg, T.V. Belysheva, E.Yu. Spiridonova, V.M. Rozenbaum, Sensor properties of nanostructured In2 O3 –CeO2 system in detecting reducing gases, Russ. J. Phys. Chem. A 88 (2014) 503–508. [76] W.C. Chueh, A.H. McDaniel, M.E. Grass, Yong Hao, N. Jabeen, Z. Liu, S.M. Haile, K.F. McCarty, H. Bluhm, F. El Gabaly, Highly enhanced concentration and stability of reactive Ce3+ on doped CeO2 surface revealed in operando, Chem. Mater. 24 (2012) 1876–1883. [77] A. Badri, C. Binet, J.-C. Lavalley, An FTIR study of surface ceria hydroxy groups during a redox process with H2 , J. Chem. Soc. Faraday Trans. 92 (1996) 4669–4673. [78] S. Bernal, J.J. Calvino, G.A. Cifredo, J.M. Rodriguez-Izquierdo, Comments on Redox processes on pure ceria and Rh/CeO2 catalyst monitored by X-ray absorption (Fast Acquisition Mode), J. Phys. Chem. 99 (1995) 11794–11796. [79] I. Kosacki, T. Suzuki, H.U. Anderson, P. Colomban, Raman scattering and lattice defects in nanocrystalline CeO2 thin films, Solid State Ion. 149 (2002) 99–105. [80] A. Katoch, S.-W. Choi, G.-J. Sun, S.-S. Kim, An approach to detecting a reducing gas by radial modulation of electron-depleted shells in core-shell nanofibers, J. Mater. Chem. A 1 (2013) 13588–13596. [81] S.-W. Choi, A. Katoch, G.-J. Sun, J.-H. Kim, S.-H. Kim, S.-S. Kim, Dual functional sensing mechanism in SnO2 -ZnO core-shell nanowires, ACS Appl. Mater. Interfaces 6 (2014) 8281–8287.

624

G.N. Gerasimov et al. / Sensors and Actuators B 240 (2017) 613–624

[82] I.-S. Hwang, S.-J. Kim, J.-K. Choi, J. Choi, H. Ji, G.-T. Kim, G. Cao, J.-H. Lee, Synthesis and gas sensing characteristics of highly crystalline ZnO-SnO2 core-shell nanowires, Sens. Actuators B 148 (2010) 595–600. [83] Y.I. Chuan, L. Pang T.T.-Yuen, Hydrogen gas sensors using ZnO-SnO2 core-shell nanostructure, Adv. Sci. Lett. 3 (2010) 548–553. [84] N. Singh, A. Ponzoni, R.K. Gupta, P.S. Lee, E. Comini, Synthesis of In2 O3 -ZnO core-shell nanowires and their application in gas sensing, Sens. Actuators B 160 (2011) 1346–1351. [85] J. Zhang, X. Liu, L. Wang, T. Yang, X. Guo, Sh. Wu, Sh. Wang, Sh. Zhang, Synthesis and gas sensing properties of ␣-Fe2 O3 -ZnO core-shell nanospindles, Nanotechnology 22 (2011) 185501–185506. [86] S. Si, C. Li, X. Wang, Q. Peng, Y. Li, Fe2 O3 /ZnO core-shell nanorods for gas sensors, Sens. Actuators B 119 (2006) 52–56. [87] C.L. Zhu, Y.J. Chen, R.X. Wang, L.J. Wang, M.S. Cao, X.L. Shi, Synthesis and enhanced ethanol sensing properties of ␣-Fe2 O3 /ZnO heteronanostructures, Sens. Actuators B 140 (2009) 185–189. [88] V. Brinzari, G. Korotcenkov, V. Golovanov, Factors influencing the gas sensing characteristics of tin dioxide films deposited by spray pyrolysis: understanding and possibilities of control, Thin Solid Films 391 (2001) 167–175. [89] Yu. Chen, F. Meng, C. Ma, Z. Yang, Ch. Zhu, Q. Ouyang, P. Gao, J. Li, Ch. Sun, In situ diffusion growth of Fe2 (MoO4 )3 nanocrystals on the surface of ␣-MoO3 nanorods with significantly enhanced ethanol sensing properties, J. Mater. Chem. 22 (2012) 12900–12906. [90] D. Fu, Ch. Zhu, X. Zhang, Ch. Li, Yu. Chen, Two-dimensional net-like SnO2 /ZnO heteronanostructures for high-performance H2 S gas sensor, J. Mater. Chem. A 4 (2016) 1390–1398.

Biographies Genrikh N. Gerasimov graduated with a Doctor of Science degree from the Karpov Institute of Physical Chemistry (Moscow) in 1984. He is currently a leading scientist at the Semenov Institute of Chemical Physics, Moscow. Vladmir F. Gromov graduated with a Doctor of Science degree from the Karpov Institute of Physical Chemistry (Moscow) in 1988. He became a Professor of Highmolecular compounds Chemistry at the same institution in 1996. He is currently a leading scientist at the Semenov Institute of Chemical Physics, Moscow. Olusegun J. Ilegbusi is a Professor of Mechanical Engineering at the University of Central Florida, Orlando, USA. Prior to that, he was a Professor at Northeastern University, Boston, and a Visiting Professor and Senior Research Faculty at the Massachusetts Institute of Technology. He graduated with a Ph.D. degree in Mechanical Engineering at Imperial College, London in 1983. Professor Ilegbusi is an expert in mathematical and physical modeling of materials processing operations. Leonid Trakhtenberg is currently a Head of Laboratory at the Semenov Institute of Chemical Physics of the Russian Academy of Science in Moscow. He graduated with a Doctor of Science degree from the Karpov Institute of Physical Chemistry (Moscow) in 1986. He became a Professor of Chemical Physics at the same institution in 1997. Professor Trakhtenberg is an expert in nanocomposite film processing and mechanism of charge transfer in such films.