CHAPTER 3

CO₂ Response of Nanostructured CoSb₂O₆ Synthesized by a Nonaqueous Coprecipitation Method

Carlos R. Michel, Alma H. Martínez and Héctor Guillén-Bonilla

Departamento de Física, CUCEI Universidad de Guadalajara, Guadalajara, Jalisco, México

Contents

1. Introduction

2. Experimental

3. Results and Discussion

4. Conclusion

Acknowledgments

References

1. Introduction

Although CO₂ has played an important role in preserving the average atmospheric temperature over centuries, the emission of this gas in large amounts during recent decades is producing a phenomenon of great concern to mankind, global warming. With the goal of developing reliable solid-state gas sensors, notable scientific research has been made worldwide. As a result, several compounds such as SnO₂, ZnO, WO₃, TiO₂, and other inorganic oxides have been intensively tested for the detection of toxic gases like CO and NO₂ [1–4]. However, less information about materials tested for sensing CO₂ and other greenhouse gases is found in the literature.

In addition to the oxides already mentioned, several oxides possessing different crystal structures have been studied for gas-sensing purposes. Recently, investigating the gas-sensing properties of CoSb₂O₆ prepared by an aqueous solution-polymerization method, our group found a good response to CO₂ and O₂ [5]. This oxide has the trirutile type structure, as well as other complex antimony oxides, and has attracted the attention due to their unique physical and chemical properties.

On the other hand, for the preparation of nanostructured materials with micro and nanoporosity, the aqueous solution-polymerization methods usually produce good results [6–8]. In these methods, the homogeneous distribution of metal ions is made using a polymer, such as polyvinyl alcohol (PVA) or polyethylene glycol dissolved in water. Then, during calcination, the polymer decomposes producing a large amount of CO₂ and other gases, which develops extensive porosity and very small particles in the resulting solid material. The synthesis of nanoceramics by these methods largely depends on the solubility of metal nitrates and polymers in water.

In the search of other polymers that facilitate the formation of nanostructured materials, the synthesis of nickel nanoparticles using polyvinyl pyrrolidone (PVP) in ethylene glycol was recently reported [9]. The authors found that the molecular weight of PVP plays an important role in the formation of the nanoparticles and that by increasing the molecular weight of PVP, the growth and agglomeration of the metal particles can be prevented.

The synthesis of ceramic nanoparticles using nonaqueous solution-polymerization methods has been less studied, probably due to the limited null solubility of some reactants in common solvents such as alcohols. For instance, attempts to obtain CoSb₂O₆ by a solution-polymerization method using PVA in ethyl alcohol did not produce good results because PVA is not soluble in this alcohol; therefore, other solvents and chemical compounds were tested for optimal results.

In this work, the synthesis of nanostructured CoSb₂O₆ by a nonaqueous coprecipitation method using PVP, cobalt nitrate, and antimony trichloride in ethyl alcohol was explored. The relationship between the PVP concentration in solution and the surface morphology was investigated. The potential application of nanostructured CoSb₂O₆, as an environmental gas-sensor material, was studied by measuring the dynamic variation of resistance in air, CO₂, and O₂. The CO₂-sensing response was also studied by recording intensity vs. voltage graphs at different CO₂ concentrations.

2. EXPERIMENTAL

For the preparation of CoSb₂O₆, stoichiometric amounts of antimony trichloride (Sigma-Aldrich, 99%) and cobalt nitrate hexahydrate (Alfa Aesar, 98–102%) were dissolved separately in 5 ml of anhydrous ethyl alcohol (J. T. Baker), and the resulting transparent solutions were colorless and red, respectively. To test the role of PVP on the morphology of CoSb₂O₆, three solutions containing, respectively, 0.05, 0.22, and 0.44 g of PVP (Sigma-Aldrich, average molecular weight = 10,000) in 5 ml of ethyl alcohol were prepared. Because PVP is soluble in ethyl alcohol, clear solutions were obtained. The next step was the mixture of each PVP solution with the solution of cobalt nitrate; this was made under strong stirring, producing transparent red liquids. After 20 min of stirring, the solution of antimony chloride was added, which formed immediately blue–turquoise suspensions, with a pH value of approximately 2; these were stirred for 18 h. The evaporation of the suspensions was made by microwave radiation using a domestic microwave oven, working at low power; the temperature of the beakers was maintained below 50 °C. The resulting precursors were annealed in static air from 200 to 600 °C, using a muffle type furnace with programmable temperature control. A heating rate of 100 °C/h was used in each calcination.

The structural characterization was made by X-ray powder diffraction (XRD) at room temperature, using a Rigaku Miniflex apparatus (Cu K_α radiation). The diffraction angle (2θ) was scanned from 10° to 70°. The surface morphology was observed by scanning electron microscopy (SEM) using a Jeol JSM-5400LV microscope. The observation of nanostructural characteristics was made by transmission electron microscopy (TEM) using a Jeol JEM-1010 microscope (operated at 100 kV). Before the TEM observation, the powders were sonicated in isopropyl alcohol for 5 min. Electrical conductivity and gas-response characterization were performed on thick films made with the powder prepared with 0.22 g of PVP. The films were made by dispersing 0.1 g of CoSb₂O₆ in 2 ml of ethyl alcohol using an ultrasonic bath; then the mixture was deposited into alumina substrates forming films with 5 mm diameter and ~300 µm thickness. Silver wires were fixed to the films as electrical contacts. The electrical measurements were carried out using the two-point probe method, from room temperature to 600 °C, using a digital data acquisition unit (Agilent 34970A) and a digital voltmeter (Agilent 34401A). The variation of intensity with voltage was recorded with a potentiostat/galvanostat Solartron 1285A. During the gas-sensing characterization, the gases were supplied by a MKS 647C mass flow controller.

3. RESULTS AND DISCUSSION

Figure 3.1 shows the XRD patterns of precursor powders obtained using the three different concentrations of PVP, after their calcination at 600 °C for 4 h. Clearly, single-phase CoSb₂O₆ was produced in each synthesis; the identification of the peaks was made by using the international centre for diffraction data (ICDD) file No. 18-0403, corresponding to CoSb₂O₆. The XRD patterns of samples synthesized with 0.05 and 0.22 g of PVP exhibit wider peaks than the sample synthesized with 0.44 g, indicating a larger average particle size in the latter.

Figure 3.1 XRD patterns of CoSb₂O₆ precursor powders prepared with 0.05, 0.22, and 0.44 g of PVP, after a calcination at 600 °C in air.

The crystal structural evolution while increasing the calcination temperature was studied using XRD on a sample prepared with 0.22 g of PVP. This sample was selected due to its particular nanostructural characteristics, which will be presented later in the SEM and TEM results. Figure 3.2 shows the XRD patterns of the precursor powder calcined from 200 to 600 °C. For the sample fired at 200 °C, some peaks of low intensity placed in 2θ from 15 to 20° and from 30 to 50° can be observed. The identification of such peaks was not possible using the ICDD database because none of them correspond to crystalline phases containing cobalt or antimony; probably, they are associated with products of decomposition of PVP. At a calcination temperature of 300 °C, the XRD pattern shows mainly an amorphous material; however, the incipient presence of the three main peaks of CoSb₂O₆, placed at 2θ = 27, 34.8, and 53°, can be observed. From 400 °C, the main peaks of single-phase CoSb₂O₆ can be clearly identified, though the calcination at 600 °C produced a better crystallized material.

Figure 3.2 Crystal structure evolution as a function of the calcination temperature for a CoSb₂O₆ precursor powder made with 0.22 g of PVP.

Figure 3.3 shows the typical surface morphology of CoSb₂O₆ prepared with 0.05 g (A), 0.22 g (B, C), and 0.44 g (D, E) of PVP, calcined at 600 °C. These observations were made by SEM at low magnification on the three samples, but in the case of the last two samples, larger magnifications are also shown, exhibiting their microstructure in more detail. When 0.05 g of PVP was used, a compact agglomeration of very fine particles was obtained; little porosity and few cracks were also observed (Fig. 3.3A). When 0.22 g of PVP was used, large scale of thin wires was developed (Fig. 3.3B and C). The length of these wires was several micrometers and about 100 nm width. When 0.44 g of PVP was used, a similar formation of wires, than the sample prepared with 0.22 g, can be observed in Fig. 3.3D and E; however, the size of the wires is considerably larger. These results show that the amount of PVP used in solution determines the microstructure of CoSb₂O₆.

To know whether the concentration of PVP solely determines the formation of wires in CoSb₂O₆, two different parameters were studied. First, the effect of stirring the suspension after their preparation was investigated. Then a new synthesis of CoSb₂O₆ using 0.22 g of PVP was made. After mixing the cobalt, PVP, and antimony solutions, the resulting suspension was under strong stirring for 1 h. Then it was maintained in calm for 18 h. This time was the same used in the previous synthesis under stirring. The evaporation of the solvent and further calcination was made using the same procedure than before. Figure 3.4 shows a typical XRD pattern of the powder obtained and its microstructure observed by SEM (inset). Clearly single-phase CoSb₂O₆ was obtained; however, the morphology shows particles with irregular shape and size smaller than ~1.3 µm. Evidently, in this case, the wires were not produced.

Figure 3.3 SEM images of CoSb₂O₆ prepared with 0.05 g (A), 0.22 g (B, C), and 0.44 g (D, E) of PVP.

The second parameter studied was the effect of the time that the sample is calcined at 600 ^°C, and this was made with the purpose of investigating a possible growth of the wires. This sample was prepared under stirring and remained at 600 °C for 24 h. Figure 3.5 shows the XRD pattern and its microstructure (inset); as it was expected, single-phase CoSb₂O₆ was obtained as shown in the XRD pattern. However, the microstructure mainly corresponds to rectangular flat particles, with size in the range 1–10 µm. This result indicates that grain growth dominates during the sintering. Therefore, controlling the PVP concentration, the time of stirring, the time of calcination, and the morphology of CoSb₂O₆ can be greatly modified using this synthesis method.

Figure 3.4 XRD pattern and SEM image (inset) of a CoSb₂O₆ sample prepared with 0.22 g of PVP, without stirring (600 °C).

Figure 3.5 XRD pattern and SEM micrograph (inset), corresponding to a CoSb₂O₆ sample synthesized with 0.22 g of PVP, calcined at 600 °C for 24 h.

On the other hand, Fig. 3.6 shows the typical TEM images, in bright-field mode, of CoSb₂O₆ prepared with 0.05 g (A), 0.22 g (B), and 0.44 g (C) of PVP calcined at 600 °C for 4 h. For the synthesis with 0.05 g, a compact arrangement of nanoparticles with few porosity can be observed; this morphology corresponds to the particles observed in Fig. 3.3A. The average particle size was measured from several TEM images resulting in 8 nm; Fig. 3.7A shows the particle size distribution. For CoSb₂O₆ prepared with 0.22 g of PVP, Fig. 3.6B exhibits the formation of nanoparticles and nanorods. In the case of nanoparticles, an average particle size of ~10 nm was found, and for nanorods an average diameter of 8 nm and length until 65 nm were observed. Figure 3.7B shows the diameter and length size distributions for this sample. It is noticeable that SEM shows wires for this sample, whereas TEM shows nanoparticles and nanorods. This can be explained by a fragmentation of the wires occurred during the sonication of the sample prior to CoSb₂O₆ observation by TEM.

For CoSb₂O₆ synthesis using 0.44 g of PVP, Fig. 3.6C shows a single wire exhibiting a bend in one of the ends, which was commonly observed by SEM (Fig. 3.3E). In this sample, the wires are composed of the aggregation of larger particles than those obtained for sample prepared with 0.22 g of PVP. The size of individual particles was measured, and Fig. 3.7C shows their corresponding particle size distribution.

Figure 3.6 TEM images of CoSb₂O₆ prepared with 0.05 g (A), 0.22 g (B), and 0.44 g (C) of PVP (calcined at 600 °C for 5 h).

Figure 3.7 Particle size distributions of CoSb₂O₆ synthesized with 0.05 g (A), 0.22 g (B), and 0.44 g (C) of PVP.

To test whether the synthesis of CoSb₂O₆ gives reproducible results, several preparations using each PVP concentration were made, obtaining similar microstructural characteristics than those presented in this work.

Comparing the nanostructural characteristics observed by TEM, clearly, an increase in particle size is observed while increasing the amount of PVP, which is in agreement with the XRD and SEM results. On the other side, although the CoSb₂O₆ powder made with 0.22 g of PVP did not have the smallest particle size, a larger specific surface area can be qualitatively inferred in this sample, because less agglomeration and plenty of nanoparticles was observed. Due to these characteristics, this powder was used for testing as environmental gas-sensing material.

To investigate the conductivity of the target CoSb₂O₆, thick films were heated from room temperature to 600 °C in dynamic air, CO₂, and O₂ while measuring the resistance. The conductivity (σ) was calculated according to the well-known equation [10]:

where R is the resistance between the two electrical contacts, separated by a length l, and S is the section of the film. Figure 3.8 shows typical Arrhenius-type plots obtained in the three gases; the increment of the conductivity with temperature observed in this figure is characteristic of the semiconductor behavior. Although the gas type had a little effect on the shape of each curve, the smallest conductivity values were registered in CO₂, in most of the temperature ranges studied. The conductivity goes from ~10^–6 to ~10^–2(Ω cm)^–1 and has two branches, with a transition temperature (T₁) at ~350 °C for air and O₂ and at ~460 °C for CO₂. The activation energies for conduction (E_A) were calculated from the slopes of each section of the curves, Table 3.1 shows these results. The transition temperature (T₁) is associated with a change in the conduction mode from a semiconductor to metallic; therefore, E_A values are smaller at higher temperatures.

For the evaluation of the environmental gas-sensing response of CoSb₂O₆ thick films, the variation of resistance with time was measured while supplying alternatively air, CO₂, and O₂. To test the response in CO₂, dry synthetic air was supplied for approximately 5 min using a flow rate of 0.22 sccm; then CO₂ was delivered for approximately 2.5 min using the same flow rate; finally, an air flow was supplied. This operation was performed several times to confirm the reproducibility. To test the films in oxygen, a similar procedure was used. In this case, the CO₂ flow was replaced by O₂.

Figure 3.8 Arrhenius plots of CoSb₂O₆ (0.22 g of PVP) thick films measured in air, O₂, and CO₂.

Table 3.1 Calculated activation energies of conduction (E_A) for CoSb₂O₆ prepared with 0.22 g of PVP

Figure 3.9 shows the typical dynamic response in CO₂ (A), and O₂ (B); these graphs were recorded at temperatures of 400 and 500 °C, respectively, where a larger response was registered. Figure 3.9A shows the introduction of CO₂ that produced a variation of resistance (ΔR) of approximately 2.7 kΩ, whereas in O₂, the opposite behavior is observed. In this gas, ΔR was approximately −0.5 kΩ. Regardless of the gas used, a reproducible response pattern was recorded in both tests.

Because the resistance of CoSb₂O₆ decreased in O₂, whereas in CO₂ has the opposite behavior, this oxide can be considered a P-type semiconductor material [11]. Besides, when oxygen surrounds the CoSb₂O₆ surface, at least two possible adsorption processes can occur:

Figure 3.9 Dynamic response of resistance of nanostructured CoSb₂O₆ (0.22 g of PVP) thick films in CO₂ (A) and O₂ (B), measured at 400 °C.

where (ads) means the adsorbed species [12]. The adsorption of oxygen produces an increase in the number of holes in the valence band, which increases the conductivity. For the detection of CO₂, a rather different mechanism may be involved because the carbon is fully oxidized and the electron transfer is less probable. Ishihara et al., have suggested that a thin carbonation layer may be formed during the CO₂ detection, which changes the electrical conductivity of the material [13]. According to the results shown in Fig. 3.9A, the carbonation layer is removed upon the introduction of air because a full recovery of the resistance is observed in each air–CO₂ cycle.

To test the ability of CoSb₂O₆ to detect CO₂ at different concentrations, current–voltage (I–V) graphs were recorded from −10 to 10 V, at 500 °C. Figure 3.10 shows the results obtained for several CO₂:air ratios. In these graphs, for each CO₂:air ratio, a curve can be easily distinguished, and these curves have a nonlinear trend in both positive and negative voltages. When a 100% CO₂ flow was used, the smaller current values were recorded at positive voltages (Fig. 3.10A). When the concentration of CO₂ was decreased, for CO₂:air ratios from 1:1 to 1:3, a clear increase in the current can be observed, and this increment in current is caused by the increment of oxygen contained in the air.

For negative voltages, the inverse behavior was obtained (Fig. 3.10B); however, the magnitude of the current was significantly smaller compared to the results at positive voltages. Moreover, comparing the values of the polarization curves at positive voltages, with those obtained in the tests of dynamic response, a good agreement can be observed.

To test the response of CoSb₂O₆ thick films to a variation in O₂ concentration, polarization curves, using several O₂:air flows, were recorded. However, the results show a superposition of the curves, and a poor sensitivity to this gas was concluded.

4. CONCLUSION

A nonaqueous coprecipitation method based on PVP in ethyl alcohol was used for the preparation of CoSb₂O₆. Using this simple method, CoSb₂O₆ was obtained at a low calcination temperature, with good control on the stoichiometry. By an appropriate control of the amount of PVP, the time of stirring, and calcination, it was possible to prepare nanostructured CoSb₂O₆ powders with an average size ~10 nm, having a narrow particle size distribution. Because PVP is an inexpensive polymer, which has been available for more than five decades, this may be a convenient method for the synthesis of this CoSb₂O₆, and other inorganic oxides.

Figure 3.10 Current-voltage curves of CoSb₂O₆ (0.22 g of PVP), measured at 400 °C in several CO₂:air proportions.

The gas-sensing properties were studied on the CoSb₂O₆ powder with the best nanostructural characteristics for which thick films were prepared with this material, without further modification of its nanostructure. According to the results obtained from the electrical characterization of O₂ and CO₂, CoSb₂O₆ can have a potential use as a CO₂ sensor material.

ACKNOWLEDGMENTS

The author is grateful to the Coordinación General Académica of the Universidad de Guadalajara and the National Council of Science and Technology of Mexico (CONACYT) for financial support under the project I-52204. A.H.M. and H.G.B want to thank CONACYT for their doctorate scholarships in the Universidad de Guadalajara.

References

[1] Yamazoe N. Toward innovations of gas sensor technology, Sens Actuators B 2005; 108:2–14.

[2] Shimizu Y, Egashira M. Basic aspects and challenges of semiconductor gas sensors, Mater Res Soc Bull 1999;6:18–24.

[3] Graf M, Barrettino D, Taschini S, Hagleitner C, Hierlemann A, Baltes H. Metal oxide-based monolithic complementary metal oxide semiconductor gas sensor microsystem, Anal Chem 2004;76:4437–4445.

[4] Azad AM, Akbar SA, Mhaisalkar SG, Birkefeld LD, Goto KS. Solid-state gas sensors: A review, J Electrochem Soc 1992;139:3690–3704.

[5] Michel CR, Martínez AH, Jiménez S. Gas sensing response of nanostructured trirutile-type CoSb₂0₆ synthesized by solution-polymerization method, Sens Actuators B 2008;132:45–51.

[6] Lee SJ, Kriven WM. Crystallization and densification of nano-size amorphous cordierite powder prepared by a PVA solution-polymerization route, J Am Ceram Soc 1998;81:2605–2612.

[7] Michel CR, Delgado E, Martínez AH. Evidence of improvement in gas sensing properties of nanostructured bismuth cobaltite prepared by solution-polymerization method, Sens Actuators B 2007;125:389–395.

[8] Michel CR, López-Mena E, Martínez AH. Improvement of the gas sensing behavior in nanostructured Gd_0.9Sr_0.1CoO₃ by addition of silver, Mater Sci Eng B 2007;141:1–7.

[9] Li D, Komarneni S. Microwave-assisted polyol process for synthesis of Ni nanoparticles, J Am Ceram Soc 2006;89:1510–1517.

[10] Schroder DK. Semiconductor material and device characterization. New York: Wiley; 2006. p. 11.

[11] Moseley PT, Stoneham AM, Williams DE. Oxide semiconductors: patterns of response behaviour according to materials type, In: Moseley PT, Norris J, Williams DE, editors. Techniques and mechanisms in gas sensing. Bristol: Adam Hilger; 1991. p. 108–38.

[12] Bochenkov VE, Sergeev GB. Preparation and chemiresistive properties of nanostructured materials, Adv Colloid Interface Sci 2005;116:245–254.

[13] Ishihara T, Kometani K, Mizuhara Y, Takita Y. Mixed oxide capacitor of CuO-BaTiO₃as a new type CO₂ gas sensor, J Am Ceram Soc 1992;75:613–618.