Nitrous oxide (N2O) is a contributor to the destruction of ozone in the stratosphere and a strong greenhouse gas [1, 2]. It is emitted from both natural and anthropogenic sources. Anthropogenic N2O emissions come mainly from chemical industries, e.g., the production of nitric acid [2, 3] and organic synthesis that use the nitric acid oxidation process, such as the production of adipic acid from the oxidation of cyclohexanol- cyclohexanone mixture [4]. The catalytic removal of N2O from anthropogenic sources is one possible solution to protect our global environment. A wide variety of catalysts have been reported for the catalytic decomposition of N2O to N2 and O2. These catalysts include noble metals [5, 6], metal oxides [7, 8], supported oxides [9, 10], ion exchanged zeolites [11, 12, 13], hexaferrites [14], perovskites [15, 16], hydrotalcites [17, 18], and spinels [19, 20, 21, 22].
Spinel oxides are a class of complex oxides with the general chemical formulas of AB2O4 (A ions are divalent cations occupying tetrahedral sites and B ions are trivalent cations that occupy octahedral sites). Cobalt oxide spinel (Co3O4) is receiving considerable interest [23, 24]. The A component of the Co3O4 spinel is often partially substituted with another divalent metal such as Mg, Ni, or Zn [19, 20, 22] to create special properties for applications [25, 26, 27]. It was reported that the catalytic activity of Co3O4spinel is affected by the preparation method, degree of Co2+ (A component) substitution, degree of spinel inversion and the presence of alkali dopants [19, 20, 22, 28]. Generally, Co3O4-based spinels can decompose N2O completely at low temperatures (below 500 °C) [19, 20]. Their activity decrease slightly with O2 and/or H2O in the reactor feed [19, 20]. The presence of many dopants like Zr4+, Ce4+, Li+, Na+, K+, Cs+, Mg2+, Ca2+, or Ba2+ enhances the Co3O4-based spinel catalysts for the decomposition of N2O [22, 29, 30, 31]. There has been only one paper on the N2O decomposition activity over Ni substituted Co3O4 [19], but this lacked information regarding the influence of calcination temperature and the role of the different parameters on the N2O decomposition activity. In this paper, a series of nickel cobaltite with the general formula NixCo1-xCo2O4 were prepared by the co-precipitation method and characterized by various techniques. The performance of the catalysts for N2O decomposition was discussed using the degree of Co2+ substitution by Ni2+, spinel crystallite size, catalyst surface area, presence of residual K+, and calcination temperature.
A catalyst series with the general formula NixCo1-xCo2O4 (x = 0.00, 0.25, 0.50, 0.75, and 1.00) were prepared by a co-precipitation method similar to that reported by Xue et al. [29]. Briefly, an aqueous solution of K2CO3 (2 mol/L) was added dropwise into an aqueous solution containing stoichiometric amounts of cobalt and nickel acetate at room temperature under mechanical stirring until pH = 9.1-9.4 was reached. The slurry was stirred for 30 min and aged for 3 h. The precipitate was filtered and washed with distilled water several times. Evaporation of excess water in the precipitate cake was done by drying in an oven at 100 °C overnight. Based on thermal analysis results (vide infra), all the dried precipitates were calcined at 500 °C for 3 h. In addition, based on the N2O decomposition activity measurement, the precipitate with x = 0.75 was calcined at 750 and 1000 °C for 3 h.
Thermogravimetry (TGA) and differential thermal analysis (DTA) curves were recorded using a Shimadzu DTG-60 instrument. 10 mg of the dried precipitate was placed in a platinum crucible and heated at a heating rate of 10 °C/min in flowing air (40 ml/min). X-ray diffraction (XRD) patterns were recorded using a Philips X-ray diffractometer (Type PW 2103/00) employing Cu Kαradiation (λ = 0.15418 nm). Fourier transform infrared (FT-IR) spectra were obtained using the KBr disk technique on a Thermo-Nicolet-6700 FT-IR spectrophotometer. N2 adsorption-desorption isotherms were measured on a NOVA 3200 automated gas adsorption system (Quantachrome) at liquid nitrogen temperature. The K+ concentration in the dried samples was measured by atomic absorption using a 210 VGP atomic absorption spectrophotometer.
Catalytic performance was evaluated with an isothermal plug flow reactor. The procedure was similar to that reported previously [6, 22, 30]. Each catalytic run was conducted using 500 mg of the catalyst and a gas mixture of N2O (500 ppm) and N2 as a balance gas at a flow rate of 200 ml/min. Before each run, the catalyst sample was heated at 500 °C for 1 h in a flow of N2, and then cooled to 150 °C and the reactant was introduced. The exit concentrations were monitored by a magnetic oxygen analyzer (ABB, AO2020-Magnos 106) and a non-dispersive infrared analyzer (ABB, AO2020-Uras 14) for N2O and NO. The steady state was reached after about 1 h. Preliminary experiments for the decomposition of N2O over all the catalysts showed the absence of NO in the exit gas.
The TGA thermogram, shown in Fig. 1(a), has two regions. The first region from ambient temperature to 400 °C was accompanied by two weight-loss steps. The first step has a maximum at 61 °C which was attributed to the dehydration of the carbonates. The second step was not a simple one. It was a composite step with maxima at 247 and 296 °C (DTG curve). The TGA-DTA thermogram of the cobalt carbonate parent (not shown) revealed that cobalt carbonate decomposed in three steps, with peaks at 73 (endothermic), 263 (exothermic), and 930 °C (endothermic). These steps were attributed to dehydration of the parent, decomposition of the anhydrous cobalt carbonate to Co3O4 spinel and the thermal reduction of Co3O4 spinel to CoO, respectively. Mansour reported on the thermal decomposition of nickel carbonate that it decomposed by two endothermic steps at 115 and 310 °C, which were attributed to the dehydration of the salt and decomposition of the anhydrous salt leading to the formation of NiO, respectively [32]. Accordingly, the second composite step with maxima at 247 and 296 °C (Fig. 1(a)) was assigned to the consecutive decomposition of cobalt and nickel carbonate. As shown in Fig. 1(b), one can observe three thermal events from ambient temperature to 400 °C. The first is endothermic with the maximum at 73 °C, and can be related to the dehydration process. The second peak (exothermic) with the maximum at 224 °C was attributed to the oxidation of Co2+ → Co3+ that accompanied the thermal decomposition of cobalt carbonate. The third peak at 309 °C was assigned to: (1) the solid state interaction between NiO and Co3O4 leading to the formation of nickel cobaltite spinel and (2) crystallization of the spinel obtained. In the second temperature range from 400 to 1000 °C, it displays one decomposition step with a maximum at 834 °C (Fig. 1(a)), which was endothermic (Fig. 1(b)). Since Co3O4 spinel oxide undergoes thermal reduction at temperatures as high as 930 °C [30], this step can be assigned to the thermal decomposition of nickel cobaltite. This suggestion is in good agreement with the results of Chi et al. [33] who reported that the thermal decomposition of NiCo2O4 spinel was at 800 °C.
XRD patterns of the nickel/cobalt precipitate mixtures calcined at 500 °C are shown in Fig. 2. The pure cobalt (oxide) calcined at 500 °C showed peaks at 2θ = 18.88°, 31.07°, 36.72°, 38.47°, 44.07°, 55.58°, 59.35°, 65.16°, 74.01°, 77.19°, and 78.36. These peaks matched well those of Co3O4 (JCPDS 78-1969). Ni0.25Co0.75Co2O4 and Ni0.5Co0.5Co2O4 samples showed the same peaks and no other phase was detected. These diffractograms matched the standard card of NiCo2O4 (JCPDS 73-1702) reported by other research groups [33, 34, 35]. From the absence of peaks due to NiO, one can state that Ni preferentially formed the nickel cobaltite spinel for the samples with x = 0.25 and 0.50. For the samples with x = 0.75 and 1.0, a further increase in the Ni content in the spinel mixture was accompanied by new peaks at 2θ = 43.25° and 62.84°, which were due to NiO (JCPDS 78-0643). Accordingly, one can conclude that in the utilized procedure, the formation of NiCo2O4 was accompanied by the formation of a trace amount of NiO as impurity. This was in good agreement with other research groups [33, 34, 35]. Lapham et al. [34] prepared NiCo2O4 by thermal decomposition of the metal nitrates and pointed out that the NiO phase initially appeared in the sample fired at 400 °C and it co-existed with NiCo2O4 in larger quantities up to the firing temperature of 600 °C. They concluded that Ni preferentially formed the spinel NiCo2O4 below 350 °C but gave a separate NiO phase above 400 °C. Chi et al. [33] have detected NiCo2O4 as a pure phase in the samples prepared by the co-precipitation method using metal nitrate and NaOH and a pretreatment temperature range of 200-350 °C. They also observed the peaks due to the NiO impurity in their XRD diffractograms for the sample calcined at 400 °C. Cabo et al. [35] have also detected the NiO impurity in their NiCo2O4 spinel calcined at 550 °C. The broadening of the diffraction lines with increasing x value (Fig. 2) may signify the nano-scale characteristic of the component crystallites. The crystallite size of the spinel phase for the 500 °C calcined catalyst was calculated using the Scherrer equation [6]. The values are listed in Table 1. It is evident that increasing the Ni content in the samples led to a crystallite size decrease until x = 0.75, and then there was a slight increase in the NiCo2O4 catalyst.
The XRD patterns of the Ni/Co mixture with x = 0.75 calcined at different temperatures are depicted in Fig. 3. Raising the calcination temperature to 750 °C was accompanied by an intensity increase of the peaks due to the spinel phase as well as those due to NiO. Increasing the calcination temperature to 1000 °C gave a marked decrease of the peaks due to the spinal phase. Meanwhile, the intensity of the peaks attributable to NiO at 2θ = 43.25°, 62.6°, and 75.1° disappeared. In addition, new peaks emerged at 2θ = 42.6°, 61.9°, 74.2°, and 78.1°. These peaks match those of CoO (JCPDS 75-0418). The diffractogram obtained for the sample calcined at 1000 °C matched that reported by Fujishiro et al. [36] for NiCo2O4 calcined at 900 °C, which suggested the formation of the rock salt (Ni,Co)O as a result of the decomposition of NiCo2O4.
Figure 4 shows the FT-IR spectra of NixCo1-xCo2O4 catalysts calcined at 500 °C for 3 h. All the spectra have two strong absorption at 564 and 655 cm-1, which were assigned to the ν1 and ν2stretching vibrations of the metal-oxygen bond in cobalt spinel oxide [30, 31]. Moreover, a weak absorption was observed at 450 cm-1 for the Ni0.75Co0.25Co2O4 and NiCo2O4 catalysts. This absorption is characteristic of NiO [32]. This observation is in good agreement with the XRD results (Fig. 2). Figure 4 revealed the presence of another set of absorption at 835, 1008, and 1384-1462 cm-1, which can be assigned to the different vibration modes of the carbonate anion [30, 31]. It is worth mentioning that the intensity of the carbonate absorption bands was higher for the Ni-containing catalysts compared to Co3O4. This finding agreed with the measured residual K+ concentrations, showing the presence of K2CO3 in the samples with higher Ni content (Table 1). For the same series of catalysts, the observed absorption spectra at 1630 and 3000-3600 cm-1 were assigned to the H-O-H bending and O-H stretching mode of adsorbed water, respectively [31, 37].
The FT-IR spectra of the Ni0.75Co0.25Co2O4 catalyst calcined at different temperatures (Fig. 5) showed that the catalyst calcined at 750 °C had a spectrum similar to that at 500 °C. This suggested that they had the same composition. Raising the calcination temperature to 1000 °C, the ν1 stretching mode appeared only as a shoulder at 561 cm-1 while the ν2 mode existed as a weak absorption at 657 cm-1. Moreover, strong absorption was observed below 600 cm-1. This absorption was in the IR spectral region expected of CoO [38]. Furthermore, two absorption peaks appeared at 785 and 1047 cm-1, which were attributed to the newly formed (Ni,Co)O phase suggested by the XRD results.
N2 adsorption-desorption isotherms of NixCo1-xCo2O4 catalysts calcined at 500 °C are shown in Fig. 6(a). Adsorption occurred at very low relative pressure suggesting a strong interaction between the pore wall and adsorbate. In this case, pore filling took place without capillary condensation at p/p0 < 0.3. The isotherms were Type I [39] at low pressure and Type II at higher p/p0 values, especially for the samples having high Ni content. This feature indicated the presence of micropores with some mesopores [13]. It showed that the desorption branch approximately overlapped the adsorption branch. The specific surface areas were obtained by BET equation and are tabulated in Table 2. The introduction of Ni with x = 0.25 to Co3O4 led to a noticeable SBET decrease. A further increase in x to 0.50 was accompanied by a sharp SBET increase, and then approximately constant values were obtained for the catalysts with x = 0.75 and 1.00. An additional set of surface areas (St) were determined from the volume-thickness curves (Va-t plots). These curves were constructed using the appropriate standard t-curves [39], and the results are shown in Fig. 6(b) and Table 2. It is seen that Co3O4 has mainly narrow pores as indicated by the downward deviation of its Va-t plot. This downward deviation with a little upward deviation indicated the presence of larger pores. The introduction of Ni into Co3O4 spinel gave Va-t plots characterized by the presence of a positive trend (upward) followed by a negative one (downward), which indicated the coexistence of both micro- and mesopores for this series of catalysts. Meanwhile, the St values showed a slight decrease for the catalyst with x = 0.25 as compared to the one without Ni. The further addition of Ni to x = 1.00 gave higher St values compared to that of Co3O4.
N2 adsorption-desorption isotherms for the Ni0.75Co0.25Co2O4 catalyst (calcined at 750 and 1000 °C not shown) indicated that the Type I isotherm became more pronounced with a calcination temperature increase and the desorption branch basically overlapped the adsorption one. The textural data are listed in Table 3. The SBET values for the 750 and 1000 °C calcined catalysts were 7.9 and 0.4 m2/g, respectively. These values were much lower than that of the sample calcined at 500 °C (40.72 m2/g). Clearly, the external surface area, micropore surface area, and micropore volume obtained from the t-method for this series of catalysts decreased as the calcination temperature increased. Moreover, the total pore volume decreased while the average pore diameter increased. These trends in the textural parameters can be attributed to sintering and densification, which increased with the calcination temperature.
Figure 7 depicts the conversion of N2O versus the x value in NixCo1-xCo2O4 catalysts calcined at 500 °C under different reaction temperatures. The reaction started at 250 and 200 °C for Co3O4 and Ni0.25Co0.75Co2O4 samples, respectively, and at 150 °C for the other catalysts. The activity increased with increasing reaction temperature over all the catalysts, and reached 100% conversion at 475-500 °C over the catalysts with x = 0.50, 0.75, and 1.00. Preliminary experiments of N2O conversion over pure NiO (SBET = 19 m2/g) prepared by co-precipitation and calcined at 500 °C gave 42% conversion at 500 °C. Thus, the results indicated that N2O conversion was much enhanced by the addition of NiO to Co3O4 spinel catalysts. In other words, the mixed NiO-Co3O4 catalysts calcined at 500 °C were significantly more active than the pure NiO and Co3O4 catalysts. This enhancement is usually referred to as a synergistic effect. Yan et al. [19] have investigated N2O decomposition over a series of MxCo1-xCo2O4 (M = Mg and Ni; x = 0.00-0.99) spinel catalysts prepared by co-precipitation, and the results showed that the catalytic activity depended on the degree of Co2+ substitution by Mg2+ or Ni2+. The highest activity was observed over the Mg0.54Co0.46Co2O4 and Ni0.74Co0.26Co2O4 catalysts. The characterization results showed that pure Co3O4 underwent thermal reduction at a high temperature, which was indicated by the enothermic peak at 930 °C [30]. The addition of Ni (x = 0.75) led to a marked decrease of this peak temperature to 835 °C (Fig. 1(b)). In other words, the presence of Ni2+ enhanced the thermal reduction of the Co3+. Recently, it was reported that the role of Mg2+ in the increase of N2O decomposition activity over a series of MgxCo1-xCo2O4 catalysts was to enhance the Co3+→ Co2+ reduction, i.e., regeneration of the Co2+ ions that are required for initiating N2O adsorption and its subsequent decomposition [22]. Based on this idea, it is plausible to suggest that the role of the added Ni2+ ions was also to enhance the Co3+→ Co2+ reduction and thus increase the catalytic activity. Also, it is worth mentioning that the presence of the residual K+ would also enhance the cobalt redox process. On comparing our results with those of Sundararajan et al. [40] for N2O decomposition over NiCo2O4 prepared by the thermal decomposition of the corresponding nitrates, our catalysts showed better performance. This could be attributed to the additional factor of the presence of residual K+. In this context, it should be mentioned that the Ni0.50Co0.50Co2O4, Ni0.75Co0.25Co2O4, and NiCo2O4 catalysts contained lower K+ concentrations compared to the Ni0.25Co0.75Co2O4 sample (Table 1). This, in turn, suggested that the samples having x values in the range 0.50 to 1.00 possessed the optimal balance between the K+ concentration and Ni content required for the highest activity. Yan et al. [19] have investigated N2O decomposition over NixCo1-xCo2O4 spinel oxide catalysts and the results indicated that the activity depended on the degree of Co2+ sub stitution, and the highest activity was exhibited by the catalysts having x = 0.36, 0.55, and 0.75. Using O2-TPD, they observed that O2 desorption, i.e., catalyst reducibility was facilitated by the presence of Ni2+ in the spinel [19]. In this work, it was also shown that the addition of Ni2+ with x = 0.50, 0.75, and 1.00 to Co3O4 increased the specific surface area from 16.4 to 40 m2/g (Table 2). Therefore, we can also attribute the positive effect of adding NiO in promoting Co3O4 activity as due to its role in increasing the catalyst surface area. Xue et al. [29] have demonstrated that the addition of CeO2 to Co3O4 led to an improvement in the activity for N2O decomposition. The catalyst was the most active when the molar ratio of Ce/Co was 0.05. Complete N2O conversion was obtained with the Co/Ce = 0.05 catalyst below 400 °C even in the presence of O2, H2O, or NO. Based on various physico-chemical characterization, these authors proposed that the addition of CeO2 increased the surface area of Co3O4, and then facilitated the reduction of Co3+ to Co2+ by facilitating the desorption of adsorbed oxygen species, which is the rate determining step in N2O decomposition over the cobalt spinel catalyst.
In the literature, many authors have correlated the N2O decomposition activity with the crystallite size of the catalyst. Recently, we reported an inverse relationship between N2O decomposition activity and the Co3O4 crystallite size of SrCO3- and BaCO3-Co3O4 catalysts [30]. Over Ag/FexAl2-xO3 catalysts, it was demonstrated that N2O decomposition reactivity increased with decreasing Ag crystallite size [6]. Accordingly, the higher N2O decomposition reactivity over the catalysts with x = 0.50, 0.75, and 1.00 can be related to their smaller spinel crystallite size as shown in Table 1. Accordingly, we conclude that the enhancement of N2O decomposition reactivity over NixCo1-xCo2O4 catalysts having x ≥ 0.25 was due to the following factors: (1) increasing amounts of Ni2+, which enhanced the thermal reduction of Co3+ ions, (2) presence of residual K+, (3) higher surface area of the nickel-containing spinels, and (4) decrease in crystallite size.
In the previous section, it was discussed that the highest N2O decomposition reactivity was exhibited over the Ni0.50Co0.50Co2O4, Ni0.75Co0.25Co2O4, and NiCo2O4 catalysts. Accordingly, we decided to check the influence of calcination temperature on the activity of Ni0.75Co0.25Co2O4. Figure 8 shows the dependence of N2O conversion on reaction temperature over the Ni0.75Co0.25Co2O4 catalyst calcined at different temperatures. It is obvious that raising the calcination temperature from 500 to 750 °C led to a noticeable decrease in the catalytic activity while increasing the calcination temperature to 1000 °C resulted in the quenching of the catalytic activity. For example, the N2O decomposition conversion (reaction at 500 °C) was decreased from 100% to 96.5% and 2.2% on raising the calcination temperature from 500 to 750 and 1000 °C, respectively. The T50 (temperature of 50% conversion) showed a 125 °C increase on increasing the calcination temperature from 500 to 750 °C. The characterization results showed that increasing the calcination temperature to 750 °C was accompanied by the persistence of the composition of the 500 °C calcination, i.e., the spinel phase was the major phase together with traces of NiO. On the other hand, it was found that a further calcination temperature rise to 1000 °C led to the disappearance of the composition of the 750 °C calcination, and the formation of a mixture of CoO and rock salt (Ni, Co)O. Thus, one can state that the observed dramatic activity decrease from the calcination temperature increase from 750 to 1000 °C was due to the decomposition of the spinel structure. In other words, it was due to the disappearance of the Co3+-Co2+ redox couple necessary for N2O decomposition. It was shown that increasing the calcination temperature from 500 to 1000 °C led to a sharp surface area decrease. This effect is directly due to sintering which dominated at high temperatures. Thus, it is plausible to relate the observed activity decrease from the pre-treatment temperature increase to the surface area decrease.
Structural analysis of the Ni/Co oxide system revealed the following points: (1) a solid state reaction between NiO and Co3O4 led to the formation of nickel cobaltite; (2) increasing Ni content led to crystallite size decrease until the catalyst with x = 0.75 which exhibited the smallest crystallite size in the 500 °C calcined series; (3) increasing the calcination temperature for this catalyst to 1000 °C led to changes in the phase composition accompanied by the formation of CoO as a product; in other words, a high temperature treatment facilitated Co3+ → Co2+ reduction; (4) residual amounts of K+ still remained in the samples, which improved the N2O decomposition activity of this catalysts series. Catalytic activity measurements for N2O decomposition demonstrated the following: (1) addition of nickel oxide significantly enhanced the catalytic activity of Co3O4 spinel oxide; (2) the high N2O decomposition reactivity over Ni/Co mixtures compared to their individual oxides was due to several parameters which operate simultaneously, viz., role of nickel cations in enhancing the Co3+ → Co2+ reduction, smaller crystallite size, larger SBET value, and the presence of residual K+; (3) increasing the calcination temperature from 500 to 750 °C led to a decrease in catalytic activity but resulted in a dramatic activity decrease at 1000 °C, which was attributed to the decomposition of the spinel structure.
The authors gratefully acknowledge the Deutscher Akademischer Austausch Dienst (DAAD) for the use of gas analyzers for these experiments.
2014, Vol. 35 Issue (7): 1105-1112
PDF (454KB)
2014, Vol. 35 Issue (7): 1105-1112
PDF (454KB)