The nanoplate-aggregate microspherical Co₃O₄ sample was prepared using the ethylenediamine-assisted hydrothermal method. Co(NO₃)₂×6H₂O (2.91 g) was dissolved in 30 ml deionized water,and 9 ml NH₃×H₂O aqueous solution (28 wt% NH₃×H₂O:H₂O = 1:3,v/v) was then added dropwise to the Co(NO₃)₂ aqueous solution. After stirring for 0.5 h,0.5,1.0 or 2.0 ml ethylenediamine was added under stirring for 0.5 h. The mixed solution was transferred to a 50-ml Teflon-lined stain­less steel autoclave,which was sealed and heated at 120,140,160,or 200 °C for 12 h. After the autoclave was cooled to room temperature (RT),the resulting precipitate was filtered out,washed with deionized water and absolute ethanol three times,dried in an oven at 60 °C for 24 h,and ground uniformly. Finally,the powder was calcined in air using a ramp of 1 °C min⁻¹ from RT to 500 °C and maintained at this temperature for 4 h,thus generating the Co₃O₄ samples. The Co₃O₄ sample obtained with 1.0 ml ethylenediamine in the hydrothermal treatment at 140 °C for 12 h was denoted as Co₃O₄ microsphere. The Co₃O₄ samples obtained with 0,0.5,and 2.0 ml ethylenediamine in the hydrothermal treatment at 140 °C for 12 h were denoted as Co₃O₄-1,Co₃O₄-2,and Co₃O₄-3,respectively. The Co₃O₄ samples obtained with 1.0 ml ethylenediamine in the hydrothermal treatment at 120,160,and 200 °C for 12 h were denoted as Co₃O₄-4,Co₃O₄-5,and Co₃O₄-6,respectively. For comparison,a bulk Co₃O₄ (Co₃O₄-bulk) sample was prepared by thermally decomposing Co(NO₃)₂ in air at 600 °C for 3 h. All of the chemicals (AR) were purchased from Beijing Chemical Reagent Company and used without further purification.

X-ray diffraction (XRD) patterns of the samples were recorded on a Bruker D8 Advance diffractometer with Cu K_α radiation and nickel filter (l = 0.15406 nm). The scanning electron microscopic (SEM) images of the samples were recorded on a Gemini Zeiss Supra 55 apparatus (operating at 10 kV). BET (Brunauer-Emmett-Teller) surface areas of the samples were measured by N₂ adsorption at -196 °C on a Micromeritics ASAP 2020 analyzer. The samples were outgassed at 250 °C for 2.5 h under vacuum before measurement. X-ray photoelectron spectroscopy (XPS,VG CLAM 4 MCD analyzer) was used to determine the Co 2p,O 1s,and C 1s binding energies (BEs) of surface species using Mg K_a (hv = 1253.6 eV) as the excitation source.

Hydrogen temperature-programmed reduction (H₂-TPR) experiments were carried out on a chemical adsorption analyzer (Autochem II 2920,Micromeritics). The sample (40-60 mesh,0.025 g) was loaded to a fixed-bed U-shaped quartz microreactor (i.d. = 4 mm) and pretreated in a O₂ flow (30 ml min⁻¹) at 500 °C for 1 h. After cooling in the same atmosphere to RT,a He flow (30 ml/min) was passed to purge the sample for 15 min. Finally,the pretreated sample was exposed to 5vol% H₂/Ar mixture (50 ml min⁻¹) and heated at a ramp of 10 °C min⁻¹ from RT to 550 °C. The H₂ concentration of the effluent was monitored online by a chemical adsorption apparatus. The reduction peak was calibrated against that of the complete reduction of well-characterized CuO powders (Aldrich,99.995%).

A continuous flow fixed-bed quartz microreactor (i.d. = 4 mm) was used to determine catalyst activity for the oxidation of toluene. To minimize the effect of hot spots,0.6 g quartz sands (40-60 mesh) was used to dilute the sample (0.1 g). The total flow rate of the reactant mixture (0.1% toluene + O₂ + N₂ (balance)) was 33.3 ml min⁻¹,giving a toluene/O₂ molar ratio of 1/400 and a space velocity (SV) of 20000 ml g⁻¹ h⁻¹. The toluene was obtained by passing a N₂ flow through a bottle containing pure toluene (AR) chilled in an ice-water isothermal bath. the SV was changed through the amount of sample. In the case of water vapor addition,3.0 vol% H₂O was introduced by passing the feed stream through a water saturator at RT. Reactants and products were analyzed online by a gas chromatograph (GC-2010,Shimadzu) equipped with a flame ionization detector (FID) and a thermal conductivity detector (TCD) using a stabilwax@-DA column (30 m in length) for VOC separation and a 1/8 in Carboxen 1000 column (3 m in length) for permanent gas separation. The carbon balance throughout the investigation was estimated to be 99.5%.

Figure 1 shows the XRD patterns of the Co₃O₄ samples. The two samples possessed a cubic Co₃O₄ crystal structure with well-indexed planes (JCPDS PDF ^#74-1657). The difference in diffraction intensity reflected the discrepancy in crystallinity. As can be seen from Fig. 2,the particles of the Co₃O₄ bulk sample were irregular polyhedra. The Co₃O₄ sample derived from the ethylenediamine-assisted hydrothermal route was composed of microspherical entities as several layers of a number of nanoplates (Fig. 2(b)-(d)). With a change in the preparation conditions (added amount of ethylenediamine or hydrothermal temperature),the morphology of the Co₃O₄ samples obtained was irregular (Fig. 3). Therefore,we investigated in detail only the Co₃O₄ bulk and Co₃O₄ microsphere samples. The surface areas of the Co₃O₄ bulk and Co₃O₄ microsphere samples were 9.8 and 66.2 m² g⁻¹,respectively,whereas the Co₃O₄-i (i = 1-6) samples showed a surface area of 33.2,43.5,50.1,39.3,54.2,and 26.9 m² g⁻¹,respectively.

Fig. 1. XRD patterns of Co3O4 bulk (1) and Co3O4 microspheres (2).

Fig. 2. SEM images of Co3O4 bulk (a) and Co3O4 microspheres (b-d).

Fig. 3. SEM images of Co3O4-1 (a),Co3O4-2 (b),Co3O4-3 (c),Co3O4-4 (d),Co3O4-5 (e),and Co3O4-6 (f).

Figure 4(a) illustrates the Co 2p_3/2 XPS spectra of the samples. The Co 2p_3/2 signal at 780.0 eV showed the presence of surface Co³⁺ species,while the Co 2p_3/2 signal at 781.9 eV together with the shake-up satellite at 785.5 eV confirms the presence of surface Co²⁺ species ^{[ 20 ]}. It can be observed from Table 1 that the surface Co³⁺/Co²⁺ molar ratio (1.44) of the Co₃O₄ microsphere was lower than that (1.65) of Co₃O₄ bulk,indicating a higher surface Co²⁺ concentration in the former. That is,the Co₃O₄ microsphere sample possessed more oxygen vacancies than the bulk counterpart. It is well known that the adsorbed oxygen species concentration is associated with the oxygen vacancy density. For an oxygen-deficient material,more oxygen vacancies give a higher oxygen adspecies concentration. This was substantiated by the O 1s XPS result. As shown in Fig. 4(b),an asymmetrical O 1s XPS signal was recorded for each sample,which was deconvoluted into the peaks of several kinds of surface oxygen species. The oxygen adspecies at 529.3,530.3,531.8,and 533.4 eV were assignable to the surface lattice oxygen (O_latt),adsorbed oxygen (O_ads,e.g.,O₂^-,O₂^2-,or O^-),hydroxyl and/or carbonate species,and adsorbed molecular water ^{[ 21, 22 ]},respectively. Compared to the Co₃O₄ b ulk sample,the O_ads/O_latt molar ratio of the microspherical Co₃O₄ sample was much higher,indicating a higher O_ads concentration.

Fig. 4. Co 2p3/2 (a) and O 1s (b) XPS spectra of Co3O4 bulk (1) and Co3O4 microspheres (2).

Table 1 Surface element compositions,H2 consumption,and catalytic activities of the Co3O4 bulk and Co3O4 microsphere samples. Sample Surface element composition H2 consumption (mmol g−1) Catalytic activity Co3+/Co2+ molar ratio Oads/Olatt molar ratio £ 250 °C > 250 °C T10% (°C) T50% (°C) T90% (°C) Co3O4 bulk 1.65 0.38 3.22 12.86 230 273 304 Co3O4 microsphere 1.44 1.03 3.10 12.38 190 230 254 Reaction conditions: toluene concentration = 0.1%,toluene/O2 molar ratio = 1/400,SV = 20000 ml g−1 h−1.

Table 1 Surface element compositions,H2 consumption,and catalytic activities of the Co3O4 bulk and Co3O4 microsphere samples.

Figure 5(a) illustrates the H₂-TPR profiles of the Co₃O₄ samples. The reduction of Co₃O₄ proceeds via the sequence of Co₃O₄→CoO→Co^{0 [ 9 ]}. The first peak at 227-230 °C was due to the reduction of Co³⁺ to Co²⁺ and the removal of oxygen adspecies. The reduction peaks at 279-292 and 316-341 °C corresponded to the reduction of Co²⁺ to Co^{0 [ 23 ]}. If the cobalt ions in Co₃O₄ were only Co³⁺ and only Co²⁺ and reduced to Co⁰,the H₂ consumption would be 18.07 and 13.33 mmol g⁻¹,respectively. Actually,the H₂ consumptions of the Co₃O₄ bulk and Co₃O₄ microsphere samples were 16.08 and 15.48 mmol g⁻¹,respectively. Furthermore,the low-temperature H₂ consumption (3.22 mmol g⁻¹) of the Co₃O₄ bulk sample was higher than that (3.10 mmol g⁻¹) of the Co₃O₄ microsphere sample,indicating that there was more Co³⁺ in Co₃O₄ bulk (i.e.,the Co₃O₄ microsphere sample possessed a higher oxygen vacancy concentration and hence a higher oxygen adspecies concen­tration). Therefore,the cobalt species in the Co₃O₄ samples were present in a mixed valence (Co ³⁺ and Co²⁺),which was in good agreement with the XPS results.

Fig. 5. H2-TPR profiles (a) and initial H2 consumption rate as a function of inverse temperature (b) of Co3O4 bulk (1) and Co3O4 microspheres (2).

It is better to evaluate the low-temperature reducibility of a sample using the initial (where less than 25% oxygen in the sample was removed for the first reduction peak) H₂ consumption rate ^{[ 24 ]}. Figure 5(b) shows the initial H₂ consumption rate as a function of inverse temperature of the Co₃O₄ samples. Obviously,the initial H₂ consumption rate of the nanoplate-aggregate Co₃O₄ microsphere sample was higher than that of the Co₃O₄ bulk sample. This difference in low-tem­perature reducibility was in consistent with the difference in their O_ads concentrations and catalytic performance.

In the quartz sand blank experiment,no conversion of toluene was observed below 400 °C,indicating that under the adopted conditions there was no occurrence of homogeneous reactions. Figure 6(a) shows catalytic activity of the samples for toluene combustion. It can be clearly seen that the microspherical Co₃O₄ sample outperformed the bulk counterpart,with the T_10%,T_50%,and T_90% (temperatures required for toluene conversion of 10%,50%,and 90%,respectively) of 190,230,and 254 °C at SV = 20000 ml g⁻¹ h⁻¹ (Table 1),respectively. Toluene was completely oxidized to CO₂ and H₂O over the Co₃O₄ samples,and there were no incomplete oxidation products,as substantiated by the carbon balance (99.5%) in each run. To better evaluate the catalytic activity of the samples,we use the specific reaction rates (normalized by weight amount and surface area of the sample) versus reaction temperature,as shown in Fig. 6(b) and (c),respectively. It is observed that the changing trend in specific reaction rate normalized by weight amount of the sample versus temperature was rather similar to that for toluene conversion versus temperature,but different from the specific reaction rate normalized by surface area of the sample versus temperature. This result indicated that the surface area had an important influence on the catalytic act ivity of the sample.

Fig. 6. Toluene conversion (a),specific reaction rate normalized by catalyst amount (b),and specific reaction rate normalized by the surface area of the catalyst (c) as a function of temperature over Co3O4 bulk (1) and Co3O4 microspheres (2). Reaction conditions: 0.1% toluene,toluene/O2 molar ratio = 1/400,SV = 20000 ml g−1 h−1.

To examine the effect of water vapor on the catalytic performance,we conducted toluene oxidation in the presence of 3.0 vol% water vapor over the microspherical Co₃O₄ sample. As shown in Fig. 7,the addition of water vapor at 260 °C did not lead to a significant decrease in toluene conversion,but introduction of water vapor at 220 °C induced a drop in toluene conversion by 4%. This result was possibly due to the competitive adsorption of water,toluene,and oxygen. When the water vapor was cut off,toluene conversion at 220 °C was restored to almost the initial value in the absence of water vapor. Therefore,the deactivation due to water vapor introd­uction was reversible.

Fig. 7. The effect of 3.0 vol% water vapor on the catalytic activity at 220 °C (1) and 260 °C (2) over Co3O4 microspheres. Reaction conditions: 0.1% toluene,toluene/O2 molar ratio = 1/400,SV = 20000 ml g−1 h−1.

Figure 8 shows the effect of SV on the catalytic activity of the Co₃O₄ microsphere sample. As expected,the toluene conversion increased with the drop in SV. In order to examine the catalytic stability,we carried out the 100 h on-stream reaction experiment over the Co₃O₄ microsphere sample at 250 °C and 20000 ml g⁻¹ h⁻¹. As shown in Fig. 9,no significant loss in activity was observed. This result demonstrated that the microspherical Co₃O₄ sample was stable under the adopted conditions.

Fig. 8. Toluene conversion as a function of reaction temperature at different SV values over the nanoplate-aggregate microspherical Co3O4 catalyst. Reaction conditions: 0.1% toluene,toluene/O2 molar ratio = 1/400.

Fig. 9. Toluene conversion as a function of on-stream reaction time over the nanoplate-aggregate microspherical Co3O4 sample. Reaction conditions: 0.1% toluene,toluene/O2 molar ratio = 1/400,SV = 20000 ml g−1 h−1,250 °C.

In the past years,many materials have been used as the catalyst for toluene combustion. The activities of these catalysts are summarized in Table 2. Usually,the specific reaction rate normalized by the sample amount is used to compare the catalytic activity of different samples. At a given reaction temperature (150 or 250 °C),the specific reaction rate (0.0376 mmol g⁻¹ h⁻¹ at 150 °C or 0.711 mmol g⁻¹ h⁻¹ at 250 °C) over the Co₃O₄ microshperes was much higher than those over Mn₃O₄ ^{[ 25 ]},a-Mn₂O₃ ^{[ 25 ]},b-MnO₂ ^{[ 25 ]},bulk Co₃O₄ ^{[ 17 ]},LaMnO₃ ^{[ 26 ]},and 5 wt% Au/CeO₂ ^{[ 27 ]},but lower than those over Co₃O₄-HT-CTAB ^{[ 2 ]},mesoporous Co₃O₄ ^{[ 17, 18 ]},and 0.5 wt% Pd/LaMnO₃ ^{[ 28 ]}.

Table 2 Comparison on the activity of various catalysts for toluene combustion. Catalyst Reaction condition Specific reaction rate (mmol g−1 h−1) Ref. Toluene concentration (%) SV (ml g−1 h−1) At 150 °C At 250 °C Co3O4 microspheres 0.1 20000 0.0376 0.711 This work Co3O4-HT-CTAB 0.1 20000 0.164 0.818 [ 2 ] Bulk Co3O4 0.1 20000 0.0164 0.246 [ 17 ] Mesoporous Co3O4 0.1 20000 0.246 0.818 [ 17 ] Mesoporous Co3O4 0.1 20000 0.778 0.818 [ 18 ] Mn3O4 0.1 15000 0 0.522 [ 25 ] a-Mn2O3 0.1 15000 0 0.00614 [ 25 ] b-MnO2 0.1 15000 0 0 [ 25 ] LaMnO3 0.19 178 h−1 0.000277 0.00554 [ 26 ] 5 wt% Au/CeO2 0.7 186 0 0.00533 [ 27 ] 0.5 wt% Pd/LaMnO3 0.1328 18000 h−1 0.293 0.880 [ 28 ]

Table 2 Comparison on the activity of various catalysts for toluene combustion.

It is well established that the oxidation of organics over a transition metal oxide involves a Mars-van Krevelen mechanism,where organic molecules are oxidized by the lattice oxygen of the metal oxide,and the partially reduced metal oxide is then re-oxidized by the gas phase oxygen ^{[ 29, 30, 31 ]}. Wang et al. ^{[ 31 ]} attributed the higher reactivity of manganese oxide to the improvement in oxygen activation ability,which enhanced the combustion of ethanol and acetaldehyde. In the case of Co₃O₄,lattice oxygen could be consumed by reaction with toluene and then replenished by gas phase oxygen ^{[ 32, 33 ]}. Thus,the lattice oxygen of cobalt oxide is of great importance in toluene oxidation. The low-temperature reducibility reflects the reactivity of the lattice oxygen in Co₃O₄. As shown in Fig. 5,the lattice oxygen reactivity (i.e.,low-temperature reducibility) of the Co₃O₄ microspheres was higher than that of Co₃O₄ bulk. Surface oxygen vacancies also have an important role to play ^{[ 34, 35 ]}. The presence of oxygen vacancies favors the activation of oxygen molecules to active oxygen adspecies. The microspherical Co₃O₄ sample possessed a higher oxygen adspecies concentration (which is related to the surface oxygen vacancy density) than the Co₃O₄ bulk sample,in good agreement with their catalytic activity sequence. In addition,the surface area (66 m² g⁻¹) of nanoplate- aggregate Co₃O₄ microspheres was much higher than that (10 m² g⁻¹) of Co₃O₄ bulk. Therefore,we conclude that the large surface area,high oxygen adspecies concentration,and good low-tem­perature reducibility accounted for the good catalytic performan­ce of the nanoplate-aggregate Co₃O₄ microspheres for toluene combustion.