CO emitted from industrial activities and transportation is a toxic gas for humans and animals [1, 2]. Among all the possible solutions for reducing CO emissions, catalytic conversion of CO into non-toxic gases has been a hot topic [2-6]. Generally, supported noble metals have excellent activity for low-temperature CO oxidation, but the high cost because of their limited global resources inhibits their practically permanent application [1, 7]. The design of alternative and cheaper catalysts for CO catalytic oxidation has therefore received much attention [8-10].
With the rapid progress in the preparation technology for nanomaterials, nanocatalysts have received wide interest because of the specificity of their nanostructures (e.g., morphologies and size effects) [11]. Nanoscale effects can cause some non-noble catalysts to have activity in CO oxidation comparable with that of noble metal catalysts. Many studies indicate that Co3O4 has remarkable activity for low-temperature CO oxidation [12-20]. It is also clear that the morphology dependence of Co3O4 nanocatalysts is a greater influence than its size dependence [21-23]. Rational design and skillful synthesis of Co3O4 catalysts with various morphologies has become one important path to improve its catalytic activity.
Generally, the synthesis of Co3O4 nanostructures can be achieved through the thermal decomposition of precursors at elevated temperatures. With such a procedure, the shape of the Co3O4 materials is largely dependent on the structure of the precursors (e.g., cobalt-hydroxide-carbonate) [24]. Several methods, such as sol-gel processing, microemulsion, forced hydrolysis, hydrothermal synthesis, and chemical precipitation have been developed for the fabrication of cobalt-hydroxide-carbonate nanoparticles [25-30]. Among these methods, the hydrothermal method offers functional control over the size and morphology of the particles with relatively simple steps and short reaction time, providing well-crystallized reaction products with high homogeneity [27, 28].
Many cobalt-hydroxide-carbonate nanostructures with varied morphologies have been synthesized by a hydrothermal approach, including nanorods, nanotubes, nanobelts, nanospheres, nanosheets, and hollow core/shell hierarchical nanostructures [4, 22, 29, 30]. In particular, Hu et al. [4] prepared nanocube and nanobelt cobalt-hydroxide-carbonate precursors via a hydrothermal process by varying the annexing agent (octadecenylamine and glycerol, respectively) in the mixture and controlling the reaction temperature. Xu et al. [22] reported an investigation of the formation of cobalt-basic-carbonate compounds [Co(OH)x(CO3)0.5(2-x)·nH2O] with dimensional and morphological control using different cobalt sources (CoCl2 and Co(NO3)2). They found that samples synthesized by urea hydrolysis consisted of large well-crystallized nanorods, and the counteranion effect was significant in the experiment. Samples obtained in Co(NO3)2 solution have uniform diameters of approximately 50 nm and lengths of 1-2 µm, whereas samples from CoCl2 solution show needle-like crystallites with a much larger size (100-200 nm in diameter, and 1-5 µm in length). Similarly, Wang et al. [29] synthesized Co(CO3)0.5(OH)·0.11H2O precursor nanowires by a facile hydrothermal method and showed how the additive (urea) concentration controlled the uniformity and the overall structure. The carbonate ions produced via the hydrolysis of urea not only participate in the formation of the intermediate compound and facilitate the anisotropic growth of the nanowires, but also behave like coordinating agents that help to organize and give rise to a more complex hierarchical structure at high concentrations. Sarawade et al. [30] also demonstrated that the type of surfactant template and the reaction temperature played crucial roles in controlling the size, shape, and morphology of the products. It is also clear that Co3O4 nanorods with preferentially exposed {110} planes are extremely active and stable for low-temperature CO oxidation [3, 31, 32]. The Co3+ sites on {110} planes over Co3O4 nanobelts have much higher turnover frequency (TOF) than those on {100} planes over Co3O4 nanocubes [4]. These results demonstrate that structure-dependent effects play a very important role in determining the catalytic activity of Co3O4 for CO oxidation.
Almost all of these studies use various methods to acquire different products. For a hydrogen thermal process, researchers studied only the effect of limited factors (e.g., one or two experimental parameters) on the formation of cobalt-hydroxide-carbonate with different morphologies. Because the relationship between the control parameters and product morphology is multifactorial in a hydrothermal process, there is still a lack of knowledge on the growth mechanism of cobalt-hydroxide-carbonate particles under various conditions. In the present work, however, considering the influence of multiple factors and their interactions, a comprehensive hydrothermal process map involving the roles of cobalt source, temperature, time, and surfactant in the growth of cobalt-hydroxide-carbonate particles is presented. It will provide insight into the nature of the formation of cobalt hydroxide carbonate nanoparticles with various morphologies (i.e., nanorods, nanosheets, and nanocubes). The corresponding Co3O4 particles were obtained by thermal decomposition of the resultant cobalt-hydroxide-carbonate precursor. Meanwhile, the physicochemical properties of these materials were characterized, and their catalytic performance for CO oxidation was also investigated.
Cobalt-hydroxide-carbonate particles with various morphologies were synthesized by a hydrothermal method [4, 27, 33]. All of the chemical reagents used in the experiment were of analytical grade and used without further purification. In a typical procedure, 0.006 mol of CoCl2·6H2O or Co(NO3)2·6H2O, 0.024 mol of urea, and 1.5 g of CTAB were dissolved in deionized water (150 mL) at room temperature with magnetic stirring. The mixture was stirred vigorously for at least 30 min to form a transparent pink solution, which was then transferred into a Teflon®-lined autoclave. The autoclave was sealed and inserted into a temperature-controlled furnace at the set reaction temperature. The controlling parameters reaction temperature (95-200 ℃), time (4-24 h), and surfactant amount (0 or 1.5 g of CTAB) were varied in a systematic fashion, as summarized in Table 1. The obtained precipitates were collected and washed with distilled water and anhydrous alcohol for several times until no Cl− could be detected by silver nitrate titration. After that, they were dried at 100 ℃ for 24 h.
The Co3O4 was readily obtained by thermal decomposition of the resultant cobalt-hydroxide-carbonate precursor. The products obtained in Section 2.1 were calcined in air at 400 ℃ for 4 h (heating rate was 2 ℃/min) to produce Co3O4 nanoparticles.
The X-ray powder diffraction (XRD) data for the prepared samples were collected on a Rigaku D/max-ШB X-ray diffractometer at a scanning rate of 2°/min with 2θ ranging from 10° to 90° (Cu Kα radiation λ = 0.15406 nm).
The thermogravimetric analysis (TGA-DTA) measurements were performed with the aid of an STA-449-F3 instrument (produced by Netzsch Co.) using a linear heating rate of 10 ℃/min and a sample mass of 9.5 mg in air flow.
The morphology of the samples was observed using a scanning electron microscopy (SEM, Hitachi S4800) instrument with an accelerating voltage of 3 kV. Samples for SEM were dusted on an adhesive conductive carbon belt attached to a copper disk and were coated with 10 nm of Au prior to measurement. Transmission electron microscopy (TEM) images of the nanoparticles were taken on a JEOL2010 transmission electron microscope.
The specific surface area of as-synthesized Co3O4 samples was calculated according to the Brunauer-Emmett-Teller (BET) method by the N2 adsorption isotherm at −196 ℃, using a Quantachrome Autosorb-iQ instrument.
X-ray photoelectron spectroscopy (XPS) studies were carried out on a PHI-5500 system equipped with a monochromatic Mg Kα (1253.6 eV) X-ray source. Spectra were registered at ambient temperature in a vacuum (residual pressure < 10-7 Pa). Sample charging during the measurement was compensated by an electron flood gun. The standard deviation in the binding energy values of the XPS line is 0.10 eV. Binding energy (BE) was determined by computer fitting of the measured spectra and calibrated using the C 1s peak (284.8 eV) as a reference. The surface composition and chemical state were evaluated based on the areas and binding energies of the Co 2p and O 1s photoelectron peaks.
Temperature-programmed reduction (TPR) experiments were performed on TPR Winv1.50 (produced by Quantachrome Instruments Co.) under a flow of a 10% H2/He mixture (25 mL/min) using 30 mg of catalyst at a heating rate of 10 ℃/min.
Temperature-programmed desorption (TPD) experiments were performed on the same instrument as the TPR experiments. Each sample was first pretreated in He (99.99% purity) at 400 ℃ for 1 h. After the O2 adsorption in pure O2 at 80 ℃ for 50 min and subsequent cooling to room temperature, pure He was flowed through the reactor for 1 h. Then, desorption profiles of O2 were obtained by ramping the reactor to 900 ℃ at a heating rate of 10 ℃/min in a He stream (30 mL/min).
The evaluation of the activity for CO oxidation was performed in a fixed-bed quartz tubular reactor at atmospheric pressure under steady-state conditions. Prior to testing, the samples (200 mg, 40-60 mesh) were treated in argon (200 mL/min) at ambient temperature for 20 min. The reactant gases (1% CO, 10% O2, and the balance Ar) were passed continuously through the sample bed at flow rates of 200, 300, or 400 mL/min (giving a gas hourly space velocity (GHSV) of 33000, 50000, or 66000 h-1, respectively). The reactor was heated through a temperature-programmed route at a heating rate of 5 ℃/min. The amounts of CO, CO2, and O2 in the outlet streams were measured on-line by a gas chromatograph.
The kinetic experiment was conducted to investigate the intrinsic activity of the catalysts, which was defined as a reasonable correlation with the active site density of catalysts. The reaction rate of CO oxidation was determined by an isothermal reaction at 65 ℃ in the kinetic regime, in which the influences of internal and external mass and heat transfer were all excluded. The reaction gas composition and total flow rate were the same as for the aforesaid CO oxidation reaction.
The conversion of CO (X) was calculated from the change in CO concentration between the inlet gases and outlet gases (Eq. (1)). For comparison of catalytic activity, the parameters reaction rate (r, mol g-1 s-1) and TOF (s-1) were calculated at a specified temperature of 65 ℃. Here, CCO is the concentration of CO; C0 (mol/s) is the initial CO concentration per second; m is the mass of catalyst; and n is the molar mass of catalyst.
The stability of the catalysts was tested under dry conditions and in the presence of moisture. In the case of water vapor introduction, 0.5 vol% H2O was introduced by passing the feed stream through a water saturator at a certain temperature.
Cobalt-hydroxide-carbonate nanoparticles with various morphologies and sizes were prepared by altering experimental conditions and parameters. To investigate the growth mechanism of the Co3O4 nanocatalysts, a series of contrasting experiments were conducted. SEM and TEM techniques were used to monitor the achievement of the particular shape and to enable a comparison with the controlled morphology of cobalt-hydroxide-carbonate precursor.
Fig. 1 shows the morphology of as-prepared cobalt-hydroxide-carbonate nanoparticles using two kinds of cobalt source (Co(NO3)2 and CoCl2) in the presence of surfactant (CTAB). Two different morphologies (nanorods and nanosheets) can be observed. The sample synthesized using CoCl2 as the cobalt source was composed of a large amount of rods, with average diameters of 40 to 80 nm and lengths of up to several micrometers (Fig. 1(a) and (b)). Fig. 1(c) and (d) display the SEM and TEM images of the sample prepared from Co(NO3)2. The obtained cobalt-hydroxide-carbonate is in the form of well-defined rectangular nanosheets, 6-8 µm long and 2-3 µm wide. This indicates that the morphologies of the samples were completely discrepant when different cobalt sources were introduced into the hydrothermal reaction at low temperature.
This result is similar to the observation by Xu et al. [22]. They also found that the crystallinity of cobalt precipitates was affected by the type of counteranion. When Co(NO3)2 was dissolved into water with urea (NH2CO), cobalt carbonate hydroxide hydrate was formed initially by slow hydrolysis of the urea precipitant. With increasing temperature, the NH4+ formation reaction consumes the proton, which makes the pH of the overall solution system over 7.0. The cobalt ions were easily precipitated as cobalt hydroxide in the aqueous solution. Subsequently, a brucite-type cobalt double hydroxide with a layered structure can be formed, suggested by the appearance of divalent metal cation hydroxides [34-36]. Moreover, anions such as NO3- could be intercalated into the interlamellar spaces, causing the interlayer distance to increase [35, 37, 38]. This is beneficial for the formation of cobalt-hydroxide-carbonate with a nanosheet morphology.
The NO3- and Cl- ions can also affect the morphology by controlling the growth orientation of the particles. In a typical cobalt-hydroxide-carbonate synthesis process, more carbonate anions and hydroxide anions are released from urea as the hydrothermal reaction proceeds. As the concentration of building blocks (carbonate anions and hydroxide anions) becomes sufficiently high, they tend to aggregate into nuclei through homogeneous nucleation. With a continuous supply of the building blocks, these nuclei can serve as seeds for further growth to form larger structures. It should be stressed that the growth of Co(NO3)2 seeds is different from the growth of CoCl2 seeds. In the growth of CoCl2 seeds, the carbonate anions may act as an inhibitor that selectively decreases the rates of crystal growth along both the {001} and {100} directions, resulting in {010}-elongated nanorods [3, 19, 29, 32]. However, in the growth of Co(NO3)2 seeds, the existence of nitrate anions may affect the location of carbonate anions and discount the effect of the inhibitor because of steric hindrance [37]. This would result in inhibition of growth along the {012} direction and the acceleration of the growth along the {110} direction, leading to the formation of nanosheets.
Temperature-resolved study of cobalt-hydroxide-carbonate intermediates was performed with respect to the individual reaction parameters. The TEM images in Fig. 2 illustrate the morphologies of the materials grown in the absence of CTAB at temperatures ranging from 95 to 200 ℃ (Samples S(Cl)1/S(N)1, S(Cl)2/S(N)2, S(Cl)3/S(N)4, S(Cl)5/S(N)5, S(Cl)7/S(N)7) with a reaction time of 16 h. As shown in Fig. 2(a)-(c), the as-synthesized materials presented rod-like structures with average diameters of 20 to 120 nm and lengths of up to 3 µm. In addition, the top of the nanorod turned from pointed to round as the hydrothermal temperature increased from 95 to 120 ℃. The nanosheet particles revealed a similar evolution with increasing reaction temperature. The anomalous sheets generally grew up and trended to become rectangular (Fig. 2(d)-(f)). However, when the temperature increased to 160 ℃, using either Co(NO3)2 or CoCl2 as the cobalt source, nanocube-like morphology was observed, as shown in Fig. 2(g) and (h). With a further increase in the temperature, the former morphologies (rod or sheet) vanished, the nanocubes gradually became larger, and the morphology tended to mature. All of the samples became nanocubes when the temperature increased to 200 ℃. These results indicate that the reaction temperature only affects the size of the samples when it is lower than 140 ℃. With increasing hydrothermal temperature, the cobalt-hydroxide-carbonate rod or sheet slightly grows from primary to mature. However, when the hydrothermal temperature is high enough (160-200 ℃), the temperature is a determining factor in the control of the formation of nanocube particles.
Generally, in the homogeneous precipitation of the hydrothermal reaction, the urea hydrolysis provides both carbonate and hydroxyl anions to react with cobalt cations. The formation of orthorhombic cobalt hydroxide carbonate can be formulated as follows [39]: Co2+ + xOH- + 0.5CO32- + 0.11H2O ⇋ Co(OH)x(CO3)0.5·0.11H2O. As the temperature rises, the reaction can be expressed as follows: 2Co2+ + 2OH- + CO32- ⇌ Co2(OH)2CO3.The control of the kinetics by the temperature and the concentration of reagents, which control the hydrolysis rate and ratio in the reaction, thus modulate the nucleation and growth processes. At higher temperatures, the crystal shape anisotropy and the growth orientation of cobalt-hydroxide- carbonate began to dominate.
Wang et al. [40] changed the reaction temperature in a certain range (95-120 ℃) to observe the changes in the morphology of synthesized cobalt-hydroxide-carbonate nanorods using CoCl2 and urea as original materials. The results indicate that the temperature only affects the size and length. As revealed by Li et al. [41], only using cobalt salts and urea, needle-like single-crystalline nanorods with the growth direction along the {010} direction were synthesized at a reaction temperature of 100 ℃. As the reaction temperature increased from 140 to 180 ℃, cobalt-hydroxide-carbonate nanorod grew epitaxially to a 3D architecture, which was quite different from the samples obtained at 100 and 140 ℃.
The TEM images suggest that layer-by-layer growth from small primary nanoparticles to the final cubicstructure occurred as the temperature increased. It is generally accepted that, at high temperature, a sudden burst occurs to the nuclei and the nuclei subsequently grow into larger crystallites by an aggregation mechanism [42]. The decomposition rate of the urea in the solution increased with increasing temperature, resulting in the combination of ions and rapid growth. Consequently, the synthesis temperature dramatically affects the crystallite morphology. The results also suggest that the optimal temperature of the products can be determined by the specific shapes.
Fig. 3 shows the variation in the morphology and size of as-synthesized samples as the reaction time increased from 4 to 24 h. In this process, the morphologies of the samples remain the same, while the size of all nanoparticles tends to increase. The diameters of rod-like nanoparticles increase to approximately 180 nm as the reaction time increases to 24 h. Also, the sizes of the sheet crystals grow to several micrometers and the thickness of plates begins to grow (Fig. 4(c) and (d)) during the hydrothermal process after 24 h. The growth of nanocubes is similar to that ofsheet crystals. There is no obvious effect on the morphology of the sample series with further increasing reaction time.
To determine the effect of the surfactant on the shape and size of the products, contrasting experiments were conducted in both the absence and presence of surfactant (CTAB). The TEM images in Fig. 4 show three typical morphologies of as-prepared cobalt-hydroxide-carbonate obtained with and without the surfactant with 16 h of reaction time at 120, 140, and 200 ℃. Well-dispersed nanoparticles were observed in the presence of CTAB surfactant. Moreover, the products were relatively uniform and well-developed. This finding suggests that nanoparticles with relatively uniform, well-developed morphology and high dispersion can be obtained in the presence of CTAB.
CTAB is an ordinary cationic surfactant and is extensively used as a “shape-inducing” agent to synthesize metal nanoparticles. CTAB can remarkably influence the dispersibility of the produced nanoparticles because of its fine surface activation performance. For metal catalysts, CTAB could enhance the dispersion of the catalyst particles because of the strong binding of CTAB molecules to the metal surface [43]. Sun et al. [44] found that the CTAB surfactant could be selectively absorbed on the crystal face of Co3O4, which finally induces the formation of nanoplates. In this synthesis process, because of the attraction between the cationic CTAB micelles and the negative ions in solution, CTAB micelles were easily absorbed on metal ions. This could enhance the distribution of the ions in the solution. In the present work, it was observed that there were no obvious changes in the morphologies of the precursor particles obtained in the absence and presence of CTAB. The presence of CTAB results in relatively uniform sizes and high dispersion of particles. In this case, the primary role of surfactant CTAB is to make particles grow in multiple directions and reduce conglomeration (dispersing agent).
Fig. 5 provides a generalized summary of the hydrothermal products as a function of cobalt source, temperature, time, and surfactant (CTAB). A comprehensive hydrothermal synthesis process map is provided for the formation of three morphologies of cobalt-hydroxide-carbonate nanoparticles from two kinds of cobaltic solution. Furthermore, the different growth mechanisms of cobalt-hydroxide-carbonate nanoparticles from the two cobalt source are also set out. Preparation along with the growth mechanism using different cobalt sources and controlling the experimental parameters (reaction temperature, reaction time, and surfactant) means that the various morphologies (nanorods, nanosheets, and nanocubes) of cobalt-hydroxide-carbonate can be accurately produced. The definite function of the experimental parameters has been stated clearly as well. CTAB additions were found to improve the dispersibility and make the products more uniform. Suitable temperature can promote the morphologies (nanosheets and nanorods) to mature, but further increasing the temperature leads to the formation of nanocubes, for both CoCl2 and Co(NO3)2 as cobalt sources. The reaction time only affects the size of the obtained nanoparticles.
By controlling the optimum synthesis conditions, cobalt-hydroxide-carbonate precursors with typical morphologies (nanosheets, nanorods, and nanocubes) were successfully synthesized. Co3O4 nanoparticles were prepared via a thermal decomposition process on the cobalt-hydroxide-carbonate precursor. The TG-DTA measurements of dried cobalt-hydroxide-carbonate precursors were performed to determine the appropriate calcination temperature. The obtained TGA-DTG curves are shown in Fig. 6. Only one weight loss step can be observed on heating the three precursors to 800 ℃, which takes place in the temperature range of 265-370 ℃. Therefore, cobalt-hydroxide-carbonate precursors were calcined at 400 ℃ in air for 4 h to obtain Co3O4 nanoparticles.
Fig. 7 shows the TEM and SAED images of various Co3O4 samples. All the samples inherit the morphology of the precursors. The lattice spacing of rod-shaped Co3O4 nanoparticles is 0.28 nm, which corresponds to the {110} crystal [3, 45, 46]. From Fig. 7(c), (f), and (i), the {110} planes, which are rich in Co3+ active sites, were exposed in all Co3O4 samples.
The obtained Co3O4 samples with three optimal morphologies were further examined by XRD (Fig. 8). The XRD pattern of each sample exhibits characteristics of the Co3O4 phase, which is perfectly in agreement with the cubic Co3O4 spinel phase (space group Fd3m; JCPDS PDF# 42-1467). No other diffraction peaks were detected, confirming the complete decomposition of the cobalt precursors after 400 ℃ calcination. For all three Co3O4 catalysts, the diffraction peaks were sharp, with high intensity, suggesting good crystallinity.
The specific surface area and pore-size distributions of as-synthesized Co3O4 samples were calculated according to the Brunauer-Emmett-Teller (BET) method by the N2 adsorption isotherm at -196 ℃, using a Quantachrome Autosorb-iQ instrument. Fig. 9 shows the representative N2 adsorption-desorption isotherms and pore-size distributions of the three samples, and their textural properties are given in Table 3. Pore size distributions were calculated using the Barrett-Joyner-Halenda (BJH) formula. The isotherms for the nanosheet and nanorod samples belong to type-II of BET's classification, with a type H3 hysteresis loop in the relative pressure (p/p0) range of 0.8-1.0, characteristic of particulate materials. These pores may originate from the interstices among the self-assembling nanoparticles, which usually present a broad pore distribution because of interparticle voids. For the nanocube sample, a small hysteresis loop in the p/p0 range of 0.6-0.8 was observed, which is related to the capillary condensation taking place in mesopores. This suggests the presence of mesopores in the nanocube sample, although the content is very limited. The specific surface area of the Co3O4 nanosheets (32.5 m2/g) is noticeably higher than those of nanorods and nanocubes (15.5 and 7.7 m2/g, respectively). The textural properties may be caused by the irregular stacking of the particles in the calcination process.
XPS analysis is a powerful means for characterizing the surface properties of catalysts. Fig. 10 presents the XPS spectra of the Co 2p and O 1s electronic levels of Co3O4 nanocubes, nanorods, and nanosheets. As shown in Fig. 10(A), two sharp peaks at 780.1 and 795.0 eV were observed in all three Co3O4 samples, which correspond to the Co 2p3/2 and Co 2p1/2spin-orbit peaks of Co3O4 spinel [5]. This result indicates that the surface of the three Co3O4 nanocatalysts comprise both Co2+ and Co3+ [47, 48]. The spectra of Co 2p3/2 can be fitted to five peaks, which can be attributed to Co3+, Co2+, Co-OH phase, and shake-up satellites of Co2+ and Co3+, respectively (as the values of binding energy peak increase from low to high) [47, 48, 49, 50, 51, 52, 53]. Table 2 shows the binding energy position of each peak and the content of the different Co species. The nanosheet sample showed the largest relative strength of the Co3+ peak among the three samples, indicating a relatively higher concentration of Co3+. For the nanocube sample, the peak attributed to the Co-OH phase disappeared.
The XPS spectra of the O 1s electronic levels of the three Co3O4 samples are shown in Fig. 10(B). Three component peaks in various positions are observed, which indicates the presence of three types of oxygen species on the surface of the Co3O4 samples. The peak with low binding energy (denoted OI) can be assigned to the atomically adsorbed oxygen on the surface of samples [54]. The signal at 530.5 eV (denoted OII) indicates the presence of lattice oxygen [55]. The third peak (denoted OIII) with low intensity can be attributed to the presence of surface hydroxyl (OH) species, which probably lead to the generation of Co-OH [53, 56]. The quantification of each surface cobalt and oxygen species was performed, as shown in Table 2. The results demonstrate that, among the three samples, Co3O4 nanocubes contain the highest concentration of surface adsorbed oxygen, while the nanosheet sample reveals the highest concentration of lattice oxygen.
Fig. 11 compares the H2-TPR and CO-TPR profiles of the three Co3O4 samples. In Fig. 11(a), all three samples exhibit two consecutive hydrogen consumption peaks, which correspond to the sequential conversion of Co3+ to Co2+ and Co2+ to Co0, respectively. For the Co3O4 nanocube and nanorod samples, similar reduction peaks of Co2+ to Co0 at 542 ℃ are observed. Nevertheless, for the nanosheet sample, the reduction temperature of Co3+ to Co2+ and Co2+ to Co0 decreases to 389 and 468 ℃, respectively. As shown in Fig. 11(b), a similar trend to the one presented in the H2-TPR was also found in the CO-TPR profiles. Two peaks can be observed for all three samples, suggesting the two-step reduction of Co3+. The changes in the reduction temperatures also confirm that the reducibility of Co3O4 nanocubes and nanosheets is superior to that of their nanorod counterparts.
By quantitatively analyzing the reduction peaks of the H2-TPR profiles, the H2 consumption for each peak can be obtained, as summarized in Table 3. Obviously, the H2 consumption of the first reduction peak increased in the order of Co3O4 nanosheets > nanocubes > nanorods. This result suggests the highest concentration of surface Co3+ sites on the Co3O4 nanosheets among the three samples, which is consistent with the results of XPS analysis. The reduction degree of Co3+ and Co2+ in the different samples was obtained by calculating the ratio of the actual H2 consumption to the theoretical values, as shown in Table 4. Apparently, 22.53% of the cobalt ions in the Co3O4 nanosheets were reduced by H2 during the H2-TPR process, and the reduction degree of Co3+ to Co2+ reached 24.80%, suggesting that Co3O4 nanosheets exhibit the best reduction performance. For the nanocubes, the reduction degree of cobalt ions reached 17.92%, which is slightly lower than that for the Co3O4 nanorods (23.66%). However, the nanocubes (17.20%) show higher reducibility of Co3+ than the nanorod samples (17.20% vs. 12.75%). This result suggests that the Co3+ in the nanocube sample is more reducible than that in the nanorod sample, which can be used to explain the relatively higher catalytic activity for CO oxidation over the nanocube sample (Fig. 13).
To gain further insight into the mobility of oxygen in the Co3O4 catalysts, O2-TPD tests were carried out (Fig. 12). The TPD profile for each catalyst showed two oxygen stripping peaks in the temperature range below 500 ℃ and higher than 850 ℃, respectively. The former peak can be attributed to desorption of surface oxygen species (e.g., adsorbed oxygen and surface lattice oxygen) of catalysts, and the latter corresponds to bulk oxygen consumption by catalysts [57]. A comparison of the tripping peaks of surface oxygen species (first peak) for all three samples indicates that Co3O4 nanosheets have more abundant surface oxygen species and higher oxygen mobility.
Fig. 13 shows the catalytic activity of the three Co3O4 samples for CO oxidation at various GHSVs. For a GHSV of 33000 h-1, the CO oxidation over the nanosheet sample took place at 47 ℃. As the temperature increased, the CO conversion rapidly increased and eventually reached 100% at 106 ℃. The CO oxidation over the Co3O4 nanocube sample began at a similar temperature, but its complete conversion temperature was delayed to 114 ℃. For the Co3O4 nanorod sample, the starting temperature for CO oxidation was as low as 130 ℃. The values of T10 (the temperature when the CO conversion equals 10%) for Co3O4 nanosheets, nanocubes, and nanorods are 65, 69, and 146 ℃, respectively. These results indicate that Co3O4 nanosheet and nanocube samples are markedly more active than the nanorod sample.
As shown in Fig. 13(b) and (c) (GHSV = 50000 and 66000 h-1, respectively), the activity of the catalysts follows the same sequence: nanosheets > nanocubes > nanorods. With an increase in GHSV, the initial reaction temperature for each sample slightly decreased. The relationships between GHSV and T50 and T90 (corresponding to CO conversion = 50% and 90%, respectively) for all the samples are summarized in Fig. 13(d). The catalytic activity of Co3O4 nanosheetsis not sensitive to the changes in GHSV, and almost no change in the T50 and T90 values was observed when theGHSV increased from 33000 to 50000 h-1. In contrast, an obvious decrease in the T50 and T90 values was detected over the Co3O4 nanocube and nanorod samples with an increase in GHSV. This indicates that CO oxidation over the Co3O4 nanosheet catalyst may not be primarily controlled by the diffusion of the reactants, which may be related to the relatively larger surface area and thinner thickness of the nanosheets. For the nanocube and nanorod catalysts, the catalytic activity decreased with increasing GHSV because of the diffusioneffect. From the perspective of T50 and T90 values, the Co3O4 nanosheet catalyst showed comparable or higher activity than the related Co3O4 catalysts (without additional pretreatments) at temperatures above room temperature under similar conditions [1, 4, 10, 14].
CO conversion over Co3O4 is achieved via Co3+ exposed active sites, which can absorb CO sufficiently [3, 58]. When the surface lattice oxygen is transferred from Co3+ to Co2+, the resulting oxygen vacancies are subsequently supplemented with reactant O2 [58]. Therefore, the formation of oxygen vacancies and the active Co3+ sites on Co3O4 nanoparticles are the rate-limiting step for determining the catalytic CO oxidation [1]. As observed by XPS, TPR, and TPD techniques, the Co3O4 nanosheets exposed many more Co3+ sites and a higher concentration of surface oxygen species. These features allow the nanosheet sample to have relatively high catalytic activity for CO oxidation. To further investigate the catalytic performance of the as-prepared catalysts, a kinetic experiment wasconducted on Co3O4 nanosheets and nanocubes at 65 ℃ under the condition of a CO conversion of no more than 10%. Moreover, the oxidation of CO in the presence of excessive oxygen (CO/O2 volume ratio = 1/10) was generally considered to obey a first-order reaction mechanism with respect to CO concentration (CCO): r = -kCCO = -A exp(-Ea/RT)CCO, where r, k, A, and Ea are the reaction rate (mol/s), rate constant (s-1), pre-exponential factor, and apparent activation energy (kJ/mol), respectively [6]. The reaction rates and their Ea over the two samples are presented in Table 3. The CO reaction rate over Co3O4 nanosheets at 65 ℃ is higher than that over nanocubes, which is in accordance with the Co3+/Co2+ molar ratio (Table 4). Moreover, the value of Co3O4 nanosheets is comparable to that of Co3O4 nanocubes with predominantly exposed {001} planes reported by Hu et al. (6.20 vs. 6.05 × 10-7 mol/g-1·s-1) [4]. This indicates that more active Co3+ species on the Co3O4 nanosheets is beneficial to catalytic oxidation. The Ea values demonstrate that oxidation of CO took place more easily over Co3O4 nanosheets than over nanocubes. For the Co3O4 nanorods, CO oxidation cannot occur at 65 ℃. As shown in Table 3, the TOF value for the nanosheet sample is 1.79 × 10-4 s-1, which is much higher than that for the nanocube sample (1.34 × 10-4 s-1), indicating more abundant active sites on the nanosheets.
Because the Co3O4 nanosheets showed the best activity for CO oxidation among the three samples, the stability of the Co3O4 nanosheet catalyst in the absence or presence of a stream (0.5 vol% water vapor) was investigated at 88 ℃. The results are shown in Fig. 14. Approximately 80% CO conversion to CO2 was observed during the initial 50 h in the absence of the stream, and no obvious decrease in the catalytic activity was observed. On the one hand, this suggests that the Co3O4 nanosheets have excellent stability for CO oxidation under dry conditions. On the other hand, the CO conversion tended to decrease when water vapor was co-introduced to the reaction system. After the water vapor flow continued for 5 h, the CO conversion decreased to 12%. However, a sharp rise in CO conversion was observed when the water vapor flow was stopped. This result indicates that Co3O4 is very sensitive to the presence of moisture, which has been reported by recent studies [3, 59, 60]. The water molecules were easily adsorbed to the surface of Co3O4 when water vapor was co-introduced to the reaction system. Subsequently, the water molecules may have reacted with CO2 to form carbonate species. With an increase in test duration, the active sites on the surface of Co3O4 were gradually covered by the formed carbonate species [61], resulting in a decrease in the catalytic performance for CO oxidation. However, there was apparent improvement in activity during the subsequent operation under dry conditions. This result suggests that the carbonate species can be gradually removed from the active sites in the absence of moisture.
Cobalt-hydroxide-carbonate with various morphologies (nanorods, nanosheets, and nanocubes) were successfully synthesized via a facile hydrothermal method. The Co3O4 nanoparticles were obtained by thermal decomposition of the cobalt-hydroxide-carbonate precursors. The samples after calcination perfectly inherited the morphology of the precursors. A “process map” for the hydrothermal synthesis of cobalt-hydroxide-carbonate with various morphologies from Co(NO3)2 or CoCl2 solution was presented via varying the experimental conditions (such as cobalt source, temperature, time, and surfactant). The formation of rod or sheet morphology was strongly affected by the cobalt source, while the cobalt-hydroxide-carbonate cubes were primarily determined by the hydrothermal temperature. CTAB addition was found to improve the dispersibility and make the products more uniform. The reaction time only affected the size of the obtained nanoparticles. The Co3O4 nanosheets contain high concentrations of Co3+, abundant adsorbed oxygen species, high reducibility, and superior oxygen desorption capacity, revealing high catalytic activity for CO oxidation.
2016,
Vol. 37 Issue (6): 908-922 
PDF (21895 KB)
2016,
Vol. 37 Issue (6): 908-922 PDF (21895 KB)