Soot particles are formed as undesired by-products during combustion, and together with NOx, CO, and unburned hydrocarbons comprise the main pollutants from diesel engines [1]. Inhaling soot particulates of size less than 2.5 μm can cause pulmonary diseases such as cancer, so soot particulates should be eliminated prior to their release into air. Diesel particulate filters (DPFs) combined with catalytic combustion technology appears to be the most practical method for eliminating soot from diesel exhaust. Soot oxidation catalysts can decrease the temperature required for DPF regeneration. Various catalysts have been used, such as metal oxides, noble metals, spinels, and perovskite-type oxides [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. Perovskite-type metal oxide catalysts have been extensively investigated and exhibit excellent catalytic performance and chemical, thermal, and structural stability [4, 16]. Several studies have reported that perovskite-type oxides are effective for the simultaneous catalytic removal of NOx and soot [4, 5, 6]. La-based perovskites have been the most extensively studied, some of which exhibit excellent performance for soot combustion and NOx reduction. Their activity follows the trend: LaCoO3 > LaMnO3 > LaFeO3 [17]. Cobalt-based perovskite catalysts doped with metal cations were recently reported for soot oxidation reactions [2, 3, 4, 16]. Supported metal catalysts employing LaCoO3 for soot removal have been rarely mentioned.
Ag-based catalysts have been used for soot oxidation because of their effective simultaneous removal of carbon particles and NOx [5, 6, 16, 17, 18]. Yamazaki et al. [5] suggested that Ag is the active species for oxygen activation, even at -70℃. Adsorbed oxygen species on the Ag surface were thought to migrate to the surface of the catalyst support via the Ag/support interface, and then migrate further to the soot particles. Aneggi et al. [19, 20] suggested that the catalytic activity of Ag increases when it is supported on CeO2 and ZrO2, and that the active temperature for both supported catalysts was around 300℃. Ag supported on Al2O3 exhibits a lower activity than Ag supported on CeO2 and ZrO2, regardless of the presence of Ag nanoparticles. CeO2 and ZrO2 are recognized oxygen storage material and oxygen anion conductor, respectively. Lattice oxygen atoms in these materials are thought to contribute to soot oxidation because of the enhanced oxygen exchange between the gas phase and solid. Selecting the appropriate catalyst support can decrease a catalyst’s active temperature. Perovskites are thought to be suitable materials for soot oxidation because of their high oxygen storage capability, high lattice oxygen activity, and thermal stability.
In this study, a series of Ag/LaCoO3 perovskite catalysts were prepared and used to remove soot particles in air and NOx. The structure of Ag/LaCoO3 after varying thermal treatments and the role of Ag species in soot oxidation were studied.
All chemicals were sourced from Sinopharm Chemical Reagent Co., Ltd., China. LaCoO3 perovskite was prepared by the citric acid complex combustion method [21]. La(NO3)3·6H2O (99%), Co(NO3)2·6H2O (99%), and citric acid (99.5%) were dissolved in deionized water. The solution was evaporated, dried, and then calcined at 800℃ for 4 h. Co3O4 was prepared by the same process and was calcined at 450℃ for 3 h in air. LaCoO3 catalysts containing 4.65 wt% Ag (denoted Ag/LaCO3) were prepared by the impregnation method (AgNO3, 99.8%), and calcined in air for 3 h. The samples are denoted as Ag/LaCoO3-400, Ag/LaCoO3-650, Ag/LaCoO3-700, Ag/LaCoO3- 750, and Ag/LaCoO3-800, where the suffix indicates the thermal treatment temperature (℃).
The catalysts were characterized by powder X-ray powder diffraction (XRD) using a D8 Advance type diffractometer (Bruker AXS LLC, Germany) with Cu Kα radiation (λ=0.15406 nm). The tube pressure was 20-60 kV, tube current was 40 mA, scanning angle was 20°-60°, and scanning rate was 2°/min.
The specific surface areas of the samples were measured by a Sorptometer Coulter SA 3100 apparatus using the Brunauer-Emmett-Teller (BET) method from N2 adsorption- desorption isotherms at -196℃. Prior to measurement, the samples were degassed at 300℃ for 3 h under vacuum.
Transmission electron microscopy (TEM) images were collected using a JEOL JEM-2000EM microscope (Japan) operated at 120 kV. Samples were suspended in ethanol by ultrasonication for 5−10 min, a few droplets of which were then deposited on a microgrid carbon polymer supported on a copper grid, and allowed to dry at room temperature.
Hydrogen temperature-programmed reduction (H2-TPR) measurements were performed using a purpose-built apparatus. The sample (25 mg) was heated from room temperature to 910℃ at a heating rate of 10℃/min, under a 5%H2-95%N2 (v/v) flow (40 mL/min). H2O produced from H2-TPR was adsorbed by a molecular sieve desiccant and allochroic silica gel, before the outlet reactor gas proceeded through to TCD.
The reactivity of adsorbed O2 with soot was determined from soot-temperature-programmed reduction (soot-TPR) measurements. In these measurements, catalyst and soot (catalyst:soot mass ratio=100:1) was placed in a quartz tube and was pretreated by heating at 300℃ in He (99.999%) for 1 h to remove adsorbed CO2, and then cooled to room temperature. O2 was then adsorbed by flowing 5%O2-95%He over the sample for 1 h, after which the sample was purged with He for 30 min. The temperature was then increased to 850℃ at 10℃/min under a He flow (50 mL/min), and the concentrations of CO2 (Mr/z=44) and CO (Mr/z=28) in the effluent were continuously monitored by mass spectrometry (MKS CIRRUS 2).
X-ray photoelectron spectroscopy (XPS) was performed using a Thermo ESCALAB 250Xi spectrometer with Mg Ka radiation (hu=1253.6 eV). Binding energies were referenced to the C 1s peak of adventitious carbon (284.6 eV).
The catalytic activity of samples was evaluated from temperature-programmed oxidation (TPO) measurements using Printex-U soot (Degussa) as the model reactant. This soot has a specific surface area of 100 m2/g, average particle size of 25±3 nm, and composition of 92.2 wt% C, 0.6 wt% H, and 6 wt% volatiles. The soot was mixed with the catalyst at a soot:catalyst mass ratio of 20:80. The mixture was heated from 200 to 600℃ at a heating rate of 1℃/min in air or in 0.5%NO-10%O2- 89.5%N2 at a flow rate of 50 mL/min. Catalytic performance was indicated by the temperature corresponding to the maximum soot oxidation rate (Tp) derived from the TPO curves, soot ignition temperature (T10), soot complete conversion temperature (T90), and temperature window △T (△T=T90-T10).
The catalytic stability was evaluated from 200 to 450℃. The used catalyst was mixed with 20 mg of soot again, to investigate the stability of Ag/LaCoO3-700 in the NO-O2-N2 feed gas at a flow rate of 50 mL/min.
XRD patterns of the Ag/LaCoO3 perovskite catalysts calcined at different temperatures are shown in Fig. 1. All catalysts exhibit the main diffraction peaks of LaCoO3 perovskite (JPCDS 75-0279). The diffraction peaks at 36.9° correspond to Co3O4 (JCPDS 72-2108) and are weaker than the diffractions of LaCoO3 perovskite. Two weak diffraction peaks at 2q of 38.2o and 44.4o detected in the XRD patterns of all Ag/LaCoO3 catalysts are attributed to the (111) and (200) Bragg reflections of face-centered cubic (fcc) Ag (JCPDS 65-2871) [22, 23]. The (200) Bragg reflection of body-centered cubic (bcc) Ag2O (JCPDS 65-6811) exhibits an XRD diffraction peak at 2q of 38.1°. Thus, it is difficult to exclude the presence of Ag2O or Ag species within the Ag/LaCoO3 catalysts from the XRD results, so XPS was used to examine the state of Ag species on the catalyst surface. The intensity of the Ag diffraction peak at 2θ of 38.2° decreases upon calcination at temperatures >700℃ because significant Ag species transfer from the surface to the lattice of the perovskite structure because of the high mobility of Ag at such temperatures [22].
The average crystallite sizes (Dc) and cell parameters of the catalysts were calculated from the XRD patterns using the Scherrer formula and are given in Table 1. Compared with pure LaCoO3, the cell parameter of Ag/LaCoO3 increases with increasing thermal treatment temperature until 700℃. Above this temperature the cell parameter decreases, as Ag+ begins to be incorporated into the perovskite lattice because of the high Ag mobility [22, 24]. The larger size of Ag+ (0.115 nm) compared with La+ (0.106 nm) and Co+ (0.061 nm) can expand the perovskite lattice. Replacing La3+ with Ag+ forms oxygen vacancies, so that the electrical neutrality of LaCoO3 can be maintained. However, more Ag+ in the perovskite structure causes structural distortion, which decreases the content of Ag particles on the surface, leads to a separate Co3O4 phase, and decreases the lattice parameter at high temperature (³750℃).
Fig. 2 shows TEM images of the Ag/LaCoO3 samples. The TEM image of Ag/LaCoO3-400 indicates a LaCoO3 particle size of 20-40 nm and an Ag particle size of 5-6 nm. Few Ag particles of 5-6 nm are observed in the TEM image of Ag/ LaCoO3-800, and LaCoO3 appears highly agglomerated. This is consistent with Ag migrating into the perovskite lattice, rather than sintering on the LaCoO3 surface.
Fig. 3 shows H2-TPR profiles of the Ag/LaCoO3 catalysts and Co3O4. A reduction peak centered at 480℃ with a small shoulder peak at lower temperature is observed in the profile of Co3O4. The two reduction peaks of LaCoO3 indicate that two reduction stages occur, namely the a (350-500℃) and b (500-700℃) peaks. The a peak is assigned to the reduction of adsorbed oxygen (Oads) and that of Co3+ and Co4+ to Co2+. The b peak is assigned to the reduction of lattice oxygen (Olat) and that of Co2+ to Co0 [2, 18].
The a peak at 350-500℃ for Ag/LaCoO3-400 is weaker and broader than that of LaCoO3, and the b peak is enhanced greatly. This indicates that Co species exist as Co2+, or that more Olat can be reduced because of the presence of Ag species. Increasing calcination temperature reduces the reduction peak intensity of Ag/LaCoO3 and leads to complexes containing shoulder peaks, compared with the H2-TPR profile for Ag/LaCoO3-400. The H2 consumption of the catalyst calcined at 400℃ is 54.2 μmol/g, which decreases to 36.0-37.3 μmol/g for Ag/LaCoO3- 650, Ag/LaCoO3-700, and Ag/LaCoO3-750. This is attributed to the absence of Co3O4 on the LaCoO3 surface, as shown by the XRD patterns in Fig. 1. The H2 consumption increases to 55.3 μmol/g for Ag/LaCoO3-800, which is similar to that of 56.4 μmol/g for LaCoO3, as shown in Table 1. The H2-TPR results indicate that the reducibility of the catalyst at low temperature is related to the presence of Co3O4. They also indicate that the migration of Ag from the LaCoO3 surface to the lattice during thermal treatment at 400-750℃ restrains the growth of Co3O4, which enhances the stability of the perovskite structure.
Fig. 4(a) shows the Ag 3d region of the XPS spectra of Ag/LaCoO3-400 and Ag/LaCoO3-800. The two main peaks with binding energy at 368.2 and 374.2 eV correspond to the Ag 3d5/2 and Ag 3d3/2 states, respectively, with a spin orbital separation of 6.0 eV. The Ag 3d5/2 peak appears at 368.2 eV in Ag0, 367.8 eV in Ag+, and at 367.3-368.6 eV in Ag2+ [25]. Therefore, Ag on the Ag/LaCoO3 surface is likely to be in the metallic state. The high symmetry of the Ag 3d5/2 peak suggests a relatively homogeneous environment of Ag0. The Ag/La ratios on the surfaces of Ag/LaCoO3-400 and Ag/LaCoO3-800 are 0.126 and 0.901, respectively, depending on the XPS analyses. This indicates a surface enrichment in Ag for Ag/LaCoO3-400, compared with a theoretical atomic ratio of 0.106 for LaCoO3 containing 4.65 wt% of Ag.
Fig. 4(b) shows the Co 2p region of the XPS spectra of Ag/LaCoO3-400 and Ag/LaCoO3-800. The binding energy of the Co 2p3/2 state at 780.0 eV is close to those of Co in LaCoO3 (779.7 eV) and Co3+ in Co2O3 (779.6 eV) [16]. Negative shifts of Co 2p3/2 XPS peaks arise from the presence of Co4+. A Co 2p3/2 state binding energy of 779.3 eV has been attributed to the presence of Co4+ in La1-xKxCoO3 perovskites [18, 26]. The positive shifts in the Co 2p3/2 peaks in Fig. 3(b) indicate that Co exists as low-valence Co2+ and Co3+. This is in accordance with the above XRD and H2-TPR results.
Fig. 4(c) shows the O 1s region of the XPS spectra of Ag/LaCoO3-400 and Ag/LaCoO3-800, with each containing two peaks. The peaks at 528.8 and 531.3-531.6 eV are assigned to Olat and Oads, respectively [18, 27]. The Oads content on the surface of Ag/LaCoO3-400 is as high as 67.7%, which is much higher than those of Ag/LaCoO3-800 (63.1%) and LaCoO3 (61.1%). This is attributed to more oxygen vacancies and more Ag0 on the surface of Ag/LaCoO3-400 [28].
Understanding the nature of surface oxygen species is important for catalytic oxidation applications. The catalytic activity of Ag/LaCoO3 for soot oxidation by surface active oxygen was studied by soot-TPR in the presence of high-purity He (99.999%). Under these conditions, soot can only be oxidized by surface oxygen species of the catalyst, including surface-adsorbed oxygen and surface lattice oxygen. The amount of active oxygen adsorbed on the catalysts was estimated by integrating the peaks associated with CO2 and CO formation at temperatures below 700℃ in Fig. 5. Only a small amount of CO is detected, compared with CO2, which indicates a high selectivity for CO2. The soot-TPR profiles are divided into three temperature regions, corresponding to the three kinds of oxygen species: O2− (300-450℃), O− (450-650℃), and O2− (>650℃) [29, 30]. The reduction temperature for lattice oxygen (>650℃) is much higher than that of other oxygen species, indicating its poor reactivity with soot. The low soot ignition temperature (<400℃) observed in this study (Fig. 6, Section 3.6) suggests that surface-chemisorbed oxygen species (O2−, O−) are the main active oxygen species for soot combustion. Adsorbed oxygen species (O2−, O−) are generally produced through the adsorption of gaseous O2 at oxygen vacancies, and active oxygen species originating from gas phase oxygen over metallic Ag, which are involved in carbon oxidation [31, 32]. The Ag loading likely enhances the catalytic activity because of the increased surface oxygen content. The amount of active oxygen increases with increasing Ag nanoparticle content [33, 34]. Ag/LaCoO3-400 possesses more adsorbed oxygen species than Ag/LaCoO3-800. This is because of its greater oxygen vacancy content, which results f rom the substitution of Ag+ and its greater surface Ag particle content.
Fig. 6(a) shows the TPO profiles for soot oxidation over Ag/LaCoO3 catalysts under an air flow. Soot combustion starts at around 350℃, with a Tp as high as 510℃ for LaCoO3. The Tp decreases by 50-70℃ when LaCoO3 is partially modified by Ag, depending on the thermal treatment temperature. A low Tp (435℃) is observed over Ag/LaCoO3-400 and Ag/LaCoO3-800. The conversion of soot in the presence of Co3O4 yields a high Tp (495℃). The addition of Ag significantly improves the catalytic performance for soot oxidation in air. The catalytic performance of Ag/LaCoO3 for soot combustion in a NOx atmosphere is shown in Fig. 6(b). The Tp of the Ag/LaCoO3 samples greatly reduces to about 140℃ over Ag/LaCoO3-400, compared with the catalyst containing no Ag. The Tp increases from 375 to 510℃ as the calcination temperature increases from 400 to 800℃.
Soot conversions as a function of temperature in air and NOx were derived from the TPO profiles in Fig. 6. The derived values of T10, T50, T90, and △T (△T=T90-T10) are shown in Table 2. Compared with the catalytic oxidation of soot over Co3O4 and LaCoO3, the soot conversion curves over Ag/LaCoO3 shift to lower temperature, and the soot conversion rates increase. The △T values of Ag/LaCoO3 fall in a narrow range (57-72℃), and are 76-91 and 43-58℃ lower than those for Co3O4 and LaCoO3, respectively. The activation energy for soot oxidation was determined for the catalysts by the Redhead method, using peak temperature data from the temperature-programmed experiments [35]. Table 2 shows the △Ea (Tp) values of the catalysts. The △Ea (Tp) values for LaCoO3 and Co3O4 are 160.0 and 156 kJ/mol, respectively, and are close to that for non-catalytic soot combustion reported previously [27, 36]. Loading with Ag reduces the activation energy for soot oxidation. The calcination temperature has a mild influence on the activ ation energy for soot oxidation in air.
Much better activity with lower Tp values for soot combustion is achieved over Ag/LaCoO3 in both NO and O2, compared with LaCoO3 and Co3O4. NO can be oxidized to NO2 by surface oxygen species (O2−, O−), and these species have higher activity for soot oxidation than NO and O2 [37, 38]. The activation energy for soot combustion is much lower than that in air atmosphere. Ag/LaCoO3-400 is the most active catalyst, with an activation energy of 131.7 kJ/mol. A slight increase in activation energy is observed with increasing calcination temperature. The presence of NO lowers the reaction temperature and improves the catalytic activity for soot oxidation. However, the higher △T indicates a slower combustion rate, as shown in Table 2. This wider temperature window may indicate a complex and as yet unknown reaction mechanism.
Catalytic soot oxidation requires oxygen transfer from the catalyst to soot to initiate soot combustion at low temperature. Two kinds of adsorbed oxygen species are involved in soot oxidation. Adsorbed oxygen species on the Ag surface migrate to the support surface via the interface to form oxygen species (partly O2-), and further migrate to soot particles [39]. The enhanced activity of Ag/LaCoO3 can be ascribed to more oxygen vacancies formed by Ag migrating from the LaCoO3 perovskite surface to the lattice to form La1-xAgxCoO3, and to the additional active sites of Ag itself.
The melting point of Ag is relatively low (961℃). The stability of supported Ag catalysts during soot oxidation is therefore an important consideration. Repetitive activity tests were carried out to investigate the stability of Ag/LaCoO3. Soot conversion profiles as a function of temperature are shown in Fig. 7. The value of T50 in NO-O2-TPO shifts from 355 to 362℃ during the first cycle and then 366℃ in the second and third cycles, respectively. This indicates that there is no obvious deactivation of Ag/ LaCoO3 after three cycles. Towata’ group [6] suggested that sintering of Ag nanoparticles and a loss of active Ag species were responsible for the deactivation of Ag catalysts for soot oxidation. No increase in the XRD peak intensity of Ag species is observed (not shown), and there is no shift to higher 2θ for the XRD peaks of LaCoO3 (not shown). Therefore, the stable activity of Ag/LaCoO3-700 in soot oxidation after three cycles is related to the high stability of the Ag/LaCoO3-700 structure.
LaCoO3 catalysts containing 4.65 wt% Ag were synthesized, characterized, and evaluated for the removal of soot in air or NOx. All catalysts exhibit lower Tp values than LaCoO3 for soot combustion. The XRD, H2-TPR, soot-TPR, and XPS characterizations indicate that the high soot combustion over the perovskite-structured Ag/LaCoO3 results from the following. First, Ag/LaCoO3 catalysts possess a uniform particle size (24-32 nm), which is similar to that of soot. This allows efficient contact between the soot particles and Ag/LaCoO3. Second, Ag/LaCoO3 exhibits a higher reducibility than LaCoO3. This facilitates redox circulation during soot combustion, which increases the catalytic activity of Ag/LaCoO3. Third, partial Ag migration from the LaCoO3 perovskite surface to the lattice allows oxygen vacancies to form in Ag/LaCoO3. More oxygen vacancies accelerate the transfer of oxygen species during soot combustion. Finally, more metallic Ag leads to higher oxidation ability because more O2- and O- are formed.
2016, Vol. 37 Issue (3): 428-435
PDF (946 KB)
2016, Vol. 37 Issue (3): 428-435
PDF (946 KB)