Owing to the environmental and health impacts of diesel particulate matter (PM) and the necessity of complying with increasingly stringent automobile emissions regulations, efforts to reduce PM emissions are ongoing, even though achieving such reductions remains challenging. PM is primarily composed of carbonaceous particles (often called soot) originating from incomplete in-cylinder oxidation of the fuel, with diameters ranging from a few nanometers (10-20 nm) to several hundred nanometers [1, 2]. Trapping PM on a filter followed by oxidation is an efficient means of eliminating this pollutant and, with the help of catalysts, the temperature required for the oxidation of soot (normally > 600 °C) can be significantly reduced. Under the conditions that typically occur in an actual catalytic diesel particulate filter (DPF), however, good contact between the PM and the catalyst is very difficult to obtain; the loose contact obtained by mixing soot and catalyst by hand has been reported to be comparable to the degree of contact obtained in a DPF [3]. Under such loose contact conditions, the majority of the carbon atoms are in contact with the surrounding gaseous phase and, consequently, the soot oxidation temperature can be significantly reduced due to the NO2 produced over the catalyst [1, 2]. For this reason, the redox properties of the Ce3+/4+ couple and the capacity of CeO2 to exchange oxygen with the gas phase make CeO2-based materials one of the most promising soot combustion catalysts.
It has been reported that CeO2-based catalysts, in addition to promoting NO2-assisted soot combustion, also allow active oxygen-assisted reactions even in the loose contact mode [4, 5, 6, 7]. In this process, active oxygen species are generated when lattice oxygen from CeO2 is replaced by gas phase oxygen [8, 9]. In the presence of NOx, the desorption of stored oxygen generated by the NO2-CeO2 reaction may also produce active oxygen species which can be transferred to the soot surface to accelerate the oxidation [4, 7]. Compared with pure CeO2, CeO2-ZrO2 mixed oxides exhibit better oxygen storage capability as well as enhanced stability during sintering, and thus Ce1-xZrxO2 solid solutions have become one of the key components in auto exhaust treatment catalysts [10]. The synergistic effect between Pt and CeO2-ZrO2 may further improve the catalytic performance, since the Pt can accelerate the mobility of the active oxygen in the CeO2-ZrO2 [6, 10]. As such, it is reasonable to suggest that Pt/CeO2-ZrO2 catalysts have potential application to the oxidation of soot and may be helpful based on the ongoing trend in which increasingly significant restrictions are imposed on diesel NOx emissions [5, 11].
Past studies have examined the applicability of CeO2-based catalysts to the soot oxidation reaction and have focused on the generation of active oxygen and the subsequent reaction promotion effect [4, 5, 6, 7], but have lacked a fundamental understanding of the manner in which the active oxygen reacts with soot, especially in the presence of NOx. The proposed reaction mechanisms have been based purely on speculation and little direct evidence has been presented. In this work, a systematic comparison between Pt/Al2O3 catalyst with Pt/CeO2-ZrO2 catalyst was carried out to explore the reaction pathways on these catalysts in the presence and absence of NOx. Efforts were made to obtain a clear insight into the roles of active oxygen in the generation and decomposition of surface oxygenated complexes on soot over Pt/CeO2-ZrO2 catalyst.
Ce0.6Zr0.4O2 (CeZr) mixed oxides were synthesized using a citric-aided sol-gel method. Ce(NO3)3·6H2O (99%, Rongruida, China) and ZrO(NO3)2·5H2O (99%, Rongruida, China) were used as the precursors while citric acid (99%, Yili, China) and polyglycol (20000 ± 200 g/mol, Yili, China) were used as complexing agents. A mixture of the precursor salts, citric acid, polyglycol, and deionized water at a molar ratio of 1:2:0.2:100 was stirred at 90 °C until a spongy yellow gel was formed. This gel was dried overnight at 110 °C and then calcined at 700 °C for 3 h in a muffle furnace. For reference, a sample of pure Al2O3 was also subjected to the same calcination process.
A solution of Pt(NO3)2 (27.82 wt%, Heraeus, Germany) was added to Ce0.6Zr0.4O2 or Al2O3 in such a manner as to ensure Pt loadings are 1%. After impregnation, the catalysts were dried at 110 °C for 12 h and calcined at 500 °C for 2 h in static air. According to a previous study, 100% of the added Pt should be present on the supports as various Pt species [6]. The resulting catalysts were denoted as PtCeZr and PtAl, respectively.
Printex-U (Degussa, Germany) was used as a model soot. It had an average particle size of 25 nm and a specific surface area of 100 m2/g. Samples of NO2-treated soot (Nsoot) containing various surface oxygen complexes were obtained by treating the Printex-U with NO2 (500 ppm) in 10% O2/N2 flowing (200 ml/min) at 350 °C for 4 h.
Powder X-ray diffraction (XRD) patterns of various samples were acquired using a diffractometer (D8 ADVANCE, Bruker, Germany) employing Cu Kα radiation (λ = 0.15418 nm) and operating at 40 kV and 30 mA. The XRD patterns were recorded at 0.02° intervals over the range of 2θ = 20°-80° at a scanning velocity of 4°/min.
The specific surface areas of the samples were obtained from N2 adsorption-desorption isotherms at -196 °C by the four-point BET method, using an automatic surface analyzer (F-Sorb 3400, Gold APP Instrument, China). The samples were degassed at 200 °C for 2 h prior to the measurements.
Pt dispersion on each of the catalysts was assessed using a Micromeritics AutoChem II 2920 apparatus with a thermal conductivity detector (TCD). During each experiment, approximately 50 mg of catalyst was placed in a U-shaped quartz tube (ɸ10 mm) and reduced under a flow of 10% H2/Ar (50 ml/min) while raising the temperature up to 400 °C at a rate of 10 °C/min, and then held at 400 °C for 30 min under a He flow (50 ml/min) to allow proper degassing. The sample was subsequently cooled to 25 °C and a loop gas consisting of 10% CO/He (20 ml/min) was introduced to the sample in pulses while the TCD signal was recorded until the peak area became constant. The percent dispersion was calculated by dividing the numbers of exposed surface Pt atoms (as determined by CO chemisorption) by the total amount of Pt in the catalyst. The mean Pt particle size was determined from CO chemisorption data by assuming all surface Pt particles were spherical in shape.
Transmission electron microscope (TEM) images were acquired using an FEI Tecnai G2 20 with an acceleration voltage of 200 kV. The size distribution of the Pt crystallites was determined by measuring 100 particles in each sample.
Infrared (IR) spectra of CO adsorbed on the samples were recorded on a Nicolet 6700 FTIR spectrometer equipped with an MCT detector. During this process, the samples were first pretreated under N2 (50 ml/min) at 500 °C for 30 min then cooled to RT, following which spectra were obtained after exposing the catalyst to 1% CO/N2 (50 ml/min) for 30 min followed by purging with N2.
H2 temperature-programmed reduction (H2-TPR) was performed using a Micromeritics AutoChem II 2920. Prior to measurements, 50 mg of each sample was pretreated under a flow of 10% O2/He (50 ml/min) at 300 °C for 30 min to allow proper degassing and then cooled to RT and flushed for 10 min. Finally, the treatment gas was switched to 10% H2/Ar (50 ml/min) and the reactor temperature was raised to 800 °C at a rate of 10 °C/min. H2 consumption was monitored by TCD.
The oxygen storage capacity (OSC) of the catalysts was measured using a flow reactor system equipped with solenoid valves, determining the composition of the effluent gases with an online mass spectrometer (Omnistar 200). Typically, 25 mg of sample was preheated under a flow of 2% O2/He (300 ml/min) prior to performing trials at 450 °C. The dynamic oxygen storage capacity (DOSC) of samples was measured by alternating CO and O2 pulses every 5 s. The OSC associated with each pulse was calculated as micromoles of CO2 released per gram of catalyst.
NO temperature-programmed oxidation (TPO) measurements were carried out in a fixed-bed reactor with monitoring the effluent gases by an IR spectrometer (Thermo Nicolet iS10). In these trials, 100 mg of sample was diluted with 300 mg of silica pellets and then sandwiched between portions of quartz wool in a tubular quartz reactor. A gas mixture consisting of NO (1000 ppm) in 10% O2/N2 (500 ml/min) was introduced into the tube and the reactor temperature was raised up to 600 °C at 10 °C/min.
Diffuse reflectance IR Fourier transform spectra (DRIFTS) were recorded on the same apparatus used for CO-IR measurements. All samples were first diluted with inert CaF2, so as to obtain a useable signal-to-noise ratio, at a soot:catalyst:CaF2 mass ratio of 1:10:100. Samples containing only soot or only catalyst were still diluted with CaF2 at the same mass ratio. After pretreatment under N2 (50 ml/min) at 500 °C for 30 min, each sample was cooled to RT and its spectrum was recorded from 100 to 550 °C at 50 °C intervals under various atmospheres. The heating rate applied during these measurements was 10 °C/min.
During activity measurements, 10 mg of soot and 100 mg of the catalyst sample were mixed by hand using a spatula for 2 min to obtain ‘‘loose contact’’ conditions. In order to prevent reaction runaway, 110 mg of the soot-catalyst mixture was then diluted with 300 mg of silica pellets. The inlet gas mixture was either NO (1000 ppm) in 10% O2/N2 or pure 10% O2/N2 at a total flow rate of 500 ml/min, applying a gas hourly space velocity (GHSV) of 30000 h-1. The activity of the catalysts for soot oxidation was evaluated in the same apparatus used for NO-TPO tests and also for TPO reaction at a heating rate of 10 °C/min. In these trials, T50 represents the temperature at which 50% of the soot was oxidized. The downstream CO2/(CO2 + CO) ratio during soot oxidation was defined as the selectivity for CO2.
Figure 1 shows the XRD patterns of the catalysts. Both of the CeO2-ZrO2-based catalysts exhibit only peaks assigned to the face centered cubic (FCC) fluorite structure of the CeO2. The diffraction peak at 2θ = 29.0°, corresponding to the FCC (111) crystal face, shows a slight shift compared to the same peak in CeO2, which appears at 28.4°. This effect is primarily due to the substitution of Zr4+ ions, which have a smaller radius of 0.086 nm, for Ce4+ ions (0.101 nm) in the fluorite-type ceria lattice, forming a Ce0.6Zr0.4O2 solid solution which results in lattice shrinkage [5]. Typical γ-Al2O3 peaks are observed in PtAl. No peaks attributed to Pt species are observed for PtCeZr and PtAl samples, indicating that Pt is highly dispersed on both supports. As shown in Table 1, PtAl sample exhibits a much higher specific surface area than the CeO2-ZrO2 mixed oxide and the corresponding supported catalyst.
The catalytic oxidation of soot over Pt catalysts is generally believed to depend strongly on the particle size of the Pt particles [12, 13]. Since the surface Pt particles on our catalysts were too small to be detected by XRD, CO chemisorption was employed to measure the Pt dispersion. The data in Table 1 demonstrate that a lesser amount of CO is adsorbed on the PtCeZr than that on the PtAl, presumably because of the reduced Pt dispersion on the low surface area CeZr support. To ascertain the Pt particle size distribution on the different catalysts and eliminate the effects of supports on CO chemisorption, high resolution TEM images of the catalysts and the corresponding Pt particle size distributions (PSD) were obtained, as presented in Fig. 2. Here it is evident that the Pt particles on the PtCeZr catalyst are larger than that on the PtAl. The PSD curve obtained with PtAl exhibits a main peak centered at around 1.5 nm, while the PtCeZr exhibits a PSD in which the center of the main maximum peak is shifted to 3 nm. These data are in good agreement with the CO chemisorption results since they indicate that larger Pt particles are formed on the CeZr support.
In addition to the Pt particle size distribution, the oxidation state of the Pt may also affect the activity of the catalysts. Figure 3 presents the IR spectra of CO adsorbed on the catalysts at RT. The PtAl catalyst exhibits three overlapping bands at 2100, 2082, and 2060 cm-1, which are assigned to CO linearly adsorbed on electron-deficient Pt clusters (Ptδ+) as well as on the terrace and step sites of Pt0 crystallites, respectively [14]. In the case of the PtCeZr catalyst, the bands assigned to Pt0 exhibit dramatically lowered intensity, indicating an obvious decrease in the Pt surface area. These results also correspond nicely with the CO chemisorption results. The band resulting from Ptδ+ on the PtCeZr (2091 cm-1) is obviously more intense than that in the PtAl spectrum (2100 cm-1) and a new band at 2162 cm-1, assigned to CO linearly adsorbed on Pt2+ or Pt3+, is observed which indicates that Pt on the CeO2-ZrO2 is more highly electron-deficient due to the alkaline nature of the support [14, 15, 16, 17]. No obvious CO adsorption bands are seen in the spectrum of the CeZr sample.
The H2-TPR profiles shown in Fig. 4 describe the redox properties of the catalysts. The PtAl exhibits a peak at 395 °C corresponding to the reduction of PtOx which has interacted with Al2O3 [18]. The Ce0.6Zr0.4O2 mixed oxide (CeZr) shows an intense peak at 520 °C, which is ascribed to the reduction of bulk oxygen in the CeO2-ZrO2 solid solution [19]. The intense peak at 98 °C obtained for PtCeZr results from the reduction of both PtOx and superficial Ce0.6Zr0.4O2 due to the strong interaction between Pt and the support. This interpretation is supported by the absence of the peak at 495 °C for the CeZr sample [20] and is attributed to the ability of the supported precious metal to activate H2 and then to spill it over onto the support [21]. In the absence of the precious metal, H2 activation is difficult and becomes the rate-determining step for the reduction reaction [19]. There is a weak H2 consumption peak at ca. 270 °C for the PtCeZr, which is assigned to the removal of oxygen from a small percentage of the unpromoted CeO2-ZrO2 powder.
The OSC data provide information concerning the utilization of various oxygen species on different catalysts, and it has been reported that H2 consumption in H2-TPR has a close connection with the total OSC of a given material [22]. At 450 °C, the DOSC values of the catalysts were 306, 858, and 492 μmol CO2/g-cat. for PtAl, PtCeZr, and CeZr, respectively. The DOSC sequence of PtCeZr > CeZr > PtAl indicates that Pt species play an important role in accelerating the release of bulk oxygen in the CeO2-based oxides. This result is in good agreement with the work of Bedrane et al. [10] who demonstrated that the Pt component may act as a porthole by which the active oxygen both enters and leaves the CeO2-ZrO2 catalyst.
Since NO2 is required to initiate and propagate soot oxidation under loose contact conditions, the oxidation of NO to NO2 is an important step during catalytic soot oxidation in the presence of NO and O2 [11, 23]. Figure 5(a) shows the effluent NO2 concentration during NO-TPO trials over different catalysts, from which it can be seen that PtAl and PtCeZr exhibit similar NO oxidation activity. The former generates a slightly larger amount of NO2 below 300 °C, while the latter produces slightly more NO2 in the range of 300-400 °C. Compared to the Pt-containing catalysts, CeZr exhibits rather low NO oxidation activity, demonstrating the important role of Pt species. Meanwhile, as shown in Fig. 5(b), almost no NO2 is observed during the soot-TPO reaction over CeZr until the temperature reaches 500 °C, at which point the majority of soot has been oxidized, while small downstream NO2 concentrations (~100 ppm) are observed during the soot-TPO process over PtAl and PtCeZr.
Figure 6(a) shows the TPO curves obtained during soot oxidation over different catalysts under either NO + O2 or O2 while the light-off temperature (T50) data are summarized in Table 2. Under both atmospheres, these catalysts uniformly exhibit high CO2 selectivity (> 98%). During the soot-TPO tests performed in 1000 ppm NO/10% O2/N2, the T50 values of the catalysts follow the order of PtCeZr < PtAl ≈ CeZr, indicating an enhancement of the catalytic activity due to the promotion effect of Pt on CeZr. These results appear, to some extent, inconsistent with the NO-TPO results, in which the NO2 productivity over the catalysts follows a different order of PtCeZr ≈ PtAl > CeZr. That is, the CeZr catalyst, while displaying lower NO oxidation activity, exhibits similar soot oxidation activity to that over the PtAl, while the NO oxidation activity over the PtCeZr and PtAl catalysts are similar even though PtCeZr obviously possesses higher soot oxidation activity. These results imply that, in addition to the important role of NO2 as a strong oxidizing agent, other factors are also involved in the soot oxidation reaction over the CeO2-ZrO2-based catalysts.
The active oxygen generated by the CeO2-ZrO2 mixed oxide and promoted by the precious metal significantly improves the soot oxidation and so the soot-TPO tests were performed in O2 to exclude the influence of NOx. As shown in Fig. 6(b) and Table 2, the T50 values over the catalysts in the absence of NO follow the order of PtCeZr < CeZr < PtAl, indicating the superior performance of the CeO2-ZrO2-based catalysts with regard to PtAl should be ascribed to the active oxygen resulting from the mixed oxide [7, 11].
It has been suggested, however, that the further oxidation of the SOCs generated by the reaction of NO2 and soot represents the rate determining step in the soot oxidation process [4, 23, 24]. As indicated in Ref. [24], the NO2-pretreated “Nsoot” contains a larger amount of surface oxygen complexes than Printex-U does, thus a comparison between Nsoot- and soot-TPO is sufficient to demonstrate the effectiveness of catalysts on the deep oxidation of soot. As shown in Fig. 6(b) and Table 2, PtAl exhibits reduced catalytic activity for Nsoot oxidation, with the T50 shifting towards higher temperatures by ca. 10 °C compared with the value obtained for soot oxidation. On the contrary, the pretreatment of soot under NOx decreases the light-off temperature of CeZr and especially of PtCeZr, demonstrating the accelerated deep oxidation of soot over the CeO2-ZrO2-based catalysts.
To confirm the accelerated dissociation of SOCs over the CeO2-ZrO2-based catalysts, in situ FTIR experiments were performed. The DRIFT spectra of adsorbed species on the soot-free catalyst and catalyst+soot mixtures under the reaction atmosphere (1000 ppm NO in 10% O2/N2) were recorded and the results are shown in Fig. 7. The results confirmed the formation of nitrites, nitrates, and some SOCs on the catalyst surface. A summary of the band assignments is presented in Table 3.
In the absence of soot, PtAl exhibits bands corresponding to chelating bidentate nitrates (1546 and 1329 cm-1) and bridging bidentate nitrates (1619 and 1229 cm-1) on Al2O3, as shown in Fig. 7(a). All the bands increase in intensity as the temperature is raised from 100 to 250 °C. At higher temperatures, with the decomposition of nitrates, the bands drop sharply in intensity and disappear almost completely at 500 °C. The band at ca. 1407 cm-1 can be assigned to nitro compounds, which are unstable and transform rapidly to nitrate species with increasing temperature. Compared with PtAl, these bands remain relatively constant in intensity in the PtCeZr and CeZr spectra acquired at high temperatures (Figs. 7(c) and 7(e)) due to the basic nature of the CeO2-ZrO2 support.
As shown in Fig. 7(b), (d), and (f), due to the pronounced light absorbance of the soot, the catalyst+soot mixtures exhibit lower band intensity compared with the soot-free samples. After mixing with the soot, an extra band appears at 1768 cm-1 in the spectra of the PtAl + soot mixture at 300 °C above. This band is generally attributed to anhydride, lactone, and carboxylic species formed on the surface of the soot, all of which are considered to be intermediate products of the partial oxidation of soot by NOx and O2 [23, 24]. These species are not observed, however, in the IR spectra of the PtCeZr + soot mixture and do not appear in the case of the CeZr + soot mixture until the temperature reaches 450 °C. These results indicate that the CeO2-ZrO2-based catalysts either inhibit the formation of SOCs or accelerate their decomposition. Based on the Nsoot-TPO results, the latter case is more plausible. This acceleration effect of the CeO2-ZrO2 mixed oxides can be further promoted with the participation of Pt.
Based on our understanding of the soot oxidation mechanism, the acceleration of the oxidation rate should involve the formation and decomposition of SOCs [4]. In the case of the Pt/Al2O3 catalyst without the participation of active oxygen from the support, the oxidation reaction will be dominated by NO2-assisted soot combustion according to the following equations.
The back spillover of the active oxygen generated (Eq. (7)) onto the Pt active sites can enhance the NO oxidation (Eq. (2)) and thereby accelerate the subsequent catalytic reactions (Eqs. (3) and (4)), resulting in a high NO ↔ NO2 recycling efficiency and more opportunity for the NO2-soot reaction to generate SOCs.
The decomposition of the SOCs is considered to be the rate-determining step during catalytic soot oxidation by some researchers [4]. The oxidation rate of the SOCs becomes obvious only at temperatures higher than 400 °C [4, 25], leading to the thermal decomposition of NO2 into inactive NO, thus making the oxidation of SOCs more difficult in the absence of the powerful oxidant NO2 [34, 35]. As indicated by Makkee et al. [9], the role of CeO2-based catalyst is to cause spillover of active oxygen onto the soot surface and its subsequent adsorption by the SOCs is an intermediate step in the soot oxidation mechanism, although only the first few reducible surface layers of the catalyst are important. Thus the soot oxidation activity is determined by the transfer capability of the active oxygen on the surface rather than the bulk OSC of the catalyst. It has been reported that the oxidation of CO on PtCeZr primarily takes place on the surface of the precious metal particles and follows an Eley-Rideal mechanism [19] as shown below.
Eq. (8) indicates the formation of CO2 on the metal surface and Eq. (9) represents the subsequent desorption of this CO2. Thus the formation and transfer of active oxygen species to the catalyst surface (Eqs. (6),(7)) are important steps in CO oxidation. In this study, the alternation of CO and O2 pulses occurred at intervals of only 5 s and thus the DOSC of the catalyst was determined to a significant extent by the active oxygen transfer. Since chemisorbed oxygen species on the Pt have been reported to play an important role in the catalytic oxidation of the SOCs, a large amount of active oxygen must be rapidly transferred from the CeO2-ZrO2 and Pt surface to the SOCs, achieving a remarkably accelerated decomposition of the SOCs on the PtCeZr, as evidenced by the Nsoot-TPO and FTIR results [36]. Because the further oxidation of the SOCs is crucial to the soot oxidation process, the PtCeZr exhibits a low T50 during soot combustion under both NO + O2 and O2. The T50 values of all the catalysts during soot-TPO in the presence of NOx as well as the T10 values obtained from Nsoot-TPO under O2 are all around 450 °C, and these values are compared with the DOSC values obtained at 450 °C in Fig. 8. It is noteworthy that a higher DOSC always results in a lower soot/Nsoot oxidation temperature.
No active oxygen is provided by Al2O3 in the case of the PtAl catalyst and so, once the chemisorbed oxygen species on the Pt are consumed, they can only be restored by gaseous O2 which has been absorbed and activated on the Pt. Thus, although this catalyst exhibits good NO oxidation activity, it shows poor NO2 utilization and O2-assisted soot oxidation activity. As for the CeZr catalyst, it has many more active oxygen species than PtAl, as indicated by the H2-TPR result. However, the lack of Pt to act as a porthole leads to low utilization of these oxygen species, resulting in an elevated H2 reduction temperature and a relatively low DOSC value at 450 °C for CeZr. Even so, the CeZr catalyst still exhibits some promotion effect during Nsoot oxidation, indicating that the active oxygen can accelerate the deep oxidation of soot even without the participation of Pt. Proposed schemes for the surface reaction processes over PtAl and PtCeZr are presented in Fig. 9.
Under loose contact conditions, a Pt catalyst supported on CeO2-ZrO2 exhibits similar catalytic activity for NO oxidation to that of an Al2O3-supported catalyst but also presents obviously higher activity for soot oxidation under NO+O2, with a decrease of 34 °C of T50. This increased activity is due to the high oxygen mobility in the mixed oxide and the synergistic effect between the precious metal and the support. Based on the high DOSC value of Pt/CeO2-ZrO2 catalyst, nearly twice that obtained with PtAl, the active oxygen on Pt/CeO2-ZrO2 can be effectively utilized to improve the NO ↔ NO2 ↔ soot recycling process. This results in ready oxidation of the SOCs that form during this recycling, leading to elevated soot oxidation activity under both O2 and NO + O2.
2014, Vol. 35 Issue (3): 407-415
PDF (962KB)
2014, Vol. 35 Issue (3): 407-415
PDF (962KB)