Because of the growing concerns regarding pollution from vehicle exhaust, emissions regulations worldwide are becoming increasingly strict. Three-way catalytic converters (TWC) are currently the primary technology used to control the various regulated compounds emitted by cars [1] and TWC formulations containing Pt/Rh, Pt/Pd/Rh, Pd/Rh and Pd-only noble metal combinations are all in commercial use [2]. However, Pd-only TWCs have received considerable attention for automotive emissions catalytic conversion and the number of vehicles equipped with Pd-only TWCs is increasing on the basis of the remarkable activity of these systems during oxidation reactions, as well as the availability of cleaner fuels and economic aspects [3]. It has long been recognized that 80%-90% of the hydrocarbon (HC) emissions from automobiles occur during cold-start conditions. Thus, a typical vehicle may fail stringent emission standards if the HC emissions from cold-start are not controlled [4]. Several potential solutions have been proposed to the cold-start problem, including HC traps, electrically heated catalysts, combustion heated catalysts, exhaust-gas ignition, microwave heating and close-coupled catalysts [5]. Among those solutions, close-coupled catalysts are the most direct and efficient, as no additional equipment or power inputs are needed. In such systems, the close-coupled catalyst is positioned close to the engine to allow for a rapid increase in the catalyst temperature. Unfortunately, the temperature in a close-coupled catalytic converter can reach in excess of 1000 °C during the practical application of the three-way catalytic conversion process, resulting in deactivation of Pd-based catalysts either through loss of the active surface by sintering or by transformation of the Pd to the metallic state [6].
Ce has traditionally been employed as a promoter [7, 8, 9] or as a support [10, 11, 12] and is well known as an excellent oxygen storage material because of its combination of facile redox cycling between the Ce3+ and Ce4+ oxidation states, good thermal stability, compatibility with noble metals and ease of insertion into alumina [2]. In addition, the ability of ceria to stabilize Pd in a highly dispersed state is well recognized, since the insertion of Pd into the cerium oxide lattice forms stable Pd-O-Ce bonds and thus leads to a high degree of dispersion [13]. Moreover, Ce can disperse on an alumina support in the form of two-dimensional regions when present at ≤ 10% by weight [14], thus serving as a bridge to connect Pd and Al2O3. Additionally, a strong interaction between CeO2 and Al2O3 under oxidizing conditions leads to the formation of CeAlO3 at the interface, which helps to stabilize the surface area of the Al2O3 [15, 16], although overly large concentrations of Ce will encapsulate the metal nanoparticles, resulting in an irreversible loss of the metal’s active sites [17]. We therefore chose to use Ce and to synthesize Pd-Ce nanoparticles supported on modified Al2O3, employing low ceria concentrations ranging from 0 to 4.0 wt%, in an attempt to determine the optimal ceria concentration.
Cargnello et al. [6, 18, 19] exploited the preorganization of variably functionalized thiol-protected Pd nanoparticles (with either 11-mercaptoundecanoic acid, 9-mercapto-1-nonanol or 1-dodecanethiol) and Ce(IV) alkoxide to form Pd@CeO2 nanoparticles. This method, however, is not suitable for commercial applications, since the structure of Pd@CeO2 is unstable at high temperatures. Wang et al. [20] prepared Pd-Ce catalysts by a stepwise impregnation method, in which the supports are first impregnated using PdCl2 acidic aqueous solutions, after which they are calcined at 500 °C and then impregnated with a Ce(NO3)3 aqueous solution using the same method. In this method, Ce is not introduced within the first step. In our work, we employed a co-adsorption impregnation method to form a Pd-Ce system, in which the Pd is initially combined with the cerium oxide. A series of Pd-Ce close-coupled catalysts with varying CeO2 contents were produced in this manner and have been characterized by means of three-way catalytic activity measurements, CO adsorption, high revolution transmission electron microscopy, hydrogen temperature programmed reduction and in situ DRIFTS. These techniques are all useful means of gaining insight into potentially important parameters involving the configuration of PdOx particles on the surface of the support and the interaction of PdOx with CeO2, which may be responsible for enhancing the catalytic performance of TWCs.
The La-modified Al2O3 support was purchased from the Rhodia Company. Pd-Ce close-coupled catalysts were prepared by a co-adsorption impregnation method at 25 °C, with a theoretical Pd content of 1 wt%. In this process, aqueous Pd(NO3)2 and Ce(NO3)3 solutions were initially mixed for 20 min (applying CeO2 contents of 0, 0.25, 0.5, 1.0, 2.0 and 4.0 wt%), following which the appropriate mass of La-modified Al2O3 was quickly added to the solution. After the mixture was stirred overnight, the powder was dried at 110 °C and then calcined at 500 °C for 2 h in static air. The resulting catalysts were termed Pd/La-A and Pd-Ce(x)/La-A (where x = 0.25, 0.5, 1.0, 2.0 and 4.0). A portion of each catalyst was further aged at 1100 °C for 4 h in static air to investigate the thermal stability of the samples, and these aged specimens were denoted as Pd/La-A-a and Pd-Ce(x)/La-A-a (x = 0.25, 0.5, 1.0, 2.0 and 4.0).
The catalytic activity tests were performed with a Bruker EQ55 FTIR spectrometer coupled with a multiple reflection transmission cell (Infrared Analysis, Inc.) before and after the simulated exhaust gas passed through the reactor. 0.2 ml catalyst was held in a quartz tube in the fixed-bed quartz reactor by packing quartz wool at both ends of the tube. The feed stream was regulated using special mass flow controllers and was composed of 0.12% NO, 0.03% NO2, 0.066% C3H6, 0.033% C3H8, 0.6% CO and 0.745% O2 with the balance Ar, at a GHSV of 43000 h−1 relative to the catalyst volume and the gas flow rate at room temperature (25 °C). Although a small quantity of H2O was generated in the three-way catalytic reaction in our study, its presence would cause serious damage to the FTIR spectrometer and therefore H2O was removed from the product stream before the gas passed through the multiple reflection transmission cell, and thus its concentration is not provided in the tabulation of results.
The air/fuel ratio (λ) was defined as equal to (2VO2 + VNO + 2VNO2 )/(VCO + 9VC3H6 + 10VC3H8), where V is the concentration of each gas in units of volume percent, and a λ value of one was used in all the activity measurements. The operational window trials were carried out at 400 °C, applying different λ values by adjusting the concentration of O2.
H2-TPR measurements were carried out on a GC-1690 chromatograph to observe the reducibility of the catalysts. During these tests, 50 mg catalyst was pretreated at 200 °C in N2 (30 ml/min) for 0.5 h and then cooled to -90 °C with liquid nitrogen. The reducing gas (5 vol% H2/Ar at 40 ml/min) was passed through the sample and the temperature was raised to 900 °C at a rate of 10 °C/min. The consumption of hydrogen during the reduction was measured by a thermal conductivity detector (TCD).
Pd dispersion and particle sizes were determined by CO chemisorption at room temperature, using a CHEMBET-3000 (Quantachrome Co.). Prior to the experimental trials, each sample was reduced under a flow of 5% H2/95% N2 at 400 °C for 1 h and then purged with He at the same temperature for 0.5 h. The sample was subsequently cooled to room temperature under an He flow and maintained at this temperature for another 0.5 h. Finally, CO pulses were injected into the sample bed every 5 min until no further consumption of CO could be detected. The Pd dispersion (D) was calculated using the following equation [21]:
Here Vco is the CO volume adsorbed (mL), f is the stoichiometric factor, CPd is the Pd metal content (wt%), and mCat. is the catalyst weight (g). The Pd particle size (d) was calculated according to the following equation:
Here MPd is the molar mass of Pd (106.42 g/mol), ρPd is the Pd density (12 g/cm3), and SPd is the molar surface area of Pd assuming equidistribution of the low index faces (S = 47780 m2/mol for Pd metal).
HRTEM images were obtained using a TECNAI G220 instrument operating at 200 kV with a perforated carbon film supported on a copper grid. Catalysts were crushed in an agate mortar and then ultrasonically dispersed in ethanol. EDS analysis was used to record elemental maps of samples, thus determining the chemical compositions of the catalysts.
Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) analyses were performed with a Nicolet 6700 FTIR spectrometer equipped with an MCT detector, a DRIFTS cell fitted with CaF2 windows and a heated reaction chamber for high-temperature trials. Spectra were collected by averaging 32 scans at a resolution of 4 cm-1. Prior to analysis, catalysts were pretreated in Ar at 450 °C for 0.5 h and then cooled to 30 °C. The composition and flow rate of the feed stream were kept constant through each test.
Fig. 1 presents the conversions of HC, CO, NO and NO2 under stoichiometric CO + HC + NOx + O2 reaction conditions (λ = 1) over fresh Pd/La-A and Pd-Ce(x)/La-A catalysts. Over the temperature range from 150 to 200 °C, HC and NO did not take part in the reaction, while the reactions CO + 1/2O2 → CO2 and 2NO2 + 4CO → N2 + 4CO2 predominated. The light-off temperatures associated with CO and NO2 conversions increased with increasing CeO2 content, indicating that the addition of CeO2 inhibited the conversion of CO and NO2 when the temperature was below 200 °C. At 200-250 °C, the CO/O2 and NO2/CO reactions were still predominant, although NO2 conversion over all catalysts decreased as the temperature increased and the CO conversion increased to 100%, suggesting that the CO/O2 reaction occurred more rapidly than the NO2/CO reaction. In contrast to the catalytic behavior below 200 °C, the conversion of NO2 in the higher temperature range was obviously enhanced with increasing CeO2 content, and Pd-Ce(2.0)/La-A showed the highest activity.
Above 250 °C, the conversion of HC, NO and NO2 increased as the temperature increased. It can be seen that HC conversion coincided with NOx conversion in this range, since NOx conversion is controlled via the reactions 10NO + 4HC → 5N2 + 4CO2 + 2H2O and 10NO2 + 8HC → 5N2 + 4CO2 + 4H2O [22]. These results show that the presence of an optimum concentration of ceria promotes the catalytic activities associated with HC and NOx removal. With increases in the CeO2 content from 0.25 to 2.0 wt%, the catalytic activity first appeared to increase and then sharply decreased at 4.0 wt% of CeO2). Thus, the Pd-Ce(2.0)/La-A catalyst was also seen to exhibit better activity than the other catalysts.
Fig. 2 presents the conversions of HC, CO, NO and NO2 under stoichiometric CO + HC + NOx + O2 reaction conditions (λ = 1) over Pd/La-A-a and Pd-Ce(x)/La-A-a catalysts calcined at 1100 °C. A large drop in catalytic activity was observed compared with the results obtained with the fresh catalysts (Fig. 1), because of the deteriorated structure and sintering of active components during the high-temperature pre-treatment [23]. In spite of this, the catalytic activity of the aged samples was still promoted with increasing CeO2 content. Pd-Ce(4.0)/La-A-a exhibited the best catalytic activity for CO and HC oxidation while Pd-Ce(2.0)/La-A-a showed the best catalytic activity for NO and NO2 elimination. The results indicate that addition of the appropriate amount of CeO2 promotes the thermal stability of the catalysts, and does not suggest any encapsulation of PdOx species by ceria during the high-temperature pre-treatment, which could have important consequences in terms of improving the catalytic performance of close-coupled catalysts in automobile exhaust elimination.
In our experimental trials, we also evaluated the conversions of CO, HC and NOx under different air/fuel ratios (λ) at 400 °C over fresh and aged catalysts. When using the fresh catalysts (data not shown here), there were no significant differences in the conversions of CO, HC and NOx; the HC conversion remained at 100% and the CO conversion was above 80% over the entire range of λ values. Moreover, the NOx conversions only showed very slight differences from one another. However, the catalytic behavior did exhibit a noticeable change after the catalysts were aged at 1100 °C. As shown in Fig. 3, the conversion of CO dropped slightly with increasing ceria content under λ < 1 (lean oxygen) conditions. The HC conversion over Pd/La-A-a was also lower than 80% even at λ > 1 (rich oxygen) conditions, which may indicate a decrease in the catalytic activity for HC oxidation due to the reduction of PdO to Pd0 under the reaction conditions [24]. The NOx conversion also dropped sharply when the λ value was above 0.95, demonstrating that the operational window of the Pd/La-A-a catalyst was significantly reduced after aging at 1100 °C. However, the operational window was also broadened by the addition of CeO2, and the conversions of HC and NOx both obviously increased with increasing CeO2 content, indicating that the presence of CeO2 stabilizes PdOx species in oxidized form because of the interaction between PdOx and CeO2. Pd-Ce(4.0)/La-A-a showed the widest operational window for HC conversion while Pd-Ce(2.0)/La-A-a exhibited the best catalytic behavior for NOx elimination. To obtain more information about the configuration of PdOx particles on the surface of the support and the inter actions between PdOx and CeO2, CO adsorption, HRTEM, H2-TPR and in situ DRIFTS results obtained for the catalysts are analyzed below.
Pd dispersion and particle size results are presented in Table 1. For each fresh catalyst, Pd was well dispersed on the support and the calculated Pd particle sizes for these samples exhibited a slight increase with increasing CeO2 contents. In contrast, the Pd particle sizes of the aged catalysts showed significant decreases as the CeO2 content increased, demonstrating that the addition of CeO2 enhances the stability of PdOx species at high temperatures and helps to slow the sintering progress. The catalyst with a CeO2 content of 2.0 wt% exhibited the highest Pd dispersion and smallest Pd particle size after being calcined at 1100 °C, results that are reflected in its catalytic performance.
Table 1 lists the calculated particle sizes of the metal Pd based on theoretical values, which will be somewhat different from the actual sizes of the PdOx species in the catalysts. To further investigate the structural differences between Pd/La-A and Pd-Ce(2.0)/La-A on the atomic scale, the materials were further studied by TEM, and micrographs of the fresh samples are shown in Fig. 4. Only La2O3 and CeO2 are observed in these samples, with interplanar distances of 0.307 and 0.312 nm, respectively. No PdOx particles were found because of the low concentration and high dispersion of this species, and since the small Pd clusters are difficult to distinguish from the support. However, as demonstrated by Fig. 4(c) and Fig. 4(d), X-ray energy dispersive spectroscopy (EDS) line profile analysis confirmed that the bright spots in these images were composed of Pd and Ce, which clearly establishes the existence of a Pd-Ce structure in the Pd-Ce(2.0)/La-A catalyst.
The micrographs of aged samples are shown in Fig. 5. In Fig. 5(d), lattice fringes with interplanar spacings of 0.262 and 0.298 nm were determined to be associated with the (101) facet of PdO and the (011) facet of Ce2O3. As shown in Fig. 5(a) and Fig. 5(c), the PdO particle sizes exhibited growth after aging at 1100 °C because of significant agglomeration. It is obvious that the Pd-Ce(2.0)/La-A-a catalyst showed much smaller diameters for its Pd-Ce mixed oxides (20-50 nm) compared with the diameters of the PdOx species in the Pd/La-A-a (70 nm). These results are consistent with the data presented in Table 1, establishing that the addition of CeO2 helps to stabilize PdOx species with smaller particle sizes. This may be the main reason that the Pd-Ce(2.0)/La-A-a catalyst maintained good catalytic performance after the aging treatment.
Fig. 6 shows the H2-TPR profiles of fresh and aged catalysts. The fresh catalysts exhibit four hydrogen consumption peaks (α, β, γ and δ) in addition to a negative peak below 400 °C. According to our previous work [25], peak α is associated with the reduction of PdOx species finely dispersed on the support, while the negative peak at 70 °C is generally attributed to the decomposition of palladium hydride when reduced Pd is exposed to H2 [10, 23]. However, the negative peak disappears when the CeO2 content reaches 0.5 wt% and peak β appears when the CeO2 content further increases. Moreover, peak α undergoes a shift to a slightly higher temperature with increasing CeO2 content. Osorio et al. [26] found that the addition of CexZr1-xO2 mixed oxides to Al2O3 promoted the palladium-support interactions and some hard-to-reduce PdOx species were formed when x = 0.5 and 0.33. Thus we suggest that there is a strong interaction between the metal and CeO2, leading to the catalysts having higher reduction temperatures, such that peak β can be attributed to the reduction of PdOx species undergoing strong interactions with CeO2. Among these catalysts, there were no significant differences in the intensities and temperature locations of peaks γ and δ, indicating that these peaks may be related to the reduction of the La-modified Al2O3 support. According to the literature [27], contact between Pd and CeO2 particles stabilizes the Pd clusters in an average Pd(I)-like oxidation state, and the presence of such oxidized Pd species serves as a source of oxygen to effectively decrease the light-off temperature during HC and NO conversions (F ig. 1). Additionally, the enhanced interactions between the PdOx species and CeO2, indicated by the shifts of both peaks α and β to higher temperatures, can be correlated to the decreased catalytic performance during CO and NO2 conversions at low temperature.
After aging at 1100 °C, the H2 consumption peaks α and β both decreased because of reduced dispersion of the active PdOx species resulting from significant sintering during the aging process. Furthermore, this agglomeration of PdOx species leads to weakening of the interactions between the PdOx species and the support, thus shifting both peaks α and β to lower temperatures compared with the fresh catalysts. In the case of the aged catalysts, however, peaks α and β also shifted to higher temperatures with increasing CeO2 content because of the strong interactions between the PdOx and CeO2.
To assess the potential relationship between the presence of particular Pd phases and catalytic behaviors, in situ DRIFTS studies were performed under different reaction conditions. Fig. 7(a) and Fig. 7(b) present a set of in situ DRIFTS spectra obtained from fresh Pd/La-A and Pd-Ce(2.0)/La-A catalysts under a stoichiometric CO + HC + O2 reaction mixture at 250 °C. Bands in the region of 1800-2100 cm-1 can be attributed to carbonyls adsorbed on Pd0 sites. In these spectra, we can observe the presence of on-top-bonded (linear) Pd-CO at 2063 cm-1 and a strong bridge-bonded Pd2-CO peak at 1972 cm-1 along with a three-fold hollow Pd3-CO at 1909 cm-1 [28, 29], indicating that some PdOx species are reduced to metallic Pd under these reaction conditions. However, the intensities of these features in the Pd/La-A spectra are much stronger than those obtained with Pd-Ce(2.0)/La-A, demonstrating that Pd-Ce(2.0)/La-A retains a large proportion of oxidized Pd species and thus presenting fewer metallic Pd sites [28]. These results show that the presence of a small amount of CeO2 promotes the reoxidation of Pd0 to PdO under reaction conditions. Furthermore, there is little evidence of the three-fold h ollow Pd3-CO species on the Pd-Ce(2.0)/La-A and the bridge/ on-top-bonded CO bands are obviously shifted to lower wavenumbers. By analogy with observations reported for Pd-Au [30], the formation of Pd-Ce nanoparticles would cause the dilution of Pd in the CeO2, thus limiting the extent of multi-bonded CO on Pd and further leading to the disappearance of the three-fold hollow Pd3-CO band. The adsorption bands in the range of 1200-1800 cm-1 can be attributed to surface carbonates/carboxylates species adsorbed on Al2O3 [31, 32]. The intensities of these bands do not undergo visible changes over time, indicating that the concentration of carboxylates/ carbonates species on the alumina surface is essentially saturated during the initial exposure to the CO + HC + O2 reaction mixture because of the complete oxidation of CO at 250 °C.
Stoichiometric NOx was abruptly added into the flow after 10 min exposure to a CO + HC + O2 mixture at 250 °C and the corresponding spectra were collected as a function of time. As shown in Fig. 7(c) and (d), bands at 2253 and 2231 cm−1 appear ascribed to isocyanate species adsorbing at octahedral and tetrahedral Al3+ cations, respectively [33]. These NCO species are formed by the interaction of unreacted CO with adsorbed N atoms (Na) [34]. Two reaction mechanisms have been proposed for the formation of Na; the first involves the unimolecular dissociative adsorption of NO to form Na and Oa, while the second involves a direct bimolecular reaction between NO and CO to form Na and CO2 [35]. The significant intensity of the NCO band and its ready detection may be used to roughly determine the onset temperature for NO dissociation, and these bands are observed to reach a maximum after 2 min and then decrease slightly because of the NCO + NO + O2 → N2 reaction. However, it is noteworthy that the intensities of the NCO bands on the Pd-Ce(2.0)/La-A are much stronger than in the Pd/La-A, indicating that more NCO species are formed on the Pd-Ce(2.0)/La-A, and thus this process is responsible for the improved NOx elimination performance seen for this catalyst at 250 °C, which is in accordance with the catalytic activity data shown in Fig. 1. Moreover, it is interesting that a new band is observed at 2175 cm-1 in the Pd-Ce(2.0)/La-A spectra. According to the results of H2-TPR analyses, there are two different PdOx species dispersed in the catalyst and so this band may be related to NCO species produced on P d0 sites having strong interactions with Ce atoms. Under such conditions, NCO species should be able to spill over onto CeO2 sites because of the close contact of the PdOx-CeO2 particles. Simultaneously, three bands at 1610, 1575 and 1459 cm-1 are also observed during the initial exposure to the CO + HC + NOx + O2 mixture, all of which reach a maximum after 5 min and then remain essentially constant. According to the literature, the band at 1610 cm-1 can be assigned to bidentate nitrates species, while the 1575 and 1459 cm-1 bands may be attributed to acetate species directly related to the reaction between surface nitrates species and HC [36]. The intensities of the nitrate/acetate bands in the Pd-Ce(2.0)/ La-A spectra are also much stronger than those of the bands obtained from Pd/La-A, which is related to the superior NO/HC elimination performance of the Pd-Ce(2.0)/La-A at lower temperatures. In addition, compared with Fig. 7(a) and (b), the intensities of the on-top/bridge-bonded CO species adsorbed on the Pd0 sites are obviously reduced under stoichiometric CO + HC + NOx + O2 reaction conditions, indicating that the presence of strongly oxidizing NO2 species favors the stability of oxidized Pd species under these conditions, especially in the case of small Pd clusters.
To elucidate the reaction mechanism, a set of in situ DRIFTS spectra were obtained for the fresh catalysts under stoichiometric CO + NOx + HC + O2 reaction conditions from 50 to 400 °C. As shown in Fig. 8, three bands in the region of 1200-1800 cm-1, assigned to nitrite/nitrates species in a variety of structures/configurations adsorbed on alumina [37, 38], can be observed at 50 °C. The intensity of these features over the Pd-Ce(2.0)/La-A catalyst is much stronger than that over Pd/La-A and almost reaches an initial maximum at 50 °C, indicating better adsorption on the support after adding CeO2, which may be favorable in terms of improving the NO reduction performance at lower temperatures. These bands decrease in intensity with increasing reaction temperature. The band at 1319 cm-1 assigned to monodentate nitrite species adsorbed on alumina [37] shifts to lower frequencies and disappears at about 270 °C, because of the effect of decreasing NOx coverage caused by the thermodynamically favored desorption of NOx and increasing CO-NOx reaction activity. Bands in the region of 2100-2300 cm-1, assigned to NCO species, and bands in the region of 1400-1800 cm-1, because of acetate and nitrates species adsorbed on support, also appear and increase in intensity with increasing reaction temperature. As noted, the changes in those bands due to the conversion of NOx and HC are in accordance with the observed catalytic performances. The intensities of the nitrate/acetate bands in the Pd-Ce(2.0)/La-A are much stronger than those in the Pd/La-A at the initial lower temperature, suggesting that the Pd-Ce(2.0)/La-A possesses better NO/HC elimination performance at lower temperatures. Moreover, in the case of the Pd/La-A catalyst, bands at 2061, 1972 and 1911 cm-1, all assigned to the adsorption of CO on metallic Pd, begin to appear at 190 °C, indicating that a portion of the PdOx species is reduced to metallic Pd under these stoichiometric reaction conditions. Also, very weak evidence of these Pd0-CO species can be observed in the Pd-Ce(2.0)/La-A catalyst at higher temperatures, demonstrating the improved stabilization of oxidized PdOx species on the Pd-Ce catalysts.
A series of Pd-Ce/La-Al2O3 close-coupled catalysts were prepared by a co-adsorption impregnation method. In the case of each unaged catalyst, the presence of an appropriate quantity of CeO2 obviously enhanced the catalytic activity during HC and NO eliminations because of interactions between PdOx species and CeO2 that improved the ability of Pd0 to reoxidize to PdO and enhanced the adsorption of nitrite/nitrates and isocyanate intermediate species under reaction conditions. The Pd-Ce(2.0)/La-A catalyst exhibited the best catalytic activity for HC and NO eliminations and the addition of CeO2 inhibited the conversions of CO and NO2. The Pd-Ce(4.0)/La-A catalyst showed the lowest catalytic activity for CO, HC and NOx eliminations, which may result from encapsulation of Pd particles by excess CeO2 and associated inhibition of its catalytic performance. However, after aging treatment, the operational window for HC and NOx conversions was obviously broadened because of the presence of optimal concentrations of CeO2 (2.0-4.0 wt%), and the catalytic activities for CO, HC and NOx eliminations were evidently promoted with increased levels of ceria content. These effects demonstrate that the presence of optimal CeO2 concentrations favors stabilization of PdOx species with smaller particle sizes because of the strong interactions between such PdOx species and CeO2.
随着汽车尾气污染排放的日益增加, 各国相应的排放法规也越来越严苛. 三效催化剂(TWC)是控制汽车尾气污染物排放的主要技术[1]. 目前投入商业应用的TWC包含了Pt/Rh, Pt/Pd/Rh, Pd/Rh和单Pd等贵金属催化剂[2]. 单Pd三效催化剂因为其显著的氧化反应活性和价格低廉等原因而广泛应用于汽车尾气的催化转化, 配置单Pd三效催化剂的机动车数量也在增加[3]. 汽车尾气中HC排放物的80%-90%是在冷启动阶段生成的. 因此, 如果HC在冷启动阶段的排放量未得到有效控制, 那传统的汽车将不能达到严格的排放标准[4]. 目前提出了众多解决冷启动问题的方案, 包括HC捕集器, 电加热催化剂, 燃烧加热催化剂, 废气点火, 微波加热和密偶催化剂(CCC) [5]. 其中, CCC因为不需要配置附加装置和额外的能量输入而成为最直接有效的解决方法. CCC非常靠近汽车发动机, 以便快速地提高催化剂的温度. 然而, CCC催化转化器在实际的三效催化转化过程中温度可以达到1000 °C以上, 导致Pd基催化剂在高温下烧结并从氧化态转变为金属态Pd从而失活[6].
众所周知, Ce通常被用作三效催化剂的助剂[7, 8, 9]或载体[10, 11, 12]. 它由于具有在Ce3+和Ce4+之间灵活的转化能力, 良好的热稳定性, 与贵金属的兼容性和容易浸渍到氧化铝上等性能而被作为最主要的储氧材料[2]. 另一方面, Pd嵌入到CeO2晶格中会形成Pd-O-Ce键, 从而使Pd在载体表面高度分散[13]. 当Ce含量小于10 wt%时, Ce可以在氧化铝载体上以二维形式分布[14], 从而起到了联系Pd与Al2O3的桥梁作用. 除此之外, 氧化条件下CeO2和Al2O3的强相互作用会使界面处生成CeAlO3, 有助于Al2O3表面的稳定[15, 16]. 然而, Ce的添加量过大可能会对贵金属颗粒形成包覆, 造成贵金属活性位点的不可逆流失[17].
Cargnello等[6, 18, 19]利用巯基十一烷酸, 巯基壬醇和十二硫醇制备了不同巯基保护的Pd纳米颗粒, 并将其与Ce(IV)醇盐反应生成Pd@CeO2纳米颗粒. 但是这种方法并不适用于商业应用, 而且Pd@CeO2在高温下并不稳定. Wang等[20]通过分布浸渍法制备了Pd-Ce催化剂: (1)用PdCl2的酸性水溶液作为前驱体将Pd浸渍在载体上; (2) 500 °C煅烧后, 用同样的方法把Ce(NO3)3的水溶液浸渍在样品上. 但是在这种方法中, Ce在第一步中没有引入. 本文使用共吸附浸渍法来形成Pd-Ce系统, 使Pd在初始阶段就可以稀释在 Ce中, 从而制备了一系列不同CeO2含量的Pd-Ce密偶催化剂, 并测量其三效催化活性. 采用CO吸附, 高分辨透射电镜, H2-程序升温还原和原位红外(in situ DRIFTS)等手段进行表征, 以期对PdOx颗粒在载体表面的状态和PdOx与CeO2之间的相互作用进行深入研究.
La-Al2O3载体购买自Rhodia公司. Pd-Ce密偶催化剂在25 °C采用共吸附浸渍法来制备, Pd含量为1 wt%. Pd(NO3)2和Ce(NO3)3的水溶液先混合搅拌20 min (CeO2含量分别为0, 0.25, 0.5, 1.0, 2.0和4.0 wt%), 然后把适量的Al2O3载体快速加入溶液中. 继续搅拌过夜后在110 °C干燥, 500 °C焙烧2 h得到新鲜的催化剂样品, 分别命名为Pd/La-A和Pd-Ce(x)/La-A (x = 0.25, 0.5, 1.0, 2.0, 4.0). 最后取部分催化剂在1100 °C焙烧4 h来研究样品的热稳定性. 老化后的样品分别命名为Pd/La-A-a和Pd-Ce(x)/La-A-a (x&# 8197;= 0.25, 0.5, 1.0, 2.0, 4.0).
催化剂活性评价仪器为自建的微反装置, Bruker EQ55型傅立叶气体红外分析仪在线检测进出反应器气体. 在石英管中填装0.2 mL催化剂, 并在其两端填充石英棉. 反应气的流量由质量流量控制器来控制, 其组成为NO 0.12%, NO2 0.03%, C3H6 0.066%, C3H8 0.033%, CO 0.6%和O2 0.745%, 载气为Ar, 空速为43000 h-1 (计算于催化剂的体积与室温下的气体流速). 虽然在反应过程中产生了少量的H2O, 但它的存在会损坏FTIR分光仪. 因此在气体进入仪器之前便除去了H2O, 它的浓度也没有列于图表中.
空燃比定义为λ = (2VO2 + VNO + 2VNO2)/(VCO + 9VC3H6 + 10VC3H8), 其中V代表各气体组分的体积百分含量, λ = 1应用于所有的活性评价中. 催化剂窗口测试在400 °C下进行, 通过调节氧含量来设置不同的λ值.
H2-TPR实验在GC-1690型色谱分析仪上进行. 将50 mg催化剂在N2气氛下(30 ml/min) 200 °C处理30 min, 用液氮冷却至-90 °C后, 通入5 vol% H2/Ar (40 ml/min)至基线稳定, 再以10 °C/min升温至900 °C. 还原过程中消耗的氢气通过热导检测器(TCD)进行检测.
CO吸附在室温下由CHEMBET-3000型(Quantachrome公司)化学吸附仪进行检测, 以此计算Pd的分散度和颗粒尺寸. 实验前, 样品在5% H2-95% N2气氛中400 °C还原1 h, 在He中保持30 min后, 样品在He中冷却至室温并保持30 min. 最后, CO每隔5 min脉冲入样品中直到没有检测到CO的消耗为止. Pd的分散度(D)计算如下[21]:
Vco, CO吸附的体积(mL); f, 化学计量数; CPd, 金属Pd的含量(wt%); mCat., 催化剂的质量(g). Pd颗粒尺寸(d)计算如下:
MPd, Pd的摩尔质量(106.42 g/mol); ρPd, Pd的密度(12 g/cm3); SPd, Pd的摩尔表面积(S = 47780 m2/mol).
HRTEM照片由TECNAI G220型透射电镜表征获得, 工作电压为200 kV. 催化剂首先在研钵中碾碎, 无水乙醇中超声分散后担载在镀有碳膜的铜网上. EDS分析用来记录元素分布, 从而得到化学组成.
原位漫反射红外傅立叶变换光谱测试(In situ DRIFTS)在配置了MCT (merbury-cadnium-tellvride)检测器的Nicolet 6700 FTIR分光仪上进行. 原位池窗片为CaF2, 有一个可加热的反应室来进行高温反应, 扫描次数32, 分辨率4 cm-1. 催化剂在Ar中450 °C下预处理30 min, 然后冷却至30 °C. 反应器的组成和流速在反应过程中保持稳定.
图1给出了模拟汽车尾气气氛下(λ = 1)新鲜Pd/La-A和Pd-Ce(x)/La-A催化剂上HC, CO, NO和NO2的转化率. 由图1可见, 在150-200 °C温度范围, HC和NO并未参与反应, 而CO + 1/2O2 → CO2和2NO2 + 4CO → N2 + 4CO2在反应中占主导地位. CO和NO2的起燃温度随着CeO2添加量的增加而增高, 说明当温度低于200 °C时, CeO2的添加抑制了CO和NO2的转化. 在200-250 °C的温度范围, CO/O2和NO2/CO反应仍然是主要的. 各催化剂上CO转化率均达到100%, 而NO2的转化率却随着温度的增加而降低, 可能是由于CO/O2反应比NO2/CO反应进行得更快. 不同的是, 在200-250 °C温度范围中各催化剂上NO2转化率随CeO2添加量的增加而增大, 其中Pd-Ce(2.0)/La-A表现出最好的NO2还原的活性.
当温度大于250 °C时, HC, NO和NO2的转化率随温度增加而增大, HC与NO转化曲线一致. 研究发现[22], NOx的转化由10NO + 4HC → 5N2 + 4CO2 + 2H2O和10NO2 + 8HC → 5N2 + 4CO2 + 4H2O两个反应来控制. 以上结果表明, 适量CeO2的添加可以提高HC和NOx的反应活性. 随着CeO2添加量从0.25增加到2.0 wt%, 催化活性明显增强, 至4.0 wt%时活性急剧下降. 其中Pd-Ce(2.0)/La-A表现出最好的催化活性.
图2为模拟汽车尾气气氛下(λ = 1)老化后的Pd/La-A和Pd-Ce(x)/La-A催化剂上HC, CO, NO和NO2的转化率. 与图1相比, 老化后催化剂的活性大大减弱, 这与催化剂在高温处理后(1100 °C)导致的活性组分烧结和载体结构塌陷有关[23]. 尽管如此, 催化剂的活性仍随着CeO2添加量的增加而增强, 其中Pd-Ce(4.0)/La-A-a呈现出最好的CO和HC的氧化活性, 而Pd-Ce(2.0)/La-A-a催化NO和NO2还原活性最高. 这说明适量CeO2的添加会在不包覆PdOx物种的情况下提高催化剂的热稳定性, 这对改善密偶催化剂的活性有重要意义.
本文还评估了400 °C时不同空燃比窗口下CO, HC和NOx在新鲜和老化催化剂上的转化. 对于所有的新鲜催化剂(数据未列出), CO, HC和NOx的转化率并没有明显差异. HC均为100%转化, CO的转化率在整个λ范围都高于80%, NOx的转化率也只有轻微的区别. 然而, 在1100 °C老化之后, 各催化剂的催化行为变化显著. 如图3所示, 在贫氧条件下(λ < 1), CO的转化率随着CeO2添加量的增加而有所下降. Pd/La-A-a上HC转化率在整个窗口中都低于80%, 这可能是因为PdO在老化过程中还原 为Pd0而导致HC氧化活性的降低[24]. 当λ > 0.95时, Pd/La-A-a的NOx转化率急剧减少. 由此可见, 1100 °C的高温老化明显破坏了Pd/La-A-a的操作窗口. 但是, CeO2的添加拓宽了催化剂的操作窗口, HC和NOx的转化率也随着CeO2添加量的增加而明显增大, 这表明PdOx和CeO2间的相互作用可以使PdOx物种稳定在氧化态. Pd-Ce(4.0)/La-A-a有最宽的HC窗口, 而Pd-Ce(2.0)/ La-A-a有最宽的NOx窗口. 为得到关于PdOx物种在载体表面的构型和PdOx与CeO2间相互作用的信息, 对催化剂进行了CO吸附, HRTEM, H2-TPR和原位红外等表征.
各催化剂的Pd分散度及其颗粒尺寸列于表1. 对所有新鲜催化剂来说, Pd在载体表面分散度都较高, Pd颗粒尺寸随着CeO2含量的增加而略有增大. 反之, 催化剂老化后Pd的颗粒尺寸随着CeO2含量的增加而减小, 说明CeO2的添加可以增强PdOx物种的高温稳定性并减缓烧结的过程, 这可以与催化剂的活性变化趋势相关联. 含2% CeO2的催化剂在1100 °C老化后呈现出最高的Pd分散度和最小的Pd颗粒尺寸.
如表1所示, 金属Pd颗粒尺寸的理论值与PdOx颗粒的实际大小并不相同. 为了从原子尺度来研究Pd/La-A和Pd-Ce(2.0)/La-A结构上的不同, 我们对催化剂进行了TEM表征. 如图4所示, 新鲜催化剂上只能观测到La2O3 (0.307 nm)和CeO2 (0.312 nm)的晶面间距. 由于PdOx颗粒含量低, 分散度高以及小的Pd簇不易与载体区分等原因而无法在图中找到. 然而对Pd-Ce(2.0)/La-A进行的EDS线扫描分析显示(图4(c)和(d)), 图中的白点由Pd和Ce共同组成, 清晰地指出了Pd-Ce结构的存在.
图5是老化催化剂的TEM照片. Pd-Ce(2.0)/La-A上也能观察到PdO的(101)晶面(d = 0.262 nm)和Ce2O3的(011)晶面(d = 0.298 nm). 如图5(a)和(c)所示, 老化后PdO的颗粒尺寸因烧结作用而急剧增大. 但Pd-Ce(2.0)/La-A-a中的Pd-Ce复合氧化物颗粒尺寸(20-50 nm)明显比Pd/La-A-a的PdO颗粒尺寸(70 nm)更小, 这与表1结果一致, 证明CeO2的添加有助于稳定PdOx物种的颗粒尺寸. 这可能是Pd-Ce(2.0)/La-A-a在老化后也能保持良好的催化活性的主要原因.
图6为新鲜和老化后催化剂的H2-TPR谱. 新鲜催化剂在400 °C以下出现了四个峰(α, β, γ和δ)以及一个负峰. α峰可归属于高度分散在载体上PdOx物种的还原, 70 °C左右的负峰则对应于氢化钯的分解[25]. 其中, 氢化钯是由PdO还原生成的金属Pd与H2进一步作用所形成[10, 23]. 当催化剂中CeO2的添加量达到0.5 wt%时, 负峰消失. 随着CeO2的添加量进一步增加, β峰开始出现. 此外, α峰逐渐往高温偏移. Osorio等[ 26]发现, CexZr1-xO2复合氧化物的添加在可促进Pd与Al2O3间相互作用, 当x = 0.5和0.33时, 会有部分难以被还原的PdO物种形成. 因此我们推测α峰往高温迁移与Pd和载体间相互作用增强有关, 而β峰也可归属于与CeO2有强相互作用的PdOx物种的还原. 对所有新鲜催化剂来说, γ和δ峰的强度与峰温都没有明显的区别, 表明这两个峰的出现可能与La改性的Al2O3载体的还原有关. 研究发现[27], Pd和CeO2的相互作用可以使Pd簇稳定在Pd+氧化态, 而Pd+的存在可以有效地降低HC和NO的起燃温度(如图1). 除此之外, α峰与β峰往高温偏移说明Pd和CeO2的相互作用逐渐增强, 这也许是低温下CO和NO2转化率随着CeO2的添加而减少的原因.
1100 °C老化后, 催化剂的高温烧结造成活性PdOx物种分散度降低, 使得α峰与β峰的耗氢量减少. PdOx物种的烧结造成PdOx与载体的相互作用减弱, 因此与新鲜催化剂相比, α峰与β峰均移向较低温度. 但是在老化催化剂中, Pd和CeO2的强相互作用仍然导致了α峰与β峰随着CeO2的添加往高温偏移.
为了观察在催化反应中特定Pd物相的存在与催化反应之间的联系, 我们在不同反应条件下对催化剂进行了原位红外表征. 图7(a)和(b)为250 °C时新鲜催化剂Pd/La-A和Pd-Ce(2.0)/La-A在CO + HC + O2气氛下采集的谱图. 1800-2100 cm-1之间出现的峰归属于羰基物种在Pd0上的吸附. 其中2063 cm-1是线式吸附Pd-CO, 1972 cm-1是桥式吸附Pd2-CO, 1909 cm-1是三原子协同吸附Pd3-CO [28, 29], 说明部分PdOx物种在反应条件下会部分还原为金属态的Pd. Pd/La-A上这三种吸附峰强度明显大于在Pd-Ce(2.0)/La- A上的峰强, 表明Pd-Ce(2.0)/La-A被还原的金属Pd较少[28], 这说明少量CeO2的存在能促进反应条件下Pd0再氧化为PdO的过程. 此外, 我们很难在Pd-Ce(2.0)/La-A上看到Pd3-CO吸附物种, 同时桥式吸附与线式吸附的峰也往低波数偏移. 从文献[30]类推, Pd-Ce复合氧化物纳米颗粒的形成使得Pd可以在CeO2中稀释, 从而限制了Pd2-CO的形成并导致Pd3-CO的消失. 1200-1800 cm-1之间的峰归属于在Al2O3表面吸附的碳酸盐与羧酸盐物种[31, 32]. 这些物种的吸附峰强度随着时间的增加并没有明显的变化, 可能是因为催化剂在250 °C刚暴露于CO + HC + O2气氛中时, 碳酸盐与羧酸盐物种在Al2O3表面的吸附就几乎达到最大值. 这可能与CO在该温度下已经完全转化有关.
催化剂暴露在CO + HC + O2气氛中10 min后, 化学计量的NOx也快速加入了反应气氛中. 如图7(c)和(d)所示, 2253与2231 cm−1处的吸收峰分别代表NCO在八面体的Al3+与四面体的Al3+上的峰[33]. NCO物种是由吸附的N原子与未反应的CO相互作用而形成[34]. Na的形成有两种机理: (1)由NO的单分子解离吸附形成Na和Oa; (2)由NO与CO的双分子反应形成Na和CO2[35]. NCO的出现时机和峰强可以对NO分解的起燃温度作粗略考量. 两种NCO吸附峰的峰强在第2 min时就达到最大值, 然后随着反应NCO + NO + O2 → N2的出现而逐渐减弱. 值得注意的是, Pd-Ce(2.0)/La-A上NCO的吸附峰远远强于在Pd/La-A上的, 表明Pd-Ce(2.0)/La-A上形成了更多的NCO物种, 从而更有利于NOx在250 °C的转化, 这与催化活性的数据一致(图1). Pd-Ce(2.0)/La-A在2175 cm-1处出现一个新峰. 由H2-TPR结果可知催化剂中有两种不同的PdOx物种, 而2175 cm-1处的峰可能是NCO吸附在与CeO2有强相互作用的Pd0上所致. 与此同时, 在1610, 1575和1459 cm-1处也出现了三个峰, 并且在第5 min达到最大峰强. 其中1610 cm-1归属于双配位的硝酸盐物种, 1575和1459 cm-1归属于表面硝酸盐物种与HC反应形成的醋酸盐[36]. Pd-Ce(2.0)/La-A上硝酸盐和醋酸盐的吸附峰强度也大于在Pd/La-A上的峰强, 这与前者有更好的NO/HC催化活性相关. 此外, 与图7(a)和(b)相比, CO在Pd0上的吸附峰强度在通入NOx后变弱, 说明具有强氧化性的NO2的存在有助于稳定氧化钯物种, 特别是小的Pd簇.
为了阐明催化剂的反应机理, 新鲜催化剂在汽车尾气气氛中从50到400 °C进行了一系列原位红外的表征. 如图8所示, 50 °C时可以观测到在1200-1800 cm-1出现三个主要的吸附峰, 归属于在Al2O3上形成的结构各不相同的亚硝酸盐/硝酸盐[37, 38]. Pd-Ce(2.0)/La-A上这三个峰的强度明显强于Pd/La-A, 并且在50 °C时就几乎达到最大峰强. 说明CeO2的添加更利于载体对亚硝酸盐/硝酸盐的吸附, 进而有利于NO在低温下的转化. 这些峰随着温度的升高而减弱. 温度的升高使CO-NOx活性增强, NOx的脱附也增多, 这使得NOx在催化剂表面的吸附量减少, 因而单齿配位的亚硝酸盐物种(1319 cm-1)[37]随着温度的升高逐渐迁移向低波数并且在270 °C消失. 亚硝酸盐物种消失的同时, NCO物种的吸附峰(2100-2300 cm-1)和醋酸盐与硝酸盐的吸附峰(1400-1800 cm-1)开始出现, 并且随着温度的升高而增强. 如前所述, 这些峰的变化与HC和NOx的催化活性相关. 与Pd/La-A相比, Pd-Ce(2.0)/La-A上这些峰出现的温度更低, 峰更强, 表明其催化HC/NOx反应的活性更高. Pd/La-A上, 三个归属于CO在Pd0< /sup>上吸附的峰(2061, 1972和1911 cm-1)出现于190 °C, 说明反应气氛中部分PdOx物种被还原为金属态Pd. 然而在Pd-Ce(2.0)/La-A上几乎不能检测到CO的吸附峰, 说明CeO2的添加有利于PdOx物种的稳定.
以La-Al2O3为载体, 采用共吸附浸渍法制备了一系列不同CeO2含量的单Pd密偶催化剂. 结果表明, CeO2的添加明显提高了催化剂对HC和NO的催化活性. 这是由于PdOx和CeO2间的强相互作用改善了Pd0再氧化为PdO的能力, 同时增强了反应条件下硝酸盐, 亚硝酸盐和异氰酸盐在载体上的吸附. Pd-Ce(2.0)/La-A呈现了最好的HC和NOx的催化活性. 然而CeO2的添加抑制了CO和NO2的转化. Pd-Ce(4.0)/La-A对CO, HC和NOx的催化活性均为最差, 这可能是由于过量的CeO2会对Pd形成包覆, 从而抑制了其催化性能. 然而, 老化后的催化剂中, 适量的CeO2 (2%-4%)的添加明显拓宽了HC和NOx的操作窗口. 同时, CeO2的添加明显提高了催化剂对三种反应物的催化活性. 这是因为Pd-Ce界面上PdOx和CeO2间的强相互作用使得PdOx物种在1100 °C高温老化后仍能以小颗粒的形式分散在载体上, 从而显著地提高催化剂的热稳定性.
2014, Vol. 35 Issue (8): 1267-1280
PDF (1343 KB)
2014, Vol. 35 Issue (8): 1267-1280
PDF (1343 KB)