The catalytic oxidation of CO is an important reaction,not only because of its wide range of practical applications,such as indoor CO and automotive exhaust treatments [1, 2],but also because of its high value in understanding some fundamental concepts in catalysis,such as structure-sensitivity relationships and reaction mechanisms. Noble metals,including Au [3, 4, 5, 6, 7, 8, 9],Pt [10, 11, 12, 13],and Pd [14, 15, 16],and transition-metal compounds,including CuO [17, 18, 19, 20, 21, 22] and Co3O4 [23, 24],are the most commonly used catalysts for CO oxidation.
Although it is generally accepted that Au nanoclusters are the most active catalysts,because they can catalyze CO conversion to CO2 at −70 °C [4],highly active Pt catalysts can also be obtained using reducible supports such as TiO2 [25] and Fe2O3 [26]. The Pt activity strongly depends on the preparation method. For example,Qiao et al. [27] synthesized a single-atom Pt1/FeOx catalyst,which was very active in CO oxidation. The enhanced activity was attributed to the partly vacant 5d orbitals of high-valent Pt atoms,which helped to reduce both the CO adsorption energy and the activation barrier for CO oxidation. Another strategy for improve the activities of Pt catalysts is promoter addition. Alkali metals such as Na and K effectively enhance the activities of Pt catalysts in various reactions such as CO oxidation and preferential CO oxidation (PROX) [28, 29, 30],the water-gas-shift reaction [31, 32, 33],HCHO oxidation [34],and oxidation of dichloromethane [35]. The role of the alkali metal depends on the reaction; it either weakens the interactions between CO and Pt and changes the CO adsorption sites in PROX over Pt/Al2O3 catalysts [28],results in the generation of alkali-stabilized Pt-OHx species in Na- or K-promoted Pt/Al2O3 and Pt/SiO2 catalysts for the water-gas-shift reaction [32],or generates surface Pt-O-Kx species in Pt/Al2O3 catalysts,which accelerate the decomposition of formate intermediates formed on the Al2O3 surface [35].
It is worth noting that evidence of the promoting effects of alkali metals on Pt catalysts is mostly based on spectroscopic results [28, 34, 35]. Although such evidence is relevant,it is desirable to perform kinetic studies,because they can provide useful information on reactions,such as the elementary steps,rate expressions,and reaction mechanisms. The roles of alkali metals could be clarified by comparing the kinetic behaviors of different catalysts (e.g.,non-promoted and promoted catalysts). In the current work,we performed detailed kinetic studies of CO oxidation over a series of K-free and K-promoted Pt/Al2O3 catalysts. The obtained power-law rate expressions showed that the addition of K significantly influenced the reaction order with respect to CO,and that the apparent activation energy was lower on the K-containing catalysts than on the K-free ones. Rate equations based on reasonable elementary steps were derived and the calculated parameters such as intrinsic rate constants and equilibrium constants for CO and O2 further illustrated that the addition of K to the catalyst weakened the interactions between CO and the catalyst surface and eased the reaction between chemisorbed CO and O2. This conclusion was supported by in situ infrared spectroscopic results.
A series of K-promoted Pt/Al2O3 catalysts were prepared using an incipient wetness impregnation method. Commercial γ-Al2O3 with a surface area of about 180 m2/g was used as the support; it was calcined 500 °C for 4 h prior to use. The detailed preparation process was as follows. The γ-Al2O3 powder (10 g) was immersed in an appropriate volume of a mixed solution containing H2PtCl6 (1 mmol) and KNO3 (2.6 mmol). The mixture was left to stand for 4 h,dried at 120 °C overnight,and calcined in static air at 500 °C for 4 h. The sample was washed three times with deionized water at room temperature,filtered,and dried at 120 °C overnight. This catalyst had a nominal Pt content of 2 wt%. Catalysts with various Pt loadings and K-free Pt/Al2O3 catalysts were prepared in a similar manner,but with no K addition for the K-free catalysts. The obtained catalysts are denoted by xK-yPt/Al2O3,where x and y are the contents (wt%) of K and Pt,respectively,in the catalyst.
Inductively coupled plasma-atomic emission spectrometry (Optima 7300DV,Perkin Elmer) was used to determine the Pt and K contents of the catalysts. The Brunauer-Emmett-Teller (BET) surface areas of the catalysts were determined based on N2 adsorption at liquid-nitrogen temperature (-196 °C),using a surface area analyzer (Quantachrome Autosorb-1). The catalysts were pretreated at 120 °C for 6 h in a vacuum before the BET measurements.
The X-ray diffraction (XRD) patterns of the catalysts were recorded using a powder diffractometer (PANalytical XxPert PRO MPD),with Cu Kα radiation,operated at 40 kV and 40 mA. The patterns were recorded in 2q range = 10°- 80°,at a scanning speed of 0.15°/s.
High-resolution transmission electron microscopy (HRTEM) was performed using a JEM-2100F microscope with a field-emissive gun,operated at 200 kV,with a point resolution of 0.24 nm.
Pulse CO chemisorption was used to investigate Pt dispersion in the catalyst (Quantachrome CHEMBET-3000 instrument). Before the measurement,the sample was reduced in a 5% H2-95% N2 mixture with a total flow rate of 30 mL/min,at 300 °C for 1 h,followed by cooling to 30 °C in a pure He flow (99.999%,30 mL/min). Pulses of CO were then fed into the stream of carrier gas (high-purity He,99.999%,30 mL/min) using a precision analytical syringe. The Pt dispersion was calculated based on the assumption that the CO/surface Pt atom ratio was 1,and the Pt particle size was calculated based on the equation dPt (nm) = 1.1/D (D = dispersion).
The oxidation states of the catalysts were determined using X-ray photoelectron spectroscopy (XPS,ESCALAB 250Xi instrument),with an Al Kα X-ray source (1486.6 eV) at a pressure of about 2 × 10−7 Pa at room temperature and a pass energy of 20 eV. The binding energy (BE) of the C 1s core level,i.e.,284.6 eV,was used as an internal standard. Before the measurements,the samples were pretreated inside the XPS chamber with an 10% O2-90% Ar mixture (30 mL/min) at 300 °C for 0.5 h,followed by a 10% H2-90% Ar mixture (30 mL/min) at 300 °C for 0.5 h. This in situ treatment was used to keep the chemical states of the catalysts the same as those under the catalyst-testing conditions (see section 2.3).
In situ Fourier-transform infrared (FTIR) spectra of the samples were recorded using a NEXUS670 spectrometer equipped with a mercury cadmium telluride detector. A self-supported wafer (diameter 16 mm) was placed in a quartz IR cell connected to the closed circulation system. The sample was pretreated under a pure O2 flow (99.95%,30 mL/min) at 300 °C for 0.5 h,followed by a pure H2 flow (99.99%,30 mL/min) at 300 °C for 0.5 h. It was then purged with pure He (99.95%,30 mL/min) at 300 °C for 10 min and cooled to 40 °C; the background spectrum was then recorded. A pure CO flow (99.95%,30 mL/min) or a mixture of 1% CO-1% O2-98% N2 (total flow rate 30 mL/min) was then introduced until saturated CO chemisorption was obtained. The sample was then purged with pure He to remove the gas-phase and physisorbed CO molecules,and the spectrum was recorded at 40 °C in a He flow. Temperature-dependent FTIR spectra were recorded during the reaction after holding each temperature point for 10 min.
Catalytic CO oxidation was performed using a fixed-bed microreactor with an inner diameter of 6 mm,using a catalyst (100 mg) of particle size 0.12-0.15 mm. The catalyst was diluted with quartz sand of the same mesh size to 0.25 mL. A thermocouple was placed in the middle of the catalyst bed to monitor the reaction temperature. Before the reaction,the sample was pretreated with an 10% O2-90% Ar mixture (30 mL/min) at 300 °C for 0.5 h (to clean the catalyst surface),followed by a 10% H2-90% Ar mixture (30 mL/min) at 300 °C for 0.5 h (to reduce the catalyst). The catalyst was then cooled to a certain temperature and a feed gas containing 1% CO-1% O2-98% N2 was introduced. The total flow rate was 40 mL/min,which corresponds to a space velocity of 9600 h−1. The CO concentration in the reactor effluent was determined using an Agilent 6850 gas chromatograph equipped with a thermal conductivity detector attached to an HP PLOT column (30 m × 0.32 mm × 12 μm).
The kinetic study was performed using the same fixed-bed reactor as was used for catalytic CO oxidation. The feed gases (i.e.,CO,O2,and N2) were controlled using mass flow controllers and mixed before they reached the reactor inlet. In general,the reactor was operated at a CO conversion of less than 15% to ensure a differential mode. The reaction conversion was controlled by changing the catalyst load; it was diluted with quartz sand to a volume of 0.1 mL. The absence of mass transport resistances was checked based on the Weisz-Prater criterion for internal diffusion and Mears’ criterion for external diffusion; the absence of heat transfer was checked based on Mears’ criterion [36]. For the 0.42K-2Pt/Al2O3 catalyst,the calculated values under kinetic conditions were 4.30 × 10−4 for the Weisz-Prater criterion for internal diffusion,7.94 × 10−4 for the Mears’ criterion for external diffusion,and 2.97 × 10−5 for the Mears’ criterion for heat transfer. The partial pressure dependences of the reaction rates were measured by adjusting the flow rate of 10% CO-90% Ar or 10% O2-90% Ar while keeping the total flow rate at 120 mL/min,by adjusting the flow rate of pure Ar. The concentrations of CO and O2 in the feed were varied between 0.1% and 3%. Each value was recorded after about 1 h as the reaction reached steady state. The CO2 concentration in the outlet gas stream was determined using the Agilent 6850 gas chromatograph mentioned above.
The definition of CO conversion is as follows:
The reaction rate (rCO) is defined as
The power-law rate expressions were derived by linear regression fitting of the data (partial pressures and reaction rates) using the POLYMATH 5.1 program [37].
The general physical properties of the catalysts are summarized in Table 1. All the catalysts have similar surface areas (166-180 m2/g). However,the actual contents of K in the catalysts are much lower than the nominal values,because most of the K salt was removed during washing. The Pt dispersion gradually decreases with increasing Pt content in the catalyst. For example,the Pt dispersion in 0.1Pt/Al2O3 is 56.1%,whereas that in 2Pt/Al2O3 is 44.0%. Moreover,the addition of K slightly enhances Pt dispersion; the 0.42K-2Pt/Al2O3 catalyst has a Pt dispersion of 56%,which is higher than that of 2Pt/Al2O3 (44%). Consequently,the calculated Pt particle size increases with increasing Pt content in the catalyst,ranging from 2.0 to 2.5 nm. However,the Pt dispersions of the K-promoted samples may have large deviations because of coverage of the Pt surface by K species.
The XRD patterns of the catalysts (not shown) show no distinct diffraction peaks from either Pt or K; this suggests that these species are highly dispersed.
Figure 1 shows representative TEM images of some samples. The presence of metallic Pt entities is verified by the measured d-space distances of the particles (Pt(111) = 0.226 nm). The detected Pt particle sizes in the investigated samples are similar (2-3 nm,based on analysis of more than 100 particles),and are consistent with the results obtained using CO chemisorption (Table 1).
The oxidation states of the catalysts were determined using XPS; the results are shown in Fig. 2. The Al 2p line overlaps with the Pt 4f one,therefore the Pt 4d line was used. Catalysts with low and high Pt contents (0.2 and 2.0 wt%) were examined. A broad band at bond energy (BE) = 310-322 eV is observed in the Pt 4d spectra of the samples (Fig. 2(a)). The peak was deconvoluted into two components with BEs = 313.8 and 316.3 eV. The peak at 313.8 eV is assigned to metallic Pt (Pt0) [38],and that at 316.3 eV is typical of PtO species [39]. On addition of K to the catalysts,the Pt0 peaks shift to higher BE (314.3 eV) compared with that of the K-free samples (313.8 eV). This shift suggests interactions between Pt and K species. The Pt0/Pt2+ ratios in these catalysts were also calculated,and are listed in Fig. 2. For the catalysts with low Pt contents,the Pt0/Pt2+ ratio in 0.05K-0.2Pt/Al2O3 is 1.20,which is essentially the same as that in 0.2Pt/Al2O3 (1.26). However,for catalysts with high Pt contents (2.0Pt/Al2O3 and 0.42K-2.0Pt/Al2O3),K addition significantly affects the surface compositions of Pt species; the Pt0/Pt2+ ratio (1.37) for 0.42K-2.0Pt/Al2O3 is much lower than that for 2.0Pt/Al2O3 (2.55). The lower Pt0/Pt2+ ratio strongly implies interactions between surface K and Pt species,resulting in Pt species remaining cationic. The broad peak in the O 1s spectra (Fig. 2(b)) can be deconvoluted into two peaks centered at BEs = 529.5 and 531.3 eV; these can be assigned to lattice oxygen species (Olatt) and adsorbed oxygen species or surface hydroxyl groups (Oads) [40],respectively. The calculated Olatt/Oads ratios in the catalysts show that for samples with low Pt contents,these ratios are similar (0.20 for 0.2Pt/Al2O3 and 0.22 for 0.05K-0.2Pt/Al2O3),whereas for samples with high Pt contents,the ratio is higher for the K-containing catalyst than the K-free one (0.08 for 2.0Pt/Al2O3 and 0.14 for 0.42K-2.0Pt/Al2O3). These results suggest that the addition of K decreases the number of hydroxyl groups on the catalyst surface.
Figure 3 shows the light-off curves for CO oxidation over various catalysts. The catalytic performances of 0.1Pt/Al2O3 and 0.02K-0.1Pt/Al2O3 are almost identical (Fig. 3(a)). With increasing K and Pt contents in the catalysts,the K-promoted samples show better performances than the K-free ones (Figs. 3(b)-(d)). This point is clearly illustrated by a comparison of the T50 temperatures (the temperature at which the CO conversion is 50%); ΔT50 (= T50 on the K-containing sample −T50 on the K-free sample) gradually increases with increasing K and Pt contents in the catalysts,strongly indicating that the addition of K significantly enhances the activity. The turnover frequencies (TOFs) were also calculated and compared with those reported in the literature. The TOF obtained on the 0.2Pt/Al2O3 at 150 °C is 0.02 s−1,which is lower than that reported by Allian et al. [41] (0.08 s−1) on Pt/Al2O3 catalysts with various Pt dispersions from 0.05 to 0.87. The different TOFs may result from the different preparation methods. In Ref. [37],the Pt/Al2O3 catalysts were prepared by slurry impregnation and colloidal methods,whereas the catalysts used in the current work were prepared using an incipient wetness impregnation method.
Kinetic investigations were conducted on four representative catalysts: 0.2Pt/Al2O3,0.05K-0.2Pt/Al2O3,2.0Pt/Al2O3,and 0.42K-2.0Pt/Al2O3. The partial pressures of CO and O2 were varied between 0.304 and 3.04 kPa,and the reaction temperatures were kept low,to obtain reliable data. The detailed results are shown in Tables 2 and 3. These data were used to determine the dependences of the reaction rates on the CO and O2 partial pressures; the results are shown in Fig. 4. Generally,the reaction rate decreases with increasing CO partial pressure,and increases with increasing O2 partial pressure. The apparent rate constants (kapp),and CO and O2 reaction orders are listed in Table 4. The obtained power-law rate expressions are r = 1.81 × 10−7 pco−0.51po20.73 for 0.2Pt/Al2O3,r = 2.52 × 10−7 pco−0.27po20.59 for 0.05K-0.2Pt/Al2O3,r = 2.56 × 10−7 pco−0.53po20.70 for 2.0Pt/Al2O3,and r = 6.55 × 10−7 pco−0.22po20.63 for 0.42K-2.0Pt/Al2O3.
The activation energies for the catalysts were also calculated based on Arrhenius plots (Fig. 5(a)). The activation energies for 0.2Pt/Al2O3 and 2.0Pt/Al2O3 are similar (60-66 kJ/mol). The activation energies of the K-promoted samples are lower than those for the K-free ones (55.2 ± 5.6 kJ/mol for 0.05K-0.2Pt/Al2O3 and 51.6 ± 3.6 kJ/mol for 0.42K-2.0Pt/Al2O3). The obtained values are higher than those for Pt/SiO2 (13-22 kJ/mol [42]) and Pt/TiO2 (41.5 kJ/mol [43]),and are comparable with those obtained for Pt/Al2O3 (80-90 kJ/mol [41]; 65 kJ/mol [44]). The higher activation energies for 2.0Pt/Al2O3 and 0.2Pt/Al2O3 compared with those for their K-promoted counterparts suggest that the former are less active than the latter. Moreover,parity plots (Fig. 5(b)) indicate that the experimental reaction rates are in good agreement with the calculated values; this validates the derived rate expressions.
The elementary steps in CO oxidation over these catalysts were derived based on the rate expressions. The mechanisms of CO oxidation over Pt catalysts have been extensively studied,and various elementary steps have been proposed [41, 45]. Classic Langmuir-Hinshelwood models have been used to deduce the elementary steps. For example,Djéga-Mariadassou and Boudart [45] proposed a reaction model in which molecular adsorption of O2 was considered to be the sole kinetically relevant step and the reaction order of O2 was 1. However,as the reaction orders of O2 are 0.6-0.7,it is unlikely that chemisorption of O2 on the Pt surface is the rate-determining step (RDS) in our study.
The following elementary steps are proposed to describe the reaction pathways over Pt/Al2O3 catalysts; they are similar to those proposed by Allian et al. [41].
A general rate reaction was derived based on the elementary steps represented by reactions 1.1-1.5:
If we take a close look at the kinetic behaviors of the catalysts,obvious differences can be observed. First,the apparent rate constants kapp (Table 4) for the K-promoted catalysts are 2.52 × 10−7 for 0.05K-0.2Pt/Al2O3 and 6.55 × 10−7 for 0.42K-2.0Pt/Al2O3; these are much larger than those for the K-free catalysts (1.81 × 10−7 for 0.2Pt/Al2O3 and 2.56 × 10−7 for 2.0Pt/Al2O3). These results suggest that K promotes the reaction. Secondly,the reaction orders of CO are about −0.5 for the K-free samples,i.e.,much lower than those for the K-containing ones (about −0.2). According to Eq. 1,a lower reaction order implies stronger interactions between CO molecules and the catalyst surface. Thirdly,the reaction orders of O2 for the K-promoted samples are slightly lower than those for the K-free ones,implying that K promotion accelerates competitive O2 chemisorption on the Pt surface. This enhancement of O2 chemisorption is probably the result of lower CO coverage of the Pt surface as a result of K promotion. The proposed elementary steps (reactions 1.1-1.5) suggest that the differences between the reaction orders of CO and O2 must be related to changes in their equilibrium constants (K1 for CO and K2 for O2) and the intrinsic rate constant k3. The kinetic parameters (K1,K2,and k3) were therefore calculated based on Eq. 1 and the data in Tables 2 and 3,using POLYMATH 5.1. The regressed results are listed in Table 5. The intrinsic rate constant k3 for 0.2Pt/Al2O3 (1.99 μmol/(g·s)) is close to that for 0.05K-0.2Pt/Al2O3 (2.09 μmol/(g·s)),suggesting similar intrinsic activation ene rgies for these catalysts. However,k3 for 2.0Pt/Al2O3 (2.90 μmol/(g·s)) is only half that for 0.42K-2.0Pt/Al2O3 (5.86 μmol/(g·s)),suggesting a higher intrinsic activation energy. These observations are consistent with the apparent activation energies (kapp). The addition of K significantly affects the CO equilibrium constant (K1). For 0.2Pt/Al2O3,K1 is 11.7 kPa−1,which is almost twice that for 0.05K-0.2Pt/Al2O3 (5.95 kPa−1). Similar results were obtained for 2.0Pt/Al2O3 and 0.42K-2.0Pt/Al2O3 (10.14 kPa−1 vs 4.84 kPa−1). The lower K1 values for the K-containing samples indicate a lower adsorption heat (ΔH) compared with those of the K-free samples. The adsorption heat for linear adsorbed CO on the Pt surface is significantly lowered by K addition to the catalyst,as reported by Derrouiche et al. [46]. For the O2 equilibrium constant (K2),addition of K slightly enhances K2. For example,K2 is 1.58 kPa−1 for 0.05K-0.2Pt/Al2O3 and that for 0.2Pt/Al2O3 is 1.34 kPa−1. These findings clearly indicate that K addition significantly suppresses CO chemisorption (a lower K1),but slightly enhances O2 chemisorption (a higher K2). Moreover,K1 is much larger than K2,suggesting that the interactions between CO and the catalyst surface are much stronger than those between O2 and the catalyst surface; this is understandable,because CO molecules are easily chemisorbed on Pt,with a high heat of adsorption [47].
Although the kinetic results give direct evidence on the promoting effect of K addition on CO oxidation,spectroscopic results on CO chemisorption are desirable because CO adsorption is often used as a probe for investigating catalyst structures and interactions between CO and metals. There are distinct differences between the kinetics of the 2.0Pt/Al2O3 and 0.42K-2.0Pt/Al2O3 catalysts; therefore,it is interesting to compare their CO chemisorption behaviors using FTIR spectroscopy. Parallel experiments on CO + O2 chemisorption were also performed to investigate the effect of O2 on the chemisorption behaviors.
Fig. 6(a) shows the spectrum for CO chemisorption on 2.0Pt/Al2O3. A strong band at 2047 cm−1 and a weak,broad band at about 1816 cm−1 are observed,together with a very weak shoulder at about 2085 cm−1; these can be assigned to linear CO on Pt0 sites (L-CO),bridging CO on Pt0 sites (B-CO),and linear CO on Pt2+,respectively [48]. These spectra are consistent with the findings reported by Alexeev et al. [48] and Kuriyama et al. [28]. CO adsorption on different Pt species is consistent with the XPS results (Fig. 2(a)),and clearly confirms the presence of both Pt0 and Pt2+ species in the catalyst. The intensities of all these bands decrease with increasing temperature,but they have not completely disappeared at 180 °C. L-CO adsorption on Pt is very strong,and it can only be completely removed at 400 °C [48]. In addition,the L-CO band at 2047 cm−1 shifts slightly to lower wavenumbers (about 4 cm−1) with increasing temperature; this can be explained by a decrease in the dipole-dipole coupling between adsorbed CO molecules resulting from a decrease in the CO surface coverage [48]. In contrast,the B-CO band shows a significant red-shift with increasing temperature. When the catalysts are exposed to a mixture of CO-O2 (Fig. 6(b)),spectra with peaks of almost identical shapes and positions are observed. The spectra of 0.42K-2.0Pt/Al2O3 are similar to those of 2.0Pt/Al2O3,for both CO and CO-O2 (Fig. 7). It is worth noting that the spectra of 0.42K-2.0Pt/Al2O3 are different from those reported for other K-promoted Pt/Al2O3 catalysts [28, 46],i.e.,a band at about 1760 cm−1 was observed; this was assigned to three-coordinated CO species on Pt atoms interacting with K species [46]. The absence of such a band at 1760 cm−1 in our work might be caused by the low K content in the catalyst. Note that the catalyst in the current work has a K/Pt molar ratio of 1,whereas the K/Pt ratios of the catalysts in refs. [28] and [46] were 10. The much higher K/Pt ratio may result in more interactive K-O-Pt species in the reported catalysts,which would make CO chemisorption on these species more distinct.
Although the features of the FTIR spectra of the investigated catalysts are similar,the peak intensities of the bands change at high temperatures,and the rates at which the band intensities decrease are different. We therefore calculated the normalized peak areas of the L-CO and B-CO bands in the catalysts. Fig. 8 compares the band areas of L-CO and B-CO under exposure to CO and CO-O2 on the 2.0Pt/Al2O3 and 0.42K-2.0Pt/Al2O3 catalysts. For the chemisorbed L-CO species (Fig. 8(a)),the area for the 0.42K-2.0Pt/Al2O3 catalyst is significantly lower than that for 2.0Pt/Al2O3. Moreover,chemisorbed L-CO is not completely removed at 180 °C from 2.0Pt/Al2O3,but is completely removed from 0.42K-2.0Pt/Al2O3. These results clearly suggest that the presence of K in the catalyst suppresses CO chemisorption and weakens the interactions between CO and the catalyst surface; this is in good agreement with the findings reported in the literature [28, 46]. For the chemisorbed B-CO species (Fig. 8(c)),the presence of K in the catalyst enhances CO chemisorption,particularly at low temperatures (<100 °C). This phenomenon has also been reported in the literature; Derrouiche et al. [46] found that the heat of adsorption of a major B-CO species was significantly larger in the presence than in the absence of K. Note that the contribution of B-CO is much lower than that of the L-CO,therefore the total L-CO + B-CO peak areas are higher for 2.0Pt/Al2O3 than for 0.42K-2.0Pt/Al2O3 (Fig. 8(e)); this observation is not consistent with the Pt dispersion measurements (Table 1),because the pulsed CO chemisorption results show that the Pt particles in 0.42K-2Pt/Al2O3 are smaller (higher amount of chemisorbed CO) than those in 2Pt/Al2O3 (lower amount of chemisorbed CO). This inconsistency may arise from the different temperatures used for CO chemisorption in the two sets of experiments (40 °C for IR spectroscopy and 30 °C for the Pt dispersion measurements); alternatively there may be considerable errors in the Pt dispersions results because of the low Pt loadings in the catalysts.
When the catalysts are exposed to a CO-O2 mixture (Figs. 8(b),(d),and( f)),the general features are similar to those in a CO environment. However,the addition of O2 further accelerates CO desorption. The temperature for complete removal of L-CO shifts to 160 °C,and that for B-CO shifts to 140 °C. These are about 20 °C lower than those in a CO environment. This difference implies that the adsorptions of CO and O2 are competitive,and coadsorption of O2 weakens the interactions between CO and the surface. In addition,note that the total area of the CO peak for 0.42K-2.0Pt/Al2O3 (Fig. 8(f)) is slightly lower than that for 2.0Pt/Al2O3,indicating that O2 coverage on the catalyst surface is quite low.
The analyses based on the FTIR results are in excellent agreement with the kinetic results. The addition of K to the catalysts results in significant suppression of CO chemisorption and consequently a lower adsorption equilibrium constant (K1) for both CO and CO-O2 adsorptions (Table 5). However,the addition of O2 only changes the CO adsorption,because the equilibrium constant (K2) for O2 chemisorption is very low,therefore O2 adsorption is weak (Table 5). Although the results of the current work provide convincing evidence for the role of K,the nature of the surface species (Pt-O-K) remains unclear. Nevertheless,some progress has been made. For example,in our previous work [35],a surface species Pt-O-Kx (x ≈ 2) was proposed for K-promoted Pt/Al2O3 catalysts,based on the surface compositions of K and Pt. Moreover,using density functional theory calculations,Zhai et al. [32] proposed that a PtK6O4(OH)2 species was the most active in the water-gas-shift reaction over a K-promoted Pt/Al2O3 catalyst. The authors [32] concluded that the role of the alkali metal (Na and K) was to generate and stabilize Pt-OHx species,which could be activated by CO at low temperatures. Although the structures of the active sites could not be determined in the current work,the participation of such Pt-associated hydroxyl groups in CO oxidation cannot be ruled out. Xu et al. [49] observed direct evidence for the interfacial oxidation of CO with hydroxyls over a FeO(111)/Pt(111) inverse model catalyst,and they concluded that an interfacial COads + OHads reaction producing CO2 occurred easily at the Pt-oxide interface at low temperatures.
A combination of detailed kinetic and spectroscopic investigations clarified the role of K in the enhancement of CO oxidation over Pt/Al2O3 catalysts. The addition of K significantly weakens CO chemisorption,leading to a lower equilibrium constant compared with that of the K-free sample. Such weakened interactions between CO and the catalyst surface lower the surface coverage by CO,facilitate competitive O2 chemisorption on the Pt surface,and significantly lower the reaction barrier between chemisorbed CO and O2 species.
2015, Vol. 36 Issue (11): 1976-1986
PDF (745 KB)
2015, Vol. 36 Issue (11): 1976-1986
PDF (745 KB)