5.2. Loss in active surface area and sintering of metal particles
The effects of ageings on the changes in precious metal dispersions and active metal surface areas of catalysts were studied by chemisorption. Figure 25 presents the metal dispersions of the catalysts as a function of ageing temperature. The corresponding active metal surface areas are presented in Table 13. As can be seen in Fig. 25 and Table 13, precious metal dispersions and metal surface areas decreased as a function of ageing temperature and time. This was observed in all ageing atmospheres. Similar behaviour in precious metal dispersion values was observed after oxidative and reductive ageings, within the limits of the measurement accuracy. The similar behaviour between these two ageing atmospheres is not self-evident because there were significant differences in the surface areas of corresponding catalysts, i.e. total surface areas remained high after the reductive ageing, as reported in section 5.1. In this respect, the chemisorption results (the loss of active metal area) do not correlate with the BET surface areas of catalysts.
Figure 25. The effect of ageing atmosphere, temperature and time on the precious metal dispersions of catalysts (based on H2-chemisorption). Comparison between reductive (3 hours (●) or 24 hours (○)) and oxidative (3 hours (▪) or 24 hours (□)) ageing atmospheres.
The precious metal dispersion and the active metal surface area measured by H2-chemisorption for the fresh catalyst were 28% and 125 m2/g, respectively. Table 13 presents a summary of the chemisorption data. As indicated, the loss of metal surface area rose with the increasing ageing temperature. The H2/800°C/3h-aged and air/800°C/3h-aged catalysts showed approximately 50% reduction in the metal surface area compared to the fresh catalyst. After reductive and oxidative ageings at 850°C for 3 hours, approximately two thirds of the original active surface area was lost. At higher temperatures (T>900°C) and after engine and vehicle ageings, the dispersion values were so small (<1%), that it is difficult to draw any reliable conclusions. It can only be stated that ageings clearly affected the changes in the active phase of the catalyst.
Table 13. Catalyst characterization by H2-chemisorption for fresh, engine-aged, vehicle-aged and thermally aged catalysts.
| Catalyst/ Ageing conditions | Dispersion (%) | Active metal surface area (m2/g metal) | Monolayer capacity (cm3/g) | |
|---|---|---|---|---|
| Fresh | 28 | 125 | 0.09 | |
| Engine-aged | < 1 | <1 | < 0.01 | |
| Vehicle-aged | < 1 | <1 | < 0.01 | |
| 800°C/3 h | Air | 16 | 70 | 0.05 |
| H2 | 14.5 | 65 | 0.05 | |
| H2O-air | < 1 | <1 | < 0.01 | |
| H2O-H2 | < 1 | <1 | < 0.01 | |
| 800°C/24 h | Air | 5.7 | 25 | 0.02 |
| H2 | 5.2 | 23 | 0.02 | |
| 850°C/3 h | Air | 10.5 | 47 | 0.03 |
| H2 | 9.1 | 41 | 0.03 | |
| 850°C/24 h | Air | 1.4 | 6.4 | 0.01 |
| H2 | 0.4 | 1.7 | < 0.01 | |
| 900°C/3 h | Air | < 1 | <1 | < 0.01 |
| H2 | < 1 | <1 | < 0.01 |
The effect of ageing time was also important in this case, as seen in Fig. 25. The results of this study indicated the rapid growth of metal particles during the first few hours of ageing, which is convergent with the observations after BET measurements. The significant differences between H2/800°C/3h- and H2/800°C/24h-aged catalysts as well as between H2/850°C/3h- and H2/850°C/24h-aged catalysts were observed. The same is true also for the air-aged catalysts (see Fig. 25). That is, the increase of ageing time from 3 hours to 24 hours led to an approximately 35% decrease in the dispersion value at an ageing temperature of 800°C.
The effect of hydrothermal ageing is rather different compared to thermal ageings with dry gases because the H2O-air/800°C/3h-aged and H2O-H2/800°C/3h-aged catalysts already showed small dispersion values (<1%). Thus, the role of water vapour is essential and it was clearly indicated that the sintering of metal particles is promoted in the presence of water vapour. This is consistent with the results of Bartholomew (2001), who concluded that the presence of water vapour increases the mobility of metal atoms or particles on the washcoat and hence accelerates the sintering of these particles. This mobility of metal particles is also related to the vapour pressures of metals or metal oxides (Barbier & Duprez 1994). This is further discussed in section 5.8, where the volatilization of Pd as a possible deactivation mechanism is considered.
The clearly lower dispersion values compared to the fresh catalyst at an ageing temperature of 800°C and above, gave evidence for the mechanism of catalyst’s deactivation during the ageing treatments. The loss in active metal surface area and low dispersion values are closely associated with the sintering of active metal particles in the washcoat. As the dispersions had already decreased after the ageings at 800°C in all ageing atmospheres, sintering of metal particles is an essential cause of deactivation, and it had already started at temperatures below 800°C. As reported by Usmen et al. (1992) and Teixeira & Giudici (1999), high temperatures favour the sintering of the active phase. Particle growth and changes in the size distribution of metal particles were thermally activated and resulted in a loss of the catalyst’s activity, and often in changes in the selectivity as well. Washcoat materials, especially thermal stabilizers such as La, enhance the catalyst’s performance by inhibiting the surface diffusion of metallic particles, which reduces the sintering of precious metals upon high temperatures. (Oudet et al. 1989, Teixeira & Giudici 1999)
Figure 26 presents the normalised active metal area as a function of ageing time. It can be noted that the rate of sintering of active metal particles on the surface was most dominant after the first few hours of ageing. On the other hand, as also shown in Fig. 26, the asymptotic approach has not yet been achieved after 42 hours of ageing below 850°C. According to the H2-chemisorption results, no differences were found on sintering rates between the reductive and oxidative ageing atmospheres with dry or wet gases. One possible explanation for this may be the low dispersion value of the fresh catalyst, which led to dispersion values below 1% at ageing temperatures of 900°C and above. Therefore, a full comparison of dispersion values cannot be carried out in the ageing temperature range of 800°–1200°C.
Figure 26. Effect of ageing atmosphere, temperature and time on the sintering behaviour of the aged catalysts. The normalised active metal surface area D/D0 as a function of ageing time (hours).
Additionally, on the basis of chemisorption results, conclusions cannot be drawn about which of the two metal particles (rhodium or palladium) were sintered. It is most likely that both Pd and Rh metal particles sintered at high ageing temperatures. Since chemisorption measurements are based on the assumption of monolayer adsorption, it can only be concluded that precious metal particles on the surface are sintered, which is observed as a decrease in dispersion values. Sintering of active metal particles will be discussed later in this thesis, in section 5.5, where the ageing-induced changes in the chemical states of the active metals are determined by the XPS measurements. There, it will be reported that rhodium metal particles in the washcoat were sintered and caused a decrease in catalytic activity. Furthermore, based only on chemisorption results, it is rather difficult to conclude that the sintering of the precious metal particles is a cause of the loss of active metal area. It is also possible that the loss of active metal area is caused by the encapsulation of metal particles in the sintered washcoat pores. In order to discriminate between these two causes of deactivation, TEM micrographs would have been necessary.
The H2-chemisorption results are considered in this thesis only qualitatively by comparing the dispersion values of the catalysts after the ageings. This is reasonable, because several assumptions have to be made considering the adsorption stoichiometry between the metal and adsorbate gas, as reported earlier in section 4.3.3. The stoichiometry ratio is also dependent on the particle size. Furthermore, in the case of catalysts containing CeO2, the extensive adsorption of H2 on CeO2 has been duly reported (Bernal et al. 1999). This spillover effect of hydrogen makes the quantitative interpretation of chemisorption results difficult. However, the amount of H2 spilled over the washcoat can be minimised by lowering the adsorption temperature (Fornasiero et al. 1997, Fornasiero et al. 1999). Recently, Graham et al. (1999) and Di Monte et al. (2000) have also shown that metal particles may be encapsulated in high temperature reduction. Therefore, the choice of an appropriate metal particle size may be an important factor to avoid deactivation.
In this work, the aged catalysts were also studied by CO-chemisorption in order to obtain results without the possible spillover effect of hydrogen. The results of CO-chemisorption are consistent with the H2-chemisorption measurements, (Lassi et al. 2002a) and no significant differences were observed between these two adsorbing gases. The precious metal dispersions for the fresh catalysts were 28% and 30% measured by H2- and CO-chemisorption respectively. In addition, no significant differences were found in the resolution (detection limits), and in both cases the precious metal dispersions were under 1% after the ageings at temperatures above 900°C. For these reasons, CO-chemisorption results will not be further discussed in this thesis.
In summary, low dispersion values of the aged catalysts proved that active metal particles were sintered. The rate of sintering of metal particles increased as a result of combined action of water vapour and high ageing temperature. Based on the chemisorption results, it is impossible to state which of the two metal particles (Rh or Pd) was most sintered. It can be concluded, however, that sintering of Pd and/or Rh particles was affected by ageing temperature, atmosphere and time. In particular, the presence of water vapour during the ageing increased the sintering rate. In contrast to the BET results, no differences were found between reductive and oxidative ageing atmospheres with dry gases. Rather, hydrothermally aged catalysts showed low dispersion values already at an ageing temperature of 800°C, which is consistent with the BET results. In order to understand the behaviour of active metals during the ageings, XPS was also used as a characterization technique in this research.