ORIGINAL PAPER Hydrothermal Aging-Induced Changes in Washcoats of Commercial Three-Way Catalysts Anna Fathali • Louise Olsson • Fredrik Ekstro¨m • Mats Laurell • Bengt Andersson Published online: 7 March 2013 � Springer Science+Business Media New York 2013 Abstract In order to quantify hydrothermal aging effects on the mono- and bi-metallic Pd and Rh supported on cerium-zirconium promoted alumina commercial three- way catalysts (TWCs), the catalysts were tested both fresh and after accelerated hydrothermal aging. The catalytic tests showed clear deactivation of the aged samples and influence on TWC’s property, such as: light-off tempera- ture, specific surface area (BET), dispersion of noble metals, oxygen storage capacity, oxygen storage capacity complete and labile oxygen storage capacity. Water inhi- bition have been investigated and confirmed for the per- formance of NOx reduction of the fresh catalyst. X-ray photoelectron spectroscopy was used to study changes in the oxidation state of the Pd and Rh present in the washcoat of the catalyst, before and after hydrothermal aging. Ele- ment maps by SEM-EDX analysis was perform in order to characterize the catalysts morphology, the surface’s com- position and element distribution. Keywords Automotive exhaust control � TWC � Aging � Palladium � Rhodium � OSC � BET 1 Introduction Deactivations of noble metals due to sintering have been investigated for Pt/Al2O3 [1], Pd/Al2O3 [2], Pt/Ce/Al2O3 [3], Pd/Rh/CeZr/Al2O3 [4] and commercial TWCs [5–8]. However, most literature focus on the characterization of the aging process. In an industrial context it is of outermost importance to be able to quantify the effect of aging. The objective with this work is characterization and quantifi- cation of three-way catalysts (TWC) properties, such as activity, selectivity, dispersion, BET and oxygen storage capacity (OSC) as well as determining design and com- position of mono- and bi- metallic catalysts. In this paper we focus only on thermal effects on catalyst deactivation. The thermal processes including changes of metallic structure, sintering, oxidation and creation of noble metal alloys are considered. The exhaust composition were selected to correspond to gasoline and NGV fueled engines. Since methane reactions were deactivated more severe, one additional study with only methane as reducing agent was performed in order to quantify the catalyst aging for methane reactions. 2 Experimental 2.1 Materials Two Euro5 commercial automotive exhaust catalysts from the same supplier were examined and compared. The first catalyst contained Pd supported on c-Al2O3 washcoat stabilized with BaO and La2O3. The second catalyst contained Pd/Rh sup- ported on BaO and La2O3 stabilized c-Al2O3 with added Ce0.75Zr0.25O. The noble metal content of the washcoat was 3 % Pd and for both catalysts with added 0.3 % Rh for the A. Fathali � F. Ekstro¨m � M. Laurell Volvo Car Corporation, 405 31 Go¨teborg, Sweden A. Fathali � L. Olsson � B. Andersson Department of Chemical Engineering, Chalmers University of Technology, 412 96 Go¨teborg, Sweden A. Fathali (&) � L. Olsson � B. Andersson Competence Centre for Catalysis, Chalmers University of Technology, 412 96 Go¨teborg, Sweden e-mail:
[email protected] 123 Top Catal (2013) 56:323–328 DOI 10.1007/s11244-013-9974-8 Pd/Rh catalyst and the washcoat was supported by a cordierite monolith (400 cpsi). 2.2 Aging Procedure Hydrothermal aging (16 h, 970 �C) was performed in oxidative (H2O/O2) aging atmosphere in the presence of 10 vol.% water vapor and 2.5 vol.% O2 by the supplier. 2.3 Characterizations Methods In order to characterize the fresh and aged catalysts, vari- ous characterization methods were used: the specific sur- face area BET (ASAP 2000 Analyser), X-ray photoelectron spectroscopy (XPS, Perkin Elmer PHI 5000C ESCA sys- tem), scanning electron microscope (SEM: Joel, model: JSM JSM-6480LV, monochromatic Al Ka radiation and 45� take off angle) with an energy dispersive X-ray ana- lyzer [EDX, Bruker system: Quantax 800, detector X-flash 4010, resolution 127 eV (Mn peak)] and dispersion mea- surements (selective CO and CO2 chemisorptions method). 2.3.1 Dispersion Measurements by the Selective Chemisorptions Method Selective chemisorptions method (CO and CO2) has been used to determine surface area of noble metals. The method has been proposed by Takeguchi et al. [9]. CO adsorption is carried out at room temperature; the error introduced by the adsorption of CO on cerium was minimized by introducing CO2 that adsorb on cerium sites. For the case of Pd/Rh a 1:1 stoichiometry (corresponding to the on-top form) is used [10]. 2.4 Activity Test The three-way catalytic activities have been examined by light off measurements performed in a temperature programmed continuous-flow. Feed gases were introduced to the reactor via mass flow controllers. The space velocity (SV) was 50,000 h-1 and heating rate (HR) 8 �C min-1 (50 ? 650 ? 50 �C). Experiments were conducted using different gas mixtures (see Table 1 below): rich (k = 0.99), stoichiometric (k = 1), lean (k = 1.01) and SAI, where SAI denotes the condition of providing ultra lean conditions (k = 2.01). 2.4.1 NOx Reduction by CH4 with and Without H2O Present in the Feed The additional methane experiments were conducted using gas mixture: NO = 0.04 vol.%, CH4 = 0.016 vol.%, O2 = 0.012 vol.% and Ar as a balance, SV = 50,000 h -1 and HR = 8 �C min-1 (200 ? 650 �C). Concentrations were monitored by a gas FTIR. 2.4.2 Oxygen Storage Capacities The amount of oxygen that can be stored in a TWC can be calculated by the amount of produced CO2. Two different methods have been applied in order to investigate OSC of our samples: CO pulse and CO–O2 cycle measurements were carried out to examine the oxygen storage complete capacity, dynamic oxygen storage capacity and labile oxygen storage capacity (LOSC). Experiments were con- ducted at 500 �C. Three oxygen storage capacities were measured: OSC—amount of CO uptake upon only one CO pulse after the samples were fully oxidized with O2, oxy- gen storage capacity complete (OSCC)—the total amount of CO uptake on a series of CO pulses until CO was no longer consumed and LOSC—amount of the most reactive O2 in the catalyst, consumed in the first CO pulse. 3 Results and Discussion 3.1 SEM-EDX The SEM-EDX mapping results are shown in Fig. 1. For the Pd/Rh catalyst washcoat was divided in two separate layers: Pd supported on Ba-stabilized c-Al2O3 (inner layer) and Rh supported on Ce0.75Zr0.25O (outer layer). Rh was supported on the outer layer in order to avoid Pd/Rh alloying during aging or formation of Rh aluminates [10]. 3.2 BET and Pore Size Distribution Decreased surface area of the aged samples was observed (Fig. 2). The BET area of aged samples was found to be 20 % decreased and average pore diameter increased 50 and 100 % for Pd/Rh and Pd only respectively. This is due to the collapsing of small pores. Table 1 Gas composition under different activity measurements CO2 CO H2 C3H6 NO CH4 O2, rich O2, stoichio O2, lean O2, SAI N2/Ar Vol.% 1 0.7 0.2 0.05 0.05 0.02 0.669 0.69 0.71 2.776 Balance 324 Top Catal (2013) 56:323–328 123 3.3 Dispersion Measurements by the Selective Chemisorptions Method Results of pulse flow CO chemisorptions measurements are presented in Table 2. Significantly decreased values of metals dispersion after thermal aging have been observed. It has been found by Rohe´ and Pitchon [11] that aging in wet oxidative atmosphere led to rhodium migration into the Al2O3 support. Noble metal particle growth has been reported for catalyst containing ceria during the aging if water vapor is present [12]. The reported sintering temperature for Pd and Rh are 1,100 and 600 �C and respectively [13, 14]. 3.4 Catalytic Activity 3.4.1 Light-off Measurements Activity was investigated by light off measurements. The effect of aging on the light-off temperature (T50) (the temperature needed for 50 % conversion of NO, CO and C3H6) was investigated for the Pd/Rh and Pd catalysts and the results are shown in Fig. 3. Further, the effect of varying the gas phase condition was investigated and the results from rich, stoichiometric, lean and SAI are also shown. As shown in Fig. 3. Light-off temperature for aged samples rises signifi- cantly compared with fresh samples. With respect to CO and HC oxidation and NOx reduc- tion, the results indicate that the deactivation of all func- tions was similar for Pd only monoliths in opposite to the bimetallic catalyst where the pronounced deactivation for NOx reduction could be seen. Based on XPS results, shift in energy band of Rh was observed (peak shift from 307.4 eV to around 310 eV) and has been assigned to RhO2 due to strong interaction between Rh and y-Al2O3 due to Rh dif- fusion from Ce to the Al2O3 support [15]. The aged sample, containing Pd only, indicate high selectivity for oxidation of CO and HC by oxygen and the Pd/Rh catalyst showed high selectivity for reduction of NO [16]. The deactivation of NOx reduction is probably due to deactivation of Rh and that is more pronounced in lean atmosphere. Fig. 1 SEM-EDX mapping results for Pd/Rh and Pd only catalysts 101 102 103 -5 0 5 10 15 20 25 30 35 Pore Diameter, Angstroms D iff er e n tia l S ur fa ce A re a Fresh Three Way Catalyst Aged Three Way Catalyst Fig. 2 Pore size distribution Pd/Rh from BET measurement Table 2 Dispersion of examined catalysts Pd/Rh fresh Pd/Rh aged Pd fresh Pd aged Dispersion % 31.2 5.36 22.9 6.42 Top Catal (2013) 56:323–328 325 123 3.4.2 NOx Reduction by Methane and Propene Conversion curves of NOx, during heating and cooling are presented in Fig. 4. NOx reacts below 250 �C with C3H6 both on Pd/Rh and Pd only catalysts. Above 300 �C decreases the selectivity for NOx reduction on Pd-only catalysts (both fresh and aged) while the high selectivity remains on Pd/Rh catalysts. Above 300 �C decreases the selectivity for NOx reduc- tion on Pd-only catalysts (both fresh and aged) while the high selectivity remains on Pd/Rh catalysts. Presence of Rh and Ce has positive influence on NOx conversion by pro- moting NOx reduction at temperatures below the temper- atures at which methane becomes active for NOx reduction (I, Fig. 4a, b). Methane is inert below 400 �C and becomes more active with increasing temperature. A clear decrease of Pd activity can be observed (II, Fig. 4a, b). For the fresh Pd-only catalyst, the reversible drop in Pd activity during extinction caused by PdO decomposition into Pd and O observed on 4a [17]. For the aged Pd-only catalyst, the irreversible decrease of Pd activity caused by aging, observed on 4b, can be explained by fact that Pd particles changed the oxidation state from ?2 to ?3 or ?4. XPS results presented changes in energy band of Pd from ?2 (336.5 and 341.5 eV) to 338 and 343 eV which could be attributed to Pd?3 and Pd?4. The high oxidation state of Pd could be explained by formation of Pd aluminates reported before by Hoost and Otto [18]. Since the flow is stoichi- ometric, highest conversion of NOx is reached at the highest examined temperature where also methane is reacting. The reaction mechanism of selective catalytic reduction of nitric oxide NO by methane in presence of water over Pd only catalyst has been investigated, which is shown in Fig. 5. With increasing temperature, the catalytic activity increased and 100 % conversion was reached above 600 �C, 80 % conversion was reached at temperature Fig. 3 Difference in light-off temperatures TLOaged–TLOfresh [C]: a Pd/Rh; b Pd Fig. 4 NOx reduction during cooling and heating; comparison between a fresh and b aged catalysts 326 Top Catal (2013) 56:323–328 123 around 450 �C without water in the feed, compared with 20 % when water was present. Negative influence of water for NOx reduction on the fresh catalyst was observed. Water inhibition increased with decreased reaction tem- perature or increased amount of water. No NOx conversion was observed when only methane was used as reducing agent under lean conditions (excess of O2). In the case of the fresh sample, the working temperature of the catalyst in the presence of water was shifted to temperature comparable to the working temperature of the aged catalyst (Fig. 5). Only small differences have been observed in the case of the aged catalyst; water seems to not influence the activity of the aged catalyst. 3.4.3 OSC, OSCC and LOSC Oxygen storage capacity (OSC and OSCC) decreased in both cases and was strongly dependent of surface area and cerium presence in the washcoat (Fig. 6.). High OSC and OSCC values can be explained by the activation of oxygen atoms on ceria that occurs above 250 �C. Pd/Rh catalyst has higher OSC capacity due to cerium-zirconium presence in the washcoat. Pd only catalyst can store oxygen but the amount of stored oxygen is much lower and storing of oxygen is only by oxygen’s chemisorption on the metal surface. We could observe that OSC and LOSC for the catalyst containing Pd only was the same. All stored oxygen was consumed in the first CO pulse due to lack of the oxygen storing components (CeZrO) in the washcoat. 4 Conclusions Hydrothermal aging caused mainly changes in the state of noble metals that are presented in the washcoat of the examined catalysts. It has been shown that the deactivation of Rh (sintering, possible formation of RhO2 and its interaction with Al2O3) affected the catalytic performance for NOx reduction. The reversible activity drop of Pd activity during cooling has been observed and explained by the decomposition of PdO into Pd and O2. The irreversible deactivation of Pd has been explained by the changes in the oxidation state of Pd, from ?2 to ?3 or ?4, as the effect of hydrothermal aging. References 1. Nagashima K, Nagata M, SAE Technical Paper Series 2007-01- 1134 2. Lampert JK, Kazi MS, Farrauto RJ (1997) Appl Catal B 14(3–4):211–223 Fig. 6 Experimental results: OSC, OSCC and LOSC Fig. 5 Water influence on NOx reduction by CH4 over the fresh a and aged b Pd only catalyst Top Catal (2013) 56:323–328 327 123 3. Gonzalez-Velasco JR et al (2003) Ind Eng Chem Res 42(2):311–317 4. Wu XD, Xu LH, Weng D (2004) Appl Surf Sci 221(1–4): 375–383 5. Baik JH et al (2007) Top Catal 42–43(1–4):337–340 6. Fernandes DM et al (2008) Catal Today 133:574–581 7. Da Costa P et al, SAE Technical Paper Series 2007-01-0039 8. Schmieg SJ, Belton DN (1995) Appl Catal B 6(2):127–144 9. Takeguchi T et al (2005) Appl Catal A 293:91–96 10. Unterhalt H et al (2001) Phys Status Solidi A 188(4):1495–1503 11. Rohe´ R, Pitchon V (2001) Top Catal 16–17(1):311–315 12. Barbier JJ, Duprez D (1994) Appl Catal B 4(2–3):105–140 13. Yao HC, Japar S, Shelef M (1977) J Catal 50(3):407–418 14. Yao HC, Stepien HK, Gandhi HS (1980) J Catal 61(2):547–550 15. Weng-Sieh Z, Gronsky R, Bell AT (1998) J Catal 174:22–33 16. Fathali A (2010) Deactivation of commercial three-way catalysts: effect of the thermal ageing. Licentiate thesis 17. Farrauto RJ, Hobson MC, Kennelly T, Waterman EM (1992) Appl Catal A 81:227–237 18. Hoost E, Otto K (1992) Appl. Catal. A 92:39 328 Top Catal (2013) 56:323–328 123 Hydrothermal Aging-Induced Changes in Washcoats of Commercial Three-Way Catalysts Abstract Introduction Experimental Materials Aging Procedure Characterizations Methods Dispersion Measurements by the Selective Chemisorptions Method Activity Test NOx Reduction by CH4 with and Without H2O Present in the Feed Oxygen Storage Capacities Results and Discussion SEM-EDX BET and Pore Size Distribution Dispersion Measurements by the Selective Chemisorptions Method Catalytic Activity Light-off Measurements NOx Reduction by Methane and Propene OSC, OSCC and LOSC Conclusions References