- Email: [email protected]
SiO2 catalysts during the selective hydrogenation of acetylene
Applied Catalysis A: General 360 (2009) 38–42
Contents lists available at ScienceDirect
Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata
Three-stage deactivation of Pd/SiO2 and Pd-Ag/SiO2 catalysts during the selective hydrogenation of acetylene In Young Ahn, Ji Hoon Lee, Seok Ki Kim, Sang Heup Moon * School of Chemical and Biological Engineering, Institute of Chemical Processes, Seoul National University, San 56-1, Shillim-dong, Kwanak-ku, Seoul 151-744, Republic of Korea
A R T I C L E I N F O
A B S T R A C T
Article history: Received 6 November 2008 Received in revised form 26 February 2009 Accepted 26 February 2009 Available online 11 March 2009
The deactivation of Pd/SiO2 and Ag-promoted Pd/SiO2 catalysts was monitored during their use in the selective hydrogenation of acetylene. Based on the analysis of green oil accumulated on the deactivated Pd/SiO2 catalysts, it was proposed that the catalyst deactivation proceeded in three stages. In the initial stage (Stage I), a large amount of relatively light green oil was deposited on, or in the vicinity of, the Pd surface, but the catalytic activity decreased only slightly. As the deactivation proceeded (Stage II), the 1,3-butadiene that had accumulated on the Pd surface was polymerized to relatively heavy green oil, a part of which moved from the Pd surface to the support. In the later stage of deactivation (Stage III), catalytic activity was drastically decreased because catalyst pores were blocked and hydrogen diffusion was limited in the thick film of the relatively heavy green oil. Although the initial activity of Ag-promoted Pd/SiO2 was slightly lower than that of the unpromoted one, the amount of green oil deposited on the former catalyst was much smaller than that deposited on the latter. Consequently, the final stage of deactivation (Stage III) was not observed with Pd-Ag/SiO2 during the reaction period of this study. Pd-Ag/SiO2 showed Stage II, characterized by the transfer of green oil from the Pd to the support, at a period earlier than in the case of Pd/SiO2, which additionally contributed to the slow deactivation of the former catalyst. The green oil that formed on the Pd-Ag/SiO2 was more volatile and mobile than that formed on Pd/SiO2, because the added Ag geometrically blocked multi-coordinated large ensembles of the Pd surface and also modified the Pd electronically such that the adsorption of 1,3-butadiene on the Pd became weaker than in the absence of promotion. ß 2009 Elsevier B.V. All rights reserved.
Keywords: Acetylene hydrogenation Pd catalyst Ag addition Deactivation Green oil Formation mechanism
1. Introduction Green oil, which is a major contributor to the deactivation of Pd catalysts in acetylene hydrogenation, has been studied extensively [1–7]. The carbonaceous deposits, which constitute green oil, modify catalytic performance by changing the catalyst surface properties: reducing the surface area [8], poisoning active sites [9], and limiting the diffusion of acetylene into the catalyst [10–12]. However, the above changes in the surface properties contribute to catalyst deactivation in different extents and in manners not necessarily proportional to the amounts of the deposits. For example, the initial activity of Pd was nearly preserved when the catalyst contained the deposits in amounts as large as 36 wt%, suggesting that the majority of active sites were still available for the reaction [7]. On the other hand, the activity was significantly lowered even by a small increase in the amounts of the deposits when the catalyst was in the later stage of deactivation [10–12].
Accordingly, the effect of the carbonaceous deposits on the activity should be investigated over the entire lifetime of the catalyst, i.e., from fresh catalyst to an almost dead one, such that their contribution to, and the underlying mechanism of, the deactivation is assessed according to the process period. In the present study, we compared the deactivation behaviors of Pd/SiO2 and Ag-promoted Pd/SiO2 by simultaneously monitoring changes in catalyst activity and in the amounts of 1,3butadiene, a known precursor of green oil [13], produced through the last stage of catalyst deactivation. The observed behaviors correlated with the properties of green-oil species that were deposited on the catalysts at different stages of deactivation. Finally, a mechanism that explains the role of Ag as a promoter for extending the lifetime of Pd/SiO2 has been proposed based on the above results. 2. Experimental 2.1. Catalyst preparation
* Corresponding author. Tel.: +82 2 880 7409; fax: +82 2 875 6697. E-mail address: [email protected] (S.H. Moon). 0926-860X/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2009.02.044
Pd/SiO2 catalyst containing 1 wt% of Pd was prepared by an ionexchange method using silica (surface area = 145 m2/g) as a
I.Y. Ahn et al. / Applied Catalysis A: General 360 (2009) 38–42
support and Pd(NH3)4(OH)2 as a Pd precursor [14]. Ag-promoted Pd catalyst was prepared by impregnating Pd/SiO2 with an aqueous solution of AgNO3. The atomic ratio of Ag/Pd was adjusted to 0.5. The catalysts were dried at 110 8C overnight and calcined in air at 300 8C for 2 h. All catalysts were reduced in H2 at 300 8C for 1 h before being used for acetylene hydrogenation. 2.2. Reaction tests A 0.06 g sample of the catalyst, contained in a quartz reactor, was used for acetylene hydrogenation at atmospheric pressure. The reactant stream contained 4.1% acetylene in ethylene flowing at 30 ml/min (space time: 4.33 10 3 min). The H2/acetylene ratio was 1 and the reaction temperature was 90 8C. The products were analyzed using an on-line GC (HP model 6890 series with FID) equipped with a capillary column (HP-AL/S). 2.3. Green-oil analysis An elemental analyzer (EA1110, CE Instrument) was used to measure the amounts of carbon and hydrogen contained in green oil deposited on the catalyst after use in the reaction test. Total amounts and types of green oil were also measured thermogravimetrically (TGA) with a Magnetic Suspension Balance (Rubotherm) in a 10% O2/N2 stream flowing at 40 ml/min. The temperature was raised from 25 8C to 700 8C at a rate of 5 8C/min. Differential thermogravimetric analysis (DTGA) results were obtained from the TGA curves by differentiating the latter with respect to temperature. 2.4. CO-IR A 0.07 g sample of catalyst was pressed into a self-supporting disc 17 mm in diameter, placed in the IR cell with CaF2 windows [15], reduced at 300 8C, and then exposed to 14.6 Torr of CO at room temperature for 5 min. The IR spectra of adsorbed CO were recorded (Midac 2100) after the removal of gaseous CO from the cell by evacuation.
39
represented by the accumulated amounts of converted acetylene such that the deactivation characteristics could be compared based on the same load of the reaction [13,14]. Data obtained at the same time-on-stream for two sample catalysts have been connected by subsidiary lines. The activity of Pd/SiO2 changed characteristically during the three reaction periods. That is, the activity remained nearly constant when the accumulated amounts of converted acetylene were smaller than 0.014 mol (designated as Period I). On the other hand, the activity drastically decreased when the accumulated amounts of converted acetylene were larger than 0.030 mol (designated as Period III). During the intermediate period (Period II), the activity decreased slightly. The activity of Pd0.5Ag/SiO2 decreased at an almost constant rate during the entire period of this study. As a result, a drastic decrease in the activity was not observed during Period III, which agreed with the wellknown information that the lifetime of the Pd catalyst was extended by Ag promotion [16]. The amounts of 1,3-butadiene, a proposed precursor of green oil [13], were plotted versus the reaction period, as shown in Fig. 2. In the case of Pd/SiO2, the amounts of 1,3-butadiene increased at relatively low rates during Period I, which was followed by an increase in the amounts that eventually showed a maximum during the next period of the deactivation process (Period II), and then decreased significantly during Period III. To the contrary, Pd0.5Ag/SiO2 produced smaller amounts of 1,3-butadiene that increased only slightly during Period II. The period required for producing the maximum concentration of 1,3-butadiene, marked in the figure, was shorter for Pd-0.5Ag/SiO2, i.e., after the conversion of acetylene in the accumulated amounts of about 0.022 mol, than for Pd/SiO2, which showed the maximum at 0.030 mol of converted acetylene. 3.2. Rates of green-oil formation
Fig. 1 shows changes in the activities of Pd/SiO2 and Agpromoted Pd/SiO2 as a function of the reaction period, which are
Fig. 3 shows the amounts of green oil deposited on the catalysts, obtained by EA and TGA, as a function of catalytic activity, which was represented by acetylene conversion in the deactivation process. The C and H contents of green oil, recovered from the catalysts after use for different periods, are also summarized in Table 1. In Fig. 3, solid and broken lines that connect TGA and EA data, respectively, show nearly the same correlations between the amounts of green oil and catalytic activity. During the initial period of the reaction (Period I), large amounts of green oil, about 34 wt%, were deposited on Pd/SiO2, even though the conversion decreased slightly by 2%. These results indicate that the reaction successfully
Fig. 1. Changes in the catalytic activity as a function of the accumulated amounts of converted acetylene on Pd/SiO2 (&) and Pd-0.5Ag/SiO2 (*).
Fig. 2. Changes in the amounts of produced 1,3-butadiene as a function of the accumulated amounts of converted acetylene on Pd/SiO2 (&) and Pd-0.5Ag/SiO2 (*).
3. Results 3.1. Deactivation behaviors
I.Y. Ahn et al. / Applied Catalysis A: General 360 (2009) 38–42
40
Fig. 3. The amounts of green oil deposited on Pd/SiO2 and Pd-0.5Ag/SiO2, analyzed by TGA(&, *) and EA (&, *).
Table 1 EA results obtained using Pd/SiO2 and Pd-0.5Ag/SiO2 after use for different periods. Catalyst
C (wt%)
H (wt%)
Green oil (wt%)
H/C atomic ratio
Pd/Period I Pd/Period II Pd/Period III Pd-0.5Ag/Period I Pd-0.5Ag/Period II Pd-0.5Ag/Period III
30.2 43.9 51.0 16.0 26.2 30.6
3.8 5.6 6.6 2.3 3.6 4.1
34.0 49.5 57.6 18.3 29.8 34.7
1.5 1.5 1.5 1.7 1.6 1.6
proceeded even in the presence of large amounts of green oil, which is in accordance with the results of previous studies [7,17– 19]. A small decrease in the conversion during the initial period suggests either that green oil, initially produced in large amounts, covered only a small fraction of the Pd active sites or that acetylene hydrogenation could proceed even when a green-oil layer was formed on the catalyst. When the catalyst was gradually deactivated during the middle period of the deactivation process (Period II), the rates of green-oil deposition became slightly lower than those during the initial period. During the later period of the reaction (Period III), the conversion decreased drastically, i.e., from 88% to 46% for Pd/SiO2, although the increase in green-oil amounts was only 5–10 wt%. In the case of Pd-0.5Ag/SiO2, amounts of about 20 wt% of green oil were deposited while the conversions decreased only slightly, by 3%, during the initial period (Period I). When the amounts of converted acetylene were larger than 0.014 mol (beyond Period I), the rates of green-oil deposition became lower than in the case of Period I and remained nearly constant afterwards. Because the activity of Pd-0.5Ag/SiO2 was higher than that of Pd/SiO2 during the later period (Period III in Fig. 1), which originated from slower deactivation of the former catalyst, the rates of green-oil deposition on Pd-0.5Ag/SiO2 remained nearly the same as in Period II (Fig. 3). The above result differs from that observed for Pd/ SiO2, which became very low in activity at the end of Period III. 3.3. DTGA of green oil To further investigate the properties of green oil deposited on the catalyst, DTGA was made by differentiating the TGA curves. Fig. 4 shows the DTGA results of deactivated catalysts (Pd/Period I denotes Pd/SiO2 used for converting 0.014 mol of acetylene). According to Larsson et al. [20], peaks observed below 300 8C (Peak I), between 300 8C and 500 8C (Peak II), and above 500 8C (Peak III)
Fig. 4. DTGA of (A) Pd/SiO2 and (B) Pd-0.5Ag/SiO2 after use for different periods.
were due to heavy hydrocarbons adsorbed onto the catalyst surface and absorbed into the catalyst pores, to green oil on or in the vicinity of Pd, and to green oil deposited on the support, respectively. In Fig. 4(A), Peaks obtained from Pd/Period I appear at temperatures below 250 8C (Peak I), between 250 8C and 400 8C (Peak II), and above 450 8C (Peak III), respectively. Peak II shows two maxima at 280 8C and 330 8C. On the other hand, Peak I shows very low intensity and Peak III is almost negligible. The peaks obtained from Pd/SiO2 changed characteristically, in their position and intensity, as the catalyst was deactivated. Peaks I and III increased in intensity, which obviously was due to the increased amounts of hydrocarbon and green oil species deposited on the catalysts. Positions of all three peaks were shifted to high temperatures, indicating that the deposited surface species gradually became heavier as the catalysts were deactivated to greater extents [21,22]. The peak shift was particularly distinct with Peak II, which initially showed two distinct peaks, in the case of Pd/SiO2, but eventually became broad and appeared at higher temperatures during the later period of deactivation. This result suggests that green oil initially deposited on or in the vicinity of the Pd surface was transformed to heavy molecules covering a large portion of the Pd surface. The polymerization of relatively light green oil formed on the Pd surface was accelerated because the
I.Y. Ahn et al. / Applied Catalysis A: General 360 (2009) 38–42
hydrogenation activity of the Pd catalyst was drastically decreased (Fig. 1). Previous studies reported that green oil spilled over from the Pd to the support as the reaction proceeded [23–25]. In the present study, green oil was present only in small amounts on the support during the initial period of deactivation, but in significant amounts during the later periods, as indicated by changes in the intensity of Peak III (Fig. 4(A)). Because green oil cannot form on the support itself, i.e., without the assistance of the Pd surface [7,9,26], the above result suggests that green oil, which was initially deposited in large amounts on or in the vicinity of Pd, migrated to the support during the later periods of deactivation. Compared with Pd/SiO2, Pd-0.5Ag/SiO2 shows peaks of much lower intensity in Fig. 4(B). However, a trend in the characteristic change of the peaks with catalyst deactivation is similar to one observed with Pd/SiO2. That is, green oil, most of which was initially deposited on or in the vicinity of the Pd surface (represented by Peak II), was polymerized to a heavier species and spilled over to the support after the middle period of the reaction (Period II). However, the amounts of green oil deposited on the Pd were very small and the polymerization of green oil was retarded, as indicated by the slight shift of Peak II with the same low intensity. These results indicate that Ag addition suppressed the polymerization of green oil as well as its formation. In the case of Pd-0.5Ag/Period III, the amounts of green oil adsorbed on the catalyst surface (Peak I) and deposited on the support (Peak III) were relatively large whereas those on the Pd surface (Peak II) were small. Changes in the properties of deposited green oil by Ag addition will be discussed in more detail in Section 4.2.
41
tion of 1,3-butadiene on the catalyst surface [13]. Large amounts of 1,3-butadiene should accelerate the formation of green oil, which will eventually decrease catalytic activity. Therefore, the activity of Pd/SiO2 decreased slightly in Stage II due to the accumulation of 1,3-butadiene (Fig. 2), whereas it remained nearly constant in Stage I. In the later stage of the process (Stage III), the catalytic activity decreased drastically (Fig. 1) even though green oil was deposited only in small amounts (Fig. 3). This is because green-oil spillover from the Pd to the support was retarded and catalyst pores were gradually blocked by the polymerized species of green oil [9,12]. A significant increase in the molecular weights of green oil in Stage III was indicated by a large shift of Peak II in Fig. 4(A). Green oil forms a thick film on the Pd surface due to the suppressed green-oil spillover and, accordingly, the rate of acetylene hydrogenation is limited by the transport of hydrogen in the surface film [9,11]. Liu et al. also reported the role of green oil as a diffusion barrier [32]. They observed that small acetylene and hydrogen molecules can diffuse through the thin green-oil layer to reach the Pd surface so the catalyst remains active, which corresponds to Stage I in this study. As the green oil builds up to form a thick layer, this growing diffusion barrier significantly reduces the activity of the catalyst, which is in accordance with the drastic activity decrease in Stage III (Fig. 1). Consequently, it can be concluded that various factors, including suppressed green-oil spillover, fast chain growth of green oil on Pd, enhanced diffusion limitation, and pore blocking, were responsible for the drastic decrease in catalytic activity in Stage III even when the amounts of green oil were only slightly increased.
4. Discussion 4.2. Effect of Ag addition 4.1. Three stages of deactivation for Pd/SiO2 Based on the above observations, it can be concluded that the deactivation of Pd/SiO2 by green-oil deposition proceeds over three stages. Periods I, II and III denoted for Pd/SiO2 can be designated as Stage I, II and III of the deactivation process, respectively. DTGA of Pd/Periods I (Fig. 4(A)) indicated that initially green oil was deposited largely on, or in the vicinity of, the Pd instead of the support. However, the green oil formed during the initial reaction period (Stage I), in amounts as large as 34 wt%, was not detrimental to the activity (Fig. 1). The above results can be explained by the hydrogen transfer mechanism, which was initially proposed by Thomson and Webb [27] and referred to by Borodzinski et al. [6,28–30]. The acetylenic species actually participated in the catalytic reaction, but was not necessarily adsorbed onto an extended Pd surface and could either be absorbed into or adsorbed onto a carbonaceous overlayer that covered the Pd [17–19,27,31]. An initial layer was formed on the Pd surface from acetylene that was dissociatively adsorbed onto the Pd. A second layer was subsequently formed on the first layer by the associative adsorption of acetylene, which was eventually converted to ethylene [17–19]. It was proposed that the carbonaceous overlayer worked as a medium to allow the transfer of hydrogen from the Pd surface to the acetylene that was associated with the overlayer [6,27–31]. Accordingly, the rates of acetylene hydrogenation was unaffected by the large amounts of green oil deposited in the initial stage (Stage I), because hydrogen adsorbed onto the Pd surface was successfully transferred to the acetylenic species that were either absorbed into or adsorbed onto the green-oil layer. It was previously reported that green-oil formation suppressed the rates of 1,3-butadiene conversion to butenes or high molecular-weight species to greater extents than those of 1,3butadiene formation from acetylene, which led to the accumula-
The above deactivation behavior changed when Pd/SiO2 was promoted with Ag to extend the catalyst lifetime. During Periods I and II, Pd-0.5Ag/SiO2 produced smaller amounts of green oil than in the case of Pd/SiO2 while the decrease in the activity was similar for both catalysts (Fig. 3). The formation of green oil was retarded because 1,3-butadiene was produced in smaller amounts during Period II on Pd-0.5Ag/SiO2 (Fig. 2). During the later period of the reaction (Period III), the decrease in the activity of Pd-0.5Ag/SiO2 was 20 times smaller than that of Pd/SiO2, but similar amounts of green oil, less than 10 wt%, were deposited on both catalysts. This result indicates that Stage III, which was characterized by a significant decease in activity due to the accumulation of heavy green oil, was not observed in the case of Pd-0.5Ag/SiO2. The added Ag species not only suppressed the formation and polymerization of green oil but also affected the properties of green oil produced. The H/C atomic ratio of green oil, listed in Table 1, was slightly higher for Pd-0.5Ag/SiO2 than it was for Pd/SiO2, indicating that green oil formed on the former catalyst was either lighter or more volatile with a short chain length. Contrary to the case of the Pd/SiO2 catalyst, all DTGA peaks obtained from Pd0.5Ag/SiO2 were observed at lower temperatures, and the shift of Peak II, representing green oil deposited on the Pd surface, was not significant. Furthermore, the amounts of green oil adsorbed onto the catalyst surface (Peak I) and deposited on the support (Peak III) were relatively large whereas those on the Pd surface (Peak II) were small. These results indicate that green oil, primarily formed on the Pd surface in small amounts, easily moved to the support in the case of Pd-0.5Ag/SiO2. The period required for producing the maximum concentration of 1,3-butadiene was also shorter for Pd0.5Ag/SiO2 than for Pd/SiO2 (Fig. 2). Accordingly, Stage II, which was responsible for the transfer of green oil from the Pd to the support, was observed at an earlier period when the catalyst was promoted with Ag.
42
I.Y. Ahn et al. / Applied Catalysis A: General 360 (2009) 38–42
The deactivation was significantly retarded when Ag was added to Pd/SiO2. Although the initial activity of Pd-0.5Ag/SiO2 was slightly lower than that of Pd/SiO2, the amount of green oil deposited on the former catalyst was much smaller than that deposited on the latter. Consequently, the final stage of deactivation (Stage III), due to diffusion limitation in catalyst pores, was not observed with Pd0.5Ag/SiO2 during the reaction period of this study. Pd-0.5Ag/SiO2 showed Stage II, characterized by the transfer of green oil from the Pd to the support, at an earlier period than in the case of Pd/SiO2, which resulted in the suppressed polymerization of green oil on the Pd. Added Ag geometrically blocked multi-coordinated large ensembles of the Pd surface and also modified Pd electronically to facilitate the desorption of 1,3-butadiene from the Pd surface. Consequently, the green oil that formed on the Pd-0.5Ag/SiO2 became more volatile and mobile than that formed on Pd/SiO2. Acknowledgements Fig. 5. Infrared spectra of CO adsorbed on Pd/SiO2 (—) and Pd-0.5Ag/SiO2 (- - - -).
To further investigate the surface properties of Pd-0.5Ag/SiO2, IR spectra of CO adsorbed onto the catalyst were obtained and compared with those observed for Pd/SiO2 (Fig. 5). When Ag was added to the Pd catalyst, the intensity of peaks observed at wavenumbers below 2000 cm 1, representing the multi-coordinated CO species, was lowered. The area ratio of linear-bound to multi-bound bands, Al/Am, increased from 0.38 to 0.68, indicating that deposited Ag effectively blocked the sites consisting of two or three adjacent Pd atoms [22,33]. Because 1,3-butadiene was produced by the dimerization of acetylenic species that were adsorbed as neighbors on the Pd surface, the blockage of adjacent Pd atoms by added Ag on the catalyst surface should suppress the formation of green oil (Fig. 3) and, consequently, the rates of catalyst deactivation (Fig. 1). It was reported that added Ag also modified the electronic properties of Pd [16,34,35], as indicated by an increase in the Pd dband electron density [16]. In previous studies, electronic modification of Pd by an added promoter reduced the adsorption strength of ethylene on the surface, thus facilitating the desorption of ethylene from the surface [13,36,37]. As butadiene was reported to follow the same trend as ethylene in the mode and energy of adsorption on Pd [38], adsorption of 1,3-butadiene would be weaker on Pd-0.5Ag/SiO2 than in the case of Pd/SiO2. Consequently, both geometric and electronic factors are responsible for the improved deactivation behavior of the Pd-0.5Ag/SiO2 catalyst. 5. Conclusions The deactivation of Pd/SiO2 in acetylene hydrogenation proceeded in three stages. In Stage I, a large amount of relatively light green oil was deposited onto the catalyst, but the activity was only slightly decreased because acetylene, which was associatively adsorbed onto the green oil, was readily hydrogenated via a hydrogen transfer mechanism. As the deactivation proceeded (Stage II), 1,3-butadiene accumulated on the Pd surface and polymerized to relatively heavy green oil, a part of which moved from the Pd surface to the support. In the later stage of the process (Stage III), the activity was drastically decreased even though the amount of green oil was only slightly increased. This was because catalyst pores were blocked by the accumulated green oil and hydrogen diffusion was limited in the thick film of the relatively heavy green oil.
This study was supported by BASF Aktiengesellschaft, Germany, by the Brain Korea 21 project, and by the Korea Science and Engineering Foundation (KOSEF) through the National Research Laboratory program of Ministry of Science and Technology (No. M10318000095-04J0000-02810), Korea. References [1] M.L. Derrien, Stud. Surf. Sci. Catal. 27 (1986) 613–666. [2] J.T. Wehrli, D.J. Thomas, M.S. Wainwright, D.L. Trimm, N.W. Cant, Catalyst Deactivation, Studies in Surface Science and Catalysis, Elsevier Science BV, 1991,, p. 203. [3] P. Albers, K. Seibold, G. Prescher, H. Muller, Appl. Catal. A 176 (1999) 135–146. [4] J.P. Boitiaux, J. Cosyns, M. Derrien, G. Leger, Hydrocarbon Process. 64 (1985) 51– 59. [5] D.K. Lee, Korean J. Chem. Eng. 7 (1990) 233–235. [6] A. Borodzinski, G.C. Bond, Catal. Rev. 48 (2006) 91–144. [7] A. Sarkany, L. Guczi, A.H. Weiss, Appl. Catal. 10 (1984) 369–388. [8] G.C. Battiston, L. Dalloro, G.R. Tauszik, Appl. Catal. 2 (1982) 1–17. [9] A. Sarkany, A.H. Weiss, T. Szilagyi, P. Sandor, L. Guczi, Appl. Catal. 12 (1984) 373– 379. [10] M.R. Stammbach, D.J. Thomas, D.L. Trimm, M.S. Wainwright, Appl. Catal. 58 (1990) 209–217. [11] S. Asplund, J. Catal. 158 (1996) 267–278. [12] J. Wood, L.F. Gladden, Appl. Catal. A 249 (2003) 241–253. [13] I.Y. Ahn, J.H. Lee, S.S. Kum, S.H. Moon, Catal. Today 123 (2007) 151–157. [14] I.Y. Ahn, W.J. Kim, S.H. Moon, Appl. Catal. A 308 (2006) 75–81. [15] S.H. Moon, H. Windawi, J.R. Katzer, Ind. Eng. Chem. Fundam. 20 (1981) 396–399. [16] D.C. Huang, K.H. Chang, W.F. Pong, P.K. Tseng, K.J. Hung, W.F. Huang, Catal. Lett. 53 (1998) 155–159. [17] A.S. Al-Ammar, G. Webb, J. Chem. Soc. Faraday Trans. I 74 (1978) 195–205. [18] A.S. Al-Ammar, G. Webb, J. Chem. Soc. Faraday Trans. I 74 (1978) 657–664. [19] A.S. Al-Ammar, G. Webb, J. Chem. Soc. Faraday Trans. I 75 (1979) 1900–1911. [20] M. Larsson, J. Jansson, S. Asplund, J. Catal. 178 (1998) 49–57. [21] P. Albers, J. Pietsch, S.F. Parker, J. Mol. Catal. A 173 (2001) 275–286. [22] W.J. Kim, E.W. Shin, J.H. Kang, S.H. Moon, Appl. Catal. A 251 (2003) 305–313. [23] G.A. Somorjai, S.M. Davis, Chem. Tech. (August 1983) 502–511. [24] A.H. Weiss, S. Leviness, V. Nair, L. Guczi, A. Sarkany, Z. Schay, Int. Congr. Catal., 8th Proceeding V5 (1984) 591–599. [25] S. Leviness, V. Nair, A.H. Weiss, Z. Schay, L. Guczi, J. Mol. Catal. 25 (1984) 131–140. [26] J. Houzvicka, R. Pestman, V. Ponec, Catal. Lett. 30 (1995) 289–296. [27] S.J. Thomson, G. Webb, J. Chem. Soc. Chem. Commun. 13 (1976) 526–527. [28] A. Borodzinski, A. Golebiowski, Langmuir 13 (1997) 883–887. [29] A. Borodzinski, Catal. Lett. 63 (1999) 35–42. [30] A. Borodzinski, Catal. Lett. 71 (2001) 169–176. [31] G. Webb, Catal. Today 7 (1990) 139–155. [32] R.-J. Liu, P.A. Crozier, C.M. Smith, D.A. Hucul, J. Blackson, G. Salaita, Appl. Catal. A 282 (2005) 111–121. [33] H. Zea, K. Lester, A.K. Daty, W. Rightor, R. Gulotty, W. Waterman, M. Smith, Appl. Catal. A 282 (2005) 237–245. [34] Q. Zhang, J. Li, X. Liu, Q. Zhu, Appl. Catal. A 197 (2000) 221–228. [35] P.A. Sheth, M. Neurock, C.M. Smith, J. Phys. Chem. B 109 (2005) 12449–12466. [36] J.H. Kang, E.W. Shin, W.J. Kim, J.D. Park, S.H. Moon, Catal. Today 63 (2000) 183– 188. [37] J.H. Kang, E.W. Shin, W.J. Kim, J.D. Park, S.H. Moon, J. Catal. 208 (2002) 310–320. [38] C. Breinlich, J. Haubrich, C. Becker, A. Valcarcel, F. Delbecq, K. Wandelt, J. Catal. 251 (2007) 123–130.
Contents lists available at ScienceDirect
Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata
Three-stage deactivation of Pd/SiO2 and Pd-Ag/SiO2 catalysts during the selective hydrogenation of acetylene In Young Ahn, Ji Hoon Lee, Seok Ki Kim, Sang Heup Moon * School of Chemical and Biological Engineering, Institute of Chemical Processes, Seoul National University, San 56-1, Shillim-dong, Kwanak-ku, Seoul 151-744, Republic of Korea
A R T I C L E I N F O
A B S T R A C T
Article history: Received 6 November 2008 Received in revised form 26 February 2009 Accepted 26 February 2009 Available online 11 March 2009
The deactivation of Pd/SiO2 and Ag-promoted Pd/SiO2 catalysts was monitored during their use in the selective hydrogenation of acetylene. Based on the analysis of green oil accumulated on the deactivated Pd/SiO2 catalysts, it was proposed that the catalyst deactivation proceeded in three stages. In the initial stage (Stage I), a large amount of relatively light green oil was deposited on, or in the vicinity of, the Pd surface, but the catalytic activity decreased only slightly. As the deactivation proceeded (Stage II), the 1,3-butadiene that had accumulated on the Pd surface was polymerized to relatively heavy green oil, a part of which moved from the Pd surface to the support. In the later stage of deactivation (Stage III), catalytic activity was drastically decreased because catalyst pores were blocked and hydrogen diffusion was limited in the thick film of the relatively heavy green oil. Although the initial activity of Ag-promoted Pd/SiO2 was slightly lower than that of the unpromoted one, the amount of green oil deposited on the former catalyst was much smaller than that deposited on the latter. Consequently, the final stage of deactivation (Stage III) was not observed with Pd-Ag/SiO2 during the reaction period of this study. Pd-Ag/SiO2 showed Stage II, characterized by the transfer of green oil from the Pd to the support, at a period earlier than in the case of Pd/SiO2, which additionally contributed to the slow deactivation of the former catalyst. The green oil that formed on the Pd-Ag/SiO2 was more volatile and mobile than that formed on Pd/SiO2, because the added Ag geometrically blocked multi-coordinated large ensembles of the Pd surface and also modified the Pd electronically such that the adsorption of 1,3-butadiene on the Pd became weaker than in the absence of promotion. ß 2009 Elsevier B.V. All rights reserved.
Keywords: Acetylene hydrogenation Pd catalyst Ag addition Deactivation Green oil Formation mechanism
1. Introduction Green oil, which is a major contributor to the deactivation of Pd catalysts in acetylene hydrogenation, has been studied extensively [1–7]. The carbonaceous deposits, which constitute green oil, modify catalytic performance by changing the catalyst surface properties: reducing the surface area [8], poisoning active sites [9], and limiting the diffusion of acetylene into the catalyst [10–12]. However, the above changes in the surface properties contribute to catalyst deactivation in different extents and in manners not necessarily proportional to the amounts of the deposits. For example, the initial activity of Pd was nearly preserved when the catalyst contained the deposits in amounts as large as 36 wt%, suggesting that the majority of active sites were still available for the reaction [7]. On the other hand, the activity was significantly lowered even by a small increase in the amounts of the deposits when the catalyst was in the later stage of deactivation [10–12].
Accordingly, the effect of the carbonaceous deposits on the activity should be investigated over the entire lifetime of the catalyst, i.e., from fresh catalyst to an almost dead one, such that their contribution to, and the underlying mechanism of, the deactivation is assessed according to the process period. In the present study, we compared the deactivation behaviors of Pd/SiO2 and Ag-promoted Pd/SiO2 by simultaneously monitoring changes in catalyst activity and in the amounts of 1,3butadiene, a known precursor of green oil [13], produced through the last stage of catalyst deactivation. The observed behaviors correlated with the properties of green-oil species that were deposited on the catalysts at different stages of deactivation. Finally, a mechanism that explains the role of Ag as a promoter for extending the lifetime of Pd/SiO2 has been proposed based on the above results. 2. Experimental 2.1. Catalyst preparation
* Corresponding author. Tel.: +82 2 880 7409; fax: +82 2 875 6697. E-mail address: [email protected] (S.H. Moon). 0926-860X/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2009.02.044
Pd/SiO2 catalyst containing 1 wt% of Pd was prepared by an ionexchange method using silica (surface area = 145 m2/g) as a
I.Y. Ahn et al. / Applied Catalysis A: General 360 (2009) 38–42
support and Pd(NH3)4(OH)2 as a Pd precursor [14]. Ag-promoted Pd catalyst was prepared by impregnating Pd/SiO2 with an aqueous solution of AgNO3. The atomic ratio of Ag/Pd was adjusted to 0.5. The catalysts were dried at 110 8C overnight and calcined in air at 300 8C for 2 h. All catalysts were reduced in H2 at 300 8C for 1 h before being used for acetylene hydrogenation. 2.2. Reaction tests A 0.06 g sample of the catalyst, contained in a quartz reactor, was used for acetylene hydrogenation at atmospheric pressure. The reactant stream contained 4.1% acetylene in ethylene flowing at 30 ml/min (space time: 4.33 10 3 min). The H2/acetylene ratio was 1 and the reaction temperature was 90 8C. The products were analyzed using an on-line GC (HP model 6890 series with FID) equipped with a capillary column (HP-AL/S). 2.3. Green-oil analysis An elemental analyzer (EA1110, CE Instrument) was used to measure the amounts of carbon and hydrogen contained in green oil deposited on the catalyst after use in the reaction test. Total amounts and types of green oil were also measured thermogravimetrically (TGA) with a Magnetic Suspension Balance (Rubotherm) in a 10% O2/N2 stream flowing at 40 ml/min. The temperature was raised from 25 8C to 700 8C at a rate of 5 8C/min. Differential thermogravimetric analysis (DTGA) results were obtained from the TGA curves by differentiating the latter with respect to temperature. 2.4. CO-IR A 0.07 g sample of catalyst was pressed into a self-supporting disc 17 mm in diameter, placed in the IR cell with CaF2 windows [15], reduced at 300 8C, and then exposed to 14.6 Torr of CO at room temperature for 5 min. The IR spectra of adsorbed CO were recorded (Midac 2100) after the removal of gaseous CO from the cell by evacuation.
39
represented by the accumulated amounts of converted acetylene such that the deactivation characteristics could be compared based on the same load of the reaction [13,14]. Data obtained at the same time-on-stream for two sample catalysts have been connected by subsidiary lines. The activity of Pd/SiO2 changed characteristically during the three reaction periods. That is, the activity remained nearly constant when the accumulated amounts of converted acetylene were smaller than 0.014 mol (designated as Period I). On the other hand, the activity drastically decreased when the accumulated amounts of converted acetylene were larger than 0.030 mol (designated as Period III). During the intermediate period (Period II), the activity decreased slightly. The activity of Pd0.5Ag/SiO2 decreased at an almost constant rate during the entire period of this study. As a result, a drastic decrease in the activity was not observed during Period III, which agreed with the wellknown information that the lifetime of the Pd catalyst was extended by Ag promotion [16]. The amounts of 1,3-butadiene, a proposed precursor of green oil [13], were plotted versus the reaction period, as shown in Fig. 2. In the case of Pd/SiO2, the amounts of 1,3-butadiene increased at relatively low rates during Period I, which was followed by an increase in the amounts that eventually showed a maximum during the next period of the deactivation process (Period II), and then decreased significantly during Period III. To the contrary, Pd0.5Ag/SiO2 produced smaller amounts of 1,3-butadiene that increased only slightly during Period II. The period required for producing the maximum concentration of 1,3-butadiene, marked in the figure, was shorter for Pd-0.5Ag/SiO2, i.e., after the conversion of acetylene in the accumulated amounts of about 0.022 mol, than for Pd/SiO2, which showed the maximum at 0.030 mol of converted acetylene. 3.2. Rates of green-oil formation
Fig. 1 shows changes in the activities of Pd/SiO2 and Agpromoted Pd/SiO2 as a function of the reaction period, which are
Fig. 3 shows the amounts of green oil deposited on the catalysts, obtained by EA and TGA, as a function of catalytic activity, which was represented by acetylene conversion in the deactivation process. The C and H contents of green oil, recovered from the catalysts after use for different periods, are also summarized in Table 1. In Fig. 3, solid and broken lines that connect TGA and EA data, respectively, show nearly the same correlations between the amounts of green oil and catalytic activity. During the initial period of the reaction (Period I), large amounts of green oil, about 34 wt%, were deposited on Pd/SiO2, even though the conversion decreased slightly by 2%. These results indicate that the reaction successfully
Fig. 1. Changes in the catalytic activity as a function of the accumulated amounts of converted acetylene on Pd/SiO2 (&) and Pd-0.5Ag/SiO2 (*).
Fig. 2. Changes in the amounts of produced 1,3-butadiene as a function of the accumulated amounts of converted acetylene on Pd/SiO2 (&) and Pd-0.5Ag/SiO2 (*).
3. Results 3.1. Deactivation behaviors
I.Y. Ahn et al. / Applied Catalysis A: General 360 (2009) 38–42
40
Fig. 3. The amounts of green oil deposited on Pd/SiO2 and Pd-0.5Ag/SiO2, analyzed by TGA(&, *) and EA (&, *).
Table 1 EA results obtained using Pd/SiO2 and Pd-0.5Ag/SiO2 after use for different periods. Catalyst
C (wt%)
H (wt%)
Green oil (wt%)
H/C atomic ratio
Pd/Period I Pd/Period II Pd/Period III Pd-0.5Ag/Period I Pd-0.5Ag/Period II Pd-0.5Ag/Period III
30.2 43.9 51.0 16.0 26.2 30.6
3.8 5.6 6.6 2.3 3.6 4.1
34.0 49.5 57.6 18.3 29.8 34.7
1.5 1.5 1.5 1.7 1.6 1.6
proceeded even in the presence of large amounts of green oil, which is in accordance with the results of previous studies [7,17– 19]. A small decrease in the conversion during the initial period suggests either that green oil, initially produced in large amounts, covered only a small fraction of the Pd active sites or that acetylene hydrogenation could proceed even when a green-oil layer was formed on the catalyst. When the catalyst was gradually deactivated during the middle period of the deactivation process (Period II), the rates of green-oil deposition became slightly lower than those during the initial period. During the later period of the reaction (Period III), the conversion decreased drastically, i.e., from 88% to 46% for Pd/SiO2, although the increase in green-oil amounts was only 5–10 wt%. In the case of Pd-0.5Ag/SiO2, amounts of about 20 wt% of green oil were deposited while the conversions decreased only slightly, by 3%, during the initial period (Period I). When the amounts of converted acetylene were larger than 0.014 mol (beyond Period I), the rates of green-oil deposition became lower than in the case of Period I and remained nearly constant afterwards. Because the activity of Pd-0.5Ag/SiO2 was higher than that of Pd/SiO2 during the later period (Period III in Fig. 1), which originated from slower deactivation of the former catalyst, the rates of green-oil deposition on Pd-0.5Ag/SiO2 remained nearly the same as in Period II (Fig. 3). The above result differs from that observed for Pd/ SiO2, which became very low in activity at the end of Period III. 3.3. DTGA of green oil To further investigate the properties of green oil deposited on the catalyst, DTGA was made by differentiating the TGA curves. Fig. 4 shows the DTGA results of deactivated catalysts (Pd/Period I denotes Pd/SiO2 used for converting 0.014 mol of acetylene). According to Larsson et al. [20], peaks observed below 300 8C (Peak I), between 300 8C and 500 8C (Peak II), and above 500 8C (Peak III)
Fig. 4. DTGA of (A) Pd/SiO2 and (B) Pd-0.5Ag/SiO2 after use for different periods.
were due to heavy hydrocarbons adsorbed onto the catalyst surface and absorbed into the catalyst pores, to green oil on or in the vicinity of Pd, and to green oil deposited on the support, respectively. In Fig. 4(A), Peaks obtained from Pd/Period I appear at temperatures below 250 8C (Peak I), between 250 8C and 400 8C (Peak II), and above 450 8C (Peak III), respectively. Peak II shows two maxima at 280 8C and 330 8C. On the other hand, Peak I shows very low intensity and Peak III is almost negligible. The peaks obtained from Pd/SiO2 changed characteristically, in their position and intensity, as the catalyst was deactivated. Peaks I and III increased in intensity, which obviously was due to the increased amounts of hydrocarbon and green oil species deposited on the catalysts. Positions of all three peaks were shifted to high temperatures, indicating that the deposited surface species gradually became heavier as the catalysts were deactivated to greater extents [21,22]. The peak shift was particularly distinct with Peak II, which initially showed two distinct peaks, in the case of Pd/SiO2, but eventually became broad and appeared at higher temperatures during the later period of deactivation. This result suggests that green oil initially deposited on or in the vicinity of the Pd surface was transformed to heavy molecules covering a large portion of the Pd surface. The polymerization of relatively light green oil formed on the Pd surface was accelerated because the
I.Y. Ahn et al. / Applied Catalysis A: General 360 (2009) 38–42
hydrogenation activity of the Pd catalyst was drastically decreased (Fig. 1). Previous studies reported that green oil spilled over from the Pd to the support as the reaction proceeded [23–25]. In the present study, green oil was present only in small amounts on the support during the initial period of deactivation, but in significant amounts during the later periods, as indicated by changes in the intensity of Peak III (Fig. 4(A)). Because green oil cannot form on the support itself, i.e., without the assistance of the Pd surface [7,9,26], the above result suggests that green oil, which was initially deposited in large amounts on or in the vicinity of Pd, migrated to the support during the later periods of deactivation. Compared with Pd/SiO2, Pd-0.5Ag/SiO2 shows peaks of much lower intensity in Fig. 4(B). However, a trend in the characteristic change of the peaks with catalyst deactivation is similar to one observed with Pd/SiO2. That is, green oil, most of which was initially deposited on or in the vicinity of the Pd surface (represented by Peak II), was polymerized to a heavier species and spilled over to the support after the middle period of the reaction (Period II). However, the amounts of green oil deposited on the Pd were very small and the polymerization of green oil was retarded, as indicated by the slight shift of Peak II with the same low intensity. These results indicate that Ag addition suppressed the polymerization of green oil as well as its formation. In the case of Pd-0.5Ag/Period III, the amounts of green oil adsorbed on the catalyst surface (Peak I) and deposited on the support (Peak III) were relatively large whereas those on the Pd surface (Peak II) were small. Changes in the properties of deposited green oil by Ag addition will be discussed in more detail in Section 4.2.
41
tion of 1,3-butadiene on the catalyst surface [13]. Large amounts of 1,3-butadiene should accelerate the formation of green oil, which will eventually decrease catalytic activity. Therefore, the activity of Pd/SiO2 decreased slightly in Stage II due to the accumulation of 1,3-butadiene (Fig. 2), whereas it remained nearly constant in Stage I. In the later stage of the process (Stage III), the catalytic activity decreased drastically (Fig. 1) even though green oil was deposited only in small amounts (Fig. 3). This is because green-oil spillover from the Pd to the support was retarded and catalyst pores were gradually blocked by the polymerized species of green oil [9,12]. A significant increase in the molecular weights of green oil in Stage III was indicated by a large shift of Peak II in Fig. 4(A). Green oil forms a thick film on the Pd surface due to the suppressed green-oil spillover and, accordingly, the rate of acetylene hydrogenation is limited by the transport of hydrogen in the surface film [9,11]. Liu et al. also reported the role of green oil as a diffusion barrier [32]. They observed that small acetylene and hydrogen molecules can diffuse through the thin green-oil layer to reach the Pd surface so the catalyst remains active, which corresponds to Stage I in this study. As the green oil builds up to form a thick layer, this growing diffusion barrier significantly reduces the activity of the catalyst, which is in accordance with the drastic activity decrease in Stage III (Fig. 1). Consequently, it can be concluded that various factors, including suppressed green-oil spillover, fast chain growth of green oil on Pd, enhanced diffusion limitation, and pore blocking, were responsible for the drastic decrease in catalytic activity in Stage III even when the amounts of green oil were only slightly increased.
4. Discussion 4.2. Effect of Ag addition 4.1. Three stages of deactivation for Pd/SiO2 Based on the above observations, it can be concluded that the deactivation of Pd/SiO2 by green-oil deposition proceeds over three stages. Periods I, II and III denoted for Pd/SiO2 can be designated as Stage I, II and III of the deactivation process, respectively. DTGA of Pd/Periods I (Fig. 4(A)) indicated that initially green oil was deposited largely on, or in the vicinity of, the Pd instead of the support. However, the green oil formed during the initial reaction period (Stage I), in amounts as large as 34 wt%, was not detrimental to the activity (Fig. 1). The above results can be explained by the hydrogen transfer mechanism, which was initially proposed by Thomson and Webb [27] and referred to by Borodzinski et al. [6,28–30]. The acetylenic species actually participated in the catalytic reaction, but was not necessarily adsorbed onto an extended Pd surface and could either be absorbed into or adsorbed onto a carbonaceous overlayer that covered the Pd [17–19,27,31]. An initial layer was formed on the Pd surface from acetylene that was dissociatively adsorbed onto the Pd. A second layer was subsequently formed on the first layer by the associative adsorption of acetylene, which was eventually converted to ethylene [17–19]. It was proposed that the carbonaceous overlayer worked as a medium to allow the transfer of hydrogen from the Pd surface to the acetylene that was associated with the overlayer [6,27–31]. Accordingly, the rates of acetylene hydrogenation was unaffected by the large amounts of green oil deposited in the initial stage (Stage I), because hydrogen adsorbed onto the Pd surface was successfully transferred to the acetylenic species that were either absorbed into or adsorbed onto the green-oil layer. It was previously reported that green-oil formation suppressed the rates of 1,3-butadiene conversion to butenes or high molecular-weight species to greater extents than those of 1,3butadiene formation from acetylene, which led to the accumula-
The above deactivation behavior changed when Pd/SiO2 was promoted with Ag to extend the catalyst lifetime. During Periods I and II, Pd-0.5Ag/SiO2 produced smaller amounts of green oil than in the case of Pd/SiO2 while the decrease in the activity was similar for both catalysts (Fig. 3). The formation of green oil was retarded because 1,3-butadiene was produced in smaller amounts during Period II on Pd-0.5Ag/SiO2 (Fig. 2). During the later period of the reaction (Period III), the decrease in the activity of Pd-0.5Ag/SiO2 was 20 times smaller than that of Pd/SiO2, but similar amounts of green oil, less than 10 wt%, were deposited on both catalysts. This result indicates that Stage III, which was characterized by a significant decease in activity due to the accumulation of heavy green oil, was not observed in the case of Pd-0.5Ag/SiO2. The added Ag species not only suppressed the formation and polymerization of green oil but also affected the properties of green oil produced. The H/C atomic ratio of green oil, listed in Table 1, was slightly higher for Pd-0.5Ag/SiO2 than it was for Pd/SiO2, indicating that green oil formed on the former catalyst was either lighter or more volatile with a short chain length. Contrary to the case of the Pd/SiO2 catalyst, all DTGA peaks obtained from Pd0.5Ag/SiO2 were observed at lower temperatures, and the shift of Peak II, representing green oil deposited on the Pd surface, was not significant. Furthermore, the amounts of green oil adsorbed onto the catalyst surface (Peak I) and deposited on the support (Peak III) were relatively large whereas those on the Pd surface (Peak II) were small. These results indicate that green oil, primarily formed on the Pd surface in small amounts, easily moved to the support in the case of Pd-0.5Ag/SiO2. The period required for producing the maximum concentration of 1,3-butadiene was also shorter for Pd0.5Ag/SiO2 than for Pd/SiO2 (Fig. 2). Accordingly, Stage II, which was responsible for the transfer of green oil from the Pd to the support, was observed at an earlier period when the catalyst was promoted with Ag.
42
I.Y. Ahn et al. / Applied Catalysis A: General 360 (2009) 38–42
The deactivation was significantly retarded when Ag was added to Pd/SiO2. Although the initial activity of Pd-0.5Ag/SiO2 was slightly lower than that of Pd/SiO2, the amount of green oil deposited on the former catalyst was much smaller than that deposited on the latter. Consequently, the final stage of deactivation (Stage III), due to diffusion limitation in catalyst pores, was not observed with Pd0.5Ag/SiO2 during the reaction period of this study. Pd-0.5Ag/SiO2 showed Stage II, characterized by the transfer of green oil from the Pd to the support, at an earlier period than in the case of Pd/SiO2, which resulted in the suppressed polymerization of green oil on the Pd. Added Ag geometrically blocked multi-coordinated large ensembles of the Pd surface and also modified Pd electronically to facilitate the desorption of 1,3-butadiene from the Pd surface. Consequently, the green oil that formed on the Pd-0.5Ag/SiO2 became more volatile and mobile than that formed on Pd/SiO2. Acknowledgements Fig. 5. Infrared spectra of CO adsorbed on Pd/SiO2 (—) and Pd-0.5Ag/SiO2 (- - - -).
To further investigate the surface properties of Pd-0.5Ag/SiO2, IR spectra of CO adsorbed onto the catalyst were obtained and compared with those observed for Pd/SiO2 (Fig. 5). When Ag was added to the Pd catalyst, the intensity of peaks observed at wavenumbers below 2000 cm 1, representing the multi-coordinated CO species, was lowered. The area ratio of linear-bound to multi-bound bands, Al/Am, increased from 0.38 to 0.68, indicating that deposited Ag effectively blocked the sites consisting of two or three adjacent Pd atoms [22,33]. Because 1,3-butadiene was produced by the dimerization of acetylenic species that were adsorbed as neighbors on the Pd surface, the blockage of adjacent Pd atoms by added Ag on the catalyst surface should suppress the formation of green oil (Fig. 3) and, consequently, the rates of catalyst deactivation (Fig. 1). It was reported that added Ag also modified the electronic properties of Pd [16,34,35], as indicated by an increase in the Pd dband electron density [16]. In previous studies, electronic modification of Pd by an added promoter reduced the adsorption strength of ethylene on the surface, thus facilitating the desorption of ethylene from the surface [13,36,37]. As butadiene was reported to follow the same trend as ethylene in the mode and energy of adsorption on Pd [38], adsorption of 1,3-butadiene would be weaker on Pd-0.5Ag/SiO2 than in the case of Pd/SiO2. Consequently, both geometric and electronic factors are responsible for the improved deactivation behavior of the Pd-0.5Ag/SiO2 catalyst. 5. Conclusions The deactivation of Pd/SiO2 in acetylene hydrogenation proceeded in three stages. In Stage I, a large amount of relatively light green oil was deposited onto the catalyst, but the activity was only slightly decreased because acetylene, which was associatively adsorbed onto the green oil, was readily hydrogenated via a hydrogen transfer mechanism. As the deactivation proceeded (Stage II), 1,3-butadiene accumulated on the Pd surface and polymerized to relatively heavy green oil, a part of which moved from the Pd surface to the support. In the later stage of the process (Stage III), the activity was drastically decreased even though the amount of green oil was only slightly increased. This was because catalyst pores were blocked by the accumulated green oil and hydrogen diffusion was limited in the thick film of the relatively heavy green oil.
This study was supported by BASF Aktiengesellschaft, Germany, by the Brain Korea 21 project, and by the Korea Science and Engineering Foundation (KOSEF) through the National Research Laboratory program of Ministry of Science and Technology (No. M10318000095-04J0000-02810), Korea. References [1] M.L. Derrien, Stud. Surf. Sci. Catal. 27 (1986) 613–666. [2] J.T. Wehrli, D.J. Thomas, M.S. Wainwright, D.L. Trimm, N.W. Cant, Catalyst Deactivation, Studies in Surface Science and Catalysis, Elsevier Science BV, 1991,, p. 203. [3] P. Albers, K. Seibold, G. Prescher, H. Muller, Appl. Catal. A 176 (1999) 135–146. [4] J.P. Boitiaux, J. Cosyns, M. Derrien, G. Leger, Hydrocarbon Process. 64 (1985) 51– 59. [5] D.K. Lee, Korean J. Chem. Eng. 7 (1990) 233–235. [6] A. Borodzinski, G.C. Bond, Catal. Rev. 48 (2006) 91–144. [7] A. Sarkany, L. Guczi, A.H. Weiss, Appl. Catal. 10 (1984) 369–388. [8] G.C. Battiston, L. Dalloro, G.R. Tauszik, Appl. Catal. 2 (1982) 1–17. [9] A. Sarkany, A.H. Weiss, T. Szilagyi, P. Sandor, L. Guczi, Appl. Catal. 12 (1984) 373– 379. [10] M.R. Stammbach, D.J. Thomas, D.L. Trimm, M.S. Wainwright, Appl. Catal. 58 (1990) 209–217. [11] S. Asplund, J. Catal. 158 (1996) 267–278. [12] J. Wood, L.F. Gladden, Appl. Catal. A 249 (2003) 241–253. [13] I.Y. Ahn, J.H. Lee, S.S. Kum, S.H. Moon, Catal. Today 123 (2007) 151–157. [14] I.Y. Ahn, W.J. Kim, S.H. Moon, Appl. Catal. A 308 (2006) 75–81. [15] S.H. Moon, H. Windawi, J.R. Katzer, Ind. Eng. Chem. Fundam. 20 (1981) 396–399. [16] D.C. Huang, K.H. Chang, W.F. Pong, P.K. Tseng, K.J. Hung, W.F. Huang, Catal. Lett. 53 (1998) 155–159. [17] A.S. Al-Ammar, G. Webb, J. Chem. Soc. Faraday Trans. I 74 (1978) 195–205. [18] A.S. Al-Ammar, G. Webb, J. Chem. Soc. Faraday Trans. I 74 (1978) 657–664. [19] A.S. Al-Ammar, G. Webb, J. Chem. Soc. Faraday Trans. I 75 (1979) 1900–1911. [20] M. Larsson, J. Jansson, S. Asplund, J. Catal. 178 (1998) 49–57. [21] P. Albers, J. Pietsch, S.F. Parker, J. Mol. Catal. A 173 (2001) 275–286. [22] W.J. Kim, E.W. Shin, J.H. Kang, S.H. Moon, Appl. Catal. A 251 (2003) 305–313. [23] G.A. Somorjai, S.M. Davis, Chem. Tech. (August 1983) 502–511. [24] A.H. Weiss, S. Leviness, V. Nair, L. Guczi, A. Sarkany, Z. Schay, Int. Congr. Catal., 8th Proceeding V5 (1984) 591–599. [25] S. Leviness, V. Nair, A.H. Weiss, Z. Schay, L. Guczi, J. Mol. Catal. 25 (1984) 131–140. [26] J. Houzvicka, R. Pestman, V. Ponec, Catal. Lett. 30 (1995) 289–296. [27] S.J. Thomson, G. Webb, J. Chem. Soc. Chem. Commun. 13 (1976) 526–527. [28] A. Borodzinski, A. Golebiowski, Langmuir 13 (1997) 883–887. [29] A. Borodzinski, Catal. Lett. 63 (1999) 35–42. [30] A. Borodzinski, Catal. Lett. 71 (2001) 169–176. [31] G. Webb, Catal. Today 7 (1990) 139–155. [32] R.-J. Liu, P.A. Crozier, C.M. Smith, D.A. Hucul, J. Blackson, G. Salaita, Appl. Catal. A 282 (2005) 111–121. [33] H. Zea, K. Lester, A.K. Daty, W. Rightor, R. Gulotty, W. Waterman, M. Smith, Appl. Catal. A 282 (2005) 237–245. [34] Q. Zhang, J. Li, X. Liu, Q. Zhu, Appl. Catal. A 197 (2000) 221–228. [35] P.A. Sheth, M. Neurock, C.M. Smith, J. Phys. Chem. B 109 (2005) 12449–12466. [36] J.H. Kang, E.W. Shin, W.J. Kim, J.D. Park, S.H. Moon, Catal. Today 63 (2000) 183– 188. [37] J.H. Kang, E.W. Shin, W.J. Kim, J.D. Park, S.H. Moon, J. Catal. 208 (2002) 310–320. [38] C. Breinlich, J. Haubrich, C. Becker, A. Valcarcel, F. Delbecq, K. Wandelt, J. Catal. 251 (2007) 123–130.