- Email: [email protected]
Precise recognition of catalyst deactivation during acetylene hydrogenation studied with the advanced TEMKIN reactor
Catalysis Communications 72 (2015) 170–173
Contents lists available at ScienceDirect
Catalysis Communications journal homepage: www.elsevier.com/locate/catcom
Short communication
Precise recognition of catalyst deactivation during acetylene hydrogenation studied with the advanced TEMKIN reactor Martin Kuhn, Martin Lucas, Peter Claus ⁎ Technische Universität Darmstadt, Department of Chemistry, Ernst-Berl-Institute, Chemical Technology II, Alarich-Weiss-Straße 8, D-64287 Darmstadt, Germany
a r t i c l e
i n f o
Article history: Received 21 July 2015 Received in revised form 30 September 2015 Accepted 1 October 2015 Available online 5 October 2015
a b s t r a c t C6 hydrocarbons were identified as an important indicator for deactivation behaviour in acetylene hydrogenation on Pd–Ag/Al2O3 egg-shell catalysts. Thereby significant differences between highly developed catalysts can be identified, helping to forecast differences in deactivation behaviour over 100 h time-on-stream in a short and cheap catalytic test. © 2015 Elsevier B.V. All rights reserved.
Keywords: Egg-shell catalysts Selective acetylene hydrogenation Deactivation Advanced TEMKIN reactor
1. Introduction Ethylene is one of the most important petrochemical bulk chemicals. It is mostly used for polymerization to polyethylene with different properties. Ethylene is mainly produced by steam cracking of naphtha at high temperatures above 800 °C; impurities (e.g. acetylene) are formed. The acetylene content has to be reduced to concentrations lower than 1 ppmv to avoid irreversible damage of the catalysts used in downstream processes of the steam cracker. The catalytic removal is carried out by selective hydrogenation of acetylene to ethylene. Pd–Ag/Al2O3 eggshell catalysts with low metal loadings are commercially used at front-end and tail-end conditions [1]. Hydrogenation of acetylene under conditions described above leads to ethylene and ethane caused by consecutive hydrogenation (Scheme 1). An essential problem of the industrially used Pd–Ag/Al2O3 eggshell catalysts in the selective hydrogenation of acetylene under tail-end conditions is the fast deactivation, induced by formation and deposition of hydrocarbons (green oil) and coke on the catalytic surface (Scheme 1) [2]. Butadiene is proposed to be a precursor of the green oil formation [3–5], mainly consisting of olefins and paraffins as well as aromatics and diolefins [6,7]. Green oil and coke form during hydrogenation of acetylene leading to a decrease of conversion and reduction of time-on-stream (TOS) [2]. A previous study of our working group demonstrates green oil and coke formation as well as the start of deactivation in the first part of the reactor that proceeds along the reactor [8]. However at the beginning of the reaction subsurface carbon is formed, suppresses subsurface hydrogen and palladium hydride, and affect the total hydrogenation of acetylene [9]. ⁎ Corresponding author. E-mail address: [email protected] (P. Claus).
http://dx.doi.org/10.1016/j.catcom.2015.10.001 1566-7367/© 2015 Elsevier B.V. All rights reserved.
Further pore filling of the catalyst by green oil and coke causes a reduction of the effective diffusion coefficient (Deff) followed by an increase of ethane formation [3]. Additionally, a spillover mechanism for ethylene hydrogenation to ethane is discussed in literature. Activated hydrogen is transferred from palladium to the support over the coke depositions and converted with ethylene to ethane [7]. Addition of Ag to Pd/Al2O3 catalysts suppresses the palladium hydride formation and spillover hydrogen and increases selectivity to ethylene [10]. Furthermore the deposited green oil serves as hydrogen reservoir especially for the consecutive hydrogenation to ethane [7]. The target improvements of the Pd–Ag/Al2O3 eggshell catalysts are to increase the selectivity to ethylene, raising the amount of ethylene gain and to force up the long-term stability. For the presented study commercial Pd/Al2O3 and Pd–Ag/Al2O3 eggshell catalysts were used. Cat B and Cat C are silver containing Pd–Ag/Al2O3 eggshell catalysts and Cat D is the silver free version of Cat C. Characterization of these catalysts by Pachulski et al. [2] is summarized in Table 1. The objective of this study is to point out the importance of the formed by-products detected via online GC to increase the significance of a catalytic test in relation to deactivation and long-term stability. 2. Material and methods The standard catalytic tests and the 100 h long-term tests were performed in the Advanced TEMKIN reactor [8] at 10 bar gauge pressure, 45 °C and a GHSV of 4000 h−1. The reaction mixture was based on tail-end conditions and contains 1 vol.% acetylene, 1 vol.% hydrogen, 1 vol.% propane (internal standard) and 30 vol.% ethylene in argon. The catalysts were reduced in situ for 1 h at 100 °C in 100% hydrogen prior to catalytic testing. For more information see our previous work [8].
M. Kuhn et al. / Catalysis Communications 72 (2015) 170–173
171
Scheme 1. Reaction network of the selective hydrogenation of acetylene.
Additionally the exhaust gas was analysed with a GC-MS (Shimadzu, GCMS-QP2010SE). 3. Results and discussion The three catalysts, described above, were tested for selective hydrogenation of acetylene under tail-end conditions. The testing procedure, the used Advanced TEMKIN reactor and detailed results were described elsewhere [8]. The used Advanced TEMKIN reactor is predestined for highly reproducible testing of egg-shell catalysts without transport limitations and effect of reactor design [8,11]. The results of a standard catalytic test at industrial conditions are summarized in Table 2. Depending on the Ag content of the catalysts, the highest conversion was found for Cat D with 90.2% followed by Cat C (76.9%) and Cat B (59.1%). The selectivity to ethylene was determined indirectly via online GC, because changes in the ethylene concentration cannot be detected accurately [12], appropriate the selectivity of the by-products ethane and the C4 compounds were used (see supplementary information). The catalytic results show the highest selectivity to ethylene for Cat B and Cat C, but no significant differences in selectivity to ethane, C4 and ethylene between these two catalysts can be determined. However, for the silver free palladium catalyst Cat D the amounts of formed ethane and C4 are much higher, resulting in a decreased selectivity to ethylene by 7% in comparison with the other two catalysts. Based on these catalytic results and the literature, assuming C4 or rather 1,3-butadiene as precursor for green oil formation [3–5], the long-term stability can be estimated. Therefore for Cat B and Cat C nearly the same deactivation behaviour can be expected but Cat D will be deactivate much faster. Long-term measurements were performed under industrial tail-end conditions over 100 h, using the three catalysts described above. The observed changes in conversion after a standard catalytic test, including run-in period and variation of the modified residence time [8], are shown in Fig. 1. The fastest decrease in activity was detected for the silver free palladium catalyst Cat D with a loss of conversion of about 28% over 75 h TOS, as expected out of the catalytic results in Table 2. The forecast for long-term stability of Cat B and Cat C cannot be confirmed. The conversion of Cat B decreases with 10% much faster than Cat C with 5% over 75 h TOS. Deactivation of palladium containing catalysts in the selective hydrogenation of acetylene is caused by deposition of higher hydrocarbons and coke [2]. Therefore the deposited amount of these by-products after 100 h long-term test was determined by weighing the three catalysts before and after each test. This method was verified by thermogravimetric analysis. For Cat D the highest increase of mass was determined with 14.3 wt.%, followed by Cat B with 5.1 wt.% and Cat C with 2.3 wt.%. This observation reflects the
Table 1 Characteristics of industrial catalysts for selective hydrogenation of acetylene. Source: [2].
Pd [wt.%] Ag [wt.%] nCO [μmolCO/gCat] Catalytic zone
determined loss in conversion over 75 h TOS. Additionally the selectivity to ethane increases for Cat D from 7% to 41% as shown in Fig. 2. For Cat B the selectivity to ethane increases by 10% to 14%. Only for Cat C the ethane selectivity slightly increases to 6%. Accordingly, the deactivation (Fig. 1) and the increased consecutive hydrogenation of ethylene to ethane (Fig. 2) can be correlated with the progress of hydrocarbon and coke deposition. Hg-porosimetry of fresh and used catalysts implicates a reduction of the pore volume from 0.44 ml g−1 to 0.42 ml g−1 for Cat D caused by deposits, while the pore volume for Cat C is constant (see supplementary information). The pore size distribution shifts to smaller pore sizes. Pore filling and blocking of the catalyst by green oil and coke causes a reduction of the effective diffusion coefficient (Deff). Asplund [3] estimates a reduction of Deff by 1/10. The rising mass transfer resistance leads to an increased consecutive hydrogenation of ethylene to ethane. This phenomenon is especially pronounced for Cat D and Cat B because of the increased green oil and coke formation, indicated by the C6 selectivity. Furthermore this can be explained by an increasing Thiele module φ: the degree of efficiency η is decreased and the conversion is decreased as well, by reduction of the Deff. Thermal gravimetric analysis (TGA) confirms the amount of deposited hydrocarbons on the catalyst surface, determined by weighing the catalysts before and after each test. The derived thermogravimetric signals (DTG) (see supplementary information) reflect the results of Pachulski et al. [2] and four ranges of mass loss can be observed: loss of humidity (up to 110 °C), a mobilization of light hydrocarbons (120 °C to 250 °C), burning off heavy hydrocarbons (280 °C to 350 °C) and burning off fixed coke (400 °C to 510 °C). The first two ranges were detected as endothermal processes and the last two peaks were detected as exothermal processes via differential scanning calorimetry (DSC), affirming the published characterizations. Further characterization of the deactivated catalysts is reported by Pachulski et al. [2]. They clearly show that the deactivation of the used Pd–Ag/Al2O3 eggshell catalysts is caused by deposition of green oil and coke on the catalytic surface. Regeneration in a steam–air-mixture and reduction in hydrogen offers the possibility to regenerate the catalysts without an observable effect on catalytic performance. For further qualitative analysis of the type of deposited hydrocarbons, the used catalysts were washed with n-pentane and the yellow solution was analysed via GC-MS. Hydrocarbons from C6 to C26 were found, but no significant differences between the used catalysts can be determined. These findings are close to the observations of LeViness [13]. The observations show, that the type of formed and deposited by-products on the palladium containing catalysts are equal, but the amount of depositions is important for deactivation and longterm stability. Nevertheless the differences between Cat B and Cat C in long-term behaviour cannot be estimated by results of a standard catalytic test Table 2 Catalytic result of industrial acetylene hydrogenation eggshell catalysts (T = 45 °C, p = 10
Cat B
Cat C
Cat D
0.035 0.079 0.45 400
0.036 0.015 0.09 300
0.036 0 0.65 300
bar, GHSV = 4000 h−1). Catalyst
Xacetylene [%]
Sethylene [%]
Sethane [%]
SΣC4 [%]
Cat B Cat C Cat D
59.1 76.9 90.2
82.6 83.0 76.1
4.1 4.8 6.5
13.3 12.2 17.3
172
M. Kuhn et al. / Catalysis Communications 72 (2015) 170–173
Fig. 1. Changes of conversion over 100 h of industrial Pd/Al2O3 and Pd–Ag/Al2O3 egg-shell catalysts (T = 45 °C, p = 10 bar, GHSV = 4000 h−1).
(Table 2). The methodology of product analysis via GC was improved and extended. In addition to the already known by-products ethane and C4 compounds further peaks of significant area were detected. The peaks were identified via GC-MS as C6 compounds particularly isomers of hexene (C6H12) and hexadiene (C6H10) as well as benzene (see supplementary information). The detected compounds were included in the determination of the improved selectivity to ethylene (Eq. (1)). Sethylene;improved ¼ 100−Sethane −SX
C4
−SX
C6
ð1Þ
Fig. 3 apportions the selectivity described above. Cat D exhibits the highest selectivity to the C6 compounds in a standard test and consequently the lowest improved ethylene selectivity. Significant differences in selectivity can be observed between Cat B and Cat C presently, which is different from what is shown in Table 2, due to widening of by-product detection. Cat B forms significantly more C6 compounds than Cat C. The differences in selectivity to C6H10 and C6H12 are very low with 0.3% and 0.5%; but the selectivity to benzene is 2.8%, much higher than Cat C with 1.4%. In summary, the C6 selectivity is 46% higher for Cat B in comparison with Cat C, giving a hint for an increased green oil formation. As described above green oil formation and the succeeding coke deposition are mainly considered as the reasons for catalyst deactivation in selective hydrogenation of acetylene [2]. Therefore the differences between Cat B and Cat C can be described by an increased green oil and coke formation, causing a blocking of the
Fig. 3. Selectivity to the observed C6 compounds and the improved ethylene selectivity of industrial Pd/Al2O3 and Pd–Ag/Al2O3 egg-shell catalysts (T = 45 °C, p = 10 bar, GHSV = 4000 h−1).
catalytic surface and filling of the pore system of the catalysts. As a result the observed conversion is decreased and the selectivity to ethane is increased. Using the improved catalytic results in selective hydrogenation of acetylene under tail-end condition of industrial Pd/Al2O3 and Pd–Ag/ Al2O3 egg-shell catalysts it is possible to point out necessary differences in selectivity and deactivation behaviour in a simple and short catalytic test of highly developed catalyst generations. Additionally time and costs can be reduced, due to economic aspects, and the carbon balance of the overall process can be completed. The formation of C6H10 and C6H12 can be described based on coupling and hydrogenation of acetylene, but butadiene is proposed to be involved in this process and in the formation of heavy hydrocarbons and green oil as well [6]. Characterization of the used catalysts via chemisorption of CO, TEM and XPS by Pachulski et al. [2] ensured a higher enrichment of silver on the catalytic surface for Cat C instead of Cat B and consequently a better isolation of the palladium surface atoms. The number of neighbored palladium is reduced and as a consequence the formation of linear C4, C6 and higher hydrocarbons is suppressed. The formation of benzene on a palladium surface in the presence of acetylene is controversially discussed in literature. Baddeley et al. [14] postulated that cyclization of acetylene to benzene takes place on large Pd ensembles. Three acetylene molecules adsorb simultaneously on a palladium ensemble with a critical size of Pd7, followed by the formation of benzene. Kaltchev et al. [15] proposed that under high pressure conditions an initial reaction of adsorbed vinylidene with acetylene forming C4H4 is followed by addition of acetylene to form benzene. A theoretical study of Pacchioni et al. [16] characterized a stable C4H4 chain as the key intermediate for benzene formation out of acetylene. The benzene formation at Cat C is reduced in comparison to Cat B due to better isolation of the palladium surface atoms. Due to the observations emphasized above, the detected unsaturated C4 and C6 compounds are important precursors and intermediate for green oil and coke formation (Scheme 1). Additionally, traces of unsaturated C8 and C10 compounds can be detected in the exhaust gas via GC-MS. 4. Conclusions
Fig. 2. Selectivity to ethane over 100 h of industrial Pd/Al2O3 and Pd–Ag/Al2O3 egg-shell catalysts (T = 45 °C, p = 10 bar, GHSV = 4000 h−1).
In summary, the formation of hexenes, hexadienes and benzene in selective hydrogenation of acetylene under tail-end conditions using Pd/Al2O3 and Pd–Ag/Al2O3 egg-shell catalysts was proved and included in determination of the improved ethylene selectivity. The formation of C6 compounds is an important indicator for long-term stability and deactivation behaviour of the used egg-shell catalysts, due to the formation and deposition of green oil and coke on the catalytic surface. Thereby significant differences between two highly developed industrial Pd–
M. Kuhn et al. / Catalysis Communications 72 (2015) 170–173
Ag/Al2O3 egg-shell catalysts can be identified, helping to forecast the differences in deactivation behaviour over 100 h TOS in a short and cheap catalytic test. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.catcom.2015.10.001. References [1] [2] [3] [4]
A. Borodziński, G.C. Bond, Catal. Rev. 48 (2006) 91–144. A. Pachulski, R. Schödel, P. Claus, Appl. Catal. A 400 (2011) 14–24. S. Asplund, J. Catal. 158 (1996) 267–278. B. Yang, R. Burch, C. Hardacre, P. Hu, P. Hughes, J. Phys. Chem. C 118 (2013) 1560–1567.
173
[5] C.S. Spanjers, J.T. Held, M.J. Jones, D.D. Stanley, R.S. Sim, M.J. Janik, R.M. Rioux, J. Catal. 316 (2014) 164–173. [6] I.Y. Ahn, J.H. Lee, S.S. Kum, S.H. Moon, Catal. Today 123 (2007) 151–157. [7] A. Sárkány, L. Guczi, A.H. Weiss, Appl. Catal. 10 (1984) 369–388. [8] M. Kuhn, M. Lucas, P. Claus, Chem. Eng. Technol. 38 (2015) 61–67. [9] D. Teschner, J. Borsodi, A. Wootsch, Z. Révay, M. Hävecker, A. Knop-Gericke, S.D. Jackson, R. Schlögl, Science 320 (2008) 86–89. [10] P. Praserthdam, B. Ngamsom, N. Bogdanchikova, S. Phatanasri, M. Pramotthana, Appl. Catal. A 230 (2002) 41–51. [11] D. Götz, M. Kuhn, P. Claus, Chem. Eng. Res. Des. 94 (2015) 594–604. [12] J. Osswald, Active-Site Isolation for the Selective Hydrogenation of Acetylene: The Pd–Ga and Pd–Sn Intermetallic Compounds, Technische Universität Berlin, 2006. [13] S.C. LeViness, PhD Thesis, Rice University, 1989. [14] C.J. Baddeley, M. Tikhov, C. Hardacre, J.R. Lomas, R.M. Lambert, J. Phys. Chem. 100 (1996) 2189–2194. [15] M. Kaltchev, D. Stacchiola, H. Molero, G. Wu, A. Blumenfeld, W.T. Tysoe, Catal. Lett. 60 (1999) 11–14. [16] G. Pacchioni, R.M. Lambert, Surf. Sci. 304 (1994) 208–222.
Contents lists available at ScienceDirect
Catalysis Communications journal homepage: www.elsevier.com/locate/catcom
Short communication
Precise recognition of catalyst deactivation during acetylene hydrogenation studied with the advanced TEMKIN reactor Martin Kuhn, Martin Lucas, Peter Claus ⁎ Technische Universität Darmstadt, Department of Chemistry, Ernst-Berl-Institute, Chemical Technology II, Alarich-Weiss-Straße 8, D-64287 Darmstadt, Germany
a r t i c l e
i n f o
Article history: Received 21 July 2015 Received in revised form 30 September 2015 Accepted 1 October 2015 Available online 5 October 2015
a b s t r a c t C6 hydrocarbons were identified as an important indicator for deactivation behaviour in acetylene hydrogenation on Pd–Ag/Al2O3 egg-shell catalysts. Thereby significant differences between highly developed catalysts can be identified, helping to forecast differences in deactivation behaviour over 100 h time-on-stream in a short and cheap catalytic test. © 2015 Elsevier B.V. All rights reserved.
Keywords: Egg-shell catalysts Selective acetylene hydrogenation Deactivation Advanced TEMKIN reactor
1. Introduction Ethylene is one of the most important petrochemical bulk chemicals. It is mostly used for polymerization to polyethylene with different properties. Ethylene is mainly produced by steam cracking of naphtha at high temperatures above 800 °C; impurities (e.g. acetylene) are formed. The acetylene content has to be reduced to concentrations lower than 1 ppmv to avoid irreversible damage of the catalysts used in downstream processes of the steam cracker. The catalytic removal is carried out by selective hydrogenation of acetylene to ethylene. Pd–Ag/Al2O3 eggshell catalysts with low metal loadings are commercially used at front-end and tail-end conditions [1]. Hydrogenation of acetylene under conditions described above leads to ethylene and ethane caused by consecutive hydrogenation (Scheme 1). An essential problem of the industrially used Pd–Ag/Al2O3 eggshell catalysts in the selective hydrogenation of acetylene under tail-end conditions is the fast deactivation, induced by formation and deposition of hydrocarbons (green oil) and coke on the catalytic surface (Scheme 1) [2]. Butadiene is proposed to be a precursor of the green oil formation [3–5], mainly consisting of olefins and paraffins as well as aromatics and diolefins [6,7]. Green oil and coke form during hydrogenation of acetylene leading to a decrease of conversion and reduction of time-on-stream (TOS) [2]. A previous study of our working group demonstrates green oil and coke formation as well as the start of deactivation in the first part of the reactor that proceeds along the reactor [8]. However at the beginning of the reaction subsurface carbon is formed, suppresses subsurface hydrogen and palladium hydride, and affect the total hydrogenation of acetylene [9]. ⁎ Corresponding author. E-mail address: [email protected] (P. Claus).
http://dx.doi.org/10.1016/j.catcom.2015.10.001 1566-7367/© 2015 Elsevier B.V. All rights reserved.
Further pore filling of the catalyst by green oil and coke causes a reduction of the effective diffusion coefficient (Deff) followed by an increase of ethane formation [3]. Additionally, a spillover mechanism for ethylene hydrogenation to ethane is discussed in literature. Activated hydrogen is transferred from palladium to the support over the coke depositions and converted with ethylene to ethane [7]. Addition of Ag to Pd/Al2O3 catalysts suppresses the palladium hydride formation and spillover hydrogen and increases selectivity to ethylene [10]. Furthermore the deposited green oil serves as hydrogen reservoir especially for the consecutive hydrogenation to ethane [7]. The target improvements of the Pd–Ag/Al2O3 eggshell catalysts are to increase the selectivity to ethylene, raising the amount of ethylene gain and to force up the long-term stability. For the presented study commercial Pd/Al2O3 and Pd–Ag/Al2O3 eggshell catalysts were used. Cat B and Cat C are silver containing Pd–Ag/Al2O3 eggshell catalysts and Cat D is the silver free version of Cat C. Characterization of these catalysts by Pachulski et al. [2] is summarized in Table 1. The objective of this study is to point out the importance of the formed by-products detected via online GC to increase the significance of a catalytic test in relation to deactivation and long-term stability. 2. Material and methods The standard catalytic tests and the 100 h long-term tests were performed in the Advanced TEMKIN reactor [8] at 10 bar gauge pressure, 45 °C and a GHSV of 4000 h−1. The reaction mixture was based on tail-end conditions and contains 1 vol.% acetylene, 1 vol.% hydrogen, 1 vol.% propane (internal standard) and 30 vol.% ethylene in argon. The catalysts were reduced in situ for 1 h at 100 °C in 100% hydrogen prior to catalytic testing. For more information see our previous work [8].
M. Kuhn et al. / Catalysis Communications 72 (2015) 170–173
171
Scheme 1. Reaction network of the selective hydrogenation of acetylene.
Additionally the exhaust gas was analysed with a GC-MS (Shimadzu, GCMS-QP2010SE). 3. Results and discussion The three catalysts, described above, were tested for selective hydrogenation of acetylene under tail-end conditions. The testing procedure, the used Advanced TEMKIN reactor and detailed results were described elsewhere [8]. The used Advanced TEMKIN reactor is predestined for highly reproducible testing of egg-shell catalysts without transport limitations and effect of reactor design [8,11]. The results of a standard catalytic test at industrial conditions are summarized in Table 2. Depending on the Ag content of the catalysts, the highest conversion was found for Cat D with 90.2% followed by Cat C (76.9%) and Cat B (59.1%). The selectivity to ethylene was determined indirectly via online GC, because changes in the ethylene concentration cannot be detected accurately [12], appropriate the selectivity of the by-products ethane and the C4 compounds were used (see supplementary information). The catalytic results show the highest selectivity to ethylene for Cat B and Cat C, but no significant differences in selectivity to ethane, C4 and ethylene between these two catalysts can be determined. However, for the silver free palladium catalyst Cat D the amounts of formed ethane and C4 are much higher, resulting in a decreased selectivity to ethylene by 7% in comparison with the other two catalysts. Based on these catalytic results and the literature, assuming C4 or rather 1,3-butadiene as precursor for green oil formation [3–5], the long-term stability can be estimated. Therefore for Cat B and Cat C nearly the same deactivation behaviour can be expected but Cat D will be deactivate much faster. Long-term measurements were performed under industrial tail-end conditions over 100 h, using the three catalysts described above. The observed changes in conversion after a standard catalytic test, including run-in period and variation of the modified residence time [8], are shown in Fig. 1. The fastest decrease in activity was detected for the silver free palladium catalyst Cat D with a loss of conversion of about 28% over 75 h TOS, as expected out of the catalytic results in Table 2. The forecast for long-term stability of Cat B and Cat C cannot be confirmed. The conversion of Cat B decreases with 10% much faster than Cat C with 5% over 75 h TOS. Deactivation of palladium containing catalysts in the selective hydrogenation of acetylene is caused by deposition of higher hydrocarbons and coke [2]. Therefore the deposited amount of these by-products after 100 h long-term test was determined by weighing the three catalysts before and after each test. This method was verified by thermogravimetric analysis. For Cat D the highest increase of mass was determined with 14.3 wt.%, followed by Cat B with 5.1 wt.% and Cat C with 2.3 wt.%. This observation reflects the
Table 1 Characteristics of industrial catalysts for selective hydrogenation of acetylene. Source: [2].
Pd [wt.%] Ag [wt.%] nCO [μmolCO/gCat] Catalytic zone
determined loss in conversion over 75 h TOS. Additionally the selectivity to ethane increases for Cat D from 7% to 41% as shown in Fig. 2. For Cat B the selectivity to ethane increases by 10% to 14%. Only for Cat C the ethane selectivity slightly increases to 6%. Accordingly, the deactivation (Fig. 1) and the increased consecutive hydrogenation of ethylene to ethane (Fig. 2) can be correlated with the progress of hydrocarbon and coke deposition. Hg-porosimetry of fresh and used catalysts implicates a reduction of the pore volume from 0.44 ml g−1 to 0.42 ml g−1 for Cat D caused by deposits, while the pore volume for Cat C is constant (see supplementary information). The pore size distribution shifts to smaller pore sizes. Pore filling and blocking of the catalyst by green oil and coke causes a reduction of the effective diffusion coefficient (Deff). Asplund [3] estimates a reduction of Deff by 1/10. The rising mass transfer resistance leads to an increased consecutive hydrogenation of ethylene to ethane. This phenomenon is especially pronounced for Cat D and Cat B because of the increased green oil and coke formation, indicated by the C6 selectivity. Furthermore this can be explained by an increasing Thiele module φ: the degree of efficiency η is decreased and the conversion is decreased as well, by reduction of the Deff. Thermal gravimetric analysis (TGA) confirms the amount of deposited hydrocarbons on the catalyst surface, determined by weighing the catalysts before and after each test. The derived thermogravimetric signals (DTG) (see supplementary information) reflect the results of Pachulski et al. [2] and four ranges of mass loss can be observed: loss of humidity (up to 110 °C), a mobilization of light hydrocarbons (120 °C to 250 °C), burning off heavy hydrocarbons (280 °C to 350 °C) and burning off fixed coke (400 °C to 510 °C). The first two ranges were detected as endothermal processes and the last two peaks were detected as exothermal processes via differential scanning calorimetry (DSC), affirming the published characterizations. Further characterization of the deactivated catalysts is reported by Pachulski et al. [2]. They clearly show that the deactivation of the used Pd–Ag/Al2O3 eggshell catalysts is caused by deposition of green oil and coke on the catalytic surface. Regeneration in a steam–air-mixture and reduction in hydrogen offers the possibility to regenerate the catalysts without an observable effect on catalytic performance. For further qualitative analysis of the type of deposited hydrocarbons, the used catalysts were washed with n-pentane and the yellow solution was analysed via GC-MS. Hydrocarbons from C6 to C26 were found, but no significant differences between the used catalysts can be determined. These findings are close to the observations of LeViness [13]. The observations show, that the type of formed and deposited by-products on the palladium containing catalysts are equal, but the amount of depositions is important for deactivation and longterm stability. Nevertheless the differences between Cat B and Cat C in long-term behaviour cannot be estimated by results of a standard catalytic test Table 2 Catalytic result of industrial acetylene hydrogenation eggshell catalysts (T = 45 °C, p = 10
Cat B
Cat C
Cat D
0.035 0.079 0.45 400
0.036 0.015 0.09 300
0.036 0 0.65 300
bar, GHSV = 4000 h−1). Catalyst
Xacetylene [%]
Sethylene [%]
Sethane [%]
SΣC4 [%]
Cat B Cat C Cat D
59.1 76.9 90.2
82.6 83.0 76.1
4.1 4.8 6.5
13.3 12.2 17.3
172
M. Kuhn et al. / Catalysis Communications 72 (2015) 170–173
Fig. 1. Changes of conversion over 100 h of industrial Pd/Al2O3 and Pd–Ag/Al2O3 egg-shell catalysts (T = 45 °C, p = 10 bar, GHSV = 4000 h−1).
(Table 2). The methodology of product analysis via GC was improved and extended. In addition to the already known by-products ethane and C4 compounds further peaks of significant area were detected. The peaks were identified via GC-MS as C6 compounds particularly isomers of hexene (C6H12) and hexadiene (C6H10) as well as benzene (see supplementary information). The detected compounds were included in the determination of the improved selectivity to ethylene (Eq. (1)). Sethylene;improved ¼ 100−Sethane −SX
C4
−SX
C6
ð1Þ
Fig. 3 apportions the selectivity described above. Cat D exhibits the highest selectivity to the C6 compounds in a standard test and consequently the lowest improved ethylene selectivity. Significant differences in selectivity can be observed between Cat B and Cat C presently, which is different from what is shown in Table 2, due to widening of by-product detection. Cat B forms significantly more C6 compounds than Cat C. The differences in selectivity to C6H10 and C6H12 are very low with 0.3% and 0.5%; but the selectivity to benzene is 2.8%, much higher than Cat C with 1.4%. In summary, the C6 selectivity is 46% higher for Cat B in comparison with Cat C, giving a hint for an increased green oil formation. As described above green oil formation and the succeeding coke deposition are mainly considered as the reasons for catalyst deactivation in selective hydrogenation of acetylene [2]. Therefore the differences between Cat B and Cat C can be described by an increased green oil and coke formation, causing a blocking of the
Fig. 3. Selectivity to the observed C6 compounds and the improved ethylene selectivity of industrial Pd/Al2O3 and Pd–Ag/Al2O3 egg-shell catalysts (T = 45 °C, p = 10 bar, GHSV = 4000 h−1).
catalytic surface and filling of the pore system of the catalysts. As a result the observed conversion is decreased and the selectivity to ethane is increased. Using the improved catalytic results in selective hydrogenation of acetylene under tail-end condition of industrial Pd/Al2O3 and Pd–Ag/ Al2O3 egg-shell catalysts it is possible to point out necessary differences in selectivity and deactivation behaviour in a simple and short catalytic test of highly developed catalyst generations. Additionally time and costs can be reduced, due to economic aspects, and the carbon balance of the overall process can be completed. The formation of C6H10 and C6H12 can be described based on coupling and hydrogenation of acetylene, but butadiene is proposed to be involved in this process and in the formation of heavy hydrocarbons and green oil as well [6]. Characterization of the used catalysts via chemisorption of CO, TEM and XPS by Pachulski et al. [2] ensured a higher enrichment of silver on the catalytic surface for Cat C instead of Cat B and consequently a better isolation of the palladium surface atoms. The number of neighbored palladium is reduced and as a consequence the formation of linear C4, C6 and higher hydrocarbons is suppressed. The formation of benzene on a palladium surface in the presence of acetylene is controversially discussed in literature. Baddeley et al. [14] postulated that cyclization of acetylene to benzene takes place on large Pd ensembles. Three acetylene molecules adsorb simultaneously on a palladium ensemble with a critical size of Pd7, followed by the formation of benzene. Kaltchev et al. [15] proposed that under high pressure conditions an initial reaction of adsorbed vinylidene with acetylene forming C4H4 is followed by addition of acetylene to form benzene. A theoretical study of Pacchioni et al. [16] characterized a stable C4H4 chain as the key intermediate for benzene formation out of acetylene. The benzene formation at Cat C is reduced in comparison to Cat B due to better isolation of the palladium surface atoms. Due to the observations emphasized above, the detected unsaturated C4 and C6 compounds are important precursors and intermediate for green oil and coke formation (Scheme 1). Additionally, traces of unsaturated C8 and C10 compounds can be detected in the exhaust gas via GC-MS. 4. Conclusions
Fig. 2. Selectivity to ethane over 100 h of industrial Pd/Al2O3 and Pd–Ag/Al2O3 egg-shell catalysts (T = 45 °C, p = 10 bar, GHSV = 4000 h−1).
In summary, the formation of hexenes, hexadienes and benzene in selective hydrogenation of acetylene under tail-end conditions using Pd/Al2O3 and Pd–Ag/Al2O3 egg-shell catalysts was proved and included in determination of the improved ethylene selectivity. The formation of C6 compounds is an important indicator for long-term stability and deactivation behaviour of the used egg-shell catalysts, due to the formation and deposition of green oil and coke on the catalytic surface. Thereby significant differences between two highly developed industrial Pd–
M. Kuhn et al. / Catalysis Communications 72 (2015) 170–173
Ag/Al2O3 egg-shell catalysts can be identified, helping to forecast the differences in deactivation behaviour over 100 h TOS in a short and cheap catalytic test. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.catcom.2015.10.001. References [1] [2] [3] [4]
A. Borodziński, G.C. Bond, Catal. Rev. 48 (2006) 91–144. A. Pachulski, R. Schödel, P. Claus, Appl. Catal. A 400 (2011) 14–24. S. Asplund, J. Catal. 158 (1996) 267–278. B. Yang, R. Burch, C. Hardacre, P. Hu, P. Hughes, J. Phys. Chem. C 118 (2013) 1560–1567.
173
[5] C.S. Spanjers, J.T. Held, M.J. Jones, D.D. Stanley, R.S. Sim, M.J. Janik, R.M. Rioux, J. Catal. 316 (2014) 164–173. [6] I.Y. Ahn, J.H. Lee, S.S. Kum, S.H. Moon, Catal. Today 123 (2007) 151–157. [7] A. Sárkány, L. Guczi, A.H. Weiss, Appl. Catal. 10 (1984) 369–388. [8] M. Kuhn, M. Lucas, P. Claus, Chem. Eng. Technol. 38 (2015) 61–67. [9] D. Teschner, J. Borsodi, A. Wootsch, Z. Révay, M. Hävecker, A. Knop-Gericke, S.D. Jackson, R. Schlögl, Science 320 (2008) 86–89. [10] P. Praserthdam, B. Ngamsom, N. Bogdanchikova, S. Phatanasri, M. Pramotthana, Appl. Catal. A 230 (2002) 41–51. [11] D. Götz, M. Kuhn, P. Claus, Chem. Eng. Res. Des. 94 (2015) 594–604. [12] J. Osswald, Active-Site Isolation for the Selective Hydrogenation of Acetylene: The Pd–Ga and Pd–Sn Intermetallic Compounds, Technische Universität Berlin, 2006. [13] S.C. LeViness, PhD Thesis, Rice University, 1989. [14] C.J. Baddeley, M. Tikhov, C. Hardacre, J.R. Lomas, R.M. Lambert, J. Phys. Chem. 100 (1996) 2189–2194. [15] M. Kaltchev, D. Stacchiola, H. Molero, G. Wu, A. Blumenfeld, W.T. Tysoe, Catal. Lett. 60 (1999) 11–14. [16] G. Pacchioni, R.M. Lambert, Surf. Sci. 304 (1994) 208–222.