Catalytic Partial Oxidation of Isooctane to Hydrogen on Rhodium

Article pubs.acs.org/IECR

Catalytic Partial Oxidation of Isooctane to Hydrogen on Rhodium Catalysts: Effect of Tail-Gas Recycling Torsten Kaltschmitt,†,‡ Claudia Diehm,§,∥ and Olaf Deutschmann*,†,§ †

Institute for Catalysis Research and Technology, ‡Institute for Nuclear and Energy Technologies, §Institute for Chemical Technology and Polymer Chemistry, and ∥Helmholtz Research School Energy-Related Catalysis at Karlsruhe Institute of Technology (KIT), P.O. Box 6980, 76049 Karlsruhe, Germany S Supporting Information *

ABSTRACT: Catalytic partial oxidation (CPOX) is a promising technology for reforming of liquid hydrocarbon fuels to hydrogen or synthesis gas for use in fuel cells. The addition of a certain amount of the tail gas of the fuel cell stack to the reformer inlet feed can increase overall efficiency and lead to higher H2 and CO selectivities and reduce coke formation. The effect of carbon dioxide or steam addition (1, 5, 10, 20, and 30 vol% of the total flow) on the performance of a CPOX reformer operated with isooctane as fuel surrogate is systematically studied over a wide range of C/O feed ratios (0.72−1.79) using a Rh/alumina honeycomb catalyst. The specific impact of the coreactants H2O and CO2 on reformer behavior can be interpreted by the water gas shift (WGS) chemistry. Production of H2 and CO2 increases with H2O addition at the expense of CO and H2O. Opposite trends are observed in case of CO2 addition. Tail gas recycling reduces formation of soot precursors up to 50% compared to the corresponding fuel feed without coreactants. However, tail-gas recycling shifts the formation of soot precursors toward lower C/O ratios.

1. INTRODUCTION Compact and autothermal reformers for the production of hydrogen and synthesis gas (H2 and CO) from liquid hydrocarbon fuels such as gasoline, diesel, and kerosene are promising on-board devices for electricity supply via fuel cells (auxiliary power units, APU) as well as for reduction of NOx emissions. Conventional fuels are attractive due to their high energy density, widespread production, and distribution and retailing infrastructure.1 Among the three reforming routes, i.e., steam reforming (SR, eq 1), partial oxidation (POX, eq 2), and autothermal reforming (ATR, combination of SR and POX, eq 3), SR is less attractive for mobile applications due to slow start-up and endothermic operation.2 In contrast, catalytic partial oxidation (CPOX) of logistic fuels offers a high throughput and autothermal route for the supply of large amounts of hydrogen at short contact times (10−2 to 10−4 s). Equation 4 shows the catalytic partial oxidation of isooctane, a commonly used model fuel for gasoline. Cx H yOz + (x − z)H2O → xCO + (x − z + (y/2))H2

(1)

Cx H yOz + ((x − z)/2)O2 → xCO + (y/2)H2

(2)

require an additional fuel processing system for CO cleaning and, actually, can even be operated with a certain amount of hydrocarbons in the feed.3−5 A scheme of an APU involving a CPOX reformer is shown in Figure 1. For higher efficiency and

Figure 1. Conception drawing of an APU. The bottom path shows the conventional principle for electricity production with combustion engine and generator and exhaust gas after-treatment. The upper path describes the combination of CPOX reformer and fuel cell stack with additional options for start-up and emission control improvement.

long-term stability of such APUs, a fundamental understanding of the chemical processes is essential, paving the way for commercialization. Even though high fuel conversion and hydrogen selectivity, both >90%, can be achieved using Rh-coated

Cx H yOz + (x /2)O2 + (x − z)H2O → xCO2 + (x − z + (y /2))H2 C8H18 + 4O2 → 8CO + 9H2

(3) (4)

Special Issue: Russo Issue Received: Revised: Accepted: Published:

Aside from proton exchange membrane (PEM) fuel cells, solid oxide fuel cells (SOFC) are also under consideration for mobile application. In contrast to PEMFCs, SOFC do not © 2012 American Chemical Society

7536

August 2, 2011 December 7, 2011 January 12, 2012 January 12, 2012 dx.doi.org/10.1021/ie201712d | Ind. Eng.Chem. Res. 2012, 51, 7536−7546

Industrial & Engineering Chemistry Research

Article

catalysts,6,7 the successful technical realization may depend on the reduction of harmful minor byproducts such as the coke precursors ethylene and propylene. It is generally accepted that the production of synthesis gas via CPOX of hydrocarbons follows the indirect route, i.e., first total oxidation of the hydrocarbon occurs until most of the oxygen is consumed, at least close to the catalytic surface. This exothermic catalytic combustion provides the heat for subsequent endothermic reforming of the remaining fuel in the oxygen-free reaction zone, resulting in the production of syngas.8,9 CPOX of methane as model natural gas feedstock was intensively studied over the group VIII noble metal catalysts, in particular over Rh, for which carbon formation was not observed.9−15 With nickel and manganese, coke formation is observed, and mainly attributed to the endothermic methane cracking (eq 5) and slightly exothermic Boudouard reaction or CO-disproportion (eq 6).16

CH4 ⇌ C + 2H2

(5)

2CO ⇌ C + CO2

(6)

The effects are discussed in comparison with product compositions reached in studies without tail gas recycling17 and with thermodynamic equilibrium data.

2. EXPERIMENT Apparatus. All experiments were carried out in a CPOX reactor, consisting of a 1-cm-long Rh-coated monolith positioned between front and back heat shields, both being 1 cm long. These heat shields (uncoated) also serve as thermocouple fixations. Two different types of thermocouples, both with an outer diameter of 1.5 mm and 500 mm length were used, type K in the front shield and type N in the back shield. The complete assembly was placed in a quartz glass tube surrounded by a furnace for heating and thermal insulation. Product composition was analyzed by online FTIR, H2-MS, and paramagnetic oxygen detection. For detailed information on the experimental setup, refer to our recent publications.6,28 Carbon dioxide was provided as gaseous reactant with mass flow control and a purity of 99.995% (CO content 0.1

Figure 9. Calculated amount of carbon deposition from carbon burnoff experiments for each ReTG/C ratio and isooctane reference. Each value represents the overall carbon amount for C/O 0.72−1.79. Error margins are calculated from the total amount of C detected in the carbon burnoff experiments performed after each measurement and its rerun.

only, CO2 is a net product in all cases. The fuel conversion in case 3 strongly depends on the amount of ReTG. Already for ReTG/C = 0.1, even in the lean area, conversion is not complete and drops precipitously to values below 0.8. For ReTG/C > 0.1, full conversion is never reached in the examined C/O ratios. 4.3. Effect on Formation of Side Products (Methane, Ethylene, Propylene). Side products, formed by thermal cracking of unconverted fuel, have in particular been observed at higher C/O ratios in CPOX of higher hydrocarbons.6 The side products ethylene, propylene, and acetylene are well-known to serve as precursors for the formation of soot particles and solid carbon deposits.22 Feeding ReTG to the reformer inlet stream significantly decreases temperature and consequently fuel conversion (Figure 6). Here, fuel conversion is reduced and shifted toward lower C/O ratios with increasing ReTG/C compared to the reference feed. A larger amount of unconverted fuel rather increases the formation of soot precursors while the lower temperature in general slows down gas-phase reactions. As long as the temperature still exceeds 850 K over the entire C/O regime, thermal cracking of unconverted fuel leads to the appearance of olefins already in the lean area. Figure 5 exhibits

Figure 8. Input referenced molar flow rates for H2O and CO2 input (dashed lines) and output (full lines) of the reactor. First = case 1, second = case 2, last two = case 3.

other ReTG/C ratios more CO2 is detected at the exit than fed to the reactor inlet for all conditions. 7543

dx.doi.org/10.1021/ie201712d | Ind. Eng.Chem. Res. 2012, 51, 7536−7546

Industrial & Engineering Chemistry Research

Article

deposited C. CO2 cofeed leads to less carbon deposition for all ReTG/C ratios and reveals a maximum at medium recycling rates. Steam and ReTG addition (cases 1 and 3) lead to clearly less carbon deposits on the catalyst and values are very similar for all ReTG/C ratios. Even though CO2 cofeed still reduces the total amount of deposited carbon on the surface compared to the reference, significantly more deposits are formed. But these are still much less than without any cofeed. It is remarkable that even very small amounts of steam and/or CO2 cofeed reduce the formation of coke significantly. Due to the network of competing reactions both on the surface and in the gas phase, and their interaction with mass and heat transfer, a purely experimental study analyzing the exit feed only is not able to satisfactorily explain the intrinsic processes leading to the product distribution observed. Also, it is not possible to eventually distinguish between chemical and thermal effects due to their strong coupling in CPOX reactors. The results discussed indicate that the reduction of olefin production is at least to a certain extent related to steam reforming and water−gas shift reactions, because the addition of steam and CO2 has a strong impact on the rate and direction of these reactions. However, for all cases studied, a drop in reactor temperature is observed due to tail gas recycling, but, the reactor temperature is still sufficiently high to thermally crack unconverted isooctane in the gas phase leading to olefins. At this point, investigations of spatially resolved profiles over the reactor, experimentally and numerically, are needed for a full understanding of chemical and thermal effects on the formation of olefins and carbonaceous overlayers.

the yields of the three major side-products methane, ethylene, and propylene. In case of steam addition (case 1), yields of all three species are reduced compared to the reference up to S/C = 0.1. Ethylene and propylene continuously decrease with increasing S/C although conversion of the fuel decreases, too. The increasing amount of steam available in the reformate may lead to additional steam reforming of olefins formed previously. Only methane increases with increasing steam content in the inlet, exceeding the reference for S/C > 0.1. One possible explanation for rising methane yields could be gasification of carbon layers formed on the catalytic surface. Thus, leading to CO and H2 followed by methanation of CO and/or CO2 with H2. Furthermore, direct methanation of surface carbon with H2 could be another possibility. Experimental data from in situ analysis are needed to verify this explanation given by the author. Comparable results are obtained with addition of carbon dioxide (case 2): The ethylene and propylene formation follows the same trends, only the absolute amount of formation of these olefins is higher than in case 1. Because steam reforming in case 2 can only be realized by steam produced in total oxidation of the fuel and dry reforming is rather slow in terms of the residence time, WGS is the path for the CO2 cofeed to interact with CPOX. The addition of both steam and carbon dioxide (case 3) exhibits a very different behavior. The yields of ethylene and propylene are qualitatively different, although a reduction in yield for both olefins is observed with increasing ReTG/C. The maxima however are shifted toward low C/O and very small amounts of olefins are observed at C/O > 0.9. In case of ReTG/C > 0.2, the maxima are shifted to very lean conditions, thus only the right shoulder of the peak is seen. These large amounts of olefins are correlated to the incomplete conversion of the fuel in high ReTG feed even at lower C/O. The different reactivities of oxygen in molecular oxygen, steam, and carbon dioxide in conversion of hydrocarbons seem to lead to a rather complex behavior. As both recycling species are present, chemistry competes among steam reforming, WGS reaction, and of course total and partial oxidation. Whereas the formation of olefins drastically changed, the formation of methane is similar to the first two cases, even though the maximum in methane yield is produced at lower C/O ratios at high ReTG/C ratios. For ReTG/C < 0.1, CH4 yield curves are similar to the ones for steam cofeed. For ReTG/C > 0.1 the methane yield dependence on C/O is similar to the one for the CO2 cofeed. Concerning formation of olefins in fairly adiabatic reactor operation, tail-gas recycling leads to a reduction of these coke precursors in all cases studied for C/O > 0.9. On the other hand, significant amounts of olefins are now seen at low C/O in case of steam and CO2 addition, i.e., in a regime where they are not formed in high-temperature CPOX reactors without tail-gas recycling. 4.4. Effect on Coke Formation. Attention has also been given to the total amount of carbon deposited on the catalyst surface. Carbon burn-offs were conducted after each full sweep of measurement over the C/O range of 0.72 to 1.79. Unfortunately, the detected amounts of carbon deposits represent the entire amount of C on the surface for this measurement cycle and do not show the carbon amount of the defined C/O ratio separately. The detected CO and CO2 signals were quantified to molar flow rates and integrated over time; values are given in mmol of C in Figure 9. The reference experiment (ReTG/C = 0) displays the largest amount of

5. CONCLUSIONS The impact of H2O and/or CO2 addition to a reformer for oxidative conversion (CPOX) of isooctane over Rh/aluminabased honeycomb catalysts to hydrogen-rich synthesis gases was studied for five different amounts of H2O, CO2, and H2O + CO2 added over a broad range of C/O ratios. Feeding these tail-gas components alters fuel conversion, product composition, reactor temperature, and coke formation rates. The formation of olefins differs quantitatively in the case of simultaneous steam and CO2 addition from all other cases. Coking of the catalyst is reduced by tail-gas recycling. The cofeed influences the reactor performance mainly due to steam reforming and the forward and reverse WGS reaction in the fuel lean area with C/O ratios up to 0.9. Reforming and WGS reactions are close to equilibrium as reaction temperature maintains at a high value due to complete fuel conversion. Once unconverted fuel is breaking through the monolithic catalyst, deviations from equilibrium are observed (Figure 7) and the appearance of smaller hydrocarbon species such as ethylene and propylene is noticeable, even at low C/O. Nevertheless, the cofeed of either H2O or CO2 can significantly reduce the amount of produced soot precursors arising from thermal cracking reactions. Tail-gas recycling in the CPOX reformer exhibits only small influence on the production of the major products hydrogen and carbon monoxide. Steam addition leads to a higher net hydrogen production over most of the C/O and S/C ratios studied. Although simultaneous steam and carbon dioxide addition (recycling of the anode exhaust of a SOFC) is not beneficial for hydrogen production, it can reduce the formation of the soot precursors ethylene and propylene at higher C/O ratios. However, an increase in olefin production is observed at low C/O ratios. Integrating over all C/O ratios 7544

dx.doi.org/10.1021/ie201712d | Ind. Eng.Chem. Res. 2012, 51, 7536−7546

Industrial & Engineering Chemistry Research

Article

F. B., Schmal, M., SousaAguiar, E. F., Eds.; Elsevier Science Bv: Amsterdam, 2007; Vol. 167, pp 319−324. (11) Cellier, C.; Clef, D. L.; Mateos-Pedrero, C.; Ruiz, P. Influence of the co-feeding of CO, H2, CO2 or H2O in the partial oxidation of methane over Ni and Rh supported catalysts. Catal. Today 2005, 106, 47−51. (12) Bitsch-Larsen, A.; Horn, R.; Schmidt, L. D. Catalytic partial oxidation of methane on rhodium and platinum: Spatial profiles at elevated pressure. Appl. Catal., A 2008, 348, 165−172. (13) Cimino, S.; Lisi, L.; Russo, G.; Torbati, R. Effect of partial substitution of Rh catalysts with Pt or Pd during the partial oxidation of methane in the presence of sulphur. Catal. Today 2010, 154, 283−292. (14) Cimino, S.; Torbati, R.; Lisi, L.; Russo, G. Sulphur inhibition on the catalytic partial oxidation of methane over Rh-based monolith catalysts. Appl. Catal., A 2009, 360, 43−49. (15) Torbati, R.; Cimino, S.; Lisi, L.; Russo, G. The Effect of Support on Sulphur Tolerance of Rh Based Catalysts for Methane Partial Oxidation. Catal. Lett. 2009, 127, 260−269. (16) DeGroote, A. M.; Froment, G. F. The role of coke formation in catalytic partial oxidation for synthesis gas production. Catal. Today 1997, 37, 309−329. (17) Hartmann, M.; Maier, L.; Minh, H. D.; Deutschmann, O. Catalytic partial oxidation of iso-octane over rhodium catalysts: An experimental, modeling, and simulation study. Combust. Flame 2010, 157, 1771−1782. (18) Dreyer, B. J.; Lee, I. C.; Krummenacher, J. J.; Schmidt, L. D. Autothermal steam reforming of higher hydrocarbons: n-Decane, nhexadecane, and JP-8. Appl. Catal., A 2006, 307, 184−194. (19) Krummenacher, J. J.; West, K. N.; Schmidt, L. D. Catalytic partial oxidation of higher hydrocarbons at millisecond contact times: decane, hexadecane, and diesel fuel. J. Catal. 2003, 215, 332−343. (20) Panuccio, G. J.; Dreyer, B. J.; Schmidt, L. D. A comparison of the catalytic partial oxidation of C-1 to C-16 normal paraffins. AIChE J. 2007, 53, 187−195. (21) Burke, N.; Trimm, D. Coke formation during high pressure catalytic partial oxidation of methane to syngas. React. Kinet. Catal. Lett. 2005, 84, 137−142. (22) Kaltschmitt, T.; Maier, L.; Hartmann, M.; Hauck, C.; Deutschmann, O. Influence of gas-phase reactions on catalytic reforming of isooctane. Proc. Combust. Inst. 2011, 33, 3177−3183. (23) Timmermann, H.; Sawady, W.; Campbell, D.; Weber, A.; Reimert, R.; Ivers-Tiffee, E. Coke formation and degradation in SOFC operation with a model reformate from liquid hydrocarbons. J. Electrochem. Soc. 2008, 155, B356−B359. (24) Lee, J. S.; Quan, N. D.; Hwang, J. M.; Lee, S. D.; Kim, H. G.; Lee, H. J.; Kim, H. S. Polymer Electrolyte Membranes for Fuel Cells. J. Ind. Eng. Chem. 2006, 12, 175−183. (25) Hoogers, G. Fuel Cell Technology Handbook (Book). Sci-Tech News 2003, 57, 51. (26) Cellier, C.; Blangy, B.; Mateos-Pedrero, C.; Ruiz, P. Modification of catalytic performances due to the co-feeding of hydrogen or carbon dioxide in the partial oxidation of methane over a NiO/[gamma]-Al2O3 catalyst. Catal. Today 2006, 112, 112−116. (27) Michael, B. C.; Donazzi, A.; Schmidt, L. D. Effects of H2O and CO2 addition in catalytic partial oxidation of methane on Rh. J. Catal. 2009, 265, 117−129. (28) Hartmann, M.; Lichtenberg, S.; Hebben, N.; Zhang, D.; Deutschmann, O. Experimental investigation of catalytic partial oxidation of model fuels under defined constraints. Chem. Ing. Tech. 2009, 81, 909−919. (29) Freund, A.; Friedrich, G.; Merten, C.; Eigenberger, G. Pulsationpoor laboratory vaporiser for small liquid streams. Chem. Ing. Tech. 2006, 78, 577−580. (30) Deutschmann, O., Tischer, S., Kleditzsch, S., Janardhanan, V. M., Correa, C., Chatterjee, D., Mladenov, N., Minh, H. D. DETCHEMTM software package. version 2.2 ed.; Karlsruhe, 2008. (31) Maier, L.; Hartmann, M.; Tischer, S.; Deutschmann, O. Interaction of heterogeneous and homogeneous kinetics with mass

studied, the total amount of coke deposited is reduced with any addition of steam and/or CO2. The effects of steam addition and the effects of CO2 addition do not allow extrapolation to the effects of both steam and CO2 addition on the reactor behavior. The results reveal a complexity of interaction between surface and gas-phase kinetics as well as mass and heat transport that is difficult to understand by integral data. Spatially resolved species and temperature profiles are needed to eventually elucidate these interactions and explain the impact of chemical and thermal effects on product distribution of the reformer.

■

ASSOCIATED CONTENT

S Supporting Information *

Experimental data discussed in the text for the three cases but not shown. This information is available free of charge via the Internet at http://pubs.acs.org/.

■

AUTHOR INFORMATION

Corresponding Author

*Fax: 0049-721-608-44805. E-mail: [email protected].

■

ACKNOWLEDGMENTS We thank Umicore AG & Co. KG for providing the model catalyst used in this work. We also gratefully acknowledge the financial support from the German Research Foundation (DFG) and the Helmholtz Association through the Helmholtz Research School Energy-Related Catalysis.

■

REFERENCES

(1) Thomas, C. E.; James, B. D.; Lomax, F. D.; Kuhn, I. F. Fuel options for the fuel cell vehicle: hydrogen, methanol or gasoline? Int. J. Hydrogen Energy 2000, 25, 551−567. (2) Ahmed, S.; Krumpelt, M. Hydrogen from hydrocarbon fuels for fuel cells. Int. J. Hydrogen Energy 2001, 26, 291−301. (3) Lawrence, J.; Boltze, M. Auxiliary power unit based on a solid oxide fuel cell and fuelled with diesel. J. Power Sources 2006, 154, 479−488. (4) Janardhanan, V. M.; Deutschmann, O. CFD analysis of a solid oxide fuel cell with internal reforming: Coupled interactions of transport, heterogeneous catalysis and electrochemical processes. J. Power Sources 2006, 162, 1192−1202. (5) Janardhanan, V. M.; Heuveline, V.; Deutschmann, O. Performance analysis of a SOFC under direct internal reforming conditions. J. Power Sources 2007, 172, 296−307. (6) Hartmann, M.; Kaltschmitt, T.; Deutschmann, O. Catalytic partial oxidation of higher hydrocarbon fuel components on Rh/Al2O3 coated honeycomb monoliths. Catal. Today 2009, 147, S204−S209. (7) Hickman, D. A.; Schmidt, L. D. Synthesis Gas-Formation by Direct Oxidation of Methane over Pt Monoliths. J. Catal. 1992, 138, 267−282. (8) Schwiedernoch, R.; Tischer, S.; Correa, C.; Deutschmann, O. Experimental and numerical study on the transient behavior of partial oxidation of methane in a catalytic, monolith. Chem. Eng. Sci. 2003, 58, 633−642. (9) Horn, R.; Williams, K. A.; Degenstein, N. J.; Bitsch-Larsen, A.; Dalle Nogare, D.; Tupy, S. A.; Schmidt, L. D. Methane catalytic partial oxidation on autothermal Rh and Pt foam catalysts: Oxidation and reforming zones, transport effects, and approach to thermodynamic equilibrium. J. Catal. 2007, 249, 380−393. (10) Tavazzi, I.; Beretta, A.; Groppi, G.; Donazzi, A.; Maestri, M.; Tronconi, E.; Forzatti, P. Catalytic partial oxidation of CH4 and C3H8: experimental and modeling study of the dynamic and steady state behavior of a pilot-scale reformer. In Natural Gas Conversion VIII, Proceedings of the 8th Natural Gas Conversion Symposium; Noronha, 7545

dx.doi.org/10.1021/ie201712d | Ind. Eng.Chem. Res. 2012, 51, 7536−7546

Industrial & Engineering Chemistry Research

Article

and heat transfer in catalytic reforming of logistic fuels. Combust. Flame 2011, 158, 796−808. (32) Tavazzi, I.; Maestri, M.; Beretta, A.; Groppi, G.; Tronconi, E.; Forzatti, P. Steady-state and transient analysis of a CH4-catalytic partial oxidation reformer. AIChE J. 2006, 52, 3234−3245. (33) Hartmann, M.; Maier, L.; Deutschmann, O. Hydrogen production by catalytic partial oxidation of iso-octane at varying flow rate and fuel/oxygen ratio: From detailed kinetics to reactor behavior. Appl. Catal., A 2011, 391, 144−152. (34) Donazzi, A.; Maestri, M.; Beretta, A.; Groppi, G.; Tronconi, E.; Forzatti, P. Microkinetic analysis of CH4 CPO tests with CO2-diluted feed streams. Appl. Catal., A 2011, 391, 350−359.

7546

dx.doi.org/10.1021/ie201712d | Ind. Eng.Chem. Res. 2012, 51, 7536−7546