1. INTRODUCTION 2. KINETIC PARAMETERS OF THE CATALYTIC REACTIONS

Transcription

1 CHEMICAL KINETICS OF HIGH TEMPERATURE HYDROCARBONS REFORMING USING A SOLAR REACTOR Jacob Yeheskel, Rachamim Rubin, Alexander Berman, and Jacob Karni Solar Research Facilities Unit, Weizmann Institute of Science, Rehovot 76100, Israel Tel , Fax rachamim.rubin@weizmann.ac.il Abstract - This study is aimed at developing a model for solar volumetric reactor for hydrocarbon reforming, operating at high temperature and pressure based on two achievements: 1. The development of the DIAPR, a volumetric receiver tested at 50000,000 suns, gas outlet temperature up to 1200 o C and a pressure of 20 atm. 2. A laboratory scale chemical kinetics study of hydrocarbons reforming reactions with Ru base catalyst. The receiver operation is simulated using PHOENICS package. The reaction kinetics is computed, using the CHEMKIN package. A chemical kinetic simulation of a CHO system based on three elements: Definition of a computational cell in the reactor by the relevant parameters: temperature, pressure, reactant compositions, residence time, and catalyst load. Using of laboratory results at K and 4 atm. to find kinetic parameters for the overall reforming of and for the Reverse Water Shift Reaction. Use of CHEMKIN with kinetic parameters found and with unit cell of the DIAPR derived from the CFD model to compute the size of the DIAPR receiver for reforming of. Results show that with achievable catalyst load in the cell and at 1600 K and 20 atm only 10 cm of cell matrix would be necessary to achieve 90% methane. This work provides an essential step in simulating the performance of high temperature and pressure volumetric solar reformer. More computational and experimental work is needed for more accurate model. 1. INTRODUCTION Solar reforming of methane and the complete Thermochemical Heat Pipe cycle ( Close-Loop reforming methanation) of the reaction CH = 2H 2 +2 have been studied at the Weizmann Institute over the last two decades. (Rosin and Levy, 1989) (Rosin and Levitan, 1989). Solar reforming at a scale of several hundred kilowatts was conducted at the Weizmann Institute using (in succession) a DLR-built volumetric reformer, and a Weizmann-built tubular reformer. Another joint project which includes a reforming high pressure reactor of DLR-Stuttgart - Germany, operating in conjunction with an electric gas turbine is presently running. Studies on developing volumetric reforming reactor receivers has been run also in Sandia National Laboratories (Skocypec and Buck, 1994.) (Hogan and Muir, 1994) and in the Boreskov Institute of Catalysis in Novosibirsk, Russia (Funken and Kuzin, 1998). In all of the developed systems, in the above mentioned sources, either the reforming temperature, or its pressure, or both, were below the optimum conditions, due to the limited capabilities of the solar receiver/reactor. For example, the maximum operating pressure and temperature of the WIS tubular reformer are 105 bars and C, respectively; the DLR volumetric reactor is limited to 3 bars and C, and the newly developed one is designed to higher pressure but similar temperature limits. Consequently, industrial solar reforming based on these devices would be neither as efficient, nor as cost-effective as non-solar, standard, industrial reformers. Feasibility studies and cost analyses (Yogev and Shemer 1996) indicate that the combination of Tower Reflector optics and the technology developed by WIS and Rotem for the DIAPR receiver can be the foundation for the development of an advanced, cost-effective solar receiver/reformer. Such a reformer will be able to operate at high sunlight concentration (1000 C 5000), temperature (800 T 1050 C) and pressure (10 p 25 bar). The R&D studies on the DIAPR receiver proved that even without chemical reactions modeling of the performance of the receiver plays a crucial role in the parametric research. In this case the Computational Fluid Dynamic (CFD) model based on PHOENICS program package provided guidelines in different stages of the experimental work as well as for the overall system optimization, to increase energy efficiencies and to assure safe standing of materials in the high temperature and high pressure environment. The scope of this work is: To gather known catalytic kinetic parameters of the CHO system from results reported in the literature, reaction kinetic databases and results of local studies carried on in the WIS catalytic laboratories. To implement the CHEMKIN II code package in analyses of reforming and side reactions of the CHO system. To link the kinetics model based on CHEMKIN II to the existing CFD model in order to have a complete simulation of the high temperature and high-pressure receiver-reformer reactor. 2. KINETIC PARAMETERS OF THE CATALYTIC REACTIONS Integrating the CFD model with chemical process using CHEMKIN II, we need to define all the elementary reactions involve and the kinetics parameter for each reaction. The DIAPR CFD model without chemical reactions is very complicate and not finishes yet. Even at this stage the run time of calculation is very long. Adding a chemical process to this model is reasonable only if we reduce the system to a very few reactions. We decided to base our model on the assumption that the catalyst is very selective and we can take into account only the two main reactions: 1. CH = 2 + 2H H 2 = + H 2 O To find the kinetic parameters for those two reactions we used the experimental data published by Berman et al (Epstein and Karni, 1996). In this work they measured the rate of the over all reactions and not the hundreds elementary reactions involve. So

2 it meets our purpose. The published results are given in l/h*g units. In order to obtain the kinetic parameters from the laboratory results we have to define a unit cell as in the CFD model. The cell composed of a central pin and four quarters of edge pins, sums to 2 pins active areas. The length of the cell is 1.56 mm the width is 0.9 mm the height is 6 cm. using the typical Ru load of 1 mgr./cm in 1-% Ru catalyst results in an overall cell load of 1.2-gr. catalyst. 2.1 Reforming Reaction Arrhenius plots based on the measured rates for some cases of the reformerming with reaction are shown in figure 1. Each case is for diferent type of catalyst. The cases 5A anb 5B are without catalyt without Ru /T [K] with Ru T30 T13 T20 T21 T24 T5A T5B Linear (T13) Figure 1: Arrhenius plots of measured rates for few catalyst cases, non-catalyst and trend line of case T13. It can be seen that within the estimate of experimental errors the Arrhenius model presents quite well the dependence of the measured rates on temperatures. Each case has different kinetic values. We chose to base our model on case named that we T13. The other cases gave lines of higher rate values but they are curved in the region of interest, the high temperatures region and meet the rate values of case T13. We assume that in those cases the system reach to equilibrium before the reactor end. In table 1 the Arrhenius kinetic parameters are noted, based on a trend line marked in figure 1 of case T13. The 2 nd row set in the table is related to non-catalyst experiments. Catalyst T Range K A[l/h] Ea [cal/mole] Ek* [K] Ru (T13) Non-cat E * R=A*exp(-Ea/RT) => R=A*exp(-Ek/T) Table 1: Experimental Arrhenius parameters. The reaction rate in table 1 and the unit cell define lead to the kinetics values using simple calculation. By basic units : r = * exp (2104/RT) [ mol/s ] Eq. (2) And considering the cell volume: V = 0.9*1.56*6 =8.424 cm^3 Eq. (3) r = r /V = 186 * exp (2104/RT) [ mol/cm^3 s] Eq. (4) Based on the assumption that the trend Arrhenius line drawn in figure 1 represents experimental results compatible to 4 atm in about feed reactant composition (/ = 1.3) and assuming a second order unidirectional reaction, we can calculate the rate constant coefficient from: r = k [] [] Eq. (5) When: [] = (P/RT) = (4/82.06 T) = 0.023/T [mol/cm^3] Eq. (5) [] = (P/RT) = (4/82.06 T) = 0.030/T [mol/cm^3] Eq. (6) And the rate constant expression, compatible with CHEMKIN II program is: k [cm^3/mol s] = 2.69 *T^2 * exp (2104 /RT) Eq. (7) 2.2 Reverse Water Shift Reaction (RWSR) Kinetic Data for the RWSR in gas phase are not reported. In the NIST Database a sole reference is found (Graven and Lang, 1954) for the equilibrium reaction: 2 + H 2 = + H 2 O However, in reference to this source it is stated that data too complex to abstract. Laboratory experiments reported by Berman et al. (1996) are the source for kinetic parameters (Epstein and Karni, 1996) in which produced H 2 O was measured besides the reforming reaction products. Figure 2 shows some Arrhenius plots based on measured rates for some cases of the RWSR /T [K] Figure 2: Arrhenius plots of measured rates of the RWSR. T30 T13 T20 T21 T24 T5A T5B We use the same mathematics as for the reforming and we get the Arrhenius plots in figure 3 for the case T13. r = - d[]/dt = (1.2*1050) * exp (2104/RT) [l(ntp)/h] Eq. (1)

3 0 The longitudinal velocity, as found by the CFD model, is 1.5 m/s and the cell displacement along the velocity vector is 1.56 cm, these two values gave a residence time of 0.01 sec. Figure 4 indicates that the implementation of the catalytic pins loaded by 6 mgr 1%-Ru Per pin in the present DIAPR design would be impractical; for a of 90%, 3.5 seconds are needed which means a reactor about 5.5 meters long. y = x R 2 = Figure 3: Arrhenius plot of the RWSR case T13. This figure shows that up to 700 K the activation energy is almost 0 and: 1/T k = A = T 2 [cm^3/mol s] Eq. (8) Above this temperature the rate is decrease and we get: k = AT 2 exp(-ea/rt) = 1.01E-4T 2 exp(5350/t) [cm^3/mol s] Eq. (9) We assume that the decrease in the RWSR rate is the result of changes in the catalyst selectivity. This mean the selectivity, in favor of the reforming reaction, is increased as the temperature increases. 2.3 Additional Reaction Databases The CHO system can include many hundreds of gas phase reactions to be included in the reaction file of CHEMKIN II. It is already concluded by many researchers that for catalytic surface reaction system the gas phase reactions seem insignificant in comparison to surface reactions (Deutschman Olaf and Smidt Lanny D, 1998), though they very much complicate the calculations. A surface reaction database has been recently used in a kinetic model for CHO system developed in the University of Minnesota Supercomputing Institute (UMSI) (Deutschman Olaf and Smidt Lanny D, 1998) (Deutschman Olaf and Smidt Lanny D, 1998) (Hickman D A and Schmidt L D, 1993). This useful database is not suit to our purpose cause it not include data on Ru as a catalyst. 3. KINETIC MPUTATION BY CHEMKIN II Using the kinetics values calculated from the laboratory results for the two reactions reforming and RWSR (Eq 7 and Eq 9) we implement the CHEMKIN II code, with constant P and T to modeling the DIAPR volumetric receiver reformer. Presently, we assume that the thermal design controls a constant temperature within a series of consequent unit cells in the flow direction. By a Lagrangian approach, computing the in one cell, using CHEMKIN II while each time segment corresponds to the residence time in a sole cell can simulate the reaction developed in the cell series. In our calculation we focus on working temperature of 1600K, at higher temperature the catalyst may invalidate. Figure 4: Mole fraction and vs. time for 6mgr. 1%-Ru catalyst load at 4 atm. The may be increased by three ways according to the present model: Decreasing the flow rates. Increasing the catalyst load on the pins. Increasing the gas pressure. Decreasing flow rates damages the reactor throughout (e.g. the windows cooling won t be sufficient). Increasing catalyst load decreases the needed cells number. Figure 5 shows that a tenfold increase in the catalyst load ( increasing only k of the reforming by 10 times) lasts in 0.3 second reaction time for about 90% ; still a 47 cm long reaction zone is needed. Figure 5: Mole fraction and vs. time for 60mgr. 1%-Ru catalyst load at 4 atm. Figure 6 indicates that increasing the reactor pressure from 4 to 20 atm can also be considered an efficient way to increase the convestion

4 Figure 6: Mole fraction and vs. time for 6mgr. 1%-Ru catalyst load at 20 atm. Figure 7 shows that reasonable increase in both the pressure to 20 atm and the catalyst load 10 times leads to only 6 cells needed for a of 90% (0.06 sec) which means about 10 cm long of the reaction zone. Figure 7: Mole fraction and vs. time for 60mgr. 1%-Ru catalyst load at 20 atm. These results indicate that it is possible to use the DIAPR receiver as a reformer. 4. NCLUSIONS The calculations done so far are of a very preliminary manner. Nevertheless we can already draw some conclusions, to the pursuance of the study. The DIAPR volumetric receiver can be used as a reformer with high load of catalyst and high pressure. It is most important to conduct laboratory scale experiments, to increase the present active catalyst load on the pins joint by an increase of surface area and to study the effect of the pressure on the overall reaction kinetics. To gradually expand the reaction to surface, gas or combined detailed elementary reactions. To consider the choice of implementing an overall reaction vis-a-vis detailed elementary reactions. To add desired and undesired side reactions like the carbon formation reactions and to back there kinetic parameters by laboratory scale experiments convesion To link, step by step the elementary cell concept to the CFD program and to feed the resulted kinetic affecting parameters back to CHEMKIN. To design the reaction area by an optimization of the available specific energy density on the surface with the area needed for chemical reaction completion. REFERENCES: Anikeev V.I., A.S. Bobrin, J. Ortner, S. Schmidt, K.H. Funken and N.A. Kuzin (1998) Catalytic Thermochemical Reactor/Receiver for Solar Reforming of Natural Gas: Design and Performance, Solar Energy, 63, pp Page in proceeding: Berman A., Levitan R., Epstein M. and Leve M. (1996) Ruthenium Methanation and Reforming Catalyst for Solar Chemical Pipe. In the 1996 International Solar Energy, Mar. 31 Apr. 3 in San Antonio, Texas, J. H. Davidson (Ed.), pp 61-69, ASME. Berman A., Epstein M. and Karni J. (July 1996) Development of Kippod Absorber for Solar Reformer of Light Hydrocarbons. Final Rpt.2, SRFU-WIS RD9-96 Report. Deutschman Olaf and Smidt Lanny D, (June 1998) Two- Dimensional Modeling of Partial Oxidation of Methane on Rhodium in a Short Contact Time Reactor, Univ. of Minnesota Supercomputing Inst. Research Rpt., UMSI98/119. Deutschman Olaf and Smidt Lanny D, (June 1998) Partial Oxidation of Methane in a Short Contact Time Reactor: Two- Dimensional Modeling With Detailed Chemistry, Univ. of Minnesota Supercomputing Inst. Research Rpt., UMSI98/120. Feasibility Study: Epstein M, Hodara I, Segal A, Yogev A and Shemer Z, (January 1996) Potential Industrial Applications of the Solar Tower Technology for Use at the Dead Sea Works (DSW), a Feasibility Study conducted together with NEPRO Engineering. Israel Ministry of Energy and Infrastructure Publication No. RD Graven WM and Lang FJ, (1954.) Kinetics and mechanisms of the two opposing reactions of the equilibrium + = +. J. Am. Chem. Soc. 76, pp , Hickman D A and Schmidt L D, (1993) Steps in Oxidation on Pt and Rh Surfaces: High-Temperature Reactor Simulation. AIChE Journal, 39 (17), pp James F. Muir, Roy E. Hogan, Jr., Russell D. Skocypec and Reiner Buck (1994) Solar Reforming of Methane in a Direct Absorption Catalytic Reactor on a Parabolic Dish: I- Test and Analysis, Solar Energy, 52, pp ,.

5 Levitan R., Rosin H. and Levy M. (1989) Chemical Reactions in Solar Furnace. Direct Heating of the Reactor in a Tubular Receiver. Solar Energy 42, pp Levy M., Rosin H. and Levitan R. (1989) Chemical Reactions in Solar Furnace by Direct Solar Irradiation of the catalyst. J. of Solar Energy Engineering, 111 pp Russel D, Skocypec, Roy E. Hogan, Jr. and James F. Muir (1994) Solar Reforming of Methane in a Direct Absorption Catalytic Reactor on a Parabolic Dish: II- Modeling and Analysis, Solar Energy, 52, pp