Low-Pressure Sorption-Enhanced Hydrogen Production - Industrial

File failed to load: https://cdn.mathjax.org/mathjax/contrib/a11y/accessibility-menu.js ..... Optimal slow pyrolysis of apple pomace reaction conditio...
0 downloads 0 Views 106KB Size
Ind. Eng. Chem. Res. 2005, 44, 1665-1669

1665

KINETICS, CATALYSIS, AND REACTION ENGINEERING Low-Pressure Sorption-Enhanced Hydrogen Production Kwang Bok Yi and Douglas P. Harrison* Gordon A. and Mary Cain Department of Chemical Engineering, Louisiana State University, Baton Rouge, Louisiana 70803

The sorption-enhanced production of hydrogen involves the simultaneous steam-methane reforming, water-gas shift, and CO2 removal reactions in the presence of a reforming catalyst and CO2 sorbent. High concentrations of H2 with low concentrations of CO are produced in a single reaction step. Pure CO2 suitable for use or sequestration may be produced during sorbent regeneration. This study examined the sorption-enhanced H2 production reactions at low pressure, which also requires a low reaction temperature so that the sorbent will retain its effectiveness for CO2 capture. H2 concentrations in excess of 96 mol % (dry basis) and CO concentrations as low as 7 ppmv (dry basis) were produced in tests between 1 and 5 bar and 400 and 460 °C using a H2O-to-CH4 feed ratio of 3. Calcined Arctic SHB dolomite proved to be an effective CO2 sorbent that did not require pretreatment for sulfur removal. Introduction

reforming:

Hydrogen is a pollution-free energy carrier that in the future may be used for electric power generation, as a transportation fuel, in industrial boilers, and even for commercial and residential heating. At the present time, the majority of hydrogen is produced in the multistep steam-methane reforming process. Synthesis gas is formed by the endothermic steam-methane reforming reaction in the first step of the process. Large quantities of supplemental energy are required at the typical reaction conditions of 850 °C and 25 bar. One or two stages of shift reaction are then used to increase the hydrogen yield. Although the shift reaction is exothermic, the reaction temperature is lower so that the energy liberated cannot be used in the reformer. Shift reaction is followed by hydrogen purification using amine scrubbing or pressure swing adsorption. Amine scrubbing for CO2 removal is often followed by methanation to reduce CO to low-ppmv levels needed to avoid catalyst poisoning when the H2 is to be used in downstream catalytic processes.1 Pressure swing adsorption is capable of producing high-purity H2 without the need for further purification, but significant fractions of H2 product are lost in the process.2 Preferential oxidation3 is being studied for CO conversion to CO2 when the downstream application does not require low CO2 concentrations. Sorption-Enhanced Hydrogen Production. Sorption-enhanced hydrogen production accomplishes reforming, shift, and purification in a single processing step. The reactions occur simultaneously in the presence of a nickel-based reforming catalyst and a CO2 sorbent, in this case calcined dolomite. The simultaneous and overall reactions are * To whom correspondence should be addressed. E-mail: [email protected].

shift:

CH4(g) + H2O(g) T CO(g) + 3H2(g) CO(g) + H2O(g) T CO2(g) + H2(g)

CO2 removal: CaO‚MgO(s) + CO2(g) T CaCO3‚MgO(s) overall:

CH4(g) + 2H2O(g) + CaO‚MgO(s) T CaCO3‚MgO(s) + 4H2(g)

Removal of CO2 from the gas phase as it is formed shifts the normal equilibrium limits of the reforming and shift reactions and, with proper choice of reaction conditions, permits high CH4 conversion with almost complete removal of CO and CO2. In addition to offering the potential for a simpler process, the combined reactions are almost thermally neutral so that no supplemental energy is required for the H2 production phase. The expensive and troublesome reforming tubes currently used may be replaced by a more common fixed- or fluidized-bed reactor. A further advantage of sorption-enhanced reaction is that no shift catalyst is required.4 Supplemental energy required to regenerate the sorbent to permit multicycle operation is estimated to be about 20% less than the supplemental energy now required in the reforming reactor.5,6 The earliest reference to sorption-enhanced H2 production was published in 1868.7 Williams8 and Gorin and Retallick9 were issued patents for fixed-bed and fluidized-bed processes involving sorption-enhanced reaction in 1933 and 1963, respectively. The concept was not developed further because of the low cost of energy and the difficulty in handling large quantities of solid sorbent. This has changed recently because of higher energy costs and because of the growing concern over

10.1021/ie048883g CCC: $30.25 © 2005 American Chemical Society Published on Web 02/19/2005

1666

Ind. Eng. Chem. Res., Vol. 44, No. 6, 2005

Figure 1. Equilibrium hydrogen and carbon monoxide concentrations as a function of temperature and pressure.

global warming resulting from greenhouse gas emissions, primarily CO2. Pure CO2 suitable for use or compression and sequestration may be produced during the sorbent regeneration step. Sorption-enhanced hydrogen production research in this laboratory began with a study of combined shift reaction and CO2 removal using simulated synthesis gas feed.4 Balasubramanian et al.10 followed with a study of methane conversion to hydrogen using a mixture of nickel reforming catalyst and high purity CaCO3 as the sorbent precursor. 95+% H2 (dry basis) was produced at 650 °C and 15 bar using a steam-to-methane feed ratio of 4. Lopez Ortiz and Harrison6 showed that commercial dolomite could be used as the sorbent precursor if pretreated for sulfur removal and that only moderate activity loss occurred when the catalyst and sorbent were subjected to 25 reaction-regeneration cycles. More recently Peng and Harrison,11 also using a steam-to-methane ratio of 4, showed that fuel cell quality H2 (97+%) containing less than 20 ppmv CO (both dry basis) could be produced at 5 bar pressure and a temperature of 480 °C. Even though the temperature and pressure required to achieve this CO concentration were relatively low, the composition of the reactor product gas closely approached combined reaction equilibrium. Studies on sorption-enhanced hydrogen production using potassium-promoted hydrotalcite sorbent have been reported by Allam et al.12 and by Ding and Alpay.13

Figure 2. Fixed-bed reactor system.

Kato14 and colleagues are studying the use of lithiumbased sorbents. Variations of the sorption-enhanced hydrogen concept using calcium-based sorbents are included in the Zero Emission Coal process under development at Los Alamos National Laboratory,15 the HyPr-Ring process in Japan,16 and the Zero Emission Gas Project in Norway.17 Recent analysis by Abanades et al.18 has suggested that naturally occurring calcium materials could be economically competitive with amine scrubbing for the capture of CO2 from flue gas because of the low intrinsic cost, about $0.05/kg of limestone. This study on low-pressure sorption-enhanced hydrogen production was undertaken in conjunction with the Institute for Energy Technology (IFE) in Norway in support of their Zero Emission Gas Project. Hydrogen produced by sorption-enhanced reaction is to be used to power atmospheric pressure PEM fuel cells and the high-temperature waste heat from solid oxide fuel cells (SOFC) is to be used for sorbent regeneration. The current study had two primary objectives: (1) to evaluate the performance of the sorption-enhanced reaction concept at or near 1 bar pressure and (2) to determine if the Arctic SHB dolomite supplied by Franzefoss AS through IFE could be used as the sorbent precursor without pretreatment for sulfur removal. Thermodynamic Analysis. Operation near atmospheric pressure requires that low reaction temperatures be used if the sorbent is to retain its effectiveness for CO2 capture. Results of a thermodynamic analysis are shown in Figure 1, where equilibrium mol fraction H2 and ppmv CO (both dry basis) are plotted as a function of temperature and pressure at a steam-tomethane feed ratio of 3. Equilibrium hydrogen content, which is almost independent of temperature and is favored by low pressure, approaches 98% and 96% at 1 and 3 bars, respectively. In contrast, the equilibrium CO content increases with increasing temperature and decreasing pressure. A product containing less than 50 ppmv CO should be possible to about 480 °C at 1 bar and to just over 500 °C at 3 bar if equilibrium can be closely approached. Reactor, Materials, and Reaction Conditions. The laboratory-scale fixed-bed reactor shown in Figure 2 was used for reaction tests. CH4 and N2 were obtained

Ind. Eng. Chem. Res., Vol. 44, No. 6, 2005 1667 Table 1. Reaction Conditions, Catalyst, and Dolomite Properties reforming catalyst (Haldor Topsøe A/S, R-67-7H) 15-20 wt % NiO 20-25 wt % MgO 55-60 wt % Al2O3 crushed and sieved to 75-150 µm dolomite (Arctic dolomite SHB from Franzefoss AS through IFE) 32.0 wt % CaO 20.6 wt % MgO 46.3 wt % loss on ignition crushed and sieved to 150-300 µm reactor packed bed dimensions solid loading catalyst sorbent (precalcined) sorbent-to-catalyst ratio

1.9 cm ID × 7.6 cm max. length (0.75 in x 3 in) 1.2-3.6 g 18-24 g 5-20

reaction conditions temperature pressure feed gas composition steam-to-carbon ratio (S/C) volumetric feed rate weight hourly space velocity (WHSV) reactor space time

400-460 °C 1-5 bar 25% CH4 and 75% H2O 3 200-600 cm3(stp)/min 6600-30 000 cm3(stp)/hr/g cat 1.1-9.3 s

from high-pressure cylinders with flow rates controlled using mass flow controllers. H2O was fed using a syringe pump, and feed lines were heat traced to ensure complete vaporization. During a reaction test, the feed gas contained only CH4 and H2O; N2 was used to purge the system before and after each test. The combined feed gas entered near the bottom of the reactor and was preheated as it flowed upward in the annular area around the reactor insert. The preheated feed gas then flowed downward through a mixture of reforming catalyst and CO2 sorbent and exited from the bottom. Excess H2O was condensed, and the product gas was analyzed by gas chromatography. The mixture of powdered catalyst and calcined dolomite was supported inside the reactor insert by a layer of quartz wool on top of a porous stainless steel disk. The source and composition of the catalyst and dolomite, the particle size ranges used, and the range of reaction conditions studied are summarized in Table 1. The catalyst from Haldor-Topsøe is a standard Nibased reforming catalyst, and the Arctic SHB dolomite supplied by Franzefoss AS through IFE contained no detectable sulfur. The dolomite was precalcined in N2 at 850 °C for 2 h in a separate furnace. Different particle sizes of catalyst and dolomite were used to facilitate separation at the conclusion of the test. A new sample of dolomite was used in each test. Catalyst was sometimes reused after being separated from the dolomite without apparent ill effects. Activation of the catalyst occurred during the initial stages of each test. Although most tests were conducted at 1 bar, limited tests were conducted at pressures to 5 bar. Pressure drop through the packed bed was negligibly small. The range of reaction temperatures, from 400 to 460 °C was lower than previous sorption-enhanced reaction tests using a calcium-based sorbent. The feed gas composition was fixed at 25% CH4 and 75% H2O for a steam-tocarbon (S/C) ratio of 3 in all tests. The maximum solid capacity of the reactor was about 28 g.

A commercial system should operate with the minimum ratio of catalyst to sorbent consistent with the product gas composition closely approaching equilibrium at reaction conditions. When sufficient catalyst is present to produce near equilibrium product composition, the quantity of sorbent affects only the duration of the fixed-bed test. The minimum quantity of catalyst tested, 1.2 g, coupled with the largest volumetric feed rate of 600 cm3(stp)/min resulted in a maximum weight hourly space velocity (WHSV) of 30 000 cm3(stp)/hr/g cat, corresponding to a reactor space time of about 1.1 s. Experimental Results A reaction system of this type is inherently unsteady state because of the fixed amount of sorbent present in the reactor. Following a brief unsteady-state period associated with flow rates reaching steady state and catalyst becoming fully activated, there is a period in which the product composition is almost constant. During this prebreakthrough period, the reforming, shift, and carbonation reactions are all active, and the H2 concentration is at its maximum value while minimum concentrations of CH4, CO2, and CO are produced. Prebreakthrough is followed by a period of active breakthrough that begins when the leading edge of the carbonation reaction front reaches the exit of the fixed bed of solid. CO2 removal becomes less effective, which limits CO conversion in the shift reaction and, in turn, reduces CH4 conversion in the reforming reaction. H2 concentration decreases while the concentrations of the other components increase with time. A steady-state postbreakthrough period is then reached when the carbonation rate approaches zero. Only the reforming and shift reactions are active, and the product composition is determined by the kinetics and thermodynamics of these reactions. The prebreakthrough period corresponding to maximum H2 and minimum concentrations of other components is of primary interest and is emphasized in the following. Typical reactor response through the beginning of the active breakthrough period is illustrated in Figure 3, parts a and b. Experimental concentrations of H2 and CH4 (mol %, dry basis) are shown in Figure 3a, whereas concentrations of CO and CO2 (ppmv, dry basis) are shown in Figure 3b. Reaction conditions, as well as prebreakthrough equilibrium concentrations of each component are shown on the figure. The experimental prebreakthrough concentrations of 92.3% H2 and 7.7% CH4 (Figure 3a) and 50 ppmv CO and 91 ppmv CO2 (Figure 3b) represent the average of the first three samples and correspond to about 87% conversion of CH4. Percent sorbent carbonation, as estimated by material balance, was approximately 90% after 50 min. Note that the CO2 concentration at 60 min exceeded 100 000 ppmv (>10%) indicating that little, if any, further carbonation was occurring at that time. Active breakthrough appeared to begin at about 50 min from Figure 3a. However, the more sensitive ppmv measurements in Figure 3b indicated that active breakthrough began somewhat earlier, at about 40 min. The clear existence of the steady-state prebreakthrough period coupled with the ability to reuse catalyst in multiple tests proved that the Arctic SHB dolomite can be used without pretreatment for sulfur removal. In earlier test results6 using a sulfur-containing dolomite without pretreatment, the activity of the reforming

1668

Ind. Eng. Chem. Res., Vol. 44, No. 6, 2005

Figure 5. Effect of temperature on prebreakthrough concentrations of hydrogen and carbon monoxide.

Figure 3. Prebreakthrough curves for (a) hydrogen and methane, and (b) carbon monoxide and carbon dioxide.

Figure 4. Effect of pressure on prebreakthrough concentrations of hydrogen and carbon monoxide.

catalyst quickly declined to the point that little CH4 conversion was achieved. Pressure is an important operating parameter as indicated by the thermodynamic analysis and confirmed by the experimental results shown in Figure 4. Prebreakthrough H2 and CO concentrations are plotted as a function of pressure at a reaction temperature of 440 °C. Other reaction conditions are shown on the figure. Discrete points indicate experimental results, whereas the solid lines represent equilibrium concentrations. The experimental CO concentration decreased from 30 ppmv at 1 bar to 11 ppmv at 3 bar and to 7 ppmv at 5 bar, whereas equilibrium values decreased from 7 to 1 ppmv over the same pressure range. The experimental H2 concentration decreased from 94.2% at 1 bar to 89.8% at 5 bar, and both values were reasonably close to the equilibrium values of 97.8% and 94.5%. CH4 concentration (not shown in the figure) increased from 5.8% at 1 bar to 10.2% at 5 bar, whereas CO2 concentrations were in the range of 5-20 ppmv. The effect of temperature on prebreakthrough H2 and CO concentrations at 1 bar is shown in Figure 5. Once

again, the solid lines represent equilibrium concentrations, whereas the discrete points designate experimental results. Equilibrium H2 varied only between 97.7% and 97.8% at these conditions. At 460 °C, the experimental value of 96.4% is quite close to equilibrium. Experimental H2 concentration, however, decreased with temperature to a minimum of 87.3% at 400 °C. Equilibrium CO concentration increased with temperature from 1 ppmv at 400 °C to 17 ppmv at 460 °C, whereas the experimental values increased from 12 ppmv to 49 ppmv over the temperature range. CH4 concentration decreased from 12.7% at 400 °C to 3.6% at 460 °C, corresponding to CH4 conversions of 65% and 87%, respectively. Experimental CO2 concentrations were in the range of 15-120 ppmv. Satisfactory reproducibility is shown from the results of duplicate tests conducted at 400 °C. Prebreakthrough CO concentrations were 12 and 15 ppmv, whereas H2 concentrations were 87.3% and 87.6% in the duplicate tests. These results suggest that the kinetics of the reforming reaction rather than the shift and carbonation reactions is responsible for lower conversion of CH4 and decreased H2 concentration at the lower temperature. The fact that CO concentration could be reduced to 12 ppmv at 400 °C indicates that the shift and carbonation rates remained quite rapid. A tradeoff between CH4 conversion and CO concentration is required since low temperature favors minimum CO concentration in the product gas, but it also decreases the conversion of CH4 significantly. This fact is important for two reasons. Unreacted CH4 means that less H2 is produced. Also, the carbon associated with unreacted CH4 in the product gas will ultimately be emitted to the atmosphere and contribute to the greenhouse gas problem. The combined effects of volumetric feed rate and mass of reforming catalyst are shown in Figure 6 where H2 and CO prebreakthrough concentrations are plotted as a function of the weight hourly space velocity (WHSV, cm3(stp)/hr/g cat). Once again, the horizontal solid lines represent equilibrium values. H2 concentration gradually decreased with increased space velocity from 96.4% at WHSV ) 6600 cm3(stp)/hr/g cat to 90.7% at WHSV ) 30 000 cm3(stp)/hr/g cat. CO concentration remained approximately constant in the range of 40-50 ppmv for 6600 cm3(stp)/hr/g cat < WHSV < 22 500 cm3(stp)/hr/g cat, but increased significantly to 90 ppmv at WHSV ) 30 000 cm3(stp)/hr/g cat. CH4 concentration increased from 3.6% at the smallest WHSV to 9.3% at the largest, whereas CO2 concentration varied between about 50 and 120 ppmv. WHSV ) 22 500 corresponds to a gas

Ind. Eng. Chem. Res., Vol. 44, No. 6, 2005 1669

Figure 6. Effect of weight hourly space velocity on prebreakthrough concentrations of hydrogen and carbon monoxide.

residence time of about 1.5 s in the packed bed of catalyst and CO2 sorbent. The level of reproducibility can be judged from the results of duplicate tests at WHSV values of 10 000 and 22 500. Conclusions To integrate an atmospheric pressure PEM fuel cell with sorption-enhanced H2 production, it is desirable that the H2 production unit operate at or near atmospheric pressure. Relatively low reaction temperatures are required to maintain CO2 removal capability at low pressures. Despite the necessity to operate at relatively unfavorable conditions, it was possible at 460 °C and 1 bar to produce a product gas containing 96 mol % H2 with CO concentration near 50 ppmv using a steam-tomethane feed gas ratio of 3. Lower temperature and/or higher pressure are effective in reducing the CO concentration, but with a sacrifice in H2 concentration. For example, at 440 °C and 3 bar, the product gas contained 92 mol % H2 and the CO was reduced to 11 ppmv. In contrast to earlier research using a sulfur-containing commercial dolomite sorbent precursor that required a pretreatment step to prevent poisoning of the Nireforming catalyst, in this study, we found that untreated, sulfur-free Arctic SHB dolomite could be used without causing catalyst poisoning. Acknowledgment The authors acknowledge personnel from the Institute for Energy Technology, and in particular Julien Meyer, for consultations on this study and for supplying the reforming catalyst and dolomite sorbent precursor. Literature Cited (1) Gary, J. H.; Handwerk, G. E. Petroleum Refining, Technology and Economics, 3rd ed.; Marcel Dekker: New York, 1994.

(2) Ruthven, D. M. Principles of Adsorption and Adsorption Processes; Wiley-Interscience: New York, 1984. (3) Korotkikh, O.; Farrauto, R. Selective Catalytic Oxidation of CO in H2 Fuel Cell Applications. Catal. Today 2000, 62, 249. (4) Han, C.; Harrison, D. P. Simultaneous Shift and Carbon Dioxide Removal for the Direct Production of Hydrogen. Chem. Eng. Sci. 1994, 49, 5875. (5) Brun-Tsekhovoi, A. R.; Zadorin, A. N.; Katsobashvili, Y. R.; Kourdyumov, S. S. The Process of Catalytic Steam-Reforming of Hydrocarbons in the Presence of a Carbon Dioxide Acceptor. In Hydrogen Energy Progress VII, Proceedings of the 7th World Hydrogen Energy Conference, Moscow, Russia, Sept 25-29, 1988; Veziroglu, T. N., Protsenko, A. N., Eds.; Pergamon Press: New York, 1988; Vol. 2, p 885. (6) Lopez Ortiz, A.; Harrison, D. P. Hydrogen Production Using Sorption-Enhanced Reaction. Ind. Eng. Chem. Res. 2001, 40, 5102. (7) Rostrup-Nielsen, J. R. Catalytic Steam Reforming. In Catalysis Science and Technology: Anderson, J. R., Boudart, M., Eds.; Springer: Berlin, 1984. (8) Williams, R. Hydrogen Production, U.S. Patent 1,938,202, 1933. (9) Gorin, E.; Retallick, W. B. Method for the Production of Hydrogen. U.S. Patent 3,108,857, 1963. (10) Balasubramanian, B.; Lopez Ortiz, A.; Harrison, D. P. Hydrogen from Methane in a Single-Step Process. Chem. Eng. Sci. 1999, 54, 3543. (11) Peng, Z.; Harrison, D. P. Low-Carbon Monoxide Hydrogen by Sorption-Enhanced Reaction. Int. J. Chem. Reactor Eng. 2003, 1, Article A37. (12) Allam, R.; Chiang, R.; Hufton, J.; Weist, E.; White, V.; Middleton, P. Power Generation with Reduced CO2 Emissions via the Sorption Enhanced Water Gas Shift Process. Proceedings of the Third Annual Conference on CO2 Capture and Sequestration, Alexandria, VA, 2004; U.S. Department of Energy, National Energy Technology Laboratory: Pittsburgh, PA, 2004 (on CDROM). (13) Ding, Y.; Alpay, E. Equilibria and Kinetics of CO2 Adsorption on Hydrotalcite Sorbent. Chem. Eng. Sci. 2000, 55, 3461. (14) Kato, M. Latest Report of NEDO and Toshiba CO2 Capturing Project, presented at the 6th Workshop of the International Test Network for CO2 Capture, Trondheim, Norway, March 2004. (15) Ziock, H.; Lackner, K.; Harrison, D. P. Zero Emission Coal Power, A New Concept, Proceedings of the First National Conference on Carbon Sequestration, Washington D.C., May 2001; U.S. Department of Energy, National Energy Technology Laboratory: Pittsburgh, PA, 2004 (on CD-ROM). (16) Lin, S.; Harada, M.; Suzuki, Y.; Hatano, H. Hydrogen Production from Coal by Separating Carbon Dioxide During Gasification. Fuel 2002, 81, 2079. (17) Zero Emission Gas Power Project: ZEG, http://www. cmr.no/avd10/nedlasting/ZEG%20Engelsk%20oktober% 202003-2.pdf. (18) Abanades, J. C.; Rubin, E. S.; Anthony, E. J. Sorbent Cost and Performance in CO2 Capture Systems. Ind. Eng. Chem. Res. 2004, 43, 3462.

Received for review November 18, 2004 Revised manuscript received January 5, 2005 Accepted January 18, 2005 IE048883G