Application of the Moving-Bed Chemical Looping Process for High

Feb 25, 2013 - William G. Lowrie Department of Chemical and Biomolecular ... Mehdi Pishahang , Yngve Larring , Martin Sunding , Marijke Jacobs , Frans...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/EF

Application of the Moving-Bed Chemical Looping Process for High Methane Conversion Andrew Tong, Liang Zeng, Mandar V. Kathe, Deepak Sridhar, and Liang-Shih Fan* William G. Lowrie Department of Chemical and Biomolecular Engineering, The Ohio State University, 125A Koffolt Laboratories, 140 West 19th Avenue, Columbus, Ohio 43210, United States ABSTRACT: The syngas chemical looping (SCL) process has been demonstrated at The Ohio State University for the conversion of gaseous fuels, such as natural gas and syngas, to sequestration-ready carbon dioxide (CO2) and high-purity hydrogen (H2) in a 25 kWth sub-pilot-scale unit operation. The present work focuses on parametric studies of the unique moving-bed reducer reactor operation for the conversion of methane to a concentrated stream of CO2. The variables studied include the reactor operating temperature, the gas hourly space velocity (GHSV), and the oxygen carrier/methane mass flow ratio. The results show that nearly full methane conversion (∼98%) can be achieved in the process with a GHSV of 395 h−1. Additionally, the post experiment oxygen-carrier analysis indicated that the oxygen-carrier conversion reached nearly 50%. Comparatively, the resulting oxygen-carrier conversion for the moving-bed reducer design is nearly 5 times greater than that theoretically achievable in a fluidized-bed reducer. The oxygen-carrier conversion profile along the height of the bed indicates that the conversion from Fe3O4 to FeO occurs at a much faster rate than the conversion from FeO to Fe in the reactor system. A multi-stage equilibrium simulation model for the reducer performance was developed using ASPEN Plus to compare against the experimental results. The simulation and experiment show a good match for oxygen-carrier and methane conversion results based on the testing conditions used. The parametric studies performed show promising results, indicating that the SCL technology can be used for the conversion of natural gas with CO2 readily separated.

I. INTRODUCTION Fossil fuels currently account for approximately 70% of the world energy consumption and are projected to supply 61% of the world electricity demand in 2020.1 The carbon dioxide (CO2) emissions from commercial fossil fuel energy conversion technologies are expected to increase 33% by 2030.2 With increasing environmental concerns on carbon emissions, natural gas will play a critical role for energy production because of its low carbon intensity per unit energy compared to all other fossil energy resources. Particularly, the United States is estimated to contain 2150 trillion cubic feet (tcf) of conventional and unconventional natural gas resources, the third largest supply worldwide.3 With the development of technologies, such as horizontal drilling, hydrofracking, and microseismic techniques, unconventional natural gas resources, in particular shale gas, are becoming an economically feasible feedstock. Additionally, H2 use is projected to increase because of the escalating demand for products, such as ethanol, gasoline, and additional hydrocarbon-based fuels. The development of novel technologies for reliable and cost-effective fossil fuels to H2 and power-production technologies is necessary to provide a practical option to satisfy the world’s energy demand, especially with potential CO2 emission regulations expected. Natural gas combined cycle (NGCC) for power generation represents one of the most advanced commercially applicable technologies for generating electricity, 2−3 times less capital intensive than coal-fired integrated gasification combined cycle (IGCC) processes.4 NGCC uses a combustion gas turbine and a heat recovery steam generator (HRSG) to produce electricity from these two units connected in series.5 In a carbonconstrained scenario, the CO2 generated from the process can be captured using three approaches: post-combustion, oxy© 2013 American Chemical Society

combustion, and pre-combustion carbon capture. Oxycombustion refers to the use of concentrated oxygen as the fuel oxidant produced via an air separation unit (ASU) and where the flue gas produced is mainly CO2 and H2O.6 The use of a cryogenic ASU requires a significant amount of parasitic energy. Independent studies show that the use of oxycombustion carbon capture on the NGCC process reduces the overall process efficiency by 8.1%.7 Additionally, current commercial gas turbines are designed for air as the feedstock. To use the conventional natural gas turbine, a flue gas recycle stream is required to prevent temperature excursions in the combustion chamber, thereby increasing the process capital cost. Post-combustion carbon capture refers to the use of the gas−gas separation technique to concentrate the dilute CO2 from the flue gas stream. Chemisorption processes, such as monoethanolamine (MEA), are employed to capture the gaseous CO2 within the liquid amine solution, which then requires significant heat to break the strong chemical bond.8 The United States Department of Energy (U.S. DOE) baseline study shows that the post-combustion carbon capture for NGCC reduces the process thermal efficiency by 7.4%.9 In pre-combustion carbon capture, catalytic reformation of the natural gas feedstock to a synthesis gas (syngas) composed mainly of CO, CO2, and H2 is used. Catalytic reformation refers to the partial oxidation of the natural gas feed with steam, Special Issue: Accelerating Fossil Energy Technology Development through Integrated Computation and Experiment Received: December 11, 2012 Revised: February 21, 2013 Published: February 25, 2013 4119

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

oxygen, or a combination thereof.10 In comparison to postcombustion carbon capture, additional steps are required to condition the syngas generated, such as a water-gas shift reactor and syngas quench cooler, before using an acid gas removal unit to capture the CO2. The added equipment operations compared to post- and oxy-combustion capture techniques result in a greater energy penalty for pre-combustion carbon capture integration with NGCC.5 Overall, a significant energy penalty is associated with carbon capture for NGCC processes. As one of the potential solutions, chemical looping technologies are projected to be economically advantageous in a carbon-constrained scenario.11−13 The chemical looping technology uses a metal oxide/sulfate material to indirectly transfer oxygen to the fuel from an air source. This reaction path design allows for inherent CO2 capture, greatly reducing the costs associated with carbon capture, utilization, and sequestration (CCUS). Extensive research has been performed to date from 10 kWth to 3 MWth pilot demonstration projects on chemical looping technologies for complete fuel combustion to CO2.14−18 Nickel-, iron-, copper-, manganese-, and calciumbased oxygen carriers have been extensively studied in these demonstration units as synthesized oxygen carriers and natural occurring ores, such as ilmenite. The majority of these demonstration projects in the field of chemical looping combustion use fluidized-bed reactors for the combustion of the fuel and the reoxidation of the metal oxide.12,19 The iron-based chemical looping technology developed at The Ohio State University (OSU) uses a unique moving-bed reactor design to ensure full fuel conversion to CO2 and the capability of co-generating hydrogen for further heat and/or chemical production. A 25 kWth sub-pilot-scale syngas chemical looping (SCL) unit has been developed for gaseous fuel conversion.20,21 As illustrated in Figure 1, the OSU chemical looping process is divided into three main reactors: the reducer, the oxidizer, and the combustor. Reactions R1−R3 summarize the main reactions in each reactor. The oxygen carrier reacts with the natural gas in the reducer to generate CO2 and steam while reducing the oxygen carrier to the Fe/FeO oxidation state. Condensing and separating the steam from the product gas results in a highly concentrated CO2 stream at the operating pressure of the process. In the oxidizer, steam is used to partially oxidize the reduced oxygen carriers and generate hydrogen via the steam−iron reaction. The combustor is used to fully regenerate the oxygen carrier with air and pneumatically convey them back to the reducer to complete the cycle. Reaction R1 is endothermic, while reactions R2 and R3 are exothermic. In a commercial-scale operation, the combustor operation will generate sufficient heat to compensate for the energy requirements of the reducer reactor as well as for power production12,21 3 CH4 (natural gas) 4 3 3 → CO2 + H 2O + 2Fe 4 2

Figure 1. Process flow diagram for the OSU chemical looping reactor system.

generation, the oxidizer step may be eliminated from the process altogether.23 In the case of liquid fuel production, the chemical looping process can be incorporated synergistically with a Fischer−Tropsch (F−T) synthesis unit for increased efficiency. The light hydrocarbons generated from the F−T reactor can be recycled to the reducer reactor to generate hydrogen from the oxidizer for enhancing the syngas feed to the F−T reactor and/or downstream processing of the heavy hydrocarbons. To date, over 800 h of operational experience has been gained from the demonstration of the 25 kWth integrated subpilot chemical looping units at the OSU research facility for both solid and gaseous fuel conversion.20−25 The sub-pilot SCL unit operated smoothly with no operational issues observed throughout all of the test cases performed. The results show that this technology has the potential to be a highly efficient means of generating electricity and/or other high-valued chemical products with 100% carbon capture. The present paper will focus on the application and validation of the moving-bed reactor design for complete natural gas conversion. A series of sub-pilot-scale tests were conducted for converting methane and determining the interplay of system design and operation parameters, including the temperature, fuel residence time, and oxygen carrier/fuel flow ratio. A multi-stage equilibrium model is used to simulate the countercurrent gas solid reactions inside the moving-bed reactor. The experimental and modeling results are compared and discussed. The integrated experimentation and modeling approach validates the feasibility and the advantages of the OSU chemical looping technology for methane conversion.

Fe2O3 +

Fe +

4 4 1 H 2O → H 2 + Fe3O4 3 3 3

Fe3O4 +

1 3 O2 (air) → Fe2O3 + heat 4 2

(reducer)

(R1)

(oxidizer)

(R2)

(combustor)

(R3)

The chemical looping process has the versatility to operate with and without hydrogen production. In the case of power 4120

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

Over 60 h of operation was conducted, analyzing the application of the OSU chemical looping process for natural gas conversion. High oxygen-carrier conversions (∼50%) and methane conversions (∼100%) were achieved. The following section will discuss the operation of the reducer test apparatus, the post-experiment solids analysis performed, and the multi-stage equilibrium model setup. 2.1. Reducer Shakedown Operation. Prior to methane injection, fully oxidized oxygen carriers are loaded into the reactor vessel. The oxygen carrier is a spherical particle nominally 1.5 mm in diameter consisting of 50% iron oxide and 50% inert support materials. Further details on the oxygen carrier used can be found in previous literature.20,21,26 The reactor is flushed with N2 at 2 standard liters per minute (slpm), and the oxygen-carrier flow is initiated at 46.20 g/min. Then, heating is commenced at a rate of 245 °C/h. The ball valves placed below the rotary solid feeder are cycled in 10 min intervals, discharging the oxygen carriers into a collection hopper. Once the vessel reaches approximately 800 °C, N2 is injected into the oxygencarrier inventory section at 7 slpm and the nitrogen flow rate into the reducer reactor is increased to match the GHSV of the test parameter. The reducer gas inlet is preheated to 200 °C. Once the process is at the operating temperature, the ball valve cycle time is reduced to 5 min intervals and the oxygen-carrier flow rate is increased to the operating condition. Measuring the amount of oxygen carrier discharged from the bottom ball valves is used to confirm the oxygen-carrier flow rate through the reducer reactor. After the system reaches the operating temperature and the oxygen-carrier and gas flow rates are steady, methane is introduced to the reducer gas inlet. Simultaneously, the nitrogen flow rate into the reducer gas inlet is reduced to maintain the GSHV of the test parameter. Table 1 summarizes six of the test cases performed in the sub-pilot unit operation. In these studies, the methane feed and GHSV ranged

2. EXPERIMENTAL SETUP AND REACTOR MODELING APPROACH Previous integrated demonstrations on the 25 kWth sub-pilot chemical looping unit illustrated the potential of this technology for natural gas conversion to CO2. Nearly 100% methane conversion was achieved at an isothermal operating temperature of 975 °C and gas hourly space velocity (GHSV) of 113.31 h−1.21 The present study analyzes the reducer performance under multiple gas and solid flow conditions to determine the effects of these operating conditions on the gas and oxygen-carrier conversions. The testing apparatus for the reducer shakedown study is illustrated in Figure 2. A 7.62 cm (3 in.) inner diameter type 310 stainless-steel

Table 1. Operating Conditions for Each Methane Reducer Shakedown Test 4/16/12 4/20/12 5/1/12 5/7/12 6/1/12 6/11/12

test

reducer temperature (°C)

wFe/wCH4

GHSV (h−1)

1 2 3 4 5 6

930 975 975 975 975 975

34.77 33.81 34.53 26.25 27.30 21.42

80.94 394.96 129.50 129.50 97.12 113.31

from 2 to 5 slpm and from 80 to 395 h−1, respectively. Two operating temperatures, 930 and 975 °C, were tested. The oxygen-carrier flow rate ranged from 145 to 156 g/min. During each experiment, the product gas stream from the reducer gas outlet is cooled and vented to the atmosphere. The gas analysis is performed for the duration of the test case. Section 3.1 provides a detailed discussion on the reducer gas profile results. 2.2. Post-experiment Oxygen-Carrier Analysis. After completion of the each test, the methane inlet feed is stopped and the system is flushed with N2 to remove any residual reactive gases in the reducer. The external heaters are shut down, and the system is allowed to cool for approximately 24 h. Throughout the cooling process, the reactor is continuously flushed with N2 to eliminate any potential reaction that can occur while the process is still warm. Once at room temperature, the bottom ball system is opened and the solids are discharged into a graduated cylinder. Oxygen-carrier samples are collected at multiple locations along the reducer height based on the measured volume. The collected oxygen-carrier samples are placed in a ceramic furnace and heated to 1000 °C for approximately 12 h to determine the extent of oxygen-carrier conversion. Air is continuously injected into the furnace at a flow rate of 11.33 slpm for the duration of the experiment. The oxygen-carrier samples are weighed before and after being placed in the furnace. Further details on the oxygen-carrier conversion study results are provided in section 3.2 below.

Figure 2. Chemical looping sub-pilot unit configuration for parametric reducer reactor studies.

cylindrical pipe is used as the moving-bed reactor. The bed height is fixed at 81.27 cm (32 in.) with a solid inventory zone above the reducer reactor gas outlet. The reactor vessel and inventory are externally heated to a desired operating temperature to maintain an isothermal reactor environment. The use of an isothermal reactor environment is based on the scale of operation and the average temperature for consistent reactor performance.12,20−23 The reducing gases are injected into the bottom of the reducer and travel upward against the flow of the oxygen carrier. A rotary solid feeder, consisting of a rotating disc and scraper, is placed below the reducer gas inlet to control the oxygen-carrier flow rate during operation. The parametric solid flow calibration under ambient and high (900 °C) temperatures is provided in previous studies.20 Ball valves placed above and below the system allow for the addition of oxidized oxygen-carrier particles and the discharge of spent particles, respectively, while preventing processing gas from leaking through either end. Additionally, a nitrogen (N2) flush is placed above the reducer gas outlet to prevent the processing gas from reacting with the oxygen carriers in the inventory zone. Mass flow controllers attached to each gas supply tank provide accurate control of the gas flow rate to the reducer. 4121

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

2.3. Reactor Modeling. ASPEN Plus was used to develop a thermodynamic model of a moving-bed reducer reactor to compare against the experimental results of each test case. In the countercurrent moving-bed reactor, the gas and oxygen-carrier profiles are evolving along the height of the reactor vessel in a contrasting manner. The methane gas gradually oxidizes to CO2 and H2O as it moves upward through the reducer, while the oxygen carrier reduces from Fe2O3 to Fe/FeO as it travels downward. A multi-stage equilibrium model has been developed in ASPEN Plus to simulate moving-bed operation.21,22 The model considers reaction equilibrium between the gas (CH4, CO, CO2, H2, and H2O) and solid (Fe2O3, Fe3O4, FeO, Fe, and C) present in the system. Five RGIBBS blocks placed in series are used to simulate the reducer reactor. Each RGIBBS block is representative of an equilibrium stage and attempts to minimize the Gibbs free energy at the given stage. Previous studies performed indicated that five RGIBBS blocks provide an accurate simulation of the countercurrent movingbed reactor in agreement with experimental results with syngas.12,21,22 Figure 3 illustrates the model design for moving-bed reducer

3.1. Gas Analysis. A gas sampling line was placed on the reducer gas outlet to characterize process performance. An inline Varian microgas chromatograph (GC) and California Analytical Instruments (CAI) infrared- and thermal-conductivity-based analyzers were used to determine the concentrations of CH4, CO2, CO, H2, and N2. The gas sampling conditioning system, consisting of a filter, manual electric gas cooler, and desiccant bed, removed any moisture in the sample line to prevent instrumentation damage. Therefore, the concentrations recorded were on a dry basis. The reducer gas outlet was sampled throughout the test runs. The methane conversion was determined on the basis of the gas concentrations at the reducer gas outlet. Equation 1 below defines the methane conversion value. χmethane =

CCO2 CCO2 + CCO + CCH4

(1)

Here, χmethane represents the methane conversion to CO2 in fractions where Ci is the volumetric concentration of the respective gaseous component i. The following equation is based on a 1:1 molar ratio between the CO2 product and CH4 reactant. Another method to determine the methane conversion is based on the H2, H2O, and CH4 concentrations. However, because of the dry gas basis requirements of the analyzer, the amount of H2O generated could not be recorded with the GC and CAI. Figures 4, 5, and 6 illustrate the gas concentration profiles for test cases 2, 3, and 6, respectively, on a dry, nitrogen-free basis from the CAI analyzer. The GC results verified the gas profile observed using the CAI continuous gas analyzer. The GC illustrated similar concentrations of H2 (0.1−0.2%) as the observed CO concentrations for all of the test cases. It can be assumed that CCO ≪ CCO2 because the CO concentration remained low (∼0.1%) for all of the cases. The assumption helps validate the use of eq 1 to represent the total methane conversion to CO2 and H2O. Table 2 summarizes the steady average methane conversion for each test case and the oxygen-carrier conversion at the discharge point from the rotary solid feeder. Further details on

Figure 3. Arrangement of the multi-stage RGIBBS model for movingbed reactor simulation. simulation. A reducer reactor model for a fluidized-bed process is also developed to compare to the moving-bed design. For the models, the reducer is maintained at isobaric and isothermal conditions similar to the experimental test conditions. The results are discussed in section 3.3 below.

3. EXPERIMENTAL AND MODELING RESULTS This section presents the post-experiment gas and solid analysis results for the fuel concentration profile and conversion, oxygen-carrier conversion profile, and modeling results. Each test case was conducted for 3 h under reactive conditions. Previous work illustrated that less than 2 h is sufficient to reach stable operating conditions.20−23 Moreover, test case 1 was performed for 3 and 6 h of operation and resulted in the same gas and solid conversion profiles. Therefore, each test case was conducted until a stable profile was reached.

Figure 4. Test case 2 reducer gas outlet profile. 4122

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

Figure 5. Test case 3 reducer gas outlet profile.

Figure 6. Test case 6 reducer gas outlet profile.

Table 2. Steady-State Methane Conversion from the Reducer and Oxygen-Carrier Conversion at the Bottom of the Reducer for Each Test Case test

steady average CH4 conversion (%)

experimental oxygen-carrier conversion (%)

1 2 3 4 5 6

98.03 97.58 99.99 96.34 98.90 80.35

39.03 32.57 32.50 42.66 43.63 48.82

χOC,exp =

mox − mred full full mox − mred

(2)

On the basis of eq 2, 0, 11.11, 33.33, and 100% represent the oxygen carrier in the Fe2O3, Fe3O4, FeO, and Fe oxidation states, respectively. Table 2 summarizes the respective oxygencarrier conversion at the outlet of the solids below the solid rotary feeder. Here, the oxygen carrier is expected to be at its lowest oxidation state (i.e., highest oxygen-carrier conversion) in the system because of the downward solids flow moving-bed design. The oxygen-carrier conversion profile with respect to the axial bed height is provided in Figure 7 for test cases 2, 3, and 6. Note that the oxygen-carrier profiles shown were taken postexperiment when steady-state conditions were reached in the reactor process prior to shutdown. Therefore, the oxygencarrier conversion profile shown is at steady-state conditions. Additionally, the heights shown in Figure 7 were approximated on the basis of the volume measured in the graduated cylinder. The design of the rotary disc solid feeder requires that the oxygen carrier travel half of the circumference of the disc before a scrapper discharges the solids onto the bottom ball valve.

the oxygen-carrier conversion results are presented in the following section. 3.2. Post-experiment Oxygen-Carrier Analysis. The oxygen-carrier conversion was determined from the oxygen carriers sampled from each test case. The experimental oxygencarrier conversion, χOC,exp, for the samples can be determined from their initial and final weight after being placed in the furnace for full reoxidation. Equation 2 was used to calculate the oxygen-carrier conversion from the experimental results. 4123

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

Figure 7. Oxygen-carrier conversion profile for test cases 2, 3, and 6.

results are illustrated in Figure 9. Note that, although the moving-bed model was developed on the basis of steady-state

Thus, the initial 20 cm of bed height of the oxygen carrier for the conversion measurement represents the residual oxygen carriers on the rotary disc feeder at the point of system shutdown. This is confirmed on the basis of the constant oxygen-carrier conversion values shown in Figure 7. In addition to the oxygen-carrier reoxidation study, carbon analyzer tests were conducted on the collected samples. The results confirmed that a negligible amount of carbon was deposited on the oxygen-carrier samples. 3.3. Modeling Results. The multi-stage equilibrium model was used to compare the thermodynamic equilibrium to the sub-pilot testing results in the countercurrent moving-bed system. A sensitivity analysis was conducted to study the effects of the oxygen carrier/methane mass flow rate ratio on the reducer performance. Figure 8 shows the sensitivity analysis results as well as five experimental results based on test conditions 2−6. All of the simulations were performed at 975 °C and 1 bar, consistent with the sub-pilot test conditions. The oxygen-carrier conversion profile with respect to the equilibrium stage was also calculated for test cases 3 and 6. The

Figure 9. Modeling results of the oxygen-carrier conversion profile for test cases 3 and 6.

equilibrium stages and does not incorporate kinetic information, the results provide some quantitative information on the progress of the reactions in the reducer. Methane and oxygencarrier conversions are calculated using eqs 1 and 2, respectively. The comparison between the simulation and experimental results is discussed in section 4.3.

4. DISCUSSION 4.1. Methane Gas Conversion. The reducer shakedown study analyzed the performance of the SCL process for methane conversion at 6 test conditions, varying the operating temperature, GHSV, and oxygen carrier/methane flow ratio. The results provide an indication of the natural gas conversion performance and optimal design parameters for the process. It is understood that longer gas residence times equate to better conversion. Test cases 2 and 3 compare the effect of GHSV as an independent parameter on gas conversion while

Figure 8. Sensitivity analysis of methane conversion with respect to the solid/gas mass flow rate ratio. 4124

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

imental methane conversion and comparing it against the available oxygen in the oxygen carrier. The calculated oxygen carrier conversion for test cases 1−6 can be determined from the methane conversion and ratio of Fe2O3/CH4. Equation 3 summarizes this approach.

keeping the temperature and oxygen carrier/fuel ratio constant. At lower oxygen carrier/methane flow ratios, shown in tests 4− 6, a lower GHSV is required to achieve high conversions. Test 5 illustrates a 98.90% methane conversion when lowering the GHSV to 97.12 h−1, while test 4 achieved 96.34% at 129.50 h−1. Both test conditions were performed at similar oxygen carrier/ methane flow ratios. Test case 3 achieved nearly full fuel conversion with a similar oxygen carrier/methane flow ratio as test 2 but with 3 times lower GHSV. It can be seen from Tables 1 and 2 that gas conversion values increase from test case 2 to test case 3 with a decrease in GHSV, validating the understanding on the effect of gas residence times. The methane oxidation with the oxygen carrier is an endothermic reaction favoring higher operating temperatures for better kinetic and thermodynamic performance. Test 1 serves to provide a baseline for analyzing the effect of the temperature on system performance. Comparing test cases 1 and 3, we see that the effect of the increased temperature outweighs the decrease in the residence time for the reaction. The comparison across test cases 1, 2, and 3 shows that the increase in the temperature can outweigh the increase in the GHSV only until a certain limit. This suggests optima present with respect to the temperature and GHSV, with system efficiency as the variable of interest. Test cases 3−6 suggest that a decreased ratio of oxygen carrier/methane flow decreases the methane conversion, outweighing the effect of a decreasing gas residence time. It can be inferred that it is critical to have a certain ratio below which the system performance drops. Comparing test cases 1 and 5 indicates that the higher operating temperature (975 °C) can produce a higher methane conversion to CO2. Test 5 operated with a slightly higher GHSV and a lower oxygen carrier/methane flow and achieved a higher methane conversation than test 1, indicating that the temperature is still a more important factor for operating the system. Effectively, the set of tests conducted go a long way in bringing out the interplay between various important parameters, which affect the system performance. Reaction R4 illustrates the potential partial oxidation reaction that can occur in the reducer reactor. 1 1 FeOx + CH4 → CO + 2H 2 + FeOy x−y x−y

(reducer)

χOC,calculated = χmethane

4Q CH × 22.4 4

3ω̇sx Fe2O3 159.69

(3)

Comparing the values shows a good match between the calculated and experimental oxygen-carrier conversions, where the difference ranged from 1.10 to 6.04% conversion. The partial inconsistency between the calculated and experimental values can be attributed to either sampling error or a possible minute air leakage through the lower ball valves as the system cooled for 24 h, affecting the oxidation state of the oxygen carrier at the lower level of the bed. Oxygen-carrier conversion analysis error because of the process valve leakage can be observed when comparing test cases 1 and 3. Previous SCL sub-pilot demonstrations indicated that mechanical valves in-line with the process are expected to undergo wear under the abrasive operating environment.21 Figure 7 illustrates the oxygen-carrier conversion profile for tests 2, 3, and 6. From these results, tests 2 and 3 show an almost identical conversion profile. The oxygen carrier/ methane flow ratio for both tests was approximately equal. Test 3 was operated at a much higher GHSV than test 2. The results confirm that a nearly sufficient gas residence time was available for both test conditions, because oxygen-carrier conversion profiles do not reflect a major difference. For test 6, the oxygen-carrier conversion profile illustrates an area in the bed where the oxygen-carrier conversion remained nearly constant at approximately 35.50% from a bed height of 28 to 57% of the total bed height. The vertical profile indicates that little to no oxygen transfer from the oxygen carrier to the gas occurred in this bed height range. The transition of the oxygen carrier from FeO (33.33%) to a lower oxidation state (Fe) can be a possible explanation for this observance. In this height range, the partial pressures of the reactant gases are in close equilibrium with the oxidation state of the oxygen carrier, resulting in a slow oxygen transfer between the two phases. As the oxygen carrier travels downward, where higher partial pressures of methane are present, the oxygen transfer rate from the solid phase increases. From the oxygen-carrier conversion profile for test 6, the transition from Fe3O4 (11.11%) to FeO (33.33%) and the reduction below FeO shows a nearly linear trend. When the oxygen carriers sampled from a bed height of 64 to 72% are compared to the oxygen carriers sampled from 24 to 45% of the total bed height, the later results in a 3.06 times greater change in the bed height with respect to the oxygen-carrier conversion. Therefore, on the basis of the comparison results, the oxygen transfer rate from Fe3O4 to FeO is approximately 3 times faster than from FeO to Fe. However, note that the conversion rate is based on the experimentally measured oxygen-carrier conversion. As mentioned above, the oxygen-carrier conversion at the lower level of the bed may have been slightly affected by an oxygen leakage through the ball valves. If one assumed that the oxygen-carrier conversion from 16.67 to 33.07% is the oxygen conversion from FeO (33.33%) to the calculated oxygen-carrier conversion, as illustrated in eq 3 (49.92%), the resulting oxygen transfer rate per change in bed height becomes half of the rate compared to the experimentally measured Fe3O4 to FeO. In

(R4)

In tests 1−6, the gas composition exiting the reducer gas outlet consisted predominantly of CO2 and unconverted methane. In test 6, the CO2 and CO concentrations reached a steady average of 80.35 and 0.30%, respectively, with the balance as unconverted methane. As mentioned in the section 3.1, on the basis of these results, it is reasonable to assume eq 1 represents the overall methane conversion. Additionally, the results indicate that the partial oxidation of the methane with the oxygen carrier is the rate-limiting step. The CO and H2 reaction with the oxygen carrier is considered relatively fast based on the gas composition from the reducer gas outlet. Additionally, previous test results with the integrated SCL sub-pilot demonstration indicated the residence time requirements, where CO/H2 was much lower than that for methane.21 4.2. Oxygen-Carrier Conversion. The oxygen-carrier conversion was determined from measuring the weight of the particle before and after reoxidation in an electric furnace. The experimental results for the oxygen-carrier conversion at the discharge point of the solids can be confirmed by calculating stoichiometric oxygen requirement for achieving the exper4125

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

from the reducer model predicts the reaction pattern in both testing conditions. Some discrepancy is found within the error range between the calculated oxygen-carrier conversion value and the experimental value. Section 4.2 discusses the possible errors in oxygen-carrier sampling during the post-experiment oxygen-carrier analysis. In a single-stage fluidized-bed reducer reactor, the iron oxide oxygen-carrier conversion is restricted to less than 11% to ensure that full conversion of methane CO2 is achieved. This has been demonstrated using modeling and experimental work in previous works.11−13 However, for a full comparison of the two reactor designs, multiple factors must be considered in addition to the oxygen-carrier conversion. Both the simulation and experimental results show that an oxygen-carrier conversion of approximately 50% is achievable in the countercurrent moving-bed reducer, significantly higher than a fluidized-bed reducer.21 The high iron oxide conversion in the moving-bed reducer enables the subsequent hydrogen generation in an oxidizer reactor via steam−iron reaction, and potentially, this approach can be integrated with the F−T process for liquid fuel synthesis. A further discussion is provided in the following section. 4.4. Chemical Looping Process Integration with a Gas-to-Liquids (GTL) Process. Figure 10 summarizes a

both approaches, the results confirm that the reduction of Fe3O4 is much faster than the reduction of FeO in this test case. 4.3. Comparison between Modeling and Experimental Results. Figure 8 illustrates the methane conversion at the reducer gas outlet calculated using the equilibrium model and the experimental demonstration results. The moving-bed reducer model predicts the methane conversion trend with respect to the solid/gas flow ratio. This model is corroborated with the experimental results at various test conditions. For example, the model predicts that, when the oxygen carrier/ methane mass flow ratio is above 20, no carbon deposition occurs. All of the post-experiment oxygen-carrier analyses have confirmed that little carbon content was found on the reduced oxygen carrier. Under the given operation conditions, a solid/ gas mass flow rate ratio of higher than 50 is required to ensure near complete fuel conversion, as shown in the modeling results as well as data from tests 2−5. In test 6, the ratio of 42.83 results in a methane conversion of less than 85%, which is also reflected by the five-stage equilibrium model. Thus, a minimal solid/gas flow ratio, namely, critical ratio, is required to fully convert methane to CO2 and H2O. At 1 bar and 975 °C, the critical ratio is 50. Deviations between the experiment and the simulation results are mainly due to the reaction kinetics in the experiments. For example, in test 6, a carbonaceous gas concentration of 80.2% CO2, 19.7% CH4, and 0.1% CO was measured. In the simulation, the equilibrium gas composition at the top of the reducer was 87.6% CO2, 0% CH4, and 12.4% CO. The comparison shows that CH 4 conversion to intermediate products, such as CO and H2, is the controlling step in the reaction system and the intermediate products can be quickly converted to CO2 and H2O. Most test cases shown in Figure 8 approach the equilibrium conditions, and a longer methane residence time will improve fuel conversion. Figure 9 shows the oxygen-carrier conversions at different equilibrium stages under test conditions 3 and 6 calculated from the moving-bed reducer model. The comparison of the experimental results shown in Figure 7 indicates that the multistage equilibrium modeling is helpful in assisting in further understanding of the solid conversion profiles in both test cases. Test 3 represents an operating condition when the solid/ gas mass flow rate ratio is higher than the critical ratio. In this case, both the sub-pilot testing data and modeling results show that no reaction takes place at the top of the reducer. In the middle part of the reducer, reduction of Fe2O3 to Fe3O4 occurs, but the conversion is limited to an extent of ∼2%. At the bottom, where methane is introduced, almost a linear oxygencarrier conversion profile is illustrated in Figures 7 and 9. Both results indicate that most of the reaction progresses at the bottom of the moving-bed reducer. Under this operating condition, the top part of the reducer can be shortened without affecting the overall reducer performance. Under test condition 6, the reaction progress is much different from test 3. In this case, an oxygen carrier/methane flow ratio of lower than the critical ratio is used. At the top of the reducer, Fe2O3 is fully reduced to FeO, as shown in both the simulation and test results. In the middle section of both Figures 7 and 9, an equilibrium stage between FeO and the resulting gas is found. At the bottom, methane introduction breaks the equilibrium condition and FeO is further reduced to a mixture of FeO and Fe. In test 6, the middle section of the moving-bed reducer can be reduced to achieve the same gas and solid conversions. The calculated solid conversion profile

Figure 10. SCL integration with GTL process configuration.

natural gas to liquid fuels production path via the SCL process. Here, the liquid fuel is produced via the hydrogenation of the CO2 produced from the reducer with the hydrogen from an oxidizer. The reaction scheme for the F−T synthesis unit is shown below.27 3H 2 + CO2 → −(CH 2) −+2H 2O

(5)

The hydrogenation of CO2 has the potential to be a highly attractive alternative over conventional F−T synthesis from CO and H2 feed stream, because the exothermicity of reaction 5 is much lower than in conventional F−T synthesis. Furthermore, previous literature illustrates that liquid hydrocarbon production yields over an iron-based catalyst perform similarly with H2/CO as the feedstock in comparison to a H2/CO2 feed.28−30 The oxygen-carrier conversion from the reducer reactor is an important factor for reducing the overall system solid circulation rate and for maximum H2 production from the oxidizer. H2 production is especially important in the SCL 4126

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

model shows that the oxidation of methane to the intermediate CO/H2 is the rate-limiting step in the process. Nearly full conversion of methane was obtained with minimal CO observed in the reducer gas outlet. Further, when the test conditions are compared, the results confirm that higher temperatures allow for the use of a higher GSHV while maintaining a nearly full methane conversion. However, as the ratio of oxygen carrier/methane flow is decreased, a lower GHSV is required for a full fuel conversion. The post-experiment solids analysis provided an indication of the oxygen-carrier conversion profile through the reducer reactor at steady operating conditions. The results confirm that the oxygen-carrier conversion reaches nearly 50%. The profile generated in the post-experiment analysis shows that the rate of oxygen transfer from lower bed heights is less than the upper bed height sections. This result indicates that the solid oxygen transfer rate per unit height in the system was greater for the oxygen-carrier reduction from Fe3O4 to FeO than from FeO to Fe. The five-stage equilibrium modeling for the reducer was conducted with the modeling results compared to the experimental results for the given test conditions. The comparison shows a good match between the experimental and modeling results for both gas and solid conversions. The sensitivity analysis of methane conversion with respect to solid/ gas mass flow rate ratio indicates that a minimal or critical solid/gas flow ratio is required to achieve full fuel conversion. The oxygen-carrier conversion profile can vary significantly depending upon the solid/gas flow ratio. Both the experimental and simulation results indicate that, when the ratio is higher than the critical ratio, a non-reacting zone is formed in the top of the moving-bed reducer. The non-reacting zone moves to the middle of the reducer when the ratio is lower than the critical ratio. The modeling results shows that a fluidized-bed reducer requires a much higher solid/gas flow ratio than the critical ratio in the moving-bed reducer to achieve the same methane conversion. Both the experimental and simulation results present a promising potential for using the SCL technology for methane conversion with nearly 100% carbon capture. The versatility of generating H2 and/or heat for chemical and power production allows the SCL process to apply for multiple applications. The scale-up of the process to a fully integrated, pressurized 250 kWth SCL pilot unit is in progress for demonstration under the Advance Research Projects Agency-Energy (ARPA-E) in Wilsonville, AL. The goal will be to verify the commercial applicability of this technology for efficient CO2 capture and H2 co-generation from natural gas and other gaseous fuels.

technology for GTL processes. Specifically, the hydrogenation reaction requires a 3:1 ratio of H2/CO2. If the oxygen-carrier conversion exiting the reducer is less than 11.11% (or Fe3O4 oxidation state), the oxidizer is thermodynamically limited to a maximum hydrogen equilibrium concentration of 50 ppmv, which would require an impractical amount of steam to generate the necessary amount of H2. However, if the solids are below 33.33% (FeO), the equilibrium H2 concentration reaches as high as 65.2%.12 A stoichiometric ratio of the H2 generation in the oxidizer corresponding to the methane processed to CO2 in the reducer can be generated on the basis of the oxygencarrier conversion exiting the reducer and the complete oxidation of the oxygen carrier with steam to Fe3O4 in the oxidizer. Figure 11 illustrates the H2 generation to CO2 ratio

Figure 11. H2/CO2 mole ratio with respect oxygen-carrier conversion at discharge of the reducer.

versus the oxygen-carrier conversion from the reducer. On the basis of this figure, the necessary oxygen-carrier conversion required to meet the target 3:1 ratio of H2/CO2 is 44.44%. Test cases 4 and 5 solid conversion results confirm the capability of the SCL process for nearly full fuel conversion while achieving the necessary 3:1 ratio for the CO2 hydrogenation process.

5. CONCLUSION The SCL process has been developed at OSU for the conversion of gaseous fuel for power and chemical production. In comparison to conventional natural gas conversion processes, the chemical looping technologies allows for inherent separation of the oxidizing gas from the fuel source by using an oxygen-carrier particle as the intermediate between the two. Thus, a high concentration stream of CO2 can be generated without energy-intensive gas−gas separation techniques or ASU. The OSU SCL process uses a unique moving-bed reducer design to co-generate highly pure streams of H2 and CO2 from natural gas. To date, over 300 h of operation have been completed with the SCL gaseous fuel-based sub-pilot demonstration unit, showing successful results for syngas and methane conversions. Reducer shakedown studies were performed on the SCL subpilot unit using natural gas to determine its performance at varied oxygen carrier/gas flow ratios, operating temperatures, and GHSVs. The results indicate that nearly full methane conversion can be achieved while sustaining high oxygen-carrier conversions greater than a fluidized-bed reducer design. The gas analysis profile in comparison to the moving-bed reducer



AUTHOR INFORMATION

Corresponding Author

*Telephone: (614) 688-3262. Fax: (614) 292-3769. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ 4127

NOMENCLATURE GHSV = gas hourly space velocity Vgas = volumetric flow rate of gas Vreactor = volume of the reactor χmethane = methane conversion CCO2 = volumetric concentration of CO2 dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128

Energy & Fuels

Article

(24) Kim, H. R.; Wang, D.; Zeng, L.; Bayham, S.; Tong, A.; Chung, E.; Kathe, M.; Luo, S.; McGiveron, O.; Wang, A.; Sun, Z.; Chen, D.; Fan, L.-S. Manuscript in review, 2013. (25) Zeng, L.; He, F.; Li, F.; Fan, L.-S. Energy Fuels 2012, 26, 3680− 3690. (26) Li, F.; Luo, S.; Sun, Z.; Fan, L.-S. Energy Environ. Sci. 2011, 4, 3661−3667. (27) Zeng, L.; Luo, S.; Sridhar, D.; Fan, L.-S. Rev. Chem. Eng. 2012, 28, 1−42. (28) Riedel, T.; Claeys, M.; Schulz, H.; Schaub, G.; Nam, S.-S.; Jun, K.-W.; Choi, M.-J.; Kishan, G.; Lee, K.-W. Appl. Catal., A 1999, 186, 201−213. (29) Centi, G.; Perathoner, S. Catal. Today 2009, 148, 191−205. (30) Borodko, Y.; Somorjai, G. A. Appl. Catal., A 1999, 186, 355− 362.

CCO = volumetric concentration of CO CCH4 = volumetric concentration of CH4 Ci = volumetric concentration of the gaseous component i χOC,exp = experimentally measured oxygen-carrier conversion based on the oxygen-carrier sample mox − mred = difference in the weight of the oxygen carrier after reduction full mfull ox − mred = theoretically maximum difference in the weight of the oxygen carrier after reduction χOC,calculated = theoretically calculated oxygen carrier conversion based on methane conversion QCH4 = volumetric flow rate of methane gas ω̇ s = mass flow rate of the iron-based oxygen carrier χFe2O3 = fraction loading of Fe2O3 in the oxygen carrier



REFERENCES

(1) U.S. Energy Information Administration (EIA). International Energy Outlook 2011; EIA: Washington, D.C., 2011. (2) Intergovernmental Panel on Climate Change (IPCC). Climate Change 2007Synthesis Report; IPCC: Geneva, Switzerland, 2007. (3) Figueroa, J. D.; Fout, T.; Plasynski, S.; McIlvried, H.; Srivastava, R. D. Int. J. Greenhouse Gas Control 2008, 2, 9−20. (4) Hammond, G. P.; Ondo Akwe, S. S. Int. J. Energy Res. 2007, 31, 1180−1201. (5) Klara, J.; Wimer, J. Natural Gas Combined Cycle Plant; National Energy Technology Laboratory (NETL): Pittsburgh, PA, 2007. (6) Kanniche, M.; Gros-Bonnivard, R.; Jaud, P.; Valle-Marcos, J.; Amann, J. M.; Bouallou, C. Appl. Therm. Eng. 2010, 30, 53−62. (7) Amann, J.-M.; Kanniche, M.; Bouallou, C. Energy Convers. Manage. 2009, 50, 510−521. (8) D’Alessandro, D.; Smit, B.; Long, J. Angew. Chem., Int. Ed. 2010, 49, 6058−6082. (9) National Energy Technology Laboratory (NETL). Cost and Performance Baseline for Fossil Energy Plants; NETL: Pittsburgh, PA, 2011. (10) Wilhelm, D. J.; Simbeck, D. R.; Karp, A. D.; Dickenson, R. L. Fuel Process. Technol. 2001, 71, 139−148. (11) Fan, L.-S.; Zeng, L.; Wang, W.; Luo, S. Energy Environ. Sci. 2012, 5, 7254−7280. (12) Fan, L.-S. Chemical Looing Systems for Fossil Fuel Conversions; Wiley-AIChE: Hoboken, NJ, 2010. (13) Li, F.; Fan, L.-S. Energy Environ. Sci. 2008, 1, 248−267. (14) Adanez, J.; Abad, A.; Garcia-Labiano, F.; Gayan, P.; de Diego, L. F. Prog. Energy Combust. Sci. 2012, 38, 215−282. (15) Andrus, H. A. E.; Chiu, J. H. M.; Lijedahl, G. N.; Stromberg, P. T.; Thibeault, P. R. Hybrid Combustion−Gasification Chemical Looping Power Technology Development; United States Department of Energy (U.S. DOE): Washington, D.C., 2006. (16) Kolbitsch, P.; Bolhar-Nordenkampf, J.; Proll, T.; Hofbauer, H. Ind. Eng. Chem. Res. 2009, 48, 5542−5547. (17) Mattisson, T.; Adanez, J.; Proell, T.; Kuusik, R.; Beal, C.; Assink, J.; Snijkers, F.; Lyngfelt, A. Greenhouse Gas Control Technol., Proc. Int. Conf., 9th 2009, 1, 1557−1564. (18) Shulman, A.; Linderholm, C.; Mattisson, T.; Lyngfelt, A. Ind. Eng. Chem. Res. 2009, 48, 7400−7405. (19) de Diego, L. F.; Garcia-Labiano, F.; Gayan, P.; Celaya, J.; Palacios, J. M.; Adanez, J. Fuel 2007, 86, 1036−1045. (20) Sridhar, D.; Tong, A.; Kim, H. R.; Zeng, L.; Li, F.; Fan, L.-S. Energy Fuels 2012, 26, 2292−2302. (21) Tong, A.; Sridhar, D.; Sun, Z.; Kim, H. R.; Zeng, L.; Wang, F.; Wang, D.; Kathe, M.; Luo, S.; Sun, Y.; Fan, L.-S. Fuel 2013, 103, 495− 505. (22) Li, F.; Zeng, L.; Velazquez-Vargas, L. G.; Yoscovits, Z.; Fan, L.-S. AIChE J. 2010, 56, 2186−2199. (23) National Energy Technology Laboratory (NETL). ChemicalLooping Process in a Coal-to-Liquids Configuration; NETL: Washington, D.C., 2008. 4128

dx.doi.org/10.1021/ef3020475 | Energy Fuels 2013, 27, 4119−4128