Chemical Looping Combustion-Research for Power and Process Heat Applications

Paper from the AFRC 2014 conference titled Chemical Looping Combustion-Research for Power and Process Heat Applications by G. Richards.

Chemical Looping Combustion (CLC) has attracted significant research interest as a method to reduce CO2 emissions from power production. The concept uses an oxygen carrier material to provide the oxygen for combustion in a fluid-bed reactor. The combustion products are ideally just CO2, water vapor, and the reduced carrier material. After condensing the water, the exhaust CO2 can be used for enhanced oil recovery, or sent to geologic storage. The reduced carrier material is re-oxidized and sent back to the fuel reactor, completing the "loop" for the oxygen carrier. The CLC process has many features that are similar to conventional fluid bed combustion boilers which are already used for coal power generation. Research and development for coal power applications may have corresponding applications in industrial boilers, and the chemical process industry. This paper presents results of ongoing CLC research at the National Energy Technology Laboratory. The paper will discuss performance results of oxygen carriers, solids flow studies, and instrumentation used to measure hot solids flow. The paper will provide a discussion of how these results apply to large-scale power plants, as well as the potential for smaller industrial process heat applications

Chemical Looping Combustion - Research for Power and Process Heat Applications Geo. A. Richards*, Doug Straub, Justin Weber, Ron Breault, Ben Chorpening, Ranjani Siriwardane, U.S. Department of Energy, National Energy Technology Laboratory American Flame Research Committee 2014 Industrial Combustion Symposium September 7 -September 10, Houston, Texas. 1. Introduction The most recent projections indicate that fossil fuels will continue to provide the majority of the domestic energy supply for at least the next 25 years1. Efforts are also in progress to reduce greenhouse gas emissions in the U.S. and world‐wide. In order to reduce emissions, technologies to capture carbon dioxide from fossil‐fuel sources are needed. Capture options may involve CO2 removal technologies (i.e., post‐combustion removal) or oxy‐combustion approaches. Although significant progress has been made to reduce the costs of these technologies, there is a need to consider transformational approaches, such as chemical looping combustion (CLC). CLC is a promising technology for carbon capture because it generates heat for steam and/or power generation while creating a relatively pure CO2 stream for storage or commercial use, for example in enhanced oil recovery (EOR)2. CLC is an indirect combustion process which means that the fuel and the air do not mix in a conventional combustion reaction. The concept supplies combustion oxygen from an "oxygen carrier" such that atmospheric nitrogen is excluded from the combustion process. Hydrocarbon combustion products are thus ideally carbon dioxide and water vapor, as in oxy‐fuel combustion. However, unlike oxy‐fuel combustion, no separate oxygen production system is needed. Instead, the oxygen is supplied by re‐oxidizing the reduced carrier in a process that "loops" the carrier between fuel and air reactors as shown schematically in Figure 1. The CLC process has features that are similar to circulating fluid bed combustion, as well as fluidized catalytic cracking (FCC) hydrocarbon processing. Various research groups are evaluating versions of the CLC concept for different applications. CLC is being studied as a low‐cost option for coal power plants with carbon dioxide capture3. Compared to typical "post‐combustion" carbon dioxide capture, no energy penalty is associated with separating carbon dioxide from flue gas, because the carbon dioxide is produced as part of the process. CLC may also be attractive for petrochemical applications where both steam and CO2 are needed for other processes. For example, where CO2 is sought for enhanced oil recovery, and steam is needed for process heat, chemical looping may be able to provide both. Relatively few studies have focused on the attributes of chemical looping for industrial boilers, yet smaller boiler applications would be a useful intermediate scale to affirm chemical looping performance while developing full‐scale (> 100MWe) coal applications. For this reason, researchers at the National Energy Technology Laboratory (NETL) are investigating CLC technology for natural‐gas fueled industrial heat. Rather than pursue step‐wise scale‐ * Contact author, [e-mail] up tests for a single chemical looping application, the research aims to accelerate the technology development of CLC using data from a suite of experiments (and literature) to calibrate numeric models for desired industrial applications. A number of papers from the NETL research have been reported elsewhere4, 5, 6 ,7, 8. In this paper, we describe recent research results from a 50kWth chemical looping reactor, as well as associated instrumentation that may be useful in other industrial combustion applications. 2. Oxygen carriers Before discussing results from the chemical looping reactor, we describe the oxygen carriers. Several oxygen carriers are being studied in research at NETL. The simplest carrier is a natural hematite Fe2O3 (i.e., iron ore) which was prepared by simply crushing and screening the iron ore material (Cliffs Natural Resources, Canada) to a particle size range of 75 to 500 microns. This material has the advantage of being relatively inexpensive, and requiring little processing to prepare it for use. The disadvantages are low oxygen transfer capacity and low reactivity , requiring long residence times and large quantities for fuel conversion. To increase reactivity, NETL developed an improvement to the natural hematite that typically incorporates less than 5% magnesium oxide (MgO) by mass. The MgO "promoter" was added by liquid impregnation of magnesium salts, with subsequent drying and calcination at 850C for three hours. The particle size distribution for this material is very similar to the raw hematite base material (i.e., 75 to 500 microns. A third NETL carrier combines copper oxide with iron oxide on an alumina support, referred to as a CuFe carrier. By itself, copper oxide has many advantages as an oxygen carrier: fast reactivity, and a large capacity for carrying oxygen. However, due to the low melting temperature of copper, agglomeration is a significant consideration9,10. The NETL CuFe carrier has been developed to avoid agglomeration problems, while retaining good capacity and high reactivity. The particle size ranges from 60 - 160 microns. Detailed comparison of the performance of these three carriers is reported in Siriwardane et al.11, including multi‐cycle testing, as well as particle attrition studies. All three carriers have been successfully prepared using commercial techniques suitable for large‐scale production. Aside from sizing the raw hematite, the MgO promoted carrier was prepared by impregnation/calcination and the CuFe carrier was manufactured by spray drying/calcining. Tests of laboratory‐prepared oxygen carrier batches (~ 1 kg) achieved the same conversion and attrition performance as samples taken from ~250kg batches manufactured with a commercial supplier. 3. Experimental Description ‐ Fluid bed testing and NETL Chemical Looping Reactor The NETL research includes several experiments and companion modeling activities. Tests in a smaller‐scale fluid bed (5.5 cm ID, 0.1kWth) are used to assess carrier reactivity without physically transporting the oxygen carrier between reactors for fuel and air. Instead, in this small reactor, the air and fuel streams are alternatively fed to the single reactor. This small reactor has the advantage that it does not add the complication of solids circulation, and provides data from an environment that is precisely controlled. Experimental results from these tests are described in Siriwardane et al.11. Compared to natural hematite, the single fluid bed tests show significant conversion and reactivity benefits for the promoted and CuFe carriers described in Section 2. In addition to small‐scale fluid bed tests, NETL has developed a nominal 50kWth chemical looping reactor to evaluate chemical looping with representative solid transport and fluid bed regimes. The experiment is shown in Figure 2. On the right‐hand side, electrically heated air supplies a fluid bed followed by a contraction to form a riser that lofts the oxidized carrier to a refractory lined cyclone. Solids pass down the standpipe to a loop seal which keeps the air from the cyclone region separate from the fuel reactor, below. The overflow of the loop seal feeds another standpipe that transports the solids down to the fuel reactor. In an actual application, the fuel reactor would be fluidized with recycled CO2 or water vapor so that the combustion products would not contain nitrogen. However, in this study, the fuel reactor is deliberately fluidized with nitrogen, making it easy to quantify the CO2 and CO produced in the fuel reactor. Fluidizing nitrogen enters through a bubble cap distributor plate at the bottom of the fuel reactor. Natural gas is mixed with the nitrogen below the distributor plate. Natural gas was diluted to (5‐20 %) in the fluidizing nitrogen. Shakedown testing in the chemical looping reactor demonstrated the importance of achieving very high cyclone efficiency . For example, given a volume of solids circulating (~ 225 kg/hr) a loss rate of even 0.01% in each pass (approx. 1/4 hour cycle time) means that 2.2 kg would exit the cyclone in 24 hours. While this level of carrier loss would be manageable, upset conditions, and changes to set points can spoil the cyclone efficiency and overload downstream particulate control filters. This issue was resolved with a secondary cyclone separator that was added to the exhaust lines. Although the process is very similar to a circulating fluidized bed combustor, the presence of reactor volumes in the recirculation loop can lead to pressure balance issues and potential process upsets 4. Instrumentation - Raman real‐time gas composition and development of microwave solids flow sensor Unlike conventional combustion processes, the operation of a chemical looping reactor requires controlling the flow of solid oxygen carrier to control the stoichiometry for the fuel and oxygen reaction. Inadequate carrier flow will result in low fuel conversion because of an inadequate supply of oxygen. In addition, the oxygen carrier flow rate is essential to supply heat to the fuel reactor via sensible energy gained by oxidation in the air reactor†. For these reasons, direct measurement of the solids flow circulation rate is desirable for good process control. The solids flow can be indirectly estimated by momentarily halting the fluidization gas that allows solids to flow in the transit between the fuel and air reactors (so‐called L‐valve), and then measuring the changing fuel reactor bed pressure drop (which increases). Because the bed pressure drop is related to the mass of fluidized material, the changing bed † For most oxygen carriers, the reaction with fuel is endothermic, and the reaction with air is exothermic. Thus, some heat from the air reactor must be carried as sensible energy to the fuel reactor. Copper/copper oxide is an exception, since the chemical looping reactions are exothermic in both the fuel and air reactors. pressure drop is a direct measure of the flow rate that adds or subtracts material from the bed. This technique has been used to estimate solid circulation rates, but deliberate interruptions to solid flow are undesirable. Halting the fluidization gas can lead to plug formation in the L‐valve, which can end the test. Thus, there is a need for direct measurement of the hot‐circulating solids. To address this need, NETL and Carnegie Mellon University are developing a prototype high‐temperature microwave mass flow sensor specifically for the chemical looping application, but likely useful in other process applications. The status of the development is reported elsewhere12,13,14, but here we describe initial tests that are being carried out as part of the chemical looping reactor studies described in Section 5. It should be noted that microwave flow sensors are commercially available for low‐temperature process conditions; the tests in chemical looping are a unique application with high temperature (800C) solids flow. In addition to solids circulation, the gas composition is a valuable control input for chemical looping. Especially in experimental development, real time diagnosis of the gas composition in the fuel reactor is helpful to understand whether fuel is being converted to H2O and CO2 versus H2 and CO. The Raman Gas Analyzer, recently developed by NETL and the University of Pittsburgh, is capable of reporting the concentrations of numerous species simultaneously with sampling times below one‐second for process control applications in energy or chemical production15, 16, 17, 18. Due to the physics of the measurement, and the aim of real‐time control, only major species (concentrations nominally greater than 0.1%) are measured. All molecular gases (N2, O2, H2, CO, CO2, CH4, and H2O) are measured simultaneously. Other light hydrocarbons (C2‐C4) can be measured as well. The sensor is based upon applying a laser to produce Raman scattering within a hollow‐core capillary waveguide with a small, micro‐sized 300 micron bore with reflective thin‐film metal and dielectric linings. The effect of using such a waveguide in a Raman process is to integrate Raman photons along the length of the sample‐filled waveguide, thus permitting the acquisition of very large Raman signals for gases in a short time. The integrated Raman signals can then be used for quick and accurate analysis of the gaseous mixture. Better performance is observed at higher sample gas pressures, due to the physics of Raman scattering, which produces higher signals at higher pressures. 5. Results from NETL Chemical Looping Reactor tests ‐ Initial tests of the hematite and MgO promoted‐hematite have been completed in the 50 kWt Chemical Looping Reactor. The reactor unit consists of refractory lined carbon steel piping components that have been designed in accordance with standard process piping codes. One objective of this unit was to include the complication of heat losses and heat management between the reactor on the air side and the reactor on the fuel side. Therefore, unlike other small‐scale chemical looping reactor systems, there is no external heat added to the solids flow path. The operation includes a preheating stage with a transition to natural gas combustion and finally a transition to Chemical Looping Combustion. Figure 3 shows the temperature and pressure history for this initial test period with the MgO‐promoted hematite oxygen carrier. The oxygen carrier circulation period is approximately 50 hours in duration and is shown as a gray box in this figure. There are three separate periods of Chemical Looping Combustion. Each chemical‐looping period is several hours in duration. The total number of chemical looping operation for this test is approximately 12 hours. It should be noted that a small amount of natural gas is also injected into the air reactor during these initial testing periods to help maintain temperatures in the reactor. Subsequent testing has been performed without injecting natural gas into the air reactor, but those results are not included this paper. If we focus on the last Chemical Looping period, Figure 4 shows the time history of the temperature and gas emissions from the fuel reactor during this third period of Chemical Looping testing. During this period, the operating conditions were being varied and these conditions are summarized in Table 1. The minimum fluidization velocity for this oxygen carrier material is approximately 0.1 m/s. It is important to note that the temperature in the fuel reactor was also decreasing during these tests. The temperature effect on the reaction rates is also an important factor in the fuel conversion. The bed temperature is a dependent variable based on reaction history in the fuel and air reactor, and in this experiment, any electrical pre‐heat supplied to the reactant gases. The history is important because the significant sensible energy associated with the circulating oxygen carrier creates a difference between measured temperature and desired temperature set‐point. As shown in the upper graph of figure 4, the gas temperature above the bed is lower than the bed temperature, indicating that gas bubbles rising through the bed do not achieve thermal equilibrium with the bed material. It is interesting to note that CH4 conversion was no better than 50% depending on operating conditions - but CO production was very low. These results suggest that unreacted CH4 can pass through with the bubbles, but any intermediate CO is more fully oxidized. Numeric models are being used to fully understand the mechanism at work. Fuel Reactor Conditions Bed Height (cm) Fluidizing Gas Velocity (Ug/Umf) NG Concentration (% vol) 50‐60 1.5Umf 3.0Umf 5 10 20 (balance is N2) Air Reactor Conditions 25‐35 7Umf Table 1: Summary of Operating Conditions For Chemical Looping Period 3. Ug is the superficial gas velocity, Umf is the minimum fluidization velocity. It was expected that fuel conversion would be highest at low superficial gas velocity because the gas would have more time to react with the solids in the bed. Conversely, at a fixed gas velocity, it was expected that the greater fuel concentration would produce less conversion, simply because there was more gas to react with the same amount of carrier. Figure 5 shows the effect of both the fuel concentration and the superficial gas velocity through the fuel reactor on the amount of methane that is converted to CO2 in the exhaust. These preliminary data indicate that increasing the inlet fuel concentration in the fluidizing gas decreases the fuel conversion. Furthermore, some of the data show an inverse relationship between the fuel conversion and the superficial gas velocity through the fuel reactor (i.e., increasing velocity decreases the fuel conversion). As noted above, these data are confounded by the dependent temperature history in the reactor, so additional analysis is in progress to understand these results. Future testing will be directed toward studying these and other effects at consistent temperature conditions. The gas composition history shown in Figure 4 was measured with the Raman Gas Analyzer described earlier. The analyzer was used in parallel to conventional optical absorption cell gas analyzers to provide a verification of Raman analyzer output for the major gas species. Figure 6 compares output from the Raman and conventional CO2 gas analyzer during a chemical looping test period. The results show good quantitative agreement, affirming the Raman Gas Analyzer performance in this process application. These data are reported to demonstrate the ability of the Raman Gas Analyzer to measure real‐time gas composition using a single instrument, in a realistic process environment. Similar levels of agreement, for multiple gas species, have been demonstrated in lab‐scale tests and in other process applications. NETL is inviting commercial evaluation of the Raman concept; contact the authors for more information. Direct measurement of the solids circulation in chemical looping reactor was attempted with the microwave sensor described earlier. The sensor was installed on the cross‐over between the air reactor and cyclone (Figure 2) and launched microwaves through a 25 mm diameter solid ceramic rod that served as the boundary between the sensor and the hot process flow. A photograph of the sensor is shown in Figure 7. Because the ceramic rod is transparent to microwave radiation, it serves as a waveguide "window" into the hot process flow. Initial data sets showed the that resulting signals were noisy, but appeared to indicate changes in flow that are consistent with estimated flows from the pressure drop measurements. Data reduction is in progress as of this writing. NETL is developing a separate hot‐flow calibration test that will provide independent validation data of the measured flow rate in a controlled environment. When fully verified, it is expected that this sensor will be useful for chemical looping control as well as other chemical processes using flowing hot solids. 6. Summary This paper reports results from a continuing study of chemical looping combustion. Data from tests of an iron oxygen carrier with natural gas fuel have demonstrated stable solids circulation for more than 50 hours, with chemical looping combustion data collected over a total of 12 hours. Results show that the relation between fuel conversion and operating conditions are complicated by the dependent reactor temperature, which varies slowly over the test period due to the large thermal inertia of the circulating oxygen carrier and vessel refractory. Additional tests and data analysis are in progress to understand these results. Testing reported here also demonstrated successful use of a new Raman gas analyzer which can provide real‐time measurement of fuel or flue gas species in a single instrument. Initial tests of a high‐temperature solids flow sensor prototype were also described. Figure 1. Chemical Looping Combustion Process. Figure 2. NETL Chemical Looping Reactor. Left: cutaway of the refractory‐lined reactor. Right: simulation of the reactor, shown slightly larger than the cutaway for clarity. Colors represent the density of solids, with blue indicating the highest void fraction, red indicating the fluid bed regions. Figure 3. Time history of chemical looping test with hematite carrier. Pressure are relative to ambient (i.e. gauge pressure). Figure 4. Measured temperatures, gas concentrations, and conversion for chemical looping test period 3. The test period is approximately four hours (x‐axis). Step changes in CO2 and CH4 conversion correspond to changing conditions listed in the table. The temperature is a dependent variable over the test period (upper figure). Figure 5. Presentation of methane conversion versus methane concentration [CH4] at different superficial gas velocities in the fuel reactor. Figure 6. Comparsion of Raman gas composition measurement to conventional CO2 gas analyzer. Figure 7. Prototype microwave hot‐solid flow sensor. The ceramic waveguide at the left is inserted through a 25mm port in the reactor vessel / refractory lining, so that the face of the waveguide is flush with the reactor interior wall. The port is angled to the flow‐path (note ~ 13 degree cant on tip). References 1 Annual Energy Outlook 2014, DOE/EIA-0383(2014), U. S. Energy Information Administration, www.eia.gov/forecasts/aeo. 2 DiPietro, J. P., (2014). Next generation Enhanced Oil Recovery, Presented at the Carbon Dioxide Utilization Congress, San Diego, California, Feb. 19, 2014. Available at http://netl.doe.gov/research/energy-analysis/ publications/details?pub=68d576c3-a2ac-4e77-8963-5ada3b04984f 3 Lyngfelt, A. (2014). Chemical-looping combustion of solid fuels - Status of development, Applied Energy, Volume 113, January 2014, Pages 1869-1873, ISSN 0306-2619, http://dx.doi.org/10.1016/j.apenergy.2013.05.043. 4 Monazam, E. R., Siriwardane, R., Breault, R. W., Richards, G. A., Shadle, L. J., Carpenter, S. (2012). "Kinetics of the Reduction of CuO/Bentonite by Methane (CH4) during Chemical Looping Combustion", Energy & Fuels, 26, pp. 2768. 5 Monazam, E. R., Breault, R. W., Siriwardane, R., Tian, H. ,Simonyi, T., Stephen Carpenter, S. (2012). "Effect of Carbon Deposition on Oxidation Rate of Copper/Bentonite in Chemical Looping Process" Energy Fuels, 2012, 26 (11), pp 6576-6583. 6 Ksepko E., Siriwardane R., Tian H., Simonyi T., and Marek S. (2012). "Effect of H2S on Chemical Looping Combustion of Coal-Derived Synthesis Gas Over Fe-Mn Oxides Supported on Sepiolite/ZrO2/Al2O3" Energy & Fuels, 26, 2461- 2472. 7 Tian, H., Siriwardane, R., Simonyi, T., Poston, J. (2013). "Application of Natural Ores as Oxygen Carriers in Chemical Looping Combustion", Energy & Fuels, 27 (8), pp 4108-4118. 8 Breault, R., Richards, G. Carpenter, S. (2014). Update on NETL's Industrial Carbon Management Initiative: A Review of 2013 Activities, The 39th International Technical Conference on Clean Coal & Fuel Systems, The Clearwater Coal Conference, Clearwater, Florida, June 1-5, 2014. 9 Siriwardane R. Tian, H., Simonyi., Poston J. (2013). "Synergetic Effect of Mixed Copper-Iron Oxides Oxygen Carriers in Chemical Looping Combustion", Fuels, Pages 319-333. 10 Cho, P. Mattisson, T., Lyngfelt, A. (2004) Comparison of iron-, nickel-, copper- and manganese-based oxygen carriers for chemical-looping combustion, Fuel, Volume 83, Issue 9, June 2004, Pages 1215-1225, ISSN 0016-2361, http://dx.doi.org/10.1016/j.fuel.2003.11.013. 11 Siriwardane, R., Tian, H., Miller, D., Richards, G. (2014). Fluidized Bed Testing of Commercially Prepared MgO-Promoted Hematite and CuO-Fe2O3 Mixed Metal Oxide Oxygen Carriers for Methane and Coal Chemical Looping Combustion, 3rd International Conference on Chemical Looping, Sept. 9-11, Goteborg, Sweden. 12 F. Gao, D. W. Greve, and I. J. Oppenheim (2012) "Launcher design for chemical looping combustion," in COMSOL Conference, Boston, Oct. 3-5, 2012. 13 F. Zhang, D. W. Greve, F. Gao, and I.J. Oppenheim (2012) "Doppler solid flow measurement for chemical looping combustion," in 2nd Global Conference on Microwave Energy Applications, Long Beach, California, July 23-27, 2012. 14 Greve, D. W., Oppenheim, I. J., Chorpening, B. T. and J. Charley (2013), "Microwave Doppler flow sensor for chemical looping combustion systems," in IEEE Sensors 2013, Baltimore, Nov. 3-6. 15 Buric, M.P., Chen, K.P.,Falk, J., Woodruff, S.D., "Multimode metal-lined capillaries for Raman collection and sensing", JOSA B, Vol. 27, Issue 12, pp. 2612-2619 (2010). 16 Buric, M.P., Mullen, J., Woodruff, S.D., Chorpening, B., "Design and industrial testing of ultra-fast multi-gas Raman spectrometer", Proc. SPIE 8726, Next-Generation Spectroscopic Technologies VI, 87260K (29 May 2013). 17 M. Buric, J. Mullen, B. Chorpening, and S. Woodruff, "Industrial Raman gas sensing for real-time system control" SPIE Defense, Sensor, and Controls, Baltimore, MD, May 2014. 18 Mullen, J.C., Buric, M.P., Chorpening, B.T., and Woodruff, S.D., "Azimuthal Polarization for Raman Enhancement in Capillary Waveguides," Optical Engineering, 52, 11, November 2013.