Hydrotalcite-Derived Copper-Based Oxygen Carrier Materials for Efficient Chemical-Looping Combustion of Solid Fuels with CO2 Capture

Chemical-looping combustion (CLC) is a promising technology that utilizes metal oxides as oxygen carriers for the combustion of fossil fuels to CO2 and H2O, with CO2 readily sequestrated after the condensation of steam. Thermally stable and reactive metal oxides are desirable as oxygen carrier materials for the CLC processes. Here, we report the performance of Cu-based mixed oxides derived from hydrotalcite (also known as layered double hydroxides) precursors as oxygen carriers for the combustion of solid fuels. Two types of CLC processes were demonstrated, including chemical looping oxygen uncoupling (CLOU) and in situ gasification (iG-CLC) in the presence of steam. The Cu-based oxygen carriers showed high performance for the combustion of two solid fuels (a lignite and a bituminous coal), maintaining high thermal stability, fast reaction kinetics, and reversible oxygen release and storage over multiple redox cycles. Slight deactivation and sintering of the oxygen carrier occurred after redox cycles at an very high operation temperature of 985 °C. We expect that our material design strategy will inspire the development of better oxygen carrier materials for a variety of chemical looping processes for the clean conversion of fossil fuels with efficient CO2 capture.


S4
The flue gas was sampled form a point 140 mm above the gas distributor via a quartz sampling probe and desiccating tubes containing CaCl2 with quartz wool as a filter at both ends. The resulting dry gas was introduced to two continuous gas analyzers placed in parallel (0.5 L min -1 , at 293 K and 1 atm) passing to each. The first analyzer (ABB EL3020) measured [CO2], [CH4], [CO] (via nondispersive infrared (NDIR)) and [O2] (using a paramagnetic method); the detection limit was 0.01 vol.%. The second analyzer (ABB EL3020) measured the mole fraction of H2 using the principle of thermal conductivity (TCD). The concentrations were recorded at a frequency of 0.5 Hz in most experiments. The time delays of gas analyzer due to mixing in the sampling system were corrected.

Cyclic oxygen release and storage
A batch of oxygen carriers (15.0 g, in fully oxidized state) was alternately exposed to inert gas (N2, 50.0 mL s -1 , STP) for 360s and oxidizing gas (air, 47.4 mL s -1 , STP) for 360-720 s in the laboratoryscale fluidized bed at a constant temperature (800-1000 °C). In some experiments, complete decomposition was performed by extending the time for N2 purge and the ultimate O2 uptake capacity was quantified. The stability of the oxygen carriers was evaluated by the apparent rate and overall amount of gaseous O2 released over multiple cycles.

Chemical-looping combustion of solid fuels
Two solid fuels were used, a lignite (Hambach) and a bituminous coal (Taldinskaya). The proximate and ultimate analyses are presented in Table S1. The coal was sieved to a size range of 1.0 -1.7 mm. A batch of oxygen carrier (15.0 g) were alternately exposed to inert gas (N2, 50.0 mL s -1 , STP) and oxidizing gas (air, 47.4 mL s -1 , STP) in the fluidized bed at a constant temperature (800-1000 °C). At each O2 release stage, a batch of coal (~0.2 g) was added to the reactor for combustion. Multiple cycles of coal combustion were also performed in this batch mode. Two sets of experiments were performed. In the first set, the oxygen carrier materials were exposed to consecutive cycles at successive temperatures of 800, 850, 900, 950, and 985 °C using different S5 fuels at each temperature. At each temperature, multiple cycles of O2 uptake and release were performed. For each cycle, one batch of coal particles was added to the bed during the O2 release stage. Detailed procedure is as follows: a batch of 15 g of oxygen carrier particles (in fully oxidized state) was added to the bed, without other inert bed materials. Then the bed was heated to the desired reaction temperature and fluidized by air (47.4 mL s -1 , STP). Then the inlet gas was switched to a flow of N2 (50.0 mL s -1 , STP) to purge the O2 from the system. After about 60 s, a batch of 0.2 g coal particles, sieved to 1.0 -1.7 mm was added to the bed. The minimum fluidization velocity was calculated using the equation developed by Wen and Yu. 2 The Umf of the oxygen carriers was around 0.10 m s -1 , with the U/Umf of oxygen carriers varying between 2.8 and 3.2 over the operation temperature of 800-985 °C. The calculated U/Umf of the raw coal particles was 1.6 -2.0, suggesting reasonable mixing of the coal particles with the oxygen carriers in the fluidized bed. Following the addition of the batch of fuel, when the mole fractions of CO, CO2 and H2 had fallen to the limit of detection of the analysers, the fluidizing gas was switched to zero grade air (47.4 mL s -1 , STP) for 6 min or extended to 12 min over temperature range of 950-985 °C until the O2 concentration in the effluent gas reached the same value as that in the inlet gas. After oxidation, the bed was purged with pure N2 before the oxygen carriers were exposed to the next cycle of reduction and oxidation. Each type of coal was tested for 3-5 cycles before the bed temperature was increased for experiments at a higher temperature. It was shown that the materials were stable over these high temperature cycles, therefore, the kinetics data of coal combustion could be compared. Further experiments using fresh particles at each temperature also gave the same combustion profiles.
In a further set of experiments, a fresh batch of oxygen carriers was exposed to 20 consecutive cycles of O2 uptake and release at 900 °C and on each O2 release cycle, 0.2 g of lignite was added to the bed. Using the same procedure, another batch of fresh oxygen carriers was tested with 20 cycles of bituminous coal at 900 °C. In both series of experiments, complete decomposition of the oxygen carriers was carried out before and after the 20 cycles, to examine the oxygen releasing capacity.

Chemical looping combustion in the presence of steam.
The combustion with solid fuels was carried out in the fluidized bed in cyclic batch mode. A batch of 20 g 60 wt% coprecipitated CuO/Al2O3 oxygen carrier particles were added to the bed. A batch of 10 cm 3 alumina sand (300-425µm) was also added to the bed. Then the bed was heated to desired reaction temperature in air (50.9 cm 3 /s, STP). Then the inlet gas was switched to a steam/N2 mixture (N2 of 53.6 cm 3 /s, STP and H2O of 49 mL/h). The molar fraction of steam corresponds to 25.6 mol% in N2. Pure steam was not used since a certain amount of carrier gas was necessary to pass through the system. After around 1-5 min, batches of 0.1-0.2 g coal particles with a size range of 1.0-1.7 mm were added to the bed using a test tube. A lower mass of coal (0.1 g) was used at higher temperatures because defluidisation occurred when too much coal was added. In most experiments the fluidizing gas during reduction was mixture of steam/N2. During the combustion stage, the temperature was varied from 850-985 °C, and the Umf of the oxygen carriers was around 0.11 m/s, with the U/Umf of oxygen carriers approximately equal to 4 using steam/N2 mixture as the fluidising gas, and 3 when using N2 alone. The oxygen carrier particles were exposed to consecutive cycles at different temperatures of 850, 900, 950, and 985 °C. At each temperature, several batches of different solid fuels were tested, after which the temperature was increased to the next temperature. When the cyclic tests were finished the inlet gas was switched to air and the heater was shut down. The oxygen S6 carrier particles were cooled in the nitrogen flow to room temperature and collected for further testing.

Gasification of coal in fluidised bed
Control experiments of solid fuels gasification over inert bed were also performed in the fluidized bed reactor using 20 cm 3 alumina sand (300-425 µm) as bed material. Rapid pyrolysis and gasification of lignite and bituminous coal were carried out over a temperature range 850-985 °C. In brief, the inert bed materials were fluidized by the steam and N2 mixture, a batch of 0.2 g of coal particles with size range of 1.0-1.7 mm was dropped into the fluidized bed. Rapid devolatilization occurred in the steam and nitrogen mixture and then the derived char was gasified to syngas. The residual char was oxidized when the feed gas was switched to air.
During the reactivity test with gases, the oxygen carrier conversion, defined as the ratio of the measured number of moles of CO2 produced to the number expected from the complete reduction of the given mass of CuO in the carrier, was used to check the mass balance of O2.
Where 2 CO y is the molar fraction of CO2 in the product gas, Since the reduction duration was long enough to completely reduce the oxygen carriers, the variation of oxygen carrier conversion could be used to determine the oxygen carrier capacity during cycling.
The carbon conversion, XC was calculated from the total amount of gaseous species containing carbon, (CO, CO2, and CH4), divided by and the amount of carbon in the coal in the ultimate analysis (Table S1) 7. Chemical-looping combustion of gaseous fuel A batch of alumina sand (20 mL, 300-425 m) served as the principal bed material to which a batch of oxygen carrier material (0.5 g, in oxidized state) was added. The bed was heated to the desired reaction temperature (950 °C) and fluidized by air. Then the oxygen carrier particles were exposed to consecutive redox cycles with each cycle consisting of four stages: (1) purging with inert N2 (50.0 mL s -1 , STP, 90 s), (2) reduction with ~2.4 vol.% CO in N2 (total flow rate of 65.8 mL s -1 , STP, 180 s), (3) purging with inert N2 (50.0 mLs -1 , STP, 90 s), (4) oxidation with zero grade air (47.4 mL s -1 , STP, 180 s). Both the fresh oxygen carriers and the samples recovered from the solid fuels experiments were exposed to 20 consecutive redox cycles. The stability of the oxygen storage materials weas evaluated by determining the apparent rate and overall amount of CO2 over multiple redox cycles.

Kinetics of gasification
For lignite coal and its derived char, the content of coal ash was quite low, therefore, the gasification of the shrinking coal particle may be described with the shrinking core model with chemical-reaction control with the resistance of diffusion in the product layer neglected. Where XC is the carbon conversion, kapp is the apparent rate constant (s -1 ), t is the time (s). Generally, En. S7 can be used to describe the kinetics of gasification in a wide pressure range. The inhibition of gasification by gasification product, i.e. H2, depends on the operating conditions. In the fluidized bed, the weight ratio of coal to the bed material was substantially small and the partial S8 pressure of H2 was much smaller than that of steam, so the inhibition effect may be neglected. Therefore, the L-H equation could be expressed as follows with the term of 2  This equation also works for the case when the gasification product was oxidised by oxygen carriers, removing the H2 and CO, as presented in CLC of coal with iron-based oxygen carriers under pressurized condition. 3 In this study the gasification of coal with and without oxygen carriers was performed in a fluidized bed at atmospheric pressure. During the gasification experiments, the partial pressure of H2 was about 0.007 MPa at the maximum and it was much smaller than 2 (0.0256 MPa at bulk). Under the low partial pressure of H2O, the gasification of coal char in steam could be assumed as first order without consideration of the inhibition effect of CO and H2. Therefore, the conversion of coal char with time could be described using a simple shrinking unreacted core model: The carbon conversion of lignite and char versus time as shown in Fig. 5a and c were simulated with this equation. For the bituminous coal, the rate of devolatilization was much faster than that of gasification of the char generated in situ. Also the relatively large particle size of bituminous coal and examination of the Weisz-Prater criterion 4 indicates that the reaction was influenced to some extent by the internal mass transfer. Nevertheless, the parameters of chemical-reaction control could be derived using Eq. S9 with only the stage of gasification of char considered.
Assuming the reaction rate constant follows the Arrhenius equation: The Arrhenius plots of reaction rate constant (ks) versus 1/T is presented in Fig. S12. The value of the apparent activation energies (Ea) were calculated from the value of the slope of the linear fitting line. The apparent activation energies for gasification of lignite, lignite char and bituminous coal char were 57, 94, and 218 kJ/mol, respectively. For the gasification of lignite and lignite char, the apparent activation energies are slightly lower than those reported in the literature for steam coal gasification, indicating that external mass transfer of steam might influence the reaction. For the bituminous coal, the value of apparent activation energy corresponds well to those in literature.

Kinetic analysis of coal combustion without steam
Here, rates of O2 release from the carrier are compared with the maximum rate of oxidation of the char, assuming it is controlled by external mass transfer. Thus, the maximum theoretical rate of S9 external mass transfer of O2 between the bed and a particle of char is given by the rate of CO2 production: is the bulk O2 concentration in the bed, mol m -3 , assumed to be equal to the equilibrium concentration of O2, e C 2 O , over the copper carrier, dPC is the diameter of a single char particle, and C , G k is the coefficient of external mass transfer for transfer of O2 to the surface of the char particle, m/s. For coal particles, the average particle size is The mass of a single coal particle is mcoal = 9.2010 -4 g per particle. At 900 °C, with fluidizing gas N2 of 53.6 cm 3 s -1 (as measured at 293 K and 1 atm), Umf =0.106 m s - m s -1 and the rate of combustion of char, or rate of CO2 production, is mol g s . The actual rate of combustion of bituminous coal char at 900 °C is 4.1×10 -4 -1 -1 coal mol g s , comparable to the theoretical external mass transfer, suggesting that the rate of combustion of the bituminous coal was controlled by external mass transfer (O2 released from the copper oxides).

Kinetics analysis of combustion of coal in the presence of steam
At lower temperatures, the maximum theoretical rate of O2 exchange between the bed and surface of a particle of char is given by the rate of CO2 production: Because the concentration of O2 was influenced by temperature, so the rate of combustion of char should be calculated individually. At 850 °C with steam as present, the fraction of O2 in the bed is estimated as 0.74% based on the dry basis fraction of 1.0% and flow rate, so the maximum rate of CO2 production is derived as mol/particle/s = 1.58  10 -4 mol/g coal/s, which is approximately the observed maximum rate of combustion of bituminous coal. The rate of combustion of lignite and lignite char is higher than the theoretical value. As presented above, the gasification of lignite and its char by steam or O2 may contribute to the reaction of gas products with the oxygen carriers, therefore, the actual rate of combustion could be higher. However, at 900 and 950 °C, the estimated maximum rate of combustion 2 CO r = 6.1  10 -4 mol/g/s is approximately the maximum measured rate of combustion of lignite and bituminous coal well. As for the lignite char the actual rate of combustion is higher than estimated, possibly owing to the temperature increase and associated higher b , O 2 p . S11  Figure S3. FTIR spectra of the Cu-Al LDH precursor. release is too fast that the O2 partial pressure drops quickly after switching the inlet gas to N2, therefore a large fraction of the O2 will be lost before the bed enters the "fuel reactor" environment, thus reaction at this temperature was not considered further.  Figure S13 shows the dry basis gas concentration profiles of gasification of three different solid fuels at typical temperatures as indicated. For all fuels, the gasification products were mainly H2, CO, and CO2. The initial release of CH4 as observed during gasification of lignite and bituminous coal was mainly from volatiles and pyrolysis gases, which was not observed in gasification of lignite char. The higher concentration of H2 was much higher than that of CO or CO2 which was due to the influence of water gas shift reaction (WGSR, CO+H2O=H2+CO2). Under higher steam concentration condition, the WGSR would be enhanced with the H2/CO2 ratio near the thermodynamics theoretical ratio of 2. Note that in all the experiments performed in this work, the off gas was sampled just above the bed, thus give a much more accurate product gas composition for evaluation of the intrinsic gasification reaction with less interference of homogeneous reaction including WGSR and methane reforming. As can be compared from the gasification profiles, the reactivity of lignite is approximate to its derived char and much higher than that of bituminous coal, which takes nearly 60 min to reach around 65mol% carbon conversion at 900 °C with the rest char accumulated in the bed. With the temperature increased, the char gasification was much faster and reached 90mol% carbon conversion after 30min. Detailed quantitative analysis on the gasification rate and carbon conversion will be compared with the data in CLC experiments. It should be noted that the low combustion rates of bituminous coal at high temperature of 950C and 985C were mainly due to the rapid O2 depletion in the bed and slow combustion during iG-CLC period.        The weight loss rates curves were further processed and deconvoluted as shown in Fig. S31. For the fresh oxygen carriers, the reduction of highly dispersed CuO was not observed at low temperatures (100-200°C) because the high loading of CuO and calcination at a high temperature yielded crystalline CuO in the sodium-stabilized Al2O3 support. The fast weight loss rate was assigned to the consumption of O2 by H2 reduction of the dominant phase, mainly the bulk CuO. The derivative of weight loss presents a major peak at 311 °C but with a small shoulder peak on the right hand side. According to the literature, the reduction of CuO to Cu at this lower temperature (300 °C) may involve the two-step reduction: CuO→Cu2O (around 300 °C) and Cu2O→Cu (320 °C) with a small shoulder at 300 °C on the peak. 8 The slow reduction during the later period accompanied with a broad low-intensity peak at about 450 °C could be attributed to the reduction of CuAl2O4 with slow weight loss of about 1.4% from 400 °C until 800 °C. The further loss of weight could be due to reduction of delafossite CuAlO2. Currently the species of the O2 can not be clearly differentiated from lattice O2 or CuAlO2.
As for the used samples in solid fuels experiments and further tested in redox cycles, the major reduction peak appeared similar but significantly different from the fresh oxygen carriers. The major reduction peak could be deconvoluted into two peaks, corresponding to the two-step of reduction of CuO→Cu2O (around 275 °C) and Cu2O→Cu (285 °C).
Interestingly the reduction reactivity of the oxygen carriers did not deactivate and instead the weight loss rate increased from 0.019 wt%/s to 0.026 wt%/s and did not change even though the sample was further tested by 20 redox cycles. However, the rates of oxidation during the temperature programmed oxidation of reduced oxygen carriers are quite different. Overall the rates of oxidation were much slower than that of reduction, S38 thus the peaks could be differentiated clearly. For the fresh oxygen carriers, there are evidently two major oxidation peaks corresponding to oxidation of Cu→Cu2O and Cu2O→CuO with center temperatures at 192 °C () and 285 °C (), respectively. The third small peak at 414 °C () is not very clear, but could be assigned to the consumption of lattice O2, as evidenced by the used samples. For the used samples, the profiles are quite different from that of fresh oxygen carriers, with much lower rate of oxidation and complete oxidation until 550 °C. Interestingly, for the samples further exposed to redox cycles,  emerged earlier at 485 °C.
The center temperatures calculated areas of fitting peaks of TPR and TPO are summarized in Table  S4. Note: a calculated by the integration of areas of fitting peaks. b calculated by overall weight loss of TG curves in Fig. S28a, including the weight loss during the 1 h isothermal period at 950 °C . c calculated by overall weight increase of TG curves in Fig. S28b, including the weight increase during the 1 h isothermal period at 950 °C . S39 Figure S32. Pore size distribution of the fresh oxygen carriers and used samples in solid fuels experiments and samples further tested in redox cycles at 950 °C. Data derived from N2 adsorption isotherms at 77K. For inset, ordinate has units of cm 3 /g nm, abscissa is in nm.