Investigation of Different Mn–Fe Oxides as Oxygen Carrier for

Nov 23, 2012 - O 2 uncoupling behaviour of ilmenite and manganese-modified ilmenite as oxygen carriers. Zhenkun Sun , Dennis Y. Lu , Robin W. Hughes ,...
0 downloads 4 Views 3MB Size
Article pubs.acs.org/EF

Investigation of Different Mn−Fe Oxides as Oxygen Carrier for Chemical-Looping with Oxygen Uncoupling (CLOU) Golnar Azimi,†,* Henrik Leion,† Magnus Rydén,‡ Tobias Mattisson,‡ and Anders Lyngfelt‡ †

Department of Environmental Inorganic Chemistry, Chalmers University of Technology, S-412 96 Göteborg, Sweden Department of Energy and Environment, Chalmers University of Technology, S-412 96 Göteborg, Sweden



ABSTRACT: The appropriate oxygen carrier for chemical-looping with oxygen uncoupling (CLOU) should be thermodynamically capable of being oxidized in the air reactor and also release gaseous O2 in the fuel reactor at appropriate temperatures and oxygen partial pressures. It should also be mechanically durable, cheap, and environmentally friendly. Iron− manganese oxides appear to be especially promising due to favorable thermodynamics. In this work, combined metal oxides of iron and manganese were investigated for the CLOU process. Particles with different ratios of Mn/Fe were produced using spray drying. The particles were calcined at 950 and 1100 °C for 4 h and then tested with respect to parameters important for CLOU. The crushing strength for these materials was between 0.1 to 1.7 N, depending on their composition and sintering temperature. The ability of the iron−manganese oxide particles to release oxygen in the gas phase was examined by decomposition of the material in a stream of N2. Moreover, the reaction with both methane and synthesis gas (50/50% CO/H2) was examined in a batch fluidized bed reactor. Here, the particles were alternately oxidized with 5% O2 and reduced in N2 or with fuel at 850 °C, 900 and 950 °C. From the results, it can be concluded that during the nitrogen period, the oxygen carriers with Mn3O4 content in the range from 20 wt % to 40 wt % release oxygen at 900 °C, whereas the materials with higher manganese content show no oxygen release. This is because they could not be oxidized to bixbyite. By decreasing the temperature from 900 to 850 °C, it was possible to oxidize oxygen carriers with manganese oxide content of 50 wt % and higher, and consequently, oxygen release during the nitrogen period was seen for these materials. This is in agreement with the phase diagram for this system. The reaction rate with methane follows the oxygen release trend very well. At the higher reaction temperature, 950 °C, oxygen carriers with manganese content in the range from 25% to 33% show the best gas conversion of methane. At 850 °C, on the other hand, high methane conversion is seen for particles with high manganese content. In fact, several particles had almost full conversion of methane to CO2 and H2O at 850 °C using a bed mass in the batch reactor corresponding to 70 kg oxygen carrier/MW.



INTRODUCTION Chemical-Looping with Oxygen Uncoupling (CLOU). Chemical-looping combustion (CLC) and chemical-looping with oxygen uncoupling (CLOU) are methods for capturing CO2 during combustion of any type of fuel, such as natural gas, coal, or biomass. In both processes, oxygen is transported from the air reactor to the fuel reactor by a solid oxygen carrier. One of the most important benefits of these techniques is that CO2 and H2O are kept separated from the other components of product gases such as excess O2 and N2 as a part of the process. By eliminating the need for the separation of gases, costly and energy consuming equipment is avoided.1 The CLOU system is composed of two fluidized bed reactors (Figure 1). One of them is an air reactor where a solid oxygen carrier, usually a metal oxide, is oxidized by air (reaction 1). The oxygen carrier will then be transported to the second reactor, the fuel reactor. In the fuel reactor, the oxygen carrier material releases gas phase O2 (reaction 2) and then added fuel reacts with the gaseous oxygen to produce CO2 and H2O (reaction 3). The reduced oxygen carrier is then again transported back to the air reactor to be reoxidized back to its original state. The total amount of heat released from the fuel reactor and the air reactor is equal to the heat released from normal combustion.2 The product gases from the fuel reactor consist almost only of CO2 and H2O. The H2O can be condensed and pure CO2 can be compressed and transported © 2012 American Chemical Society

Figure 1. Schematic diagram of chemical-looping with oxygen uncoupling (CLOU).

to an appropriate storage location. The product gases from the air reactor consist mainly of nitrogen and a small amount of oxygen which can be released to the atmosphere. The major difference between CLOU and CLC is the mechanism for oxidation of the fuel in the fuel reactor. In CLC, the oxygen carrier does not release gas phase oxygen, and hence, the fuel is oxidized by a direct gas−solid reaction between fuel and the oxygen carrier. Subsequently, if a solid fuel is applied it has to Received: July 4, 2012 Revised: November 6, 2012 Published: November 23, 2012 367

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

be gasified in order to be able to react with the oxygen carrier. However, in CLOU, the fuel can react directly with released O2 and does not need to be gasified. Therefore, the main benefit with CLOU, as compared to CLC, is that the slow gasification of the solid fuel in CLC is eliminated.2 O2 + MexOy − 2 → MexOy

(1)

MexOy → MexOy − 2 + O2 (g)

(2)

CnH 2m + (n + m /2)O2 (g) → nCO2 + mH 2O

(3)

The combined manganese oxides have been examined as oxygen carriers in CLC or CLOU. Shulman et al.28,29 tested several combinations, for example, Mn/Mg, Mn/Ni, Mn/Si, as well as the Fe/Mn oxide system. Work that has focused only on the Fe/Mn system has been performed by Azimi et al.,20,35 Ksepko et al.,36 Lambert et al.,37 and Rydén et al.38 Thermodynamics. Figure 2 shows a phase diagram of the (MnyFe1−y)Ox system calculated with the software FactSage

It should be noted that oxides with CLOU properties may react as both conventional CLC oxygen carriers and CLOU oxygen carriers, and this depends on the rates of the individual reactions. When such materials react with gaseous fuel, it may be difficult to distinguish between the mechanisms. The basic idea of chemical looping combustion was first presented in a patent by Lewis and Gilliland in 1954, where it was proposed as a technique to produce pure carbon dioxide.3 In 1983, Richter and Knoche also suggested the concept of chemical looping combustion as a power production technique for increasing the thermal efficiency.4 Later in 1994, Ishida and Jin proposed chemical looping combustion as a technique for CO2 capture in power plant.5 In 2001, Lyngfelt et al. suggested two interconnected fluidized beds as a reactor design for the chemical looping combustion process.1 Substantial research has been performed on CLC in the last years. Recently, comprehensive reviews of the development of CLC and CLOU with respect to oxygen carrier, reactor design, and prototype testing have been done by Lyngfelt6 and Adanez et al.7 The majority of the published works on CLC have focused on gaseous fuel.8−14 Solid fuels, such as coal, are more abundant and cheaper than gaseous fuel. Consequently, it would be beneficial if the CLC process could be adapted to solid fuels. This is an ongoing development both with laboratory batch reactor15−20 and circulating systems.21−25 Oxygen Carriers. The appropriate oxygen carrier for CLOU should be thermodynamically capable to be oxidized in the air reactor and also release gaseous O2 in the fuel reactor at appropriate temperature and oxygen partial pressures.2 At the same time, it has to be mechanically durable, cheap, and environmentally reliable. Some metal oxides of manganese, copper, and cobalt have appropriate equilibrium pressure of gaseous oxygen within the range from 700 to 1200 °C. However, Co3O4/CoO is unsuitable due to its high cost and high toxicity. CuO/Cu2O appears promising,26,27 but the fairly high cost and the low melting point of metallic Cu, 1085 °C, are disadvantages. Although Cu will not be formed during a pure CLOU reaction, it is likely that there will be some formation of Cu in the fuel reactor due to direct reaction with reactive gases. Applying pure manganese oxide in CLOU is troublesome because the relevant equilibrium concentrations applicable for CLOU mean operation at relatively low temperatures. For example, oxidizing Mn3O4 to Mn2O3 in the air reactor with an oxygen concentration of maximum 5% is only possible at temperatures below 800 °C.2 However, this temperature limitation can be raised by combining of active manganese oxide with other materials. Iron, nickel, silicon, magnesium, and calcium are materials that can be combined with manganese oxides to improve its characteristics.28−32 The system Fe−Mn appears to be especially promising due to favorable thermodynamics,33,34 which is also confirmed by experimental work.35

Figure 2. Phase diagram of (MnyFe1−y)Ox in an atmosphere with an O2 partial pressure of 5%.

using the FToxid database. The diagram is calculated for a partial pressure of 5% O2, which may be an appropriate basis with respect to the exiting O2 concentration in the air reactor in the CLOU process. The phase diagram of iron−manganese oxide has also been investigated experimentally by Kjellqvist et al.,39 Muan and So̅miya,40 Wickham,34 and Crum et al.,33 although in air. Figure 2 indicates that the stable phases at low temperature are hematite and bixbyite, both with the general formula Fe 2−x Mn x O 3 or (Fe,Mn) 2 O 3 , whereas spinel phases (Fe,Mn)3O4 are stable at high temperatures. There are also two-phase areas in which both forms (i.e., bixbyite/hematite and spinel) coexist at intermediate temperatures. Moving from low to high temperatures will give phase change from (Fe,Mn)2O3 to (Fe,Mn)3O4, which is accompanied by oxygen release (reaction 4) equivalent to 3.3−3.4% change of mass. A similar release of O2 is expected to occur when moving from a high to a low partial pressure of oxygen, very similar to when an oxygen carrier is transported from the air to the fuel reactor of a CLOU-system. Thus, decomposition of bixbyite to spinel, reaction 4, is the expected reaction in the fuel reactor. Then, the fuel would react with released O2. In the air reactor, reaction 4 is reversed; that is, bixbyite is recreated by oxidation with air. 1 3(Mn, Fe)2 O3 ↔ 2(Mn, Fe)3 O4 + O2 (g) (4) 2 Figure 2 shows that when the oxygen partial pressure in the air reactor is 5%, pure Mn2O3 can be oxidized at a temperature below 800 °C. Experiments with Mn3O4 suggest that this temperature is too low to have a sufficiently high reactivity.41 However, the combination of iron oxide and manganese oxide has different properties, and it should be possible to oxidize by 5% O2 at higher temperature. This can be seen in Figure 2, as the phase transition boundary between bixbyite and the two phase region of bixbyite and spinel occurs at higher 368

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

Experimental Procedure. The CLOU property of iron− manganese oxide is examined by decomposition in N2 and moreover the reaction with both methane and synthesis gas (50/50% CO/H2) was examined. The experiments were done in a fluidized bed quartz reactor that has a length of 820 mm and a porous quartz plate of 22 mm in diameter placed 370 mm from the bottom. A sample of 15 g of oxygen carrier particles with diameter of 125−180 μm was placed on the porous plate and the reactor was then heated to the temperature of interest in a flow of 900 mLn/min containing 5% O2 in N2 in order to prevent uncontrolled release of oxygen and to ensure that the oxygen carriers are adequately oxidized prior to the experiments. As the required conditions were reached, the particles were fluidized by 600 mLn/min of pure N2, and the outlet oxygen concentration was measured during the inert period. The particles were exposed to consecutive cycles of oxidizing and inert periods at a temperature of 900 °C and also periods in which the temperature for oxidation was still 900 °C, but the temperature was raised to 1000 °C during the inert period. It should be mentioned that, in this paper, the periods in which the pure nitrogen is the only fluidizing gas in the reactor are called nonfuel periods or inert periods. The nonfuel period helps to better understanding the O2 uncoupling behavior, since N2 is inert and does not interfere in oxygen release. For reactivity evaluation, the particles were exposed to 365 mLn/min CH4 or 450 mLn/min synthesis gas (syngas) at 950 °C. The oxidation and the reduction periods were separated by an inert period during which the reactor was purged from reactive gases and gaseous products by introduction of N2. Some of the particles were also examined at a temperature of 850 °C for both decomposition in nitrogen and reactivity test with methane. Table 2 presents a detailed plan of experiments. It should be noted that the particles with calcination temperature of 1100 °C were selected as a basis for all experiments and were examined at three different temperatures, 850, 900, and 950 °C. However, for materials calcined at 950 °C, the two most interesting temperatures were examined. Thus, the samples with manganese content of lower than 50% were examined at 900 and 950 °C , and the material with manganese content higher than 60% were tested at 850 °C. This is motivated both by the results with the materials calcined at 1100 °C and also by the thermodynamic analysis, which clearly shows why lower temperature should be used for high manganese content and vice versa, cf. Figure 2. The gas from the reactor was led to an electric cooler for removing water and then to a Rosemount NGA 2000 multi-component gas analyzer, which measured the concentrations of CO, CO2, CH4, and O2 and the gas flow. The temperature was measured 5 mm under and 10 mm above the porous quartz plate using Pentronic CrAl/NiAl thermocouples with inconel-600 enclosed in quartz shells. The temperature presented in the paper is the set-point temperature, that is, the temperature at the beginning of the reduction. From high frequency measurements of the pressure drop, it was possible to see if the bed was fluidized. Data Evaluation. The degree of conversion, X, describes the extent to which the oxygen carriers are oxidized and is defined as follows:

temperature as the amount of Fe is increased. However, as Mn and Fe are mixed the phase transition between fully oxidized phase (Mn,Fe)2O3 and fully reduced phase needs to pass a two phase area where both phases coexist. This means that, for a constant oxygen partial pressure, a certain temperature change is needed to accomplish the phase change between the fully oxidized and fully reduced phases. The same will also apply to the needed change in oxygen concentration, if a change in oxygen concentration is used to achieve this phase change. Figure 2 also indicates that the smallest temperature change is needed where the manganese fraction is 60−80%. In this work, a systematic investigation of oxygen carriers of iron−manganese oxide was performed. This was done by preparing the particles with different combinations of iron and manganese oxide, with the Mn/Fe molar ratios varying between 4:1 and 1:4. The particles were calcined at two different temperatures, 950 and 1100 °C, and investigated with respect to oxygen release and up-take potential and also their reactivity with respect to methane and synthesis gas (50/50% CO/H2).



EXPERIMENTAL SECTION

Materials. The oxygen carriers studied in this work are particles with different molar ratios of Fe/Mn that were produced using spray drying at VITO in Belgium. Powder mixtures of Mn3O4 (supplied by Chemalloy) and Fe2O3 (supplied by Alfa Aesar) were dispersed in deionized water containing organic additives. Polyethyleneoxide (PEO, type PEO-1 Sumitomo Seika, Japan) and/or polyvinylalcohol (PVA 1500 Fluka, Switzerland) and/or polyethyleneglycol (PEG 6000, Merck-Schuchardt, Germany) were used as organic binder; Darvan (type C, RT Vanderbilt, U.S.A.) and/or Dolapix (types A88, PC75 and PC80, Zschimmer & Schwarz, Germany) and/or Targon 1128 (BK Giulini Chemie, Germany) were used as dispersants. The water-based suspension was continuously stirred with a propeller blade mixer while being pumped to the 2-fluid spray-dry nozzle, positioned in the lower cone part of the spray-drier. After spray-drying, the fraction within the required particle size range was sieved to be separated from the rest of the spray-dried product. To obtain oxygen carrier particles with sufficient mechanical strength, calcination was performed in air at 1100 or 950 °C for 4 h. After calcination, the particles were sieved again and are in the size range 125−180 μm.42 Table 1 shows specifications of the different particles used in this work. Here, M denotes Mn3O4, and the digits after M represents the manganese oxide mass fraction of the sample. Further, F denotes the iron oxide mass fraction, and the digit after F denotes calcination temperature of the sample.

Table 1. Composition of Iron−Manganese Oxide Materials denotation

molar ratio (Fe/Mn)

mass ratio (Fe2O3/Mn3O4)

calcination temp. (°C)

M20F1100 M20F950 M25F1100 M25F950 M33F1100 M33F950 M40F1100 M40F950 M50F1100 M50F950 M67F1100 M67F950 M75F1100 M75F950 M80F1100 M80F950

4:1 4:1 3:1 3:1 2:1 2:1 1.5:1 1.5:1 1:1 1:1 1:2 1:2 1:3 1:3 1:4 1:4

80.7:19.3 80.7:19.3 75.8:24.2 75.8:24.2 67.7:32.3 67.7:32.3 60:40 60:40 51.1:48.9 51.1:48.9 34.4:65.6 34.4:65.6 25.9:74.1 25.9:74.1 20.7:79.3 20.7:79.3

1100 950 1100 950 1100 950 1100 950 1100 950 1100 950 1100 950 1100 950

X=

m − mred mox − mred

(5)

Here, m is the actual mass of the sample, mox is the mass of the fully oxidized sample (i.e., bixbyite), and mred is the mass of the sample in its fully reduced form (i.e., spinel). The degree of conversion of oxygen carriers as a function of time during reduction with methane and syngas is calculated from the outlet gas concentrations using eqs 6 and 7, respectively.

Xi = Xi − 1 −

∫t

t1 0

1 nout ̇ (4pCO ,out + 3pCO,out − pH ,out ) dt 2 2 n0Ptot (6)

369

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

Table 2. Experimental Plan for Testing with Decomposition and Gaseous Fuela

a

no. cycles

reducing gas

FOx (mLn/min)

FIn (mLn/min)

tIn (s)

3 1 3 3 3 3 3 3 3

nitrogen nitrogen methane syngas nitrogen nitrogen nitrogen methane nitrogen

900 900 900 900 900 900 900 900 900

600 600 600 600 600 600 600 600 600

360 360 60 60 360 360 360 60 360

FRed (mLn/min)

tRed (s)

365 450

20 80

365

20

TOx (°C)

TRed (°C)

900 900 950 950 900 900 850 850 850

900 900 → 1000 950 950 900 900→1000 850 850 850

Fx is flow in period x, (i.e. Ox(idation), Red(uction) and In(ert)).

Figure 3. Oxygen concentration vs (a) oxygen carrier conversion, X, and (b) time during the final nonfuel periods after the fuel cycles at 900 °C for 360 s.

Xi = Xi − 1 −

∫t

t1 0

1 nout ̇ (2pCO ,out + pCO,out − pH ,out ) dt 2 2 n0Ptot

γCO = 2

t1 0

1 (nout ̇ pO ,out ) dt 2 n0Ptot

(8)

where Xi is the conversion as a function of time for a period i, Xi−1 is the degree of conversion after the foregoing period; t0 and t1 are the times for the start and the end of the period; n0 is the moles of active oxygen in the fully oxidized sample; ṅout is the molar flow of dry gas entering the analyzer; Ptot is the total pressure; pCO2,out, pH2,out, and pCO,out are the outlet partial pressures of CO2, H2, and CO after removal of water vapor; pO2,out is the partial pressure of exiting oxygen. The value of pH2,out was not measured online with gas analyzer, but it was calculated by the assumption of having equilibrium water−gas shift reaction. CO + H 2O ↔ CO2 + H 2

m = 1 + R 0(X − 1) mox

(11)



RESULTS Reactivity of the Oxygen Carriers. The different iron− manganese oxide particles were tested by exposure to nitrogen, methane, and syngas according to the previously described experimental scheme. Figure 3 illustrates oxygen concentration as a function of (Figure 3a) oxygen carrier conversion, X, and (Figure 3b) time during the final nonfuel periods after the fuel cycles for 360 s at a temperature of 900 °C for different combinations of iron− manganese oxide. In Figure 3b, a blank run with 15 g sand has

(9)

To be able to compare oxygen carrier materials that contain different amounts of oxygen, a mass-based conversion, ω, is defined as follows:

ω=

2

Characterization of Oxygen Carriers. The crushing strength (i.e., the force needed to fracture the particles) was examined using a Shimpo FGN-5 crushing strength apparatus. For each sample, 30 different particles of size 180−250 μm were tested. The value of the crushing strength is achieved by calculating the mean value of the 30 collected values. The analysis of the phase compositions of the oxygen carrier particles was performed on a Siemens D5000 powder X-ray diffractometer (Cu Kα1, k = 1.54056 Å). The shape and morphology of fresh and tested oxygen carriers were observed using a FEI, Quanta 200 environmental scanning electron microscope FEG. The bulk density of all materials, sized 125−180 μm, was measured by weighing 5 mL of particles filled in a graduated cylinder. The Brunauer− Emmett−Teller (BET) surface area of the particles was measured by N2-absorption using Micromeritics, ASAP 2020.

Correspondingly, the degree of conversion for the inert period is determined using the relationship

∫t

2

pCH + pCO + pCO 4

(7)

Xi = Xi − 1 −

pCO

(10)

For analysis of gas conversion, the fraction of CO2 in the outlet gas flow was calculated on a dry basis as follows: 370

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

been added to show the transition profile of oxygen gas in the applied experimental setup. As can be seen, there is some backmixing that was also considered in the calculation of X in Figure 3a. During the nonfuel periods, some of the particles release oxygen and some of them are not able to release oxygen. Particles with manganese (Mn3O4) content of 20−40% released oxygen at the end of the nonfuel periods. The levels of oxygen for these materials after 360 s are in the range from 0.2% to 0.4%. The other materials did not release any oxygen during the nonfuel periods. The latter can be explained by the phase diagram, Figure 2, which shows that the reduced oxygen carriers with manganese content of more than 40% would be difficult to oxidize to bixbyite in 5% of oxygen at 900 °C because they are very close to the phase region of bixbyite + spinel or spinel. The most likely reason is that the reaction is kinetically hindered when conditions are close to those where the reduced phase is stable. Consequently, according to the results from the nonfuel periods, M25F950, M33F950, and M33F1100 show the best behavior in terms of its release of oxygen. According to the mentioned experimental scheme, the temperature was raised from 900 to 1000 °C during nonfuel periods to investigate the release of oxygen during temperature increase. According to the phase diagram (Figure 2), by increasing temperature, the oxygen carriers are expected to release oxygen at higher oxygen partial pressure. The results for the last performed periods with temperature increase are presented in Figure 4. As is evident, the temperature increase leads to a significant oxygen release for materials with 20−40% manganese.

Figure 5. Dry gas concentration for reduction of M20F950 at 950 °C using methane for 20 s.

of CO2, CH4, and CO and the up-take of oxygen vary for each material. In Figure 5, the air is shifted to nitrogen at the time 20 s. The oxygen carrier decomposes and releases oxygen. At the time 80 s, methane is inserted for 20 s. Methane reacts with oxygen from the oxygen carrier and produces CO2 and small amounts of CO. The amount of oxygen decreases to 0% when the nitrogen is switched to methane. As seen in Figure 5, the material cannot produce a sufficient amount of O2 to attain full conversion of CH4 to CO2, and hence, there is still unconverted CH4 and CO leaving the reactor. There is some backmixing of the gas before it reaches the analyzer giving a transient of approximately 10 s. This backmixing can explain the overlapping period in Figure 5 when O2 and CH4 are measured simultaneously during 10 s. Oxidation is exothermic for all the materials that were also confirmed by a temperature increase during the oxidation. The overall reaction in the fuel reactor for Mn−Fe material with manganese content higher than 35% is exothermic, as calculated by HSC Chemistry software. Further, Rydén et al.43 has calculated the enthalpy for O2 release of materials with Mn/Fe molar ratios 80:20 and 50:50 to be 254 kJ/mol O2 and 344 kJ/mol O2 respectively. Compared to the enthalpy of combustion of methane at 900 °C (−401 kJ/mol O2), it can be determined that the reaction with methane is exothermic in the fuel reactor. Figure 6 shows the gas conversion, γ, from eq 11, as a function of mass-based oxygen carrier conversion at 950 °C with (Figure 6a) methane and (Figure 6b) syngas, both for the final cycle in Table 2. The value for ω does not start at 1 because it decreases slightly due to release of oxygen during the short inert period before the reduction. The methane conversion for M25F950 and M33F950 are higher than for the others. The oxygen carriers with 50−80% manganese that did not release any oxygen during nonfuel periods in inert atmosphere, showed some gas conversion during reduction phase, but it was generally lower compared to the others. According to Figure 6b, almost all of the particles except M67F1100 showed a full conversion of syngas. As shown previously in the phase diagram, Figure 2, the oxidation of the oxygen carriers with manganese content higher

Figure 4. Oxygen concentration vs oxygen carrier conversion, X, during the last nonfuel period with temperature increase from 900 to 1000 °C.

The reactivity of oxygen carriers was examined by reaction with methane and syngas. Figure 5 demonstrates the outlet dry gas concentration during reduction and the following inert period at 950 °C using methane as fuel under 20 s reduction of M20F950. It should be mentioned that all the other oxygen carriers showed a similar type of behavior, although the amount 371

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

Figure 6. Gas conversion, γ, vs mass-based conversion, ω, at 950 °C using (a) methane and (b) syngas.

Figure 7. (a) Oxygen concentration vs oxygen carrier conversion, X, during nonfuel periods and at 850 °C for 360 s and (b) gas conversion,γ, vs X, at 850 °C using methane for 20 s.

From the phase relationships presented in Figure 2, applying a temperature lower than 900 °C in the air rector should make it possible to oxidize these materials to bixbyite. Hence, the materials were also tested at a lower temperature, i.e. 850 °C. Figure 7a shows the oxygen concentration as a function of oxygen carrier conversion during nonfuel periods for 360 s at 850 °C. In Figure 7b, gas conversion is plotted against oxygen carrier conversion using methane for 20 s at 850 °C. As seen in Figure 7, both oxygen release and reactivity for oxygen carriers with manganese content of more than 50%, increase when reducing the temperature to 850 °C and decrease for material with manganese content of less than 50%. The particles with calcination temperature of 950 °C show better oxygen release and methane conversion than the particles calcined at 1100 °C. These observations were expected since the lower calcination temperature gives softer particles with more porosity, which results in higher reactivity. The oxygen carriers M67F950, M75F950, and M80F950 have almost full conversion of methane to CO2 and H2O at 850 °C.

than 50% would be difficult or impossible in the air reactor with 5% of oxygen at 900 or 950 °C. This is because the oxygen carriers are in or very close to the phase region of bixbyite + spinel or spinel. Therefore, at 950 °C, these particles are introduced to the fuel reactor in the reduced spinel phase, (Mn,Fe)3O4, and hence, the methane conversion shown in Figure 6a is likely due to further reduction of (Mn,Fe)3O4 to MnO and FeO by gaseous fuel. This reduction step cannot release O2 in gas phase and is also not capable of providing full conversion of methane due to equilibrium constraints. The mass-based conversion, ω, is used in Figure 6 instead of X, because the X is defined based on the conversion of (Mn,Fe)2O3 to (Mn,Fe)3O4, while the reaction likely taking place for particles with higher Mn contents is Mn3O4 to MnO. For comparison to the other figures, the mass-based conversion, ω, can be converted to X using eq 10. Thus, a full conversion in the other figures corresponds to a change in ω of R0 = 0.0336. 372

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

Analysis of the Oxygen Carrier Particles. Besides the reactivity of particles, mechanical strength and fluidization properties of the particles are important in a CLC reactor system. Measuring the pressure drop in the reactor, determine whether the bed in the reactor is fluidized or not. Table 3

particles defluidized and a plus sign shows that the particles fluidized more or less for the whole experiments. As can be seen from Table 3, the crushing strength varies between 0.1−1.7 N and generally increases as a function of Fe content. There is less of an effect of the calcination temperature, although for high-Fe materials the crushing strength is somewhat higher for materials treated at 1100 °C. Although the exact correlation between crushing strength and attrition is not known, it can be expected that material with too low crushing strength will break apart in a real unit. For comparison, the crushing strength of a Ni-based oxygen carrier successfully operated in continuous fluidized bed CLC reactor for 1000 h was 2.3 N.44 The surface areas of the investigated materials are all below 2.2 m2/g, with the highest surface area seen for the Mn-rich oxygen carriers calcined at 950 °C. There were some tendencies for defluidization, which also can be noted in Table 3, and this effect was most prominent at the higher reaction temperature. However, most of the materials that defluidized at 950 °C actually fluidized well at the lower reaction temperature, 850 °C, the exception being M20F1100 and M25F1100. For some experiments at 950 °C, a part of the material was also agglomerated after the experiment, for instance M20F950, M25F1100, M33F1100, M33F950, and M75F1100. During some experiments, some of the particles turned to fine dust and a thin layer of dust covered the quartz wall of the reactor, which made the reactor look black. This was observed for M50F1100 and M75F1100 for both temperatures and for M67F1100 for tests at 950 °C. This behavior could be associated with the very low crushing strength of these oxygen carriers. All materials were investigated using scanning electron microscopy (SEM). Figures 8 and 9 shows the results for M25F1100, M40F950, and M67F950 both for fresh and used materials. All materials were spherical although they showed

Table 3. Crushing Strength, Bulk Density, BET Surface Area, and Fluidization Properties of Different Oxygen Carriers fluidizing properties denotation

crushing strength (N)

bulk density (kg/m3)

BET surface area (m2/g)

950 °C

850 °C

M20F1100 M20F950 M25F1100 M25F950 M33F1100 M33F950 M40F1100 M40F950 M50F1100 M50F950 M67F1100 M67F950 M75F1100 M75F950 M80F1100 M80F950

1.7 1.6 1.4 1.1 0.9 0.8 0.7 0.7 0.2 0.5 0.1 0.4 0.3 0.4 0.6 0.5

2090 1475 1862 1254 1563 1145 1549 1215 1150 1087 1149 1040 1244 980 1790 1043

0.22 1.02 0.39 1.28 0.56 1.62 0.47 1.7 0.71 1.83 0.55 1.83 0.38 1.94 0.22 2.23

− + − + + + + + − + −



− −

− + + + + + + + + +

presents the measured crushing strength, bulk density, BET surface area, and also fluidization properties of different materials at both temperatures 950 and 850 °C. A minus sign in the column for fluidization properties shows that the

Figure 8. SEM images of the fresh particles: (a) M25F1100, (b) M40F950, and (c) M67F950. The bars at the bottom right of the upper and lower figures are 100 and 10 μm, respectively. 373

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

Figure 9. SEM images of the used particles: (a) M25F1100, (b) M40F950, and (c) M67F950. The bars at the bottom right of the higher and lower figures are 100 and 10 μm, respectively.

some structural variations on the surface. From the SEM images, it can be seen that particles with higher crushing strength (e.g. M25F1100) have a smoother surface and coarser structure whereas particles with lower crushing strength (e.g. M67F950) have a rough surface structure of fine granules. This is well correlated with the BET measurements in Table 3, which showed that the oxygen carriers calcined at the higher temperature had lower BET surface areas, which suggests the formation of larger primary particles, as also seen in the SEM images. As the reaction rate with methane was also dependent upon the calcination temperature, with higher rates of reaction for materials sintered at 950 °C, this may suggest that the physical changes may have affected the limiting rates of reaction, for example, from kinetic to diffusion control. The crystalline phase composition of the fresh oxygen carrier samples was examined with X-ray powder diffraction. The identified phases were bixbyite and for high iron content material, hematite. However, there is still a difficulty in safe analyses of this system as the XRD peaks of several of these compounds are close (since iron and manganese are neighbors in the periodic table). Some additional tests were performed with M33F950 particles in order to better elucidate the reduction mechanism. Here, a fresh sample was first reduced in 900 °C in N2 until the measured oxygen concentration approached zero, at which point the sample was cooled in nitrogen and analyzed with XRD. The same sample was then again heated in an inert atmosphere to 950 °C, and there exposed to methane during only a 10 s reduction. As can be seen in in Figure 10, the identified phases for fresh M33F950 are bixbyite and hematite structure of (Mn,Fe)2O3 as expected. The sample reduced in nitrogen contained two phases: oxidized hematite (Mn,Fe) 2 O 3 and the reduced spinel phase (Mn,Fe)3O4. This is also expected from the phase relationships in Figure 2, as a complete reduction to spinel will occur only at

Figure 10. Diffractograms for fresh M33F950, reduced particles in N2 at 900 °C, and reduced particles with CH4 at 950 °C.

a very high temperature or very low partial pressure of oxygen for material with this composition. Hence, as the partial pressure of oxygen is very low at 900 °C, complete reduction in nitrogen may take a long period of time, and thus, the two identified species are expected. When this sample was heated to 950 °C, there was a clear release of oxygen seen due to decomposition of the remaining hematite to spinel at the higher temperature. Further reduction with a 10 s pulse with CH4 completely reduced the hematite to spinel, as can be seen in Figure 10.



DISCUSSION This work is the first comprehensive investigation of the Fe− Mn−O system for chemical-looping with oxygen uncoupling 374

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

(CLOU), where materials with varying Mn/Fe ratios have been investigated with respect to oxygen release and reactivity. The results clearly illustrate that this system could have great promise, as many of the materials release a fair amount of oxygen at relevant conditions. Figure 11 illustrates the concentration of O2 as a function of manganese content at the end of the 300 s nonfuel periods at 900 and 850 °C.

Figure 12. Methane conversion,γ, at 950 and 850 °C at ω = 0.997 versus manganese content.

ratio of about 1.2. The results presented showed that the reduced oxygen carriers with manganese content of more than 50% at 950 and 900 °C would be difficult or impossible to oxidize to (Mn,Fe)2O3 because they are in or very close to the phase regions of bixbyite + spinel or spinel at a partial pressure of O2 of 0.05. Therefore, applying a lower temperature of 850 °C will improve the reoxidation of these oxygen carriers. It should be mentioned that these thermodynamic restrictions are just valid for the air reactor, and the temperature of the fuel reactor could be higher, which would likely enhance the overall kinetics and the equilibrium oxygen partial pressure. This would be a possibility since the overall reactions in the fuel reactor for high manganese content oxygen carrier are exothermic.

Figure 11. O2 concentration as a function of manganese content at the end of the 300 s nonfuel periods at 900 and 850 °C.

From Figure 11, it can be concluded that the oxygen carriers with manganese content in the range from 20% to 40% release oxygen at 900 °C, whereas the materials with higher manganese content show no oxygen release. This indicates that the oxygen carriers with high manganese content could not be oxidized to bixbyite at 900 °C at any feasible rate. The latter is likely explained by the fact that the materials with manganese content of 50% and higher are in a phase region which is near to the spinel phase, cf. Figure 2. Figure 11 shows that, by decreasing the temperature from 900 to 850 °C, the oxygen release of material with manganese content of 50% and higher is significantly increased from 0% to 0.1−0.45%. Also, the particles with a calcination temperature of 950 °C show better oxygen release than the particles calcined at 1100 °C. In Figure 12, the methane conversion, γ, from eq 11, for an oxygen carrier conversion, ω, equal to 0.997, is plotted as a function of manganese content. As is evident from Figure 12, the methane reactivity shows the same trend as the oxygen uncoupling properties in Figure 11. At the higher temperature, 950 °C, oxygen carriers with manganese content in the range from 25% to 33% show the best gas conversion. At 850 °C, on the other hand, high methane conversion is seen for high manganese content oxygen carrier. In a CLC system, it is important to keep the air ratio low in order to improve efficiency, which means that the oxygen concentration from the air reactor should be as low as possible. In this work, 5% O2 was employed, which corresponds to an air



CONCLUSION The CLOU property of iron−manganese oxide particles with different molar ratios of Fe/Mn was examined by decomposition in N2 at different temperatures. Further, the reaction of these combined oxides with both methane and synthesis gas (50/50% CO/H2) was examined. For decomposition in N2, the oxygen carriers with manganese content in the range from 20% to 40% release oxygen when cycles of oxidation and release were performed at 900 °C, whereas the materials with higher manganese content show no oxygen release at this temperature. By decreasing the temperature of the cycles from 900 to 850 °C, the oxidation of material with manganese content of 50% and higher was possible with 5% of oxygen leading to release of oxygen during the nitrogen periods. Also, the particles with calcination temperature of 950 °C show better oxygen release than the particles calcined at 1100 °C. During the methane period, at the higher temperature, 950 °C, oxygen carriers with manganese content in the range from 25% to 33% show the best gas conversion. At 850 °C, on the other hand, high methane conversion is seen for particles with a high ratio of Mn 375

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

(15) Leion, H.; Mattisson, T.; Lyngfelt, A. The use of petroleum coke as fuel in chemical-looping combustion. Fuel 2007, 86 (12−13), 1947−1958. (16) Scott, S. A.; Dennis, J. S.; Hayhurst, A. N.; Brown, T. In situ gasification of a solid fuel and CO2 separation using chemical looping. AIChE J. 2006, 52 (9), 3325−3328. (17) Brown, T. A.; Dennis, J. S.; Scott, S. A.; Davidson, J. F.; Hayhurst, A. N. Gasification and chemical-looping combustion of a lignite char in a fluidized bed of iron oxide. Energy Fuels 2010, 24, 3034−3048. (18) Mattisson, T.; Leion, H.; Lyngfelt, A. Chemical-looping with oxygen uncoupling using CuO/ZrO(2) with petroleum coke. Fuel 2009, 88 (4), 683−690. (19) Eyring, E.; Konya, G.; Lighty, J.; Sahir, A.; Sarofim, A.; Whitty, K. Chemical looping with copper oxide as carrier and coal as fuel. Oil Gas Sci. Technol. 2011, 66, 209−221. (20) Azimi, G.; Leion, H.; Mattisson, T.; Lyngfelt, A. Chemicallooping with oxygen uncoupling using combined Mn−Fe oxides, testing in batch fluidized bed. Energy Procedia 2011, 4, 370−377. (21) Berguerand, N.; Lyngfelt, A. Design and pperation of a 10 kWth chemical-looping combustor for solid fuelsTesting with South African coal. Fuel 2008, 87 (12), 2713−2726. (22) Shen, L.; Wu, J.; Gao, Z.; Xiao, J. Reactivity deterioration of NiO/Al2O3 oxygen carrier for chemical looping combustion of coal in a 10 kWth reactor. Combust. Flame 2009, 156 (7), 1377−1385. (23) Cuadrat, A.; Abad, A.; García-Labiano, F.; Gayán, P.; de Diego, L. F.; Adánez, J. The use of ilmenite as oxygen-carrier in a 500 Wth chemical-looping coal combustion unit. Int. J. Greenhouse Gas Control 2011, 5 (6), 1630−1642. (24) Adánez-Rubio, I.; Gayán, P.; Abad, A.; García-Labiano, F.; de Diego, L. F.; Adánez, J., CO2 capture in coal combustion by chemicallooping with oxygen uncoupling (CLOU) with a Cu-based oxygencarrier. In Proceedings of the 5th International Conference on Clean Coal Technologies (CCT2011), Zaragoza, Spain, 2011. (25) Rydén, M.; Lyngfelt, A.; Mattisson, T. Combined manganese/ iron oxides as oxygen carrier for chemical looping combustion with oxygen uncoupling (CLOU) in a circulating fluidized bed reactor system. Energy Procedia 2011, 4, 341−348. (26) Mattisson, T.; Leion, H.; Lyngfelt, A. Chemical-looping with oxygen uncoupling using CuO/ZrO2 with petroleum coke. Fuel 2009, 88, 683−690. (27) Leion, H.; Mattisson, T.; Lyngfelt, A. Using chemical-looping with oxygen uncoupling (CLOU) for combustion of six different solid fuels. Energy Procedia 2009, 1 (1), 447−453. (28) Shulman, A.; Cleverstam, E.; Mattisson, T.; Lyngfelt, A. Manganese/Iron, manganese/nickel, and manganese/silicon oxides used in chemical-looping with oxygen uncoupling (CLOU) for combustion of methane. Energy Fuels 2009, 23, 5269−5275. (29) Shulman, A.; Cleverstam, E.; Mattisson, T.; Lyngfelt, A. Chemical-Looping with oxygen uncoupling using Mn/Mg-based oxygen carriersOxygen release and reactivity with methane. Fuel 2011, 90 (3), 941−950. (30) Leion, A.; Larring, Y.; Bakken, E.; Bredesen, R. Use of CaMn0.875Ti0.125O3 as oxygen carrier in chemical-looping with oxygen uncoupling. Energy Fuels 2009, 23, 5276−5283. (31) Bakken, E.; Norby, T.; Stølen, S. Nonstoichiometry and reductive decomposition of CaMnO3−δ. Solid State Ionics 2005, 176 (1−2), 217−223. (32) Rydén, M.; Lyngfelt, A.; Mattisson, T. CaMn0.875Ti0.125O3 as oxygen carrier for chemical-looping combustion with oxygen uncoupling (CLOU)Experiments in a continuously operating fluidized bed reactor system. Int. J. Greenhouse Gas Control 2011, 5, 356−366. (33) Crum, J. V.; Riley, B. J.; Vienna, J. D. Binary phase diagram of the manganese oxide−iron oxide system. J. Am. Ceram. Soc. 2009, 92, 2378−2384. (34) Wickham, D. The chemical composition of spinels in the system Fe3O4−Mn3O4. J. Inorg. Nucl. Chem. 1969, 31, 313−320.

to Fe. The oxygen carriers with manganese content of 67%, 75%, and 80% calcined at 950 °C have almost full conversion of methane to CO2 and H2O at 850 °C.



AUTHOR INFORMATION

Corresponding Author

*Tel.: +46 31 7722887. Fax: +46 31 7722853. E-mail: golnar. [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This study was financed by the Swedish Energy Agency, Project No. 32368-1. REFERENCES

(1) Lyngfelt, A.; Leckner, B.; Mattisson, T. A fluidized-bed combustion process with inherent CO2 separation: application of chemical-looping combustion. Chem. Eng. Sci. 2001, 56 (10), 3101− 3113. (2) Mattisson, T.; Lyngfelt, A.; Leion, H. Chemical-looping with oxygen uncoupling for combustion of solid fuels. Int. J. Greenhouse Gas Control 2009, 3 (1), 11−19. (3) Lewis, W. K.; Gilliland, E. R.Production of pure carbon dioxide. U.S. Patent No. 2665971, 1954. (4) Richter, H. J.; Knoche, K. F. Reversibility of combustion process . efficiency and Costing, Second Law Analysis of ProcessACS Symposium series 235; American Chemical Society: Washington, DC, 1983; pp 71−85 (5) Ishida, M.; Jin, H. A new advanced power-generation system using chemical-looping combustion. Energy (Oxford, United Kingdom) 1994, 19 (4), 415−422. (6) Lyngfelt, A. Oxygen carriers for chemical looping combustion 4000 h of operational experience. Oil Gas Sci. Technol. 2011, 66 (2), 161−172. (7) Adanez, J.; Abad, A.; Garcia-Labiano, F.; Gayan, P.; Diego, L. F. Progress in chemical-looping combustion and reforming technologies. Prog. Energy Combust. Sci. 2012, 38, 215−282. (8) Lyngfelt, A.; Thunman, H. Construction and 100 h of operational experience of a 10-kW chemical-looping combustor. Carbon Dioxide Capture for Storage in Deep Geologic FormationsResults from the CO2 Capture Project; 2005; Vol. 1, Chapter 36, pp 625−645. (9) Ryu, H.-J.; Jin, G.-T.; Bae, D.-H.; Yi, C.-K., Continuous operation of a 50 kWth chemical-looping combustor: Long-term operation with Ni- and Co-based oxygen carrier particles. Presented at the 5th China− Korea Joint Workshop on Clean Energy Technology, Qingdao University, China, Oct. 25−28, 2004, pp. 221−230. (10) Abad, A.; Mattisson, T.; Lyngfelt, A.; Ryden, M. Chemicallooping combustion in a 300 W continuously operating reactor system using a manganese-based oxygen carrier. Fuel 2006, 85 (9), 1174− 1185. (11) Adanez, J.; Gayan, P.; Celaya, J.; de Diego, L. F.; GarciaLabiano, F.; Abad, A. Chemical looping combustion in a 10 kWth prototype using a CuO/Al2O3 oxygen carrier: Effect of operating conditions on methane combustion. Ind. Eng. Chem. Res. 2006, 45, 6075−6080. (12) Johansson, E.; Mattisson, T.; Lyngfelt, A.; Thunman, H. A 300 W laboratory reactor system for chemical-looping combustion with particle circulation. Fuel 2006, 85 (10−11), 1428−1438. (13) de Diego, L. F.; Garcia-Labiano, F.; Gayan, P.; Celaya, J.; Palacios, J. M.; Adanez, J. Operation of a 10 kWth chemical-looping combustor during 200 h with a CuO-Al2O3 oxygen carrier. Fuel 2007, 86, 1036−1045. (14) Linderholm, C.; Abad, A.; Mattisson, T.; Lyngfelt, A. 160 h of chemical-looping combustion in a 10 kW reactor system with a NiObased oxygen carrier. Int. J. Greenhouse Gas Control 2008, 2 (4), 520− 530. 376

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377

Energy & Fuels

Article

(35) Azimi, G.; Rydén, M.; Leion, H.; Mattisson, T.; Lyngfelt, A. MnzFe1−z)yOx combined oxides as oxygen carrier for chemical-looping with oxygen uncoupling. AIChE J. 2012, DOI: 10.1002/aic.13847. (36) Ksepko, E.; Siriwardane, R.; Tian, H.; Simonyi, T.; Poston, J.; Sciazko, M., Effect of H2S on chemical-looping combustion of coalderived synthesis gas over Fe2O3−MnO2 supported on ZrO2/sepiolite. In 1st International Conference on Chemical Looping, Lyon, France, 2010. (37) Lambert, A.; Delquié, C.; Clémeneçon, I.; Comtea, E.; Lefebvrea, V.; Rousseau, J.; Durand, B. Synthesis and characterization of bimetallic Fe/Mn oxides for chemical looping combustion. Energy Procedia 2009, 1 (1), 375−381. (38) Rydén, M.; Lyngfelt, A.; Mattisson, T. Combined manganese/ iron oxides as oxygen carrier for chemical looping combustion with oxygen uncoupling (CLOU) in a circulating fluidized bed reactor system. Energy Procedia 2011, 4, 341−348. (39) Kjellqvist, L.; Selleby, M. Thermodynamic assessment of the Fe−Mn−O System. J. Phase Equilib. Diffus. 2010, 31 (2), 113−134. (40) Muan, A.; So̅miya, S. The system of iron oxide−manganese oxide in air. Am. J. Sci. 1962, 260, 230−240. (41) Zafar, Q.; Abad, A.; Mattisson, T.; Gevert, B.; Strand, M. Reduction and oxidation kinetics of Mn3O4/Mg−ZrO2 oxygen carrier particles for chemical-looping combustion. Chem. Eng. Sci. 2007, 62, 6556−6567. (42) Jerndal, E.; Mattisson, T.; Thijs, I.; Snijkers, F.; Lyngfelt, A. Investigation of NiO/NiAl2O4 oxygen carriers for chemical-looping combustion produced by spray-drying. Int. J. Greenhouse Gas Control 2010, 4, 23−35. (43) Rydén, M.; Leion, H.; Mattisson, T.; Lyngfelt, A. Combined oxides as oxygen carrier material for chemical-looping with oxygen uncoupling. In 2nd International Conference on Chemical Looping, Darmstadt, Germany, 26−28 Sept. 2012. (44) Linderholm, C.; Mattisson, T.; Lyngfelt, A. Long-term integrity testing of spray-dried particles in a 10 kW chemical-looping combustor using natural gas as fuel. Fuel 2009, 88, 2083−2096.

377

dx.doi.org/10.1021/ef301120r | Energy Fuels 2013, 27, 367−377