- Email: [email protected]
Experimental study of copper modified manganese ores as oxygen carriers in a dual fluidized bed reactor
Applied Energy 162 (2016) 940–947
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Experimental study of copper modified manganese ores as oxygen carriers in a dual fluidized bed reactor Lei Xu, Hongming Sun, Zhenshan Li ⇑, Ningsheng Cai Key Laboratory for Thermal Science and Power Engineering of Ministry of Education, Beijing Municipal Key Laboratory for CO2 Utilization & Reduction, Department of Thermal Engineering, Tsinghua University, Beijing 100084, China
h i g h l i g h t s Reactivity of manganese ore was significantly improved by impregnating copper. Two 88 h continuous run were performed using both oxygen carriers. Attrition data was obtained for both materials based on long operations.
a r t i c l e
i n f o
Article history: Received 28 August 2015 Received in revised form 13 October 2015 Accepted 29 October 2015
Keywords: Manganese ore Chemical looping combustion Oxygen carrier
a b s t r a c t Chemical-looping combustion (CLC) is a developing CO2 capture technology. CLC makes use of the repeated oxidation/reduction reactions of metal oxide (oxygen carrier, OC) to separate CO2 from fuel combustion and to obtain a pure CO2 stream suitable for compression and storage. Low cost materials, such as natural ores, are required for coal-fueled CLC because the lifetime of the oxygen carrier (OC) is lowered by side reactions with the fuel ash or carryover losses. In this study, five manganese ores were examined as oxygen carriers using CO as the fuel gas in a laboratory batch fluidized bed reactor. All five of the ores were impregnated by copper nitrate solution to evaluate the reactivity enhancement of copper impregnation. The period with full CO conversion can be enhanced 2–100 times for different ores in the single fluidized bed test, which indicated that the Cu impregnation may be a general method to enhance the reactivity of manganese ores. Finally, one manganese ore and the corresponding Cu-modified particles were tested in a dual fluidized bed reactor. The attrition rates of both materials were measured as 0.13 wt.%/h during the 88 h operation in the dual fluidized bed. Both the manganese ore and the Cuimpregnated ore exhibited stable and high reactivity during the continuous test in the dual fluidized bed reactor, even at a low temperature (310 °C). Copper impregnation had no obvious influence on the attrition property of the manganese ore. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Chemical looping combustion (CLC) is a promising technology for inherent CO2 separation at a low cost [1]. CLC generally consists of two reaction steps: reduction and oxidation. During the reduction step, the fuel is oxidized by the oxygen provided by the reduction of a solid oxygen carrier, typically a metal oxide. In the oxidation step, the reduced oxygen carrier is re-oxidized by air. These two steps are commonly performed in two separate reactors: the fuel reactor and the air reactor, respectively. Both reactors are connected by a circulating stream of solid oxygen carrier. The gases from both reactors never mix, and a stream of H2O and ⇑ Corresponding author. Tel./fax: +86 10 62789955. E-mail address: [email protected] (Z. Li). http://dx.doi.org/10.1016/j.apenergy.2015.10.167 0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.
CO2 is obtained at the outlet of the fuel reactor, which stays separated from the air nitrogen. The principle of using a solid oxygen carrier material to provide oxygen for thermochemical fuel conversion was proposed in the 1940s [2] to reform methane. The idea of CLC was patented as a process for CO2 production from carbonaceous sources in the 1950s [3]. The same principle was proposed during the 1980s with the goal to lower the irreversibility of combustion [4,5]. With the increasing recognition of anthropogenic climate change, the concept was investigated to recover CO2 from the exhaust of combustion systems for power generation [6,7]. Since then, obtaining CO2 capture with a low energy penalty for the power sector has been the main driving force for CLC research. Gaseous fuels [8–10], such as natural gas or syngas, and solid fuels [11–17], such as coal or biomass, have been widely
941
L. Xu et al. / Applied Energy 162 (2016) 940–947
investigated in the literature as possible fuels for CLC. The application of gaseous fuels is less problematic, whereas for solid fuels, losses of oxygen carriers due to possible side reaction with ash and impurities in the gas stream [18], as well as material withdrawn from the reactor with the ash, may lead to high expected amounts of oxygen carrier makeup [17]. This increases the demand for abundant, low cost and environmental friendly oxygen carrier materials [13]. Possible materials under investigation include natural ores [19–23] and industrial residues or by-products [24–27]. Among the natural ores, ilmenite has been most widely investigated [20,28–31]. Most recently, Arjmand et al. [12,32] investigated several manganese ores as oxygen carriers for CLC, and these manganese ores showed better reactivity than ilmenite. This suggests that manganese ore could be a better choice than ilmenite as an oxygen carrier. It is generally accepted that low cost oxygen carriers, as mentioned above, tend to have lower conversion rates. Low reactivity of the oxygen carrier directly leads to an increase of the required solids inventory [33]. This in turn leads to a larger pressure drop within the reactor and inevitably to increased power consumption to maintain the solid fluidization [34]. Therefore, it is important to search for oxygen carriers with high reactivity or for methods allowing increased reactivity of the existing low cost oxygen carriers. Previous research has found that the reactivity of manganese ore can be enhanced by introducing a very small amount of copper ions [35]. The redox stability of the Cu-modified manganese ore and the effects of the loading amount on the reduction reactivity of Cu-modified manganese ore were investigated in a single fluidized bed reactor [35]. However, the more practical use of the oxygen carrier is in a dual fluidized bed reactor in continuous operation. Additionally, more knowledge about the copper impregnation method for improving the manganese ore reactivity would be useful. Accordingly, in this study, five manganese ores were impregnated by copper nitrite solution to examine the effect of the impregnating copper on different manganese ores. Furthermore, both the manganese ore and the Cu-modified manganese ore were evaluated in a dual fluidized bed reactor to examine the long-term chemical stability and attrition properties of the oxygen carriers.
mass ratio of the copper ion to manganese ore was 2 wt.% and was controlled during preparation. 2.2. Characterization of oxygen carriers The composition of the manganese ore was analyzed by X-ray fluorescence (XRF), and the results are summarized in Table 1. The corresponding mass fraction of single chemical element can be calculated based on the molecular formula. For example, the mass fraction of Mn can be calculated as 40.45%/71⁄55 = 31.33%. The BET (Brunauer, Emmett and Teller) surface area and BJH (Barrett, Joyner and Halenda) pore volume of the particles were measured with a Micromeritics micropore analyzer (Autosorb-iQ2MP, NOVA4000). The crushing strength for all particles was determined using an HP-10 digital force gauge. The value given is the average force in N from 30 individual measurements needed to crush a particle in the size fraction of 125–300 lm. The size distribution was measured by a laser particle size analyzer (Mastersizer 2000). The particles were observed under a scanning electron microscope (SEM, JSM-7001F). The bulk density, BET surface area, BJH pore volume and crushing strength of the particles are listed in Table 2. 2.3. Test of OC in the single fluidized bed reactor All oxygen carriers were tested in a single fluidized bed. The reactor was made of quartz tube (i.d. 30 mm) and was heated by an electric furnace. The bed material for all the tests was 30 g with a size of 125–300 lm placed on the porous distributor. The flow rates of the fluidizing gases for both reduction and oxidation were 2 L/min (STP). Both the reduction and oxidation temperature were 800 °C for the redox cycles. The gas velocity was 0.17 m/s at 800 °C, and the fluidization number U/Umf was 3.3, where Umf was calculated on the basis of the relations by Kunii and Levenspiel [36] with the mean radius of the particle as 212.5 lm. For each test, the bed material was first heated to 800 °C in 10 vol% O2 (90 vol% N2). Then, the reducing gas was introduced. The reducing agent was 10 vol% CO (90 vol% N2). The oxidizing agent was 10 vol% O2 (N2 90 vol%). Inert gas (N2) was introduced between the reduction and the oxidation period. The product gases were introduced into the gas analyzer after passing through a filter.
2. Experimental details 2.4. Test of OC in the dual fluidized bed reactor 2.1. Preparation of oxygen carriers The five natural manganese ores (designated as Mn1–Mn5) were provided by Mugui Manganese Ore Company from Guangxi Province in China. The ores were first crushed and then calcined in a muffle oven for 4 h at 950 °C and sieved to 125–300 lm. The Cu-modified manganese ores (designated as Mn1Cu2%–Mn5Cu2%) were prepared by impregnation of copper nitrate solution. For every 50 g of manganese ore, solid Cu(NO3)23H2O (3.8 g) was used as a precursors and was dissolved in water (15 ml) to give the solution. The ore particles were soaked in the solution for 6 h at ambient temperature and were then calcined in air for 2 h at 500 °C. After cooling, all of the particles were sieved to 125–300 lm. The
The continuous test was performed in a dual fluidized bed reactor, as illustrated in Fig. 1. The system consisted of an air reactor (AR), a fuel reactor (FR), a riser, a cyclone and two loop seals (LS). The riser was used to transport oxygen carriers from the air reactor to the fuel reactor. The solid circulation rate was controlled by the lower loop seal. The fuel reactor was a bubbling fluidized bed with a length of 1.3 m. The inner diameters of both the air reactor and the fuel reactor were 50 mm. The inner diameter of the riser was 20 mm. The height of the riser was 3.3 m. The secondary air was introduced at the lower part of the riser to assist particle entrainment. The air and fuel reactors were heated by electric furnaces. Ktype thermocouples were inserted into the reactor to monitor the
Table 1 Chemical analysis of the manganese ores. wt.%
MnO
Fe2O3
SiO2
Al2O3
BaO
K2O
P2O5
TiO2
MgO
CaO
Mn1 Mn2 Mn3 Mn4 Mn5
40.45 54.13 73.37 39.41 64.75
7.36 24.71 10.23 11.24 16.53
44.09 10.54 7.12 32.94 5.82
1.68 7.80 4.01 13.09 2.33
4.71 0.09 0.65 0.24 0.50
0.23 1.45 0.58 1.33 0.90
0.44 – 0.22 0.29 0.06
– 0.30 0.27 0.56 8.50
0.18 0.46 0.31 0.20 0.13
0.36 0.21 1.61 0.22 0.07
942
L. Xu et al. / Applied Energy 162 (2016) 940–947
Table 2 Physical properties and characteristics of the oxygen carriers used in this work. Sample
Bulk density (g/cm3)
BET surface area (m2/g)
BJH pore volume (cm3/g)
Crushing strength (N)
Mn1/Mn1Cu2% Mn2/Mn2Cu2% Mn3/Mn3Cu2% Mn4/Mn4Cu2% Mn5/Mn5Cu2%
1.71/1.72 0.94/0.93 1.45/1.45 1.28/1.32 1.67/1.69
2.82/3.64 6.32/6.39 1.70/2.34 3.21/4.75 2.34/3.42
0.016/0.204 0.029/0.031 0.007/0.010 0.020/0.024 0.013/0.012
2.0/1.9 ± 0.4 0.8/0.8 ± 0.1 1.8/1.7 ± 0.3 1.7/1.7 ± 0.4 1.9/1.9 ± 0.3
has been used for testing ilmenite and cement-supported CuO oxygen carriers in our previous studies, and more details about the reactor system can be found in the relevant publications [37,38]. Four continuous tests were conducted in the dual fluidized bed reactor. The first two were 88 h run using Mn5 and Mn5Cu2% as the oxygen carriers separately in the same operation conditions. The initial solid inventory of the system was 5000 g, and 10 vol% CO (90 vol% N2) as the reducing agent and air as the oxidizing agent for both tests. N2 was used as the fluidizing gas for the loop seals. The fuel reactor was maintained at 950 °C and the air reactor was maintained at 900 °C. The operation conditions are summarized in Table 3. The third continuous test lasted 200 min using Mn5 as the oxygen carrier at different operation conditions and with different concentrations of CO as fuel. The fourth continuous test also lasted 200 min using Mn5Cu2% as the oxygen carrier at the same operation conditions as the third one. The initial solid inventory of the system was 3500 g for both the third and fourth continuous tests. The reactor system was designed to work stably in a solid circulation rate range of 23–76 kg/h. In the first two operations, a solid circulation rate of 23 kg/h was used. In the third and fourth operations, a solid circulation rate of 76 kg/h was used. The detailed operation conditions are summarized in Table 4. 2.5. Data evaluation
Fig. 1. Schematic diagram of the dual fluidized bed reactor for CLC.
temperature. Fine particles from both reactors were collected by filters. O1–O8 and I0–I2 were pressure measurement points used to monitor the solid circulation of the entire loop. The reactor
CO conversion (X) and oxygen carrier-to-fuel ratio (u) are two parameters used to evaluate the result of the dual fluidized bed tests. The CO conversion was calculated according to Eq. (1),
X¼
C CO2 C CO2 þ C CO
ð1Þ
Table 3 Gas flow rates (LN/min, STP) at different locations during the first two continuous operations. Location
Fluidizing gas
Gas flow rates (LN/min)
Gas velocity (m/s)
U/Umf
AR Riser FR Lower LS Upper LS
Air Air 10 vol% CO N2 N2
4 30 4 2 2
0.15 5 0.15 0.07 0.07
2.9 94 2.8 1.3 1.3
Table 4 Fuel gas flow rates (LN/min, STP) and CO conversion at the different conditions in Figs. 5 and 6.
1 2 3 4 5 6 7 8
N2
CO
Fuel power
Temperature
(LN/min)
(LN/min)
(W)
(°C)
3.6 3 2.5 2 2 1 0 4
0.4 1 1.5 2 2 3 4 0
85 213 319 425 425 638 850
305–315 305–315 305–315 305–315 400–410 500–510 500–510
u 108 43 29 22 22 14 11
X Mn
Mn5Cu2%
0.98 0.91 0.91 0.87 0.97 1 1
1 1 1 1 1 1 1
L. Xu et al. / Applied Energy 162 (2016) 940–947
where C CO2 and C CO are the gas concentrations detected by the gas analyzer. The oxygen carrier-to-fuel ratio (u), defined by Eq. (2), was used to describe the state of solid circulation and fuel feed rate during the operation.
u¼
F Mn2 O3 3F CO
ð2Þ
where F Mn2 O3 is the molar flow rate of Mn2O3, which is calculated from the solid circulation rate and the ore composition, assuming that the manganese oxide exists in the form of Mn2O3 before reduction (u = 1 corresponds to the stoichiometric Mn2O3 amount required for the full conversion of CO to CO2 and the full reduction of Mn2O3 to Mn3O4). 3. Results and discussion 3.1. Test of Cu-modified manganese ore in the single fluidized bed reactor The performance of oxygen carriers was investigated in a single fluidized bed using CO as fuel. All of the oxygen carriers showed similar profiles of outlet CO and CO2. CO was fully converted at the beginning of each redox cycle. CO2 was detected immediately after CO was introduced into the reactor. As the reaction proceeded, oxygen carriers were consumed and CO started to come out. The period with full CO conversion is defined as breakthrough time, as illustrated in Fig. 2a. For example, the breakthrough time of Mn1Cu2% was 426 s for the second cycle. The reactivity of oxygen carriers in the first cycle was higher than the reactivity in the successive cycles. However, the reactivity decay for the successive cycles was very slow and tended to be stable. Therefore, the breakthrough time of the second cycle was used to compare the reactivity of the oxygen carriers and is summarized in Fig. 2b. The copper addition enhanced the reactivity of all five ores. However, the extent to which the reactivity was improved was different from Mn1 to Mn5. The breakthrough time was enhanced by less than 2 times for Mn1 but by 100 times for Mn5. Before impregnation, Mn1 showed the best reactivity, whereas Mn2, Mn4 and Mn5 showed relatively lower reactivity. After impregnation, the reactivity of Mn2, Mn4 and Mn5 was enhanced by 8–100 times. Therefore, although only five ores were tested in our study, it can be postulated that copper impregnation could be an effective method to improve the reactivity of low-reactive manganese ores. Fig. 3 shows that impregnation with copper can enhance the surface area of the particles. The enhancement of the surface area contributed to the higher reactivity of the particles. Pore volume was also enhanced by copper impregnation, except for Mn5. Increased pore volume facilitated the diffusion of reactant gas
943
inside the particles and thus increased the overall reactivity. However, neither the increase of the surface area nor the pore volume was as significant as the enhancement of the breakthrough time, especially for Mn2, Mn4 and Mn5. This result suggests that the contribution of copper to the reactivity enhancement was mainly due to the interaction of copper with manganese oxide, which is consistent with our previous study [35].
3.2. Continuous operation using Mn5 and Mn5Cu2% as oxygen carriers Mn5 and Mn5Cu2% were used for the continuous tests in the dual fluidized bed. No sintering and agglomeration were found during the operations. The reactor system was run for 88 hours continuously for each test. Fig. 4 shows the outlet gas composition and pressure drop of the fuel reactor during the continuous run. In the first 7 h, the whole reactor system was heating, and the fuel reactor was fluidized by N2. After the fuel reactor and the air reactor reached the target temperatures, 950 °C and 900 °C, respectively, the fuel gas was switched to 10 vol% CO. The corresponding fuel power is 85 W, with a specific inventory of 59 kg/kW. The operation conditions were then kept constant, as summarized in Table 3. The solid circulation rate was 23 kg/h, calculated according to Eq. (1) in previous work [37], i.e. 20 kg/m2 s based on the cross-section area of riser. Both materials presented similar results during the continuous run. The detected gas concentrations of CO and CO2 were mixed with N2 from the loop seals. Generally, the off-gas from FR was mainly diluted with N2 from the upper loop seal and the off-gas from AR was mainly diluted with N2 from the lower loop seal [37]. The CO2 concentration varied between 6.2 vol% and 7 vol% during the whole operation. No CO was detected, which meant that CO was fully converted. The result shows that the manganese ore has much higher reactivity than ilmenite. Ilmenite only reached a conversion of 60% in the same reactor system at similar conditions [37]. The pressure drop through the fuel reactor, denoted as DP, decreased rapidly in the first 20 h and was then stable. During the first 10 h, the pressure drop was beyond the gauge range; therefore, it appears as a straight line in Fig. 4. From the 10th hour to the 19th hour, DP dropped at a rate of 470 Pa/h. From the 19th hour to the end, DP dropped only 300 Pa. The rapid decline of DP during the first 19 h was probably due to the rapid attrition and fragmentation rate of the particles. Mn5 showed good and stable reactivity during the 88 h continuous run, especially compared to the results from ilmenite [37]. Considering the relatively low reactivity of Mn5 among the five manganese ores, it could be speculated that manganese ores are much better oxygen carriers than ilmenite with respect to reactivity. However, the mechanical properties of manganese ores are worse than ilmenite. The mechanical properties of both materials are discussed in detail in Section 3.4.
Fig. 2. Outlet CO and CO2 profile of Mn1 in the single fluidized bed test (a) and breakthrough time comparison of the 2nd cycle of ten oxygen carriers (b).
944
L. Xu et al. / Applied Energy 162 (2016) 940–947
Fig. 3. Change of the BET surface area (a) and BJH pore volume (b) of five manganese ores by impregnating with copper.
Fig. 5. Flue gas concentrations of the fuel reactor as a function of time with changing operation conditions and the manganese ore as the oxygen carrier.
nents from CLC reactors [38]. The upper loop seal used N2 with a flow rate of 4 LN/min. The lower loop seal used N2 with a flow rate of 3.5 LN/min. The air reactor was fluidized by air with a flow rate of 5.5 LN/min, and the flow rate of the secondary air was 45 LN/min. The flow rate and gas composition of the fuel gases were varied during the two operations. The whole period for each run was divided into eight stages, and the fuel gas flow rates, temperatures and CO conversions in each stage are summarized in Table 4. Using Mn5 as the oxygen carrier, the CO conversion reached 0.98 at 310 °C when the CO flow rate was 0.4 LN/min (u 108). The conversion dropped to 0.87 when the CO flow rate increased to 2 LN/min (u 22). When the temperature increased to 400 °C, the conversion increased to 0.97. At 500 °C, CO was fully converted, even after the CO flow rate increased to 4 LN/min (u 11). The Cuimpregnated manganese ore showed much better reactivity than the original ore. CO was almost fully converted in the same operation conditions, as can be seen in Table 4 and Fig. 6.
3.4. Fragmentation and attrition of the oxygen carriers Oxygen carriers are continuously lost in the fluidized bed reactor due to abrasive wear or attrition and fragmentation resulting
Fig. 4. Flue gas concentrations and pressure drop of the fuel reactor versus time during the 88 h continuous operation with Mn5 (a) and (b), and Mn5Cu2% (c) and (d).
3.3. Comparison of manganese ore and Cu-modified manganese ore in dual fluidized bed reactor Continuous operations of Mn5 and Mn5Cu2% were performed at lower temperatures in the dual fluidized bed reactor for two reasons. One is that the reactivity difference of both materials can be distinguished at lower temperatures in the reactor system, the other reason is that oxygen carriers could have potential use at lower temperatures for eliminating the unburnt gaseous compo-
Fig. 6. Flue gas concentrations of the fuel reactor as a function of time with changing operation conditions and the Cu-modified manganese ore as the oxygen carrier.
945
L. Xu et al. / Applied Energy 162 (2016) 940–947
Fig. 7. Particle size distribution of Mn5 and Mn5Cu2% before and after the continuous test in the reactor and filter.
from particle–particle interactions and particle–wall interactions. Resistance to attrition and fragmentation is an important property of particles used as oxygen carriers. The particle size distribution, crushing strength and attrition rate were used to evaluate the mechanical properties of the oxygen carriers. The particle size distributions of fresh Mn5 and Mn5Cu2% particles remaining in the reactor after the 88 h operation and the particles in the AR filter are presented in Fig. 7. Both materials presented a similar attrition property. This suggested that impregnation by 2 wt.% copper did not have an obvious effect on the attrition property of manganese ore. The size distribution of both materials were narrower, and the average size were smaller after the test. The d(0.9) of Mn5 in the reactor decreased from 508 lm to 349 lm after the test while 505–341 lm for Mn5Cu2%. The crushing strengths of the ten oxygen carriers were also tested to evaluate the mechanical properties of the materials. A crushing strength larger than 2 N is considered sufficiently hard according to a study by Rydén et al. [39]. Four ores (Mn1, Mn3, Mn4 and M5) showed similar crushing strengths, ranging from 1.7 to 2.0 N, and Mn2 showed a much smaller crushing strength of 0.8 N, as illustrated in Table 2. The attrition rate is another important parameter used for estimating the mechanical properties of oxygen carriers. The attrition rate can be evaluated using a parameter called the loss of fines, Lf, which is defined as
Lf ¼
Dmfines 1 100 Dt mtotal
Particles of Mn5 elutriated from the dual fluidized bed reactor during the 88 h operation were collected at different times (10 h, 24 h, 48 h, 72 h, and the end of the operation) from the filters and fine particles collector and were weighed to determine Lf. The loss of fines for the oxygen carriers was assumed to be those with size of <125 lm as determined by passing through a sieve. The mass of the particles collected in the AR filter (1610 g) was much greater than that in the FR filter (190 g), as shown in Table 5. Moreover, the amount of particles with size of <125 lm in the AR filter (1170 g) was much greater than that in the FR filter (11 g). This suggests that attrition and fragmentation of particles mainly occurred in the riser and cyclone rather than the other parts of the reactor system, such as the air reactor and the fuel reactor. This is reasonable because the much higher gas velocity in the riser due to the introduction of secondary air and the swirling flow inside the cyclone would lead to fiercer collisions between particles and impacts of particles against the reactor wall. During the operation, the loss of fine particles was high at the beginning and rapidly decreased, as shown in Fig. 8. In the first 10 h, the Lf was 1.17 wt.%/h for Mn5 and 1.28 wt.%/h for Mn5Cu2%. The Lf decreased rapidly to 0.34 wt.%/h for Mn5 and 0.36 wt.%/h for Mn5Cu2% in the next 14 h. Finally, it dropped to 0.13 wt.%/h for both materials and stabilized. The same trend was also found in the study of a CuO/Al2O3 oxygen carrier [8]. The high attrition rate at the beginning was partially attributed to the rounding effect, which means the sharp edges of the particles are more easily scraped, and the shape of the particle becomes closer and closer to a sphere. Fig. 9 shows that the particles after the 88 h operation were not as sharp as the fresh particles. Another possible reason could be that the particles became more robust as the reaction went on. This was indicated by the higher crushing strength of the particles after the 88 h operation, which was 2.1 N for each material. It can be concluded that the ore particles had a high attrition rate in the beginning and a much lower attrition rate after tens of hours; therefore, the lifetime is estimated as 770 h according to the Lf of 0.13 wt.%/h.
ð3Þ
where Dmfines is the amount of fines produced during the period of time, Dt, and mtotal is the total amount of materials input before the operation, i.e., 5000 g in this study. The lifetime of the oxygen carrier, t life , is calculated as
t life ¼
1 Lf
ð4Þ
Fig. 8. Loss of fines of Mn5 and Mn5Cu2% versus time during the 88 h operation.
Table 5 Mass distribution in the system for Mn5 and Mn5Cu2% after the 88 h tests. Sample
Mn5 Mn5Cu2%
Reactor (g)
2813 2766
FR filter (g)
AR filter (g)
<125 lm
>125 lm
<125 lm
>125 lm
11 15
179 176
1170 1165
440 470
Collector (g)
Total (g)
Lifetime (h)
136 142
5000 5000
770 770
946
L. Xu et al. / Applied Energy 162 (2016) 940–947
Fig. 9. SEM images of the Mn5 particles before (a) and after (b) the 88 h operation.
4. Conclusions Wet impregnation of manganese ore by copper nitrate solution enhanced the reactivity of manganese ore as an oxygen carrier. Five manganese ores and five copper impregnated ones were examined as oxygen carriers using CO as the fuel gas in a laboratory batch fluidized bed reactor. The breakthrough time of CO was enhanced 2–100 times on different ores by impregnating with 2 wt.% Cu. Four manganese ores presented similar crushing strengths between 1.7 and 2.0 N. Mn5 was chosen to perform an 88 h run in a dual fluidized bed to evaluate the reactivity, fragmentation and attrition properties of the material. The manganese ore showed much higher reactivity than ilmenite. Both the manganese ore and the Cu-impregnated ore showed stable and high reactivity during the continuous test in the dual fluidized bed reactor, even at a low temperature (310 °C). The attrition rate was high in the beginning of the 88 h operation and became stable at 0.13 wt. %/h after 24 h. The lifetimes of both the manganese ore and Cumodified manganese ore are estimated as 770 h. Manganese ores are good candidates as oxygen carriers for chemical looping combustion because of the higher reactivity compared to ilmenite. Copper impregnation can enhance the reactivity of manganese ore, especially those ores with a low reactivity. Acknowledgments This work was supported by the National Natural Science Foundation of China (51376105, 91434124), the National Key Basic Research and Development Program (2011CB707301), the Tsinghua University Initiative Scientific Research Program, and the Program for New Century Excellent Talents in University (NCET-120304). References [1] Ekström C, Schwendig F, Biede O, Franco F, Haupt G, de Koeijer G, et al. Technoeconomic evaluations and benchmarking of pre-combustion CO2 capture and oxy-fuel processes developed in the European ENCAP project. Energy Proc 2009;1:4233–40. [2] Lewis WK, Gilliland ER, Reed WA. Reaction of methane with copper oxide in a fluidized bed. Ind Eng Chem 1949;41:1227–37. [3] Lewis WK. Production of pure carbon dioxide. U.S. Patent 2,665,972. 1954-112. [4] Knoche KF, Richter H. Verbesserung der reversibilität von verbrennungsprozessen. Brennstoff-Wärme-Kraft 1968;20:205–11. [5] Richter HJ, Knoche KF. Reversibility of combustion processes. ACS Symp Ser 1983;235:71–86. [6] Ishida M, Zheng D, Akehata T. Evaluation of a chemical-looping-combustion power-generation system by graphic exergy analysis. Energy 1987;12:147–54. [7] Ishida M, Jin H. A new advanced power-generation system using chemicallooping combustion. Energy 1994;19:415–22.
[8] Adánez J, Gayán P, Celaya J, de Diego LF, García-Labiano 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;45:6075–80. [9] Jin H, Ishida M. Reactivity study on natural-gas-fueled chemical-looping combustion by a fixed-bed reactor. Ind Eng Chem Res 2002;41:4004–7. [10] Mattisson T, Garcia-Labiano F, Kronberger B, Lyngfelt A, Adánez J, Hofbauer H. Chemical-looping combustion using syngas as fuel. Int J Greenhouse Gas Control 2007;1:158–69. [11] Lyngfelt A. Chemical-looping combustion of solid fuels – status of development. Appl Energy 2014;113:1869–73. [12] Arjmand M, Leion H, Mattisson T, Lyngfelt A. Investigation of different manganese ores as oxygen carriers in chemical-looping combustion (CLC) for solid fuels. Appl Energy 2014;113:1883–94. [13] Leion H, Mattisson T, Lyngfelt A. Solid fuels in chemical-looping combustion. Int J Greenhouse Gas Control 2008;2:180–93. [14] Shen L, Wu J, Xiao J, Song Q, Xiao R. Chemical-looping combustion of biomass in a 10 kWth reactor with iron oxide as an oxygen carrier. Energy Fuels 2009;23:2498–505. [15] Gu H, Shen L, Xiao J, Zhang S, Song T. Chemical looping combustion of biomass/coal with natural iron ore as oxygen carrier in a continuous reactor. Energy Fuels 2010;25:446–55. [16] Song T, Shen T, Shen L, Xiao J, Gu H, Zhang S. Evaluation of hematite oxygen carrier in chemical-looping combustion of coal. Fuel 2013;104:244–52. [17] Berguerand N, Lyngfelt A. Design and operation of a 10 kWth chemical-looping combustor for solid fuels–testing with South African coal. Fuel 2008;87:2713–26. [18] Keller M, Arjmand M, Leion H, Mattisson T. Interaction of mineral matter of coal with oxygen carriers in chemical-looping combustion (CLC). Chem Eng Res Des 2014;29:1753–70. [19] Wen Y, Li Z, Xu L, Cai N. Experimental study of natural Cu ore particles as oxygen carriers in chemical looping with oxygen uncoupling (CLOU). Energy Fuels 2012;26:3919–27. [20] Leion H, Lyngfelt A, Johansson M, Jerndal E, Mattisson T. The use of ilmenite as an oxygen carrier in chemical-looping combustion. Chem Eng Res Des 2008;86:1017–26. [21] Schwebel GL, Leion H, Krumm W. Comparison of natural ilmenites as oxygen carriers in chemical-looping combustion and influence of water gas shift reaction on gas composition. Chem Eng Res Des 2012;90:1351–60. [22] Moldenhauer P, Rydén M, Mattisson T, Younes M, Lyngfelt A. The use of ilmenite as oxygen carrier with kerosene in a 300W CLC laboratory reactor with continuous circulation. Appl Energy 2014;113:1846–54. [23] Linderholm C, Lyngfelt A, Cuadrat A, Jerndal E. Chemical-looping combustion of solid fuels–operation in a 10 kW unit with two fuels, above-bed and in-bed fuel feed and two oxygen carriers, manganese ore and ilmenite. Fuel 2012;102:808–22. [24] Fossdal A, Bakken E, Øye B, Schøning C, Kaus I, Mokkelbost T, et al. Study of inexpensive oxygen carriers for chemical looping combustion. Int J Greenhouse Gas Control 2011;5:483–8. [25] Leion H, Mattisson T, Lyngfelt A. Use of ores and industrial products as oxygen carriers in chemical-looping combustion. Energy Fuels 2009;23:2307–15. [26] Mendiara T, García-Labiano F, Gayán P, Abad A, de Diego LF, Adánez J. Evaluation of the use of different coals in chemical looping combustion using a bauxite waste as oxygen carrier. Fuel 2013;106:814–26. [27] Moldenhauer P, Rydén M, Lyngfelt A. Testing of minerals and industrial byproducts as oxygen carriers for chemical-looping combustion in a circulating fluidized-bed 300W laboratory reactor. Fuel 2012;93:351–63. [28] Schwebel GL, Filippou D, Hudon G, Tworkowski M, Gipperich A, Krumm W. Experimental comparison of two different ilmenites in fluidized bed and fixed bed chemical-looping combustion. Appl Energy 2014;113:1902–8. [29] Cuadrat A, Abad A, García-Labiano F, Gayán P, de Diego LF, Adánez J. Ilmenite as oxygen carrier in a chemical looping combustion system with coal. Energy Procedia 2011;4:362–9.
L. Xu et al. / Applied Energy 162 (2016) 940–947 [30] Abad A, Adánez J, Cuadrat A, García-Labiano F, Gayán P, de Diego LF. Kinetics of redox reactions of ilmenite for chemical-looping combustion. Chem Eng Sci 2011;66:689–702. [31] Ströhle J, Orth M, Epple B. Design and operation of a 1 MWth chemical looping plant. Appl Energy 2014;113:1490–5. [32] Arjmand M, Leion H, Lyngfelt A, Mattisson T. Use of manganese ore in chemical-looping combustion (CLC) effect on steam gasification. Int J Greenhouse Gas Control 2012;8:56–60. [33] Lyngfelt A, Leckner B, Mattisson T. A fluidized-bed combustion process with inherent CO2 separation; application of chemical-looping combustion. Chem Eng Sci 2001;56:3101–13. [34] Bao J, Li Z, Cai N. Promoting the reduction reactivity of ilmenite by introducing foreign ions in chemical looping combustion. Ind Eng Chem Res 2013;52:6119–28.
947
[35] Xu L, Edland R, Li Z, Leion H, Zhao D, Cai N. Cu-modified manganese ore as an oxygen carrier for chemical looping combustion. Energy Fuels 2014;28:7085–92. [36] Kunii D, Levenspiel O. Butterworth-Heine-mann series in chemical engineering. Fluidization engineering. Oxford, U.K.: Butterworth-Heinemann; 1991. p. 69. [37] Bao J, Li Z, Sun H, Cai N. Continuous test of ilmenite-based oxygen carriers for chemical looping combustion in a dual fluidized bed reactor system. Ind Eng Chem Res 2013;52:14817–27. [38] Xu L, Li Z, Sun H, Bao J, Cai N. Low-temperature chemical looping combustion for removing unburnt gaseous components with a cement-supported CuO oxygen carrier. Energy Fuels 2013;27:6872–9. [39] Rydén M, Leion H, Mattisson T, Lyngfelt A. Combined oxides as oxygen carrier material for chemical-looping combustion with oxygen uncoupling. Appl Energy 2014;113:1924–32.
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Experimental study of copper modified manganese ores as oxygen carriers in a dual fluidized bed reactor Lei Xu, Hongming Sun, Zhenshan Li ⇑, Ningsheng Cai Key Laboratory for Thermal Science and Power Engineering of Ministry of Education, Beijing Municipal Key Laboratory for CO2 Utilization & Reduction, Department of Thermal Engineering, Tsinghua University, Beijing 100084, China
h i g h l i g h t s Reactivity of manganese ore was significantly improved by impregnating copper. Two 88 h continuous run were performed using both oxygen carriers. Attrition data was obtained for both materials based on long operations.
a r t i c l e
i n f o
Article history: Received 28 August 2015 Received in revised form 13 October 2015 Accepted 29 October 2015
Keywords: Manganese ore Chemical looping combustion Oxygen carrier
a b s t r a c t Chemical-looping combustion (CLC) is a developing CO2 capture technology. CLC makes use of the repeated oxidation/reduction reactions of metal oxide (oxygen carrier, OC) to separate CO2 from fuel combustion and to obtain a pure CO2 stream suitable for compression and storage. Low cost materials, such as natural ores, are required for coal-fueled CLC because the lifetime of the oxygen carrier (OC) is lowered by side reactions with the fuel ash or carryover losses. In this study, five manganese ores were examined as oxygen carriers using CO as the fuel gas in a laboratory batch fluidized bed reactor. All five of the ores were impregnated by copper nitrate solution to evaluate the reactivity enhancement of copper impregnation. The period with full CO conversion can be enhanced 2–100 times for different ores in the single fluidized bed test, which indicated that the Cu impregnation may be a general method to enhance the reactivity of manganese ores. Finally, one manganese ore and the corresponding Cu-modified particles were tested in a dual fluidized bed reactor. The attrition rates of both materials were measured as 0.13 wt.%/h during the 88 h operation in the dual fluidized bed. Both the manganese ore and the Cuimpregnated ore exhibited stable and high reactivity during the continuous test in the dual fluidized bed reactor, even at a low temperature (310 °C). Copper impregnation had no obvious influence on the attrition property of the manganese ore. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Chemical looping combustion (CLC) is a promising technology for inherent CO2 separation at a low cost [1]. CLC generally consists of two reaction steps: reduction and oxidation. During the reduction step, the fuel is oxidized by the oxygen provided by the reduction of a solid oxygen carrier, typically a metal oxide. In the oxidation step, the reduced oxygen carrier is re-oxidized by air. These two steps are commonly performed in two separate reactors: the fuel reactor and the air reactor, respectively. Both reactors are connected by a circulating stream of solid oxygen carrier. The gases from both reactors never mix, and a stream of H2O and ⇑ Corresponding author. Tel./fax: +86 10 62789955. E-mail address: [email protected] (Z. Li). http://dx.doi.org/10.1016/j.apenergy.2015.10.167 0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.
CO2 is obtained at the outlet of the fuel reactor, which stays separated from the air nitrogen. The principle of using a solid oxygen carrier material to provide oxygen for thermochemical fuel conversion was proposed in the 1940s [2] to reform methane. The idea of CLC was patented as a process for CO2 production from carbonaceous sources in the 1950s [3]. The same principle was proposed during the 1980s with the goal to lower the irreversibility of combustion [4,5]. With the increasing recognition of anthropogenic climate change, the concept was investigated to recover CO2 from the exhaust of combustion systems for power generation [6,7]. Since then, obtaining CO2 capture with a low energy penalty for the power sector has been the main driving force for CLC research. Gaseous fuels [8–10], such as natural gas or syngas, and solid fuels [11–17], such as coal or biomass, have been widely
941
L. Xu et al. / Applied Energy 162 (2016) 940–947
investigated in the literature as possible fuels for CLC. The application of gaseous fuels is less problematic, whereas for solid fuels, losses of oxygen carriers due to possible side reaction with ash and impurities in the gas stream [18], as well as material withdrawn from the reactor with the ash, may lead to high expected amounts of oxygen carrier makeup [17]. This increases the demand for abundant, low cost and environmental friendly oxygen carrier materials [13]. Possible materials under investigation include natural ores [19–23] and industrial residues or by-products [24–27]. Among the natural ores, ilmenite has been most widely investigated [20,28–31]. Most recently, Arjmand et al. [12,32] investigated several manganese ores as oxygen carriers for CLC, and these manganese ores showed better reactivity than ilmenite. This suggests that manganese ore could be a better choice than ilmenite as an oxygen carrier. It is generally accepted that low cost oxygen carriers, as mentioned above, tend to have lower conversion rates. Low reactivity of the oxygen carrier directly leads to an increase of the required solids inventory [33]. This in turn leads to a larger pressure drop within the reactor and inevitably to increased power consumption to maintain the solid fluidization [34]. Therefore, it is important to search for oxygen carriers with high reactivity or for methods allowing increased reactivity of the existing low cost oxygen carriers. Previous research has found that the reactivity of manganese ore can be enhanced by introducing a very small amount of copper ions [35]. The redox stability of the Cu-modified manganese ore and the effects of the loading amount on the reduction reactivity of Cu-modified manganese ore were investigated in a single fluidized bed reactor [35]. However, the more practical use of the oxygen carrier is in a dual fluidized bed reactor in continuous operation. Additionally, more knowledge about the copper impregnation method for improving the manganese ore reactivity would be useful. Accordingly, in this study, five manganese ores were impregnated by copper nitrite solution to examine the effect of the impregnating copper on different manganese ores. Furthermore, both the manganese ore and the Cu-modified manganese ore were evaluated in a dual fluidized bed reactor to examine the long-term chemical stability and attrition properties of the oxygen carriers.
mass ratio of the copper ion to manganese ore was 2 wt.% and was controlled during preparation. 2.2. Characterization of oxygen carriers The composition of the manganese ore was analyzed by X-ray fluorescence (XRF), and the results are summarized in Table 1. The corresponding mass fraction of single chemical element can be calculated based on the molecular formula. For example, the mass fraction of Mn can be calculated as 40.45%/71⁄55 = 31.33%. The BET (Brunauer, Emmett and Teller) surface area and BJH (Barrett, Joyner and Halenda) pore volume of the particles were measured with a Micromeritics micropore analyzer (Autosorb-iQ2MP, NOVA4000). The crushing strength for all particles was determined using an HP-10 digital force gauge. The value given is the average force in N from 30 individual measurements needed to crush a particle in the size fraction of 125–300 lm. The size distribution was measured by a laser particle size analyzer (Mastersizer 2000). The particles were observed under a scanning electron microscope (SEM, JSM-7001F). The bulk density, BET surface area, BJH pore volume and crushing strength of the particles are listed in Table 2. 2.3. Test of OC in the single fluidized bed reactor All oxygen carriers were tested in a single fluidized bed. The reactor was made of quartz tube (i.d. 30 mm) and was heated by an electric furnace. The bed material for all the tests was 30 g with a size of 125–300 lm placed on the porous distributor. The flow rates of the fluidizing gases for both reduction and oxidation were 2 L/min (STP). Both the reduction and oxidation temperature were 800 °C for the redox cycles. The gas velocity was 0.17 m/s at 800 °C, and the fluidization number U/Umf was 3.3, where Umf was calculated on the basis of the relations by Kunii and Levenspiel [36] with the mean radius of the particle as 212.5 lm. For each test, the bed material was first heated to 800 °C in 10 vol% O2 (90 vol% N2). Then, the reducing gas was introduced. The reducing agent was 10 vol% CO (90 vol% N2). The oxidizing agent was 10 vol% O2 (N2 90 vol%). Inert gas (N2) was introduced between the reduction and the oxidation period. The product gases were introduced into the gas analyzer after passing through a filter.
2. Experimental details 2.4. Test of OC in the dual fluidized bed reactor 2.1. Preparation of oxygen carriers The five natural manganese ores (designated as Mn1–Mn5) were provided by Mugui Manganese Ore Company from Guangxi Province in China. The ores were first crushed and then calcined in a muffle oven for 4 h at 950 °C and sieved to 125–300 lm. The Cu-modified manganese ores (designated as Mn1Cu2%–Mn5Cu2%) were prepared by impregnation of copper nitrate solution. For every 50 g of manganese ore, solid Cu(NO3)23H2O (3.8 g) was used as a precursors and was dissolved in water (15 ml) to give the solution. The ore particles were soaked in the solution for 6 h at ambient temperature and were then calcined in air for 2 h at 500 °C. After cooling, all of the particles were sieved to 125–300 lm. The
The continuous test was performed in a dual fluidized bed reactor, as illustrated in Fig. 1. The system consisted of an air reactor (AR), a fuel reactor (FR), a riser, a cyclone and two loop seals (LS). The riser was used to transport oxygen carriers from the air reactor to the fuel reactor. The solid circulation rate was controlled by the lower loop seal. The fuel reactor was a bubbling fluidized bed with a length of 1.3 m. The inner diameters of both the air reactor and the fuel reactor were 50 mm. The inner diameter of the riser was 20 mm. The height of the riser was 3.3 m. The secondary air was introduced at the lower part of the riser to assist particle entrainment. The air and fuel reactors were heated by electric furnaces. Ktype thermocouples were inserted into the reactor to monitor the
Table 1 Chemical analysis of the manganese ores. wt.%
MnO
Fe2O3
SiO2
Al2O3
BaO
K2O
P2O5
TiO2
MgO
CaO
Mn1 Mn2 Mn3 Mn4 Mn5
40.45 54.13 73.37 39.41 64.75
7.36 24.71 10.23 11.24 16.53
44.09 10.54 7.12 32.94 5.82
1.68 7.80 4.01 13.09 2.33
4.71 0.09 0.65 0.24 0.50
0.23 1.45 0.58 1.33 0.90
0.44 – 0.22 0.29 0.06
– 0.30 0.27 0.56 8.50
0.18 0.46 0.31 0.20 0.13
0.36 0.21 1.61 0.22 0.07
942
L. Xu et al. / Applied Energy 162 (2016) 940–947
Table 2 Physical properties and characteristics of the oxygen carriers used in this work. Sample
Bulk density (g/cm3)
BET surface area (m2/g)
BJH pore volume (cm3/g)
Crushing strength (N)
Mn1/Mn1Cu2% Mn2/Mn2Cu2% Mn3/Mn3Cu2% Mn4/Mn4Cu2% Mn5/Mn5Cu2%
1.71/1.72 0.94/0.93 1.45/1.45 1.28/1.32 1.67/1.69
2.82/3.64 6.32/6.39 1.70/2.34 3.21/4.75 2.34/3.42
0.016/0.204 0.029/0.031 0.007/0.010 0.020/0.024 0.013/0.012
2.0/1.9 ± 0.4 0.8/0.8 ± 0.1 1.8/1.7 ± 0.3 1.7/1.7 ± 0.4 1.9/1.9 ± 0.3
has been used for testing ilmenite and cement-supported CuO oxygen carriers in our previous studies, and more details about the reactor system can be found in the relevant publications [37,38]. Four continuous tests were conducted in the dual fluidized bed reactor. The first two were 88 h run using Mn5 and Mn5Cu2% as the oxygen carriers separately in the same operation conditions. The initial solid inventory of the system was 5000 g, and 10 vol% CO (90 vol% N2) as the reducing agent and air as the oxidizing agent for both tests. N2 was used as the fluidizing gas for the loop seals. The fuel reactor was maintained at 950 °C and the air reactor was maintained at 900 °C. The operation conditions are summarized in Table 3. The third continuous test lasted 200 min using Mn5 as the oxygen carrier at different operation conditions and with different concentrations of CO as fuel. The fourth continuous test also lasted 200 min using Mn5Cu2% as the oxygen carrier at the same operation conditions as the third one. The initial solid inventory of the system was 3500 g for both the third and fourth continuous tests. The reactor system was designed to work stably in a solid circulation rate range of 23–76 kg/h. In the first two operations, a solid circulation rate of 23 kg/h was used. In the third and fourth operations, a solid circulation rate of 76 kg/h was used. The detailed operation conditions are summarized in Table 4. 2.5. Data evaluation
Fig. 1. Schematic diagram of the dual fluidized bed reactor for CLC.
temperature. Fine particles from both reactors were collected by filters. O1–O8 and I0–I2 were pressure measurement points used to monitor the solid circulation of the entire loop. The reactor
CO conversion (X) and oxygen carrier-to-fuel ratio (u) are two parameters used to evaluate the result of the dual fluidized bed tests. The CO conversion was calculated according to Eq. (1),
X¼
C CO2 C CO2 þ C CO
ð1Þ
Table 3 Gas flow rates (LN/min, STP) at different locations during the first two continuous operations. Location
Fluidizing gas
Gas flow rates (LN/min)
Gas velocity (m/s)
U/Umf
AR Riser FR Lower LS Upper LS
Air Air 10 vol% CO N2 N2
4 30 4 2 2
0.15 5 0.15 0.07 0.07
2.9 94 2.8 1.3 1.3
Table 4 Fuel gas flow rates (LN/min, STP) and CO conversion at the different conditions in Figs. 5 and 6.
1 2 3 4 5 6 7 8
N2
CO
Fuel power
Temperature
(LN/min)
(LN/min)
(W)
(°C)
3.6 3 2.5 2 2 1 0 4
0.4 1 1.5 2 2 3 4 0
85 213 319 425 425 638 850
305–315 305–315 305–315 305–315 400–410 500–510 500–510
u 108 43 29 22 22 14 11
X Mn
Mn5Cu2%
0.98 0.91 0.91 0.87 0.97 1 1
1 1 1 1 1 1 1
L. Xu et al. / Applied Energy 162 (2016) 940–947
where C CO2 and C CO are the gas concentrations detected by the gas analyzer. The oxygen carrier-to-fuel ratio (u), defined by Eq. (2), was used to describe the state of solid circulation and fuel feed rate during the operation.
u¼
F Mn2 O3 3F CO
ð2Þ
where F Mn2 O3 is the molar flow rate of Mn2O3, which is calculated from the solid circulation rate and the ore composition, assuming that the manganese oxide exists in the form of Mn2O3 before reduction (u = 1 corresponds to the stoichiometric Mn2O3 amount required for the full conversion of CO to CO2 and the full reduction of Mn2O3 to Mn3O4). 3. Results and discussion 3.1. Test of Cu-modified manganese ore in the single fluidized bed reactor The performance of oxygen carriers was investigated in a single fluidized bed using CO as fuel. All of the oxygen carriers showed similar profiles of outlet CO and CO2. CO was fully converted at the beginning of each redox cycle. CO2 was detected immediately after CO was introduced into the reactor. As the reaction proceeded, oxygen carriers were consumed and CO started to come out. The period with full CO conversion is defined as breakthrough time, as illustrated in Fig. 2a. For example, the breakthrough time of Mn1Cu2% was 426 s for the second cycle. The reactivity of oxygen carriers in the first cycle was higher than the reactivity in the successive cycles. However, the reactivity decay for the successive cycles was very slow and tended to be stable. Therefore, the breakthrough time of the second cycle was used to compare the reactivity of the oxygen carriers and is summarized in Fig. 2b. The copper addition enhanced the reactivity of all five ores. However, the extent to which the reactivity was improved was different from Mn1 to Mn5. The breakthrough time was enhanced by less than 2 times for Mn1 but by 100 times for Mn5. Before impregnation, Mn1 showed the best reactivity, whereas Mn2, Mn4 and Mn5 showed relatively lower reactivity. After impregnation, the reactivity of Mn2, Mn4 and Mn5 was enhanced by 8–100 times. Therefore, although only five ores were tested in our study, it can be postulated that copper impregnation could be an effective method to improve the reactivity of low-reactive manganese ores. Fig. 3 shows that impregnation with copper can enhance the surface area of the particles. The enhancement of the surface area contributed to the higher reactivity of the particles. Pore volume was also enhanced by copper impregnation, except for Mn5. Increased pore volume facilitated the diffusion of reactant gas
943
inside the particles and thus increased the overall reactivity. However, neither the increase of the surface area nor the pore volume was as significant as the enhancement of the breakthrough time, especially for Mn2, Mn4 and Mn5. This result suggests that the contribution of copper to the reactivity enhancement was mainly due to the interaction of copper with manganese oxide, which is consistent with our previous study [35].
3.2. Continuous operation using Mn5 and Mn5Cu2% as oxygen carriers Mn5 and Mn5Cu2% were used for the continuous tests in the dual fluidized bed. No sintering and agglomeration were found during the operations. The reactor system was run for 88 hours continuously for each test. Fig. 4 shows the outlet gas composition and pressure drop of the fuel reactor during the continuous run. In the first 7 h, the whole reactor system was heating, and the fuel reactor was fluidized by N2. After the fuel reactor and the air reactor reached the target temperatures, 950 °C and 900 °C, respectively, the fuel gas was switched to 10 vol% CO. The corresponding fuel power is 85 W, with a specific inventory of 59 kg/kW. The operation conditions were then kept constant, as summarized in Table 3. The solid circulation rate was 23 kg/h, calculated according to Eq. (1) in previous work [37], i.e. 20 kg/m2 s based on the cross-section area of riser. Both materials presented similar results during the continuous run. The detected gas concentrations of CO and CO2 were mixed with N2 from the loop seals. Generally, the off-gas from FR was mainly diluted with N2 from the upper loop seal and the off-gas from AR was mainly diluted with N2 from the lower loop seal [37]. The CO2 concentration varied between 6.2 vol% and 7 vol% during the whole operation. No CO was detected, which meant that CO was fully converted. The result shows that the manganese ore has much higher reactivity than ilmenite. Ilmenite only reached a conversion of 60% in the same reactor system at similar conditions [37]. The pressure drop through the fuel reactor, denoted as DP, decreased rapidly in the first 20 h and was then stable. During the first 10 h, the pressure drop was beyond the gauge range; therefore, it appears as a straight line in Fig. 4. From the 10th hour to the 19th hour, DP dropped at a rate of 470 Pa/h. From the 19th hour to the end, DP dropped only 300 Pa. The rapid decline of DP during the first 19 h was probably due to the rapid attrition and fragmentation rate of the particles. Mn5 showed good and stable reactivity during the 88 h continuous run, especially compared to the results from ilmenite [37]. Considering the relatively low reactivity of Mn5 among the five manganese ores, it could be speculated that manganese ores are much better oxygen carriers than ilmenite with respect to reactivity. However, the mechanical properties of manganese ores are worse than ilmenite. The mechanical properties of both materials are discussed in detail in Section 3.4.
Fig. 2. Outlet CO and CO2 profile of Mn1 in the single fluidized bed test (a) and breakthrough time comparison of the 2nd cycle of ten oxygen carriers (b).
944
L. Xu et al. / Applied Energy 162 (2016) 940–947
Fig. 3. Change of the BET surface area (a) and BJH pore volume (b) of five manganese ores by impregnating with copper.
Fig. 5. Flue gas concentrations of the fuel reactor as a function of time with changing operation conditions and the manganese ore as the oxygen carrier.
nents from CLC reactors [38]. The upper loop seal used N2 with a flow rate of 4 LN/min. The lower loop seal used N2 with a flow rate of 3.5 LN/min. The air reactor was fluidized by air with a flow rate of 5.5 LN/min, and the flow rate of the secondary air was 45 LN/min. The flow rate and gas composition of the fuel gases were varied during the two operations. The whole period for each run was divided into eight stages, and the fuel gas flow rates, temperatures and CO conversions in each stage are summarized in Table 4. Using Mn5 as the oxygen carrier, the CO conversion reached 0.98 at 310 °C when the CO flow rate was 0.4 LN/min (u 108). The conversion dropped to 0.87 when the CO flow rate increased to 2 LN/min (u 22). When the temperature increased to 400 °C, the conversion increased to 0.97. At 500 °C, CO was fully converted, even after the CO flow rate increased to 4 LN/min (u 11). The Cuimpregnated manganese ore showed much better reactivity than the original ore. CO was almost fully converted in the same operation conditions, as can be seen in Table 4 and Fig. 6.
3.4. Fragmentation and attrition of the oxygen carriers Oxygen carriers are continuously lost in the fluidized bed reactor due to abrasive wear or attrition and fragmentation resulting
Fig. 4. Flue gas concentrations and pressure drop of the fuel reactor versus time during the 88 h continuous operation with Mn5 (a) and (b), and Mn5Cu2% (c) and (d).
3.3. Comparison of manganese ore and Cu-modified manganese ore in dual fluidized bed reactor Continuous operations of Mn5 and Mn5Cu2% were performed at lower temperatures in the dual fluidized bed reactor for two reasons. One is that the reactivity difference of both materials can be distinguished at lower temperatures in the reactor system, the other reason is that oxygen carriers could have potential use at lower temperatures for eliminating the unburnt gaseous compo-
Fig. 6. Flue gas concentrations of the fuel reactor as a function of time with changing operation conditions and the Cu-modified manganese ore as the oxygen carrier.
945
L. Xu et al. / Applied Energy 162 (2016) 940–947
Fig. 7. Particle size distribution of Mn5 and Mn5Cu2% before and after the continuous test in the reactor and filter.
from particle–particle interactions and particle–wall interactions. Resistance to attrition and fragmentation is an important property of particles used as oxygen carriers. The particle size distribution, crushing strength and attrition rate were used to evaluate the mechanical properties of the oxygen carriers. The particle size distributions of fresh Mn5 and Mn5Cu2% particles remaining in the reactor after the 88 h operation and the particles in the AR filter are presented in Fig. 7. Both materials presented a similar attrition property. This suggested that impregnation by 2 wt.% copper did not have an obvious effect on the attrition property of manganese ore. The size distribution of both materials were narrower, and the average size were smaller after the test. The d(0.9) of Mn5 in the reactor decreased from 508 lm to 349 lm after the test while 505–341 lm for Mn5Cu2%. The crushing strengths of the ten oxygen carriers were also tested to evaluate the mechanical properties of the materials. A crushing strength larger than 2 N is considered sufficiently hard according to a study by Rydén et al. [39]. Four ores (Mn1, Mn3, Mn4 and M5) showed similar crushing strengths, ranging from 1.7 to 2.0 N, and Mn2 showed a much smaller crushing strength of 0.8 N, as illustrated in Table 2. The attrition rate is another important parameter used for estimating the mechanical properties of oxygen carriers. The attrition rate can be evaluated using a parameter called the loss of fines, Lf, which is defined as
Lf ¼
Dmfines 1 100 Dt mtotal
Particles of Mn5 elutriated from the dual fluidized bed reactor during the 88 h operation were collected at different times (10 h, 24 h, 48 h, 72 h, and the end of the operation) from the filters and fine particles collector and were weighed to determine Lf. The loss of fines for the oxygen carriers was assumed to be those with size of <125 lm as determined by passing through a sieve. The mass of the particles collected in the AR filter (1610 g) was much greater than that in the FR filter (190 g), as shown in Table 5. Moreover, the amount of particles with size of <125 lm in the AR filter (1170 g) was much greater than that in the FR filter (11 g). This suggests that attrition and fragmentation of particles mainly occurred in the riser and cyclone rather than the other parts of the reactor system, such as the air reactor and the fuel reactor. This is reasonable because the much higher gas velocity in the riser due to the introduction of secondary air and the swirling flow inside the cyclone would lead to fiercer collisions between particles and impacts of particles against the reactor wall. During the operation, the loss of fine particles was high at the beginning and rapidly decreased, as shown in Fig. 8. In the first 10 h, the Lf was 1.17 wt.%/h for Mn5 and 1.28 wt.%/h for Mn5Cu2%. The Lf decreased rapidly to 0.34 wt.%/h for Mn5 and 0.36 wt.%/h for Mn5Cu2% in the next 14 h. Finally, it dropped to 0.13 wt.%/h for both materials and stabilized. The same trend was also found in the study of a CuO/Al2O3 oxygen carrier [8]. The high attrition rate at the beginning was partially attributed to the rounding effect, which means the sharp edges of the particles are more easily scraped, and the shape of the particle becomes closer and closer to a sphere. Fig. 9 shows that the particles after the 88 h operation were not as sharp as the fresh particles. Another possible reason could be that the particles became more robust as the reaction went on. This was indicated by the higher crushing strength of the particles after the 88 h operation, which was 2.1 N for each material. It can be concluded that the ore particles had a high attrition rate in the beginning and a much lower attrition rate after tens of hours; therefore, the lifetime is estimated as 770 h according to the Lf of 0.13 wt.%/h.
ð3Þ
where Dmfines is the amount of fines produced during the period of time, Dt, and mtotal is the total amount of materials input before the operation, i.e., 5000 g in this study. The lifetime of the oxygen carrier, t life , is calculated as
t life ¼
1 Lf
ð4Þ
Fig. 8. Loss of fines of Mn5 and Mn5Cu2% versus time during the 88 h operation.
Table 5 Mass distribution in the system for Mn5 and Mn5Cu2% after the 88 h tests. Sample
Mn5 Mn5Cu2%
Reactor (g)
2813 2766
FR filter (g)
AR filter (g)
<125 lm
>125 lm
<125 lm
>125 lm
11 15
179 176
1170 1165
440 470
Collector (g)
Total (g)
Lifetime (h)
136 142
5000 5000
770 770
946
L. Xu et al. / Applied Energy 162 (2016) 940–947
Fig. 9. SEM images of the Mn5 particles before (a) and after (b) the 88 h operation.
4. Conclusions Wet impregnation of manganese ore by copper nitrate solution enhanced the reactivity of manganese ore as an oxygen carrier. Five manganese ores and five copper impregnated ones were examined as oxygen carriers using CO as the fuel gas in a laboratory batch fluidized bed reactor. The breakthrough time of CO was enhanced 2–100 times on different ores by impregnating with 2 wt.% Cu. Four manganese ores presented similar crushing strengths between 1.7 and 2.0 N. Mn5 was chosen to perform an 88 h run in a dual fluidized bed to evaluate the reactivity, fragmentation and attrition properties of the material. The manganese ore showed much higher reactivity than ilmenite. Both the manganese ore and the Cu-impregnated ore showed stable and high reactivity during the continuous test in the dual fluidized bed reactor, even at a low temperature (310 °C). The attrition rate was high in the beginning of the 88 h operation and became stable at 0.13 wt. %/h after 24 h. The lifetimes of both the manganese ore and Cumodified manganese ore are estimated as 770 h. Manganese ores are good candidates as oxygen carriers for chemical looping combustion because of the higher reactivity compared to ilmenite. Copper impregnation can enhance the reactivity of manganese ore, especially those ores with a low reactivity. Acknowledgments This work was supported by the National Natural Science Foundation of China (51376105, 91434124), the National Key Basic Research and Development Program (2011CB707301), the Tsinghua University Initiative Scientific Research Program, and the Program for New Century Excellent Talents in University (NCET-120304). References [1] Ekström C, Schwendig F, Biede O, Franco F, Haupt G, de Koeijer G, et al. Technoeconomic evaluations and benchmarking of pre-combustion CO2 capture and oxy-fuel processes developed in the European ENCAP project. Energy Proc 2009;1:4233–40. [2] Lewis WK, Gilliland ER, Reed WA. Reaction of methane with copper oxide in a fluidized bed. Ind Eng Chem 1949;41:1227–37. [3] Lewis WK. Production of pure carbon dioxide. U.S. Patent 2,665,972. 1954-112. [4] Knoche KF, Richter H. Verbesserung der reversibilität von verbrennungsprozessen. Brennstoff-Wärme-Kraft 1968;20:205–11. [5] Richter HJ, Knoche KF. Reversibility of combustion processes. ACS Symp Ser 1983;235:71–86. [6] Ishida M, Zheng D, Akehata T. Evaluation of a chemical-looping-combustion power-generation system by graphic exergy analysis. Energy 1987;12:147–54. [7] Ishida M, Jin H. A new advanced power-generation system using chemicallooping combustion. Energy 1994;19:415–22.
[8] Adánez J, Gayán P, Celaya J, de Diego LF, García-Labiano 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;45:6075–80. [9] Jin H, Ishida M. Reactivity study on natural-gas-fueled chemical-looping combustion by a fixed-bed reactor. Ind Eng Chem Res 2002;41:4004–7. [10] Mattisson T, Garcia-Labiano F, Kronberger B, Lyngfelt A, Adánez J, Hofbauer H. Chemical-looping combustion using syngas as fuel. Int J Greenhouse Gas Control 2007;1:158–69. [11] Lyngfelt A. Chemical-looping combustion of solid fuels – status of development. Appl Energy 2014;113:1869–73. [12] Arjmand M, Leion H, Mattisson T, Lyngfelt A. Investigation of different manganese ores as oxygen carriers in chemical-looping combustion (CLC) for solid fuels. Appl Energy 2014;113:1883–94. [13] Leion H, Mattisson T, Lyngfelt A. Solid fuels in chemical-looping combustion. Int J Greenhouse Gas Control 2008;2:180–93. [14] Shen L, Wu J, Xiao J, Song Q, Xiao R. Chemical-looping combustion of biomass in a 10 kWth reactor with iron oxide as an oxygen carrier. Energy Fuels 2009;23:2498–505. [15] Gu H, Shen L, Xiao J, Zhang S, Song T. Chemical looping combustion of biomass/coal with natural iron ore as oxygen carrier in a continuous reactor. Energy Fuels 2010;25:446–55. [16] Song T, Shen T, Shen L, Xiao J, Gu H, Zhang S. Evaluation of hematite oxygen carrier in chemical-looping combustion of coal. Fuel 2013;104:244–52. [17] Berguerand N, Lyngfelt A. Design and operation of a 10 kWth chemical-looping combustor for solid fuels–testing with South African coal. Fuel 2008;87:2713–26. [18] Keller M, Arjmand M, Leion H, Mattisson T. Interaction of mineral matter of coal with oxygen carriers in chemical-looping combustion (CLC). Chem Eng Res Des 2014;29:1753–70. [19] Wen Y, Li Z, Xu L, Cai N. Experimental study of natural Cu ore particles as oxygen carriers in chemical looping with oxygen uncoupling (CLOU). Energy Fuels 2012;26:3919–27. [20] Leion H, Lyngfelt A, Johansson M, Jerndal E, Mattisson T. The use of ilmenite as an oxygen carrier in chemical-looping combustion. Chem Eng Res Des 2008;86:1017–26. [21] Schwebel GL, Leion H, Krumm W. Comparison of natural ilmenites as oxygen carriers in chemical-looping combustion and influence of water gas shift reaction on gas composition. Chem Eng Res Des 2012;90:1351–60. [22] Moldenhauer P, Rydén M, Mattisson T, Younes M, Lyngfelt A. The use of ilmenite as oxygen carrier with kerosene in a 300W CLC laboratory reactor with continuous circulation. Appl Energy 2014;113:1846–54. [23] Linderholm C, Lyngfelt A, Cuadrat A, Jerndal E. Chemical-looping combustion of solid fuels–operation in a 10 kW unit with two fuels, above-bed and in-bed fuel feed and two oxygen carriers, manganese ore and ilmenite. Fuel 2012;102:808–22. [24] Fossdal A, Bakken E, Øye B, Schøning C, Kaus I, Mokkelbost T, et al. Study of inexpensive oxygen carriers for chemical looping combustion. Int J Greenhouse Gas Control 2011;5:483–8. [25] Leion H, Mattisson T, Lyngfelt A. Use of ores and industrial products as oxygen carriers in chemical-looping combustion. Energy Fuels 2009;23:2307–15. [26] Mendiara T, García-Labiano F, Gayán P, Abad A, de Diego LF, Adánez J. Evaluation of the use of different coals in chemical looping combustion using a bauxite waste as oxygen carrier. Fuel 2013;106:814–26. [27] Moldenhauer P, Rydén M, Lyngfelt A. Testing of minerals and industrial byproducts as oxygen carriers for chemical-looping combustion in a circulating fluidized-bed 300W laboratory reactor. Fuel 2012;93:351–63. [28] Schwebel GL, Filippou D, Hudon G, Tworkowski M, Gipperich A, Krumm W. Experimental comparison of two different ilmenites in fluidized bed and fixed bed chemical-looping combustion. Appl Energy 2014;113:1902–8. [29] Cuadrat A, Abad A, García-Labiano F, Gayán P, de Diego LF, Adánez J. Ilmenite as oxygen carrier in a chemical looping combustion system with coal. Energy Procedia 2011;4:362–9.
L. Xu et al. / Applied Energy 162 (2016) 940–947 [30] Abad A, Adánez J, Cuadrat A, García-Labiano F, Gayán P, de Diego LF. Kinetics of redox reactions of ilmenite for chemical-looping combustion. Chem Eng Sci 2011;66:689–702. [31] Ströhle J, Orth M, Epple B. Design and operation of a 1 MWth chemical looping plant. Appl Energy 2014;113:1490–5. [32] Arjmand M, Leion H, Lyngfelt A, Mattisson T. Use of manganese ore in chemical-looping combustion (CLC) effect on steam gasification. Int J Greenhouse Gas Control 2012;8:56–60. [33] Lyngfelt A, Leckner B, Mattisson T. A fluidized-bed combustion process with inherent CO2 separation; application of chemical-looping combustion. Chem Eng Sci 2001;56:3101–13. [34] Bao J, Li Z, Cai N. Promoting the reduction reactivity of ilmenite by introducing foreign ions in chemical looping combustion. Ind Eng Chem Res 2013;52:6119–28.
947
[35] Xu L, Edland R, Li Z, Leion H, Zhao D, Cai N. Cu-modified manganese ore as an oxygen carrier for chemical looping combustion. Energy Fuels 2014;28:7085–92. [36] Kunii D, Levenspiel O. Butterworth-Heine-mann series in chemical engineering. Fluidization engineering. Oxford, U.K.: Butterworth-Heinemann; 1991. p. 69. [37] Bao J, Li Z, Sun H, Cai N. Continuous test of ilmenite-based oxygen carriers for chemical looping combustion in a dual fluidized bed reactor system. Ind Eng Chem Res 2013;52:14817–27. [38] Xu L, Li Z, Sun H, Bao J, Cai N. Low-temperature chemical looping combustion for removing unburnt gaseous components with a cement-supported CuO oxygen carrier. Energy Fuels 2013;27:6872–9. [39] Rydén M, Leion H, Mattisson T, Lyngfelt A. Combined oxides as oxygen carrier material for chemical-looping combustion with oxygen uncoupling. Appl Energy 2014;113:1924–32.