Article pubs.acs.org/est
Process Modeling of an Advanced NH3 Abatement and Recycling Technology in the Ammonia-Based CO2 Capture Process Kangkang Li,†,‡ Hai Yu,*,‡ Moses Tade,† Paul Feron,‡ Jingwen Yu,§ and Shujuan Wang§ †
Department of Chemical Engineering, Curtin University of Technology Australia, GPO Box U1987, Perth, Western Australia 6845, Australia ‡ CSIRO Energy Centre, 10 Murray Dwyer Circuit, Mayfield West, NSW 2304, Australia § Department of Thermal Engineering, Tsinghua University, Beijing 100084, People’s Republic of China S Supporting Information *
ABSTRACT: An advanced NH3 abatement and recycling process that makes great use of the waste heat in flue gas was proposed to solve the problems of ammonia slip, NH3 makeup, and flue gas cooling in the ammonia-based CO2 capture process. The rigorous rate-based model, RateFrac in Aspen Plus, was thermodynamically and kinetically validated by experimental data from open literature and CSIRO pilot trials at Munmorah Power Station, Australia, respectively. After a thorough sensitivity analysis and process improvement, the NH3 recycling efficiency reached as high as 99.87%, and the NH3 exhaust concentration was only 15.4 ppmv. Most importantly, the energy consumption of the NH3 abatement and recycling system was only 59.34 kJ/kg CO2 of electricity. The evaluation of mass balance and temperature steady shows that this NH3 recovery process was technically effective and feasible. This process therefore is a promising prospect toward industrial application.
1. INTRODUCTION The issue of ammonia slip is becoming one of the most critical technical and economical difficulties for the commercial application of ammonia-based CO2 capture technology.1 During the CO2 capture process, ammonia is inevitably lost to the flue gas due to its intrinsically high volatility. This requires ammonia makeup to balance solvent ammonia concentration, as well as introducing a high energy penalty to reuse the slipped ammonia. Ammonia itself is a hazardous gas that strongly irritates the throat, eyes and respiratory system, thereby threatening the health of human beings and animals. The United States National Institute of Occupational Safety and Health (NIOSH) recommends that NH3 concentrations in workplace air should not exceed 25 ppmv.2,3 For the industrial emission level, the NH3 concentration should be controlled to less than 50 ppmv according to the Korean government regulation of industrial exhaust gas.1 Currently, the most common method of controlling ammonia loss is by installing a separate ammonia abatement system, which consists of an NH3 absorber and NH3 stripper, to capture and recycle vaporized NH3 using large amounts of washing water. This method has been widely employed in many processes, such as the Alstom’s chilled ammonia process(CAP)4,5 and the KIER process.6 However, this conventional ammonia recovery method places an extra energy consumption burden on the entire CO2 capture system. Specifically, the thermodynamic analysis of Alstom’s chilled ammonia process by Mathias et al.7 shows that the NH3 abatement regenerator duty reaches © 2014 American Chemical Society
2377 kJ/kg CO2, while the CO2 stripper duty is only 2291 kJ/kg CO2 under the conditions of 26 wt % NH3 and 10 °C absorption temperature. Niu et al.8 carried out a rigorous simulation of the CO2 capture process by aqueous ammonia at room temperature and normal pressure. Their results show that the energy consumption for the NH3 abatement system (1703 kJ/kg) will be much higher than for CO2 regeneration (1285 kJ/kg CO2). Both studies indicate that the energy penalty for recovering slipped ammonia might be equivalent to or more than the energy penalty for CO2 capture alone. Although Darde et al.9 claims the heat requirement for the NH3 stripper in the CAP process can be reduced to 167 kJ/kg CO2 after configuration retrofit, the result is based on the rather low temperature cooling water available and the process faces the crystallization problem caused by the low temperature absorption process. Therefore, the substantial energy cost or the critical conditions of NH3 recovery is causing ammonia-based CO2 capture technology to lose its competitive edge of low energy consumption. The development of effective approaches is imperative to solve the ammonia slip problem and reduce the energy loads of the NH3 recovery system. In this work, we presented a simple but effective process for ammonia abatement and recycling by installing an NH3 Received: Revised: Accepted: Published: 7179
March 10, 2014 May 20, 2014 May 21, 2014 May 21, 2014 dx.doi.org/10.1021/es501175x | Environ. Sci. Technol. 2014, 48, 7179−7186
Environmental Science & Technology
Article
absorber and greatly utilizing the waste heat in flue gas. An accurate, reliable rate-based model using the commercial simulator-Aspen Plus was proposed for process optimization and improvement. The objectives of the new process are to recycle over 99% of slipped ammonia, control vent gas NH3 concentration to below 25 ppmv for the consideration of human being respiratory safety, and substantially reduce the energy consumption and capital cost of an NH3 recovery system.
(120 °C) in Australian coal-fired power stations required extra energy/ cooling water to cool it before entering the CO2 capture plant. The waste heat in hot flue gas could instead be used for NH3 recovery through heat exchange between the flue gas and washing solution in the pretreatment stage, which also saves energy consumption for flue gas cooling. Third, the PCC plant required a large amount of makeup ammonia to maintain ammonia balance during CO2 capture. The new process was therefore designed to significantly reduce the scale of ammonia makeup process by utilization of the heat in the flue gas and directly recycling almost all the slipped ammonia to the CO2 absorber. Figure 1 shows schematic flow-sheet of an ammonia-based CO2 capture process, which integrated the advanced ammonia
η% =
NH3 to CO2 absorber × 100% NH3 from CO2 absorber
3. PROCESS SIMULATION 3.1. Model Specification and Validation. In this work, the Pitzer property method, which was based on the aqueous electrolyte activity coefficient model, was applied to determine the fugacity coefficient, Gibbs energy, enthalpy, and entropy in the liquid phase; the Redlich−Kwong−Soave equation of state was used for the vapor phase fugacity coefficient.11 The model validation included two steps. The first used the FLASH simulator embedded in Aspen Plus V7.3 to build and validate the thermodynamic model of the NH3CO2H2O system, which was fundamental to the ensuing rate-based modeling work. In the thermodynamic simulation, the reactions, equilibrium and rate constants, Henry constants and other parameters of the NH3CO2H2O system can be retrieved from the Aspen Plus databank, and further details have been published elsewhere.12,13 The study of vapor−liquid equilibrium using FLASH module was then carried out, and experimental data from Göppert et al.14 and Kurz et al.15 was used to verify the model. The results in Figure SI-1 in the Supporting Information (SI) show that the proposed thermodynamic model agreed well with the experimental results. The second step was to establish a rigorous rate-based model using the RateFrac simulator in Aspen Plus. This model gives a more accurate description and characterization of CO2 absorption and ammonia loss in a packed column, including data such as mass and heat transfer, materials and energy balance, and chemical kinetic, hydraulic, and interface properties of the CO2 capture process. The packing material (25 mm Pall ring), flow model (countercurrent), packed height (7.8 m in total), and column diameter (0.6 m) were consistent with the column construction in the Delta Electricity PCC plant at Munmorah Power Station,16 which is being relocated from Munmorah to Vales Pt Power Station, NSW. The absorption column was divided into 25 stages. The interfacial area factor for the pall ring packing was set as 1.2 according to the experimental measurement results from Munmorah pilot trails.13,16 The calculation of mass transfer and heat transfer properties of the packed column were determined by the correlation method
Figure 1. Schematic flow-sheet of the ammonia-based CO2 capture process.
abatement and recycling system with a typical CO2 capture system. The CO2 capture system comprised a CO2 absorber and a CO2 stripper, while the ammonia abatement and recycling system consisted of an NH3 absorber and a pretreatment column (NH3 stripper). Initially, fresh water (Lean-in) was fed to the top of the NH3 absorber to scrub the slipped ammonia. After NH3 absorption, the vent gas will meet the set NH3 exhaust level 7180
dx.doi.org/10.1021/es501175x | Environ. Sci. Technol. 2014, 48, 7179−7186
Environmental Science & Technology
7181
43.3 39 39.9
50.9
43.3 42.8 35.7
47.2
11.7 8.0 9.4
10.1
912 898 799 799
4.96 ± 0.12 4.49 ± 0.39 1.92 ± 0.14 0.31 ± 0.001 0.28 ± 0.01 0.22 ± 0.002 5.82 n.a. 0.36 n.a.
17.1 16.3 16.4 19.9
134
key test conditions solvent flow134 134 134 134 134 134 134 100 67 67 67 134 rate, L/min liquid inlet 23.9 27 32.3 16.4 25.8 16.8 17.0 15.5 15.2 15.5 15.6 14.5 temperature, °C NH3 wt % 4.91 ± 0.30 4.08 ± 0.10 4.21 ± 0.12 3.79 ± 0.24 3.56 ± 0.41 4.19 ± 0.30 3.98 ± 0.23 4.24 ± 0.43 4.37 ± 0.32 4.37 ± 0.25 4.00 ± 0.12 4.97 ± 0.20 0.24 ± 0.025 0.24 ± 0.014 0.23 ± 0.026 0.25 ± 0.012 0.24 ± 0.027 0.26 ± 0.014 0.22 ± 0.02 0.21 ± 0.04 0.22 ± 0.02 0.23 ± 0.06 0.22 ± 0.003 0.41 ± 0.002 lean CO2 c loading flue gas flow646 646 632 750 780 760 821 817 906 915 916 799 rate, kg/h flue gas inlet 8.6 9.4 9.8 7.6 8.8 10.8 8.1 8.0 10.1 9.4 9.4 9.8 CO2, vol % comparison of test results and simulation results test CO2 ab32.2 59.1 55.2 38.3 44.5 49.2 46.8 51.4 34.4 33.2 31.5 29.5 sorption rate g (absorber1), kg/h simu. CO2 ab48.6 50.5 53.4 38.6 46.2 53.4 44.9 40.4 41.4 39 38.2 31.2 sorption rate (absorber1), kg/h
36 34R2 34R1 34 33 32B 32A 32 31B 31R 31 30
b
4. RESULTS AND DISCUSSION 4.1. Base-Case Scenario. In the NH3 abatement and recycling process, the slipped NH3 from the CO2 absorber (4.7 kg/h) went to three destinations: one was to be discharged into the atmosphere after the NH3 absorber; the second was to be recycled to the CO2 absorber after washing and the pretreatment column; and the third was to be absorbed or accumulated in the solution through the reaction with CO2 in the form of aqueous N-containing species, such as NH3 molecule, NH4+ ions and NH2COO− carbamate. The flow rate of NH3 absorbed by liquid (the third part) can be calculated
test ID
Table 1. Comparison between Pilot Plant Test and Rate Based Model Simulation Results under a Variety of Experimental Conditionsa
35B
35
134
39
134
38
proposed by Onda et al.17 and Chilton−Colburn,18 respectively. The correlation proposed by Stichlmair et al.19 was adopted for the holdup calculation. The validity of this ratebased model was verified by comparing the results between the simulation and CSIRO’s pilot trials. As shown in Table 1, the average relative error for the overall CO2 absorption rate in 30 tests was only ±6.0%, and for NH3 loss rate in 24 tests was ±11.1%. Figure 2 also gives a clear perception of the result comparison between the pilot plant and the rate-based model. The simulation results of CO2 absorption rate and ammonia loss rate were just within the error range of the experimental results. These results imply that the rate-based model is able to satisfactorily predict the CO2 absorption and ammonia loss rate for the pilot plant tests. Therefore, the established model is thermodynamically and kinetically valid. 3.2. Simulation Description of NH3 Abatement and Recycling System. Since the NH3-based CO2 absorption process and the proposed NH3 abatement and recycling process share the same chemical system of NH3H2OCO2, particularly the reaction equilibrium and kinetic constants, the validated model was therefore reliable and practical to guide process development and optimization of NH3 abatement and recycling process. The model settings were consistent with the above validated model, except that the stage number of the NH3 absorber was set to 25, in which the lean stream (Lean-in) to the fifth stage of the NH3 absorber and the makeup water was fed to the first stage for further NH3 removal. The typical hightemperature flue gas from the Munmorah Power Station and typical flue outgas after the CO2 absorber of CSIRO pilot test ID-32A containing 12 000 ppmv NH3 were used (details in SI Table SI-1). In reality, the flue gas also included 200−300 ppmv SO2 and NOX, and other trace species. The previous postcombustion capture pilot plant tests have confirmed that the presence of NOX (mainly NO) and other trace gases have minimal impact on the CO2 capture process and that the existence of SO2 can facilitate the capture of ammonia. A detailed study of SO2 capture will be conducted in next paper. The simulation process was divided into three steps. First, the base-case scenario without chiller and heat exchanger was set up to give a detailed description of the NH3 abatement and recycling process. Second, a thorough sensitivity analysis of the main process parameters, such as liquid flow rate, column size, and packing materials, was performed for process improvement and optimization. Third, the chilling process with chiller and heat exchanger was applied to further improve NH3 removal and recycling efficiency. The temperature approach between hot inlet and cold outlet streams was set as 5 °C for the heat exchanger. The chilling duty at different temperatures was converted into electricity consumption using a coefficient of performance (COP) of 5. The same value of temperature difference and COP has been used by Darde et al.9
134
Article
dx.doi.org/10.1021/es501175x | Environ. Sci. Technol. 2014, 48, 7179−7186
30
31
31R
7182
65.9 ± 2.8 69.8 5.6% 7.70 ± 1.5 4.8 –
67.3 ± 3.6 71.2 5.8% 7.19 ± 0.7 4.3 –
9.3% 46
3.6% 45
15−20 97.5 0.5 4.4
6.21
3.79
15−20 98.4 0.44 4.6
1.1% 5.68
9.5% 3.93
67 18.9 3.95 ± 0.10 0.22 ± 0.02 638
75
61.7
67 15.9 4.04 ± 0.21 0.21 ± 0.02 641
3.8% 75.9 ± 5.0
32
0.8% 68.2 ± 2.3
31B
0.3% 11.7 ± 2.0 9.5 18.8%
71.0
71.2 ± 2.2
15−20 103.0 1.2 6.0
67 28.0 4.53 ± 0.45 0.23 ± 0.01 667
12.6% 47
3.9
0.7% 4.46
87.2
8.5% 87.8 ± 3.3
32A
2.9% 13.6 ± 3.6 9.5 –
82.2
79.9 ± 6.0
15−20 105.7 1.3 6.7
67 27.0 3.83 ± 0.66 0.14 ± 0.02 665
10.4% 47 B
5.00
10.6% 5.58
71.9
4.0% 80.5 ± 4.3
32B
5.7% 12.4 ± 1.7 11.5 7.2%
81.4
86.2 ± 4.2
15−20 150.5 0.7 7.2
34
50
3.3% 6.21 ± 2.5 8.9 –
74.5
72.1 ± 1.1
15−20 102.5 1.3 6.8
51
1.0%
4.14
2.8% 4.08
72.9
14.1% 11.7 ± 2.2 12.6 7.7%
72.1
83.9 ± 2.1
15−20 158.3 1.3 8.8
52
15.3%
2.10
4.6% 2.48
57.6
4.8% 12.1 ± 2.5 11.0 9.1%
94.8
99.6 ± 6.8
15−20 157 2.1 6.6
53
4.78
3.2% 3.74
68.3
2.6% 9.4 ± 1.2 11.3 20.2%
63.6
65.3 ± 3.9l
15−20 131.3 0.6 6.5
9.0% 54
4.11
20.8% 3.77
64.3
8.8% 81.2 ± 4.1
35
5.4% 10.4 ± 2.0 12.3 18.2%
84.5
80.2 ± 3.5l
15−20 132.7 0.5 6.6
67 23.1 4.84 ± 0.30 0.19 ± 0.01 925
11.7% 70.6 ± 1.0
35B
50 23.1 4.08 ± 0.23 0.18 ± 0.01 972
5.7% 60.4 ± 3.0
36
100 22.4 4.33 ± 0.54 0.22 ± 0.05 1060
21.2% 70.0 ± 2.4
34R2
50 24.1 4.22 ± 0.34 0.18 ± 0.02 1003
16.2%
4.45
2.7% 5.31
72.4
17.4% 74.4 ± 3.0
34R1
50 23.0 4.77 ± 0.256 0.17 ± 0.01 661
0.3%
3.89
3.2% 3.88
75.8
20.3% 78.3 ± 4.3
67 23.2 4.56 ± 0.307 0.22 ± 0.005 1000
17.3% 48
4.46
19.5% 5.39
67.1
21.4% 85.9 ± 4.8
33
8.0% 8.2 ± 2.0 12.1 –
82.5
89.7 ± 4.17l
15−20 127 0.6 n.a
67 23.4 4.62 ± 0.18 0.18 ± 0.02 903
11.6% 55
4.41
1.7% 4.99
87.1
7.8% 85.6 ± 1.7
39
11.3% 9.8 ± 3.1 9.1 7.1%
75.4
85
15−20 122 n.a n.a
67 22.8 4.41 ± 0.27 0.22 ± 0.025 837
6.3% 56
2.03
11.8% 1.91
83.8
0% 74.9 ± 3.5
38
a
“b”: More details on test 30−38 in Qi G., et al;13 “c”: defined as the molar ratio of total C-containing species to the total N-containing species (C/N molar ratio, mol/mol). “--”: great error between test and simulation; “g”: CO2 absorption rate based on gas analysis ; “l”: CO2 absorption rate based on liquid analysis; “n.a”: not available.
key test conditions solvent flow-rate, L/min 67 67 liquid inlet temperature, °C 15.8 12.7 3.30 ± 0.27 4.40 ± 0.26 NH3 wt % 0.22 ± 0.01 0.24 ± 0.013 lean CO2 loading c flue gas mass flow-rate, 795 774 kg/h gas temperature, °C 15−20 15−20 inlet CO2 flow-rate, kg/h 125.9 113.1 0.9 0.28 inlet NH3 flow-rate, kg/h 6.8 5.3 inlet H2O flow-rate, kg/h comparison of test results and simulation results 69.6 ± 4.5 69.1 ± 1.7 test CO2 absorption rate g, kg/h simu. CO2 absorption rate, 71.9 71.4 kg/h relative error, % 3.3% 3.3% test NH3 loss rate, kg/h 5.27 ± 2.3 6.72 ± 2.3 4.1 5.2 simu. NH3 loss rate, kg/h relative error, % 22.2% 22.6%
comparison of test results and simulation results relative error – 14.5% 3.3% test overall CO2 66.6 ± 5.0 79.6 ± 1.6 77.7 ± 3.6 absorption rate g, kg/h simu. overall 70.1 74.8 77.3 CO2 absorption rate, kg/h relative error 5.2% 6.0% 0.5% 7.10 6.10 10.08 test NH3 loss rate g (absorber 1), kg/h simu. NH3 loss 7.00 6.81 9.23 rate (absorber 1), kg/h relative error 1.4% 11.6% 8.4% test ID 44−1 44−2
test ID
Table 1. continued
Environmental Science & Technology Article
dx.doi.org/10.1021/es501175x | Environ. Sci. Technol. 2014, 48, 7179−7186
Environmental Science & Technology
Article
Figure 2. Comparison of (a) CO2 absorption rate and (b) NH3 loss rate between pilot plant tests and rate-based model under the conditions listed in Table 2.
Figure 3. (a) NH3 distribution in gas and liquid phases and (b) flow rate of N-containing species as a function of number of cycles at a solvent circulation rate of 400 kg/h.
because in the NH3 desorption process, part of the carbonate/ carbamate/bicarbonate in the NH3-rich solution was also decomposed to release part of the CO2. This is proven by the study of CO2 profile in gas streams in SI Figure SI-2, which shows a 0.85 kg/h CO2 gas was released from the solution to flue gas in the pretreatment column. The study of CO2 profile also suggests that CO2 absorption/desorption in the washing/ pretreatment column reached steady state after 12 cycles. In terms of temperature profile, the study in SI Figure SI-3 suggests that temperatures in all streams can quickly reach steady state. We therefore conclude that the simulation has reached stable levels of species concentrations and stream temperatures over 12 cycles. The following simulations were conducted in the same way as the base case, and NH3 flow rates, CO2 flow rates, and temperatures at steady sate were employed for parameters optimization. 4.2. Process Optimization. We analyzed the effects of the washing solvent flow rate, column size, and packing materials on the NH3 recycling efficiency and emission concentration and present the results in Figure 4. The flow rate of washing solvent has a clear influence on the ammonia recycling efficiency and emission concentration. A low liquid flow rate cannot provide sufficient water to absorb ammonia, resulting in the slippage of more ammonia into the environment. While at a high washing solvent flow rate, the heat in the flue gas is insufficient to heat the solution to the temperature required for desorbing most of the NH3, leaving more to stay in solution and therefore increasing the ammonia emission concentration. As shown in Figure 4(a), the NH3 recycling efficiency first rose and then fell as the liquid flow rate increased, reaching a maximum of 94.9% at a rate of 350 kg/h. At this point, NH3 emission concentration reached a minimum value of 592 ppmv. The flow rate of 350 kg/h in the NH3
using the following equation which is based on NH3 mass balance for the NH3 abatement and recycling system. NH3 absorbed in solution = NH3 emitted from absorber − NH3 in vent gas − NH3 recycled to CO2absorber = NH3 scrubbed in NH3 absorber (absorption) − NH3 desorbed from the pretreatment column (desorption)
As shown in Figure 3(a), the NH3 distributed in the three destinations became constant and the process reached steady state after 12 cycles. The flow rate of NH3 to the CO2 absorber was consistently higher than those of the other NH3-containing streams, and stabilized at 4.4 kg/h with a NH3 recycling efficiency of 93.61%. In terms of the NH3 absorbed by the liquid, the flow rate decreased with increasing circulation cycles because of the increasing ammonia concentration in the circulation solvent. At steady state, the flow rate of NH3 absorbed by the liquid was close to “zero”. This indicates that NH3 absorption/ desorption in the liquid phase had reached a dynamic equilibrium where all the absorbed NH3 by the lean solvent was recycled back as gaseous NH3 to the CO2 absorber. However, the “zero” did not mean that there was no ammonia in the solution. Figure 3(b) shows the total flow rates of N-containing species in the lean stream and rich stream. Consistent with the trend in Figure 3(a), the flow rates of N-containing species in the lean and rich streams reached constant values after 12 cycles, these being 12.6 and 17.8 kg/h, respectively. It is interesting that the difference in flow rates between the two streams at steady state was 5.2 kg/h, which was higher than the total scrubbed NH3 of 4.4 kg/h. This is 7183
dx.doi.org/10.1021/es501175x | Environ. Sci. Technol. 2014, 48, 7179−7186
Environmental Science & Technology
Article
Figure 4. Effect of parameters such as (a) liquid flow rate, (b) NH3 absorber size, (c) pretreatment column size, and (d) packing materials on the NH3 recycling efficiency and emission concentration.
of 25 mm Pall ring, 341 m2/m3 of 16 mm Pall ring, 500 m2/m3 of Wire-Pack-500Y, and 708 m2/m3 of SULZER-CY (details in SI Table SI-3). The results on the basis of constant 80% flood and 0.6 mbar/m pressure drop, show that the recycling efficiencies increased and NH3 emissions fell significantly as the surface area increased, because the higher effective surface area is beneficial for greater gas−liquid mass transfer. When using SULZER-CY packing material for the washing and pretreatment column, the NH3 recycling efficiency reached 98.9% with 131 ppmv NH3 exhaust concentration. 4.3. Further Improvement by Chilling Process. The above simulation results were based on utilizing flue gas heat without additional energy onto the system. Although there was a high recycling efficiency of ammonia, the ammonia concentration discharged into the environment was 131 ppmv, which far exceeded the aimed emission level of 25 ppmv. So the chilling process of NH3-lean solution was included to decrease NH3 volatility, thereby ensuring a better NH3 capture efficiency. As shown in Table 2, NH3 emissions dropped significantly from 131 to 15.4 ppmv when the washing solvent was chilled to 10 °C, and to 3.9 ppmv at 5 °C. Meanwhile, the NH3 recycling efficiency experienced an increase to over 99.5% when chilling process applied. In terms of energy consumption, this advanced NH3 recovery system required no heat duty input because of making use of heat contained in the hot NH3-lean stream from the pretreatment stream (46−47 °C) by using high efficient heat exchanger, and required a very small electricity input due to the small amount of solvent needed in this system. Compared to Darde’s work,9 this promising system has a better ability to recover high NH3 concentration gas (12 000 ppmv) with the advantages of low energy penalty and simple facility requirement. Considering the technical and energy-saving targets, a chilling temperature of 10 °C was chosen, at which NH3 recycling efficiency was 99.87%, and vent gas NH3 concentration was 15.2 ppmv and the total energy consumption was only 59.34 kJ/kg CO2 captured, which is extremely small
recycle system is very small compared to the 8040 kg/h solvent circulation rate in the CO2 capture system, which is able to significantly reduce the scale of NH3 recovery and energy input created by solvent pumping. From the viewpoint of facility capital, it is very important to minimize the column size, as the construction cost is also one of the biggest barriers for CO2 capture commercial application. The design of column size should consider both technical efficiency and capital investment. As presented in Figure 4(b), (c), with the increasing packing height of the washing and pretreatment column from 1.5 to 3.0 m, the NH3 recycling efficiency rose sharply and the NH3 emission concentrations dropped quickly, due to the increasing residence time for the gas−liquid reactions alongside the columns. But no obvious improvement was found at a packed height over 3.0 m. In contrast, the column diameter has an insignificant effect on the NH3 abatement and recycling process. Basically, the diameter of column is strictly limited by flooding problem (80% of flood in this study). During the simulation, a warning of flooding appeared using a 0.4 m column diameter, where the gas encountered great resistance from the falling liquid. This will lead to a high pressure drop and make the operation difficult to carry out in plant process. Although increasing the column diameter to 0.5 m and 0.6 m eliminated the flooding risk, it increased the capital cost without obvious improvement in NH3 recycling efficiency in this study. Therefore, a balanced packed height and inner diameter were determined at 3.0 and 0.5 m, respectively, for both NH3 absorber and pretreatment column, at which NH3 recycling efficiency can reach 96.46% and 434 ppmv NH3 emission level. Packing materials with different geometries also influence the NH3 absorption and recycling process, because they are closely relevant to the mass/heat transfer coefficients, interfacial surface area and pressure drop. Figure 4(d) shows the NH3 recycling efficiency and emission concentration as a function of four different packing materials with effective surface areas of 205 m2/m3 7184
dx.doi.org/10.1021/es501175x | Environ. Sci. Technol. 2014, 48, 7179−7186
Environmental Science & Technology
Article
Table 2. Comparison of Darde’s Equilibrium Simulation of Two-Stage NH3 Washing with the Results of Chilling Process in This Study Darde’s simulationa NH3 conc. gas out absorber (ppmv) NH3 emission conc. (ppmv) NH3 recycle efficiency (%) heat requirement (kJ/kg CO2) electricity requirement (kJ/kg CO2) a
chilling temperature (°C)
case A
case B
22.5
10 105 14.6 99.85 408 143
1860 8.7 99.53 167 88
12 000 131 98.9 0 0
15
10
5
36.6 99.69 0 42.3
15.2 99.87 0 59.34
3.8 99.96 0 76.5
10 °C chilling water was used for ammonia recovery.
components NH3, CO2, and H2O in the gas phase is complementary to that of liquid phase. That is to say, if the mass balance in the gas streams can be achieved, the balance of liquid streams can also be maintained. Figure 5 shows how the system runs with the mass balance at steady state. (I) CO2 balance: a small amount of CO2 (1.5 kg/h) was absorbed during the NH3 capture process, while the same flow rate of CO2 was regenerated by high-temperature flue gas in the pretreatment column; (II) NH3 balance: the majority of NH3 recycled to the CO2 absorber together with exhaust NH3 gas was equal to the source NH3 vaporized from CO2 absorber, indicating that almost all of the slipped ammonia was recycled to the CO2 absorber for recapture of CO2; (III) H2O balance: a H2O makeup stream was used to maintain water balance for a steady state of operation because more moisture was vaporized by the hot flue gas and flowed into CO2 absorber. It has to be pointed out that the makeup water consumption is substantial when this process is applied to a full-scale CO2 capture plant. The excessive water enters the CO2 absorber, which will lead to the imbalance of water in the CO2 capture system. One solution to addressing the makeup water consumption and water balance in the entire system is to separate water from the gas leaving the pretreatment column and recycle a certain amount of water as the makeup water to the NH3 absorber. The rest of water is discharged from the system for further treatment before it can be reused. Further research will be carried out to identify the best way for water separation and recycle.
compared to the energy requirement of CO2 capture system 2000−3000 kJ/kg CO2 captured20 4.4. Detailed Analysis of Stream Temperatures and Mass Balance. The temperature profile of each stream is particularly important to understand the heat flow and transport among streams. The streams temperature at steady state in Figure 5 shows how the NH3 recovery process makes
5. CONCLUSIONS In this work, the combination of a validated rate-based model and process simulation have demonstrated that our proposed NH3 abatement and recycling process can successfully solve the problem of ammonia slip in the ammonia-based CO2 capture process. The proposed technology provides the following technical, environmental, and economic advantages: (i) a simplified NH3 abatement system; (ii) over 99.5% ammonia recycling efficiency; (iii) 15.2 ppmv NH3 content in the vent gas; (iv) 59.34 kJ/kg CO2 electricity consumption by making great use of flue gas waste heat; and (v) system mass balance of NH3, CO2 and H2O. Although this advanced process addresses the issue of ammonia loss, NH3-based CO2 capture technology still faces challenges, such as a low CO2 absorption rate and a relatively high energy input for regeneration and chilling the ammonia solution. Future work will focus on (i) increasing aqueous ammonia concentration to improve CO2 absorption performance and reduce the regeneration duty of CO2 desorption; (ii) increasing the temperature of NH3 solvent to room temperature to save the substantial chilling energy input; (iii) performing an energy-integrated study using power station to evaluate the
Figure 5. Mass balance of NH3, CO2, H2O, and streams temperature in the ammonia abatement and recycling system.
use of the heat: (I) the waste heat in high-temperature flue gas was transferred to the NH3-rich solvent for heating up solvent and recovering almost all the captured NH3; (II) the relatively high-temperature solvent from the pretreatment column transported the heat to heat up the solvent from the NH3 absorber through the high-efficient heat exchanger for a better NH3 recycling efficiency. Then the flue gas is cooled to 44.5 °C after the NH3 recovery process, which is a suitable temperature for CO2 capture. It is worth mentioning that the temperature of flue gas to CO2 absorber ranged from 25 to 45 °C, and has a negligible effect on the CO2 capture system according to our simulation results, with a CO2 absorption rate from 87.61 to 87.58 kg/h and an NH3 loss rate from 4.700 to 4.704 kg/h. This indicates that the flue gas cooling process could be removed with the help of this NH3 abatement and recycling system. The mass balance is also particularly important to ensure the normal operation of NH3 abatement and recycling process. On the whole system, the balance of three most important 7185
dx.doi.org/10.1021/es501175x | Environ. Sci. Technol. 2014, 48, 7179−7186
Environmental Science & Technology
Article
(12) Que, H.; Chen, C. Thermodynamic modeling of the NH3 CO2H2O system with electrolyte NRTL model. Ind. Eng. Chem. Res. 2011, 50 (19), 11406−11421. (13) Qi, G. J.; Wang, S. J.; Yu, H.; Wardhaughb, L.; Feron, P.; Chen, C. H. Development of a rate-based model for CO2 absorption using aqueous NH3 in a packed column. Int. J. Greenhouse Gas Control 2013, 17, 450−461. (14) Göppert, U.; Maurer, G. Vapor-liquid equilibria in aqueous solutions of ammonia and carbon dioxide at temperatures between 333 and 393 K and pressures up to 7 MPa. Fluid Phase Equilib. 1988, 41, 153−185. (15) Kurz, F.; Rumpf, B.; Maurer, G. Vapor−liquid−solid equilibria in the system NH3CO2H2O from around 310 to 470 K: New experimental data and modeling. Fluid Phase Equilib. 1995, 104, 261− 275. (16) Yu, H.; Morgan, S.; Allport, A.; Cottrell, A.; Do, T.; Mcgregor, J.; Wardhaugh, L.; Feron, P. Results from trialling aqueous NH3 based post-combustion capture in a pilot plant at Munmorah power station: Absorption. Chem. Eng. Res. Des. 2011, 89, 1204−1215. (17) Onda, K.; Takeuchi, H.; Okumuto, Y. Mass transfer coefficients between gas and liquid phases in packed columns. J. Chem. Eng. Jpn. 1968, 1, 56−62. (18) Chilton, T. H.; Colburn, A. P. Mass transfer (absorption) coefficients prediction from data on heat transfer and fluid friction. Ind. Eng. Chem. 1934, 26, 1183−1187. (19) Stichlmair, J.; Bravo, J. L.; Fair, J. R. General model for prediction of pressure drop and capacity of countercurrent gas/liquid packed columns. Gas Sep. Purif. 1989, 3, 19−28. (20) Jilvero, H.; Normann, F.; Andersson, K.; Johnsson, F. Heat requirement for regeneration of aqueous ammonia in post-combustion carbon dioxide capture. Int. J. Greenhouse Gas Control 2012, 11, 181− 187.
energy requirements of the entire CO2 capture system. We hope that the ammonia-based CO2 capture process can be technically, environmentally, and economically friendly in future commercial use to enable large-scale reductions in CO2 emissions.
■
ASSOCIATED CONTENT
S Supporting Information *
Detailed description of flue gas composition and parameters of different packing materials (Table SI-1 and Table SI-2); total pressure of NH3CO2H2O mixture for various ammonia and CO2 molalities with simulation and experimental result at different temperatures (Figure SI-1); CO2 flow rate profile and temperature profile in the gas and liquid streams (Figure SI-2 and FigureSI-3). This material is available free of charge via the Internet at http://pubs.acs.org/
■
AUTHOR INFORMATION
Corresponding Author
*Tel: +61-2-49606201; fax: +61-2-49606021; e-mail: hai.yu@ csiro.au. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors are grateful for Guojie Qi’s contribution to the model validation and for the Australian IPRS-APA scholarship and CSIRO Top-up scholarship to support K.Li’s research.
■
REFERENCES
(1) Han, K.; Ahn, C. K.; Lee, M. S.; Rhee, C. H.; Kim, J. Y.; Chun, H. D. Current status and challenges of the ammonia-based CO2 capture technologies toward commercialization. Int. J. Greenhouse Gas Control 2013, 14, 270−281. (2) Busca, G.; Pistarino, C. Abatement of ammonia and amines from waste gases: A summary. J. Loss Prevent. Proc. 2003, 16, 157−163. (3) Zhang, M.; Guo, Y. Process simulations of NH3 abatement system for large-scale CO2 capture using aqueous ammonia solution. Int. J. Greenhouse Gas Control 2013, 18, 114−127. (4) Darde, V.; Thomsen, K.; van Well, W. J. M.; Stenby, E. H. Chilled ammonia process for CO2 capture. Int. J. Greenhouse Gas Control 2010, 4, 131−136. (5) Kozak, F.; Petig, A.; Morris, E.; Rhudy, R.; Thimsen, D. Chilled ammonia process for CO2 capture. Energy Proc. 2009, 1, 1419−1426. (6) Yi, K. B.; Kim, J. N. Performance analysis of aqueous ammonia process for CO2 capture. In Presentation Materials at PCC S&T Seminar, CSIRO, 2008. (7) Mathias, P. M.; Reddy, S.; O’Connell, J. P. Quantitative evaluation of the chilled ammonia process for CO2 capture using thermodynamic analysis and process simulation. Int. J. Greenhouse Gas Control 2010, 4, 174−179. (8) Niu, Z. Q.; Guo, Y. C.; Zeng, Q.; Lin, W. Y. A novel process for capturing carbon dioxide using aqueous ammonia. Fuel Proc. Technol. 2013, 108, 154−162. (9) Darde, V.; Mogensen, B. M.; van Well, W. J. M.; Stenby, E. H.; Thomsen, K. Process simulations of CO2 capture with aqueous ammonia using the Extended UNIQUAC model. Int. J. Greenhouse Gas Control 2012, 10, 74−87. (10) Yu, H.; Qi, G. J.; Wang, S. J.; Morgan, S.; Allport, A.; Cottrell, A.; Do, T.; McGregor, J.; Wardhaugh, L.; Feron, P. Results from trialling aqueous ammonia-based post-combustion capture in a pilot at Munmorah Power Station: gas purity and solid precipitation in the stripper. Int. J. Greenhouse Gas Control 2012, 10, 15−25. (11) Aspen Technology. Aspen Physical Property System: Physical Property Models; Burlington, MA, 2010. 7186
dx.doi.org/10.1021/es501175x | Environ. Sci. Technol. 2014, 48, 7179−7186
