Energy Performance of Stripper Configurations for CO2 Capture by

Cellulose, Lignin, Paper, and Other Wood Products · Chemistry of Synthetic High Polymers ... This technology is energy-intensive and has been applied ...
0 downloads 0 Views 416KB Size
Download PDF
Ind. Eng. Chem. Res. 2006, 45, 2457-2464

2457

Energy Performance of Stripper Configurations for CO2 Capture by Aqueous Amines Babatunde A. Oyenekan and Gary T. Rochelle* Department of Chemical Engineering, The UniVersity of Texas at Austin, Austin, Texas 78712

Aqueous absorption/stripping is the state-of-the-art technology for the capture of CO2 from coal-fired power plants. This technology is energy-intensive and has been applied to CO2 removal from natural gas, ammonia, and hydrogen gas streams. Energy requirements can be reduced by the use of a more-reactive solvent, operating the cross exchanger at a lower temperature, optimizing the stripper operation, and using innovative stripper configurations (vacuum and multipressure). This work calculates stripper performance with an algorithm in Aspen Custom Modeler (ACM) that incorporates thermodynamic studies, reaction rate measurements, physical properties, and contactor-specific information for three stripper configurationssa simple stripper operating at 160 kPa, a multipressure stripper operating at three pressures (330/230/160 kPa), and a vacuum stripper (30 kPa) for two solvents: 7m (30 wt %) monoethanolamine (MEA) and 5m K+/2.5m piperazine. The temperature approach is varied from 5 to 10 °C. With some approximations, we predict the influence of using solvents with varying heats of desorption (∆Hdes) on the reboiler duty and the equivalent work for stripping (reboiler duty as equivalent Carnot work plus compression work). With a rich solution giving PCO2* ) 2.5 kPa at 40 °C, the vacuum stripper is favored for solvents with ∆Hdes e 21 kcal/(gmol of CO2) while the multipressure configuration is attractive for solvents with ∆Hdes g 21 kcal/(gmol of CO2). 1. Introduction Aqueous absorption/stripping with aqueous solvents such as alkanolamines and promoted potassium carbonate has been used effectively for removing acid gases (CO2 and H2S) from natural gas, oil refineries, and the production of ammonia and synthesis gas.1-8 Figure 1 shows a typical flow diagram of the process for a simple reboiled stripper. The system consists of two columns: the absorber, in which the CO2 is absorbed into an amine solution via a fast chemical reaction, and a stripper, where the amine is regenerated and then sent back to the absorber for further absorption. Prior to CO2 removal, particulates, sulfur dioxide, and NOx are removed from the flue gas. The flue gas from the power plant is typically cooled before the absorber from 150 to 55 °C (its adiabatic saturation temperature) or to 40 °C if cooling water is used. Typical target CO2 removal efficiency in the absorber is 90%, though efficiencies ranging from 70% to 99% could be achieved in a well-designed absorber. The exiting liquid, the rich solution, is then pumped to the stripper through a cross-heat exchanger, where it is heated by the lean solution from the stripper bottom. If the partial pressure of the CO2 in the rich solution is higher than the operating pressure of the stripper, flashing occurs at the stripper inlet, desorbing some of the CO2. Further desorption occurs within the contactor (a series of trays or height of packing) by normal mass transfer, and some desorption should occur in the reboiler. Gaseous CO2 and water vapor exit the top of the stripper, where water is condensed and thereafter sent to the top stage of the stripper. The lean solution exiting the stripper is sent through the cross-heat exchanger and is cooled before it is returned to the absorber. The dominant operating cost of this system is the steam that is extracted from the power plant to run the reboiler. This reduces the amount of electricity/work produced by the powerplant turbines. Several authors7,9-15 have modeled absorption/ * To whom correspondence should be addressed. Tel.: (512) 4717230. Fax: (512) 475-7824. E-mail: [email protected].

Figure 1. Typical absorber/stripper configuration for 7m MEA (rich [CO2]T ) 3.68m, lean [CO2]T ) 2.39m at 40 °C).

desorption systems. Most of these models predict performance for a single solvent without optimizing the system or comparing solvents. Bishnoi16 showed that the second-order rate constant for piperazine (k225C ) 53 000 L/(mol‚s)) is an order of magnitude greater than that of monoethanolamine (k225C ) 5 000L/(mol‚ s)). Cullinane & Rochelle17 and Cullinane18 found that 5m K+/ 2.5m piperazine (PZ) in a wetted-wall column provides an absorption rate that is 1.5-4× faster than that of 30 wt % MEA and a heat of absorption that is 10-25% less. With this solvent, an absorber/stripper system may use 1.5-2.5× less packing or it may be run at a closer approach to saturation, which could result in energy savings. We used past thermodynamic studies, reaction rate measurements, physical properties, and contactor-specific information to create a stripper model in Aspen Custom Modeler (ACM). Comparisons of reboiler duty and equivalent work for stripping are made between 7m MEA and 5m K+/2.5m PZ and generic solvents. The equivalent work is the sum of compression work and the work content of the reboiler heat estimated from a Carnot cycle. This model helps us understand the stripping operation, reduce energy requirements for stripping, and understand the phenomenon of mass transfer with chemical

10.1021/ie050548k CCC: $33.50 © 2006 American Chemical Society Published on Web 11/11/2005

2458

Ind. Eng. Chem. Res., Vol. 45, No. 8, 2006 Table 1. Adjustable Constants in VLE Expression 2

ln PCO2* ) a + b[CO2]T +

a b c d e f

Figure 2. Multipressure stripper (rich [CO2]T ) 5.33m, lean [CO2]T ) 4.51m at 40 °C).

[CO2]T [CO2]T [CO2]T c +d +e +f 2 2 T T T T

7m MEA

5m K+/2.5m PZ

35.12 -6.43 -14 281 -11 148.5 -485 777 4 667.14

-0.263 0.148 -5 306 -16 995.5 -469 758 2 808

reboiler runs at 60-80 °C. The CO2 is compressed in five intercooled stages to 1000 kPa. Vacuum stripping has the following features: 1. Lower-temperature (less valuable) steam is used to run the reboiler so more electricity can be extracted before the steam is used in the stripper. 2. Additional compression is required for the CO2. 3. The mass transfer is not as fast as that of the simple stripper because the lower temperature results in slower kinetics. 3. Model Development

reactions at stripper conditions. The model creates opportunities in the use of alternative solvents, contactor selection, and innovative stripper designs. 2. Stripper Configurations 2.1. Simple Stripper. In the conventional configuration, the simple reboiled stripper is run at 160 kPa. Pressure drop across the stripper was neglected since it might not be critical for this configuration. The vapor leaving the top of the stripper is cooled and the condensed water is refluxed. The CO2 is compressed in five stages (intercooled to 40 °C) to 1000 kPa. The reboiler runs at 110-120 °C in this configuration. Five compression stages were selected for all configurations. 2.2. Multipressure Stripper. In this configuration (Figure 2), the stripper is divided into three sections, each operating at a different pressure. The CO2 compressor is integrated with the stripper. The vapor from a lower-pressure stage is compressed and subsequently used as stripping vapor in a higher-pressure section. Water vapor condenses with the increased pressure, and the latent heat of water is recovered. This leads to lower reboiler duties, and CO2 is produced at a greater pressure than with the simple (isobaric) stripper. However the compression work is greater than that of the simple stripper, because some water vapor is compressed with the CO2. The pressure levels are 160, 230, and 330 kPa from the bottom to the top of the stripper. The vapor exiting the stripper is cooled, and water is refluxed. The CO2 is further compressed in three stages (intercooled to 40 °C) to 1000 kPa. Therefore, the five compression stages include two integrated with the stripper and three downstream of the stripper. Multipressure stripping has the following features: 1. The latent heat of water is recovered at the rich end. 2. It makes use of the high-temperature preheat in the highpressure flash, thereby rewarding a closer approach temperature in the cross exchanger. 3. CO2 can be recovered at a greater concentration and pressure. This leads to less compression work downstream of the stripper. 4. This configuration should be best with high ∆Hdes solvents such as 7m MEA. 2.3. Vacuum Stripper. This configuration is identical to the simple stripper. The stripper is operated at 30 kPa, and the

A stripper model for aqueous solutions of 7m MEA and of 5m K+/2.5m PZ was developed in ACM. This model divides the stripper into 10 sections with Murphree efficiencies assigned to CO2, water, and temperature. In the multipressure configuration, four sections are at 160 kPa, four are at 230 kPa, and two are at 330 kPa. An empirical expression with six adjustable constants was used to represent the vapor-liquid equilibrium (VLE) and heat of absorption/desorption. The three stripper configurations were modeled with variations of the rich and lean CO2 loading, temperature approach (5-10 °C), and stripper operating pressure. The equivalent work consumed by the process was calculated. 3.1. Modeling Assumptions. (a) The sections are well-mixed in the liquid and vapor phases. (b) The reaction takes place in the liquid phase. (c) The reboiler is assumed to be in equilibrium. (d) There is negligible vaporization of the amine. (e) The top flash was 40% efficient. The flash region in the column is being quantified in terms of actual section performance. The CO2 vapor pressure (kPa) under stripper conditions was represented by the empirical expression in Table 1. The total CO2 concentration, [CO2]T, is the concentration of CO2 in all forms including free CO2, bicarbonate, carbonate, and carbamate. The adjustable constants in Table 1 were obtained by regressing the points from the rigorous model for 5m K+/2.5m PZ by Cullinane18 and for 30 wt % MEA using equilibrium flashes in AspenPlus based on the rigorous model developed by Freguia19 from the data of Jou et al.20 The heat of absorption/desorption is calculated by differentiating the equation in Table 1 with respect to 1/T: 2

-

[CO2]T [CO2]T ∆H ) c + 2d + 2e + f[CO2]T R T T

(1)

Table 2 gives calculated values for the equilibrium CO2 partial pressures (PCO2*) for both solvents at 40 and 60 °C as a function of CO2 loading. The rich-solvent composition can be expressed as total CO2, [CO2]T, CO2 loading [(mol of CO2)/((mol of MEA) + (mol of K+) + (mol of PZ))] or PCO2*. Since typical flue gas has 12 kPa of CO2, it should be possible to achieve PCO2* of 2.5-10 kPa at typical absorber temperatures of 40-60 °C. The

Ind. Eng. Chem. Res., Vol. 45, No. 8, 2006 2459 Table 2. Predicted CO2 Solubility at Absorber Conditions CO2 loading

[CO2]T (m) solvent

[

(mol of CO2)

(mol of MEA) + (mol of K+) + (mol of PZ)

rich PCO2* (kPa)

60 °C

40 °C

60 °C

40 °C

1.25 2.5 5 10 1.25 2.5 5 10

2.73 2.99 3.26 3.53 4.49 4.72 4.92 5.21

3.43 3.68 3.94 4.21 5.05 5.33 5.61 5.91

0.390 0.427 0.466 0.504 0.599 0.629 0.656 0.695

0.490 0.526 0.563 0.601 0.673 0.711 0.748 0.788

7m MEA

5m K+/2.5m PZ

]

heat of vaporization of water, the partial pressure of water, and the heat capacities of steam, CO2, and the solvent (essentially water) were calculated with equations from the Design Institute for Physical Properties (DIPPR) database.21 The partial pressures of CO2 and water in each section were calculated by

Pn ) Emv(Pn* - Pn-1) + Pn-1

(2)

Murphree efficiencies (Emv) of 40% and 100% were assigned to CO2 and water, respectively. The model assumed that temperature equilibrium is achieved in each section. The model inputs were the rich [CO2]T and liquid rate (1 kg/s), the temperature approach in the cross exchanger (difference between the temperature of the rich stripper feed and the lean solution leaving the bottom of the stripper), and the column pressure. Initial guesses of the lean [CO2]T, section temperatures, partial pressures, and concentrations were provided. The model solves equations for calculating VLE and for material and energy balances. It calculates temperature and composition profiles, reboiler duty, and equivalent work. The total energy required by the stripper is given as total equivalent work:

Weq ) 0.75Qreb

[

]

(Treb + 10) - 313 (Treb + 10)

+ Wcomp

Figure 3. Optimized lean loading for minimum equivalent work with 7m MEA (rich [CO2]T ) 3.68m, lean [CO2]T ) 2.39m, simple stripper, ∆T ) 10 °C, abs. rich T ) 40 °C).

(3)

Wcomp constitutes the isentropic work of compression of the gas exiting the top of the stripper to 1000 kPa. An efficiency of 75% was assumed for the compressor. The first term in eq 3 accounts for the electricity generation lost by extracting steam from a turbine, while the second term is the compressor work. The condensing temperature of the steam is assumed to be 10 K higher than that of the reboiler fluid. The turbine assumes condensing steam at 313 K. 4. Results and Discussion 4.1. Predicted Stripper Performance. The optimization of the lean loading in a simple stripper using 7m MEA for a rich CO2 loading of 0.525 (mol of CO2)/(mol of MEA) (3.68m) is shown in Figure 3. The minimum equivalent work (8.01 kcal/ (gmol of CO2)) occurs at a CO2 loading of 0.33 (mol of CO2)/ (mol of MEA) (2.39m) with a reboiler duty of 33.5 kcal/(gmol of CO2). The lean loading required to minimize reboiler duty does not coincide with that required to minimize equivalent work. The equilibrium partial pressure of CO2 in the rich solution leaving the absorber is 2.5 kPa. The lean partial pressure leaving the stripper bottom is 0.11 kPa at 40 °C. This implies that >90% removal can be achieved with the equivalent work minimized.

Figure 4. Reboiler duty for different configurations with 7m MEA (∆T ) 10 °C, abs. rich T ) 40 °C).

The reboiler duty with the lean loading required to minimize the total equivalent work for 7m MEA using the three configurations is shown in Figure 4. The multipressure stripper gives the least reboiler duty for all rich CO2 partial pressures. The reboiler duty is reduced by 20-27%, depending on the rich partial pressure of CO2, when the multipressure stripper is

2460

Ind. Eng. Chem. Res., Vol. 45, No. 8, 2006

Figure 5. Total equivalent work for different configurations with 7m MEA (∆T ) 10 °C, abs. rich T ) 40 °C).

Figure 6. Reboiler duty for different configurations with 5m K+/2.5m PZ (∆T ) 10 °C, abs. rich T ) 40 °C).

used as compared to the simple stripper. With the vacuum stripper, the reboiler duty is 22-62% more than that with the simple stripper. Figure 5 shows the minimum equivalent work for 7m MEA using the three configurations. The multipressure stripper gives the least equivalent work over the entire rich PCO2* range. The simple stripper is the least attractive configuration with the highest work over most of the rich PCO2* range. The multipressure stripper offers 8% energy savings when compared to the simple stripper. The vacuum stripper requires 6% less energy at high rich PCO2*. The reboiler duty with the lean loading required to minimize the total equivalent work for 5m K+/2.5m PZ using the three configurations is shown in Figure 6. The multipressure stripper gives the least reboiler duty for all rich CO2 partial pressures. The reboiler duty is reduced by 12-32%, depending on the rich partial pressure of CO2, when the multipressure stripper is used as compared to the simple stripper. With the vacuum stripper, the reboiler duty is 20-50% more than that with the simple stripper.

Figure 7. Total equivalent work for different configurations with 5m K+/ 2.5m PZ (∆T ) 10 °C, abs. rich T ) 40 °C).

Figure 8. Total equivalent work for different ∆T for 5m K+/2.5m PZ for a simple stripper (abs. rich T ) 40 °C).

Figure 7 shows the minimum equivalent work for 5m K+/ 2.5m PZ using the three configurations. The vacuum stripper gives the least equivalent work over most of the rich PCO2* range, with the multipressure stripper competitive at higher rich PCO2*. The simple stripper is the least attractive configuration at high rich PCO2*. In comparison to the simple stripper, the vacuum stripper requires 18% less energy at a lower rich PCO2* and offers savings up to 8% at higher PCO2*. Operating the stripper using 5m K+/2.5m PZ with a closer temperature approach, 5 °C instead of 10 °C, offers 2-6% savings over the practical range of rich PCO2*, as evident from Figure 8. An economic analysis is desirable before additional investment in the heat transfer area is made. Comparing the equivalent work for 7m MEA and 5m K+/ 2.5m PZ at fixed rich PCO2* shows that 5m K+/2.5m PZ gives less equivalent work at richer PCO2* (PCO2* > 5 kPa). This suggests that the solvent could be an alternative to 7m MEA at higher concentrations, where both energy savings and capital cost savings as a result of its faster nature are attractive. The condition of the stripper feed determines the mechanism of stripping and affects column profiles. McCabe-Thiele diagrams for the three configurations were constructed to give

Ind. Eng. Chem. Res., Vol. 45, No. 8, 2006 2461

Figure 9. McCabe-Thiele plot for 5m K+/2.5m PZ for a simple stripper (rich [CO2]T ) 5.33m, lean [CO2]T ) 4.46m, Qreb ) 36.4 kcal/(gmol of CO2), Weq ) 8.8 kcal/(gmol of CO2), ∆T ) 10 °C, abs. rich T ) 40 °C).

Figure 10. McCabe-Thiele plot for 5m K+/2.5m PZ for a multipressure stripper (rich [CO2]T ) 5.33m, lean [CO2]T ) 4.51m, Qreb ) 28.5 kcal/ (gmol of CO2), Weq ) 8.2 kcal/(gmol of CO2), ∆T ) 10 °C, abs. rich T ) 40 °C).

insight into different phenomena. Figure 9 shows the McCabeThiele plot for a simple stripper using 5m K+/2.5m PZ. The rich feed is a supersaturated liquid with a CO2 loading of 0.711 mol/((mol of K+) + (mol of PZ)) and a temperature of 101 °C. The liquid flashes to 97 °C at the stripper inlet. Its CO2 loading decreases to 0.686 mol/((mol of K+) + (mol of PZ)) by the time it leaves the first section. The liquid temperature increases steadily to the reboiler. The driving force (PCO2* - PCO2) suggests a pinch at the rich end. The pinch experienced may be as a result of the large contacting capability inherent in the model. Very little change in loading occurs over three sections, so the column could provide almost equivalent performance with seven sections rather than ten. After the top three sections, there is an evenly distributed driving force in the column, which should correspond to good mass transfer. Figure 10 shows the McCabe-Thiele plot for a multipressure stripper (330/230/160 kPa) using 5m K+/2.5m PZ. The pressure levels decrease from the top (rich end) to the bottom (lean end). The rich feed is a subcooled liquid with a CO2 loading of 0.711 mol/((mol of K+) + (mol of PZ)) and a temperature of 100 °C.

Figure 11. McCabe-Thiele plot for 5m K+/2.5m PZ for a vacuum stripper (rich [CO2]T ) 5.33m, lean [CO2]T ) 4.64m, Qreb ) 45.1 kcal/(gmol of CO2), Weq ) 7.65 kcal/(gmol of CO2), ∆T ) 10 °C, abs. rich T ) 40 °C).

The large flow of subcooled liquid condenses water, and CO2 absorption occurs initially in the top of the column at 330 kPa. CO2 loading increases to 0.726 mol/((mol of K+) + (mol of PZ)), and the liquid temperature increases to 102.5 °C by the time it leaves the first section. The drop in temperature in moving between pressure sections results from flashing. The McCabe-Thiele plot for a vacuum stripper (30 kPa) using 5m K+/2.5m PZ is shown in Figure 11. The rich feed is a subcooled liquid with a CO2 loading of 0.711 mol/((mol of K+) + (mol of PZ)) and a temperature of 57 °C. On entering the column, the liquid is heated quickly. There is no apparent pinch in the column as there is an evenly distributed driving force. 4.2. Mass Transfer Calculations. Some mass transfer calculations were performed using IMTP #40 packing. This is random metal packing with a nominal diameter of 0.04 m, a dry area-to-volume ratio of 165 m2/m3, and a packing factor of 79 m-1. Mass transfer calculations are essential to determine the height of packing required for a specific separation as well as to determine the reaction mechanisms. The cross-sectional area of the column was 0.147 m2, which is the diameter of a pilot plant on which simultaneous testing is being carried out. The mass transfer model used was that originally developed by Bishnoi16 and modified for a potassium carbonate/piperazine solution by Cullinane.17 The model is a rigorous rate model based on the eddy diffusivity theory.22 It integrates a series of differential equations for the thermodynamics in the bulk liquid using the electrolyte non-random two liquid (ENRTL) model proposed by Chen and co-workers,23-25 the diffusion across the liquid film, and the reaction in the boundary layer, and it calculates the liquid-phase mass transfer coefficient with a partial-pressure driving force, kg′. The liquid-phase mass transfer coefficient consists of kinetic and diffusion terms. The flux of CO2 is given by the expression

NCO2 ) KG (PCO2* - PCO2)

(4)

The overall mass transfer coefficient (KG) is the sum of the gas phase (kg) and liquid phase (kg′) components.

1 1 1 ) + KG kg kg′

(5)

2462

Ind. Eng. Chem. Res., Vol. 45, No. 8, 2006 Table 3. Constants for Generic Solvents VLE Expression

Figure 12. Rate in stripper for 5m K+/2.5m PZ with a simple stripper (rich [CO2]T ) 5.33m, lean [CO2]T ) 4.46m, ∆T ) 10 °C, abs. rich T ) 40 °C).

∆Hdes (kcal/(gmol of CO2))

a

15 20 22 25 27 35 40

8.92 16.48 19.50 24.03 27.05 39.14 46.69

Figure 13 shows the minimum total equivalent work for the generic solvents using the three configurations. The vacuum stripper requires the least equivalent work with solvents with ∆Hdes e 21 kcal/(gmol of CO2), while the multipressure stripper requires the least equivalent work for solvents with ∆Hdes g 21kcal/(gmol of CO2). The reboiler duty for generic solvents at a rich PCO2* ) 2.5 kPa at 40 °C using a vacuum stripper is shown in Figure 14. In these cases, the lean loading is adjusted to minimize the total

kg is obtained from Onda et al.26 and Wilson.27 kg′ is obtained from Cullinane.18 The CO2 desorption rate is

rate ) KGA(PCO2* - PCO2)

(6)

The wetted area of contact, A, depends on the equipment and hydraulics in the column. The volume of a section is found by dividing the wetted area of the column by the specific wetted area of the packing. The height of a segment is found by dividing the volume of the segment by the cross-sectional area of the column. The height of packing required for a particular operation is found by summing the heights of all the segments. Figure 12 shows the mass transfer rate modeling results for a simple stripper with a rich PCO2* ) 2.5 kPa leaving the absorber at 40 °C. The fractional gas resistance to mass transfer ranges from 62% at the rich end to 86% at the lean end. This shows that the relative contribution of the gas-phase resistance to the overall mass transfer coefficient increases as we go from the top (rich end) to the bottom (lean end) of the column. The fraction gas-phase resistance decreases because the liquid-phase mass transfer coefficient, kg′, decreases from 2.46 × 10-8 kmol/ (Pa‚m2‚s) at the lean end to 3.12 × 10-9 kmol/(Pa‚m2‚s) at the rich end. The greater amount of free amine at the lean end leads to an increased rate from the rich to the lean end. The small driving force (PCO2* - PCO2) suggests a pinch at the rich end. Since very little change in CO2 loading occurs in the top three sections, the stripper can be run with seven sections without compromising performance. For a 16.8 in. diameter column, based on model results, we would need 30 m of packing to achieve this separation; however, since two-thirds of the column is pinched, we could actually run the stripper with 9 m of packing.

Figure 13. Total equivalent work for generic solvents (rich PCO2 ) 2.5 kPa at 40 °C, ∆T ) 10 °C).

5. Generic Solvent Modeling A three-parameter expression for the vapor-liquid equilibrium was used to model generic solvents:

ln P ) a + b[CO2]T -

∆H RT

(7)

The constant b was set to 3.07, while the constant a was varied. The value of the constant a used in eq 7 for the generic solvents is shown in Table 3.

Figure 14. Reboiler duty for generic solvents (rich PCO2 ) 2.5 kPa at 40 °C, ∆T ) 10 °C).

Ind. Eng. Chem. Res., Vol. 45, No. 8, 2006 2463 Table 4. Optimum ∆HDes to Minimize Energy Requirement (Absorber Rich T ) 40 °C) optimum ∆Hdes (kcal/(gmol of CO2)) equivalent work

reboiler duty

rich PCO2* (kPa)

simple

multi P

vacuum

simple

multi P

vacuum

1.25 2.5 5 10

40 40 40 40

40 40 40 40

27 22 20 15

22 20 15 e15

20 e15 e15 e15

26 23 19 e15

Table 5. Optimum Capacity that Minimizes Equivalent Work for Different Solvents and Stripper Configurations. (Absorber Rich T ) 40 °C, Rich PCO2* ) 2.5 KPa,TApp ) 10 °C) optimum capacity (m) 7m MEA 5m K+/2.5m PZ generic solvents ∆Hdes (kcal/gmol) 15 20 22 25 27 35 40

simple

multi P

vacuum

1.29 0.87

1.26 0.82

0.80 0.69

0.85 1.39 0.75 0.71 0.70 0.66 0.64

0.83 1.19 0.97 0.85 0.82 0.75 0.72

0.55 0.75 0.84 0.94 0.71 0.57 0.54

equivalent work for each value of ∆Hdes. This minimum reboiler duty occurs at ∆Hdes e 15 kcal/(gmol of CO2) for the multipressure stripper, at ∆Hdes ≈ 20 kcal/(gmol of CO2) for the simple stripper, and at ∆Hdes ≈ 23 kcal/(gmol of CO2) for the vacuum stripper. The temperature swing in moving from the absorber to the stripper is only advantageous in reducing equivalent work for high ∆Hdes solvents. Table 4 shows the optimum ∆Hdes of a solvent required to minimize reboiler duty and equivalent work for different configurations and rich PCO2*. For the simple and multipressure configurations, equivalent work is minimized by the solvent with ∆Hdes ) 40 kcal/(gmol of CO2). For a rich PCO2* of 2.5 kPa at 40 °C, the lean PCO2* is 0.44 kPa (83% removal) for the simple stripper and 0.31 kPa (87% removal) for the multipressure stripper. At a rich PCO2* of 2.5 kPa at 40 °C, the lean PCO2* with the vacuum stripper is 0.25 kPa (90% removal). The ∆Hdes for a solvent to reduce equivalent work using the vacuum stripper decreases with increasing rich PCO2*. The corresponding reboiler duty decreases with an increase in rich PCO2* for all configurations. A comparison of the optimum operating capacity of the different solvents was carried out. Table 5 compares the optimum operating capacity to minimizes equivalent work for different solvents and stripper configurations. The capacity of a solvent is defined as

capacity (m) ) rich [CO2]T (m) - lean [CO2]T (m) (8) The results show that 7m MEA has a greater capacity than 5m K+/2.5m PZ for the three configurations. This capacity is a property of the solvent. The constant, b, in eq 7 is an indication of the capacity of the different generic solvents. For a fixed inherent capacity (i.e., constant b), the actual maximum capacity is seen to correspond to the solvent that minimizes reboiler duty, as in Figure 14. This capacity is the optimum operating capacity of the solvent. We can hypothesize that the solvent with the maximum capacity for each of the stripper configurations will minimize the reboiler duty that corresponds to the minimum equivalent work.

Figure 15. Equivalent work for generic solvents for a vacuum stripper, ∆T ) 10 °C.

Figure 15 shows the minimum equivalent work for the generic solvents using a vacuum stripper with rich PCO2* ) 1.25-10 kPa. Higher ∆Hdes solvents give the least equivalent work at lower rich PCO2*. 6. Conclusions 1. For 7m MEA, the multipressure configuration gives the least equivalent work at a fixed absorber rich PCO2*. 2. For 5m K+/2.5m PZ, the vacuum stripper is the most attractive option. Lower temperatures, 60-80 °C, will help reduce corrosion and allow for alternative materials of construction like fiber-reinforced plastic. 3. With 5m K+/2.5m PZ, operating the cross exchanger with a 5 °C approach offers 2-6% energy savings over one with a 10 °C approach. 4. For generic solvents, the optimum ∆Hdes for the solvent that will minimize equivalent work is a function of the stripper configuration used. The vacuum stripper is favored for solvents with ∆Hdes e 21 kcal/(gmol of CO2), while the multipressure configuration is attractive for solvents with ∆Hdes g 21 kcal/ (gmol of CO2) at a rich PCO2* ) 2.5 kPa when the rich absorber temperature is 40 °C. Acknowledgments Aspen Technology provided the Aspen Plus and Aspen Custom Modeler software. This paper was prepared with the support of the U.S. Department of Energy (DOE), under Award No. DE-FC26-02NT41440. However, any opinions, findings, conclusions, or recommendations expressed herein are those of the authors and do not necessarily reflect the views of the DOE. Nomenclature A ) wetted area of contact between gas and liquid phases (m2) [CO2]T ) total concentration of CO2 (molal) Emv ) Murphree section efficiency defined in terms of partial pressures ∆H ) heat of absorption/desorption (kcal/(gmol of CO2)) kg ) gas-phase mass transfer coefficient (kmol/(Pa‚m2‚s)) kg′ ) liquid-phase mass transfer coefficient based on a partial pressure driving force (kmol/(Pa‚m2‚s))

2464

Ind. Eng. Chem. Res., Vol. 45, No. 8, 2006

KG ) overall mass transfer coefficient based on the gas phase (kmol/(Pa‚m2‚s)) ldg ) CO2 loading [(mol of CO2)/((mol of MEA) + (mol of K+) + (mol of PZ))] m ) molality ((mol of CO2)/(kg of H2O)) NCO2 ) flux of CO2 (kmol/(m2‚s)) PCO2 ) partial pressure of CO2 in the bulk gas (kPa) PCO2* ) equilibrium partial pressure of CO2 (kPa) Pn ) partial pressures on sections n (kPa) Pn-1 ) partial pressures sections n - 1 (kPa) Pn* ) equilibrium partial pressure leaving section n (kPa) Qreb ) reboiler duty (kcal/(gmol of CO2)) R ) universal gas constant (cal/(K‚mol)) T ) temperature (K) Wcomp ) isentropic work of compression (kcal/(gmol of CO2)) Weq ) equivalent work (kcal/(gmol of CO2)) Literature Cited (1) Benson, H. E.; Field, J. H.; Haynes, W. P. Improved process for CO2 absorption uses hot potassium carbonate solutions. Chem. Eng. Prog. 1956, 10, 433-438. (2) Benson, H. E.; Field, J. H.; Jimeson, R. M. CO2 absorption employing hot potassium carbonate solutions. Chem. Eng. Prog. 1954, 50, 356-364. (3) Bottoms, R. R. Organic bases for gas purification. Ind. Eng. Chem. 1931, 23, 501-504. (4) Savage, D. W.; Astarita, G.; Joshi, S. Chemical absorption and desorption of carbon dioxide from hot carbonate solutions. Chem. Eng. Sci. 1980, 35, 1513-1522. (5) Shier, A. L.; Danckwerts, P. V. Carbon dioxide absorption into amine-promoted potash solutions. Ind. Eng. Chem. Fundam. 1969, 8, 415423. (6) Teller, A. J.; Ford, R. E. Packed column performance of carbon dioxide-monoethanolamine system. Ind. Eng. Chem. 1958, 50, 1201-1206. (7) Trass, O.; Weiland, R. H. Absorption of carbon dioxide in ethylenediamine solutions. II. Pilot plant study of absorption and regeneration. Can. J Chem. Eng. 1971, 49, 773-781. (8) Wilson, M.; Tontiwachwuthikul, P.; Chakma, A.; Idem, R.; Veawab, A.; Aroonwilas, A.; Gelowitz, D.; Barrie, J.; Mariz, C. Test results from a CO2 extraction pilot plant at boundary dam coal-fired power station. Energy 2004, 29, 1259-1267. (9) Alatiqi, I.; Sabri, M. F.; Bouhamra, W.; Alper, E. Steady-state ratebased modelling for CO2/amine absorption-desorption systems. Gas Sep. Purif. 1994, 8, 3-11. (10) Alie, C.; Backham, E.; Croiset, E.; Douglas, P. L. Simulation of CO2 capture using MEA scrubbing: A flowsheet decomposition method. Energy ConVers. Manage. 2005, 46, 475-487. (11) Desideri, U.; Paolucci, A. Performance modeling of a carbon dioxide removal system for power plants Energy ConVers. Manage. 1999, 40, 1899-1915.

(12) Freguia, S.; Rochelle, G. T. Modeling of CO2 capture by aqueous monoethanolamine. AIChE J. 2003, 49, 1676-1686. (13) Kucka, L.; Muller, I.; Kenig, E. Y.; Gorak, A. On the modeling and simulation of sour gas absorption using aqueous amine solutions Chem. Eng. Sci. 2003, 58, 3571-3578. (14) Suenson, M. M.; Georgarkis, C.; Evans, L. B. Steady state and dynamic modeling of a gas absorber-stripper system. Ind. Eng. Chem. Res. 1985, 24, 288-295. (15) Weiland, R. H.; Rawal, M.; Rice, C. G. Stripping of carbon dioxide from monoethanolamine solutions in a packed column. AIChE J. 1982, 28, 963-973. (16) Bishnoi, S. Carbon dioxide absorption and solution equilibrium in piperazine activated methyldiethanolamine. Ph.D. Dissertation, The University of Texas at Austin, Austin, TX, 2000. (17) Cullinane, J. T.; Rochelle, G. T. Carbon dioxide absorption with aqueous potassium carbonate promoted by piperazine. Chem. Eng. Sci. 2004, 59, 3619-3630. (18) Cullinane, J. T. Thermodynamics and kinetics of aqueous piperazine with potassium carbonate for carbon dioxide absorption. Ph.D. Dissertation, Department of Chemical Engineering, The University of Texas at Austin, Austin, TX, 2005. (19) Freguia, S. Modeling of CO2 removal from flue gases with monoethanolamine. M.S. Thesis, The University of Texas at Austin, Austin, TX, 2002. (20) Jou, F. Y.; Mather, A. E.; Otto, F. D. The solubility of CO2 in a 30 mass percent monoethanolamine solution. Can. J. Chem. Eng. 1995, 73, 140-147. (21) Design Institute for Physical Properties (DIPPR). American Institute of Chemical Engineers: New York, 2004. (22) King, C. J. Turbulent liquid-phase mass transfer at a free gasliquid interface. Ind. Eng. Chem. Fundam. 1966, 5, 1-8. (23) Chen, C.; Evans, L. B. A local composition model for excess Gibbs energy of aqueous electrolyte systems. AIChE J. 1986, 32, 444-454. (24) Chen, C.; Britt, H. I.; Boston, J. F.; Evans, L. B. Local composition model for excess Gibbs energy of electrolyte systems. Part I: Single solvent, single completely dissociated electrolyte systems. AIChE J. 1982, 28, 588596. (25) Mock, B.; Evans, L. B.; Chen, C. Thermodynamic representation of phase equilibria of mixed solvent electrolyte systems. AIChE J. 1986, 32, 1655-1664. (26) Onda, K.; Takeuchi, H.; Okumoto, Y. Mass transfer coefficients between gas and liquid phases in packed columns. J. Chem. Eng. Jpn. 1968, 1. (27) Wilson, I. Gas-liquid contact area of random and structured packing. M.S. Thesis, The University of Texas at Austin, Austin, TX, 2004.

ReceiVed for reView May 11, 2005 ReVised manuscript receiVed October 10, 2005 Accepted October 12, 2005 IE050548K