Removal of CO2 by Single and Blended Aqueous Alkanolamine

2576

Ind. Eng. Chem. Res. 2007, 46, 2576-2588

Removal of CO2 by Single and Blended Aqueous Alkanolamine Solvents in Hollow-Fiber Membrane Contactor: Modeling and Simulation Subham Paul, Aloke K. Ghoshal,* and Bishnupada Mandal† Department of Chemical Engineering, Indian Institute of Technology Guwahati, Guwahatis781039, Assam, India

A theoretical analysis to capture CO2 using different aqueous single and blended alkanolamine solutions in a hollow-fiber membrane contactor has been performed. The aqueous solutions of monoethanolamine (MEA), diethanolamine (DEA), N-methyldiethanolamine (MDEA), and 2-amino-2-methyl-1-propanol (AMP), as well as aqueous blends of MEA or DEA with AMP or MDEA are considered. The aqueous solution of MEA has the highest CO2 absorption flux among the single amine solutions. For the absorption in the blends of (MEA + MDEA), (DEA + MDEA), and (MEA + AMP), the flux increases as the concentration of MEA or DEA increases. For the (DEA + AMP) system, the flux of CO2 is observed to be almost constant with different blend compositions. The absorption fluxes of CO2 in the blends of (MEA + MDEA) and (MEA + AMP) are higher, compared to those in other blends. 1. Introduction There is a growing concern about the removal of CO2, which has been considered to be the major greenhouse gas. The major sources of CO2 emission are the combustion of fossil fuel, natural gas, and refinery offgases. The alkanolamine technology is the most widely used procedure to capture CO2 in conventional equipment such as packed, spray, or bubble column absorption towers. Approximately 90% of the acid-gas-treating processes in operation today use alkanolamine solvents, because of their versatility and their ability to remove acid gases to very low levels. Alkanolamines with commercial significance for the acid-gas-treating processes include monoethanolamine (MEA), diethanolamine (DEA), di-2-propanolamine (DIPA), and Nmethyldiethanolamine (MDEA). Sterically hindered amines such as 2-amino-2-methyl-1-propanol (AMP) have also been proposed as important candidates for acid-gas purification, because they have not only greater capacity for CO2 but also a lower tendency to form carbamate and an appreciable absorption rate for CO2.1 The advantages of sterically hindered amines result from the bulkiness of the substituent attached to the amino group, which causes an increased cyclic capacity and increased energy consumption in industrial operations. Blends of primary or secondary amines with a tertiary or sterically hindered amine combine the higher CO2 reaction rates of the primary or secondary amine with the higher CO2 loading capacity of the tertiary or sterically hindered amine.2 Phase dispersion and limited mass-transfer area are the major drawbacks of the conventional equipment. Microporous hollowfiber membrane contactors (HFMCs) can overcome the disadvantages of the conventional equipment when incorporated into the acid-gas-treating processes.3 One characteristic of microporous membrane contactors is that the gas streams flow on one side and the absorbent liquid flows on the other side of the membrane without phase dispersion, thus avoiding the problems that are often encountered in the conventional equip* To whom correspondence should be addressed. Tel.: 91-3612582251 (O)/2584252 (R). Fax: 91-361-2690762. E-mail: aloke@ iitg.ernet.in. † Present address: Department of Chemical & Biomolecular Engineering, The Ohio State University, 140 West 19th Ave., Columbus, OH 43210-1180, USA. E-mail: [email protected].

ment, such as flooding, foaming, and entrainment. The compact module of the membrane contactor also provides much-larger gas/liquid interfaces with known area at the pore mouth of the membrane and also helps to facilitate easy scaleup or scaledown.4 However, a disadvantage of the membrane contactor is the presence of additional diffusional resistance through the gasfilled membrane pores.5 Given the extensive use of alkanolamines for the absorption of acid gases in industry and the substantial energy requirement of acid-gas-treating plants, there is considerable incentive for the development of more-efficient and more-flexible methods for acid-gas separation. The absorption of CO2 in different single and blended alkanolamine solvents using conventional gas-liquid contactors has been studied by several researchers.2,6-20 However, there are few literature reports available for the absorption of CO2 in alkanolamine solvents using HFMCs. Zhang et al.21 studied the absorption of CO2 in aqueous DEA solution using HFMCs. They built models for the absorption of CO2 with varying CO2 concentration in the gas phase. Wang et al.4 theoretically studied the absorption of CO2 in HFMCs using three typical alkanolamines solutions of AMP, DEA, and MDEA. Their simulation results indicate that AMP and DEA solutions have much higher CO2 absorption fluxes than MDEA solution. Gong et al.22 have reported the experiments and simulation of CO2 removal using aqueous blends of MDEA and MEA in HFMCs. The fractional removal of CO2 and CO2 absorption flux increase as the MEA content in the blend increases. With the advancement of membrane research, membrane contactors are attracting considerable attention to be used for acid-gas treatment, exploiting the advantages of the HFMCs as discussed. Therefore, researchers worldwide are starting to devote considerable attention to membrane contactors, for the purpose of developing a potential method to capture CO2 with reduced energy consumption. Because there are several single and blended alkanolamines already used for CO2 absorption in different traditional contactors, it has become necessary to analyze the techno-economic aspects of the amine systems that will guide the selection of a proper alkanolamine solvent to capture CO2 in HFMCs. To the best of our knowledge, detailed theoretical analysis on the performances of the HFMC, particularly for blended amines, toward absorption of CO2 have not yet been reported in the literature. In the present work, a

10.1021/ie061476f CCC: $37.00 © 2007 American Chemical Society Published on Web 03/17/2007

RA )

k2,AmineCACAmine

∑

1

1+ [(kH2O/k-1)CH2O] +

∑ [(kAmine/k-1)CAmine]

(4)

(The derivation is given in the Appendix.) However, a tertiary amine cannot undergo reactions 1 and 2. The reaction mechanism is essentially a base-catalyzed hydration of CO2, forming a protonated amine and a bicarbonate anion:14,15 k2,Amine

R1R2R3N + H2O + CO2 98 R1R2R3NH+ + HCO3 (5)

7.0 × 10-4 1.90 × 10-5

1.50 × 10-2

8.0 × 10-4

1.58 × 10-3 2.58 × 10-1 7.23 × 10-4 1.84 × 10-3 9.58 × 10-6 9.80 × 10-5 1.68 × 10-6 1.11 × 10-7

4.34 × 10-4 6 × 10-2 3.54 × 10-4 4.83 × 10-3 1.58 × 10-3 0.437 × 10-3 2.335 × 10-3 9.58 × 10-6 2.20 × 10-6 2.64 × 10-6

where the reaction rates on the right-hand side of the equation are CO2 with the amine, OH-, and H2O, respectively. The reaction of CO2 with OH- and H2O can be neglected, because of their weak contribution.10-12 Based on the assumption of quasi-steady-state conditions for the zwitterion concentration, the rate of CO2 reaction with primary or secondary amines is expressed by

6.358 × 100 2.375 × 100 8.1 × 10-1 5.21 × 10-3 6.358 × 100 5.60 × 100 3.13 × 100 8.3 × 10-1

(3)

MEA(B) + H2O DEA(B) + H2O AMP(B) + H2O MDEA(B) + H2O MEA(B) + MDEA(C) + H2O MEA(B) + AMP(C) + H2O DEA(B) + MDEA(C) + H2O DEA(B) + AMP(C) + H2O

+ RA-H2O

k2,BkB/k-1 (m6 mol-2 s-1)

-

Table 1. Kinetic Parameters (at 298 K) Used in Simulation

∑ RA-Amine) + RA-OH

RA ) (

k2,BkC/k-1 (m6 mol-2 s-1)

where R1 is an alkyl group and R2 is H for primary amines and an alkyl group for secondary amines; b is a base that could be an amine, OH-, or H2O, although the contribution of OH- can be neglected because its concentration is very low, compared with that of the amine and H2O.13 Any base present in the solution may contribute to the deprotonation of zwitterion. The contribution of each base would be dependent on its concentration as well as on the strength of the base. For the absorption of CO2 into aqueous alkanolamine solvents, the overall reaction rate for CO2 can be expressed as follows:

5.41 × 10-3 5.6 × 10-1 5.21 × 10-3 5.5 × 10-1

k2,C (m3 mol-1 s-1)

(2)

k2,BkH2O/k-1 (m6 mol-2 s-1)

kb

R1R2NH+COO- + b 98 R1R2NCOO- + bH+

(1)

k2,B (m3 mol-1 s-1)

-1

system

k2,Amine

k2,CkH2O/k-1 (m6 mol-2 s-1)

2.1. Reaction Mechanism of CO2 with Amines. The proposed mechanism for the reaction between CO2 with MEA, AMP, and DEA involves the formation of zwitterion as in eq 1, followed by the deprotonation of zwitterion by a base to produce carbamate and a protonated base, according to eq 2:12,19,23-25

1.80 × 10-5

k2,CkC/k-1 (m6 mol-2 s-1)

2. Modeling of CO2 Capture in Hollow-Fiber Membrane Contactors (HFMCs)

CO2 + R1R2NH y\ z R1R2NH+COOk

9.08 × 10-2

k2,CkB/k-1 (m6 mol-2 s-1)

reference(s)

numerical simulation to capture CO2 in HFMCs using different single and blended aqueous alkanolamines has been presented as a part of techno-economic aspect. The emphasis of the present work is to apply the fundamental conservation equation (the component mass balance, for the present case) to understand the local and average fluxes during the absorption of CO2 in different single amines. The same is being extended for different blends of amines to estimate the average flux of CO2. The alkanolamine solvent systems considered here are the aqueous solutions of monoethanolamine (MEA), diethanolamine (DEA), N-methyldiethanolamine (MDEA), and 2-amino-2-methyl-1propanol (AMP), as well as aqueous blends of (MEA + MDEA), (MEA+AMP), (DEA + MDEA), and (DEA + AMP).

Liao and Li17 Xu et al.13 Xu et al.13 Littel et al.10 Liao and Li17 Ali30 Littel et al.10,15 Ali30

Ind. Eng. Chem. Res., Vol. 46, No. 8, 2007 2577

2578

Ind. Eng. Chem. Res., Vol. 46, No. 8, 2007

Table 2. Rate Expressions of Amines RB (mol m-3 s-1)

aqueous amine solution

RC (mol m-3 s-1)

Single Amines k2,BCACB MEA(B), DEA(B), and AMP(B)

1+

MDEA(B)

1 [(kH2O/k-1)CH2O] + [(kB/k-1)CB] k2,BCACB Blended Amines

MEA(B) + MDEA(C) and DEA(B) + MDEA(C)

MEA(B) + AMP(C) and DEA(B) + AMP(C)

k2,BCACB 1+

k2,BCACB

k2,CCACC

1 1+ [(kH2O/k-1)CH2O] + [(kB/k-1)CB] + [(kC/k-1)CC]

Therefore, the rate of CO2 reaction with tertiary amines is expressed by

RA ) k2,AmineCACAmine

(6)

The kinetic constants for all the single amine and blended amine systems are listed in Table 1, and the rate expressions for all the amines are given in Table 2. 2.2. Equations Describing the Diffusion-Reaction Process. In terms of the mass transfer and chemical reactions that occur in a HFMC, the analysis is performed here for the cases of pure CO2 and a CO2/N2 mixture. The CO2 inlet concentration, in the case of the CO2/N2 mixture, is assumed to be 20 vol %. The gas-phase concentration was assumed constant in the simulation. The total amine concentrations in all cases are assumed to be 10 wt %. For the liquid that is flowing through the hollow fibers, laminar flow with a parabolic velocity profile is considered and the pores of the membrane are assumed to be filled with gas flowing in the shell side, as shown schematically in Figure 1. Thus, for liquid flowing in the fiber lumen, the component mass-balance equations can be written as

Vz

k2,CCACC

1 [(kH2O/k-1)CH2O] + [(kB/k-1)CB] + [(kC/k-1)CC]

[ ( )]

∂Ci 1 ∂ ∂Ci r ) Di ∂z r ∂r ∂r

- Ri

(7)

where i represents components A, B, and C. A denotes CO2 and B and C denote alkanolamines for blended amine systems. For the single amine solution, only B denotes the amine. Equation 7 is deduced based on the following assumptions: (i)

1+

1 [(kH2O/k-1)CH2O] + [(kB/k-1)CB] + [(kC/k-1)CC]

steady-state and isothermal conditions; (ii) fully developed velocity profile and axis symmetry; (iii) the velocity component in the radial direction (Vr) is neglected; and (iv) negligible axial diffusion, which is reasonably assumed in the membrane contactor, because the concentration gradient in the axial direction is much smaller than that in the radial direction. In a laminar flow through a tube of radius R, a fully developed axial velocity profile can be described as

[ (Rr ) ] 2

Vz ) 2VL 1 -

(8)

where VL is the average velocity. The respective boundary conditions in the axial and radial directions are

at z ) 0; for all r; CA ) 0, CB ) CB0, CC ) CC0

(9)

and

at r ) 0; for z > 0;

( )

∂Ci )0 ∂r

(10)

For a nonvolatile liquid-phase component flowing through the fiber, the boundary condition at the gas/liquid interface is given by

at r ) R; for z > 0;

( ) ( )

∂CB ∂CC ) 0, )0 ∂r ∂r

(11)

At the gas/liquid interface, i.e., the membrane or fiber wall mass transfer of the gas-phase solute to the liquid phase occurs, which is described by

DA

( )| ∂CA ∂r

r)R

) kext(CAg - CAg,i)

(12)

Henry’s law is applied to relate the CO2 interfacial concentrations in the gas and liquid phases:

CA,i ) mCAg,i

Figure 1. Schematic diagram of the non-wetted mode of absorption in hollow-fiber membrane contactors (HFMCs).

(13)

The overall gas-phase mass-transfer coefficient (kext) is the series of two resistances that are comprised of the resistances to mass transport of species due to the gas phase and the microporous membrane (because the gas is assumed to be filling

Ind. Eng. Chem. Res., Vol. 46, No. 8, 2007 2579 Table 3. Physicochemical Parameters of Amines at 298 K Used in the Simulation aqueous amine solution

DA (m2/s)

m (mol/mol)

10-9

DB (m2/s)

reference(s) and comments

10-10

MEA DEA AMP

0.80 0.79 0.80

1.51 × 1.47 × 10-9 1.18 × 10-9

9.32 × 6.32 × 10-10 5.67 × 10-10

MDEA

0.82

1.44 × 10-9

6.21 × 10-10

the membrane pores):

all data taken from Versteeg and van Swaiij28 all data taken from Versteeg and van Swaiij28 the value for m was assumed; the value for DA was taken from Saha et al.;31 the value for DB was calculated the data for m and DA were taken from Versteeg and van Swaiij;28 the value for DB was taken from Hagewiesche et al.18

4. Method of Solution

1 1 1 ) + kext kg km

(14)

However, in the present case, the resistance to transport is assumed to be offered by the microporous membrane only (radius ) 2 × 10-4 m, length of fiber ) 0.2 m) and the gasphase resistance is assumed to be negligible, which reduces eq 14 to kext ) km. 3. Physicochemical Properties

The set of partial differential equations, along with the boundary conditions, as well as the reaction rates of CO2 with amine solutions were transformed to dimensionless form and solved in MATLAB (The MathWorks, Natick, MA), using a solver called pdepe. After the concentration profiles of CO2 and different alkanolamines were obtained, the local absorption flux JA local of CO2 along the length of the fiber was subsequently calculated using Fick’s law. The average absorption flux (JA) was obtained from the integration of the local fluxes along the length of the fiber:

The diffusion coefficients of CO2 in the blended amine solutions are calculated using the N2O analogy:26,27

( ) DN2O DCO2

)

amine soln

( ) DN2O DCO2

(15)

water

Diffusion coefficients of CO2 and N2O in water are calculated using the following equations of Versteeg and van Swaaij:28

( 2119.0 T ) 2371.0 exp(T )

DCO2 ) 2.35 × 10-6 exp -

(16)

DN2O ) 5.07 × 10-6

(17)

The diffusion coefficients of N2O in the amine solutions are calculated according to the modified Stokes-Einstein relation:

(DN2Oµ )amine soln ) (DN2Oµ )water 0.6

0.6

(18)

The diffusion coefficients of the mixed amines in water were calculated from the following correlation:16

Damine, water ) 2.5 × 10-10

JA )

1 L

∫0L JA local(z) dz

(21)

4.1. Dimensionless Forms of the Component Mass-Balance Equations for the Single and Blended Amine Systems. The set of partial differential equations, along with the boundary conditions, as well as the reaction rates of CO2 with amine solutions were transformed to dimensionless form. For the single amine systems of (MEA(B) + H2O), (DEA(B) + H2O), and (AMP(B) + H2O), the dimensionless forms of eq 7, combined with the reaction rates, are

[ ( )] [ ( )]

(1 - x2)

∂U1 ∂U1 1 ∂ )R x ∂y x ∂x ∂x

(1 - x2)

∂U2 ∂U2 DB 1 ∂ R x ) ∂y DA x ∂x ∂x

-

U1U2

β + [1/(γη + λU2)] -

(22)

U1U2 θ + [1/(ση + ψU2)] (23)

and, for the single amine of (MDEA(B) + H2O), the

-0.54

(MF )

(19)

The diffusion coefficients were correlated for viscosity and temperature using the modified Stokes-Einstein relation:2

(

)

µ H2 O T Damine, soln ) Damine, water 298 K µamine soln

0.6

(20)

The solubilities of CO2 in MEA, DEA, and MDEA were taken from the literature.28 The solubility of CO2 in an aqueous AMP solution and mixed amine solutions is assumed to be a constant value of 0.8 mol/mol, understanding the fact that there is a very negligible difference among the solubilities of CO2 in different alkanolamine solutions, as reported in the literature.28 The diffusivity and the solubility data for CO2 into alkanolamines solution used in the simulation are listed in Table 3.

Figure 2. Comparison of the average CO2 absorption flux in DEA and MDEA over the fiber length.

2580


Figure 3. Radial concentration profile of protonated and consumed MDEA for the absorption of pure CO2 in aqueous MDEA solution at the liquid exit of the fiber. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

dimensionless forms are

(1 - x2)

[ ( )]

∂U1 ∂U1 1 ∂ )R x ∂y x ∂x ∂x

(1 - x2)

-

U1U2 β + [1/(γη + λU2)]

[ ( )]

∂U2 DB 1 ∂ ∂U2 R x ) ∂y DA x ∂x ∂x

(24) - ωU1U2

(25)

where

CH2O CA CB r z x ) , y ) , U1 ) , U2 ) , η) , R L mCAg CB0 CB0

( )

LCB02 kH2Ok2,B 2VL , γ) , R) 2 , β) LCB0k2,B 2VL k-1 2R VL D AL

Figure 4. Average CO2 absorption flux over the fiber length for different external mass-transfer coefficients. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, and VL ) 0.1 m/s.

( ) ( )

LCB02 kBk2,B 2VL , θ) , λ) 2VL k-1 L(mCAg)k2,B L(mCAg)CB0 kBk2,B L(mCAg)CB0 kH2Ok2,B , ψ) , σ) 2VL k-1 2VL k-1 L(mCAg)k2,B ω) 2VL

( )

For all single amine systems, the boundary conditions in the axial and radial directions, respectively, are

at y ) 0; for all x; U1 ) 0, U2 ) 1

( ) ( ) ( ) ( )

at x ) 0; for y > 0; at x ) 1; for y > 0;

Figure 5. Local CO2 absorption flux for 20% CO2 and pure CO2 in a single amine solution over the fiber length. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

∂U1 ∂U2 ) 0, )0 ∂x ∂x

∂U1 ∂U2 ) ζ(1 - U1,i), )0 ∂x ∂x

where U1,i is the dimensionless CO2 liquid-side interfacial concentration and ζ ) kextR/(DAm). Here, R, β, γ, θ, λ, σ, ψ, η, ω, and ζ are dimensionless constants. For the blended amine systems of (MEA(B) + MDEA(C)) and (DEA(B) + MDEA(C)), the dimensionless forms of


Figure 6. Radial concentration profile of amine for the absorption of CO2 in the single amine solution at the liquid exit of the fiber: (A) concentration profile of amine for 20% CO2; (B) concentration profile of amine for pure CO2. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

eq 7 are

U3 )

∂U1 ) ∂y ∂U1 1 ∂ R x x ∂x ∂x

(1 - x2)

[ ( )] ( -

∂U2 ) ∂y ∂U2 DB 1 ∂ R x DA x ∂x ∂x

(1 - x2)

U1U2

β + [1/( + τU2 + EU3)]

[ ( )]

where

)

(26)

-

U1U2 θ + [1/(G + HU2 + IU3)]

[ ( )]

∂U3 DC 1 ∂ ∂U3 R x ) ∂y DA x ∂x ∂x

(1 - x2)

+ FU1U3

- φU1U3

(27) (28)

( ) ( ) ( )

CC L kH2Ok2,B , ) (CH2OCB0), CC0 2VL k-1

LCB02 kBk2,B L kCk2,B , E) (CB0CC0), 2VL k-1 2VL k-1 LCC0k2,C 2VL F) , θ) , 2VL L(mCAg)k2,B L kH2Ok2,B (mCAg)CH2O, G) 2VL k-1 L(mCAg)CC0 kCk2,B L(mCAg)CB0 kBk2,B , I) , H) 2VL k-1 2VL k-1 L(mCAg)k2,C φ) 2VL τ)

( ) ( )

( )

For the blended amine systems of (MEA(B) + AMP(C)) and (DEA(B) + AMP(C)), the dimensionless forms of eq 7 are

2582


Figure 7. Radial concentration profile of CO2 in the single amine solution at the liquid exit of the fiber: (A) concentration profile of 20% CO2; (B) concentration profile of pure CO2. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

[ ( )]

∂U1 ∂U1 1 ∂ x )R ∂y x ∂x ∂x U1U2

(1 - x2)

{

β + [1/( + τU2 + EU3)]

+

U1U3

[ ( )]

U1U2

θ + [1/(G + HU2 + IU3)]

[ ( )]

∂U3 DC 1 ∂ ∂U3 ) R x (1 - x2) ∂y DA x ∂x ∂x

}

K + [1/(ξ + σU3 + PU2)] (29)

∂U2 DB 1 ∂ ∂U2 ) R x ∂y DA x ∂x ∂x

(1 - x2)

where

-

(30)

K)

( ) ( )

2VL L kH2Ok2,C , ξ) (CH2OCC0), Lk2,CCC0 2VL k-1

Q)

2VL

U1U3 Q + [1/(S + δU3 + WU2)]

(31)

, S)

Lk2,C(mCAg)

δ)

-

( )

LCC02 kCk2,C LCC0CB0 kBk2,C σ) , P) , 2VL k-1 2VL k-1

( )

L kH2Ok2,C (mCAg)CH2O, 2VL k-1

( )

( )

L(mCAg)CB0 kBk2,C LCC0(mCAg) kCk2,C , W) 2VL k-1 2VL k-1

For all blended amine systems, the boundary conditions in the axial and radial directions, respectively, are


at y ) 0; for all x; U1 ) 0, U2 ) 1, U3 ) 1 at x ) 0; for y > 0;

( ) ( ) ( )

∂U1 ∂U2 ∂U3 ) 0, ) 0, )0 ∂x ∂x ∂x

at x ) 1; for y > 0; ∂U1 ∂U2 ∂U3 ) ζ(1 - U1,i), ) 0, )0 ∂x ∂x ∂x

( )

( ) ( )

where U1,i is the dimensionless CO2 liquid-side interfacial concentration and ζ ) kextR/(DAm). Here, δ, , ξ, σ, τ, and φ are dimensionless constants. 5. Results and Discussion To validate the model for the absorption of CO2 in different aqueous alkanolamine solutions, literature data were compared with the calculated results. As shown in Figure 2, the present simulation result agrees well with the literature data of Wang et al.4 The parameters used in the calculation are the same as those used by Wang et al.4 Radial concentration profiles of protonated MDEA (R1R2R3NH+ in eq 5) and consumed MDEA at the liquid exit of the fiber are simulated and shown in Figure 3. The overlapping profiles confirm that MDEA has only been transformed to protonated MDEA and the component balance is matching. We have also analyzed the mass balance in the inlet and the exit points of the fiber for the absorption of CO2 in aqueous solution of MDEA and the deviation between the total inlet and outlet mass flow rates was determined to be insignificant. Figure 4 shows the variation of the average CO2 absorption flux over the fiber length with increasing overall gas-phase masstransfer coefficient (kext ) km, in the present case), for the absorption of 20% and pure CO2 in aqueous solutions of highly reactive MEA and less-reactive MDEA. When kext is small, the mass-transfer resistance has a significant influence on flux. Kreulen et al.29 estimated km to be in the range of 0.012-0.077 m/s for the absorption of different gas mixtures in a liquid. According to Figure 4, the CO2 absorption flux already reached the steady value within this range for both the cases of highly reactive MEA and less-reactive MDEA solutions. However, we have assumed a kext value of 100 m/s during the simulation studies, which is much higher than the values reported by Kreulen et al.,29 to address the fact that the efficiency of the membrane for different processes does not influence the fluxes obtained in the latter part of the study. 5.1. CO2 Absorption with a Single Amine Solvent. The performance of different single aqueous alkanolamine solvents for the absorption of 20% and pure CO2 is analyzed in terms of local flux of CO2 along the length of the HFMC and the liquid-phase CO2 and amine concentration profiles. Figure 5 shows the variation of the local flux of 20% and pure CO2 along the length of the fiber. As expected, the absorption flux is greater for pure CO2 than for 20% CO2 for all amines. The flux for the absorption of 20% CO2 is ∼50% of that for the absorption of pure CO2. In both cases, there is no significant variation in the flux for absorption in MDEA under simulation conditions, because of its lower reaction rate. However, the CO2 absorption fluxes for MEA, DEA, and AMP decrease along the length of the fiber in both cases. With increases in the fiber length, more amine is consumed, because of the continuous supply of CO2, leading to a decrease in amine concentration. This, in turn, leads to a decrease in CO2 flux, because the reaction rate is a function of amine concentration. Figure 5 also shows that the aqueous solution of MEA has the best CO2 absorption capacity, followed

Figure 8. Average CO2 absorption flux over the fiber length, as a function of amine blend composition: (A) average CO2 absorption flux in (MEA + MDEA) and (DEA + MDEA) blends; (B) average CO2 absorption flux in (MEA + AMP) and (DEA + AMP) blends. L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

by AMP, DEA, and MDEA in sequence. This trend can be justified from the reaction kinetics of amines with CO2, based on eqs 4 and 6, as well as Table 1. It is also understood that, when the CO2 concentration in the gas phase at the gas/liquid interface remains constant, which is assumed in the present study, the reaction rate is the dominating factor over other transport and physicochemical properties during the absorption of CO2 into alkanolamines. Figures 6A and 6B describe the radial concentration profiles of four alkanolamines in the liquid phase for the absorption of 20% and pure CO2, respectively. In both cases, there are significant decreases in the concentrations of MEA, AMP, and DEA. MEA was depleted more in the interface, whereas the depletions for DEA and AMP are almost the same. However, the change in MDEA concentration is much less, as compared to the others in both cases, because of its high equilibrium loading capacity and low reaction rate, in comparison to that of other amines. The concentration of CO2 is much less in MEA, AMP, and DEA, compared to that in the case of MDEA for both the absorption of 20% and pure CO2, as shown in Figures 7A and 7B, respectively. 5.2. CO2 Absorption into a Blended Amine. The performance of absorption of 20% and pure CO2 into aqueous blends

2584


Figure 9. Radial concentration profile of CO2 in the (MEA + MDEA) blended amine solution at the liquid exit of the fiber: (A) concentration profile of 20% CO2; (B) concentration profile of pure CO2. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

of (MEA + MDEA), (MEA + AMP), (DEA + MDEA), and (DEA + AMP) is analyzed in terms of the average flux of CO2 over the fiber length, and the liquid-phase radial concentration profile of CO2 and the alkanolamines. The total amine concentration was always maintained at 10 wt %. The concentration of each and every amine in the blend was varied from 0 wt % to 10 wt % to examine the effect of different alkanolamine compositions on the absorption of CO2. Figures 8A and 8B show the average CO2 absorption flux in different aqueous alkanolamine blends. The flux for the absorption of pure CO2 is much higher than that for the absorption of 20% CO2. This is due to the fact that pure CO2 gives a greater driving force for absorption. As shown in Figure 8A, a greater amount of MEA or DEA gives a higher flux, because of their higher reaction rate constant. However, for (MEA + MDEA) blends, the rate of increase is much higher than that for (DEA + MDEA) blends for both 20% and pure CO2. According to Figure 8B, the behavior of the (MEA + AMP) systems are similar to that of the (MEA + MDEA) systems. However, for

the (DEA + AMP) systems, the flux of pure CO2 as well as 20% CO2 is almost constant with different blend compositions. This is due to the comparable reaction rates for CO2 in DEA and AMP. The behavior of 20% CO2 and pure CO2 absorption in aqueous blends of (MEA + MDEA) has been described in Figures 9A and 9B, respectively. In both cases, the liquid-phase radial CO2 concentration profile at the liquid exit shifts to the right side as the concentration of MEA in the blend increases and the difference is gradually decreased with increasing MEA, because of the higher reaction rate of MEA. A similar result was obtained for the absorption of CO2 in (DEA + MDEA) blends, as shown in Figures 10A and 10B, except for the difference in the magnitude of absorption. The average absorption flux for pure CO2 increased from 0.0027 mol m-2 s-1 for the (1 wt % MEA + 9 wt % MDEA) blend to 0.0102 mol m-2 s-1 for the (9 wt % MEA + 1 wt % MDEA) blend and becomes equal to the flux in 10 wt % MEA, whereas the flux for pure CO2 increased from 0.0021 mol m-2 s-1 for the (1 wt % DEA


Figure 10. Radial concentration profile of CO2 in the (DEA + MDEA) blended amine solution at the liquid exit of the fiber: (A) concentration profile of 20% CO2; (B) concentration profile of pure CO2. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

+ 9 wt % MDEA) blend to 0.0042 mol m-2 s-1 for the (9 wt % DEA + 1 wt % MDEA) blend. Therefore, from a technical viewpoint, the (MEA + MDEA) blend is better than the (DEA + MDEA) blend for the absorption of CO2. Figures 11 and 12 show the variation of the MEA and DEA concentrations with radius for the absorption of 20% CO2 in (MEA + MDEA) and (DEA + MDEA) blends. Similar behavior has been observed for the absorption of pure CO2, except for the changes in magnitude, and, therefore, the figures are not reported here. According to these figures, at higher MEA or DEA concentrations, the depletion of amine is less, because more amine is available to react with CO2. The concentration profiles of MDEA for all the blends, which show insignificant changes in the depletion of MDEA, are not reported here. From an analysis of the liquid-phase radial concentration profile of 20% CO2 and pure CO2 in aqueous blends of (MEA + AMP), it has been observed that there is no significant difference in the performance of absorption of CO2 between

10 wt % MEA and all (MEA + AMP) blends. This observation is because there is not much difference in the reaction rate constants between CO2 with MEA and CO2 with AMP, unlike the case of MEA and MDEA discussed earlier. The average absorption flux for pure CO2 increased from 0.0058 mol m-2 s-1 to 0.0102 mol m-2 s-1 for the (1 wt % MEA + 9 wt % AMP) blend to the (9 wt % MEA + 1 wt % AMP) blend. The liquid-phase radial MEA and AMP concentration profiles for the absorption of 20% and pure CO2 at the liquid exit are not changed significantly. From the radial liquid-phase concentration profiles of CO2, DEA and AMP at the liquid exit for the absorption of 20% CO2 and pure CO2 in aqueous blends of (DEA + AMP), it is evident that there is practically no change in the absorption performance among the blends. The average absorption fluxes of both 20% CO2 and pure CO2 in all the blends are literally the same. Therefore, the (DEA + AMP) blends are less suitable for the absorption of CO2, compared to the other alkanolamine blends that have been discussed earlier.

2586


Figure 11. Radial concentration profile of MEA for 20% CO2 in the (MEA + MDEA) blended amine solution at the liquid exit of the fiber. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

Figure 12. Radial concentration profile of DEA for 20% CO2 in the (DEA + MDEA) blended amine solution at the liquid exit of the fiber. Conditions: L ) 0.2 m, R ) 2 × 10-4 m, VL ) 0.1 m/s, and kext ) 100 m/s.

Thus, from the aforementioned study, it has been observed that either the (MEA + MDEA) or (MEA + AMP) blends should be the recommended combination over others for the removal of CO2, because they show similar extents of CO2 absorption. However, further studies are necessary with these blends to determine the effects of different operating and membrane parameters for suitable design of a HFMC for the removal of CO2. However, the better of the two would be the blend that enables easier recovery of the alkanolamines. 6. Conclusion The absorption of 20% CO2 and pure CO2 was studied theoretically in aqueous solutions of four single alkanolamine solvents, as well as four blended alkanolamine solvents, using hollow-fiber membrane contactors (HFMCs). The total alkanolamine concentrations in all cases were maintained at 10 wt %. Among the single amine solutions, the aqueous solution of MEA

was determined to be the most suitable for the absorption of CO2, if only the average flux of CO2 in the amine is considered. For the absorption in the (MEA + MDEA), (DEA + MDEA), and (MEA + AMP) blends, the flux increases as the concentration of MEA or DEA increases in the blends. The absorption performance in the different blends of (DEA + AMP) was literally the same and, technically, this blend is not suitable for the absorption of CO2. The absorption flux of CO2 in the (MEA + MDEA) and (MEA + AMP) blends is greater, compared to that in other blends. The successful design of a HFMC demands further investigation on the effects of operating parameters such as flow rates of both gas and liquid streams, membrane parameters (such as pore size and fiber radius), and other effects (such as gas-phase resistance and number of fibers per unit shell) to optimize the performances of the HFMC. However, the ultimate choice of the most appropriate solvent for the absorption of CO2 should be dependent on the regenerability and cost of the different alkanolamines.


Acknowledgment This work was supported by the Department of Science and Technology (DST) (D. O. No.: SR/FT/ETA-07/2004), New Delhi, Government of India, India. Appendix: The Derivation of eq 4 If [Z] is the concentration of zwitterion (at quasy-steadystate), then

RA ) k2,AmineCACAmine - k-1[Z] ) [Z]

∑ kb[b]

(A1)

The term ∑ kb[b] indicates the contribution of the various bases present to the rate of removal of protons. Thus,

[Z] ) concentration of intermediate zwitterionic product (mol/ m3) Greek Letters µ ) viscosity (Pa s) F ) density (kg/m3) Subscripts g ) gas phase i ) interface i ) component i A ) CO2 B, C ) alkanolamine Literature Cited

RA

k2,Amine

)

CACAmine

1+

1 [(kH2O/k-1)CH2O] +

∑ [(kAmine/k-1)CAmine]

(A2)

For the absorption of CO2 in the blends of primary and secondary amines,

RA )

k2,AmineCACAmine

∑

1

1+ [(kH2O/k-1)CH2O] +

∑ [(kAmine/k-1)CAmine] (A3)

Nomenclature C ) concentration of components in liquid (mol/m3) D ) diffusion coefficient (m2/s) E-I ) dimensionless constants J ) average absorption flux along the fiber length (mol m-2 s-1) K, P, S, W ) dimensionless constants Jlocal ) local absorption flux (mol m-2 s-1) k-1 ) reverse first-order reaction rate constant (m3 mol-1 s-1) k2 ) second-order forward reaction rate constant (m3 mol-1 s-1) kb ) second-order reaction rate constant for base b (m3 mol-1 s-1) kext ) overall gas-phase mass-transfer coefficient (m/s) kg ) gas-phase mass-transfer coefficient (m/s) km ) membrane phase mass-transfer coefficient (m/s) L ) length of hollow fiber (m) m ) distribution coefficient (mol/mol) M ) molar mass (kg/kmol) r ) distance in the radial direction (m) R ) inner radius of hollow fiber (m) Ri ) rate of reaction (mol m-3 s-1) T ) temperature (K) U1 ) dimensionless concentration of CO2 in liquid U2,U3 ) dimensionless concentration of amine VL ) average velocity of liquid (m) Vr ) velocity of liquid in radial direction (m/s) Vz ) velocity of liquid in the z-direction (m/s) x ) dimensionless radial distance y ) dimensionless axial distance z ) distance in the axial direction from liquid inlet (m) [R1R2R3N] ) concentration of MDEA (mol/m3) [R1R2R3N]0 ) initial concentration of MDEA (mol/m3) [R1R2R3NH+] ) concentration of protonated MDEA (mol/m3)

(1) Sartori, G.; Savage, D. W. Sterically Hindered Amines for CO2 Removal from Gasses. Ind. Eng. Chem. Fundam. 1983, 22, 239. (2) Mandal, B. P.; Biswas, A. K.; Bandyopadhayay, S. S. Absorption of Carbon Dioxide into Aqueous Blends of 2-Amino-2-Methyl-1-Propanol and Diethanolamine. Chem. Eng. Sci. 2003, 58, 4137. (3) Gabelman, A.; Hwang, S.-T. Hollow Fiber Membrane Contactors. J. Membr. Sci. 1999, 159, 61. (4) Wang, R.; Li, D. F.; Liang, D. T. Modelling of CO2 Capture by Three Typical Amine Solutions in Hollow Fibre Membrane Contactors. Chem. Eng. Process. 2004, 43, 849. (5) Sirkar, K. K. Other New Membrane Processes. In Membrane Handbook; Winstone Ho, W. S., Sirkar, K. K., Eds.; Van Nostrand Reinhold: New York, 1992. (6) Mandal, B. P.; Guha, M.; Biswas, A. K.; Bandyopadhyay, S. S. Removal of Carbon Dioxide by Absorption in Mixed Amines: Modelling of Absorption in Aqueous MDEA/MEA and AMP/MEA Solutions. Chem. Eng. Sci. 2001, 56, 6217. (7) Mandal, B. P.; Bandyopadhyay, S. S. Absorption of Carbon Dioxide into Aqueous Blends of 2-Amino-2-Methyl-1-Propanol and Monoethanolamine. Chem. Eng. Sci. 2006, 61, 5440. (8) Rinker, E. B.; Ashour, S. S.; Sandall, O. C. Kinetics and modelling of carbon dioxide absorption into aqueous solutions of N-methyldiethanolamine. Chem. Eng. Sci. 1995, 50, 755. (9) Rinker, E. B.; Ashour, S. S.; Sandall, O. C. Absorption of carbon dioxide into aqueous blends of diethanolamine and methyldiethanolamine. Ind. Eng. Chem. Res. 2000, 39, 4346. (10) Littel, R. J.; Versteeg, G. F.; Van Swaaij, W. P. M. Kinetics of CO2 with primary and secondary amines in aqueous solutionssI. Zwitterion deprotonation kinetics for DEA and DIPA in aqueous blends of alkanolamines. Chem. Eng. Sci. 1992, 47, 2027. (11) Xiao, J.; Li, C. W.; Li, M. H. Kinetics of Absorption of Carbon Dioxide into Aqueous Solutions of 2-Amino-2-Methyl-1-Propanol + Monoethanolamine. Chem. Eng. Sci. 2000, 55, 161. (12) Blauwhoff, P. M. M.; Versteeg, G. F.; van Swaaij, W. P. M. A Study on the Reaction between CO2 and Alkanolamines in Aqueous Solutions. Chem. Eng. Sci. 1984, 39, 207. (13) Xu, S.; Wang, Y. W.; Otto, F. D.; Mather, A. E. Kinetics of the Reaction of CO2 with 2-Amino-2-Methyl-1-Propanol Solutions. Chem. Eng. Sci. 1996, 51, 841. (14) Versteeg, G. F.; van Swwaaij, W. P. M. On the Kinetics between CO2 and Alkanolamines Both in Aqueous and Non-Aqueous Solutionss II. Tertiary Amines. Chem. Eng. Sci. 1988, 43, 587. (15) Littel, R. J.; Van Swaaij, W. P. M.; Versteeg, G. F. Kinetics of Carbon Dioxide with Tertiary Amines in Aqueous Solution. AIChE J. 1990, 36, 1633. (16) Glasscock, D. A. Modeling and Experimental Study of Carbon Dioxide Absorption in Aqueous Alkanolamines, Ph.D. dissertation, The University of Texas at Austin, Austin, TX, 1990. (17) Liao, C. H.; Li, M. H. Kinetics of Absorption of Carbon Dioxide into Aqueous Solutions of Monoethanolamine + N-Methyldiethanolamine. Chem. Eng. Sci. 2002, 57, 4569. (18) Hagewiesche, D. P.; Ashour, S. S.; Al-Ghawas, H. A.; Sandall, O. C. Absorption of Carbon Dioxide into Aqueous Blends of Monoethanolamine and N-Methyldiethanolamine. Chem. Eng. Sci. 1995, 50, 1071. (19) Glasscock, D. A.; Critchfield, J. E.; Rochelle, G. T. CO2 Absorption/ Desorption in Mixtures of Methyldiethanolamine with Monoethanolamine or Diethanolamine. Chem. Eng. Sci. 1991, 46, 2829. (20) Idem, R.; Wilson, M.; Tontiwachwuthikul, P.; Chakma, A.; Veawab, A.; Aroonwilas, A.; Gelowitz, D. Pilot Plant Studies of the CO2 Capture Performance of Aqueous MEA and Mixed MEA/MDEA Solvents at the

2588


University of Regina CO2 Capture Technology Development Plant and the Boundary Dam CO2 Capture Demonstration Plant. Ind. Eng. Chem. Res. 2006, 45, 2414. (21) Zhang, H.-Y.; Wang, R.; Liang, D. T.; Tay, J. H. Modeling and Experimental Study of CO2 Absorption in a Hollow Fiber Membrane Contactor. J. Membr. Sci. 2006, 279, 301. (22) Gong, Y.; Wang, Z.; Wang, S. Experiments and Simulation of CO2 Removal by Mixed Amines in a Hollow Fiber Membrane Module. Chem. Eng. Process. 2006, 45, 652. (23) Caplow, M. Kinetics of Carbamate Formation and Breakdown. J. Am. Chem. Soc. 1968, 90, 6795. (24) Danckwerts, P. V. The Reaction of CO2 with Ethanolamines. Chem. Eng. Sci. 1979, 34, 443. (25) Versteeg, G. F.; Kuipers, J. A. M.; Van Beckum, F. P. H.; Van Swaaij, W. P. M. Mass Transfer with Complex Reversible Chemical ReactionssII. Parallel Reversible Chemical Reactions. Chem. Eng. Sci. 1990, 45, 183. (26) Mandal, B. P.; Kundu, M.; Padhiyer, N. U.; Bandyopadhyay, S. S. Physical Solubility and Diffusivity of N2O and CO2 into Aqueous Solutions of (2-Amino-2-Methyl-1-Propanol + Diethanolamine) and (N-Methyldiethanolamine + Diethanolamine). J. Chem. Eng. Data 2004, 49, 264. (27) Mandal, B. P.; Kundu, M.; Bandyopadhyay, S. S. Physical Solubility and Diffusivity of N2O and CO2 into Aqueous Solutions of (2-Amino-2-

Methyl-1-Propanol + Monoethanolamine) and (N-Methyldiethanolamine + Monoethanolamine). J. Chem. Eng. Data 2005, 50, 352. (28) Versteeg, G. F.; Van Swwaaij, W. P. M. Solubility and Diffusivity of Acid Gases (Carbon Dioxide, Nitrous Oxide) in Aqueous Alkanolamine Solutions. J. Chem. Eng. Data 1988, 33, 29. (29) Kreulen, H.; Smolders, C. A.; Versteeg, G. F.; Van Swwaaij, W. P. M. Determination of Mass Transfer Rates in Wetted and Non-wetted Microporous Membranes. Chem. Eng. Sci. 1993, 48, 2093. (30) Ali, S. H. Kinetics of the Reaction of Carbon Dioxide with Blends of Amines in Aqueous Media Using the Stopped-Flow Technique. J. Chem. Kinetics 2005, 37, 391. (31) Saha, A. K.; Bandyopadhyay, S. S.; Biswas, A. K. Solubility and Diffusivity of Nitrous Oxide and Carbon Dioxide in Aqueous Solutions of 2-Amino-2-Methyl-1-Propanol. J. Chem. Eng. Data 1993, 38, 78.

ReceiVed for reView November 19, 2006 ReVised manuscript receiVed February 11, 2007 Accepted February 14, 2007 IE061476F

Removal of CO2 by Single and Blended Aqueous Alkanolamine

Recommend Documents