Chemical Engineering Science 64 (2009) 1185 -- 1194 Contents lists available at ScienceDirect Chemical Engineering Science journal homepage: w w w . e l s e v i e r . c o m / l o c a t e / c e s Absorption of carbon dioxide into aqueous solutions of piperazine activated 2-amino-2-methyl-1-propanol Arunkumar Samanta, S.S. Bandyopadhyay ∗ Separation Science Laboratory, Cryogenic Engineering Centre, IIT Kharagpur, W.B. 721 302, India A R T I C L E I N F O Article history: Received 13 May 2008 Received in revised form 24 October 2008 Accepted 25 October 2008 Available online 7 November 2008 Keywords: Absorption Kinetics Carbon dioxide AMP Piperazine A B S T R A C T In this work, new experimental data on the rate of absorption of CO2 into piperazine (PZ) activated aqueous solutions of 2-amino-2-methyl-1-propanol (AMP) are reported. The absorption experiments using a wetted wall contactor have been carried out over the temperature range of 298–313 K and CO2 partial pressure range of 2–14 kPa. PZ is used as a rate activator with a concentration ranging from 2 to 8 wt%, keeping the total amine concentration in the solution at 30 wt%. The CO2 absorption into the aqueous amine solutions is described by a combined mass transfer-reaction kinetics-equilibrium model, developed according to Higbie's penetration theory. Parametric sensitivity analysis is done to determine the effects of possible errors in the model parameters on the accuracy of the calculated CO2 absorption rates from the model. The model predictions have been found to be in good agreement with the experimental results of rates of absorption of CO2 into aqueous (PZ+AMP). The good agreement between the model predicted rates and enhancement factors and the experimental results indicates that the combined mass transfer-reaction kinetics-equilibrium model with the appropriate use of model parameters can effectively represent CO2 mass transfer in PZ activated aqueous AMP solutions. © 2008 Elsevier Ltd. All rights reserved. 1. Introduction The removal of acid gas impurities, such as carbon dioxide (CO2 ) and hydrogen sulfide (H2 S), from natural gas, refinery off-gases, synthesis gas and other industrial gases is an important operation in industrial gas processing. Due to the need to exploit even poorer quality natural gas, extensive research activities to develop less energy intensive efficient gas processing technologies are essential. Improved solvents for gas sweetening with reduced cost will ensure the utilization of tertiary natural gas resources. Besides, growing environmental concerns today for global warming and climate change have motivated extensive research activities towards developing more efficient and improved processes for CO2 capture from stationary CO2 emission sources for economic CO2 sequestration. Thus, besides CO2 removal from natural gas, improved processes for CO2 capture from the flue gas streams of fossil fuel based power plants are also essential today. Aqueous alkanolamine solutions are widely used for the removal of acid gas impurities from natural gas and industrial gas streams. Industrially important alkanolamines for the regenerative chemical ∗ Corresponding author. Tel.: +91 3222 283580; fax: +91 3222 282258. E-mail address: ssbandyo@hijli.iitkgp.ernet.in (S.S. Bandyopadhyay). 0009-2509/$ - see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.ces.2008.10.049 absorption processes are monoethanolamine (MEA), diethanolamine (DEA), N-methyldiethanolamine (MDEA), and 2-amino-2-methyl-1propanol (AMP) (Kohl and Nielsen, 1997). Recently, mixing of alkanolamines, e.g., mixture of a primary (e.g., MEA) or secondary (e.g., DEA) alkanolamine with a tertiary alkanolamine (e.g., MDEA) is suggested to capitalize on the advantages of each amine. The blended amine solvents combine the higher equilibrium capacity of the tertiary amine for CO2 with the higher CO2 reaction rate of the primary or secondary amine (Chakravarty et al., 1985). The use of blended amines in gas treatment brings about a significant improvement in absorption capacity and absorption rate and a great saving in solvent regeneration energy requirement. Besides, it also offers the advantage of setting the selectivity of the solvent toward CO2 by judiciously mixing the amines in varying proportions which results in an additional degree of freedom for achieving the desired separation to meet the required specification of the treated gas. As with MDEA, the sterically hindered amine, AMP, also provides an equilibrium CO2 capacity about twice that of any primary or secondary amine. Besides, similar to the CO2 -MDEA reaction product, the ultimate product of the CO2 -AMP reaction is bicarbonate. Hence, the regeneration energy requirement, when AMP is used as a component of a blended amine solvent, will be lower as in the case of using MDEA. On the other hand, AMP offers an additional advantage over MDEA particularly for CO2 removal, due to the fact that the CO2 -AMP reaction rate is much faster than the CO2 -MDEA reaction rate. 1186 A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 More recently, there is an interest in using activated alkanolamine solvents by employing a reaction rate accelerator e.g., piperazine (PZ), in the aqueous alkanolamine solution. PZ activated aqueous MDEA or AMP for CO2 removal take advantage of the high rate of reaction of CO2 with the activating agent, e.g., PZ, combined with the advantage of high CO2 loading capacity of MDEA or AMP and relatively lower regeneration energy requirement. PZ is used as an activator in the activated MDEA process of BASF (Appl et al., 1982) and it is reported that PZ is more effective than the conventional activators (Bishnoi and Rochelle, 2002). While PZ activated aqueous AMP solutions, like PZ activated aqueous MDEA, can be highly efficient solvents for CO2 removal, published literature on this solvent system is very limited. Seo and Hong (2000) experimentally studied the effect of PZ on the absorption of CO2 in aqueous solutions of AMP at 303 and 313 K using a wetted-sphere contactor. They used AMP concentrations in the range of 0.55–3.35 kmol m−3 along with PZ concentrations of 0.058, 0.115, and 0.233 kmol m−3 . However, they performed their experiments at relatively higher CO2 partial pressure resulting in substantial depletion of PZ at the gas–liquid interface. As a result, the kinetic data reported by them from their absorption experiments appear to be incorrect (Bishnoi and Rochelle, 2000). The kinetics of the absorption of CO2 into aqueous solutions of (AMP+PZ) were also investigated by Sun et al. (2005) using a wetted wall contactor. They performed the absorption experiments over the temperature range of 303–313 K and CO2 partial pressure range of 2.63–4.55 kPa using solutions containing 1.0–1.5 kmol m−3 AMP and 0.1–0.4 kmol m−3 PZ. A hybrid kinetic model, a second order reaction for CO2 with PZ and a zwitterion reaction mechanism for CO2 with AMP have been used by Sun et al. (2005) to interpret their results. More recently, Lin et al. (2008) studied the performance of microporous polyvinylidinefluoride hollow fiber membrane contactor using the aqueous solutions of PZ and AMP as the solvents. Experimental results showed that the CO2 absorption rate was much enhanced by the addition of PZ promoter. However, there is no study reported in the open literature to interpret the experimental absorption data of CO2 in PZ activated aqueous AMP solutions using a numerically solved coupled mass transfer-kinetics-equilibrium model that takes into account all possible reversible reactions of CO2 -(AMP+PZ+H2 O). In this work, the absorption of CO2 into PZ activated aqueous AMP solutions has been studied experimentally and theoretically at various temperatures, different relative compositions of AMP/PZ in the solutions, and various CO2 partial pressures to gain better understanding of the absorption of CO2 into PZ activated aqueous alkanolamine solutions. Following the work of Hagewiesche et al. (1995) and Mandal et al. (2001), the diffusion-reaction processes for the chemical absorption of CO2 are modeled according to Higbie's (1935) penetration theory with the assumption that all reactions are reversible. Parametric sensitivity analysis has also been presented using the mathematical model developed in this work to determine the effects of possible errors in the important model parameters on the accuracy of the calculated absorption rates and enhancement factors from the model. 2. Model development 2.1. Reaction scheme and reaction mechanism When CO2 is absorbed into an aqueous mixed amine solution of AMP(R R R N, where R = C2 H2 (CH3 )2 OH and R = H) and PZ, the following reactions may take place in the liquid phase: Base-catalyzed hydration reaction: K1 ,k21 CO2 + R R R N + H2 O ←→ R R R NH+ + HCO− 3 (1) Formation of monocarbamate: K2 ,k22 CO2 + PZ + H2 O ←→ PZCOO− + H3 O+ (2) Formation of monocarbamate by PZ/AMP: K3 ,k23 CO2 + R R R N + PZ ←→ PZCOO− + R R R NH+ (3) Formation of dicarbamate: K4 ,k24 CO2 + PZCOO− + H2 O ←→ PZ(COO− )2 + H3 O+ Formation of dicarbamate by (4) PZCOO− /AMP: K5 ,k25 CO2 + R R R N + PZCOO− ←→ PZ(COO− )2 + R R R NH+ (5) Formation of bicarbonate: K6 ,k26 CO2 + OH− ←→ HCO− 3 (6) Formation of carbonate: K7 2− + HCO− 3 + H2 O ←→ CO3 + H3 O (7) Protonation of PZ: K8 PZ + H3 O+ ←→ PZH+ + H2 O (8) Protonation of monocarbamate: K9 PZCOO− + H3 O+ ←→ PZH+ COO− + H2 O (9) Protonation of AMP: K10 R R R N + H3 O+ ←→ R R R NH+ + H2 O (10) Dissociation of water: K11 2H2 O ←→ H3 O+ + OH− (11) Reactions (1)–(6) have finite reaction rates and are reversible. Reactions (7)–(11) are reversible and instantaneous with respect to mass transfer and at equilibrium, since they involve only proton transfer. In view of the very low carbamate stability constant of the sterically hindered amine AMP (Sartori and Savage, 1983), the only reaction of importance between CO2 and AMP is suggested to be the formation of bicarbonate ion. Hence, bicarbonate ions may be present in the solution in much larger amounts than the carbamate ions (Yih and Shen, 1988; Hagewiesche et al., 1995; Mandal et al., 2001). Thus, the CO2 -AMP reaction, suggested to be similar to that of CO2 -MDEA, may be represented by reaction (1), neglecting the formation of carbamate by AMP. The proposed mechanism for the reaction between CO2 and PZ involves the formation of a zwitterion followed by the deprotonation of the zwitterion by a base to produce PZ-carbamate and protonated base (Bishnoi and Rochelle, 2000; Derks et al., 2006; Samanta and Bandyopadhyay, 2007). Any base present in the solution may contribute to the deprotonation of the zwitterion (Caplow, 1968; Danckwerts, 1979). The contribution of each base would depend on its concentration as well as how strong a base it is (Hagewiesche et al., 1995; Bishnoi, 2000). Hence, the main contribution to the deprotonation of the zwitterion in an aqueous solution of a mixture of AMP and PZ would come from PZ, AMP and to a lesser extent from H2 O and OH− . The presence of significantly more AMP appears to catalyze the reaction of CO2 and PZ to form carbamate and the AMP deprotonation will dominate. In the case of PZ, formation of zwitterion should be the rate-determining step, while the deprotonation step involves only a proton transfer and is considered to be very fast. Reaction (2) represents the reaction between CO2 A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 and PZ for producing carbamate (Bishnoi and Rochelle, 2000; Derks et al., 2006; Samanta and Bandyopadhyay, 2007). The rate constant, k22 , is considered the global rate coefficient for the formation of zwitterion and zwitterion deprotonation (reaction (2)). Hence, this representation does not rule out the possible formation of a zwitterion reaction intermediate. Also, the rate constant, k24 , is viewed as the global rate coefficient for the formation of PZ-dicarbamate, PZ(COO− )2 , by reaction (4). The third order rate constants, k23 and k25 , for AMP catalyzed PZ-carbamate (reaction (3)) and PZdicarbamate (reaction (5)) formations have been considered to include the contribution of AMP in zwitterion deprotonation. 2.2. Bulk liquid equilibrium model Liquid bulk concentrations of all chemical species must be known in order to solve the diffusion-reaction model developed in this work. The liquid bulk concentrations of all chemical species can be estimated from the initial concentration of AMP and PZ solution; the initial CO2 loading, 1 , of the solution and the assumption that all reactions are at equilibrium. The concentration of water is assumed to remain constant because its concentration is much larger than the concentration of other chemical species. Here, for convenience the chemical species have been renamed as follows: u3 = [R R R NH+ ] u2 = [R R R N], u1 = [CO2 ], u4 = [HCO− 3 ], u5 = [OH− ], u7 = [H3 O+ ], u8 = [PZ], u10 = [PZCOO− ], u9 = [PZH+ ] u11 = [PZH+ COO− ] − u12 = [PZ(COO )2 ] We have the following 12 equations for 12 liquid bulk concentrations: AMP balance: (12) PZ balance: (13) CO2 balance: u01 + u04 + u06 + u010 + u011 + 2u012 = 1 {[AMP]initial + [PZ]initial } (14) Electroneutrality balance: + u07 + u09 − u04 − u05 − 2u06 − u010 − 2u012 =0 K4 = K6 = K7 = K8 = u07 u010 u01 u08 u07 u012 u01 u010 u04 u01 u05 u06 u07 u04 u09 u07 u08 K10 = u03 (22) u02 u07 K11 = u05 u07 (23) The 12 simultaneous nonlinear algebraic equations (12)–(23) have been solved using a subroutine called DNEQNF documented in the IMSL Math/Library (Visual Numerics, Inc., 1994) in FORTRAN 90 for the 12 unknowns (u01 , . . . , u012 ) of the liquid bulk concentrations. The routine solves a system of nonlinear equations using a modified Powell hybrid algorithm and a finite-difference approximation to the Jacobian. This algorithm is a variation of Newton's method, which uses a finite-difference approximation to the Jacobian and takes precautions to avoid large step sizes or increasing residuals (Visual Numerics, Inc., 1994; More et al., 1980). As finite-difference approximation is used to estimate the Jacobian, double precision calculation has been adopted in order to obtain an accurate Jacobian. It has been found that the algorithm converged to solution even when the initial guesses are not close to the solution (Aboudheir et al., 2003; Samanta and Bandyopadhyay, 2007). The mathematical model developed to interpret the experimental results of absorption of CO2 into aqueous solutions of AMP and PZ is based on the concept of gas absorption accompanied by multiple reversible chemical reactions in a thin liquid film. It takes into account the coupling between mass transfer, chemical equilibria and chemical kinetics of all liquid phase chemical reactions. The mathematical model describing the diffusion-reaction processes consists of partial differential equations formulated according to Eq. (24) for the reactants and products present in the liquid phase: j ui j2 u = Di 2i + i Ri jt jx (24) All reactions are numerically treated as reversible reactions with finite rate, in which the reverse reaction rate constants are calculated using the equilibrium constant of the corresponding reaction. The rate equations of the finite rate reactions (Eqs. (1)–(6)) are then given by Eqs. (25)–(30). R1 = −k21 u1 u2 + k21 u3 u4 K1 (25) R2 = −k22 u1 u8 + k22 u7 u10 K2 (26) (15) All reactions are in equilibria: K2 = (21) u07 u010 i u08 + u09 + u010 + u011 + u012 = [PZ]initial u03 u011 2.3. Diffusion-reaction model u6 = [CO2− 3 ] u02 + u03 = [AMP]initial K9 = 1187 k23 u3 u10 K3 (16) R3 = −k23 u1 u2 u8 + (17) R4 = −k24 u1 u10 + (18) R5 = −k25 u1 u2 u10 + (19) (20) R6 = −k26 u1 u5 + k24 u12 u7 K4 k25 u3 u12 K5 k26 u4 K6 (27) (28) (29) (30) The following Eqs. (31)–(37), based on Eq. (24) govern the diffusion and reaction processes for CO2 absorption into an aqueous solution containing AMP and PZ. 1188 A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 CO2 balance: 6 ju1 j 2 u1 = D1 + Ri jt jx2 (31) i=1 Total carbon (from CO2 ) balance: ju1 ju4 ju6 ju10 ju11 ju12 + + + + +2 jt jt jt jt jt jt j 2 u1 j2 u4 j 2 u6 j2 u10 = D1 + D4 + D6 + D10 2 2 2 jx jx jx jx2 2 2 j u11 j u12 + D11 + 2D12 jx2 jx2 (32) (33) −D1 (34) (35) dx at x = 0 and t > 0 (44) = kg (p1 − H1 u1 (0, t)) at x = 0 and t > 0 (45) For negligible mass-transfer resistance in the gas phase, Eq. (45) reduces to (46) Absorption rate and enhancement factor: The differential equations are integrated from t = 0 to , the contact time, to obtain the concentration profile of CO2 in the liquid film. The time averaged absorption rate per unit interfacial area is obtained by Eq. (47) over the contact time, . R=− j u1 (0, t) dt 0 jx D1 (47) For wetted wall contactor, the contact time, , is calculated using = (37) K7 = u6 u7 u4 (38) K8 = u9 u8 u7 (39) K9 = u11 u10 u7 (40) u3 u2 u7 (41) K11 = u5 u7 (42) Thus, there are 12 partial differential–algebraic equations which can be solved for the concentration profiles of the 12 chemical species (u1 , . . . , u12 ) present in the aqueous solutions of (AMP+PZ). 2/3 2h 3 1/3 d 3 g VL (48) The enhancement factor, E, defined as the ratio of the rate of absorption of a solute gas to that if there is no reaction, is given by E= Instantaneous reactions assumed to be at equilibrium: K10 = j u1 (36) Electroneutrality Balance: ju3 ju7 ju9 ju4 ju5 ju6 ju10 ju12 + + − − −2 − −2 jt jt jt jt jt jt jt jt 2 2 2 2 2 j u3 j u7 j u9 j u4 j u5 = D3 + D7 + D9 − D4 − D5 jx2 jx2 jx2 jx2 jx2 2 2 2 j u6 j u10 j u12 − 2D6 − D10 − 2D12 jx2 jx2 jx2 =0 u1 (0, t) = u∗1 = p1 /H1 Dicarbamate balance: ju12 j2 u12 = D12 − R4 − R5 jt jx2 (43) For the volatile chemical species i = 1 (CO2 ), the mass-transfer rate in the gas near the interface is equal to the mass-transfer rate in the liquid near the interface: Carbamate balance: ju10 ju11 j2 u10 j2 u11 + = D10 + D11 − R2 − R3 + R4 + R5 2 jt jt jx jx2 j ui dx Total PZ balance: ju8 ju9 ju10 ju11 ju12 + + + + jt jt jt jt jt j 2 u8 j2 u9 = D8 + D9 jx2 jx2 2 j u10 j2 u11 j2 u12 + D10 + D11 + D12 2 2 jx jx jx2 ui = u0i Boundary conditions at gas–liquid interface (x = 0): At x = 0, the fluxes of all non-volatile chemical species, i.e., i = 2,3, . . . ,12, are equal to zero. Total AMP balance: ju2 ju3 j 2 u2 j2 u3 + = D2 + D3 2 jt jt jx jx2 The diffusion coefficients of various ionic species in solution are assumed to be equal to that of PZ. This amounts to neglecting the effect of electrostatic potential gradients on ion diffusion. The more rigorous approach taking into account the electrostatic potential gradients with unequal diffusion coefficients of the ionic species requires much higher computation time with little influence on the accuracies of the model calculated rates of absorption (Glasscock and Rochelle, 1989; Hagewiesche et al., 1995; Rinker et al., 1996). Initial and boundary conditions at x = ∞ : At t = 0 (for all x 0) and at x = ∞ (for all t 0), the concentration of chemical species, i = 1, . . . ,12, are equal to their liquid bulk concentrations, i.e., R kL (u∗1 − u01 ) (49) 2.4. Method of solution The suggested model consists of partial differential and algebraic equations. The partial differential equations are transformed into ordinary differential equations in t by discretizing the spatial variable x using method-of-lines. Equally spaced nodes are used to discretize the spatial variable x. The typical number of nodes used in this work is 450 and corresponding nodal spacings are of the order of 10−8 m. The resulting system of ordinary differential equations coupled with the nonlinear algebraic equations is solved by using the subroutine DDASSL (Petzold, 1983; Brenan et al., 1989) in double precision FORTRAN 90 on a Pentium IV processor. The subroutine DDASSL uses the backward-differentiation formulas of orders one through five in a variable step integration mode to integrate the nonlinear systems of differential/algebraic equations. A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 1189 Table 1 Physicochemical properties aqueous (AMP+PZ) solutions. Mass % AMP Mass % PZ [AMP] (kmol m−3 ) [PZ] (kmol m−3 ) T (K) 28 28 28 28 25 25 25 25 22 22 22 22 30 2 2 2 2 5 5 5 5 8 8 8 8 0 0.23 0.23 0.23 0.23 0.58 0.58 0.58 0.58 0.94 0.94 0.94 0.94 0.0 3.14 3.14 3.14 3.14 2.81 2.81 2.81 2.81 2.49 2.49 2.49 2.49 3.33 298 303 308 313 298 303 308 313 298 303 308 313 313 a a a×103 (kg m−1 s) H1 b (kPa m3 kmol−1 ) D1 b×109 (m2 s−1 ) 998.72 995.93 993.07 990.30 1000.9 998.41 995.61 992.67 1003.7 1001.0 998.2 995.0 988.44 3.737 3.005 2.524 2.060 3.879 3.123 2.591 2.119 4.063 3.312 2.708 2.235 2.043 3870 4226 4674 4949 3753 4088 4620 4888 3715 4044 4513 4829 4720c 0.62 0.73 0.90 1.17 0.58 0.72 0.82 1.05 0.55 0.67 0.75 0.95 1.46c (kg m−3 ) Samanta and Bandyopadhyay (2006). b Estimated using N2 O analogy; this study. c Mandal et al. (2005). 3. Physicochemical properties The densities and viscosities of the aqueous (AMP+PZ) have been measured using standard procedures described earlier (Samanta and Bandyopadhyay, 2006) and are presented in Table 1. The solubility and diffusivity of CO2 in the aqueous amine and activated amine solutions containing 30 wt% total amine and with PZ concentrations of 2, 5 and 8 wt% are estimated in the temperature range of 298–313 K using the N2 O-analogy method. The procedures were similar to that discussed earlier by Samanta et al. (2007). Since CO2 reacts with amines, physical solubility and diffusivity of CO2 in amine solutions cannot be determined directly. As a result, one must use a nonreacting gas, e.g., nitrous oxide (N2 O) as a surrogate to CO2 in estimating the physical solubility and diffusivity of CO2 in these solvents. Clarke (1964) proposed that N2 O, a molecule with similar molecular structure, molecular weight, and similar electronic configuration as CO2 , could be used to represent CO2 behavior in reactive systems. It has been established by Laddha et al. (1981) that the ratios of CO2 /N2 O physical solubilities and CO2 /N2 O diffusivities in water and in several non-reacting organic solvents are constant. Hence, the physical solubility and diffusivity of CO2 in solvents with which it reacts can be determined by using these constant factors as shown in Eqs. (50) and (51) for H1 and D1 . H1 = HN2 O-amine D1 = DN2 O-amine HCO2 -water HN2 O-water DCO2 -water DN2 O-water ln k21 = 23.69 − 5176.49 T (50) (51) 2360.7 − 24.727 × 10−5 [MDEA] T (52) (53) The rate coefficients for the reactions of CO2 with aqueous PZ were determined by these authors and presented elsewhere (Samanta and Bandyopadhyay, 2007). The rate coefficients for reactions (2) and (4) were fitted as functions of temperature by the following Arrhenius equations: 3.5 × 104 1 1 k22 = 5.8 × 104 exp − − (54) R T 298 3.55 × 104 k24 = 5.95 × 10 exp − R 4 1 1 − T 298 (55) Values of the forward rate coefficient k26 of reaction (6) was calculated using Eq. (56) presented by Pinsent et al. (1956) for the temperature range of 273–313 K: where HN2 O-amine and DN2 O-amine are the physical solubility and diffusivity of N2 O in the amine solution. This is known as “N2 O analogy”. This analogy has been frequently used to estimate the physical solubility and diffusivity of CO2 in amine solvents (Laddha et al., 1981; Haimour and Sandall, 1984; Versteeg and van Swaaij, 1988). The estimated H1 and D1 values are presented in Table 1. The diffusion coefficients of chemical species in the liquid phase are also needed for the model. The diffusion coefficients of the ionic species have been assumed to be equal to that of PZ. The diffusion coefficients of PZ in solution have been estimated using the diffusion coefficient of MDEA, corrected for the molecular weight by multiplying with a factor of 1.38. The diffusion coefficient of MDEA has been calculated from Eq. (52) as given by Snijder et al. (1993): ln DMDEA = −13.088 − where [MDEA] is the concentration of MDEA in aqueous solution, mol m−3 . The reaction rate constant k21 for the reaction of CO2 with aqueous AMP (Eq. (1)) was calculated from the following correlation presented by Saha et al. (1995) over the temperature range 294–318 K: log10 k26 = 13.635 − 2895 T (56) The values of the independent equilibrium constants have been found out using reliable correlations and the dependent ones have been estimated by appropriate combination of the independent equilibrium constants. The values of K2 , K4 and K9 were calculated from the following correlations of Bishnoi (2000): ln K2 = −29.31 − 5615 T (57) ln K4 = −30.78 − 5615 T (58) ln 1 K9 = −8.21 − 5286 T (59) Danckwerts and Sharma (1966) correlated the data for K7 over the temperature range of 273–323 K with the following equation: log10 K7 = 6.498 − 0.0238T − 2902.4 T (60) 1190 A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 Pagano et al. (1961) reported data for K8 according to the following equation: 4351 1 = −11.91 − (61) ln K8 T Data for K11 were reported by Posey (1996) for the temperature range 273–523 K and were correlated according to the following equation: ln K11 = 132.899 − 13445.9 − 22.4773 ln T T (62) Data for (K6 K11 ) were reported by Read (1975) for the temperature range 273–523 K and were correlated according to the following equation: log10 (K6 K11 ) = 179.648 + 0.019244T − 67.341 log10 T 7495.441 − T (63) Value of K10 for aqueous AMP-PZ was calculated from the following correlation reported by Silkenbäumer et al. (1998) for the temperature range 313–353 K: ln(K10 K11 ) = 7261.78 − 22.4773 ln T + 142.58612 T (64) Reagent grade PZ ( > 99% pure) and AMP ( > 97% pure) were obtained from E. Merck, Germany, and were used without further purification. Distilled water degassed by prolonged boiling and cooled to ambient temperature under vacuum, was used for preparing the amine solutions. The CO2 gas ( > 99.9% pure) and nitrogen gas ( > 99.999% pure) were obtained from Chemtron Science Pvt. Ltd., India. A volumetric gas flow rate of about 180×10−6 m3 s−1 was used throughout all runs using mass flow controllers (Sierra Instruments, USA) for CO2 and N2 . Nitrogen gas was used as a diluent for CO2 to obtain various desired partial pressures of CO2 . The liquid flow rate was maintained at about 2×10−6 m3 s−1 by a precalibrated rotameter. The CO2 concentrations at the inlet and outlet of the wetted wall contactor was determined with a HORIBA NDIR on-line CO2 analyzer (Model: VA 3000, Japan) connected through its continuous sampling unit (Model: HORIBA VS 3000, Japan). Liquid samples were collected at the bottom of the absorption zone at fixed intervals after the absorption measurements reached steady state. The total CO2 content of each liquid sample was determined by acidulating a known volume of the sample using 6 N HCl in a glass cell placed in a thermostated bath and accurately measuring the volume of the evolved gas, as described earlier by Saha et al. (1995). The corresponding rate of absorption of CO2 in the PZ activated AMP solvent was determined from the liquid phase CO2 concentration. 4. Experimental 5. Results and discussion A 2.81×10−2 m o.d. stainless-steel wetted wall contactor was used for the absorption measurements. The apparatus and method are similar to that described by Saha et al. (1995). Absorption measurements were performed at 298, 303, 308 and 313 K. The total pressure in the absorption chamber was about 100 kPa with the CO2 partial pressures in the range of 2–14 kPa. PZ concentrations in the activated aqueous AMP solutions were in the range of 2–8 wt%, while the total amine concentration in the solution was kept at 30 wt%. The temperature of absorption was controlled within ± 0.2 K of the desired level with two circulator temperature controllers (JULABO FP 55 and Julabo F 32, FRG). Table 2 presents the measured rates and enhancement factors for the absorption of CO2 into (28 mass% AMP+2 mass% PZ), (25 mass% AMP+5 mass% PZ) and (22 mass% AMP+8 mass% PZ) at 298, 303, 308 and 313 K and 30 wt% AMP at 313 K. As shown in Table 2 and Fig. 1, the addition of small amounts of PZ to an aqueous solution of AMP results in significant enhancement in the rates of absorption. For example, for T = 313 K and CO2 partial pressure of about 5 kPa, the enhancement factor for absorption in aqueous solutions of 30 mass% AMP, (2 mass% PZ+28 mass% AMP), (5 mass% PZ+25 mass% AMP), and (8 mass% PZ and 22 mass% AMP) are about 45.1, 150.1, 209.7 and 250.5, respectively. So, by replacing 2 mass% Table 2 Experimental and model predicted results for the absorption of CO2 into aqueous (AMP+PZ) in the wetted wall contactor. [AMP] (kmol m−3 ) [PZ] (kmol m−3 ) T (K) 3.14 3.14 3.14 3.14 3.14 3.14 3.14 2.81 2.81 2.81 2.81 2.81 2.81 2.81 2.49 2.49 2.49 2.49 2.49 2.49 2.49 3.326 3.326 3.326 0.23 0.23 0.23 0.23 0.23 0.23 0.23 0.58 0.58 0.58 0.58 0.58 0.58 0.58 0.94 0.94 0.94 0.94 0.94 0.94 0.94 0.0 0.0 0.0 298 303 308 313 313 313 313 298 303 308 313 313 313 313 298 303 308 313 313 313 313 313 313 313 p1 (kPa) 4.85 4.88 4.87 4.8 1.93 6.83 13.78 4.92 4.86 4.93 4.82 1.97 6.88 13.69 4.84 4.79 4.79 4.56 1.93 6.65 13.61 5.04 7.02 14.06 (s) 0.60 0.63 0.56 0.48 0.44 0.43 0.43 0.55 0.53 0.46 0.44 0.42 0.44 0.41 0.58 0.53 0.49 0.45 0.53 0.43 0.48 0.43 0.43 0.45 kL×105 (m s−1 ) 3.64 3.83 4.52 5.57 5.82 5.87 5.88 3.66 4.18 4.78 5.51 5.64 5.50 5.70 3.46 4.03 4.43 5.20 4.77 5.28 5.00 6.56 6.61 6.46 Experimental results Predicted results R×106 (kmol m−2 s) E R×106 (kmol m−2 s) E 6.33 7.080 7.34 8.11 4.8 12.30 20.60 8.84 9.75 10.70 11.40 5.84 15.38 27.80 10.20 10.70 11.75 12.30 6.70 17.10 31.60 3.16 3.98 8.14 138.9 159.9 155.7 150.1 211.4 151.7 125.8 184.5 196.4 210.0 209.7 257.0 198.7 174.0 226.1 224.4 249.7 250.5 351.7 235.3 224.3 45.1 40.5 42.3 5.83 6.15 6.98 8.46 3.98 11.57 19.15 7.70 8.85 9.72 11.60 5.21 16.17 29.48 8.90 9.93 10.92 12.21 5.49 16.79 32.3 2.91 3.74 7.51 128.0 138.9 148.1 156.5 175.3 142.7 116.9 160.7 178.2 190.7 213.4 229.3 208.9 184.5 197.3 208.2 232.0 248.7 288.2 231.0 227.4 41.6 38.0 39.0 A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 Fig. 1. Effect of molar ratio of PZ/AMP on the enhancement of CO2 absorption into aqueous solution of (AMP+PZ): , T = 313 K. AMP with 2 mass% PZ, the enhancement factor increased by about 233%. Replacing an additional 3 mass% AMP with PZ increased the enhancement factor by an additional 40%. A further replacement of 3 mass% AMP with an equal amount of PZ resulted in increasing the enhancement factor by additional 20%. The mathematical model has been used to predict the rates of absorption of CO2 into aqueous blends of PZ and AMP. For this system, the rate constants, k23 and k25 , have been obtained by adjusting the values of rate constants in the mathematical model until the predicted rates of absorption agreed with the experimentally measured rates of absorption to within an average absolute deviation of 5% for each temperature and for a particular concentration of the (AMP+PZ) solution. It has been observed that, if reaction (3) is not considered, the model under predicts the rate of absorption by about 36% compared with the measured rates for CO2 partial pressure above 5 kPa. From this observation it appears that the role of AMP in proton transfer is also important. It is also observed that if reactions (4) and (5) are not considered, the model under predicts the rate of absorption by about 11% for CO2 partial pressure above 5 kPa. Thus, formation of PZ-dicarbamate (reactions (4) and (5)) is also important in the overall kinetics of CO2 -AMP-PZ. The estimated values of k23 and k25 for CO2 -(AMP+PZ+H2 O) have been correlated by the following the Arrhenius equations (65) and (66). 1 7.2 × 104 1 4 k23 = 3.516 × 10 exp − − (65) R T 298 k25 = 1.836 × 104 exp − 7.6 × 104 R 1 1 − T 298 (66) The estimates of k23 from this work is plotted as function of temperature in Fig. 2. The measured and model predicted rates of CO2 absorption are compared in the parity plot shown in Fig. 3. The model predicted rates of absorption for CO2 into (AMP+PZ+H2 O), as presented in Table 2 and Fig. 3, are in excellent agreement with the experimental results. The average absolute deviation (AAD) between the experimental and model results of rates of absorption of CO2 into aqueous (AMP+PZ) is about 7.7%. Typical calculated concentration profiles for absorption of CO2 into aqueous (25 mass% AMP+5 mass% PZ) are shown in Fig. 4. The concentrations of AMP and PZ near the gas–liquid interface are lower 1191 Fig. 2. Arrhenius plot for k23 for (CO2 +AMP+PZ). Fig. 3. Parity plot of model predicted rates and experimental rates of absorption of CO2 into aqueous (AMP+PZ). by about 4.08% and 30.24%, respectively, than their liquid bulk concentrations after a gas–liquid contact time of about 0.44 s. From this observation, it appears that the assumption of fast pseudo-first reaction regime for absorption of CO2 in PZ activated AMP may not be correct. 5.1. Parametric sensitivity analysis Parametric sensitivity analysis has been performed using the mathematical model developed in this study for the absorption of CO2 into PZ activated AMP solution to examine the effects of important parameters on the rates of absorption and to determine the effects of possible inaccuracies in the values of the pertinent model parameters on the accuracy of the calculated CO2 absorption rates from the model. The parameters considered for these analyses are interfacial concentration of CO2 (CO2 partial pressure), Henry's law constant for CO2 , diffusion coefficients of CO2 1192 A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 Fig. 4. Calculated concentration profile of species in the liquid film near the gas–liquid interface at the end of the absorption length: [AMP] = 2.81 kmol m−3 , [PZ] = 0.58 kmol m−3 , T = 313 K, p1 = 4.82 kPa and = 0.44 s. Fig. 6. The effect of errors in Henry's Law constant on the predicted rate of absorption of CO2 into an aqueous (AMP+PZ) solution. Fig. 5. Predicted enhancement factor for the absorption of CO2 into aqueous (AMP+PZ) as a function of CO2 interfacial concentration. Fig. 7. The effect of errors in the diffusion coefficient of CO2 on the predicted rate of absorption of CO2 into an aqueous (AMP+PZ) solution. in amine solution, and reaction rate constants, k23 . The parametric sensitivity analysis, presented in Figs. 5–9 has been performed for T = 313 K, [AMP] = 2.81 kmol m−3 , [PZ] = 0.58 kmol m−3 , p1 = 4.82 kPa, = 0.44 s. For Figs. 5–9, most of the required physicochemical and transport parameters have been taken from Tables 1 and 2 and reaction rate constants and other data are determined from the correlations presented in Section 3. Fig. 5 shows the effect of the CO2 interfacial concentration (i.e., CO2 partial pressure) on the enhancement factor. It is clear that the enhancement factor at lower interfacial concentrations (i.e., lower partial pressures) is greater than that at higher interfacial concentrations (i.e., higher partial pressures). This may be due to the fact that as the CO2 interfacial concentration decreases, the contribution of physical absorption to the overall absorption of CO2 decreases and the contributions of the chemical reactions between CO2 and (AMP+PZ+H2 O) increases. Fig. 6 shows the effect of any possible errors in the value of Henry's law constant of CO2 on the accuracy of the model pre- dicted rate of absorption of CO2 into aqueous (AMP+PZ) solution. Here, errors of −50% to +50% have been assumed in the value of the Henry's law constant and the corresponding errors in the model calculated rate of absorption have been determined to be from +87.3% to −32.6%. Hence, it is evident that the rate of absorption of CO2 in PZ activated amine is very sensitive to the value of Henry's law constant. The reason may be due to the fact that it determines the amount of CO2 that dissolves into the aqueous amine solution for the given temperature, pressure and amine concentration. It is therefore important to have accurate values of the Henry's law constant of CO2 in the (AMP+PZ) solution in order to accurately predict the rates of absorption of CO2 in aqueous (AMP+PZ) from the model. The effect of errors in the diffusion coefficient of CO2 in aqueous (AMP+PZ) on the predicted rate of absorption is shown in Fig. 7. Errors of −50% to +50% have been introduced into the value of the diffusion coefficient of CO2 , and the corresponding errors in the predicted rate of absorption has been determined to be from −27.6% to A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 1193 of −50% to +50% have been introduced into the value of k23 , and the corresponding errors in the predicted rate of absorption have been determined to be from −13.2% to +10.4%. Thus, for accuracy in the calculated rate of absorption and enhancement factor, it is also important to have accurate values of the rate constant of the reaction between CO2 and AMP in presence of PZ. 6. Conclusion Fig. 8. The effect of errors in the diffusion coefficients of all liquid phase chemical species on the predicted rate of absorption of CO2 into an aqueous (AMP+PZ) solution. In this work, the absorption of CO2 into aqueous solutions of mixture of small amounts of PZ and much larger amounts of AMP was studied experimentally and theoretically. Absorption measurement over the temperature range of 298–313 K and CO2 partial pressure range of 2–14 kPa were done using a wetted wall contactor. It has been found that the addition of small amounts of PZ to an aqueous solution of AMP, e.g., 2 wt% PZ in 28 wt% aqueous AMP, significantly enhances the rate of absorption of CO2 and enhancement factor. A comprehensive coupled mass transfer-reaction kinetics-equilibrium model has been developed to describe absorption of CO2 into PZ activated aqueous AMP solutions incorporating all important reversible reactions in the liquid phase. The model is validated by comparing the model predicted rates of absorption and enhancement factors with the experimental results of absorption of CO2 into aqueous (AMP+PZ). New kinetic parameters, k23 and k25 , for the reactions of CO2 with aqueous (AMP+PZ) have been obtained using the developed mathematical model and the measured absorption data of this work. The model has also been used for parametric sensitivity analysis to determine the effects of various important parameters on the rates of absorption and enhancement factors. Notation d Di E h H1 kL p1 [PZ] Fig. 9. The effect of errors in k23 on the predicted rate of absorption of CO2 into an aqueous (AMP+PZ) solution. +21.4%. Therefore, it is also important to have accurate values of the diffusion coefficient of CO2 in aqueous (AMP+PZ) in order to predict the rates of absorption from the model with good accuracy. The effect of errors in the values of the diffusion coefficients of all chemical species in the liquid phase except CO2 , on the predicted rate of absorption has been shown in Fig. 8. Here, errors of −50% to +50% have been considered in the values of these diffusion coefficients simultaneously, and the corresponding errors in the predicted rate of absorption have been determined to be −5.3% to +2.3%. Hence, it can be seen that relatively large errors in the values of the diffusion coefficients of all chemical species other than CO2 in the liquid phase, are not likely to introduce significant errors in the predicted rate of absorption. This observation justifies the assumption that the diffusion coefficients of all ionic species in solution are equal to that of PZ (Section 2.3). Fig. 9 shows the effect of errors in the value of the rate constant, k23 (Eq. (3)) on the predicted rate of absorption. Here, errors [PZ]initial R Ri t ui ui 0 ui * VG VL x outer diameter of the wetted wall column, m diffusion coefficient of species i in the aqueous solution, m2 s−1 enhancement factor for absorption of CO2 absorption length of the wetted wall column, m Henry's law constant for CO2 , kPa m3 kmol−1 liquid-phase mass transfer coefficient for physical absorption of a gas is defined as kL =2 D1 / , m s−1 partial pressure of CO2 in the gas phase, kPa concentration of PZ in the aqueous solution, kmol m−3 initial liquid bulk concentration of PZ, kmol m−3 average rate of absorption of CO2 per unit interfacial area, kmol m−2 s−1 reaction rate expression for reaction i independent time variable, s concentration of species i in the liquid phase, kmol m−3 initial liquid bulk concentration of species i, kmol m−3 interfacial concentration of CO2 in the liquid, kmol m−3 volumetric flow rate of gas, m3 s−1 volumetric flow rate of solution, m3 s−1 independent spatial variable, m Greek letters 1 initial CO2 loading of the aqueous amine solution, kmol CO2 , kmol total amine−1 1194 i A. Samanta, S.S. Bandyopadhyay / Chemical Engineering Science 64 (2009) 1185 -- 1194 thickness of diffusion film, m viscosity of the solution, kg m−1 s−1 stoichiometric factor density of the solution, kg m−3 References Aboudheir, A., Tontiwachwuthikul, P., Chakma, A., Idem, R., 2003. Kinetics of the reactive absorption of carbon dioxide in high CO2 -loaded, concentrated aqueous monoethanolamine solutions. Chemical Engineering Science 50, 1071–1079. Appl, M., Wagner, U., Henrici, H.J., Kuessner, K., Volkamer, F., Ernst-Neust, N., 1982. Removal of CO2 and/or H2 S and/or COS from gases containing these constituents. US Patent 4336233. Bishnoi, S., 2000. CO2 absorption and solution equilibrium in piperazine activated methyldiethanolamine. Ph.D. Dissertation, The University of Texas at Austin. Bishnoi, S., Rochelle, G.T., 2000. Absorption of carbon dioxide into aqueous piperazine: reaction kinetics, mass transfer and solubility. Chemical Engineering Science 55, 5531–5543. Bishnoi, S., Rochelle, G.T., 2002. Absorption of carbon dioxide in aqueous piperazine/methyldiethanolamine. A.I.Ch.E. Journal 48, 2788–2799. Brenan, K.E., Campbell, S.L., Petzold, L.R., 1989. Numerical Solution of Initial-Value Problems in Differential–Algebraic Equations. North-Holland, New York. Caplow, M., 1968. Kinetics of carbamate formation and breakdown. Journal of American Chemical Society 90, 6795–6803. Chakravarty, T., Phukan, U.K., Weiland, R.H., 1985. Reaction of acid gases with mixtures of amines. Chemical Engineering Progress 81, 32–36. Clarke, J.K.A., 1964. Kinetics of absorption of carbon dioxide in monoethanolamine solution at short contact times. Industrial and Engineering Chemistry Fundamentals 3, 239–245. Danckwerts, P.V., 1979. The reaction of CO2 with ethanolamines. Chemical Engineering Science 34, 443–446. Danckwerts, P.V., Sharma, M.M., 1966. The absorption of carbon dioxide into aqueous amine solutions of alkalis and amines (with some notes on hydrogen sulphide and carbonyl sulphide). The Chemical Engineer 10, CE244–CE280. Derks, P.W.J., Kleingeld, C., van Aken, C., Hogendoorn, J.A., Versteeg, G.F., 2006. Kinetics of absorption of carbon dioxide in aqueous piperazine solutions. Chemical Engineering Science 61, 6837–6854. Glasscock, D.A., Rochelle, G.T., 1989. Numerical simulation of theories for gas absorption with chemical reaction. A.I.Ch.E. Journal 35, 1271–1281. Hagewiesche, D.P., Ashour, S.S., Al-Ghawas, H.A., Sandall, O.C., 1995. Absorption of carbon dioxide into aqueous blends of monoethanolamine and Nmethyldiethanolamine. Chemical Engineering Science 50, 1071–1079. Haimour, N., Sandall, O.C., 1984. Absorption of carbon dioxide into aqueous methyldiethanolamine. Chemical Engineering Science 39, 1791–1796. Higbie, R., 1935. The rate of absorption of a pure gas into a still liquid during short periods of exposure. Transactions of the American Institute Chemical Engineers 31, 365–389. Kohl, A.L., Nielsen, R.B., 1997. Gas Purification. fifth ed. Gulf Publishing Company, Houston. Laddha, S.S., Diaz, J.M., Danckwerts, P.V., 1981. The nitrous oxide analogy: the solubilities of carbon dioxide and nitrous oxide in aqueous solutions of organic compounds. Chemical Engineering Science 36, 228–229. Lin, S.-H., Chiang, P.-C., Hsieh, C.-F., Li, M.-H., Tung, K.-L., 2008. Absorption of carbon dioxide by the absorbent composed of piperazine and 2-amino-2-methyl1-propanol in PVDF membrane contactor. Journal of the Chinese Institute of Chemical Engineers 39, 13–21. Mandal, B.P., Biswas, A.K., Bandyopadhyay, S.S., 2001. Removal of carbon dioxide by absorption in mixed amines: modeling of absorption in aqueous MDEA/MEA and AMP/MEA solutions. Chemical Engineering Science 56, 6217–6224. Mandal, B.P., Kundu, M., Bandyopadhyay, S.S., 2005. Physical solubility and diffusivity of N2 O and CO2 into aqueous solutions of (2-amino-2methyl-1-propanol+monoethaolamine) and (N-methyldiethanolamine+monoethanolamine). Journal of Chemical and Engineering Data 50, 352–358. More, J., Garbow, B., Hillstrom, K., 1980. User Guide for MINPACK-1. Argonne National Labs Report ANL-80-74, Illinois. Pagano, J.M., Goldberg, D.E., Fernelius, W.C., 1961. A thermodynamic study of homopiperazine, piperazine, and N-(2-aminoethyl)-piperazine and their complexes with copper(II) ion. Journal of Physical Chemistry 65, 1062–1064. Petzold, L.R., 1983. A Description of DASSL: A Differential/Algebraic System Solver in Scientific Computing. North-Holland, Amsterdam. pp. 65–68. Pinsent, B.R.W., Pearson, L., Roughton, F.J.W., 1956. The kinetics of combination of carbon dioxide with hydroxide ions. Transactions of the Faraday Society 52, 1512–1520. Posey, M.L., 1996. Thermodynamic model for acid gas loaded aqueous alkanolamine solutions. Ph.D. Dissertation, The University of Texas at Austin. Read, A.J., 1975. The first ionization constant of carbonic acid from 25 to 250 ◦ C and to 2000 bar. Journal of Solution Chemistry 4, 53–70. Rinker, E.B., Ashour, S.S., Sandall, O.C., 1996. Kinetics and modeling of carbon dioxide absorption into aqueous solutions of diethanolamine. Industrial and Engineering Chemistry Research 35, 1107–1114. Saha, A.K., Bandyopadhyay, S.S., Biswas, A.K., 1995. Kinetics of absorption of CO2 into aqueous solutions of 2-amino-2-methyl-1-propanol. Chemical Engineering Science 50, 3587–3598. Samanta, A., Bandyopadhyay, S.S., 2006. Density and viscosity of aqueous solutions of piperazine and (2-amino-2-methyl-1-propanol+piperazine) from 298 to 333 K. Journal of Chemical and Engineering Data 51, 467–470. Samanta, A., Bandyopadhyay, S.S., 2007. Kinetics and modeling of carbon dioxide adsorption into aqueous solutions of piperazine. Chemical Engineering Science 62, 7312–7319. Samanta, A., Roy, S., Bandyopadhyay, S.S., 2007. Physical solubility and diffusivity of N2 O and CO2 in aqueous solutions of piperazine and (NMethyldiethanolamine+piperazine). Journal of Chemical and Engineering Data, 1381–1385. Sartori, G., Savage, D.W., 1983. Sterically hindered amines for CO2 removal from gases. Industrial and Engineering Chemistry Fundamentals 22, 239–249. Seo, D.J., Hong, W.H., 2000. Effect of piperazine on the kinetics of carbon dioxide with aqueous solutions of 2-amino-2-methyl-1-propanol and piperazine. Industrial and Engineering Chemistry Research 39, 2062–2067. Silkenbäumer, D., Rumpf, B., Lichtenthaler, R.N., 1998. Solubility of carbon dioxide in aqueous solutions of 2-amino-2-methyl-1-propanol and Nmethyldiethanolamine and their mixtures in the temperature range from 313 to 353 K and pressure up to 2.7 MPa. Industrial and Engineering Chemistry Research 37, 3133–3141. Snijder, D.E., te Riele, M.J.M., Versteeg, G.F., van Swaaij, W.P.M., 1993. Diffusion coefficients of several aqueous alkanolamine solutions. Journal of Chemical and Engineering Data 38, 475–480. Sun, W.-C., Yong, C.-B., Li, M.-H., 2005. Kinetics of the absorption of carbon dioxide into mixed aqueous solutions of 2-amino-2-methyl-1-propanol and piperazine. Chemical Engineering Science 60, 503–516. Versteeg, G.F., van Swaaij, W.P.M., 1988. Solubility and diffusivity of acid gases (CO2 , N2 O) in aqueous alkanolamine solutions. Journal of Chemical and Engineering Data 33, 29–34. Visual Numerics Inc., 1994. IMSL MATH/LIBRARY: FORTRAN Subroutines for Mathematical Applications. Visual Numerics Inc., Texas. Yih, S.-M., Shen, K.-P., 1988. Kinetics of carbon dioxide reaction with sterically hindered 2-amino-2-methyl-1-propanol aqueous solutions. Industrial and Engineering Chemistry Research 27, 2237–2241.