Correlations for Equilibrium Solubility of Carbon Dioxide in Aqueous 4

ARTICLE pubs.acs.org/IECR

Correlations for Equilibrium Solubility of Carbon Dioxide in Aqueous 4-(Diethylamino)-2-butanol Solutions Teerawat Sema,† Abdulaziz Naami,† Raphael Idem,*,‡ and Paitoon Tontiwachwuthikul†,‡ †

International Test Centre for CO2 Capture (ITC), Faculty of Engineering and Applied Science, University of Regina, 3737 Wascana Parkway, Regina, Saskatchewan S4S0A2, Canada ‡ Joint International Center for CO2 Capture and Storage (iCCS), Department of Chemical Engineering, Hunan University, Changsha, 410081, P.R. China ABSTRACT: The experimental results of the equilibrium solubility of CO2 in aqueous 4-(diethylamino)-2-butanol (DEAB) solutions are reported over the temperature range of 298
333 K, CO2 partial pressure range of 10
100 kPa, and DEAB concentration range of 1
2.5 M. These results were then used to calculate, with the help of various thermodynamic models, the equilibrium solubility constant of CO2 in aqueous DEAB solutions as a function of various operating parameters. Subsequently, data for the chemical equilibrium constant that governs the CO2
DEAB
water reaction were fitted to the correlation models of Kent
Eisenberg, Austgen, Li
Shen, and Hu
Chakma and used to predict the equilibrium solubility as a function of temperature and CO2 partial pressure. The results show that these models did not represent the CO2 equilibrium solubility in aqueous DEAB solution very well, with absolute average deviations of 7.3%, 7.3%, 7.1%, and 8.3%, respectively.

1. INTRODUCTION It is generally accepted that carbon dioxide (CO2) is to blame for global warming and climate change problems. Thus, the removal of CO2 from large point exhaust gas streams is considered to be a promising step to mitigate these problems for many industrial processes, such as electric power plants, refineries, and natural gas processing plants. Amine-based CO2 capture technology has been wildly used for postcombustion CO2 removal due to its wide commercial acceptance. A newly developed amino alcohol solvent, 4-(diethylamino)-2-butanol (DEAB), is considered to be a promising solvent to capture CO2 due to its energy efficiency for regeneration, high absorption capacity, and cyclic capacity for CO2 removal.1,2 By comparing the boiling point at 10 mmHg of MEA (71 °C), MDEA (128 °C), and DEAB (110 °C), it can be seen that the volatility of DEAB is lower than that of MEA and a bit higher than that of MDEA. This implies that DEAB is a good solvent in terms of the solvent volatility. However, several investigations on DEAB performance, such as reaction kinetics, mass transfer, corrosion, and degradation behaviors, still need to be carried out before it can be used for commercial application. The vapor
liquid equilibria of the aqueous CO2/DEAB system is essential for reaction kinetics model development, which is required for process simulation and design for a CO2 treating plant.3,4 In order to obtain the vapor
liquid equilibria of the aqueous CO2/DEAB system, the chemical equilibrium constant that governs the CO2
DEAB
water reaction is required. However, it is difficult and costly to obtained the chemical equilibrium constant experimentally. Kent and Eisenberg5 proposed a predictive method for determining the chemical equilibrium constant via predicting the equilibrium solubility of acid gas at various temperatures, partial pressures of acid gas, and amine concentrations. In order to develop the mathematical model for predicting the equilibrium solubility of CO2 in aqueous DEAB solutions, the equilibrium solubility of CO2 at various operating r 2011 American Chemical Society

conditions of temperatures, DEAB concentrations, and CO2 partial pressures are required. Moreover, the development of correlations to predict the equilibrium solubility of CO2 in aqueous DEAB solutions is also important, because experimental measurements are not only costly but also time-consuming. In this work, the experimental results of the equilibrium solubility of CO2 in aqueous DEAB solutions are reported over ranges of temperatures, DEAB concentrations, and CO2 partial pressures. These results were then used to test existing mathematical correlation models for predicting the equilibrium solubility and chemical equilibrium constant of CO2 absorption in aqueous DEAB solutions for different conditions.

2. EXPERIMENTAL SECTION 2.1. Chemicals. DEAB was synthesized according to the procedure described by Tontiwachwuthikul et al.1 in our solvent synthesis laboratory in the International Test Centre for CO2 Capture (ITC) at the University of Regina. The purity of synthesized DEAB was determined by GC
MS to be in the range of 91
93%. Aqueous solutions of DEAB of desired concentrations were prepared by adding a known amount of deionized water and predetermined amounts of DEAB. CO2 and nitrogen (N2) with purities of 99.9% were supplied by Praxair Inc. 2.2. Equilibrium Solubility of CO2. The apparatus and experimental technique used for determining the equilibrium solubility of CO2 were similar to those reported in the work of Tontiwachwuthikul et al.1 The main features are a saturation cell, an absorption reactor, a mass flow meter, and a water bath with a temperature controller. For each experiment, the saturated cell and Received: April 18, 2011 Accepted: November 3, 2011 Revised: October 19, 2011 Published: November 03, 2011 14008

dx.doi.org/10.1021/ie2008345 | Ind. Eng. Chem. Res. 2011, 50, 14008–14015

Industrial & Engineering Chemistry Research

ARTICLE

Figure 1. Equilibrium solubility of CO2 in aqueous solutions of 2 M MEA, 2 M DEA, 2 M MDEA, 2 M AMP, 2 M DEAB, and 2 M PZ (lines are trend lines of the experimental results obtained from this study).2,7
10

the absorption reactor were immersed in a water bath (ColeParmer) with the temperature controller (Cole-Parmer, within the temperature range from
20 to 200 °C with a ( 0.01 °C accuracy). Two gas streams (CO2 and N2) were mixed and controlled by a flow controller (electronic Aalborg GFM-17 gas flow meter ranging from 5 mL to 5 L/min with a ( 0.15%/°C accuracy) and fed to the system. Initially, the amine solution of the desired concentration was introduced into the absorption reactor. Then, the mixed gas stream of the desired CO2 partial pressure was introduced into the saturation cell and bubbled into the absorption reactor at atmospheric pressure. The water-saturated stream from the absorption reactor was then passed through a condenser in order to recover all condensable species back to the system. The condenser was continually fed with cold water in order to keep the amine loss at a minimum. The CO2 concentration in the mixed gas stream was measured by a portable infrared (IR) CO2 gas analyzer (Nova Analytical System Inc., Hamilton, Canada, model 302 HWP ranging from 0.0 to 100.0% CO2 with a (1% accuracy). The liquid sample was then taken for the measurement of CO2 loading using the acidification technique.6 The amine concentration was determined by titration with a 1 M HCl solution to the 0.1% wt methyl orange end point. For the CO2 loading measurement, the amine sample was also treated with the excess 1 M HCl solution and then the released absorbed CO2 was collected in a gas buret containing the displacement solution. The CO2 loading measurement was repeated every 2 h until the system reached equilibrium. The equipment was validated with 2 M monoethanolamine (MEA), 2 M diethanolamine (DEA), 2 M N-methyldiethanolamine (MDEA), and 2 M 2-amino-2-methyl-1propanol (AMP), as shown in Figure 1. The results are found to be in good agreement with the literature2,7
10 with an absolute average deviation (AAD) of 3.2%. Therefore, it is concluded that the equipment and procedures for this study are considered to be properly applicable for measuring the equilibrium solubility of CO2. Also, some experiments were repeated for the repeatability. It was found that the deviation of equilibrium solubility measurement is 0.6%.

3. CALCULATING EQUILIBRIUM SOLUBILITY PARAMETERS OF CO2 IN AQUEOUS DEAB SOLUTIONS The chemical equilibrium in CO2/DEAB aqueous system is governed by the following equations: K1

RR 02 N þ Hþ r s f RR 02 NHþ

ð1Þ

K2

ð2Þ

CO2 þ RR 02 N þ H2 O r s f RR 02 NHþ þ HCO3
K3

H2 O þ CO2 r s f Hþ þ HCO3


ð3Þ

K4

ð4Þ

s f HCO3
CO2 þ OH
r K5

s f Hþ þ CO3 2
HCO3
r K6

H2 O r s f Hþ þ OH


ð5Þ ð6Þ

RR0 2N represents DEAB, where R is (CH2)2CH(CH3)OH and R0 is CH2CH3. Ki is the chemical equilibrium constant for reaction i. DEAB is considered to be a tertiary amine because it has three carbon atoms attached to the nitrogen atom. In this case, DEAB does not react directly with CO2 but it acts as a base which catalyzes the hydration of CO2,11 as presented in eq 2. The expressions for the chemical equilibrium constants can be described as follows:

14009

K1 ¼

½RR 02 NHþ  K2 ¼ ½Hþ ½RR 02 N K3

ð7Þ

K2 ¼

½RR 02 NHþ ½HCO3
 ½CO2ðaqÞ ½RR 02 N

ð8Þ

dx.doi.org/10.1021/ie2008345 |Ind. Eng. Chem. Res. 2011, 50, 14008–14015

Industrial & Engineering Chemistry Research

ARTICLE

K3 ¼

½Hþ ½HCO3
 ½CO2ðaqÞ 

ð9Þ

K4 ¼

½HCO3
 ½CO2ðaqÞ ½OH


ð10Þ

K5 ¼

½Hþ ½CO3 2
 K2 ¼
½HCO3  K1 K6

ð11Þ

K6 ¼ ½Hþ ½OH


ð12Þ

On the basis of eqs 7
12, it can be seen that not all the chemical equilibrium constants are independent. Only four chemical equilibrium constants (i.e., K2, K3, K4, and K6) are independent. The remaining two (K1 and K5) can be obtained by the combination of the independent chemical equilibrium constants, as shown in eqs 7 and 11. Also, water concentration is considered to be constant due to the presence of an excess amount of water in the solution (the water concentration is much higher than the DEAB concentration). The total DEAB balance, total carbon balance, and charge balance based on the chemical equilibrium equations (eqs 1
6) can be expressed as ½RR 02 N0 ¼ ½RR 02 NHþ  þ ½RR 02 N α½RR 02 N0




ð13Þ

¼ ½CO2ðaqÞ  þ ½HCO3  þ ½CO3  2


½RR 02 NHþ  þ ½Hþ  ¼ ½OH
 þ ½HCO3
 þ 2½CO3 2


dðln PCO2 Þ ΔHabs   ¼ 1 R d T

ð17Þ

where ΔHabs is heat of CO2 absorption (J/mol), R is the universal gas constant (J/mol K), and K2 is the chemical equilibrium constant that governs the CO2
amine reaction as given in eq 2.

4. RESULTS AND DISCUSSION

ð14Þ

4.1. Equilibrium Solubility of CO2 in Aqueous DEAB Solutions. The experimental determination of the equilibrium solu-

ð15Þ

bility of CO2 in aqueous DEAB solutions was done over a temperature range of 298
333 K, a CO2 partial pressure range of 10
100 kPa, and a DEAB concentration range of 1
2.5 M. For 5.0 M DEAB, the experiment was done at 298 K and partial pressure range of 9
30 kPa. On the basis of our previous study,2 it was shown that DEAB has a very high CO2 absorption capacity compared with conventional amines such as MEA, even at practical partial pressures of CO2 used in the CO2 absorption process, as shown in Figure 1. Also, the molecular weight of DEAB is high, 145 g/mol. This also implies high viscosity at very high DEAB concentrations. Bearing in mind the viscosity limitation and the fact that it does not need a high concentration of DEAB to achieve the same CO2 recovery efficiency as conventional amines, we decided to work with DEAB concentrations in the range of 1
2.5 M. The experimental results are shown in Tables 1
4 and plotted in Figures 2 and 3. By comparing the equilibrium solubility of CO2 in 2 M DEAB with those in 2 M MEA, 2 M DEA, 2 M MDEA, 2 M AMP, 2 M PZ (Figure 1), it is found that the equilibrium solubility of CO2 in 2 M DEAB is the highest (and comparable with those in 2 M PZ). The equilibrium solubility of CO2 in 2 M DEAB is clearly higher than those in 2 M AMP, 2 M MDEA, 2 M MEA, and 2 M DEA, respectively. In addition, it can be observed from Figures 2 and 3 that the equilibrium solubility of CO2 decreases as temperature increases and increases as CO2 partial pressure increases, as expected. These results are similar to those of Maneeintr et al.,2 Tontiwachwuthikul et al.,10 and Baek et al.,18 who used aqueous solutions of MEA, AMP, AMPD, DEAB, 4-isopropylamino-2butanol, 4-piperidino-2-butanol, 4-propylamino-2-butanol, and 4(ethylmethylamino)-2-butanol. The experimental results for equilibrium solubility of CO2 in aqueous DEAB solutions at 298 K is given in Figure 2. It is found that the equilibrium solubility of CO2 in aqueous DEAB solutions also increases as DEAB concentration increases within the DEAB concentration range of 1
2.5 M. The results for 313 and 333 K follow the same trend as those of 298 K,

where [RR0 2N]0 and α are the initial DEAB concentration and CO2 loading in the aqueous DEAB solution, respectively. In addition, the physical solubility of CO2 in aqueous DEAB solution, which is described in Sema et al.,12 can be related to Henry’s law as PCO2 ¼ HeCO2 ½CO2ðaqÞ 

the eight unknowns, namely, α, [RR0 2N], [H+], [RR0 2NH+], [HCO3
], [CO2(aq)], [OH
], and [CO32
] can be calculated. In this case, [RR0 2N]0 and PCO2 are considered as the operating condition. The different calculated values of α are obtained for various values of [RR0 2N]0 and PCO2. The calculated results of the equilibrium solubility of CO2 (α) were then compared with the experimental values, and the %AAD was determined. 3.3. Heat of CO2 Absorption in Aqueous DEAB Solutions. The heat of acid gas absorption [mainly CO2 and hydrogen sulfide (H2S)] is an important property required for an acid gas removal plant because it is directly related to the steam requirements for amine regeneration. The heat of CO2 absorption can be measured experimentally using a calorimeter. It can also be estimated on the basis of the Gibbs
Helmholtz equation15,16 as shown in eq 17.

ð16Þ

where PCO2 and HeCO2 represent the partial pressure of CO2 in the gas phase and Henry’s law constant of CO2 in aqueous DEAB solution, respectively. 3.1. Correlation for K2 with Various Operating Parameters Using Existing Models. A solution of the eight coupled nonlinear algebraic equations (i.e., eqs 8
10 and 12
16) will yield the values of the eight unknowns (namely, K2, [RR0 2N], [H+], [RR0 2/NH+], [HCO3
], [CO2(aq)], [OH
], and [CO32
]). On the other hand, [RR0 2N]0, α, and PCO2 can be obtained from the experiment involving the measurement of the equilibrium solubility of CO2. HeCO2 can be estimated using the polynomial model provided in Sema et al.12 The chemical equilibrium constants K3, K4, and K6 are calculated using the correlations mentioned in Austgen et al.8,13 The values of K2 at various operating conditions, including initial DEAB concentration, temperature, CO2 loading, and CO2 partial pressure, are collected and then correlated (using a nonlinear regression package called NLREG at a confidence level of 90%) with several existing models in the literature, namely, the Kent
Eisenberg model,5 Austgen model,8,13 Li
Shen model,4 and Hu
Chakma model.14 3.2. Calculating the Equilibrium Solubility of CO2 in Aqueous DEAB. At this point, the equilibrium solubility of CO2 or α is considered as the unknown instead of K2 (since the K2 is known through a previous set of calculation). By solving the eight coupled nonlinear algebraic equations (eqs 8
10 and 12
16),

14010

dx.doi.org/10.1021/ie2008345 |Ind. Eng. Chem. Res. 2011, 50, 14008–14015

Industrial & Engineering Chemistry Research

ARTICLE

Table 1. Equilibrium Solubility of CO2 in 1.0 M Aqueous DEAB Solutions 298 K

313 K

333 K

CO2

CO2

CO2

partial

partial

partial

298 K

pressure

CO2

pressure

CO2

pressure

CO2

(kPa)

loading

(kPa)

loading

(kPa)

loading

11

0.83

11

0.78

11

0.52

13

0.87

15

0.80

13

0.56

35

0.89

36

0.83

33

0.63

60

0.90

57

0.89

62

0.71

71

0.92

66

0.89

72

0.75

99

0.93

78

0.90

99

0.81

100

0.91

Table 2. Equilibrium Solubility of CO2 in 2.0 M Aqueous DEAB Solutionsa 298 K

313 K

333 K

CO2

CO2

CO2

partial

partial

partial

pressure

CO2

pressure

CO2

pressure

CO2

(kPa)

loading

(kPa)

loading

(kPa)

loading

10

0.92

10

0.83

10

0.76

15 31

0.94 0.95

15 30

0.88 0.92

15 31

0.79 0.81

51

0.97

51

0.95

51

0.83

76

0.98

76

0.97

76

0.88

100

1.00

100

0.99

100

0.93

a

The equilibrium solubility of CO2 in 2.0 M DEAB results obtain from Maneeintr.17

Table 3. Equilibrium Solubility of CO2 in 2.5 M Aqueous DEAB Solutions 298 K

313 K

333 K

CO2 partial

CO2 partial

Table 4. Equilibrium Solubility of CO2 in 5.0 M Aqueous DEAB Solutions

CO2 partial

pressure

CO2

pressure

CO2

pressure

CO2

(kPa)

loading

(kPa)

loading

(kPa)

loading

11

0.95

11

0.89

11

0.83

13

0.96

15

0.92

13

0.85

35

1.00

36

0.98

33

0.89

60 70

1.01 1.02

57 66

0.98 0.99

62 72

0.94 0.96

99

1.03

78

1.00

99

0.98

100

1.02

but the equilibrium solubility of CO2 in 5 M aqueous DEAB solutions at 298 K is found to be lower than the equilibrium solubility of CO2 in 1.0 M DEAB at 298 K. The observed increase in equilibrium solubility of CO2 with DEAB concentration over the concentration range of 1
2.5 M could be explained as follows. At 1 M DEAB, the DEAB molecule

CO2 partial pressure (kPa)

CO2 loading

9

0.59

16

0.61

21 30

0.62 0.64

is diluted by water molecule, so there is ineffectiveness in the CO2 molecule contacting the DEAB molecule. By increasing the concentration of DEAB, the probability of a CO2 molecule coming into contact with a DEAB molecule increases, resulting in the increase of the equilibrium solubility of CO2 as DEAB concentration increases. However, at a high DEAB concentration, e.g., 5 M, the equilibrium solubility of CO2 decreased drastically and was below the value for 1 M DEAB, as shown in Figure 2. This is because at this high DEAB concentration, the solution is not only viscous (limiting mass transfer) but also saturated with the DEAB molecule; thus, not all DEAB molecules can have contact with CO2, resulting in the observed drastic decrease in equilibrium solubility of CO2 with an increase in DEAB concentration at high DEAB concentrations. 4.2. Calculation of Equilibrium Solubility of CO2 in Aqueous DEAB Solutions. In order to obtain the predicted equilibrium solubility of CO2, a correlation for K2 as a function of solubility parameters is required. By simultaneously solving eqs 8
10 and 12
16, the eight unknowns, namely, K2, [RR0 2N], [H+], [RR0 2NH+], [HCO3
], [CO2(aq)], [OH
], and [CO32
], are obtained. The K2 values obtained at different conditions were then fitted to several models, namely, Kent
Eisenberg, Austgen, Li
Shen, and Hu
Chakma models, using a nonlinear regression (NLREG) program. Similar to K2, α was estimated as a function of different operating parameters. In this case, α was the unknown while K2 was known, and all the unknowns are α, [RR0 2N], [H+], [RR0 2NH+], [HCO3
], [CO2(aq)], [OH
], and [CO32
]. The values of these unknown variables were obtained by simultaneously solving eqs 8
10 and 12
16. 4.2.1. Kent
Eisenberg Model. The Kent
Eisenberg model has been used to estimate the equilibrium solubility of H2S and CO2 in aqueous solutions of MEA and DEA by fitting the experimental solubility results with the model. This model is simple4 and depends only on temperature in kelvin (T), as shown in eq 18.   B C D E þ 2 þ 3 þ 4 Ki ¼ exp A þ T T T T

ð18Þ

This model was used for the CO2
DEAB system. The values of the parameters (A, B, C, D, and E) for K2 for this system, which were determined after regression analysis using the NLREG program, are given in Table 5. The COMSOL software was applied for solving eqs 8
10 and 12
16 for α and bulk concentrations ([RR0 2N], [H+], [RR0 2NH+], [HCO3
], [CO2(aq)], [OH
], and [CO32
]), while K2 was calculated by the Kent
Eisenberg model. The predicted results using the Kent
Eisenberg model were found not to represent the equilibrium solubility of CO2 in aqueous solution of DEAB very well, with an AAD of 7.3%. 14011

dx.doi.org/10.1021/ie2008345 |Ind. Eng. Chem. Res. 2011, 50, 14008–14015

Industrial & Engineering Chemistry Research

ARTICLE

Figure 2. Equilibrium solubility of CO2 in 1, 2, 2.5, and 5 M aqueous DEAB solution at 298 K (solid lines are trend lines).

Figure 3. Equilibrium solubility of CO2 in 2 M aqueous DEAB solution at 298, 313, and 333 K (solid lines are trend lines).

4.2.2. Austgen Model. Austgen et al.8,13 proposed a thermodynamic model for prediction of H2S and CO2 solubility in aqueous solutions of MEA, DEA, blended MEA/MDEA, and blended DEA/MDEA using the electrolyte
NRTL equation. In this model, the activity coefficient was applied as a representative of long-range ion
ion interaction and short-range binary interaction. The Austgen model is also a function of only the temperature in kelvin (T), as shown in eq 19.   C2 þ C3 ln T þ C4 T Ki ¼ exp C1 þ T

ð19Þ

This model was also applied to the CO2 DEAB system from which the values of K2 were correlated with the Austgen model in eq 19 to determine the values of the parameters (C1, C4, C1, and C4) using the NLREG program. The estimated values of these parameters are presented in Table 6. As with the previous model, the predicted equilibrium solubility results using the Austgen model were found not to represent the equilibrium solubility of CO2 in aqueous solution of DEAB very well, with an AAD of 7.3%. 4.2.3. Li
Shen Model. The Li
Shen model4 takes not only the temperature into account but it also considers the effects of CO 2 loading and free amine concentration, as 14012

dx.doi.org/10.1021/ie2008345 |Ind. Eng. Chem. Res. 2011, 50, 14008–14015

Industrial & Engineering Chemistry Research

ARTICLE

Table 5. Parameters for the Kent
Eisenberg Model for K2 in eq 18 parameters for K2

Table 8. Parameters for the Hu
Chakma Model for K2 in eq 21 parameters For K2

value

value

A


109.3

D1


6.903

B C

6.88  104
1.008  107

D2 D3

3.56  10
2
84.8

D

1.0

D4


1.92

E

1.0

Table 6. Parameters for the Austgen Model for K2 in eq 19 parameters for K2

values

C1


193

C2 C3


1.685  104 67.85

C4


0.4325

Table 7. Parameters for the Li
Shen Model for K2 in eq 20 parameters For K2

values

A1

188.7

A2 A3


581 0.916

B1


81.5

B2


132.4

B3

31.05

B4


1.314

presented in eq 20.

  A 2 A3 B2 B3 Ki ¼ exp A1 þ þ 2 þ B1 α þ þ 2 þ B4 ln½R3 N T T α α

ð20Þ

where α and [R3N] are the equilibrium solubility of CO2 and free DEAB concentration, respectively. The parameters (A1, A2, A3, B1, B2, B3, and B4) for K2 in eq 20 for the CO2
DEAB system were determined using the NLREG program. The values of these parameters are presented in Table 7. The predicted equilibrium solubility of CO2 results (using the Li
Shen model) were found to have an AAD of 7.1% compared with the experimental results. 4.2.4. Hu
Chakma Model. Hu and Chakma’s14 mathematical model has been used for the H2S and CO2 equilibrium solubility in aqueous solutions of diglycolamine (DGA). The model represents the equilibrium constant in terms of temperature, physically dissolved H2S (or CO2) concentration, and free amine concentration, as shown in eq 21. For the present study, the physically dissolved CO2 concentration in aqueous DEAB solutions was calculated using eq 16.   PCO2 Ki ¼ exp D1 þ D2 T þ D3 þ D4 ln½R3 N ð21Þ HeCO2 where PCO2/HeCO2 represents [CO2(aq)], which is the physically dissolved CO2 concentration in aqueous DEAB solutions. HeCO2 can be calculated from the work of Sema et al.12 The values of D1, D2, D3, and D4 calculated fromK2 using the Hu
Chakma model of eq 21 are presented in Table 8. By introducing Hu
Chakma model, the predicted equilibrium solubility of

CO2 results were found not to represent the experimental results very well, with an AAD of 8.3%. Several predictive models for K2(the Kent
Eisenberg, Austgen, Li
Shen, and Hu
Chakma models) were applied separately for the prediction of equilibrium solubility of CO2 in aqueous solutions of DEAB. By comparing the predicted results of CO2 equilibrium solubility, it can be said that, of all the models, the Li
Shen model seems to provide the lowest deviation from the experimental results with an AAD of 7.1%. Also, the Kent
Eisenberg model and the Austgen model provide almost the same predicted results for the equilibrium solubility of CO2 in aqueous DEAB solutions, which corroborates the assertion by Hu and Chakma19 that that there is no significant different in using the Kent
Eisenberg model and the Austgen model, because both Kent
Eisenberg and Austgen models consider that the chemical equilibrium constant (Ki) depends only on temperature in kelvin. However, it has been mentioned by various researchers4,10,14,18,20 that in order to predict the equilibrium solubility of acid gas in an aqueous amine solution, the equilibrium constant, which governs amine reaction, should not only be considered as a function of temperature but also as a function of other parameters, such as acid gas loading, free amine concentration, and physically dissolved CO2 concentration. The Li
Shen model and the Hu
Chakma model take these parameters into consideration in developing their models. The Li
Shen model correlates Ki with temperature, CO2 loading, and free DEAB concentration, as shown in eq 20. The results from this study showed that the CO2 equilibrium solubility results using K2 calculated from the Li
Shen model provide the lowest deviation from the experimental results, with an AAD of 7.1%. Even though the Hu
Chakma model considers Ki as a function of temperature, physically dissolved CO2 concentration, and free DEAB concentration as shown in eq 21, the predicted CO2 equilibrium solubility results using K2 calculated from the Hu
Chakma model provide the highest deviation from the experimental results, with an AAD of 8.3%. One parameter that makes the Hu
Chakma model different from the Li
Shen model is the dependency on physically dissolved CO2 concentration (which is CO2 loading in the case of the Li
Shen model). CO2 loading represents the capacity of the amine to react with CO2 in terms of mole of absorbed CO2 per mole of amine, and therefore, CO2 loading directly relates with the CO2
amine reaction. This shows that the physically dissolved CO2 concentration might not be a very good parameter for correlation with Ki. It is generally accepted that a chemical equilibrium constant is strongly related to the chemical reaction and equilibrium concentration of the involved species. Thus, using physically dissolved CO2 concentration instead of CO2 loading results in a bigger deviation of the predicted CO2 equilibrium solubility. After comparing the predicted CO2 equilibrium solubility results using several predictive models for K2 (the Kent
Eisenberg, Austgen, Li
Shen, and Hu
Chakma models), it can be seen 14013

dx.doi.org/10.1021/ie2008345 |Ind. Eng. Chem. Res. 2011, 50, 14008–14015

Industrial & Engineering Chemistry Research

ARTICLE

Figure 4. pKa of DEAB at various temperatures of 298, 313, and 333 K (solid line is predicted line).

that none of the models represent the equilibrium solubility of CO2 in aqueous DEAB solutions very well, with AADs of 7.3%, 7.3%, 7.1%, and 8.3%, respectively. The pKa or pK1 was then calculated using eq 7 while K2 was calculated by the Li
Shen model of eq 20. A plot of the pKa against T over the temperature range of 298
333 K was then constructed. This plot of pKa vs T is presented in Figure 4. It is seen that the pKa value of aqueous DEAB solution is inversely proportional to the temperature according to the following equation: pKa ¼ ð
2:3  10
2 ÞT þ 18:54

ð22Þ

4.3. Heat of CO2 Absorption in Aqueous DEAB Solutions. Kim and Svendsen,15 Rho et al.,16 and Lee et al.21 estimated the heat of CO2 absorption in MEA, DEA, and MDEA using the Gibbs
Helmholtz equation (eq 17). They observed that the estimated results were close to the experimental calorimeter results of Carson et al.22 In the present study, the heat of CO2 absorption in aqueous DEAB solution (ΔHabs) were calculated using eq 17. The estimated heat of CO2 absorption in aqueous DEAB solution was found to be
41.4 kJ/mol, as presented in Table 9 together with the experimental and the estimated results for MEA, DEA, and MDEA obtained from the works of Kim and Svendsen,15 Rho et al.,16 and Lee et al.21 The minus symbol represents that the reaction of CO2 with MEA, DEA, MDEA, or DEAB is exothermic. By comparing the heat of CO2 absorption in aqueous DEAB with those in aqueous MEA, DEA, and MDEA, it is found that the heat of CO2 absorption in aqueous DEAB solution is lower than those in MEA, DEA, and MDEA. On the other hand, it can be said that the heat of desorption or regeneration for DEAB is 41.4 kJ/mol. Conversely, the regeneration energy for DEAB is lower than for MEA, DEA, and DEAB, respectively, which corresponds well with the work of Maneeintr et al.,2 who compared the regeneration energy of MEA and DEAB in terms of cyclic capacity at different temperature ranges. They found that DEAB has a lower regeneration energy than MEA.

Table 9. Heat of CO2 Absorption in Aqueous Solutions of MEA, DEA, MDEA, and DEAB heat of CO2 absorption, ΔHabs(kJ/mol) amine

estimated

experimental

MEA


84.3b


82e

DEA


66.9c


69e

MDEA


54.6d


49e

DEAB


41.4

N/A

b

Esimated values are obtained from Kim and Svendsen.16 c Esimated values are obtained from Lee et al.21 d Esimated values are obtained from Rho et al.17 e Experimental values are obtained from Carson et al.22

Generally, the reboiler heat duty (Q) required for stripping can be approximated as the sum of three terms, the heat required for CO2 desorption (ΔHdes), the heat required to generate water vapor at the top of the column (QH2Ogen), and the sensible heat requirement (Qsens), as shown in the equation below: Q ¼ ΔHdes þ QH2 Ogen þ Qsens

ð23Þ

Even though Oyenekan23 mentioned that a greater heat of absorption will always provide better overall energy performance in CO2 absorption with thermal swing regeneration, generally, the values of QH2Ogen will be similar for all amine systems, while Qsens for different amines are small compared to Qwater and Qgen. The one parameter that changes the most with type of amine is ΔHdes.24 Therefore, reboiler heat duty (Q) is dominated by ΔHdes. Also, if the ΔHdes is high, then reboiler heat duty (Q) will also be high. In the case of DEAB, it has been observed that the ΔHdes of DEAB is lower than those of MEA, DEA, and MDEA, respectively. Therefore, it can be mentioned that the reboiler heat duty (Q) for regeneration of DEAB is lower than that for MEA, DEA, and MDEA, respectively. 14014

dx.doi.org/10.1021/ie2008345 |Ind. Eng. Chem. Res. 2011, 50, 14008–14015

Industrial & Engineering Chemistry Research

5. CONCLUSIONS This study presents experimental data on the equilibrium solubility of CO2 in aqueous DEAB as a function of CO2 partial pressure (in the range of 10
100 kPa), DEAB concentration (in the range of 1
2.5 M), and temperature (in the range of 298
333 K). Ot was found that the predicted CO2 equilibrium solubility results using several correlation models for K2 (the Kent
Eisenberg (7.3% AAD), Austgen (7.3% AAD), Li
Shen (7.1% AAD), and Hu
Chakma (8.3% AAD) models) do not represent the equilibrium solubility of CO2 in aqueous DEAB solution very well. The estimated heat of CO2 absorption in aqueous DEAB solution using the derived Gibbs
Helmholtz equation is found to be
41.4 kJ/mol which is lower than for MEA, DEA, and MDEA, respectively. Therefore, the regeneration energy for DEAB is lower than that for MEA, DEA, and MDEA, respectively. ’ AUTHOR INFORMATION Corresponding Author

*Tel.: +1 306 585 4470. Fax: +1 306 585 4855. E-mail: Raphael. [email protected].

’ ACKNOWLEDGMENT The financial support from International Test Centre for CO2 Capture (ITC) at the University of Regina and the Natural Sciences and Engineering Research Council of Canada (NSERC) is gratefully acknowledged. ’ NOMENCLATURE AMP = 2-amino-2-methyl-1-propanol AMPD = 2-amino-2-methyl-1,3-propanediol DEA = diethanolamine DEAB = 4-(diethylamino)-2-butanol DGA = diglycolamine FEM = finite element method ΔHabs = heat of CO2 absorption, kJ/mol HeCO2 = solubility of CO2, kPa m3/kmol H2S = hydrogen sulfide Ki = chemical equilibrium constant MEA = monoethanolamine MDEA = N-methyldiethanolamine PCO2 = partial pressure of CO2, kPa PZ = piperazine [R3N] = free amine concentration [R3N]0 = initial amine concentration R = universal gas constant, J/mol K T = temperature, K Greek Letters

α = CO2 loading, mol CO2/mol amine

’ REFERENCES (1) Tontiwachwuthikul, P.; Wee, A. G. H.; Idem, R. O.; Maneeintr, K.; Fan, G. J.; Veawab, A.; Aroonwilas, A.; Chakma, A. Method for Capturing Carbon Dioxide from Gas Streams. US Patent Application. US Patent Application, No. US 2008/0050296 A1, 2008. (2) Maneeintr, K.; Idem, R. O.; Tontiwachwuthikul, P.; Wee, A. G. H. Synthesis, Solubilities, and Cyclic Capacity of Amino Alcohols for CO2 Capture from Flue Gas Streams. Energy Procedia 2009, 1, 1327. (3) Rochelle, G. T. Research Needs for Acids Gas Kinetics and Equilibria in Alkanolamine Systems. Annu. Conv. Proc.—Gas Proc. Assoc. 1991, 66.

ARTICLE

(4) Li, M. H.; Shen, K. P. Calculation of Equilibrium Solubility of Carbon Dioxide in Aqueous Mixtures of Monoethanolamine and Methyldiethanolamine. Fluid Phase Equilib. 1993, 85, 129. (5) Kent, R. L.; Eisenberg, B. Better Data for Amine Treating. Hydrocarbon Process. 1976, 55, 87. (6) Horwitz, W. Association of Official Analytical Chemists (AOAC) Methods; George Banta Co.: Menasha, WI, 1975. (7) Benamar, A.; Aroua, M. K. Modeling of CO2 Solubility and Carbamate Concentration in DEA, MDEA and Their Mixtures Using the Deshmukh
Mather Model. Fluid Phase Equilib. 2005, 231, 150. (8) Austgen, D. M.; Rochelle, G. T.; Chen, C. C. Model of Vapor
Liquid Equilibria for Aqueous Acid Gas
Alkanolamine System. 2. Representation of H2S and CO2 Solubility in Aqueous MDEA and CO2 Solubility in Aqueous Mixtures of MDEA with MEA or DEA. Ind. Eng. Chem. Res. 1991, 30, 543. (9) Jou, F. Y.; Mather, A. E.; Otto, F. D. Solubility of H2S and CO2 in Aqueous Methyldiethanolamine Solutions. Ind. Eng. Chem. Process Des. Dev. 1992, 21, 539. (10) Tontiwachwuthikul, P.; Meisen, A.; Lim, J. C. Solubility of CO2 in 2-Amino-2-methyl-1-propanol Solutions. J. Chem. Eng. Data 1991, 36, 130. (11) Kohl, A.; Nielsen, R. B. Gas Purification; Gulf Publishing Co.: Houston, TX, 1997. (12) Sema, T.; Edali., M.; Naami, A.; Idem, R.; Tontiwachwuthikul, P. Solubility and Diffusivity of N2O in Aqueous 4-(Diethylamino)-2butanol Solutions. Ind. Eng. Chem. Res. 2011, submitted. (13) Austgen, D. M.; Rochelle, G. T.; Peng, W.; Chen, C. C. Model of Vapor
Liquid Equilibria for Aqueous Acid Gas
Alkanolamine Systems Using the Electrolyte
NRTL Equation. Ind. Eng. Chem. Res. 1989, 26, 1060. (14) Hu, W.; Chakma, A. Modelling of Equilibrium Solubility of CO2 and H2S in Aqueous Diglycolamine (DGA) Solutions. Can. J. Chem. Eng. 1990, 68, 523. (15) Kim, I.; Svendsen, H. F. Heat of Absorption of Carbon Dioxide (CO2) in Monoethanolamine (MEA) and in 2-(Aminoethyl)ethanolamine (AEEA) Solutions. Ind. Eng. Chem. Res. 2007, 46, 5803. (16) Rho, S. W.; Lee, Y. K.P.; Nam, J. S.; Son, S. C.; Min, J. E. B.M. Solubility of CO2 in Aqueous Methyldiethanolamine Solutions. J. Chem. Eng. Data 1997, 42, 1161. (17) Maneeintr, K. New Solvents for Acid Gas Treating. Ph.D. thesis, University of Regina, Regina, Canada, 2010. (18) Baek, J. I.; Yoon, J. H.; Eum, H. M. Prediction of Equilibrium Solubility of Carbon Dioxide in Aqueous 2-Amino-2-methyl-1,3-propanediol Solutions. Korean J. Chem. Eng. 2000, 17, 484. (19) Hu, W.; Chakma, A. Modelling of Equilibrium Solubility of CO2 and H2S in Aqueous Amino Methyl Propanol (AMP) Solutions. Chem. Eng. Commun. 1990, 94, 53. (20) Beak, J. I.; Yoon, J. I. Solubility of Carbon Dioxide in Aqueous Solutions of 2-Amino-2-methyl-1,3-propanediol. J. Chem. Eng. Data 1998, 43, 635. (21) Lee, J. I.; Otto, F. D.; Mather, A. E. Solubility of Carbon Dioxide in Aqueous Diethanolamine Solutions at High Pressures. J. Chem. Eng. Data 1972, 17, 465. (22) Carson, J. K.; Marsh, K. N.; Mather, A. E. Enthalpy of Solution of Carbondioxide in (Water+Monoethanolamnine, or Diethanolamine, or NMethylidiethanolamine) and (Water + Monoethanolamine + N-Methylidiethanolamine) at T = 298.15 Ka. J. Chem. Thermodyn. 2000, 32, 1285. (23) Oyenekan, B. A. Modeling of Strippers for CO2 Capture by Aqueous Amines. Ph.D. thesis, University of Texas. Austin, TX, 2007. (24) Gelowitz, D., Idem, R. O., Tontiwachwuthikul, P. Method and Apparatus for Recovering a Gaseous Component from an Incoming Gas Stream. US provisional patent, May 29, 2007; PCT application, May 29, 2008; PCT/CA2008/001029.

14015

dx.doi.org/10.1021/ie2008345 |Ind. Eng. Chem. Res. 2011, 50, 14008–14015