Modelling Reactive Absorption of CO2 in Packed Columns [PDF]

Zitiervorschau

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

Contents lists available at ScienceDirect

Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd

Modelling reactive absorption of CO2 in packed columns for post-combustion carbon capture applications F.M. Khan 1 , V. Krishnamoorthi 2 , T. Mahmud ∗ Institute of Particle Science and Engineering, School of Process, Environmental and Materials Engineering, The University of Leeds, Leeds LS2 9JT, UK

a b s t r a c t A rate-based process model for the reactive absorption of carbon dioxide (CO2 ) from a gas mixture into an aqueous monoethanolamine (MEA) solution in a packed column is developed. The model is based on the fast second-order kinetics for the CO2 –MEA reactions and takes into account the mass transfer resistances. The heat effects associated with the absorption and chemical reaction are included through energy balances in the gas and liquid phases. Appropriate correlations for the key thermodynamic and transport properties and for the gas–liquid mass transfer are incorporated into the model to ensure reliable predictions. The model predictions are validated by simulating a series of experiments conducted in pilot and industrial scale absorption columns with random and structured packings reported in the literature. Comparisons between the simulation results and the experimental data reveal good quality predictions of the gas phase CO2 and MEA concentrations and the liquid temperature along the column height. The sensitivity studies reveal that the correlations for the gas- and liquid-film mass transfer coefficients given by Onda et al. (1968) provide better predictions than the penetration theory of Higbie (1935) and the correlation of Bravo et al. (1985). © 2010 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. Keywords: Carbon capture; CO2 –MEA absorption; Reactive absorption; Process modelling

1.

Introduction

In recent years, there has been increasing demand for a significant reduction of carbon dioxide (CO2 ) emissions from industrial sources to alleviate the global warming problem. Following the 1997 Kyoto Protocol, the European Union has set a target of 20% reduction of CO2 emission by the year 2020. In order to meet this target, a significant reduction of CO2 release from fossil fuel fired thermal power plants, which contribute approximately 25% to the total global emissions (WRI, 2007), will be required. This can be achieved by adopting an effective strategy for carbon capture and storage (CCS) such as the pre-combustion CO2 capture technology as used in the integrated gasification and combined-cycle (IGCC) plants, oxyfuel combustion for the production of sequestration-ready CO2 or the post-combustion CO2 capture

from flue gases which would otherwise be vented to the atmosphere (Wilson and Gerard, 2007). The latter technology is being considered by the power generation industry as one of the potential options. In principle, CO2 can be removed from a gas stream using several separation processes including physical/chemical absorption into a liquid solvent, adsorption on solids, permeating through membranes, and chemical conversion. Reactive absorption using aqueous alkanolamine solutions, including monoethanolamine (MEA) solutions, in packed columns are widely used for the separation of CO2 from process gas streams in the chemical and petroleum industries. This is a well-developed technology and is currently the most preferred process approach for the CO2 capture from power plant flue gases (Freund, 2003; Idem and Tontiwachwuthikul, 2006). However, packed-bed absorption columns are known to suffer from various operational problems including high gas

∗ Corresponding author at: Institute of Particle Science and Engineering, School of Process, Environmental and Materials Engineering, The University of Leeds, Engineering (Houldsworth) Building, Leeds LS2 9JT, UK. E-mail address: [email protected] (T. Mahmud). Received 15 May 2010; Received in revised form 14 September 2010; Accepted 30 September 2010 1 BP Exploration, 1–4 Wellheads Avenue, Dyce, Aberdeen, AB21 7PB, UK. 2 Hindustan Unilever Research Centre, 64 Main Road, Whitefield, Bangalore-560066, India. 0263-8762/$ – see front matter © 2010 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.cherd.2010.09.020

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

Gas-liquid interface Gas film

pA

Liquid film

CB

pAi A

Bulk liquid

CAi

Bulk gas

CBi

CA

Fig. 1 – Gas–liquid reaction model based on the two-film theory of mass transfer (actual concentration profiles in the liquid phase depend on the reaction regime). Adapted from Levenspiel (1999). phase pressure drop, liquid channelling and flooding of the packing materials resulting in a poor gas–liquid contact. These problems are prone to occur in the flue gas scrubbing operations, which involve low solvent flow rates and high superficial gas velocities because of the large gas volumes (e.g., a typical 400 MWe coal-fired power plant produces approximately 1.1 × 106 Nm3 /h of flue gas (Raynal and Royon-Lebeaud, 2007)) with relatively low CO2 concentrations (typically 3–5% for natural gas and 10–15% for coal combustion) and require low pressure drop as the flue gas fan typically operates slightly above the atmospheric pressure. In order to achieve high CO2 capture efficiency, there is a need for improved design of packed-bed absorption columns and optimisation of operating conditions, which can be facilitated via the use of advanced process models for reactive absorption. Rigorous theories of absorption with chemical reactions and the rate expressions for different gas–liquid reaction regimes are well documented by Astarita (1967) and Danckwerts (1970), which can be used as a basis for devising a process model for the design and simulation of packed columns for the CO2 absorption processes. In the past, different modelling approaches (see review in Kenig et al., 2001) ranging from a relatively simple equilibrium stage model to more rigorous rate-based model with electrolyte solution chemistry have been used for various gas–liquid systems. A rate-based methodology accounting for the interfacial mass transfer and finite rate chemistry is considered as an appropriate approach for the modelling of industrial absorption columns as phase equilibrium is difficult to achieved in such systems (Lawal et al., 2009). This approach also allows the effects of column hydrodynamics and internals on the absorption process to be included via the correlations for the mass transfer coefficients, gas–liquid interfacial area and hold-up (Kenig et al., 2001). A general rate-based modelling methodology for the gas absorption with chemical reactions based on the steady-state, adiabatic physical absorption model of Treybal (1969) was proposed by Pandya (1983). Over the last two decades, a number of mathematical models of reactive absorption of CO2 have been developed (e.g., Tontiwachwuthikul et al., 1992; Pintola et al., 1993; Aroonwilas et al., 2003; Freguia and Rochelle, 2003; Tobiesen et al., 2007; Lawal et al., 2009) based on Pandya’s (1983) approach for the simulation of CO2 capture processes using alkaline and MEA solutions. The work reported in this paper focuses on the development and detail validation of a rate-based process model for

1601

the simulation of reactive absorption of CO2 from a mixture of air–CO2 into aqueous MEA solutions in packed columns under adiabatic conditions. The model accounts for the mass transfer resistances in the gas and liquid films and is based on the fast second-order kinetics for the CO2 –MEA reactions in the liquid film. The heat effects associated with the absorption and chemical reaction are taken into account through energy balances in the gas and liquid phases. Appropriate correlations for the key thermodynamic and physical properties and for the mass transfer coefficients and the gas–liquid interfacial area are incorporated into the model to ensure reliable predictions. The model predictions are validated by simulating a number of experiments conducted in packedbed absorption columns of different scale sizes, including pilot-plant columns packed with structured packing of 2.21 m height (Aroonwilas et al., 2003) and ceramic Berl saddles of 6.6 m height (Tontiwachwuthikul et al., 1992) and an industrial column of 14.1 m packing height containing stainless steel Pall rings (Pintola et al., 1993).

2.

Reactive absorption model formulation

In order to develop a comprehensive mathematical model for the absorption of CO2 in aqueous MEA solutions, it is essential to understand the underlying mechanism of the absorption process coupled with chemical reactions. This encompasses mass transfer of CO2 from the bulk gas to the MEA solution, reactions between CO2 and MEA species and the associated kinetics regimes. The transport of CO2 from the gas to the liquid phase can be modelled using the two-film mass transfer theory of Whitman (1923), as depicted in Fig. 1. CO2 is transported through the gas-film by molecular diffusion, absorbed in the liquid and then diffuses through the liquid-film. In principle, the reaction between a dissolved gas and a solvent can take place at any location in the liquid phase, which depends on the kinetics regime. However, the CO2 –MEA reaction is characterised as a fast reaction (Astarita, 1967) which occurs within the liquidfilm only, consequently the concentration of CO2 reduces to zero somewhere within this region. The reactive absorption process model developed in this study following the approaches used in the earlier work of Treybal (1969) and Pandya (1983) consists of a reaction model using the two-film model for fast reaction incorporated into a gas–liquid contactor model based on the mass and energy balances. The model is described below.

2.1.

Chemistry of CO2 –MEA system

The mechanism of reaction between CO2 and amines is highly complex and in spite of being studied extensively both theoretically and experimentally (see for details Astarita, 1967; Danckwerts, 1970), it is not well understood (Pintola et al., 1993). The reaction mechanism for the CO2 –MEA system may be represented by the following overall reactions (Astarita, 1967): Carbamate formation: CO2 + 2RNH2 → RNHCOO− + RNH3 +

(R1)

Bicarbonate formation: CO2 + RNH2 + H2 O → HCO3 − + RNH3 +

(R2)

1602

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

Reversion of carbamate to bicarbonate: RNHCOO− + CO2 + 2H2 O → HCO3 − + 2RNH3 +

(R3)

where RNH2 is MEA and R is OH·CH2 ·CH2 . Astarita (1967) has suggested that for CO2 /MEA mole ratio less than 0.5, the rate of bicarbonate formation is insignificant and the overall reaction may be expressed by R1, which is considered to be approximately irreversible. The overall reaction rate is second-order, first-order with respect to both CO2 and MEA, and is expressed in terms of the molar concentrations of CO2 and MEA as: r = k2 [CO2 ][RNH2 ]

(1)

Danckwerts (1970) has suggested that the use of molar concentrations in the rate equation instead of thermodynamic activities is valid. The second-order rate constant, k2 , depends on the temperature and is given by Hikita et al. (1977): log(k2 ) = 10.99 −

2.2.

2152 T

(2)

Reaction model

The overall rate of absorption of CO2 in an aqueous solution of MEA taking into account the mass transfer and chemical reaction rates can be expressed based on the two-film model for a fast second-order reaction as (Levenspiel, 1999; Doraiswamy, 2001): rCO2 =

pCO2 (1/kg a) + (HCO2 /kol aE)

(3)

where a is the gas–liquid interfacial area per unit volume of packing, E is the enhancement factor, HCO2 is the Henry’s law constant, kg and kol are the gas- and liquid-film physical mass transfer coefficients, respectively, and pCO2 is the partial pressure of CO2 in bulk gas phase. The enhancement factor is defined as the ratio of the liquid-film mass transfer coefficient for chemical absorption, kl , to the liquid-film mass transfer coefficient for physical absorption, kol . One of the most important yet difficult aspects of modelling reactive absorption processes is the determination of the enhancement factor (Tontiwachwuthikul et al., 1992), which requires solutions of coupled diffusion-reaction differential equations in the liquid film. As it stands, exact analytical solutions do not exist for second-order kinetics. Approximate solutions have been provided by a number researchers (e.g., van Krevelen and Hoftijzer, 1948; Wellek et al., 1978; DeCoursey, 1982). In this work, following previous studies (Tontiwachwuthikul et al., 1992; Pintola et al., 1993), the explicit relations developed by Wellek et al. (1978) in terms of the enhancement factor for an instantaneous reaction, E∞ , and Hatta modulus, M, has be adopted: E=1+

1 [(1/(E∞ − 1))

where E∞ = 1 +

1.35

 C D  Bl B bDAl CAi

+ (1/(E1 − 1))

1.35 1/1.35

(4)

]

√ M D k2 C and M2 = Al 2 Bl , E1 = √ tanh M (kol )

where CAi is the CO2 concentration at the gas–liquid interface, CBl is the MEA concentration in bulk liquid, and DAl and DBl are the diffusivity of CO2 and MEA in liquid, respectively. The

values of E given by Eq. (4) deviate by less than 3% from the most accurate solution of van Krevelen and Hoftijzer (1948) (Tontiwachwuthikul et al., 1992). The gas- and liquid-film mass transfer coefficients and the interfacial area for random packings are obtained from the widely used correlations provided by Onda et al. (1968): kg RT = C1 av DAg

kol

  1/3 l l g



0.7 

G av g

g g DAg

1/3 (av dp )

−2.0

 L 2/3   −0.5 l

= 0.0051



aw l

 0.75 

aw = 1 − exp −1.45 av

c l

l DAl

L av l

0.1 

L2 av l 2 g

(5a)

(av dp )

−0.05 

0.4

L2 l l av

(5b)

0.2  (5c)

where av is the total packing surface area per unit volume of packing, aw is the surface area of wetted packing per unit volume of packing, C1 is a dimensionless constant which depends on the packing size, DAg is the gas phase diffusivity of CO2 , dp is the nominal packing size, G and L are the flow rates of gas and liquid, R is the universal gas constant, T is the gas temperature,  c is the critical surface tension of packing material and  l is the surface tension of liquid, g and l are the viscosity of gas and liquid, and g and l are the density of the gas and liquid. The surface area of wetted packing per unit volume of packing was used in Eq. (3). These parameters for structured packing can be estimated using the correlations given by Bravo et al. (1985).

2.3.

Physical and thermodynamic properties

The calculation of absorption rates requires data for the physical properties of the gas and liquid phases, as well as the solubility of gas and diffusivity in the liquid phase. The diffusivity of a solute gas in a liquid is strongly dependent on both temperature and viscosity of the liquid. The following semi-empirical correlations that relate CO2 diffusivity to both of these parameters given by Versteeg and van Swaaij (1988) was used: DCO2 ,l = DCO2 ,w

  0.8 w

(6a)

l

DCO2 ,w = 2.35 × 10−6 exp

 −2119  T

(6b)

where DCO2 ,l and DCO2 ,w are the diffusivity of CO2 in the solution and in water, l and w are the viscosity of the solution and water, and T is the solution temperature. The viscosity ratio in Eq. (6a) was obtained from the following correlation (Weiland et al., 1998) in terms of the weight percent of MEA in the solution, w, CO2 loading, ˛, and the solution temperature: l = exp w



(21.186w + 2373)[˛(0.01015w + 0.0093T − 2.2589) + 1]w T2

(7)

The Henry’s law constant was determined using the correlation proposed by Danckwerts (1970) to calculate the

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

1603

solubility of CO2 in MEA solutions as a function of MEA concentration, CMEA , and solution temperature: HCO2 = 10(5.3−0.035CMEA −1140/T)

(8)

This approach was adopted instead of a rigorous method based on the electrolyte NRTL model (Austgen et al., 1989) for the vapour-liquid equilibrium of CO2 –MEA system for the ease of computation.

2.4.

Gas–liquid contactor model

A gas–liquid contactor model is developed using appropriate forms of the species and energy conservation equations for the predictions of CO2 and MEA concentrations and temperatures in the gas and liquid phases as a function of packing height. The model is based on the following assumptions: • Steady-state and adiabatic operations. • Plug-flow of gas and liquid. • Diffusion in the axial direction is negligible compared with convective transport. • The gas and liquid flow rates are constant throughout the column (i.e., dilute gas–liquid system). • Evaporation of solvent is negligible. • Pressure is constant.

2.4.1.

Mass balance

dNCO2 ,g d(uz CCO2 ,g ) = = SCO2 dz dz

(9)

where CCO2 ,g is the molar concentration of CO2 in the gas phase, NCO2 ,g (= uz CCO2 ,g = GYCO2 ) is the molar flux of CO2 , G is the molar flow rate of gas per unit cross-sectional area of the column, SCO2 is the source term representing the overall rate of absorption of CO2 , uz is the gas velocity in the axial direction (along the column height), and YCO2 is the mole ratio of CO2 in the gas phase. The source term for a fast second-order reaction is given by Eq. (3). Eq. (9) is discretised over an incremental height of the column (z), as shown in Fig. 2, using a forward difference scheme. The MEA concentration in the solution can be obtained by performing a mass balance over the incremental height: (YCO2 i − YCO2 i+1 )bG L

(10)

where b is the stoichiometric factor of reaction (R1), XMEA is the mole ratio of MEA in water, and L is the molar flow rate of water.

2.4.2.

Gas phase energy equation: GYCO2 Cpg

dTg = hg a(Tg − Tl ) dz

(11)

Liquid phase energy equation:

The variation of CO2 concentration in the gas phase along the height of the column is given by the following steady-state, one-dimensional species conservation equation:

XMEAi = XMEAi+1 +

Fig. 2 – Schematic of a packed-bed CO2 absorption column with an infinitesimal element for mass and energy balances.

Energy balance

The absorption of CO2 in a MEA solution results in the release of heat of solution followed by heat of reaction due to the exothermic chemical reaction. Both of these effects increase the liquid temperature, and consequently the gas temperature due to it being in direct contact with the liquid, from the inlet to the outlet of the column. The variations of the gas and the liquid temperature are determined from the thermal energy equations given below.

LCpl

dTg dYCO2 dTl = GYCO2 Cpg + G(HR + HS ) dz dz dz

(12)

where Tg and Tl are the gas and the liquid temperature, respectively, hg is the gas phase heat transfer coefficient, and HR and HS are the heat of reaction and heat of solution, respectively. The values of HR and HS given by Akanksha et al. (2007) were used in the calculation. Eqs. (11) and (12) were discretised using the forward difference scheme over the incremental height as shown in Fig. 2.

2.5.

Numerical solution

The absorption column was divided into a number of infinitesimal elements of height z and the discretised species conservation and thermal energy equations were solved for each element using Microsoft Excel spreadsheet. For the simulation of counter-current absorption columns, the mass and energy balance equations for the gas phase were solved from the bottom to the top of the column, whilst the liquid phase equations were solved from the top to the bottom of the column. Calculations were performed using the given inlet concentration, temperature and flow rate of the gaseous stream at the bottom and the liquid stream at the top of the column (see Fig. 2). Although the compositions and temperatures of the streams at the outlets were known for the experimental cases simulated in this study, they were not specified as the exit conditions needed for the calculation. Instead, these data were specified as initial guess and iterative calculations were performed until a converged solution was achieved. For the estimation of the enhancement factor using Eq. (4), an iterative procedure was adapted to determine the unknown CO2 concentration at the gas–liquid interface in each element of the absorption column following an approach similar to that used by Aroonwilas et al. (2003).

1604

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

Table 1 – Experimental conditions for the CO2 absorption using MEA solutions in counter-current packed-bed absorption columns. Scale size/packing type Pilot-plant column with random packinga

Gas flow rate [mol/m2 s] Inlet CO2 conc. [%mol] Liquid flow rate [m3 /m2 h] Inlet MEA conc. [kmol/m3 ] Liquid temperature [◦ C] CO2 removal efficiency [%] Column pressure [kPa]

T13d

T15d

T19d

14.8 19.1 9.5 3.00 19.0 100.0 103.15

14.8 15.3 13.5 2.00 19.0 100.0 103.15

14.8 19.5 13.5 2.03 19.0 100.0 103.15

14.8 11.5 13.5 2.00 19.0 100.0 103.15

Industrial column with random packingc I06d

10.7 15.0 15.9 5.20 19.7 – 101.32

89.32 1.53 13.27 4.10 38.7 99.92 724.47

Tontiwachwuthikul et al. (1992). Aroonwilas et al. (2001). Pintola et al. (1993). Run number.

3.

Experimental cases

The rate-based process model was validated by simulating a number of experiments on the absorption of CO2 in aqueous MEA solutions carried out in counter-current packedbed absorption columns ranging from a small pilot-plant to a large industrial scale size. The following experimental studies (Tontiwachwuthikul et al., 1992; Pintola et al., 1993; Aroonwilas et al., 2001) reported in the literature were selected. Aroonwilas et al. (2001) carried out experiments in a small scale pilot-plant absorption column of 2.21 m packing height and 0.1 m in internal diameter containing cylindrical elements of stainless steel Gempak 4A structured packings (element height = 0.245 m, specific surface area = 466 m2 /m3 and void fraction = 0.92). The gas mixture, air–CO2 , was introduced into the column at the bottom and the solvent was introduced at the top, thus providing a counter-current flow. Having reached a steady-state, the gas phase CO2 concentrations and solution temperatures were measured at different locations along the column height. Tontiwachwuthikul et al. (1992) carried out a series of experiments in a larger countercurrent pilot-plant absorption column consisting of six packed beds with a total packing height of 6.6 m and 0.1 m in internal diameter. The column was packed with 12.7 mm ceramic Berl saddles. Steady-state gas phase CO2 and MEA concentrations and solution temperatures were measured along the column height for different gas/liquid ratio, inlet CO2 and MEA concentrations. Pintola et al. (1993) reported data taken from an acid gas treatment unit of an olefin plant. The absorption column, 26.6 m high and 1.9 m internal diameter, was packed with 50 mm stainless steel Pall rings with a total packing height of 14.1 m. The packing was divided into three sections with heights of 4.6, 4.6, and 4.9 m. The CO2 and MEA concentrations, temperatures, and flow rates were only measured at the top and the bottom of the column. The operating conditions of all the experimental runs simulated in this study are given in Table 1.

4.

Results and discussion

4.1.

Base case simulation

For detail validation of the model described in Section 2, the predictions are compared with the pilot-plant experimental

data for run #T22 (Tontiwachwuthikul et al., 1992). Fig. 3(a) shows the comparison between the predicted and measured gas phase CO2 concentration profiles along the column height. As can be seen, overall agreement between the prediction and measurement is generally good; however, the CO2 concentration is somewhat underpredicted in the initial 2.5 m of the column. This may be due to the use of empirical correlations for the estimation of thermodynamic and transport 25.00

(a)

20.00

CO2 Concentration (mol%)

c d

T22d

15.00

10.00

5.00

0.00 0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

Distance from the bottom of the column (m) 3500.00

MEA Concentration (mol/m3)

a b

Pilot-plant with structured packingb

(b)

3000.00 2500.00 2000.00 1500.00 1000.00 500.00 0.00 0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

Distance from the bottom of the column (m)

Fig. 3 – Comparison between the predicted and measured (a) CO2 and (b) MEA concentration profiles along the column height. (–) Prediction and () Expt. data of run #T22 (accuracy of CO2 data: ±2%). Tontiwachwuthikul et al. (1992).

1605

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

160.00

25.00

CO2 Concentration (mol%)

Enhancement Factor

140.00 120.00 100.00 80.00 60.00 40.00

20.00

15.00

10.00

5.00

20.00 0.00 0.00

1.00

2.00

3.00

4.00

5.00

6.00

0.00 0.00

7.00

Distance from the bottom of the column (m) Fig. 4 – The predicted enhancement factor as a function of column height for run #T22. properties and parameters associated with mass transfer in the absorption rate Eq. (3). The error associated with CO2 concentration measurement might have also contributed towards the discrepancy between the prediction and measurement. The predicted and measured MEA concentration distributions along the column are shown in Fig. 3(b). The level of agreement between the predicted and measured CO2 concentrations revealed in Fig. 3(a) is also reflected here. The predicted outlet concentrations of both CO2 and MEA exactly match the experimental data. It is evident from these figures that the concentrations decrease rapidly in the bottom part of the column, indicating that most CO2 removal from the gas takes place in that particular region. The remaining height of the column serves to further reduce the CO2 concentration to a very small value. Fig. 4 shows the variation of the enhancement factor along the height of the column, which increases rapidly in the bottom half of the column. Fig. 5 shows the comparison between the predicted and measured solution temperature along the column height. Although the predicted temperature profile is in good qualitative agreement with measurement, there are discrepancies in the bottom section of the column where most CO2 absorption occurs. The solvent temperature near the bottom exit is overpredicted in spite of the accurate prediction of the MEA concentration (see Fig. 3(b)). This may be due to the neglect

Liquid Temperature (ºC)

60.00

4.2.

40.00 30.00 20.00 10.00

3.00

4.00

5.00

6.00

4.00

5.00

6.00

7.00

Sensitivity study

The sensitivity of the predictions to the gas- and liquid-film mass transfer coefficients was assessed by simulating the experimental run #T22 (Tontiwachwuthikul et al., 1992). In the base case predictions shown in Fig. 3, the gas (kg ) and liquid (kol ) film mass transfer coefficients were obtained from Onda et al.’s correlation (Eqs. (5a) and (5b)). The influence of kol on the model predictions was assessed by estimating this parameter using the penetration theory of Higbie (1935), according to which kol is given by:



2.00

3.00

of solvent evaporation in the energy balance. As shown in Fig. 5, the heat effects associated with the CO2 absorption and reaction with MEA cause a rapid variation of the solvent temperature in the bottom half of the column with a maximum temperature at the exit. This is due to a significant amount of CO2 absorption in this region as revealed in Fig. 3. This emphasises that the variation of temperature due to the exothermic nature of the absorption and reaction within the absorption column should not be neglected as the reaction rate as well as the gas solubility and transport properties is a strong function of temperature.

kol

1.00

2.00

Fig. 6 – Comparison between the predicted CO2 concentration profiles obtained using klo from Onda et al. (1968) correlation (–) and penetration theory (- - -) for run #T22. () Expt. data (accuracy ± 2%). Tontiwachwuthikul et al. (1992).

50.00

0.00 0.00

1.00

Distance from the bottom of the column (m)

7.00

Distance from the bottom of the column (m) Fig. 5 – Comparison between the predicted and measured solution temperature profiles along the column height. (–) Prediction and () Expt. data of run #T22. Tontiwachwuthikul et al. (1992).

=2

DAl t

(13)

where t is the contact time. The calculated CO2 concentration profiles using kol obtained from the penetration theory and from the correlation of Onda et al. (1968) are compared in Fig. 6. As can be seen, the predicted concentrations in the bottom half of the absorption column are somewhat different, with the prediction obtained from Onda et al.’s correlation being in close agreement with measurements, while it diminishes in the upper region of the column. It is interesting to note that the sensitivity of the predictions to kol suggests that the diffusive transport of CO2 from the gas–liquid interface to the reaction zone influences the overall absorption rate. A number of empirical correlations are reported in the literature for the estimation of the gas-film mass transfer coef-

1606

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

CO2 Concentration (mol%)

CO2 Concentration (mol%)

25.00

20.00

15.00

10.00

5.00

18.00 16.00

Run #T13

14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

0.00 0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

Distance from the bottom of the column (m)

Fig. 7 – Comparison between the predicted CO2 concentration profiles obtained using kg from Onda et al. (1968) (–) and Bravo et al. (1985) (- - -) correlations for run #T22. () Expt. data (accuracy ± 2%). Tontiwachwuthikul et al. (1992).

CO2 Concentration (mol%)

Distance from the bottom of the column (m)

where Reg is the Reynolds number, Scg is the Schmidt number, and Shg is the Sherwood number. The predicted CO2 concentrations obtained using kg given by Eq. (14) and the correlation (Eq. (5a)) of Onda et al. (1968) are shown in Fig. 7, which reveal that the mass transfer rate in the gas phase is considerably overestimated by the former correlation. Cleary, the liquid-film mass transfer coefficient obtained from the penetration theory and the gas-film coefficient given by Bravo et al.’s correlation result in the CO2 concentration predictions which deviate from the base case simulation results obtained using the correlations of Onda et al. (1968) for both parameters as well as the experimental data. A possible cause for this may lie in the fact that the film mass transfer coefficients in the expressions developed by Onda et al. (1968) are correlated as functions of packing type (e.g., Berl saddles or Raschig rings) as well as their dimensions. The packing type affects the hydrodynamic conditions in the column, in addition to the gas–liquid interfacial area, and hence the mass transfer rate between the two phases. In the penetration theory and the correlation of Bravo et al. (1985), the film coefficients are not expressed as a function of packing geometry and therefore not able to properly account for its effect on the absorption rate. The penetration theory also neglects convective mass transfer due to the liquid motion, which can also influence the absorption rate.

4.3.

15.00 10.00 5.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

Distance from the bottom of the column (m) CO2 Concentration (mol%)

(14)

Run #T15 20.00

0.00 0.00

ficient. One such correlation proposed by Bravo et al. (1985) is given by: 0.333 Shg = 0.0338Re0.8 g Scg

25.00

14.00 12.00

Run #T19

10.00 8.00 6.00 4.00 2.00 0.00 0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

Distance from the bottom of the column (m)

Fig. 8 – Comparison between the predicted and measured CO2 concentration profiles. (–) Prediction and () Expt. data of run #T13, T15 and T19 (accuracy ± 2%). Tontiwachwuthikul et al. (1992). of process conditions if these models are to be used for the process design. The performance of the developed model was examined by simulating three experimental runs (T13, T15 and T19) carried out by Tontiwachwuthikul et al. (1992) by varying the inlet CO2 and MEA concentrations and the solution flow rate (i.e., the gas/liquid ratio) from those used in the base case run #T22 (see Table 1). The predicted CO2 concentration profiles for these runs are compared with the experimental data in Fig. 8. As can be seen, for all three runs good levels of agreement with experimental data are achieved, with the predicted trends similar to that observed for run #T22 in Fig. 3.

Simulation of the effects of process parameters

In post-combustion CO2 capture plants, the process conditions such as the flue gas and solution flow rates (and hence the gas/liquid ratio) and inlet CO2 and solvent concentrations can vary depending on the type of fuel burnt and the load on the boilers. Thus, the absorption columns must be designed and operated to handle the variations in the inlet conditions without adversely affecting the CO2 capture efficiency. It is therefore important that the absorption process models must demonstrate their ability to simulate accurately a wide range

4.4. Simulation of an absorber with structured packing Structured packings are used to provide lower gas phase pressure drops than random packings while maintaining high mass transfer rates, which can be advantageous for the postcombustion CO2 capture applications. Although these types of packings are used in a number of applications, their industrial use for CO2 absorption has not yet been reported in the open literature (Aroonwilas et al., 2003). In contrast with randomly

1607

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

1.80

CO2 Concentration (mol%)

CO2 Concentration (mol%)

16.00 14.00 12.00 10.00 8.00 6.00 4.00

(a)

1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20

2.00 0.00 0.00

0.50

1.00

1.50

2.00

0.00 0.00

2.50

2.00

packed absorption columns, modelling of gas absorption in structured packings has received less attention (Aroonwilas et al., 2003). The present absorption model was used to simulate an experiment carried out by Aroonwilas et al. (2001) in a column of 2.21 m packing height and 0.1 m internal diameter packed with stainless-steel Gempak 4A structured packing. The experimental conditions used in the simulation are listed in Table 1. Fig. 9 shows the comparison between the predicted and measured CO2 concentrations along the absorption column. The predictions are in very good agreement with experimental data. It is interesting to note that the level of agreement between the predictions and measurements shown in Fig. 9 is similar to that obtained using a more rigorous absorption model incorporating a mechanistic model of liquid distribution over the packings by Aroonwilas et al. (2003).

4.5.

Simulation of an industrial absorber

Finally, the model was applied to simulate an industrial absorption column of an acid gas treatment unit in an olefin plant for which data was reported by Pintola et al. (1993). The column packing consisted of three sections with heights of 4.6, 4.6 and 4.9 m. The absorber operating conditions for the trial run (I06) simulated in this study are given in Table 1. As can be seen from the table, the gas and liquid flow rates are much larger and MEA concentration is higher (∼4 kmol/m3 ) in the industrial absorber than those used in the pilot-plant by Tontiwachwuthikul et al. (1992). Fig. 10 shows the predicted CO2 and MEA concentration profiles together with the measured respective concentrations at the top and the bottom exit of the column. Unfortunately, concentration data along the height of the column is not available for detail validation of the predictions. However, it is evident from the figure that there is good agreement between the predicted and measured exit concentrations. It should also be noted that the simulated concentration profiles are very similar to those obtained for the pilot-plant runs. The predicted CO2 concentration profile reveals that over 98% of CO2 is absorbed in the bottom two packed beds of the column, which represents about 67.4% of the total packing height, while the top bed reduces the CO2 concentration to a ppm level.

4200.00

MEA Concentration (mol/m3)

Fig. 9 – Comparison between the predicted and measured CO2 concentration profiles in an absorption column with structured packing. (–) Prediction and () Expt. data (accuracy ± 2%). Aroonwilas et al. (2001).

4.00

6.00

8.00

10.00 12.00 14.00 16.00

Distance from the bottom of the column (m)

Distance from the bottom of the column (m)

(b)

4000.00 3800.00 3600.00 3400.00 3200.00 3000.00 0.00

2.00

4.00

6.00

8.00

10.00 12.00 14.00 16.00

Distance from the bottom of the column (m) Fig. 10 – The predicted (a) CO2 and (b) MEA concentration profiles. (–) Prediction and measured exit concentrations () in an industrial absorption column. Pintola et al. (1993).

5.

Conclusions

A rate-based absorption model has been developed for the simulation of post-combustion amine based CO2 capture processes in packed-bed absorption columns. The model predictions are validated against wide ranging experimental data of gas phase CO2 and liquid phase MEA concentrations and the solution temperature reported in the literature. Good agreement is obtained between the predictions and measurements in pilot and industrial scale CO2 absorption columns for a range of operating conditions. The predicted concentration profiles along the column height, in agreement with the measurement, show that most CO2 removal typically occurs in the bottom region of the absorber. The top region merely reduces the CO2 concentration down to a ppm level. The sensitivity studies have revealed that amongst the various correlations for the gas- and liquid-film mass transfer coefficients, the ones given by Onda et al. (1968) provides better predictions. Further validation of the rate-based model using data from industrial scale columns is need before they can be used reliably for the design and simulation of absorption columns for CO2 capture from power plant flue gases.

References Akanksha, Pant, K.K., Srivastava, V.K., 2007. Carbon dioxide absorption into monoethanolamine in a continuous film contactor. Chemical Engineering Journal 133, 229–237.

1608

chemical engineering research and design 8 9 ( 2 0 1 1 ) 1600–1608

Aroonwilas, A., Tontiwachwuthikul, P., Chakma, A., 2001. Effects of operating and design parameters on CO2 absorption in columns with structured packings. Separation and Purification Technology 24, 403–411. Aroonwilas, A., Chakma, A., Tontiwachwuthikul, P., Veawab, A., 2003. Mathematical modelling of mass transfer and hydrodynamics in CO2 absorbers packed with structure packings. Chemical Engineering Science 58, 4039–4053. Astarita, G., 1967. Mass Transfer with Chemical Reactions. Elsevier, Amsterdam. Austgen, D.M., Rochelle, G.T., Peng, X., Chen, C., 1989. Model of vapor–liquid equlibria for aqueous acid gas–alkanolamine systems using the electrolyte-NRTL equation. Industrial & Engineering Chemistry Research 28, 1060–1073. Bravo, J.L., Rocha, J.A., Fair, J.R., 1985. Mass transfer in gauze packings. Hydrocarbon Processing 64, 91–95. Danckwerts, P.V., 1970. Gas–Liquid Reactions. McGraw-Hill, New York. DeCoursey, W.J., 1982. Enhancement factors for gas absorption with reversible reaction. Chemical Engineering Science 37, 1483–1489. Doraiswamy, L.K., 2001. Organic Synthesis Engineering. Oxford Press. Freguia, S., Rochelle, G.T., 2003. Modeling of CO2 capture by aqueous monoethanolamine. Journal of American Institute of Chemical Engineers 49 (7), 1676–1686. Freund, P., 2003. Making deep reductions in CO2 emissions from coal-fired power plant using capture and storage of CO2 . Proceedings of Institute of Mechanical Engineers, Part A: J. Power and Energy 217, 1–7. Higbie, R., 1935. The rate of absorption of a pure gas into a still liquid during short periods of exposure. Transactions of American Institute of Chemical Engineers 31, 365–389. Hikita, H., Asal, S., Ishikawa, H., Honda, M., 1977. The kinetics of reactions of CO2 with MEA, DEA and TEA by a rapid mixing method. Chemical Engineering Journal 13, 7–12. Idem, R., Tontiwachwuthikul, P., 2006. Preface for the special issue on the capture of carbon dioxide from industrial sources: technological developments and future opportunities. Industrial & Engineering Chemistry Fundamentals 45 (8), 2413. Kenig, E.Y., Schneider, R., Gorak, A., 2001. Reactive absorption: optimal process design via optimal modelling. Chemical Engineering Science 56 (2), 343–350. Lawal, A., Wang, M., Stephenson, P., Yeung, H., 2009. Dynamic modelling of CO2 absorption for post combustion capture in coal-fired power plants. Fuel 88, 2455–2462.

Levenspiel, O., 1999. Chemical Reaction Engineering, 3rd ed. John Wiley & Sons. Onda, K., Takeuchi, H., Okumoto, Y., 1968. Mass transfer coefficients between gas and liquid phases in packed column. Journal of Chemical Engineering of Japan 1 (1), 56–61. Pandya, J.D., 1983. Adiabatic gas absorption and stripping with chemical reaction in packed towers. Chemical Engineering Communications 19 (4), 343–361. Pintola, T., Tontiwachwuthikul, P., Meisen, A., 1993. Simulation of pilot plant and industrial CO2 –MEA absorbers. Gas Separation & Purification 7 (1), 47–52. Raynal, L., Royon-Lebeaud, A., 2007. A multi-scale approach for CFD calculations of gas–liquid flow within large size column equipped with structured packing. Chemical Engineering Science 62, 7196–7204. Tobiesen, F.A., Svendsen, H.F., Juliussen, O., 2007. Experimental validation of a rigorous absorber model for CO2 postcombustion capture. Journal of American Institute of Chemical Engineers 53 (4), 846–865. Tontiwachwuthikul, P., Meisen, A., Lim, J., 1992. CO2 absorption by NaOH, monoethanolamine and 2-amino-2-methyl-1-propanol solutions in a packed column. Chemical Engineering Science 47 (2), 381–390. Treybal, E., 1969. Adiabatic gas absorption and stripping in packed towers. Industrial & Engineering Chemistry 61 (7), 36–41. van Krevelen, D.W., Hoftijzer, P.J., 1948. Kinetics of gas–liquid reactions: part 1. General theory. Recueil des Travaux Chimiques des Pays-Bas 67, 563–586. Versteeg, G.F., van Swaaij, W.P., 1988. Solubility and diffusivity of acid gases (CO2 , N2 O) in aqueous alkanolamine solutions. Journal of Chemical & Engineering Data 33, 29–34. Weiland, R.H., Dingman, J.C., Cronin, D.B., Browning, G.J., 1998. Density and viscosity of some partially carbonated aqueous alkanolamine solutions and their blends. Journal of Chemical & Engineering Data 43, 378–382. Wellek, R.M., Brunson, R.J., Law, F.H., 1978. Enhancement factors for gas absorption with second-order irreversible chemical reaction. Canadian Journal of Chemical Engineering 56, 181–186. Whitman, W.G., 1923. The two-film theory of gas absorption. Chemical and Metallurgical Engineering 29, 146–148. Wilson, E.J., Gerard, D., 2007. Carbon Capture and Sequestration: Integrating Technology, Monitoring and Regulation. Blackwell Publications, Oxford. WRI, 2007. Climate Analysis Indicators Tool., http://www.wri.org/.