Improved CO2 Absorption in a GasâLiquid Countercurrent Column

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Improved CO2 Absorption in a Gas-Liquid Counter Current Column using a Ceramic Foam Contactor Zhen Wang, Mayank Gupta, Sumedh S. Warudkar, Kenneth R. Cox, George J Hirasaki, and Michael S. Wong Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.5b03600 • Publication Date (Web): 15 Jan 2016 Downloaded from http://pubs.acs.org on January 18, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Improved CO2 Absorption in a Gas-Liquid Counter Current Column using a Ceramic Foam Contactor Zhen Wanga, Mayank Guptaa, Sumedh S.Warudkara+, Kenneth R. Coxa, George J. Hirasakia, Michael S.Wonga,b,c,d* a Department of Chemical and Biomolecular Engineering, Rice University, 6100 Main St, MS 362, Houston, TX 77005, USA b Department of Chemistry, Rice University, 6100 Main St, MS 60, Houston, TX 77005, USA c Department of Civil and Environmental Engineering, Rice University, 6100 Main St, Houston, TX 77005, USA d Department of Materials Science and NanoEngineering, Rice University, 6100 Main St,, Houston, TX 77005, USA + Author Present Address: Shell Global Solution (US), Inc. * Corresponding Author E-mail address:[email protected] Tel: 1-713-348-3511 Fax: 1-713-348-5478 Abstract Solid foams are porous, monolithic materials with higher specific surface areas compared to random packings that are commonly used in amine-based CO2 capture processes. In this work, the hydrodynamic characteristics (e.g., pressure drop, flooding point and liquid holdup) and CO2 absorption performance of α-Al2O3 ceramic foam packings of different porosities were investigated experimentally in a gas-liquid counter-current column. With a 30 wt% diglycolamine (DGA) solvent as the CO2 absorbent, the foams allowed higher flow rates of gas and liquid compared to a random packing before undesirable flooding is reached. Ceramic foams with lower porosity have larger operating capacities than those with higher porosity. A parametric study of a one-dimensional flow model was performed by investigating the effects of gas velocity, liquid velocity, and CO2 solvent loading on CO2 removal performance. Lower gas velocities and higher liquid velocities increased CO2 removal efficiency. CO2 removal efficiency decreased as initial CO2 loading increases. Initial CO2 loading of DGA solutions is recommended to be less than 0.35 mol CO2/mol DGA to provide efficient CO2 removal. Ceramic foams improve CO2 absorption using liquid amines, which can lead to smaller carbon capture units. Keywords: Carbon dioxide; Absorption; Ceramic foam; Packed column; Diglycolamine; 1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 42

Hydrodynamics

2


Page 3 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction The excessive emission of carbon dioxide (CO2), resulting from combustion of fossil fuels (e.g., coal-fired power plants), contributes significantly to global warming1. 60% of global CO2 emission has been proven to be produced by the power and industrial sectors2, and its reduction from the coal-fired power plants represents a critical effort to mitigate climate change. Among various CO2 capture approaches, solvent-based post-combustion capture (PCC) technologies are believed to have the highest potential for commercial success due to their successful use in the chemical process industry to sweeten acid gases such as CO2 and SOX 3. Such amine scrubbers are effective for gas streams of low CO2 partial pressures, and they can be retrofitted to existing power plants4. The world’s first commercial-scale carbon capture and storage (CCS) facility based on solvent-based PCC technology began its operations in Canada (at the Boundary Dam power station, Saskatchewan) in October 20145. A typical solvent-based CO2 separation process consists of absorbing CO2 from the gas phase into the aqueous solvent phase in the absorber unit, and releasing CO2 from the spent solvent in the stripper unit. The absorber unit is generally operated at slightly above atmospheric pressure, and the stripper unit is recommended to be operated at pressures ranging from 75 kPa to 300 kPa6, 7. In this process, reaction and mass transfer are regarded as the two most important factors. Alkanolamines such as monoethanolamine (MEA), diglycolamine (DGA) and diethanolamine (DEA) are the most extensively used solvents for CO2 capture8. Amines with higher CO2 absorption reaction rates do not necessarily lead to increased CO2 separation efficiency, as mass transfer rates can be limiting if there is inefficient contact between the flue gas and liquid absorbent. The type of tower packing can 3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

increase the rate of mass transfer by increasing the gas-liquid contact area and by enhancing the turbulent mixing of the fluid phases. For gas-liquid counter-current flow configurations, some structured packings, e.g., monoliths, wire gauze packings, and Sulzer MellapakTM, have been shown to improve hydrodynamic performance compared to random or irregular packings (of Raschig rings or Pall rings, for example), in terms of pressure drop, gas-liquid contact and flow distribution9, 10. Solid foam materials may be more advantageous as a type of structured packing due to their open-cell structure with very high porosities (up to 90%), high surface area per unit volume, and low pressure drop in the system11. So far, solid foam packings for CO2 absorption have not been reported. However, there are several reports on hydrodynamic and mass transport behavior of commercial solid foams. Stemmet et al. investigated the use of metal foams as packing and studied liquid holdup in countercurrent flow experimentally and through modeling9, 12-14. Laveque et al. performed measurements of liquid holdup and mass transfer in distillation column15. Grosse et al. investigated the hydrodynamic properties of ceramic foams with a range of porosities and pore sizes, in which they showed the pressure drop measured through solid foams is always small compared to that through conventional random packings16. Solid foams are commercially available in materials such as aluminum, silicon carbide, and aluminum oxide9, 17-19. In this study, we studied CO2 absorption performance of solid foam packings using a liquid amine, choosing α-Al2O3 ceramic foams as a model material.

4


Page 4 of 42

Page 5 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The liquid amine absorbent was a 30 wt% DGA solution in water. The hydrodynamic characteristics (e.g., pressure drop, flooding point and liquid holdup) of α-Al2O3 ceramic foams with different pores-per-inch (PPI) numbers were studied in gas-liquid counter current operation. Besides, we developed a one-dimensional (1-D) mathematical model and numerically simulated CO2 absorption through a foam column. After validating the model with the experimental results, we performed a parametric study on the effects of gas velocity, liquid velocity, and solvent CO2 loading on CO2 removal efficiency, and the influence of liquid-to-gas ratio (L/G) on temperature distribution within the column.

2. Experimental 2.1 Materials Water solutions of diglycolamine (DGA, C4H11NO2, 30% by weight) solution were prepared using DGA from Huntsman Chemicals (99.5%). CO2 and N2 gases were supplied by Matheson TriGas with purities of >99.9 mol%. The solid foams (commercially available from ASK-Chemicals, Alfred, NY) were cylindrical pieces of α-Al2O3 (99.5 wt%) with a diameter of 2.54 cm and a length of 5.1 cm (Fig. 1). Three different grades of ceramic foams were used: 20, 30 and 45 PPI (number of pores per inch). Table 1 summarizes their structural parameters in comparison to other packings.

5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 42

Figure 1 Endview of the α-Al2O3 ceramic foam.

Table 1. Structural parameters for different packings

Packing Type

Porosity (%)

Specific surface area (m2/m3)

Bulk densitya (g/cm3)

Equivalent pore diameter (mm)

Permeability

20-PPI

85

700

0.60

1.28

8.0×10

30-PPI

85

900

0.65

1.00

7.3×10

45-PPI

84

1400

0.71

0.60

6.2×10

Raschig Ring

62.6

239

0.58

1.5

3.87×10-8

Pall Ring

94.2

232

0.48

2.5

3.53×10-7

Structure

b

(m2) -9

α-Al2O3 Ceramic Foam

Random Packing20

-9

-9

(a) Data are available at www.ask-chemicals.com and http://www.tower-packing.com; (b) Permeability of packing was calculated using k =

3εde2 , the parameters and units of this equation can 50

be found in the Nomenclature section.

6


Page 7 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


2.2 Experimental apparatus and procedure 2.2.1 Hydrodynamic behavior experiments A counter-current, single-column ceramic foam unit was constructed from a glass tube with an inner diameter (ID) of 2.8 cm (Fig. 2). Six foam pieces of the same grade were wrapped tightly, end-to-end in heat-shrunk Teflon tubing to prevent the bypass of gas or liquid on the side of ceramic foam plugs. O-rings were used to provide a liquid-tight seal between the glass column and the ceramic foam unit to prevent liquid bypass. N2 gas was introduced from the bottom of the ceramic foam column. In order to prevent any escape of the entering gas with the exiting liquid, a plastic graduated cylinder was added to the experimental setup to provide a liquid seal. The liquid seal works by providing a static head of water, which acts as a back pressure regulator and prevents the escape of inlet gas through the liquid outlet port.

Figure 2 Experimental setup for measuring pressure drop across ceramic foam, left: schematic diagram of experimental setup, right: photograph of experimental setup. The pressure drop across the column was measured using a water manometer built in-house. Water was delivered to the column using a peristaltic pump (FPU500, FPUMT, Omega Engineering). Liquid was distributed on the ceramic foam with a piece of stainless

7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

steel wool at the top of the ceramic foam column. N2 flow was controlled using a mass flow controller (FMA6528ST, Omega Engineering). Before experiments, the pressure drops of external sources (e.g., tubing and fittings) were measured by conducting measurements on the empty glass tube. When performing the experiments, the gas flux was held constant at a specific value and the liquid flux was varied. For each flux, it was ensured that the system was operationally stable for at least 20 min. Pressure drops in the column corresponding to these operating conditions were recorded after the water levels in the manometer reached steady state.

2.2.2 CO2 absorption experiments The ceramic foam unit was modified for CO2 absorption, in which the water manometer was removed (Fig. 3). The length of ceramic foam packings was changed by using different numbers of foam pieces. A CO2/N2 gas mixture (13% v/v CO2) injected at the bottom of the unit was used as the flue gas stimulant. The 30 wt% DGA solvent (~3 M DGA) was delivered at the column top from the solvent reservoir using a peristaltic pump. Spent absorbent was collected in a carboy for discarding. The outlet gas was analyzed for its CO2 concentration by a non-dispersive infrared detector (LI 820, LI-COR Biosciences; dual wavelength infrared detection = 3.95 and 4.26 µm).

8


Page 8 of 42

Page 9 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3 Schematic diagram of experimental setup for CO2 absorption in ceramic foam

3. Chemistry of CO2 absorption by DGA 3.1. Reaction rates DGA, as a primary amine, is widely used for acid gas treatment. The kinetics for the reaction of CO2 and DGA has previously been studied extensively21-24. The following two reversible reactions are usually used to represent the reactions between CO2 and DGA solution: CO 2 + H 2 O+ DGA ⇔ DGACOO - + H 3 O +

(1)

CO 2 + OH - ⇔ HCO 3 -

(2)

The contribution of reaction (2) is usually ignored due to the very low OH- concentration (10-5-10-4 M) relative to DGA concentration (~3 M), and the CO2 reaction rate is assumed to follow the following elementary rate expression:

rCO2 = kDGA[CO2 ][DGA]- kDGA / KDGA,K [DGACOO- ][H3O+ ] 9


(3)


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 42

or rCO 2 = k DGA [DGA]([CO 2 ] − [CO 2 ]eq )

(4)

where [CO 2 ]eq is the equilibrium concentration of CO2 in solvent. The rate constant kDGA can be expressed as kDGA (m3kmol−1s−1 ) = k DGA(298K)exp(−

Ea 1 1 ( − )) R T 298.15

(5)

in which kDGA(298K) = 6.66 ×103 m3kmol-1s-1 ; Ea=40.1 kJ/mol, and T is the absolute temperature, K [21]. 3.2. Equilibrium concentrations Five reactions were used to describe the equilibrium concentrations of all chemical species in the DGA-H2O-CO2 system25: KW 2H 2 O ← → H 3 O + +OH -

(6a)

K DGAH+ DGAH + +H 2 O ←  → DGA+H 3 O +

(6b)

K

DGACOO DGACOO− +H2O ← → DGA+HCO3−

(6c)

K CO 2 CO 2 +2H 2 O ← → H 3 O + +HCO 3−

(6d)

K

−

−

HCO3 HCO3− +H2O ← →H3O+ +CO32−

(6e)

The chemical equilibrium constants, which are a strong function of temperature T, can be expressed as: ln K eq = A + B / T + C ln(T ) + DT

(7)

where Keq is the equilibrium constant for equations 6a-e; A, B, C and D are adjustable parameters, which are available in Table 2. The relationship between gaseous CO2 and CO2 in aqueous DGA-H2O-CO2 system can be expressed by Henry’s Law, 10


Page 11 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


PCO2 = H CO 2 , L × [CO 2 ]

(8)

where PCO2 is CO2 partial pressure; HCO2 ,L is the Henry’s law constant of CO2 in aqueous DGA solution; and [CO2] is CO2 concentration in liquid phase. Calculations of HCO2 ,L at different temperatures are detailed in the Appendix.

Table 2. Parameters A, B, C and D for equilibrium constants in Eq. 7 Reaction

A

B

C

D

Reference

6a

132.899

-13445.9

-22.4773

0

26

6b

-13.3373

-4218.708

0

0

27

6c

3.661096

-3696.1689

0

0

27

6d

231.465

-12092.1

-36.7816

0

28

6e

216.05

-12431.7

-35.4819

0

26

4. Numerical model development The ceramic foam column was modeled as a 1-D counter-current flow CO2 absorption process (Fig. 4). The following assumptions were applied: (1) steady state operation; (2) the solvent was an incompressible and Newtonian fluid; (3) negligible vapor pressure of water and DGA; (4) uniform liquid velocity and gas velocity fields throughout column; and (5) Henry’s law to describe the relation between the solution and gas-phase concentrations for CO2 in the gas-liquid interface.

11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 42

Figure 4 Schematic of CO2 absorption process in the ceramic foam. 4.1. Momentum equations For stationary, creeping and incompressible flow, the inertial terms can be neglected due to low Reynolds number. Therefore, the momentum balance equations for the gas and liquid flows can be simplified into Darcy’s Law: (9)

(10)

Under the assumption of having a uniform velocity throughout the ceramic foam column and the liquid saturation being uniformly distributed, the pressure gradients of respective gas phase and liquid phase are equal. Therefore,

dPG dPL = dz dz

(11)

Within the ceramic foam, the pressure difference between the gas and liquid phases (also 12


Page 13 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


called gas-water capillary pressure Pc) is a function of liquid saturation (SL),

PG − PL = Pc ( S L )

(12)

in which, subscripts G or L represents gas or liquid phase, respectively; ρ is the density; g is gravitational acceleration; µ is the viscosity; P is pressure, k is the permeability. k rw and krg are the relative permeability of liquid phase and gas phase, respectively. SL is liquid saturation, which presents the fraction of the pore space occupied by liquid in the porous medium. The relative permeability is assumed to be a function of the saturation of each phase29. Therefore, gas phase relative permeability (krg) and liquid phase relative permeability (krw) can be expressed in terms of gas saturation (SG) and liquid saturation (SL) in the following way30:

krg = SGn1 = (1− SL )n1 k rw

(13)

S L − S L0 n 2 =( ) 1 − S L0

(14)

The parameters n1 and n2 for different types of ceramic foams with different PPI number are reported by Stemmet et al.9. In equation (14), SL0 is the liquid saturation at the static holdup condition and is correlated by Saez and Carbonell30, S L0 = [ε (20 + 0.9 Eo * )]−1

(15)

in which, the dimensionless parameter Eo* is defined as, Eo* =

ρ gd e 2ε σ (1 − ε ) 2

(16)

where ε is the porosity of ceramic foam; g is gravity acceleration; de is equivalent diameter of foam; σ is surface tension of liquid. The liquid saturation can be estimated by the capillary pressure based on the Thomeer 13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 42

model31,

J < J e = C1 1.0,    S L − S L0   =  C2  S Li − S L0 1.0 − exp   , J > J e = C1  ln  C1         J  

(17)

SLi is initial liquid saturation, which is equal to 1 in this study. S L0 is the residual liquid saturation, which is defined here as the asymptotic saturation where the capillary pressure approaches infinity. J is the Leverett J function32, which is a dimensionless function that accounts for capillary pressure as a function of liquid saturation, J =

Pc

k

σ

ε

(18)

where Pc is capillary pressure, σ is surface tension, k is permeability, ε is porosity. 4.2. Mass transport equations Mass balance of species CO2 in gas phase and liquid phase can be respectively expressed as dCCO2 ,G d ( − DCO2 ,G + CCO2 ,GU G ) = RCO2 ,G dz dz

(19)

dCCO2 ,L d (−DCO2 ,L + CCO2 ,LUL ) = RCO2 , L dz dz

(20)

in which DCO

and DCO

2 ,G

respectively; C CO

2 ,G

are CO2 diffusion coefficient in gas phase and liquid phase,

2 ,L

and C CO

2

,L

are CO2 concentration in gas phase and liquid phase,

respectively. The boundary conditions for these equations are as follows, 0 at z=0, CCO2 ,G = C CO , 2 ,G

at z=L,

dCCO2 ,G dz

dCCO2 , L dz

=0

0 = 0 , CCO2 ,L = C CO 2 ,L

Mass balance of specie DGA in liquid phase can be expressed as,

14


Page 15 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


dC d (− DDGA,L DGA,L + CDGA,LU L ) = RDGA, L dz dz

(21)

subject to the following initial and boundary conditions at z=0,

dCDGA, L dz

=0

0 at z=L, CDGA,L = C DGA ,L

in which DDGA,L is DGA molecular diffusion coefficient in liquid phase; C DGA ,L is DGA concentration in liquid phase. The source term RCO

2 ,G

for equation (19) is the net mass transfer rate of CO2 into gas

phase, which can be expressed as,

RCO2 ,G = − Kov aeff [

CCO2 ,G H

The source term R CO

2

,L

− CCO2 ,L ]

(22)

for equation (20) represents the net CO2 generation rate in the

liquid phase, which is the combination of CO2 mass transfer and CO2 reaction,

RCO 2, L = K ov aeff [

CCO 2,G H

− CCO 2, L ] − rCO 2

(23)

The source term R DGA,L for equation (21) is the net reaction rate of DGA absorbent in liquid phase, which can be expressed as,

RDGA, L = −2rCO 2

(24)

In equations (22) and (23), K ov is overall mass transfer coefficient for CO2, which can be described as the combination of liquid phase mass transfer coefficient (KL) and gas phase mass transfer coefficient (KG),

1 1 1 = + K ov K G K L

(25)

Gas phase mass transfer coefficient (KG) in packing column was correlated as33

ρU d KG de µG = 0.054( )0.333 ( g G e )0.8 ρ g DCO2 ,G µG (1 − ε ) DCO2 ,G

(26)

in which de is equivalent pore diameter; µG is gas viscosity; and ε is porosity. 15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 42

Liquid phase mass transfer coefficient (KL) can be calculated by the following equation,

K L = EK L,0 = K L,0 1 + Ha 2 = K L,0 1 +

kDGACDGA, L DCO2, L K L ,0 2

(27) in which K L ,0 is the physical mass transfer coefficient in liquid phase, E is the chemical enhancement factor, E = 1 + Ha2 ; Ha is Hatta number, Ha =

k DGACDGA, L DCO 2, L K L ,0 2

kDGA is

reaction rate constant, which has been defined in Eq. 5. For laminar liquid flow, the effective interfacial gas-liquid area per unit volume ( a eff ) can be correlated as20,

aeff = 6.49a 0.333

( ρ L − ρ g )0.5 g1/6υ L1/3

σ L1/2

U L1/3

(28)

in which a is geometric surface are of ceramic foam per unit volume, υ L is kinematic viscosity of solvent, and σ L is surface tension of liquid. 4.3. Energy equations The differential equations for heat transfer in steady-state in ceramic foam can be described as For liquid phase:

ρLCp, LU L

dTL d dT = (λL L ) + haeff (TG − TL ) + RCO 2,L ∆H R dz dz dz

(29)

The boundary conditions imposed on the thermal energy equation for liquid phase are as follows: at z=0,

dTL =0 dz

at z=L, TL = TL0 For gas phase:

16


Page 17 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ρGC p,GUG

dTG d dT = (λG G ) + haeff (TL − TG ) dz dz dz

(30)

The boundary conditions imposed on Eq. 30 are as follow at z=0, TG = TG0 at z=L,

dTG =0 dz

Since CO2 physical absorption heat is much smaller compared to CO2 reaction heat, the last term in Eq. 29 only refers to the heat generation due to the chemical reactions that occur in the solvent phase. The C p , L and C p ,G are specific heat capacities for liquid and gas, respectively; λL and λG are the thermal conductivities for liquid and gas, respectively;

∆H R is enthalpy change of CO2 reaction with DGA; and h is heat transfer coefficient, which can be approximately calculated by following correlation34:

h = 1.195GC p ,G [

deG ]−0.36 PrG −0.667 µG G (1 − ε ')

(31)

where G is the gas mass flow rate, kg/(m2s); ε ' is operating void space fraction in the 3 3 packing, mvoid /mpacked volume ; PrG is Prandtl number, and PrG =

µGCp,G . λG

4.4. Numerical solution The partial differential equations associated with appropriate boundary conditions, physical and chemical properties, and reaction rate equations were numerically solved via the finite element method using the commercial software COMSOL. The detailed calculations for some physical and chemical properties needed to solve the above model equations are presented in the Appendix. They include mass transfer coefficients in gas and liquid phase, Henry’s Law constants, viscosities, and diffusion coefficients. 17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

5. Results and discussions 5.1. Hydrodynamic behavior study 5.1.1. Liquid holdup and flow regime prediction Fig. 5 presents the change of liquid holdup and flow regimes at different gas and liquid fluxes. The trickle flow regime, i.e., liquid flows down as a laminar film on the foam surface while the gas passes through the remaining void space, only occurs at relatively low liquid holdup. In this regime, gas and liquid have high contact area favorable for mass transfer. With the increase of gas or liquid flow, the liquid holdup increases since the flow of gas opposes the down-flow of liquid in the ceramic foam. As the gas flux or liquid flux increases to a certain value, the trickle flow behavior will begin to change to pulse flow behavior. The trickle-to-pulse flow regime transition point is usually referred to as flooding point.

Figure 5 Liquid holdup and flow regimes for different gas and liquid fluxes. (Dashed and solid lines represent the contour lines of liquid velocity).

18


Page 18 of 42

Page 19 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In bubble flow, gas flows up as discrete bubbles through a continuous liquid phase within the foam packing. In this regime, the liquid holdup is usually very high and gas velocity is relatively low. As the gas flux increases or the liquid flux decreases, these bubbles will become larger and coalesce, eventually spanning the width of packing column and developing to the pulse flow regime. For both bubble flow and pulse flow regimes, the mass transfer characteristics deteriorate due to lowered gas-liquid contact area. Fig. 6 shows calculated liquid holdup values for the three ceramic foams with different porosities at varying gas and liquid flow rates. Under trickle flow, we can find that the liquid holdup increases with increasing liquid flow rate at a fixed gas flow rate. As the gas flow rate increases, it opposes the flow of the liquid in the packing column, giving rise to a higher liquid holdup as well. The highest liquid holdup that can be obtained is equal to the porosity of ceramic foam as gas flow rate reduces to zero. The trickle flow regime falls below the loci of points.

19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 6 Modelling results of the liquid holdup versus superficial gas velocity:(a) 20-PPI ceramic foam (b) 30-PPI ceramic foam (c) 45-PPI ceramic foam. (Dashed lines are constant liquid velocities, and solid lines are flooding lines) Operating conditions: packing height: 30.5 cm; liquid phase: water; gas phase: N2; 25 °C. By comparing with Fig. 6 (a)-(c), we can further find the difference of flooding lines between 20, 30 and 45 PPI ceramic foams. Ceramic foam with a higher PPI number presents a higher risk to reach flooding than ceramic foam with lower PPI. For example, in case of 20-PPI ceramic foam in Fig.6 (a), as the liquid flow rate is 0.0001 m/s, the flooding occurs as gas flow rate increases to 1.6 m/s. However, for 45-PPI ceramic foam in Fig. 6(c), the flooding occurs as gas flow rate only exceed 1.125 m/s. Fig. 7 further gives a clear comparison of flooding points for different ceramic foam 20


Page 20 of 42

Page 21 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


packings with gas and liquid flow velocities. In a packed column, the gas and liquid flow rates are limited by the tendency of the packing to the liquid flooding. At the flooding point, the pressure drop rises sharply and much of the liquid is carried off mechanically by the gas leaving the top of the packing. As the PPI number of ceramic foam increases, flooding is encountered at lower liquid and gas flow rates. The reason is that the higher PPI ceramic foam has more tortuous flow channels and smaller pore sizes, increasing the restriction to gas and liquid flow. Therefore, the ceramic foam with lower PPI number may have larger operating flexibility than ceramic foam with higher PPI number.

Figure 7 Modelling results of gas and liquid velocities at the flooding point for 20, 30 and 45 PPI ceramic foams Operating conditions:packing height: 30.5 cm; liquid phase: water; gas phase: N2; temperature: 25 °C.

We further investigated the flooding characteristics of the ceramic foam by comparing with other conventional packings via generalized pressure drop correlation (GPDC) charts, as shown in Fig. 8. GPDC chart describes the balance between the vapor momentum and gravity forces acting on the liquid droplets, which is commonly used to predict packing flooding and 21



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 42

pressure drops in industry35. The abscissa and ordinate of the GPDC chart are the flow parameter and capacity factor, respectively36:

flow parameter = (

capacity factor =

L ρG 0.5 )( ) G ρL

(32)

G 2 Fpξ 2 µ L0.2

(33)

ρG ρ L g

where Fp is the wet packing factor; and ξ is the ratio between the density of water and the density of liquid. The flow parameter represents the ratio of liquid kinetic energy to vapor kinetic energy, and the capacity factor describes the balance between the up-flowing vapor momentum force, that acts to entrain swarms of liquid droplets, and the gravity force, that resists the upward entrainment. As can be seen, ceramic foam performed better than the random packing in terms of avoiding liquid flooding, with twice the capacity factor limit of the latter. However, the ceramic foam had a lower capacity factor limit than the structured packing. Thus, ceramic foam has intermediate hydrodynamic performance based on GPDC studies.

22


Page 23 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8 Comparison of different packings in generalized pressure drop correlation (GPDC) chart.

5.1.2. Pressure drop prediction of ceramic foams with different PPI number The pressure drop measurements were performed using N2-water system at room temperature in the hydrodynamic test setup as shown in Fig. 2. In the measurement, 20, 30 and 45 PPI ceramic foam packings were used (Table 1). The gas and liquid flow were controlled in the trickle flow regime. Fig. 9(a) shows the predicted and experimental pressure drops in 20, 30 and 45 PPI ceramic foams as a function of gas flow rate. The predicted pressure drops of ceramic foams had fair agreement with measured pressure drops, considering there were no adjustable parameters in the pressure drop model. Pressure drops increased with faster gas flow velocities or with higher PPI numbers (or lower permeabilities) of the foam. Fig. 9(b) shows a similar comparison of the predicted and experimental pressure drops at different liquid and gas flow rates in a 20 PPI ceramic foam column. Higher liquid flow rates increase the pressure drops 23



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of the ceramic foam, which can be explained by the decrease of the gas phase relative permeability (krg) with faster liquid flow velocities. Higher liquid flow rates result in the increase of liquid saturation (SL) and decrease of gas saturation (SG), which will decrease the gas phase relative permeability in the ceramic foam, as shown in Equation (13) and (14).

Figure 9 Predicted and experimental the pressure drops in ceramic foams as a function of gas flow rate: (a) ceramic foam with different PPI number, liquid flow rate: 50 mL/min; (b) 20-PPI ceramic foam at different liquid flow rates. Operating conditions: Packing height: 30.5 cm; liquid phase: water; gas phase: N2; temperature: 25 °C. 24


Page 24 of 42

Page 25 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


5.2. CO2 absorption in ceramic foam column 5.2.1. Model validation CO2 absorption using 30 wt% DGA was conducted experimentally in a counter-current column packed with a 20-PPI ceramic foam. Fig. 10 shows a comparison of experimentally determined and simulated CO2 removal efficiencies at various gas and liquid flow rates. The predictions agreed well with the experimental results as the height of ceramic foam column is 10.2 cm in Fig.10(a). CO2 removal efficiency increased with higher ceramic foam height (Fig. 10(b)). When the packing height doubled to 20.4 cm, a general good accordance between simulation and experiments was seen, with the exception of CO2 removal efficiencies at the lowest gas flow rate tested (0.25 SLPM), which were under-predicted.

Figure 10 Simulated and experimental CO2 removal efficiency, as a function of gas flow rate 25



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

at different liquid flow rates: (a) ceramic foam height = 10.2 cm; (b) ceramic foam height = 20.4 cm. Operating conditions: ceramic foam type: 20 PPI; liquid phase: 30 wt.% DGA, gas phase: 13% CO2/87% N2; absorption temperature: 25 °C; lean loading: 0.2 mol CO2/mol DGA. 5.2.2. Parametric study of CO2 absorption in ceramic foam CO2 absorption in ceramic foam column is affected by several operating parameters such as gas/liquid velocity, CO2 loading of absorbent, absorption temperature, inlet CO2 concentration, and absorbent concentration. We found most parameters influenced CO2 absorption quite similarly to absorption via other traditional packings, and so the effects of only the three most important parameters (gas velocity, liquid velocity and CO2 loading) on CO2 absorption via ceramic foam were reported in this study. Fig. 11 shows the simulated concentrations and temperatures profiles along the ceramic foam column. Ordinate describes the dimensionless height of ceramic foam packing. The CO2 loading of lean DGA solvent is 0.2 mol CO2/mol DGA at the top of the column (z=H); gas-phase CO2 concentrations at the bottom of column (z=0) is 13 v/v%. A typical concentration profile of CO2 in gas phase decreasing along the column height can be seen in Fig. 11(a). The CO2 loading of DGA solvent has a similar distribution along the column, as shown in Fig. 11(b). Fig. 11(c) presents the temperature profiles of gas and liquid phases along the column. A maximum temperature (a "temperature bulge ") can be observed near the bottom of the column.

26


Page 26 of 42

Page 27 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 11 Simulation results for CO2 capture using DGA in a ceramic foam column: (a) CO2 concentration in gas phase; (b) liquid-phase CO2 loading; (c) column temperature. Operating conditions:ceramic foam height: 25.4 cm; ceramic foam type: 20 PPI; gas flow velocity: 0.01 m/s; liquid flow velocity: 0.01 cm/s; liquid phase: 30% DGA solvent; gas phase: 13% CO2/87% N2; absorption temperature: 40 °C; lean loading: 0.2 mol CO2/mol DGA. 5.2.2.1.Effects of liquid and gas velocities Fig. 12 plots the CO2 concentration profiles in gas phase along the column’s length under different liquid flow velocities. A decreasing trend of CO2 concentration is found along the column’s length. But, for the cases with high liquid flow rates, the CO2 concentration is decreasing slowly gradually. This phenomenon becomes more significant with slower gas flow rate. This is because near the top end of the column, CO2 concentration is becoming extremely low and close to CO2 equilibrium, which will lead to a very low CO2 absorption driving force. The optimal liquid velocities should be those at which CO2 concentration decreases linearly along the column.

27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 12 Simulated CO2 concentration profiles of gas phase along ceramic foam column for four different liquid velocities at three different gas velocities. Operating conditions: ceramic foam height: 25.4 cm; ceramic foam: 20 PPI; liquid phase: 30% DGA sovlent; gas phase: 13% CO2/87% N2; absorption temperature: 40°C; lean loading: 0.2 mol CO2/mol DGA.

Fig. 13 shows simulated CO2 removal efficiencies as the function of liquid velocity and gas velocity for DGA solvents. CO2 removal efficiency increased as gas flow rate decreased. This can be explained by the change of residence time of CO2 gas in the ceramic foam column as the gas flow rate increases. A higher gas flow rate will result in a shorter residence time, which decreases the reaction time between CO2 and the DGA absorbent. CO2 removal efficiency also increased as the liquid flow rate became faster. This is due to more reactive absorbent to gaseous CO2 at faster liquid velocity. On the other hand, faster liquid flow will also lead to the increase of effective gas-liquid contacting area and the intensified mass 28


Page 28 of 42

Page 29 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


transfer in the liquid phase. However, as the CO2 removal efficiency exceeds 90%, increased liquid flow rate has minor effect on CO2 removal efficiency since CO2 equilibrium is achieved between the gas and liquid phase.

Figure 13 Change of CO2 removal efficiencies under different gas and liquid flow velocities (Inset: gas and liquid velocities yielding >80 % CO2 removal efficiency). Operating conditions: ceramic foam height: 25.4 cm; ceramic foam type: 20 PPI; liquid phase: 30% DGA solvent; gas phase: 13% CO2/87% N2; absorption temperature: 40°C; lean loading: 0.2 mol CO2/mol DGA.

5.2.2.2.Effect of CO2 loading Fig. 14 investigates the change of CO2 removal efficiency for DGA absorbent with different 29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

CO2 loadings at gas flow velocity of 0.01 m/s. CO2 removal efficiency decreased as initial CO2 loading increased. The maximum removal efficiency was achieved when the DGA solvent was free of CO2. As CO2 loading increased from 0 to approximately 0.35 mol CO2/mol DGA, CO2 removal efficiency was slightly decreased. However, as CO2 loadings above 0.35 mol CO2 /mol DGA, CO2 removal efficiency declined sharply with CO2 loading. Therefore, the recommended initial CO2 loading of DGA solution should be less than 0.35 mol CO2/mol DGA. This phenomenon can be explained by two negative effects of CO2 loading: the reduction of free amine, and the increase of CO2 equilibrium pressure in the liquid phase, respectively. Increasing CO2 equilibrium pressure will decrease CO2 driving force. As the CO2 loading is relatively low, the second negative effect is limited, and the decrease of free amine plays a more important role. As CO2 loading increases to a high level, CO2 equilibrium pressure will climb sharply. At this moment, the second negative effect becomes significant, which leads to a rapid drop of CO2 removal efficiency. In Fig. 14, a negative CO2 removal efficiency can even be seen as CO2 loading reached to 0.5 mol CO2/mol DGA. This indicates that CO2 equilibrium pressure is greater than the inlet CO2 gas phase pressure, thus CO2 desorption occurs in this condition.

30


Page 30 of 42

Page 31 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 14 Change of CO2 removal efficiency for DGA solvent with different CO2 loadings. Operating conditions: ceramic foam height: 25.4 cm; ceramic foam type: 20 PPI; liquid phase: 30 wt.% DGA solvent; gas phase: 13% CO2/87% N2; gas flow velocity: 0.01 m/s; absorption temperature: 40 °C. 5.2.2.3.Temperature profile in ceramic foam column Fig. 15 shows the temperature profiles for gas phase and liquid phase at different liquid flow velocities and gas flow velocities. It can be clearly seen that for all cases, a temperature bulge can be found in the column. The gas phase and liquid phase temperature profiles look similar in shape but will be lagged due to the difference in heat capacities of the two phases and the solvent-to-gas ratio (L/G). Varying solvent-to-gas ratio (L/G) will change the magnitude of temperature bulge. Lower L/G will result in a more significant temperature bulge in the column. Fig. 15 also reveals the information about the effect of varying the L/G with respect to the 31



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

temperature bulge location. The primary negative effect of temperature bulge is the reduction of the equilibrium driving force. If there is insufficient solvent related to gas, the greatest absorption rate will occur at the top part of the column. Excess solvent relative to the gas will make the greatest absorption rate near the bottom of the column. The greater heat capacity of the liquid relative to the gas will also tend to push the heat of reaction to the bottom37. Therefore, with the increase of solvent flow rate or the decrease of gas flow rate, the location of the temperature bulge shifts from the top of the column to the bottom.

Figure 15 Temperature profiles for gas and liquid phases at different liquid flow rate, solid line: liquid phase; dashed line: gas phase. Operating conditions: ceramic foam height: 25.4 cm; ceramic foam type: 20 PPI; liquid phase: 30 wt.% DGA sovlent; gas phase: 13% CO2/87% N2; absorption temperature: 40°C; lean loading: 0.2 mol CO2/mol DGA.

6. Conclusion 32


Page 32 of 42

Page 33 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


We investigated the hydrodynamic characteristics of α-Al2O3 ceramic foam and the performance of CO2 absorption using aqueous 30 wt% DGA in a gas-liquid counter current column packed with α-Al2O3 ceramic foams. The 1-D mathematical model we developed to further analyze pressure drop and liquid flooding point and CO2 absorption compared well with our experimental results. Ceramic foams allowed for higher flow rates of gas and liquid to be used before flooding occurred, compared to random packings. Ceramic foams with lower PPI numbers were less susceptible to flooding. Liquid velocity improved CO2 removal efficiency, but the flooding issue resulted in an upper velocity limit. Decreasing the gas velocity improved CO2 removal efficiency but at the cost of reducing operating capacity of CO2 treatment. A minor "hot spot" (

Improved CO2 Absorption in a GasâLiquid Countercurrent Column

Recommend Documents

Improved CO2 Absorption in a GasâLiquid Countercurrent Column