International Journal of Greenhouse Gas Control 103 (2020) 103179

Contents lists available at ScienceDirect

International Journal of Greenhouse Gas Control

journal homepage: www.elsevier.com/locate/ijggc

Influence of Particles on Amine Losses During CO2 Capture: A Process

Simulation Coupled Aerosol Dynamics Model
David Dhanraj , Pratim Biswas *
Aerosol and Air Quality Research Laboratory, Center for Aerosol Science and Engineering, Department of Energy, Environmental and Chemical Engineering, Washington
University in St. Louis, MO 63130, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Aerosol-driven solvent losses have been identified as one of the major challenges of amine-based post-com­
Process simulation bustion CO2 capture. In this work, a multi-component aerosol dynamics model based on the discrete-sectional
Post combustion CO2 capture approach, accounting for condensation and coagulation was coupled with a process simulation model devel­
Aerosol dynamics
oped using ASPEN PLUS v10® to account for the multi-component mass and heat transfer, hydrodynamics,
Amine losses
thermodynamics, and electrolytic-solution chemistry. Particle losses inside the absorber were incorporated into
the model based on a cut-off size determined from experiments reported in the literature. Based on the results, it
was shown that neglecting particle losses inside the absorber resulted in a significant over-prediction of amine-
based solvent losses, while coagulation of particles resulted in reduced (~10%) amine emissions. Furthermore,
amine emissions increased when the number concentration and the geometric standard deviation of inlet par­
ticles in the flue gas were increased. Moreover, it was shown that amine emissions were lower at lower solvent
concentrations and temperatures. CO2 capture efficiency dropped for decreasing solvent concentration, but
remained unchanged for lower temperatures, suggesting that decreasing the solvent temperature is an efficient
strategy to reduce amine emissions.

(Sreedhar et al., 2017) and increased environmental impacts, because

1. Introduction
Driven by escalating concerns about global warming and climate
change (Wang et al., 2011), significant efforts have recently been initi­
ated to reduce the concentration of atmospheric CO2. The generation of
power from fossil fuels is the largest source of anthropogenic CO2
emissions (Freund, 2003). Nonetheless, fossil fuel-fired power plants are
critical in meeting the energy demands of developing countries. For
instance, China, the largest consumer of coal, will still depend on coal
for 50 % of its energy demands in 2030 (He, 2015). Developing CO2
emission abatement strategies such as carbon capture and storage (CCS)
will ensure that fossil-fuel powered power plants can continue to operate
until renewable energy becomes cost-effective.
Post-combustion CO2 capture using amine-based solvents is
currently the most economical retrofit process for large power plants,
(Majeed et al., 2017a) because it does not require radical changes to the
existing power plants. However, this advantage comes with an efficiency
penalty: the separation process consumes 75-80 % of the energy of the
CCS process (Davison, 2007). Solvent losses pose a primary challenge for
amine-based CO2 capture, resulting in higher operational costs
amines react with NOx in the atmosphere to form carcinogens, such as
nitrosamines (Pitts et al., 2020). The suggested design criterion for the
technology to become applicable is to limit amine emissions below 12
mg/Nm3 (Khakharia et al., 2013).
In this process, liquid amine is sprayed at the top of a packed bed
absorber, with the flue gas contacting the solvent in counter-current
flow. A major challenge during this process is the loss of amine due to
emissions. Amine emissions are of two types: vapor-based solvent
emissions due to amine’s volatility, and aerosol-based amine emissions
(Harsha et al., 2019). The volatility of amine and the exothermicity of
the liquid phase reactions between CO2 and amine, vaporize amine
within the absorber, while fine particles in the flue gas acts as conden­
sation nuclei leading to aerosol-based emissions. Consequently, these
aerosols grow by condensation and coagulation, and exit the absorber
with condensed amine. Although countermeasures such as water wash
can abate vapor-based emissions efficiently (Mimura et al., 2004),
designing efficient ways to abate aerosol-based emissions requires
further understanding of the underlying mechanisms governing the
evolution of aerosols. Given the complexity of the problem, many re­
searchers (Khakharia et al., 2013; Harsha et al., 2019; Khakharia et al.,

* Corresponding author.
E-mail address: pbiswas@wustl.edu (P. Biswas).

https://doi.org/10.1016/j.ijggc.2020.103179
Received 24 November 2019; Received in revised form 9 August 2020; Accepted 27 September 2020
Available online 23 October 2020
1750-5836/© 2020 Elsevier Ltd. All rights reserved.
D. Dhanraj and P. Biswas
Nomenclature
a specific area (m2/m3)
AC cross-section area of column, m2
ct total molar concentration, mol⋅m-3
Deff effective diffusion coefficient, m2⋅s-1
F Faraday’s constant, 96500 C⋅mol-1
G total gas molar flow, mol⋅s-1
h molar enthalpy, J⋅mol-1
I particle condensation rate, cm3⋅s-1
lh liquid hold-up, -
L total liquid molar flow, mol⋅s-1
dp particle diameter, nm
dpg geometric mean diameter, nm
kB Boltzmann constant, m2⋅kg⋅s-2⋅K-1
KD phase distribution coefficient, -
NTOT total number concentration of particles, cm-3
qi particle population function of discrete size i
qi,sat saturation particle population function on the surface of
discrete size i
Qk particle population function of section k
Qk,sat saturation particle population function on the surface of
section k
QG volumetric flowrate of gas, m3⋅s-1
t time, s
T ambient temperature, K
v particle volume, cm3
V Volume, cm3
x liquid phase mole fraction, -
y gas phase mole fraction, -
z valence of ionic species, -
2015; Mertens et al., 2014a; Mertens et al., 2017; Mertens et al., 2012;
Moser et al., 2009; Moser et al., 2017) have performed both bench- and
pilot-scale experiments that illuminate the phenomena involved. A good
summary on these experiments was provided by Harsha et al., 2019
(Harsha et al., 2019). The experiments of Khakharia et al., 2013 (Kha­
kharia et al., 2013) clearly distinguish vapor-based and aerosol-based
emissions. Two important conclusions from their work were that
introducing particles (H2SO4 droplets) to the flue gas increases amine
emissions from 45 mg/Nm3 (vapor-based) to 1100 mg/Nm3 (aero­
sol-based + vapor-based) and that amine emissions increase with an
increase in the number concentration of the particles. It can be inferred
from these results that particles act as condensation nuclei for amine
condensation, and as amine in the vapor phase is depleted, more amine
from the liquid phase vaporizes, leading to increased emissions. How­
ever, it is critical to note that the vapor phase is not saturated with
amine. As water vapor in the column is near saturation, water will be
also condense (Fulk and Rochelle, 2013). Hence, the concentration of
amine on the particle’s surface is very small relative to its gas phase
concentration, driving amine condensation. Furthermore, some experi­
ments conducted in the past (Harsha et al., 2019; Brachert et al., 2013)
have shown reasonable evidence of aerosol coagulation inside the
absorber. Harsha et al., 2019 (Harsha et al., 2019) showed that in the
case of H2SO4 aerosols, the number concentration reduced from 6.24 ×
107 #/cm3 at the inlet of the absorber to 2.30 × 107 #/cm3 at the outlet.
They attribute this 63 % decrease in number concentration to coagula­
tion and depositional losses.
Several researchers (Majeed et al., 2017a; Fulk and Rochelle, 2013;
Majeed et al., 2018; Majeed et al., 2017b; Majeed and Svendsen, 2018;
Majeed et al., 2017c; Zhang et al., 2017) have developed models to
investigate this system by coupling a process simulation package (to
model the multi-component reactive mass transfer inside the absorber)
2
International Journal of Greenhouse Gas Control 103 (2020) 103179
Greek letters
β coagulation collision frequency function, cm3⋅s-1
β
D
coagulation collision frequency function between discrete
size and section, cm3⋅s-1
DD
β coagulation collision frequency function between discrete
sizes, cm3⋅s-1
DDS
β coagulation collision frequency function between discrete
sizes to form section, cm3⋅s-1
δ film thickness, m
η dimensionless film co-ordinate, -
ζ population index: 0-number, 1-volume, 2-volume square
σ surface tension, N⋅m-1
τ residence time, s
σg geometric standard deviation
λ thermal conductivity, W⋅m-1⋅K-1
φ electrical potential, V
ϕ volumetric hold-up, -
Subscripts
i component index
m mth species
n number of components
t total
Superscripts
b bulk
f film
g gas phase
i interface
k kth section
l liquid phase
with an aerosol dynamics model. However, only Majeed et al., 2018
(Majeed et al., 2018) consider the particle size distribution in their
model, while the rest model the growth of single particle/droplet inside
the absorber solely by condensation. However, Majeed et al., 2018
(Majeed et al., 2018) use a Moment-based model to track the average
particle size and standard deviation, while setting the total number
concentration as a constant, because the model considers only conden­
sation. No models account for coagulation or particle loss, and most of
these models are not validated against experimental data.
Therefore, in this work a coupled model is developed that integrates
a multicomponent aerosol dynamics approach that accounts for the time
evolution of the particle size distribution with a process simulation
model. The model predictions are validated with published experi­
mental data (Khakharia et al., 2013). The multi-component mass and
heat transfer between bulk gas and bulk liquid was modelled using
ASPEN PLUS v10®, and the multi-component aerosol dynamics was
modelled using the discrete-sectional approach. An empirical approach
to account for the particle losses inside the absorber, developed based on
the experimental data of Khakharia et al., 2013 (Khakharia et al., 2013),
is presented. Furthermore, the influence of considering the vapor con­
centrations in each stage of the absorber as a constant vs. depleting, was
investigated. The integrated model was used to investigate the effects of
particle number concentration and initial particle size distribution on
amine emissions. Particle number concentrations between 1.0 × 1014 -
1.8 × 1014 #/m3 (corresponding to the experiments of Khakharia et al.,
2013 (Khakharia et al., 2013)) were investigated and the predicted
amine emissions were compared with the experimental data. Three
distributions: monodisperse, log-normal, and bi-lognormal, were
considered. Further, the integrated model was used to study the influ­
ence of solvent (30 wt.% Mono Ethanol Amine (MEA)) concentration,
and solvent temperature on vapor- and aerosol-based amine emissions,
D. Dhanraj and P. Biswas International Journal of Greenhouse Gas Control 103 (2020) 103179

and CO2 capture efficiency. Solvent concentrations between 0.06 – model, 16 stages were used in APP to describe the column of Khakharia
0.125 (MEA mole fraction) and temperatures between 34 – 44 ◦ C, were et al., 2013 (Khakharia et al., 2013). Also, 20 discretization points were
considered in the study. used to model the liquid film. In the DISC model, the vapor phases
(amine and water) were considered as discrete sizes, while the particles
2. Model description - The framework of DISC_APP were treated as sections with a maximum of 400 sections. The DISC
simulations were performed in the Engineering Cloud Cluster of Wash­
Amine-based CO2 capture is a thermal-swing reactive absorption ington University in St. Louis.
process. A schematic of the various physicochemical processes occurring
inside the absorber is shown in Fig. 1. The present model accounts for 2.1. ASPEN PLUS v10® model (APP)
these processes. The flue gas containing particles contacts the liquid
aqueous amine counter-currently, while the aerosols flowing with the The multicomponent mass and heat transfer between gas and liquid
gas grow by condensation and coagulation, and exit the column. The phases, the hydrodynamics, the thermodynamics, and the solution
reactive mass transfer process is kinetically limited (Schneider et al., chemistry were all modelled using APP. RadFrac was used to model the
1999) and hence, a rate-based model is required. Also, since the disso­ reactive mass transfer between the two phases, and the ELECNRTL
lution of CO2 in aqueous amine generates electrolytes, the non-ideality property method was used. In the rate-based model, separate mass and
due to ionic interactions should be accounted for. Since RadFrac, a heat balance equations (Eqn. 1 – 4) are written for each phase, while the
rate-based-model in ASPEN PLUS v10® (APP), has been extensively mass and heat transfer resistance are estimated by the two-film theory
used (Fulk and Rochelle, 2013; Zhang et al., 2017; Zhang and Chen, (Taylor and Krishna, 1993) through implicit calculations of interfacial
2013) and shown to predict this system reasonably well, it was used to fluxes (Eqns. 5 – 8). The bulk phase equations consider the axial change
simulate the multi-component mass and heat transfer between gas and of the total molar stream, composition, and enthalpy, and are written as
liquid phases.
The time-evolution of aerosols can be predicted by solving the gen­ dLxlb
i
= nlb i lb i l
(1)
i a AC + Ri ϕ AC , i = 1, ..., n
eral dynamics equation (GDE) for aerosol growth (Friedlander, 1977). dz
Although there are multiple modeling approaches to solve the GDE, the
discrete-sectional (DISC) approach was used to solve the partial dGygb
i
= ngb i
i a AC , i = 1, ..., n
g
(2)
integro-differential equations, given that it is the most versatile (in terms dz
of computational cost and accuracy) aerosol dynamics model (Wu and
dLhlb
Biswas, 1998; Wu and Flagan, 1988). Because the concentration and = qlfi ai AC (3)
dz
temperature of the different species in the vapor phase govern the
growth of aerosols, the predictions from APP were used as inputs to the dGhgb
DISC. The modelling framework of the integrated model (DISC_APP) is = qgf i
i a AC (4)
dz
shown in Fig. 2. The experimental system of Khakharia et al., 2013
(Khakharia et al., 2013) was used as the base case for simulation. In the In the film regions, differential mass and heat balances in the film
coordinate (ηi ) were performed, wherein the balance equations and
interfacial fluxes are written as

1 dnlfi
= Rlfi , i = 1, ..., nl (5)
δl d ηl

1 dngfi
= 0, i = 1, ..., nl (6)
δg d ηg
( )
clft Dleff ,i dxlfi F dφ
nlfi =− lf
+ xi zi + xlfi nlft , i = 1, ..., nl (7)
δl dηl RG T d η l

( )
cgf g
t Deff ,i dygf
ngf
i =− g
i
+ ygf gf
i nt , i = 1, ..., n
g
(8)
δ d ηg

lf
( ) nl
∑
λlf dTil
lf
q = − lf + n hlfi (9)
δ d ηl i=1
i

gf
( ) ng
∑
λgf dT g
gf
q = − gf + n hgf
i (10)
δ d ηg i=1
i

The flux due to ionic chemical species is accounted using the Nernst-
Plank expression (in Eqn. 7). Further, the thermodynamic equilibrium at
the gas-liquid interface, described by the relation below, relates the gas
and liquid phase compositions at the interface (Schneider et al., 1999
(Schneider et al., 1999)):

yii = KD,i xii (11)

Fig. 1. Operating conditions of the flue gas and liquid solvent, the evolution of
aerosols inside the absorber, and the important aerosol phenomena that affect Wherein, the phase distribution coefficient (KD,i ) in Eq. 11 comprises
solvent losses during post-combustion CO2 capture. fugacities in both phases and activity coefficients in the liquid phase.

3
D. Dhanraj and P. Biswas
Fig. 2. The modelling framework developed in this work, showing the multi-component mass and heat transfer between bulk gas and liquid phases (modelled using
ASPEN PLUS v10®) and multi-component aerosol dynamics (modelled using the discrete-sectional approach).
The calculations for the non-idealities are based on the three-parametric
Electrolyte-NRTL method. The thicknesses of the films were estimated
using the empirical correlations suggested by Taylor and Krishna, 1993
(Taylor and Krishna, 1993). The holdup and pressure drop were esti­
mated using the Billet and Schulte’s (Billet and Schultes, 1999) corre­
lation. The interfacial area and mass transfer coefficients were estimated
using the correlation suggested by Hanley et al., 2011 (Brian Hanley,
2011), and the heat transfer coefficient was estimated using the Chilton
and Colburn correlation (Chilton and Colburn, 1934).
The equilibrium and kinetic reactions in the liquid phase between
MEA, CO2 and H2O were considered in the model. Although there are
different studies on the reaction kinetics, the reactions suggested by
Zhang et al., 2013 (Zhang and Chen, 2013) provide reasonable agree­
ment with experimental data, and were hence included in this model.
The equilibrium (12-16) and kinetic (17,18) reactions are shown below:
2H2 O = H3 O+ + OH − (Waterhydrolysis) (12)
(
RNH2 OH + H3 O+ = RNH3+ OH + H2 O Amineprotonation
)
(13)
( )
CO2 + 2H2 O = H3 O+ + HCO−3 Bicarbonateformation (14)
( )
HCO3 − + H2 O = H3 O+ + CO2−3 Bicarbonatedissociation (15)
( )
RNH2 OH + HCO−3 = RNHOHCOO− + H2 O Carbamateprotonation
(16)
RNH2 OH + CO2 + H2 O = RNHOHCOO− + H3 O+ (17)
CO2 + OH − = HCO−3 (18)
The activity-based kinetic expressions suggested by Zhang et al.,
2013 (Zhang and Chen, 2013) were used to describe the rate of the ki­
netic reactions.
4
International Journal of Greenhouse Gas Control 103 (2020) 103179
2.1.1. The discrete sectional model (DISC)
The general dynamics equation for aerosol growth due to conden­
sation and coagulation can be written as (Friedlander, 1977)
∫
∂n(v, t) ∂(Gn) 1 v
+ = β(v − ṽ, v)n(v
∂t ∂v 2 0
∫∞
− ṽ, t)n(ṽ, t)dṽn(v, t) β(ṽ, v)n(ṽ, t)dṽ (19)
0
where, the second term on the L.H.S. accounts for condensation at rate
G, while the terms on the R.H.S account for formation and loss due to
Brownian coagulation. The discrete-sectional approach is a mathemat­
ical simplification to the GDE wherein balance equations are written for
two general aerosol properties qi,m in the discrete size regime, and Qk in
the sectional size regime (Eqns. 20 and 21). For a multicomponent
aerosol of m species, with its size characterized by its volume v, a general
aerosol property can be written as
qi,m = nm (iv1 ) ∙(iv1 )ζ (20)
where nm (iv1 ) represents the number concentration of species m, while
v1 represents the volume of the monomer or vapor molecule or the first
discrete size. ζ is the aerosol property index such that, when ζ = 0, qi,m is
the size distribution function, and at ζ = 1, qi,m is the aerosol volume
concentration. Similarly
Qk = N (vk ) ∙(vk )ζ (21)
where N (vk ) denotes the number concentration of section k. In this
work, the vapor species (amine and water) were modelled as discrete
sizes, the particles were considered as sections and number concentra­
tion was chosen as the conserved property (ζ = 0). Time step sizes were
chosen as a function of time, wherein the size was finer when the driving
force for condensation was higher and vice-versa. The time-step sizes
ranged between 10-8 - 10-6. Some numerical diffusion was observed
D. Dhanraj and P. Biswas International Journal of Greenhouse Gas Control 103 (2020) 103179

when the particle sizes reached around 10-6 m, and hence the choice of
QG
the finest time-step size was made such that the results were not influ­ QG,eff = (30)
(1 − lh)∗vfrac
enced by a further decrease in time-step size.
The balance equations are written as: It should be noted that the mass transfer of CO2 in the particles
dqvapor,m (amine droplets) has been neglected. As the droplets are rather small
=0 (22) (micrometer sized), this assumption is fairly accurate and will not
dt
impact the results. Moreover, for the system considered, the mole frac­
dqvapor,m ∑
max ∑
max
4 D ( ) tion of amine in the aerosol phase is in the order of 1 × 10-4, due to the
=− β1,k,m q1,m − Qk,sat,m Qk (23) significantly higher amounts of water in the aerosol phase. The gradients
dt
in concentration are rather small for the small droplets and can be
t=1 k=1

∑
MAX ∑
mmax assumed to be well mixed for the dilute concentrations. For conditions in
dQ1 13 16 2 D ( )
which the amine concentration is reasonably high (5 M, not this work),
4
= − β1 Q1 2 + β1 Q1 2 − Q1 β1,i Qi − β1,i,m q1,m − Qi,sat,m Qi
dt 2 2 i=2 m=1 the estimation and impact of radial concentration profiles of carbamate
∑ and amine are discussed in detail by Majeed et al., (Majeed et al., 2017b;
mmax
5 D ( )
+ β1,i,m q1,m − Qi,sat,m Qi
m=1 Majeed et al., 2017c).
(24)
3. Results and Discussion
dQk 1∑k− 1 ∑
k− 1
1 D ∑
k− 1
2 ∑
k− 1
5

dt
=+
2 i=1 j=1
βi,j,k Qi Qj + − Qk βi,k Qi + + Qk βi,k Qi 3.1. Simulation plan
i=1 i=1

∑ This section describes the different simulation cases performed in

MAX
13 16 2 2 D ( )
+− βk Qk 2 + βk Qk 2 − Qk βi,k Qi − β1,k,m q1,m − Qk,sat,m Qk
2 2 i=k+1 this work. A complete list of simulations is tabulated in Table 1. Section
5 D ( ) 1 D ( ) 3.2 describes the model development to account for particle losses, the
+ β1,k,m q1,m − Qk,sat,m Qk + β1,k− 1,m q1,m − Qk− 1,sat,m Qk− 1 influence of a constant vs. depleting vapor profile in the absorber stages,
(25) and the comparison of vapor- and amine-based emissions to the exper­
imental measurements of Khakharia et al., 2013 (Khakharia et al.,
where 2013). The effects of particle size distribution and number concentration
[ (
4σ m v1m
)] on amine emissions are described in Section 3.3. Section 3.3 also de­
Qk,sat,m = γ m xm nsat,m exp vk ζ (26) scribes the influence of different aerosol phenomena such as coagulation
dk,m KT
and condensation on amine emissions. The respective effects of amine
[( ∑max 4 D ( ) )] concentration and temperature on amine emission and CO2 capture ef­
∑
max
k=1 β1,k,m q1,m − Qk,sat,m Qk
xm = (27) ficiency are described in Sections 3.4 and 3.5.
∑max ∑max 4 D ( )
k=1 β1,k,m q1,m − Qk,sat,m Qk
t=1
m=1

The condensed concentration of water and amine were estimated as

Table 1
dqcond,m ∑max ∑
max
4 D ( )
= β1,k,m q1,m − Qk,sat,m Qk (28) Simulation plan
dt t=1 k=1
Section Range/condition Description
Eqn. 22 is the balance equation for vapors for the case in which the Comparison of predicted
concentration of the vapors is treated as constant, whereas Eqn. 23 is the temperature, MEA, and water
Base-line case of Khakharia
balance equation for the case wherein the vapors deplete. Eqn. 24 and vapor profiles with
et al., 2013 (Khakharia
25 are the balances on the size distribution of the particles (treated as experimental measurements of
et al., 2013)
Khakharia et al., 2013 (
sections) with Eqn. 24 being the balance for the first section and the
Model validation Khakharia et al., 2013)
latter being the balance for the other sections. The first and second terms Comparison of predicted amine
on the R.H.S of Eqn. 1 are respectively the rate of loss and gain of par­ emissions with experimental
Number concentration:
ticles in section 1 due to coagulation. The 3rd term is the loss due to (1.0 - 1.8) x 108 #/cm3
measurement of Khakharia
coagulation of section 1 particles with particles in other sections. The 4th et al., 2013 (Khakharia et al.,
2013)
and 5th terms are the loss and gain of section 1 particles due to Development of an empirical
condensation. Eqn. 25 includes similar terms for the other sections. The approach to account for particle
n D Operating conditions as per
collision frequency functions are represented as βi,j,m , and for the Model losses inside the absorber
(Khakharia et al., 2013) for
development Study the influence of depleting
complete list, refer to Biswas et al., 1997 (Biswas et al., 1997). H2SO4 droplets
vs. constant vapor profile in
Condensation was assumed to be transport limited, and the condensa­ each absorber stage
tional growth law corrected using the Fuchs interpolation formula for Effect of particle number
(1.0 - 1.8) x 108 #/cm3 concentration on amine
the entire Knudsen Number range was used (Friedlander, 1977). The Effect of particle
emissions
coagulation collision frequency functions for the free molecular and size
Effect of geometric standard
distribution
continuum regime were used for corresponding regimes, and for the 1 – 2.4 deviation of particle size
transition regime, the interpolation formula suggested by Pratsinis, distribution on amine emissions
1988 (Pratsinis, 1988) was used. Effect of aerosol Number concentration: Effect of condensation and
phenomena (1.0 - 1.8) x 108 #/cm3 coagulation on amine emissions
The vapor pressure of amine and water was calculated using Antoine Effect of inlet solvent
equation. The residence time for particles inside the column was esti­ 0.06 – 0.125 (mole fraction concentration on amine
mated based on the following equations (Fulk, 2016): of MEA) emissions and CO2 capture
Effect of solvent efficiency
(void fraction)∗Vpacked bed inlet condition Effect of inlet solvent
τstage = (29)
QG,eff temperature on amine
34 – 44 0C
emissions and CO2 capture
efficiency

5
D. Dhanraj and P. Biswas
3.2. Model development and validation
The experimental data of Khakharia et al., 2013 (Khakharia et al.,
2013) were chosen for validation, because it can be used to separately
validate the APP and APP_DISC results. The parameters and operating
conditions relevant to this work are mentioned here, however a detailed
description can be found in (Khakharia et al., 2013). They used an ab­
sorption column that was 3.5 m high and consisted of four packed beds
(Sulzer Mellapak 2X, height = 0.5 m, diameter =4.5 cm). The column
did not include a water wash or demister, and the solvent was 30 wt.%
aqueous MEA at 40 ◦ C at a typical lean loading of 0.25 mol CO2/mol
amine. The comparison of model predictions with the corresponding
experiments are discussed in the following sections.
3.2.1. Vapor-based emissions
In the base line study, Khakharia et al. 2013 (Khakharia et al., 2013)
measured vapor-based amine emissions using a FTIR analyzer (GASMET
CX 4000) after the flue gas exiting the absorber was heated to 180 ◦ C.
The amine concentration was measured to be 45 mg/Nm3, with a re­
ported error of ±10%. In this case, the simulated flue gas did not include
particles, and it consisted of CO2 (12.8 vol. %), O2 (14.6 vol.%) and CO
(43.5 mg/Nm3). They reported a typical temperature profile inside the
bed which shows a temperature bulge at the upper-middle section of the
absorber, and reported the concentration of CO2 in the exiting flue gas to
be 2.5 vol.%. The APP system in our work was setup based on these
conditions (considering all the conditions mentioned in their work), and
a comparison of simulated vapor-phase MEA, CO2, and temperature
profiles with the experimental data is shown in Fig. 3. While the APP
simulations predict the measured temperature, concentration of MEA
and CO2 at the outlet of the absorber very well, the simulations capture
Fig. 3. Comparison of ASPEN PLUS v10® simulation results using the experimental data of Khakharia et al., 2013 (Khakharia et al., 2013) showing (a) vapor phase
temperature profile (b) MEA vapor concentration, and (c) gas phase CO2 volume fraction inside the absorber.
6
International Journal of Greenhouse Gas Control 103 (2020) 103179
the temperature profile in the upper-middle section of the absorber
satisfactorily well, wherein the simulation under-predicts the tempera­
ture near the second stage of the absorber. Note that Khakharia et al.,
2013 (Khakharia et al., 2013) describe these measurements as typical
temperature profiles and do not explicitly mention how and for which
case the profile was obtained. However, the temperature profiles
measured and predicted by Zhang et al., 2013 (Zhang and Chen, 2013)
are very similar to the APP model predictions (see Fig. 6 (a) of Zhang
et al., 2013). The gas phase concentration of MEA and CO2 predicted by
the model agrees well with the profiles predicted by Majeed et al., 2017
(Majeed et al., 2017a). They have also used ASPEN to calculate the gas
phase compositions. While CO2 concentration showed a steady decrease
with respect to increasing absorber height, MEA concentration increased
along the height of the column with the maximum increase near the top
of the column. These trends can be observed in the work of Majeed et al.,
2017 (Majeed et al., 2017a). Further, note that a counter-current flow
model was used in APP to describe the vapor and liquid phase flow in the
APP model because it predicted the extent of back-mixing in the column
quite well. Also, the choice of the empirical correlations for interfacial
area, mass transfer coefficients (Hanley et al., 2011; (Chilton and Col­
burn, 1934)) and heat transfer coefficient (Chilton and Colburn, 1934)
was made such the model represents the experimental system reason­
ably well.
3.2.2. Accounting for vapor depletion
The simulated concentration of the vapors and the gas-phase tem­
perature profiles by the APP model for the base-line case of Khakharia
et al., 2013 (Khakharia et al., 2013) is used as input to the DISC model.
The amine vapor phase concentration predicted by the APP model is
shown in Fig. 4 (a). The mass transfer of vapors from the gas phase to the
D. Dhanraj and P. Biswas International Journal of Greenhouse Gas Control 103 (2020) 103179

Fig. 4. (a) The MEA vapor profile using APP model without including vapor losses, (b) the depleted MEA vapors calculated using a DISC model (with depleting
vapors), and (c) the MEA vapor profile using APP model while considering vapor loss compared with case (a).

particles surface due to condensation can lead to depletion of vapors
inside the absorber when the rate of depletion of vapors due to
condensation is significantly greater than the rate of vaporization from
the bulk liquid. However, if the rate of condensation is comparable or
smaller than that of vaporization, then the vapor profiles inside the
absorber will remain constant during the process. In order to verify this,
DISC simulations were performed corresponding to the base-line case
discussed previously. The size distribution of the particles was chosen
such that they are relevant to the experiments performed by Khakharia
et al., 2013 (Khakharia et al., 2013) wherein they introduced H2SO4 acid
droplets into the flue gas. The number concentration was 1.016 × 108
#/cm3, monodisperse at 100 nm. The DISC simulations were performed

7
D. Dhanraj and P. Biswas
for depleting vapors (Eqn. 23) and mass of vapors depleted in each stage
per unit volume of gas was computed, and shown in Fig. 4 (b). It can seen
that most of the amine vapor depletion is observed at the top of the
column. These results agree with the model predictions of Majeed et al.,
2017 (Majeed et al., 2017b) wherein they have simulated the changes in
amine partial pressure due to the condensation of amine onto droplets
that are initially 0.15 μm in size (Case 5) and at different number con­
centrations. They have shown that at NTOT = 107 #/cm3, amine deple­
tion is observable near the top of the column. Further, APP model was
modified to account for the depleting vapors at each stage (see Fig. 4 (c))
and the steady-state vapor and temperature profiles were computed. It
can be seen in Fig. 4 (c) that the vapor profiles before and after ac­
counting for vapor depletion are similar. This result is because the sur­
face area available for condensation (0.0406 m2) is two orders of
magnitude lower than that available for gas-liquid contact (1.87 m2).
Therefore, for further cases, the DISC simulations were performed for a
constant vapor system (Eqn. 22).
3.2.3. Particle losses and aerosol-based emissions
Khakharia et al., 2013 (Khakharia et al., 2013) measured
aerosol-based emissions by introducing particles (H2SO4 acid droplets)
Fig. 5. (a) Simulated amine mass in sections and the estimated cut-off size to account for particle losses based on the experimental data of Khakharia et al., 2013
(Khakharia et al., 2013), (b) estimated cut-off sizes at different initial number concentrations, and (c) the comparison of predicted amine emissions with experimental
measurements.
8
International Journal of Greenhouse Gas Control 103 (2020) 103179
in their flue gas and similarly measuring the heated exiting flue gas. In
this case, the amine emissions increased to 600 mg/Nm3 when the
number concentration of the particles was 1.016 × 1014 #/m3, and
further increased to 853 mg/Nm3 and 1163 mg/Nm3 when the number
concentration was increased to 1.3 × 1014 #/m3 and 1.73 × 1014 #/m3,
respectively. They measured the particle number concentration using a
condensation particle counter (CPC; PALAS UFCPC with Sensor 200) at
the inlet to the absorber. Note that these amine emission values corre­
spond to total amine emissions (vapor + aerosol). However, the
aerosol-based emissions can be calculated by subtracting the known
vapor-based emissions (45 mg/Nm3). The temperature, vapor phase
amine, and water concentration profiles calculated using the APP model
were input to the DISC model. An initial particle size of 100 nm was used
because it was reported by Khakharia et al., 2013 (Khakharia et al.,
2013) (based on their simulations). The inlet particle size distribution
was assumed to be monodisperse, as there was no mention of the dis­
tribution. However, in our work the effect of initial particle size distri­
bution was investigated separately (Section 3.2). We simulated the time
evolution of particle size distribution using DISC for the conditions of
the Khakharia et al., 2013 (Khakharia et al., 2013) wherein they studied
the effect of H2SO4 droplets’ number concentration on amine emissions.
D. Dhanraj and P. Biswas International Journal of Greenhouse Gas Control 103 (2020) 103179

The simulated amine emissions resulted in significant over-prediction. Table 2

The simulated amine emission at NTOT = 1.016 × 1014 #/m3 was 3.85 Estimated cut-off sizes at different experimental data points of Khakharia et al.,
× 105 mg/m3. However, this was expected as the particle losses were not 2013 [8].
considered. Since there are no accurate experimental data/models to Data Measured Measured Amine Estimated Amine Estimated Cut-
account for particle losses in this system, the measured amine emission Point NTOT emission emission off Size
data points (of Khakharia et al., 2013 (Khakharia et al., 2013)) corre­ (-) (#/m3) (mg/m3) (mg/m3) (μm)

sponding to each number concentration was used to determine a cut-off 1 1.02 x 1014 601.58 556.58 29.7434
size such that the simulated amine emissions predict the experimentally 2 1.11 x 1014 603.17 558.17 29.7442
3 1.17 x 1014 831.13 786.13 29.7442
measured values. The cut-off size is defined in this work as the size (see
4 1.21 x 1014 846.97 801.97 29.7442
Fig. 5 (a)) at which the cumulative mass concentration of amine 5 1.31 x 1014 853.30 808.30 29.7434
condensed on these particles equals the measured aerosol-based amine 6 1.43 x 1014 1105.01 1060.01 29.7442
mass concentration (emission). The measured amine mass concentration 7 1.77 x 1014 1163.59 1118.59 29.7442
for different H2SO4 droplet number concentration was fitted such that
few representative data points (number concentration and amine
simulations, the aforementioned value was used as the cut-off size. It
emission) can be chosen for determining the cut-off size. The specific
should be noted for the case in which the solvent concentration was
reason for fitting the data was because at the same number concentra­
decreased, the cut-off size approach did not work as the mean size of the
tion, Khakharia et al., 2013 (Khakharia et al., 2013) have reported
particles at the final time was smaller than the cut-off size. Hence for this
different amine emissions (see Fig. 5 (b)). The chosen number concen­
case, the percent loss in the particles estimated from the base case at the
tration data points, measured amine emissions, fitted amine emissions
same number concentration was used to determine the cut-off size.
and estimated cut-off sizes are shown in Table 2. It can be seen in Fig. 5
Further simulations were performed to predict the effect of the number
(b) that the cut-off size for the different number concentrations was very
concentration of particles on amine emissions, and Fig. 5 (c) shows that
similar and it was found to be 29.74 μm (average cut-off size), based on
the simulations predict the amine emissions reasonably well. The
the simulations performed for the seven cases (see Fig. 5 (b)). The
number concentrations corresponding to the experimental measure­
physical significance of this value is that it can be considered that the
ments of Khakharia et al. 2013 (Khakharia et al., 2013) were used, while
particles greater than this cut-off size will be lost inside the absorber and
the distribution was considered to be monodisperse at 100 nm. The
will not contribute to amine-based emissions. Therefore, for further
simulations predict the amine emissions reasonably well. It can be

Fig. 6. (a) Lognormal, bi-lognormal (b) experimentally measured (Mertens et al., 2014) size distributions considered to investigate the effect of the distribution on
amine emission. Effect of coagulation of particles on (c) amine emission and (c) water emission.

9
D. Dhanraj and P. Biswas International Journal of Greenhouse Gas Control 103 (2020) 103179

inferred from these results that aerosol-based amine emissions scale correspondingly higher.
proportionally with particle number concentration, due to the increase In order to understand the influence of coagulation on amine emis­
in surface area available for condensation. This increase in amine sions, simulations were performed with and without coagulation terms
emission due to the increase in total number concentration agrees well in the balance equations. In both cases, condensation was considered.
with the model predictions of Majeed et al., 2017 (Majeed et al., 2017b). These simulations were performed for two different initial number
They have shown that as NTOT increases, the depletion of amine in the concentrations (lowest and highest in the range of interest). For these
gas phase increases due to condensation to the droplets. It can therefore simulations, the size distribution was considered to be monodisperse,
be inferred that at higher number concentrations, the surface area with particle size of 100 nm. Fig. 6 (b), shows that the model considering
available for condensation is higher, leading to increased amine deple­ coagulation and condensation predicted a lower amine emission
tion which results in higher amine emissions. (~10%) than a pure condensation model. This difference was observed
at all number concentrations considered, and it increased with an in­
crease in number concentration. A very similar trend was also observed
3.3. Effect of particle size distribution and aerosol phenomena on amine
in the predicted water emissions (see Fig. 6 (c)). These results can be
emissions
explained by the data provided in Table 4, wherein the final number
concentration is smaller for the model that considers coagulation than
We investigated the influence of initial particle size distribution on
for the pure condensation model. As mentioned earlier, amine emissions
amine emissions, considering monodisperse, log-normal, bi-lognormal
are directly proportional to the surface area available for condensation.
distributions and experimentally measured size distributions reported
Coagulation decreases surface area hence there is a decrease in amine
by Mertens et al., 2104 (Mertens et al., 2014b) (for the system of Kha­
emissions. Note that particle losses are also considered in these cases and
kharia et al., 2013 (Khakharia et al., 2013)). We have scaled down the
hence there is a decrease in number concentration in the pure conden­
measured size distribution such that the total number concentrations
sation case also. Therefore, coagulation of particles in flue gas decreases
match the measurements of Khakharia et al, 2103 (Khakharia et al.,
amine emissions and hence it is important to include coagulation in the
2013). Also, Mertens et al., 2014 (Mertens et al., 2014b) have reported
model to accurately predict solvent emissions.
that their measurements possibly include artifacts due to dilution, in­
clusion of a fan to assist ELPI sampling, and inclusion of a filter stage.
3.4. Effect of solvent concentration on vapor- and aerosol-based amine
Dilution and fan inclusion would shrink the particles and the inclusion of
emissions
the filter stage would over-estimate number concentration of particles in
the size < 10 nm. Table 3 shows the average parameters describing the
The DISC_APP model was further used to investigate the effect of
size distribution at the initial time, along with the amine emissions. For
solvent concentration on amine emissions. The mole fraction of MEA in
the bi-lognormal case, two log-normal distributions were added. For
the solvent was variously 0.06, 0.1 and 0.125. Fig. 7 (a) shows that the
each of the distributions, the number concentration was set equal to half
temperature profiles exhibit a bulge in each of the cases, while the
the NTOT of the other cases, while σg was set to 1.2. The dpg values were
temperature values showed an increase with increased MEA concen­
set to 50 nm and 200 nm so that the dpg of the bi-lognormal distribution
tration with significant changes towards the bottom of the column. In­
equaled 100 nm. The log-normal, bi-lognormal and the scaled-down
creases in the mole fraction of MEA resulted in higher absorption of CO2
version of experimentally measured distributions considered in the
(see Fig. 7 (d)), and hence more CO2 reacted with MEA in the liquid
simulation are shown in Fig. 6 (a). The results for these sets of simula­
phase. Because the reactions between MEA and CO2 are exothermic,
tions are shown in Table 3. Between the monodisperse, log-normal and
more heat was released, increasing the temperature inside the column,
bi-lognormal cases amine emissions increased with σg in all the cases
which in turn increased the vapor phase concentrations of MEA and
and the maximum increase occurred in the bi-lognormal case wherein
water (Fig. 7 (b) and (c). The increased MEA concentration in the vapor
the initial σg was the largest. The amine emissions are proportional to
phase (vapor-based emissions), caused the increased amine emissions
the initial surface area concentration (STOT), and hence the bi-lognormal
(aerosol-based) shown in Fig. 7 (e). This critical result shows that
case shows the highest emissions, while the monodisperse case shows
aerosol-based emissions increase when vapor-based emissions increase,
the lowest emissions. The calculated STOT is shown in Table 3. From
and is a consequence of increased MEA concentration.
these results, it can be inferred that amine emissions increase with
Vapor-based water emissions show a bulge very similar to the tem­
increasing standard deviation of the initial size distribution of the par­
perature profile. However, as the MEA concentration was increased, the
ticles in the flue gas for a given initial total number concentration and
concentration of water in the solvent decreased (so that the mole frac­
geometric mean diameter. However, the amine emissions for experi­
tions add up to unity). However, vapor-based emissions of water are
mental case, showed significantly higher emissions which is because the
more sensitive to temperature profile than to concentration, and hence
size distribution includes extremely high concentration of particles in
an increase in vapor-based emission was seen for a decrease in water
the size range 10 – 40 nm. Since a cut-off size of 29.74 μm was used, the
concentration in the solvent. This is because the relative change in water
smaller particles would take-up a significantly higher volume of amine
concentration is lesser than the relative increase in temperature values.
to reach that size and hence the estimated amine emissions are
Consequently, the aerosol-based water emissions increase even as the as
water concentration in the solvent is decreased.
Table 3
Effect of initial particle size distribution on amine emission.
Initial Final

Size distribution NTOT dpg σg STOT Amine

(#/m3) (nm) (-) (m2/m3) emission Table 4
(mg/m3) Effect of coagulation and condensation on amine emission.
Monodisperse 1.43 x 100 1.0 4.49 x 10- 9.08 x 102 Initial Final
1014 4
5 Pure condensation Coagulation and condensation
Lognormal 1.43 x 100 1.2 4.53 x 2.20 x 10
1014 100 NTOT NTOT Amine emission NTOT Amine emission
Bi-Lognormal 1.43 x 100 2.04 1.02 x 9.95 x 105 (#/m3) (#/m3) (mg/m3) (#/m3) (mg/m3)
1014 102 1.02 x 1014 1.50 x 1011 798.0 1.19 x 1011 749.0
Mertens et al., 1.44 x 19.77 2.013 1.12 x 2.84 x 108 1.43 x 1014 2.12 x 1011 1105.1 1.56 x 1011 957.3
2014 1014 100 1.78 x 1014 2.63 x 1011 1358.1 1.83 x 1011 1119.2

10
D. Dhanraj and P. Biswas
Fig. 7. Effect of solvent concentration on (a) vapor phase temperature profile, (b) MEA vapor concentration, (c) water vapor concentration, (d) gas-phase CO2 mole
fraction and (e) MEA and water emissions.
3.5. Effect of solvent temperature on vapor- and aerosol-based amine
emissions
The influence of the solvent’s inlet temperature was varied to
investigate its influence on vapor- and aerosol- based amine and water
emissions. The temperature was varied as 34, 39 and 44 ◦ C. Fig. 8 (a)
shows that the temperature profiles inside the absorber show an increase
11
International Journal of Greenhouse Gas Control 103 (2020) 103179
in temperature magnitudes when the solvent temperatures was
increased, which is a direct consequence of the adiabatic absorber. The
increase in the temperature profile inside the absorber results in
increased vapor- and aerosol-based amine and water emissions and are
shown in Fig. 8 (b), (c) and (e), while the uptake efficiency of CO2 did
not change significantly (Fig. 8 (d)). It should be noted that the increase
in aerosol-based amine emission due to higher inlet solvent temperature
D. Dhanraj and P. Biswas International Journal of Greenhouse Gas Control 103 (2020) 103179

Fig. 8. Effect of solvent temperature on (a) vapor phase temperature profile, (b) MEA vapor concentration, (c) water vapor concentration, (d) gas-phase CO2 mole
fraction and (e) MEA and water emissions.

is in agreement with the model predictions of Zhang et al., 2017 (Zhang
et al., 2017). Given that the reactive mass transfer process is kinetically
limited, it can be inferred that the reaction rates of MEA and CO2 are
stronger functions of MEA concentration (based on the conclusions of
previous sections) than the gas-phase temperature (in the temperature
range considered). Therefore, decreasing the MEA’s temperature does
not affect CO2 capture efficiency, while it can potentially reduce
aerosol-based amine emissions.
4. Conclusions
An ASPEN PLUS v10® model accounting for multi-component mass
and heat transfer, column hydrodynamics, solution thermodynamics,
and electrolytic-solution chemistry was coupled with a multi-

12
D. Dhanraj and P. Biswas
component discrete-sectional-based aerosol dynamics model. The
aerosol dynamics model was developed in FORTRAN 90, accounting for
condensation, and coagulation, to predict amine emissions during post-
combustion CO2 capture. Particle losses were considered by fitting the
experimental data Khakharia et al., 2013 (Khakharia et al., 2013). The
model’s outputs were validated using their experimentally measured
vapor- and aerosol-based amine emission data. The simulations showed
that neglecting particle losses resulted in an order of magnitude
over-prediction of amine-based solvent losses. The effect of including
coagulation in the model was also investigated and it was found that
including coagulation in the model resulted in reduced (~10%) emis­
sions. Furthermore, it was shown that amine emissions increased with
an increasing number concentration of inlet particles. Different particle
size distributions of inlet particles such as monodisperse, lognormal and
bi-lognormal were considered, and it was shown that distributions with
higher geometric standard deviations at the same number concentration
resulted in higher emissions. Furthermore, amine emissions were
reduced when the solvent’s concentrations and temperatures were
decreased. However, CO2 capture efficiency dropped for decreasing
solvent concentration but remained unchanged at lower temperatures. It
can be concluded that aerosol-based emissions scale proportionally to
vapor-based emissions, which increases with increasing temperatures in
the absorber. Given that decreasing solvent temperatures reduce amine
emissions without affecting CO2 capture efficiency, this efficient
approach can reduce the absorber’s temperature and thereby reduce
amine emissions during post-combustion CO2 capture.
CRediT authorship contribution statement
David Dhanraj: Conceptualization, Methodology, Software, Data
curation, Writing - original draft, Visualization, Investigation, Valida­
tion, Writing - review & editing. Pratim Biswas: Project administration,
Resources, Funding acquisition, Conceptualization, Methodology, Su­
pervision, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgements
The authors would like to thank the sponsors: U.S. D.O.E and Linde,
for funding this work through the grant entitled: Flue Gas Aerosol Pre­
treatment Technologies to Minimize PCC Solvent Losses (001365-
39847). Further, the discrete-sectional simulations were performed in
Engineering Cloud Cluster of Washington University in St. Louis.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.ijggc.2020.103179.
References
Wang, M., Lawal, A., Stephenson, P., Sidders, J., Ramshaw, C., 2011. Post-combustion
CO2 capture with chemical absorption: A state-of-the-art review. Chemical
Engineering Research and Design 89, 1609–1624.
Freund, P., 2003. Making deep reductions in CO2 emissions from coal-fired power plant
using capture and storage of CO2. Proc. Inst. Mech. Eng. A: J. Power Eng 217, 1–8.
He, J.-K., 2015. Objectives and strategies for energy revolution in the context of tackling
climate change. Advances in Climate Change Research 6, 101–107.
Majeed, H., Knuutila, H.K., Hillestad, M., Svendsen, H.F., 2017a. Characterization and
modelling of aerosol droplet in absorption columns. International Journal of
Greenhouse Gas Control 58, 114–126.
Davison, J., 2007. Performance and costs of power plants with capture and storage of
CO2. Energy 32, 1163–1176.
13
International Journal of Greenhouse Gas Control 103 (2020) 103179
Sreedhar, I., Nahar, T., Venugopal, A., Srinivas, B., 2017. Carbon capture by absorption –
Path covered and ahead. Renewable and Sustainable Energy Reviews 76,
1080–1107.
J. N. Pitts, D. Grosjean, K. Van Cauwenberghe, J. P. Schmid, and D. R. Fitz,
"Photooxidation of aliphatic amines under simulated atmospheric conditions:
formation of nitrosamines, nitramines, amides, and photochemical oxidant,"
Environmental Science & Technology, vol. 12, pp. 946-953, 1978/08/01 1978.
Khakharia, P., Brachert, L., Mertens, J., Huizinga, A., Schallert, B., Schaber, K., et al.,
2013. Investigation of aerosol based emission of MEA due to sulphuric acid aerosol
and soot in a Post Combustion CO2 Capture process. International Journal of
Greenhouse Gas Control 19, 138–144.
Harsha, S., Khakharia, P., Huizinga, A., Monteiro, J., Goetheer, E., Vlugt, T.J.H., 2019.
In-situ experimental investigation on the growth of aerosols along the absorption
column in post combustion carbon capture. International Journal of Greenhouse Gas
Control 85, 86–99.
Mimura, T., Nojo, T., Ishida, K., Nakashoji, H., Tanaka, H., Hirata, T., 2004. Amine
recovery method and apparatus and decarbonation apparatus having same. ed:
Google Patents.
Khakharia, P., Brachert, L., Mertens, J., Anderlohr, C., Huizinga, A., Fernandez, E.S.,
et al., 2015. Understanding aerosol based emissions in a Post Combustion CO2
Capture process: Parameter testing and mechanisms. International Journal of
Greenhouse Gas Control 34, 63–74.
Mertens, J., Brachert, L., Desagher, D., Schallert, B., Khakharia, P., Goetheer, E., 2014a.
Predicting Amine Mist Formation Based on Aerosol Number Concentration and Size
Measurements in Flue Gas. Energy Procedia 63, 893–901.
Mertens, J., Khakharia, P., Rogiers, P., Blondeau, J., Lepaumier, H., Goetheer, E., et al.,
2017. Prevention of Mist Formation in Amine Based Carbon Capture: Field Testing
Using a Wet ElectroStatic Precipitator (WESP) and a Gas-Gas Heater (GGH). Energy
Procedia 114, 987–999.
Mertens, J., Knudsen, J., Thielens, M.-L., Andersen, J., 2012. On-line monitoring and
controlling emissions in amine post combustion carbon capture: A field test.
International Journal of Greenhouse Gas Control 6, 2–11.
Moser, P., Schmidt, S., Sieder, G., Garcia, H., Ciattaglia, I., Klein, H., 2009. Enabling post
combustion capture optimization–The pilot plant project at Niederaussem. Energy
Procedia 1, 807–814.
Moser, P., Wiechers, G., Stahl, K., Stoffregen, T., Vorberg, G., Lozano, G.A., 2017. Solid
Particles as Nuclei for Aerosol Formation and Cause of Emissions – Results from the
Post-combustion Capture Pilot Plant at Niederaussem. Energy Procedia 114,
1000–1016.
Fulk, S.M., Rochelle, G.T., 2013. Modeling Aerosols in Amine-based CO2 Capture. Energy
Procedia 37, 1706–1719.
Brachert, L., Kochenburger, T., Schaber, K., 2013. Facing the Sulfuric Acid Aerosol
Problem in Flue Gas Cleaning: Pilot Plant Experiments and Simulation. Aerosol
Science and Technology 47, 1083–1091.
Majeed, H., Hillestad, M., Knuutila, H., Svendsen, H.F., 2018. Predicting aerosol size
distribution development in absorption columns. Chemical Engineering Science 192,
25–33.
Majeed, H., Knuutila, H., Hillestad, M., Svendsen, H.F., 2017b. Gas phase amine
depletion created by aerosol formation and growth. International Journal of
Greenhouse Gas Control 64, 212–222.
Majeed, H., Svendsen, H.F., 2018. Effect of water wash on mist and aerosol formation in
absorption column. Chemical Engineering Journal 333, 636–648.
Majeed, H., Knuutila, H., Hillestad, M., Svendsen, H.F., 2017c. Characterization and
modelling of aerosol droplet in absorption columns. International Journal of
Greenhouse Gas Control 58, 114–126.
Zhang, Y., Kang, J.-L., Fulk, S., Rochelle, G., 2017. Modeling Amine Aerosol Growth at
Realistic Pilot Plant Conditions. Energy Procedia 114, 1045–1060.
Schneider, R., Kenig, E.Y., Górak, A., 1999. Dynamic Modelling of Reactive Absorption
with the Maxwell-Stefan Approach. Chemical Engineering Research and Design 77,
633–638.
Zhang, Y., Chen, C.-C., 2013. Modeling CO2 Absorption and Desorption by Aqueous
Monoethanolamine Solution with Aspen Rate-based Model. Energy Procedia 37,
1584–1596.
Friedlander, S.K., 1977. Smoke, dust and haze: Fundamentals of aerosol behavior. Wiley-
Interscience, New York, 333 p., 1977.
Wu, C.-Y., Biswas, P., 1998. Study of Numerical Diffusion in a Discrete-Sectional Model
and Its Application to Aerosol Dynamics Simulation. Aerosol Science and
Technology 29, 359–378.
Wu, J.J., Flagan, R.C., 1988. A discrete-sectional solution to the aerosol dynamic
equation. Journal of Colloid and interface Science 123, 339–352.
Taylor, R., Krishna, R., 1993. Multicomponent mass transfer, vol. 2. John Wiley & Sons.
Billet, R., Schultes, M., 1999. Prediction of Mass Transfer Columns with Dumped and
Arranged Packings. Chemical Engineering Research and Design 77, 498–504.
Brian Hanley, C.-C.C., 2011. New Mass-Transfer Correlations for Packed Towers. AIChE
58, 132–152.
Chilton, T.H., Colburn, A.P., 1934. Mass transfer (absorption) coefficients prediction
from data on heat transfer and fluid friction. Industrial & engineering chemistry 26,
1183–1187.
Biswas, P., Wu, C.Y., Zachariah, M.R., McMillin, B., 1997. Characterization of iron oxide-
silica nanocomposites in flames: Part II. Comparison of discrete-sectional model
predictions to experimental data. Journal of Materials Research 12, 714–723.
D. Dhanraj and P. Biswas International Journal of Greenhouse Gas Control 103 (2020) 103179

Pratsinis, S.E., 1988. Simultaneous nucleation, condensation, and coagulation in aerosol Mertens, J., Brachert, D., Desagher, M.L., Thielens, P., Khakharia, E., Goetheer,
reactors. Journal of Colloid and Interface Science 124, 416–427. Schaber, K., 2014. ELPI+ measurements of aerosol growth in an amine absorption
Fulk, S.M., 2016. Measuring and Modeling Aerosols in Carbon Dioxide Capture by column. International Journal of Greenhouse Gas Control 23 (2014), 44–50.
Aqueous Amines. PhD Thesis. The University of Texas at Austin.

14