Mathematical Modeling of a Moving Bed Reactor for

Post-Combustion CO2 Capture

Hosoo Kim1, David C. Miller2, Srinivasarao Modekurti3, Benjamin Omell3, Debangsu Bhattacharyya3,*,
Stephen E. Zitney2,3

1
National Energy Technology Laboratory, 3610 Collins Ferry Road, Morgantown, West Virginia, 26507, United
States

2
National Energy Technology Laboratory, 626 Cochrans Mill Road, Pittsburgh, Pennsylvania, 15236, United States

3
Department of Chemical Engineering, West Virginia University, Morgantown, West Virginia 26506, United States

ABSTRACT

A mathematical model for a moving bed reactor with embedded heat exchanger has been
developed for application to solid sorbent-based capture of carbon dioxide (CO2) from flue gas
emitted by coal-fired power plants. The reactor model is one-dimensional, non-isothermal, and
pressure-driven is flexible enough for use with both adsorption and desorption operations. The
two-phase (gas and solids) model includes rigorous kinetics and heat and mass transfer between
the two phases. Flow characteristics of the gas and solids in the moving bed are obtained by
analogy with correlations for fixed and fluidized bed systems. From the steady-state perspective,
this work presents the impact of key design variables that can be used for optimization. From the
dynamic perspective, the paper shows transient profiles of key outputs that should be taken into
account while designing an effective control system.

Keywords: CO2 capture, Solid Sorbent, Moving Bed, Dynamic Model, Regenerator

*Corresponding Author: Tel: +1-3042939335, Fax: +1-3042934139

E-mail Address: [email protected]

INTRODUCTION

Fossil fuels currently supply a significant portion of the world’s energy needs, and their
utilization is the major source of the anthropogenic CO2 emissions1. Within the U.S., existing
coal-fired power plants annually emit approximately two billion tons of CO2 which is about two-
thirds of the total emissions from the U.S. power generation sector2. Therefore, it is critical to
develop cost-effective capture technologies to mitigate this environmental problem. The U.S.
Department of Energy (DOE) has set a goal to demonstrate technology capable of capturing 90%
of the CO2 with less than 30% increase in cost of electricity for existing coal power plants2.
Ciferno et al.2 described the advantages and disadvantages of the three main options for
capturing CO2 from fossil energy plants: post-combustion capture, pre-combustion capture, and
oxy-combustion. Among them, post-combustion capture has the greatest near-term potential for
reducing CO2 emission because it can be retrofitted with relative ease to the existing fleet of
fossil-fired power plants. Existing commercially available solvent-based technologies (i.e.,
aqueous amines) have the advantages of fast kinetics and favorable equilibrium in the absorption
reactor; however, they give rise to unacceptably high capital costs and operating costs resulting
from the energy requirements to regenerate the solvent. In addition, aqueous amines have several
other drawbacks, such as thermal and oxidative degradation of the solvents, solvent loss, and
corrosion problems3.

For post-combustion capture, solid sorbent-based adsorption processes have been reported to
have certain technical and economic advantages in comparison to conventional aqueous amine
solvent-based absorption processes4–7. Several potential candidate solid sorbent materials have
been shown to exhibit higher CO2 carrying capacity and lower regeneration duty than traditional
solvents3. Sorbents are typically regenerated by two techniques: pressure swing adsorption (PSA)
and temperature swing adsorption (TSA). In TSA, the sorbent is regenerated by raising the
temperature, whereas in PSA, the sorbent is regenerated by lowering the pressure. Vacuum
swing adsorption (VSA) is a special case of PSA. Sorbent regeneration can be enhanced by
purging the contactor with a non-adsorbing inert gas or displacing the adsorbed species using
other competitive adsorbable species8.
Efficient, flexible, and easy-to-operate gas-solid contactors are absolutely necessary for
commercial application of solid sorbent-based CO2 capture processes. Three types of beds are
typically used for gas-solid contacting processes: fixed beds, moving beds, and fluidized beds7.
Fixed bed contactors require more complicated operation and control procedures in which the
bed is alternatively saturated and regenerated in a cyclical manner. Diffusion and conduction are
the dominant heat and mass transfer mechanisms in a fixed bed reactor, which results in
significant resistances compared with other contactors. For post-combustion CO2 capture
applications, a large volume of flue gas must be treated. For a nominal 500 MW power plant, this
is greater than 34,000 m3/min4. Due to limitations on gas velocity to keep the bed from
fluidizing, a large number of fixed beds would be needed adding to the capital cost. Furthermore,
if PSA or VSA is considered for sorbent regeneration, it would likely add to more number of
beds because of short cycle time for a given bed due to the large volume of CO2 that needs to be
handled. While thermal regeneration has the advantage of lowering the solids CO2 loading to a
low value resulting in some increase in the cycle time, TSA has significant disadvantage due to
the large thermal mass of the solids and the reactor shells for the fixed beds. Furthermore, cycle
time is still expected to be relatively short because of large volume of CO2. Therefore, in
addition to the complexity of cyclic operation, fixed beds seem infeasible for post-combustion
CO2 capture. In bubbling as well as circulating fluidized bed reactors, the solids and gas are well
mixed, which results in excellent heat and mass transfer characteristics; however, this can also
cause significant attrition of sorbent particles and erosion of the reactor vessel and internals.
Moreover, a well-mixed system has the disadvantage for CO2 capture applications where true
countercurrent contacting can be beneficial due to a larger driving force resulting in higher CO2
loading in the adsorber and lower CO2 loading exiting the regenerator. Moving beds can provide
nearly countercurrent contacting and therefore have strong potential for post-combustion CO2
capture applications.

In moving bed reactors, fresh solids can be fed from the top while a gas stream that is much
lower in flowrate in comparison to a fluidized bed operating under similar solids flowrate, is fed
from the bottom. These solid particles slowly move down the reactor and are withdrawn from the
bottom of the bed. In addition to providing near counter-current contacting, this approach can
also reduce attrition of the particles compared with fluidized beds.
To achieve counter-current operation in a moving bed, the gas velocity must be kept low enough
to prevent the solids from becoming fluidized. This presents a significant challenge for the
utilization of moving bed reactors for CO2 capture due to the large amount of nitrogen present in
the flue gas. For such applications, it is anticipated that the diameter of a moving bed adsorber
would be impractically large to keep the superficial gas velocity in the moving bed regime. For
example, a 650 MW power plant would require approximately 90 adsorbers in parallel, each with
a diameter of 9 m, to keep the superficial gas velocity below that of the minimum fluidization
velocity. Due to this anticipated limitation, a moving bed reactor is specifically suitable for
regenerator.

Gas velocity is not the only limitation for practical industrial application of moving beds. It can
be difficult to maintain a desired rate of solids flow. Further, the beds typically have a low radial
heat conductivity9. To address these design and operational challenges, a rigorous process model
of a moving bed contactor that includes heat and mass transfer is required. While there are a
large number of models for fixed and fluidized beds, the current literature provides little
information on mathematical models for moving bed reactors, especially for regenerating solid
sorbents used for CO2 capture, where a large amount of heat transfer is needed. While the
literature on moving beds is limited for CO2 capture applications, numerous other applications of
moving beds have been studied.

Parisi and Laborde10 have modeled a gas-solid countercurrent moving bed furnace in which the
reduction of iron ore pellets is performed at high temperature. They have simultaneously solved
a set of ordinary differential equations for mass and energy balances. Valipour and Saboohi11
have presented a general moving bed model for porous pellets to simulate high temperature
noncatalytic gas–solid reactions. They have taken into account the external mass transfer, the
internal diffusion transport through the pores, the chemical reactions, the heat of reaction and
heat transfer due to effective conduction in the entire solid matrix. Dussoubs et al.12 have
developed a similar model for the production of uranium tetrafluoride. Simultaneous heat and
mass transfer between air and soybean seeds in a countercurrent moving bed dryer has been
studied by Lacerda et al.13. Lira et al.14 have analyzed the effect of air velocity on drying in co-
current moving beds. Gomes15 et al. have presented a moving bed adsorption reactor model with
a detailed particle model accounting for intraparticle radial profiles to analyze the conversion of
sucrose to glucose and fructose and subsequent separation. There is hardly any model model in
the open literature for moving beds with an embedded heat exchanger for providing the large
amount of heat required to maintain effective solids regeneration.

While analyzing the applicability of moving beds for CO2 capture and sorbent regeneration,
Yang and Hoffman7 have noted that a moving bed may not be effective because of poor heat
transfer and a very high pressure drop when small sorbent particles are used. However, recent
advances in moving bed design offer a solution to these problems making moving bed contactors
an option for solid sorbent-based carbon capture. Knaebel16 have suggested a novel moving bed
system for TSA-based carbon capture, in which attrition can be reduced by introducing
perforated plates at relatively short vertical intervals and carefully choosing the dimensions. In
addition, Knaebel and Knaebel17 have showed that the pressure drop of the flue gas could be
adjusted by varying internal dimensions, while simultaneously minimizing the plant footprint.

To analyze the performance of a counter-current moving bed for commercial-scale CO2 capture
applications, the DOE’s Carbon Capture Simulation Initiative (CCSI) is using modeling and
simulation to investigate tradeoffs between various design alternatives, process operating
conditions, and sorbent types. CCSI is a partnership among DOE laboratories, industry, and
universities that is developing, demonstrating and deploying a suite of computations tools to be
used by industry to accelerate the deployment of post-combustion carbon capture technologies in
power generation applications18–20. Recently, CCSI researchers have developed a general one-
dimensional, three region model for a solid sorbent-based CO2 capture process using a bubbling
fluidized bed adsorber reactor2122. The CCSI team has also studied the tradeoffs for various types
of contactors for use in solid sorbent-based CO2 capture processes19.

In this study, a rigorous, one-dimensional, non-isothermal model of a moving bed reactor stage is
presented with application to the regeneration of solid sorbents used in a post-combustion CO2
capture process. The model can be used in both a steady-state mode for design and optimization
as well as a pressure-driven dynamic mode for development of operational and control strategies.
Additionally, the model is flexible enough to be applied in adsorption applications along with
regeneration. Due to the current lack of literature in this area, the moving bed model has been
developed by drawing analogy to the fixed and fluidized bed processes as appropriate. One more
feature of this work is consideration of an embedded heat exchanger that is essential for TSA-
based regeneration applications because the heat of regeneration is quite large. The model has
been used to study the impacts of key design variables. Furthermore, transient response of the
bed due to disturbances that are typical in power plants are presented.

MODEL DEVELOPMENT

General Model Description

In a counter-current moving bed reactor, a continuous stream of solid sorbent particles moves
downward under the influence of gravity, contacting a rising gas stream in a counter-current
manner (Figure 1). Temperature swing desorption of CO2 is considered in this work. A vertical
shell and tube reactor has been considered where gas-solid contacting takes place in the shell
side and steam is used to control the temperature on the tube side. The solid particles pass over a
distributor at the top and fall through a series of perforated trays. The perforated trays are
assumed to ensure even distribution of solid particles across the cross section and to help control
the residence time. Stripping steam can be introduced through the distributor at the bottom of the
reactor to decrease the partial pressure of CO2 throughout the bed.

The two-phase model of the shell side of the reactor consists of mass and energy balance
equations. Mass transfer between gas and solids occurs by reaction. Heat exchange between the
two phases and the immersed tubes is represented by general heat transfer equations. Gas flow is
considered to be convective with axial dispersion. Solids flow is approximated as uniform with
constant velocity. Steam slightly above its saturation temperature is considered to be the heating
medium.

Flow characteristics of gases and solids in moving beds have not been widely studied, so most
coefficients were derived by analogy with correlations for fixed and fluidized bed systems. The
main assumptions of the moving bed reactor model used in this study are as follows:
• The system is represented by a set of one-dimensional partial differential
equations (PDEs) in the axial direction.
o The reactor is fully mixed in the radial direction.
 o Particles are uniformly dispersed through the reactor with constant
voidage.
o Imaginary internals (e.g. plates) are assumed to provide the uniform solids
distribution.
 Flow of the particles through the plates is unrestricted at all times.
 Pressure drop across plates is negligible.
• Each particle has a uniform temperature (i.e., there is no temperature profile
within particles).
• Particle attrition is negligible.

The model equations along with the boundary conditions are presented in the following section.

Mass Balance

Flow of gas and solids in a moving bed can be generally represented by the following plug flow
model with axial dispersion.

In the gas phase, the mass balance is written as

𝜕𝜕𝐶𝐶𝑖𝑖 𝜕𝜕 2 𝐶𝐶𝑖𝑖 𝜕𝜕�𝑣𝑣𝑔𝑔 𝐶𝐶𝑖𝑖 � 𝜕𝜕𝑄𝑄𝑗𝑗

𝜀𝜀𝑏𝑏 = 𝜀𝜀𝑏𝑏 𝐷𝐷𝑧𝑧 2 − + (1 − 𝜀𝜀𝑏𝑏 ) � 𝜈𝜈𝑖𝑖,𝑗𝑗 (1)
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
𝑗𝑗

where 𝜀𝜀𝑏𝑏 is an average voidage through the reactor and 𝐷𝐷𝑧𝑧 is an effective axial dispersion
coefficient that lumps the effects of all mechanisms that contribute to axial mixing. 𝐶𝐶𝑖𝑖 is the mole
concentration for component 𝑖𝑖 in gas phase, 𝑣𝑣𝑔𝑔 is the superficial velocity of gas, 𝑡𝑡 and 𝑧𝑧 are the
time and the reactor axial coordinates, respectively, 𝜈𝜈𝑖𝑖,𝑗𝑗 is the stoichiometric coefficient for gas
species 𝑖𝑖 with respect to the adsorption reaction for adsorbed species 𝑗𝑗, 𝜌𝜌𝑠𝑠 is the mass density for
solid particle, and 𝑄𝑄𝑗𝑗 is the molar amount of adsorbed species j that is transferred between gas
and solids.

When the adsorption within porous particles is sufficiently strong and the Reynolds number is
low, the Peclet number, Pe', for calculating an effective axial dispersion coefficient is obtained
using Equation 28.
 1 𝐷𝐷𝑧𝑧 20 1
= = + (2)
𝑃𝑃𝑃𝑃 ′ 𝑣𝑣𝑑𝑑𝑝𝑝 𝑅𝑅𝑅𝑅𝑃𝑃 𝑆𝑆𝑆𝑆𝑃𝑃 2

where v is the superficial velocity of the particle and Rep and Scp are the particle Reynolds
number and Schmidt number, respectively.

In the solid phase, the mass balance is expressed as

𝜕𝜕𝑤𝑤𝑗𝑗 𝜕𝜕𝑤𝑤𝑗𝑗 𝜕𝜕𝑄𝑄𝑗𝑗

(1 − 𝜀𝜀𝑏𝑏 )𝜌𝜌𝑠𝑠 = 𝐽𝐽𝑠𝑠 − (1 − 𝜀𝜀𝑏𝑏 ) (3)
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕

where 𝑤𝑤𝑗𝑗 is the mole concentration for adsorbed species j per unit mass of solid particle and 𝐽𝐽𝑠𝑠
denotes the axial solid flux that is assumed as constant.

Reactions and Mass Transfer

Various solid sorbents have been studied as potential candidates for CO2 capture applications,
including zeolites, carbon-based sorbents, carbonates, and supported amines23. In order to
demonstrate and test the current moving bed model, a solid sorbent developed by the U.S. DOE’s
National Energy Technology Laboratory (NETL) 24, has been selected. This sorbent is a mixture
of polyethyleneimine and aminosilanes impregnated into the mesoporous structure of a silica
substrate. A simple lumped-parameter kinetic and equilibrium model was developed for this
sorbent by Lee et al.24. The model assumes that the adsorption of CO2 and water occurs through
three simple reactions, resulting in the formation of physisorbed water (Equation 4), bicarbonate
ions (Equation 5) and carbamate ions (Equation 6), respectively.

𝜕𝜕𝑄𝑄𝐻𝐻2 𝑂𝑂 1
= 𝑘𝑘𝐻𝐻2 𝑂𝑂 �𝑐𝑐𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠,𝐻𝐻2 𝑂𝑂 𝑅𝑅𝑇𝑇𝑔𝑔 − 𝑉𝑉 � (4)
𝜕𝜕𝜕𝜕 𝐾𝐾𝐻𝐻2 𝑂𝑂 𝑚𝑚,𝐻𝐻2 𝑂𝑂

𝜕𝜕𝑄𝑄𝐵𝐵𝐵𝐵𝐵𝐵 2𝑉𝑉𝑚𝑚,𝐶𝐶𝐶𝐶𝐶𝐶 + 𝑉𝑉𝑚𝑚,𝐵𝐵𝐵𝐵𝐵𝐵 1 𝑉𝑉𝑚𝑚,𝐶𝐶𝐶𝐶𝐶𝐶 + 𝑉𝑉𝑚𝑚,𝐵𝐵𝐵𝐵𝐵𝐵

= 𝑘𝑘𝐵𝐵𝐵𝐵𝐵𝐵 ��1 − � 𝑉𝑉𝑚𝑚,𝐻𝐻2 𝑂𝑂 𝑐𝑐𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠,𝐶𝐶𝐶𝐶2 𝑅𝑅𝑇𝑇𝑔𝑔 − 𝑉𝑉 � �� (5)
𝜕𝜕𝜕𝜕 𝑛𝑛𝑣𝑣 𝐾𝐾𝐵𝐵𝐵𝐵𝐵𝐵 𝑚𝑚,𝐵𝐵𝐵𝐵𝐵𝐵 𝑛𝑛𝑣𝑣

𝜕𝜕𝑄𝑄𝐶𝐶𝐶𝐶𝐶𝐶 2𝑉𝑉𝑚𝑚,𝐶𝐶𝐶𝐶𝐶𝐶 + 𝑉𝑉𝑚𝑚,𝐵𝐵𝐵𝐵𝐵𝐵 2 𝑚𝑚 1 𝑉𝑉𝑚𝑚,𝐶𝐶𝐶𝐶𝐶𝐶 𝑉𝑉𝑚𝑚,𝐶𝐶𝐶𝐶𝐶𝐶 + 𝑉𝑉𝑚𝑚,𝐵𝐵𝐵𝐵𝐵𝐵

= 𝑘𝑘𝐶𝐶𝐶𝐶𝐶𝐶 ��1 − � �𝑐𝑐𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠,𝐶𝐶𝐶𝐶2 𝑅𝑅𝑇𝑇𝑔𝑔 � − � �� (6)
𝜕𝜕𝜕𝜕 𝑛𝑛𝑣𝑣 𝐾𝐾𝐶𝐶𝐶𝐶𝐶𝐶 𝑛𝑛𝑣𝑣 𝑛𝑛𝑣𝑣

where 𝑘𝑘𝑗𝑗 is the rate constant for solid species j, 𝑐𝑐𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠,𝑖𝑖 is the concentration of gaseous species i
on the surface of the sorbent, 𝑛𝑛𝑣𝑣 is the concentration of available amine sites in sorbent, 𝑚𝑚 is an
empirical non-ideality coefficient, 𝑇𝑇𝑔𝑔 is the temperature of the gas, 𝐾𝐾𝑗𝑗 is the equilibrium constant,
and 𝑉𝑉𝑚𝑚,𝑗𝑗 is the concentration of the absorbed species.

Lee et al.24 determined equilibrium and rate constants (Equations 7 and 8) from experimental
thermogravimetric data. Using additional data, the recently determined improved values of the
kinetic parameters are obtained as shown in Table 1.

−𝐸𝐸𝑗𝑗
𝑘𝑘𝑗𝑗 = 𝐴𝐴𝑗𝑗 (𝑇𝑇𝑠𝑠 + 273.15)𝑒𝑒𝑒𝑒𝑒𝑒 � � (7)
𝑅𝑅𝑇𝑇𝑠𝑠

∆𝑆𝑆𝑗𝑗 −∆𝐻𝐻𝑗𝑗
𝐾𝐾𝑗𝑗 = 𝑒𝑒𝑒𝑒𝑒𝑒 � � 𝑒𝑒𝑒𝑒𝑒𝑒 � � /(𝑃𝑃 × 105 ) (8)
𝑅𝑅 𝑅𝑅𝑇𝑇𝑠𝑠

where 𝐴𝐴𝑗𝑗 is the Arrhenius pre-exponential factor for reaction j, 𝑇𝑇𝑠𝑠 is the temperature of the solid
phase, 𝐸𝐸𝑗𝑗 is the activation energy, ∆𝑆𝑆𝑗𝑗 is the entropy of reaction, ∆𝐻𝐻𝑗𝑗 is the enthalpy of reaction,
and 𝑃𝑃 is the pressure of the gas.

The reaction enthalpy of each reaction is assumed to be constant within the model.

The mass transfer rates, 𝑅𝑅𝑥𝑥𝑥𝑥𝑥𝑥,𝑖𝑖 , are calculated using the following equations:

10−4 𝐷𝐷𝑖𝑖
𝑘𝑘𝑓𝑓𝑓𝑓,𝑖𝑖 = (9)
𝑑𝑑𝑝𝑝 𝑆𝑆ℎ

6
𝑅𝑅𝑥𝑥𝑥𝑥𝑥𝑥,𝑖𝑖 = −𝑘𝑘𝑓𝑓𝑓𝑓,𝑖𝑖 � � �𝑐𝑐𝑖𝑖 − 𝑐𝑐𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠,𝑖𝑖 � (10)
𝑑𝑑𝑝𝑝

𝜕𝜕𝑄𝑄𝐶𝐶𝐶𝐶𝐶𝐶 𝜕𝜕𝑄𝑄𝐵𝐵𝐵𝐵𝐵𝐵
𝑅𝑅𝑥𝑥𝑥𝑥𝑥𝑥,𝐶𝐶𝐶𝐶2 = + (11)
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕

𝜕𝜕𝑄𝑄𝐻𝐻2 𝑂𝑂
𝑅𝑅𝑥𝑥𝑥𝑥𝑥𝑥,𝐻𝐻2 𝑂𝑂 = (12)
𝜕𝜕𝜕𝜕

where 𝑘𝑘𝑓𝑓𝑓𝑓,𝑖𝑖 is the mass transfer coefficient which is calculated based upon the Sherwood
number, 𝑆𝑆ℎ, d p is the diameter of a solid particle, and 𝑐𝑐𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠,𝑖𝑖 is the concentration of the gas on

the solid surface which determines the concentration gradient that drives the mass transfer rate.
In Equation 9, Di denotes the diffusivity of the species i in a mixture and is calculated using
Equations 13 and 1425:
 𝑐𝑐𝑖𝑖1 −𝑖𝑖2
𝑎𝑎𝑖𝑖1 −𝑖𝑖2 𝑇𝑇𝑔𝑔 𝑏𝑏𝑖𝑖1−𝑖𝑖2 � �
𝐷𝐷𝑖𝑖1 −𝑖𝑖2 = 𝑒𝑒 𝑇𝑇𝑔𝑔 (13)
𝑃𝑃

1 − 𝑧𝑧𝑖𝑖
𝐷𝐷𝑖𝑖 = 𝑧𝑧 𝑧𝑧 𝑧𝑧𝑖𝑖 ; 𝑁𝑁 = 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 𝑜𝑜𝑜𝑜 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
𝑖𝑖1 𝑖𝑖 (14)
+ 2 … + 𝑁𝑁−1
𝐷𝐷𝑖𝑖−𝑖𝑖1 𝐷𝐷𝑖𝑖−𝑖𝑖2 𝐷𝐷𝑖𝑖−𝑖𝑖𝑁𝑁−1

where N is the number of gaseous species. 𝐷𝐷𝑖𝑖1 −𝑖𝑖2 is the binary diffusivity between two species
and is considered symmetric, i.e. 𝐷𝐷𝑖𝑖1 −𝑖𝑖2 = 𝐷𝐷𝑖𝑖2 −𝑖𝑖1 . The values for the binary diffusivity
parameters; 𝑎𝑎𝑖𝑖1 −𝑖𝑖2 , 𝑏𝑏𝑖𝑖1 −𝑖𝑖2 and 𝑐𝑐𝑖𝑖1 −𝑖𝑖2 are given in Table 1.

Pressure Drop

Yoon and Kunii26 suggested that the Ergun equation, a well-known correlation of pressure drop
for fixed beds, can be applied to moving beds using the slip velocity, 𝜀𝜀∆𝑣𝑣 , between the
interstitial fluid velocity and the interstitial particle velocity, instead of the superficial fluid
velocity, 𝑣𝑣𝑔𝑔 . Absolute slip velocity is used in the following equation with positive fluid velocity
being upward and positive particle velocity being downward.

𝜕𝜕𝜕𝜕 150 × 10−5 𝜇𝜇𝑔𝑔 (1 − 𝜀𝜀𝑏𝑏 )2 𝑣𝑣𝑔𝑔 𝑣𝑣𝑠𝑠 1.75 × 10−5 𝜌𝜌𝑔𝑔 (1 − 𝜀𝜀𝑏𝑏 ) 2 𝑣𝑣𝑔𝑔 𝑣𝑣𝑠𝑠 2
= −� 2 𝜀𝜀𝑏𝑏 ( + ) + 3 𝜀𝜀𝑏𝑏 ( + ) � (15)
𝜕𝜕𝜕𝜕 �𝑑𝑑 𝜓𝜓� 𝜀𝜀 3 𝜀𝜀𝑏𝑏 1 − 𝜀𝜀𝑏𝑏 (𝑑𝑑𝑝𝑝 𝜓𝜓)𝜀𝜀𝑏𝑏 𝜀𝜀𝑏𝑏 1 − 𝜀𝜀𝑏𝑏
𝑝𝑝 𝑏𝑏

where 𝜓𝜓 is the particle shape factor, 𝜌𝜌𝑔𝑔 is the mass density of gas, and 𝑀𝑀𝑀𝑀𝑔𝑔 is the average
molecular weight of gas.

Energy Balance

The energy balance equation for the gas phase is

 𝜕𝜕𝑇𝑇𝑔𝑔 𝜕𝜕 2 𝑇𝑇𝑔𝑔 𝜕𝜕𝑇𝑇𝑔𝑔 𝜕𝜕𝑣𝑣𝑔𝑔
𝐶𝐶𝑝𝑝,𝑔𝑔 𝜀𝜀𝑏𝑏 𝜌𝜌𝑔𝑔 = 𝜀𝜀𝑏𝑏 𝑘𝑘𝑔𝑔 − 𝐶𝐶𝑝𝑝,𝑔𝑔 𝑣𝑣𝑔𝑔 𝜌𝜌𝑔𝑔 − 𝑃𝑃 + (1 − 𝜀𝜀𝑏𝑏 )ℎ𝑔𝑔𝑔𝑔 𝑎𝑎𝑠𝑠 �𝑇𝑇𝑠𝑠 − 𝑇𝑇𝑔𝑔 � (16)
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 2 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕

where 𝑇𝑇𝑔𝑔 and 𝑇𝑇𝑠𝑠 denote the temperature of gas and solid, respectively, ℎ𝑔𝑔𝑔𝑔 is the heat transfer
coefficient between gas and solid, and 𝑎𝑎𝑠𝑠 is the surface area of the solids.

The energy balance for the solid phase is expressed as

𝜕𝜕𝑇𝑇𝑠𝑠 𝜕𝜕𝑇𝑇𝑠𝑠 𝜋𝜋𝐷𝐷𝑡𝑡 𝑁𝑁𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

(1 − 𝜀𝜀𝑏𝑏 )𝜌𝜌𝑠𝑠 𝐶𝐶𝑝𝑝,𝑠𝑠 = 𝐶𝐶𝑝𝑝,𝑠𝑠 𝐽𝐽𝑠𝑠 − (1 − 𝜀𝜀𝑏𝑏 )ℎ𝑔𝑔𝑔𝑔 𝑎𝑎𝑠𝑠 �𝑇𝑇𝑠𝑠 − 𝑇𝑇𝑔𝑔 � + ℎ𝑤𝑤𝑤𝑤 (𝑇𝑇𝑤𝑤 − 𝑇𝑇𝑠𝑠 )
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝐴𝐴𝑏𝑏𝑏𝑏𝑏𝑏
𝜕𝜕𝑄𝑄𝑗𝑗 (17)
+ � ∆𝐻𝐻𝑟𝑟𝑟𝑟𝑟𝑟,𝑗𝑗 𝜌𝜌𝑠𝑠
𝜕𝜕𝜕𝜕
𝑗𝑗

where ∆𝐻𝐻𝑟𝑟𝑟𝑟𝑟𝑟,𝑗𝑗 is the enthalpy of 𝑗𝑗th reaction and 𝑇𝑇𝑔𝑔 denotes the temperature of the tube wall.

Heat transfer across the tube wall yields the following equation:

𝜋𝜋(𝐷𝐷𝑡𝑡 −2𝑤𝑤𝑡𝑡ℎ𝑥𝑥 )𝑁𝑁𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ℎ𝑤𝑤𝑤𝑤𝑤𝑤 (𝑇𝑇𝑤𝑤 − 𝑇𝑇𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ) − 𝜋𝜋𝐷𝐷𝑡𝑡 𝑁𝑁𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ℎ𝑡𝑡 (𝑇𝑇𝑤𝑤 − 𝑇𝑇𝑠𝑠 ) = 0 (18)

The energy balance equation for the tube side of the embedded heat exchanger can be written as

𝜕𝜕𝑣𝑣𝑓𝑓
𝑣𝑣𝑡𝑡 ∆𝐻𝐻𝑣𝑣𝑣𝑣𝑣𝑣 − 𝜋𝜋(𝐷𝐷𝑡𝑡 −2𝑤𝑤𝑡𝑡ℎ𝑥𝑥 )𝑁𝑁𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ℎ𝑤𝑤𝑤𝑤𝑤𝑤 (𝑇𝑇𝑡𝑡 − 𝑇𝑇𝑤𝑤 ) = 0 (19)
𝜕𝜕𝜕𝜕

where ℎ𝑡𝑡 denotes the heat transfer coefficient between tube-side wall and solid, ℎ𝑤𝑤𝑤𝑤𝑤𝑤 denotes the
tube-side internal heat transfer coefficient, and 𝑣𝑣𝑡𝑡 denotes the velocity of tube fluid. ∆𝐻𝐻𝑣𝑣𝑣𝑣𝑣𝑣 and
𝑣𝑣𝑓𝑓 denote the enthalpy change of steam and the degree of saturation of steam, respectively.

Heat transfer coefficients

The heat transfer coefficients between the wall and solid and gas phases are calculated by using
the modified surface renewal theory. The gas-to-solid ℎ𝑔𝑔𝑔𝑔 , wall-to-gas ℎ𝑤𝑤𝑤𝑤 , wall-to-solid ℎ𝑤𝑤𝑤𝑤 ,
and steam-wall ℎ𝑡𝑡 heat transfer coefficients are obtained from the following correlations8,27,28:

ℎ𝑤𝑤𝑤𝑤 𝑑𝑑𝑝𝑝 1 1 (20)

𝑁𝑁𝑁𝑁 = = 0.009𝐴𝐴𝐴𝐴 �2 𝑃𝑃𝑃𝑃 �3
𝑘𝑘𝑔𝑔
 ℎ𝑔𝑔𝑔𝑔 𝑑𝑑𝑝𝑝 1 3 (21)
𝑁𝑁𝑁𝑁𝑝𝑝 = = 2.0 + 1.1𝑃𝑃𝑃𝑃 �3 𝑅𝑅𝑅𝑅𝑝𝑝 �5
𝑘𝑘𝑔𝑔

𝑘𝑘
0.46(1−𝑒𝑒) (22)
𝑘𝑘𝑝𝑝𝑝𝑝 = (3.58 − 2.5𝑒𝑒)𝑘𝑘𝑔𝑔 � 𝑠𝑠�𝑘𝑘 �
𝑔𝑔

1� (23)
1 − 𝑒𝑒 2
ℎ𝑤𝑤𝑤𝑤 = 2 �𝑘𝑘𝑝𝑝𝑝𝑝 𝜌𝜌𝑠𝑠 𝐶𝐶𝑝𝑝,𝑠𝑠 �
𝜋𝜋𝜋𝜋

ℎ𝑡𝑡 = 𝑓𝑓𝑏𝑏 ℎ𝑤𝑤𝑤𝑤 + (1 − 𝑓𝑓𝑏𝑏 ) ℎ𝑤𝑤𝑤𝑤 (24)

where 𝑘𝑘𝑔𝑔 and 𝑘𝑘𝑠𝑠 denote the gas and sorbent thermal conductivities, respectively, 𝐴𝐴𝐴𝐴 is the
Archimedes number, 𝑃𝑃𝑃𝑃 is the Prandtl number, 𝑘𝑘𝑝𝑝𝑝𝑝 is the bed’s thermal conductivity at
minimum fluidization velocity, 𝑓𝑓𝑏𝑏 is the fraction of time that the heat exchanger surface contacts
the solids, and τ is the average residence time of the solids contacting the heat exchanger surface.
The parameters 𝑓𝑓𝑏𝑏 and τ are given by the following relations

0.14
2
(𝑓𝑓𝑛𝑛 − 𝑎𝑎ℎ )2
𝑓𝑓𝑏𝑏 = 0.33 �𝑣𝑣𝑚𝑚𝑚𝑚 � � � (25)
9.8𝑑𝑑𝑝𝑝

𝑣𝑣𝑔𝑔
𝑓𝑓𝑛𝑛 = (26)
𝑣𝑣𝑚𝑚𝑚𝑚

0.14 0.225
9.8𝑑𝑑𝑝𝑝 𝑑𝑑𝑝𝑝
𝜏𝜏 = 0.44 �� 2 � � � � (27)
𝑣𝑣𝑚𝑚𝑚𝑚 (𝑓𝑓𝑛𝑛 − 𝑎𝑎ℎ )2 𝐷𝐷𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

Auxiliary Equations and Conditions

Maximum gas velocity

The behavior of a falling particle in the moving bed can be estimated by analogy to a fluidized
bed. For maintaining the bed in the moving bed region, the internal gas velocity through the bed
should be less than the minimum fluidization velocity, vmf , given by the equation below from

Kunii and Levenspiel29

 1.75  d p vmf ρ g ,0  150 (1 − ε mf ) d p vmf ρ g ,0 d p ρ g ,0 ( ρ s − ρ g ,0 ) g
2 3

  + = (28)
φsε mf 3  µ g ,0  φs 2ε mf 3 µ g ,0 µ g ,0 2

where 𝑔𝑔 is the gravitational acceleration and φs is the sphericity of the particle. Therefore, the
following constraint should be satisfied along the axial coordinates

𝑣𝑣𝑔𝑔 < 𝑣𝑣𝑚𝑚𝑚𝑚 (29)

Tube pitch

When the tubes are too close to each other (𝑃𝑃𝑡𝑡 /𝑑𝑑𝑡𝑡 less than 1.25), the header plate (tube sheet)
becomes too weak for proper rolling of the tubes and causes leaky joints30. In Equation 30, 𝑃𝑃𝑡𝑡 is
the tube pitch and 𝐷𝐷𝑥𝑥 and 𝑑𝑑𝑡𝑡 are the diameters of reactor and tube, respectively. The total number
of tubes, 𝑁𝑁𝑡𝑡 , can be calculated by the cross sectional area of the reactor divided by the projected
area of the tube layout pertaining to a single tube, 𝐴𝐴1 .

𝜋𝜋𝐷𝐷𝑥𝑥2
𝑁𝑁𝑡𝑡 = (𝐶𝐶𝑇𝑇𝑇𝑇) , 𝐴𝐴1 = (𝐶𝐶𝐶𝐶)𝑃𝑃𝑡𝑡2 (30)
4𝐴𝐴1

CTP is the tube count calculation constant and CL is the tube layout constant, where CTP=0.93
for one tube pass and CL=0.87 for 30 and 60 equilateral tri pitch 30.

Boundary Conditions

The boundary conditions are given in Table 2. At the top of the bed (z=L), both the axial
dispersion flux and the conductive heat flux are zero, while the loading and temperature at z=L
are equal to the solid sorbent inlet loading and temperature, respectively. At the bottom of the
bed (z=0) where the steam enters the stage, the boundaries are written considering the effect of
axial dispersion.

NUMERICAL SIMULATION

The one-dimensional model of the moving bed is implemented in Aspen Custom Modeler®
(ACM) (Aspen Technology, Inc.) and gPROMS® (Process Systems Enterprise). In both cases,
the mixed system of PDEs and algebraic equations is solved using the well-known method of
lines. To discretize the spatial dimension, z, two domains are used, one for forward flow
direction and one for reverse flow. The first-order forward finite difference method is applied to
the equations involving the reverse flows of the solid phase and tube-side medium and the first-
order backward finite difference method was used for all other PDEs.

Chemical Components and Thermodynamic Properties

The moving bed reactor model uses three main chemical components in the gas phase, namely
CO2, H2O and N2. Among these, CO2 and H2O are adsorbed and exist as ions in the solid phase.
Therefore, “Bic”, “Car” and “H2O”, which represent bicarbonate, carbamate and physisorbed
water, respectively, are specified as the components in the solid phase. The thermodynamic
properties of the gas phase are calculated using the Predictive Soave-Redlich-Kwong equation-
of-state (EOS) method. Steam properties utilize steam tables. In the ACM version property
calculations utilize Aspen Properties®. In the gPROMS version, property calculations utilize
Multiflash. Physical properties of the solids are obtained from experimental measurements20.

CASE STUDIES

To effectively demonstrate capture technology capable of 90% CO2 removal from existing coal
power plants, a 650 MWe net (before capture) supercritical pulverized coal power plant is used
as the reference plant31. The high volume of flue gas produced by this power plant makes it
impossible to perform CO2 capture to the desired degree using a single train of adsorbers and
regenerators since the maximum reactor size is limited, mainly by mechanical factors. Due to
these mechanical limitations, it is assumed in this study that the moving bed reactors have a
maximum allowable diameter of 9 m based on comparison with the largest reactors reported in
the literature (Table 3).

In the moving bed regenerator, CO2 and H2O are released as the solid sorbent is heated;
therefore, gas velocity increases and reaches its maximum at the top of the bed. To avoid
fluidizing the bed, especially at the top, one possible approach is to develop a multi-stage
moving bed. In this case, a portion of the evolved CO2 and water vapor is removed between
stages, ensuring that the superficial gas velocity remains below the minimum fluidization
velocity at the top of the bed. To verify the applicability of this approach, a single stage
representing the top of the bed will be used. An inlet stream is presumed to have a high
concentration of CO2 and comes from the stage below with a portion diverted before entering the
top stage as can be seen in Figure 2. The inlet conditions for the solid sorbent were determined
based upon a previously developed two-stage bubbling fluidized bed model22 that serves as the
adsorber. The base case for the moving bed regenerator is simulated using the reactor parameters
and operating conditions shown in Table 3.

Steady-State Simulation and Design Considerations

The steady-state simulation results of the moving bed stage are shown in Figure 3, which shows
the results from gPROMS. Note that gPROMS and ACM give similar results as shown in Table
4. Figure 3a shows that the flowrate of CO2 increases from 7.638 kmol/hr at the inlet (z/L = 0) to
44.09 kmol/hr at the exit. The carbamate molar flowrate, shown in Figure 3b, reduces from the
inlet flowrate of 204 kmol/hr to 168 kmol/hr at the bottom of the stage (z/L = 0), which is
equivalent to reducing the carbamate loading of the sorbent from 2.00 mol/kg to 1.64 mol/kg.
The bicarbonate loading remains very low throughout the stage. It can be seen in Figure 3c that
the concentration of the CO2 gas on the sorbent’s surface closely mirrors that of the molar
fraction of CO2 gas in the reactor (Figure 3a), which reduces the driving force of the reaction as
CO2 is released from the sorbent. The pressure drop, shown in Figure 3d, for the entire stage is
found to be 0.006 bar. Figure 3e shows that the gas and sorbent maintain thermal equilibrium
while the temperature has a steady increase from the top to the bottom of the reactor, notably
with a small decrease in heat transfer between z/L of 0.6 to 1.0. This is due to the gas velocity
approaching minimum fluidization velocity near the top of the stage. Note that while this
decrease in heat transfer may not be easily spotted in Figure 3e, the decrease is evident in Figure
5 when compared to other designs). Table 4 shows the base case comparison of ACM and
gPROMS.

The results presented in Figure 3 are from a single stage of the moving bed reactor; therefore, the
regeneration of the solid sorbent is not at the desired level for being used in the adsorber. While
this work is intended to demonstrate the model of a single stage in a moving bed reactor, it will
be necessary to have a multi-stage design to reach the necessary level of regeneration, where the
number of stages could be determined by evaluating appropriate design variables and operating
conditions and by conducting techno-economic optimization of the moving bed regenerator.
While an optimized design is beyond the scope of this work, the importance of such design
considerations is shown by performing an analysis of the example steady-state case where the
original operating conditions in Figure 3 are modified and compared to new sets of operating
conditions in Figure 4 and Figure 5.

Figure 4 shows the steady-state results from doubling the solid sorbent flowrate from the base
case value of 102,200 kg/hr to a new value of 204,400 kg/hr. It is observed that the CO2 removed
from the solid sorbent remains relatively constant despite a dramatic increase of solid-sorbent
flowrate. As seen in Figure 4a, the increase in solid sorbent flowrate results in a higher driving
force along the regenerator; however, this increased driving force is countered by a decrease in
solids residence time resulting in only a minor increase in CO2 released as shown in Figure 4b.

Another operating condition that can easily be examined is an increase of the inlet gas flowrate.
Figure 5 compares the base case, the increased inlet gas flowrate case, and a redesigned MB case
that accounts for the change in the gas residence time. The need for a redesigned MB stage can
be seen in Figure 5b, which shows that an increase of the inlet gas flowrate results in a
corresponding increase in the carbamate loading of the outgoing solids. At the exit (z/L = 0), the
carbamate loading goes from 1.64 mol/kg in the base case to 1.68 mol/kg for the increased inlet
gas case, corresponding to a reduction in the amount of CO2 released. This occurs due to a
reduction in the average temperature profile of the stage, corresponding to a reduction in overall
heat transfer coefficient. The effect on temperature can be clearly seen in Figure 5c which shows
a reduction in the temperature profile throughout the reactor for the increased inlet gas case in
comparison to the base case. This result for the increased inlet gas case makes it apparent that it
would be beneficial to redesign the stage to maintain the residence time of the gas and
temperature profile in the stage. To accomplish this, the heat exchanger surface area and cross-
sectional area of the bed can be modified to maintain similar conditions to the base case. By
doing this, however, the reduced gas velocity in the redesigned stage results in a higher average
temperature in the bed, even though the inlet and outlet temperatures are the same as the base
case. The results for the redesigned stage are shown with the dotted lines in Figure 5. The
sorbent regeneration improves in comparison to the base case and is reflected by higher amount
of CO2 released as seen in Figure 5b. An improvement in solid sorbent regeneration can
potentially decrease the operating cost by increasing the working capacity of the sorbent.
However, this approach would require an increase in capital cost due to the larger diameter of the
reactor, more heat exchanger area, and higher amount of live steam contained in the evolved gas.
These studies demonstrate that the design parameters for the moving bed should be optimized by
performing techno-economic studies.

Dynamic Simulation and Results

The steady-state solutions to the base case given in Table 3 are used as the initial conditions for
the dynamic simulation cases.

Change in solid inlet temperature

To investigate the effect of a disturbance in the inlet temperature of the solid sorbent, a 30
second ramp increase of 5 ⁰C (from 112.6 to 117.6 ⁰C) is implemented at 10th second followed
by a 30 second ramp decrease of 10 ⁰C (from 117.6 to 107.6 ⁰C) implemented at 200th second.
The effects of the inlet temperature change propagating through the bed can be clearly seen in
Figure 6a. Figures 6b and c show the decrease in the physisorbed water and carbamate loading
propagating through the bed that takes approximately 400 second to reach the bottom of the
stage. Likewise, the ramp introduced at 200th second has little effect on the previous ramp until
the temperature change propagates though the reactor. It is also interesting to note that the molar
flowrate of gas changes immediately as seen in Figure 6d. This is due to the fact that an increase
in the solids inlet temperature results in an immediate increase in reaction rates at the top of the
bed, resulting in a rapid release of gas from the solid sorbent at the top of the bed. This study
shows that if the regenerator temperature is manipulated to control the extent of regeneration of
the solid sorbent, which in turn, affects the adsorber performance, the associated delay must be
taken into account while designing the control system.

Change in inlet solid sorbent carbamate loading

Due to changes in the adsorber operating conditions such as change in the flowrate of the flue
gas to the adsorber or partial pressure of CO2 at the inlet of the adsorber, the CO2 loading of the
solid sorbent can change. To simulate this disturbance, starting at a time of 10th second, the
carbamate loading of the inlet solid sorbent is ramped by 0.5 mol/kg over 30 second. Figures 7a
and b show the transients in the molar flow of the adsorbed solids, and give a residence time of
about 400 second that is similar to what has been reported in the previous section. However, it is
interesting to note that the physisorbed water experiences a spike that propagates down the stage
exiting at about 400th second, at which point the reactor reaches a new steady state which can be
seen in Figure 7b. Figures 7c and d show the transients of the molar flow of the gas and shows
that the gas profile reaches steady state much more quickly than the absorbed solid profile.

Change in inlet gas composition

The initial condition for the inlet gas contains a high concentration of CO2 due to the assumption
that the gas comes from the moving bed stage below. Due to changes in the operating conditions
of the lower stages, the composition of the inlet gas can change. Figure 8 shows the transient
effect of changing the inlet gas composition. Starting at a time of 10th second, a ramp decrease of
0.55 in the mass fraction of CO2 in the inlet gas (from 0.6 to 0.05 CO2) over 30 second is
implemented. To ensure the sum of the gas mass fractions is unity, the mass fraction of water is
increased accordingly. One of advantages of the moving bed is that the counter-current flow
allows for maximizing the driving force, especially at the boundaries. Therefore a reduction in
the partial pressure of CO2 will have a dramatic effect on the driving force at the bottom of the
stage, resulting in a greater release of CO2 from the solid sorbent. Figures 8a and 8b show the
molar flowrate of carbamate and physisorbed water at the bottom of the stage. It is interesting to
note that there is an immediate jump in the solids flow rate as soon as the ramp is introduced;
however, the trend continues for another 400 second before reaching steady state. Figures 8c and
8d show that the molar flowrates of the gas species reach the new steady state fairly quickly.
Even though the residence time of the gas species is fairly small, the residence time of the solid
sorbent is considerably larger, resulting in a longer settling time. These results give an interesting
perspective and suggest that the extent of regeneration of the solid sorbent can be efficiently
controlled by potentially injecting steam between stages, for the tradeoff of additional utility
cost. The trade-off between cost and extent of sorbent regeneration would need to be evaluated to
obtain an economic optimum.

Examination of process constraints on velocity

It will be desired that the moving bed’s operating gas velocity is as close as possible to the
minimum fluidization velocity as applications to CO2 capture systems will require that the
moving bed be capable of handling large volumes of gas. Generally, this requires a large cross-
sectional bed area to keep the gas velocity constraint satisfied, and any increases in capital cost
for the purpose of providing an operating cushion could prove to be economically infeasible.
This section examines what operating conditions are important in keeping the gas velocity below
that of the minimum fluidization velocity.
Figure 9 shows transient gas velocity responses to numerous ramps in process variables, and
compares the gas velocity response (shown as a multicolored grid) to that of the vmf constraint
(shown as a wire mesh grid). Figure 9a shows a ramp in the solid carbamate loading
demonstrated in Section 4.2.2 and Figure 9b shows a ramp in the inlet solid sorbent temperature
demonstrated in Section 4.2.1. Both of these process variables are likely disturbances to the
moving bed and in both cases the gas velocity increases to a point that violates the vmf constraint,
and therefore demonstrates a need for a control strategy. One of the effective means for reducing
the gas velocity is to limit the reactions taking place in the moving bed and therefore limiting the
amount of CO2 gas evolved. This leaves many possible options for control variables, including
the inlet temperatures of the gas or solid, inlet gas composition, inlet feed rates of both the gas
and solid, and steam temperature of the integrated heat exchanger. However, some of these
process variables have small effects on the overall gas velocity profile or the vmf profile. In many
cases, these variables only have an effect near the boundary. An example of this can be seen in
Figure 10a, as the left plot shows the response to a ramped increase in inlet gas CO2 composition
from a mass fraction of 0.6 to 1. While it reduces the amount of evolved CO2 gas in the reactor,
the effect on the gas velocity is not dominant as the main contribution to the gas flowrate is from
the released CO2 as shown in Figure 3a. It can be seen that this has negligible effect on the vmf. A
ramped decrease in the inlet temperature of the solids has a similar effect on the vmf as the
previous case, however this results in a large reduction in gas velocity especially at the top of the
bed as shown in Figure 10b. However, a decrease in the inlet temperature will result in a
decreased regeneration of the sorbent. Thus, the moving bed systems that will comprise of
several moving beds for large-scale applications need to be designed and operated by
considering such disturbances and tradeoffs.
CONCLUSIONS

A one-dimensional, non-isothermal model of a moving bed reactor has been developed for
steady-state and pressure-driven dynamic operation of temperature swing regeneration of solid
sorbents used for post-combustion CO2 capture from a pulverized coal power plant. This reactor
model provides a means to investigate tradeoffs between various design alternatives and
determine optimal configurations and conditions. It is flexible over a wide range of operating
conditions and is capable of identifying various concepts and exploring the feasibility of
commercial-scale application. A steady-state, base case of a single-stage moving bed is
simulated and axial profiles of temperature, pressure, and molar flowrates of gaseous and
adsorbed species are generated. The results show that an increase in the solids circulation rate
can result in negligible increase in the flowrate of CO2 released if the bed is not appropriately
designed for the increased flowrate. The results also show the bed should be properly designed
by considering the amount of evolved gas flowrate. Thus, simultaneous optimization of the
operating conditions and the bed design parameters are necessary by considering the techno-
economic tradeoffs.

Additionally, the transient response to important process disturbances, including solid sorbent
loading, inlet temperature, and inlet gas composition, are examined. By introducing a ramp in the
solid sorbent inlet temperature and loading, it is found that the effect on the solid sorbent exiting
the bed takes approximately 400 second before any change is observed. If multiple beds in series
are considered for commercial applications, such a delay would add up and should be accounted
for while designing the control system. Introduction of a change in the inlet gas compositions
shows the high impact of the partial pressure of CO2 and H2O on the reaction rates, and that
changes in gas composition have nearly instantaneous effects. The results suggest that injection
of steam between the stages can be used to improve working capacity of a moving bed and can
possibly be used as a control strategy to control the extent of regeneration of the solid sorbent.
Additionally, it is demonstrated how the model can be used to determine effective control
strategies to ensure that the minimum fluidization velocity is nor violated. These options will be
investigated further in our future work when the design of the entire CO2 capture system is
undertaken.

Acknowledgments

This work was supported by the U.S. Department of Energy, Office of Fossil Energy as part of
the Carbon Capture Simulation Initiative (CCSI). H.K. acknowledges the support by an
appointment to the U.S. Department of Energy (DOE) Postgraduate Research Program at the
National Energy Technology Laboratory administered by the Oak Ridge Institution for Science
and Education.

S.M., B.O., and D.B. acknowledges the National Energy Technology Laboratory’s Regional
University Alliance (NETL-RUA), a collaborative initiative of the NETL. This technical effort
was performed under the RES contract DE-FE0004000.

Disclaimer

This project was funded by the Department of Energy, National Energy Technology Laboratory,
an agency of the United States Government, through a support contract with URS Energy &
Construction, Inc. Neither the United States Government nor any agency thereof, nor any of their
employees, nor URS Energy & Construction, Inc., nor any of their employees, makes any
warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product, or process disclosed, or
represents that its use would not infringe privately owned rights. Reference herein to any specific
commercial product, process, or service by trade name, trademark, manufacturer, or otherwise,
does not necessarily constitute or imply its endorsement, recommendation, or favoring by the
United States Government or any agency thereof. The views and opinions of authors expressed
herein do not necessarily state or reflect those of the United States Government or any agency
thereof.
Nomenclature
𝑎𝑎 External surface area per unit volume, 1/𝑚𝑚
𝐴𝐴𝐴𝐴 Archimedes number, 𝜌𝜌𝑔𝑔 �𝜌𝜌𝑠𝑠 − 𝜌𝜌𝑔𝑔 �𝑔𝑔𝑑𝑑𝑝𝑝3 �𝜇𝜇𝑔𝑔2
𝐶𝐶 Mole concentration, 𝑚𝑚𝑚𝑚𝑚𝑚/𝑚𝑚3
𝐶𝐶𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 Mole concentration of gas on the surface of the particle
𝐶𝐶𝑝𝑝 Specific heat capacity, k𝐽𝐽/𝐾𝐾𝐾𝐾𝐾𝐾
𝐷𝐷 Mass diffusivity, 𝑚𝑚2 /𝑠𝑠
𝑑𝑑𝑝𝑝 Diameter of particle, 𝑚𝑚
𝐷𝐷𝑒𝑒 Vapor diffusivity, 𝑚𝑚2 /𝑠𝑠
𝐷𝐷𝑖𝑖1 −𝑖𝑖2 Binary diffusivity between two species, m2/s
𝑑𝑑𝑡𝑡 Diameter of tube, 𝑚𝑚
𝐷𝐷𝑥𝑥 Reactor diameter, 𝑚𝑚
𝑓𝑓𝑏𝑏 Fraction of time during which heat exchanger contacts the emulsion region
𝑓𝑓𝑛𝑛 Fluidization Number (𝑣𝑣/𝑣𝑣𝑚𝑚𝑚𝑚 )
𝑔𝑔 Gravitational acceleration, 𝑚𝑚/𝑠𝑠 2
∆𝐻𝐻𝑟𝑟𝑟𝑟𝑟𝑟 Enthalpy of adsorption, 𝑘𝑘𝑘𝑘/𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘
∆𝐻𝐻𝑣𝑣𝑣𝑣𝑣𝑣 Enthalpy for steam condensation, 𝑘𝑘𝑘𝑘/𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘
ℎ Heat transfer coefficient, 𝑘𝑘𝑘𝑘/𝑚𝑚2 𝐾𝐾
𝐽𝐽 Axial mass flux, 𝑘𝑘𝑘𝑘/𝑚𝑚2 𝑠𝑠
𝑘𝑘 Thermal conductivity, 𝑘𝑘𝑘𝑘/𝑚𝑚𝑚𝑚
𝑘𝑘𝐿𝐿𝐿𝐿𝐿𝐿 Overall effective rate coefficient, 1/𝑠𝑠
𝑘𝑘𝑓𝑓,𝑔𝑔 Fluid film mass transfer coefficient, 𝑚𝑚/𝑠𝑠
𝐿𝐿 Reactor height, 𝑚𝑚
𝑀𝑀𝑀𝑀 Average molecular weight, 𝑘𝑘𝑘𝑘/𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘
𝑛𝑛𝑣𝑣 Total amine loading on sorbent, 𝑚𝑚𝑚𝑚𝑚𝑚 ⁄𝑚𝑚3 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
𝑁𝑁𝑡𝑡 Number of tubes
𝑃𝑃 Pressure in gas phase, 𝑏𝑏𝑏𝑏𝑏𝑏
𝑃𝑃𝑃𝑃𝑝𝑝 Peclet number
𝑃𝑃𝑃𝑃 Prandtl number, 𝐶𝐶𝑝𝑝,𝑔𝑔 𝜇𝜇𝑔𝑔 ⁄𝑘𝑘𝑔𝑔
𝑃𝑃𝑡𝑡 Tube pitch, 𝑚𝑚
𝑄𝑄 Transferred amount of adsorbent in moles per unit volume, 𝑚𝑚𝑚𝑚𝑚𝑚/𝑚𝑚3
𝑅𝑅𝑅𝑅𝑝𝑝 Particle Reynolds number, 𝜌𝜌𝑔𝑔 �𝑣𝑣𝑔𝑔 + 𝑣𝑣𝑠𝑠 �𝑑𝑑𝑝𝑝 �𝜇𝜇𝑔𝑔
𝑆𝑆𝑆𝑆𝑝𝑝 Particle Schmidt number, 𝜇𝜇𝑔𝑔 ⁄𝜌𝜌𝑔𝑔 𝐷𝐷𝑔𝑔
𝑇𝑇 Temperature, ℃
𝑡𝑡 Time, 𝑠𝑠
 𝑈𝑈𝑐𝑐 Transition velocity, 𝑚𝑚/𝑠𝑠
𝑣𝑣 Superficial velocity, 𝑚𝑚/𝑠𝑠
𝑣𝑣𝑚𝑚𝑚𝑚 Minimum fluidization velocity, 𝑚𝑚/𝑠𝑠
𝑣𝑣𝑣𝑣 Degree of saturation of steam
𝑉𝑉𝑚𝑚,𝑗𝑗
𝑤𝑤 Mole concentration of adsorbent per unit mass of particle, 𝑚𝑚𝑚𝑚𝑚𝑚 ⁄𝑘𝑘𝑘𝑘 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
𝑦𝑦 Gas phase mole fraction, 𝑚𝑚𝑚𝑚𝑚𝑚/𝑚𝑚𝑚𝑚𝑚𝑚
𝑧𝑧 Axial coordinates, 𝑚𝑚
𝛼𝛼 Intraparticle void fraction
𝜀𝜀𝑏𝑏 Average bed voidage
𝜇𝜇 Dynamic viscosity, 𝑃𝑃𝑃𝑃 ∙ 𝑠𝑠
𝜈𝜈 Stoichiometric coefficient
𝜌𝜌 Mass density of solid, 𝑘𝑘𝑘𝑘/𝑚𝑚3
τ Average residence time of emulsion packets at the heat exchanger surface, s
𝜓𝜓 Particle shape factor

Subscript
𝑔𝑔 Gas phase

𝑖𝑖 Components in gas phase

𝑗𝑗 Adsorbed species in solid phase

𝑠𝑠 Solid phase

𝑠𝑠𝑠𝑠 Inner cylinder in the Zehner-Schlunder unit cell model

𝑡𝑡 Tube side

𝑤𝑤 Tube-wall

𝑧𝑧 Effective axial quantity

Literature Cited
1. Figueroa JD, Fout T, Plasynski S, McIlvried H, Srivastava RD. Advances in CO2 capture
technology—The U.S. Department of Energy’s Carbon Sequestration Program. Int J
Greenh Gas Control. 2008;2(1):9-20.

2. Ciferno JP, Fout TE, Jones AP, Murphy JT. Capturing carbon from existing coal-fired
power plants. Chem Eng Prog. 2009;105(4):33-41.

3. Arunkumar Samanta AZ. Post-Combustion CO2 Capture Using Solid Sorbents: A Review.
Ind Amp Eng Chem Res. 2011;51(4). doi:10.1021/ie200686q.

4. Tarka TJ, Ciferno JP, Gray ML, Fauth D. CO2 capture systems using amine enhanced solid
sorbents. In: Fifth Annual Conference on Carbon Capture and Sequestration, Alexandria,
VA.; 2006.

5. Ma X, Wang X, Song C. “Molecular basket” sorbents for separation of CO2 and H2S from
various gas streams. J Am Chem Soc. 2009;131(16):5777-5783.

6. Zhang Z, Ma X, Wang D, Song C, Wang Y. Development of silica-gel-supported

polyethylenimine sorbents for CO2 capture from flue gas. AIChE J. 2012;58(8):2495-2502.

7. Yang W-C, Hoffman J. Exploratory design study on reactor configurations for carbon
dioxide capture from conventional power plants employing regenerable solid sorbents. Ind
Eng Chem Res. 2008;48(1):341-351.

8. Ruthven DM. Principles of Adsorption and Adsorption Processes. John Wiley & Sons;
1984.
https://books.google.com/books?hl=en&lr=&id=u7wq21njR3UC&oi=fnd&pg=PR18&dq=
%22Principles+of+adsorption+and+adsorption+processes%22&ots=waUoQvykxX&sig=36
 K_Zxz-sKkQihBFMsN7JA69dV4. Accessed January 19, 2015.

9. Roes AWM, Van Swaaij WPM. Hydrodynamic behaviour of a gas—solid counter-current

packed column at trickle flow. Chem Eng J. 1979;17(2):81-89.

10. Parisi DR, Laborde MA. Modeling of counter current moving bed gas-solid reactor used in
direct reduction of iron ore. Chem Eng J. 2004;104(1–3):35-43.

11. Valipour MS, Saboohi Y. Modeling of multiple noncatalytic gas–solid reactions in a

moving bed of porous pellets based on finite volume method. Heat Mass Transf.
2007;43(9):881-894.

12. Dussoubs B, Jourde J, Patisson F, Houzelot J-L, Ablitzer D. Modelling of a moving bed
furnace for the production of uranium tetrafluoride Part 1: formulation of the model. Chem
Eng Sci. 2003;58(12):2617-2627.

13. Lacerda AF, Lisboa MH, Barrozo MA. Heat and mass transfer in a countercurrent moving
bed dryer. Appl Therm Eng. 2005;25(16):2641-2652.

14. Lira TS, Barrozo MAS, Assis AJ. Concurrent moving bed dryer modelling: Sensitivity of
physicochemical parameters and influence of air velocity profiles. Appl Therm Eng.
2009;29(5–6):892-897.

15. Sá Gomes P, Leão CP, Rodrigues AE. Simulation of true moving bed adsorptive reactor:
Detailed particle model and linear driving force approximations. Chem Eng Sci.
2007;62(4):1026-1041.

16. Knaebel KS. Temperature swing adsorption system. September 2009.

http://scholar.google.com/scholar_url?url=https%3A%2F%2Ffanyv88.com%3A443%2Fhttps%2Fwww.google.com%2Fpatents
%2FUS7594956&hl=en&sa=T&ct=res&cd=0&ei=3YG9VI3fDcS70AH8hYCIBQ&scisig=
AAGBfm3OlkNEiWNbUgNlhc3EB7TF9nXL3A&nossl=1&ws=1440x761. Accessed
January 19, 2015.

17. Kent S. Knaebel, Seth P. Knaebel. CO2 Capture by Moving Bed Temperature Swing
Adsorption. In: Vol Philadelphia, PA, USA; 2008.

18. David C. Miller, M Syamlal, J C Meza, et al. Overview of the U.S. DOE’s Carbon Capture
Simulation Initiative for Accelerating the Commercialization of CCS Technology. 36th
 International Technical Conference on Clean Coal & Fuel Systems, Clearwater, FL, United
States; 2011.

19. David C. Miller, N. V. Sahinidis, A. Cozad, et al. Computational Tools for Accelerating
Carbon Capture Process Development. In: Vol Clearwater, Florida, USA; 2013.

20. Miller DC, Syamlal M, Mebane DS, et al. Carbon Capture Simulation Initiative: A Case
Study in Multiscale Modeling and New Challenges. Annu Rev Chem Biomol Eng.
2014;5(1):301-323. doi:10.1146/annurev-chembioeng-060713-040321.

21. Lee A, Miller DC. A One-Dimensional (1-D) Three-Region Model for a Bubbling
Fluidized-Bed Adsorber. Ind Eng Chem Res. 2012;52(1):469-484.

22. Modekurti S, Bhattacharyya D, Zitney SE. Dynamic Modeling and Control Studies of a
Two-Stage Bubbling Fluidized Bed Adsorber-Reactor for Solid–Sorbent CO2 Capture. Ind
Eng Chem Res. 2013;52(30):10250-10260.

23. Sjostrom S, Krutka H. Evaluation of solid sorbents as a retrofit technology for CO2 capture.
Fuel. 2010;89(6):1298-1306.

24. Lee A, Mebane D, Fauth D, Miller D. A Model for the Adsorption Kinetics of CO2 on
Amine-Impregnated Mesoporous Sorbents in the Presence of Water. National Energy
Technology Laboratory-In-house Research; National Energy Technology Laboratory
(NETL), Pittsburgh, PA, and Morgantown, WV (United States); 2011.
http://www.osti.gov/scitech/biblio/1036497. Accessed January 19, 2015.

25. Marrero TR, Mason EA. Gaseous Diffusion Coefficients. J Phys Chem Ref Data.
1972;1(1):3-118.

26. Yoon SM, Kunii D. Gas Flow and Pressure Drop through Moving Beds. Ind Eng Chem
Process Des Dev. 1970;9(4):559-565.

27. Chen JC, Grace JR, Golriz MR. Heat transfer in fluidized beds: design methods. Powder
Technol. 2005;150(2):123-132.

28. Baskakov AP, Berg BV, Vitt OK, et al. Heat transfer to objects immersed in fluidized beds.
Powder Technol. 1973;8(5–6):273-282.
29. Kunii D, Levenspiel O. Fluidization engineering. 1969. http://agris.fao.org/agris-
search/search.do?recordID=US201300671282. Accessed January 20, 2015.

30. Kakac S, Liu H, Pramuanjaroenkij A. Heat Exchangers: Selection, Rating, and Thermal
Design. CRC press; 2012.
https://books.google.com/books?hl=en&lr=&id=sJXpvP6xLZsC&oi=fnd&pg=PP1&dq=He
at+exchangers:+selection,+rating,+and+thermal+design&ots=MJhZ2ToGcy&sig=I_MO4k
WG2BbQITjK5m1VCqin0rc. Accessed January 19, 2015.

31. David C. Miller, Y. Chang, A Cozad, H Kim, Andrew Lee. Synthesis of Optimal
Adsorptive Carbon Capture Processes. In: AIChE Annual Meeting., Minneapolis, MN,
USA; 2011.
 Solid Inlet

Gas Out
Utility Out

Solid Particles
Heat Exchanger
Tubes
Perforated trays

Utility In
Gas Inlet
Distributor

Solid Out

Figure 1. General schematic of a moving bed reactor.

Figure 2. Two stage moving bed reactor with CO2 draw-off between stages.
Figure 3. Steady-state profiles for the base case design. The x-axis, z/L, represents the
normalized axial position of the bed, where 1.0 is the top of the bed where the sorbent
enters the bed, and 0.0 is the bottom of the bed where the gas enters.
 (a) Evolved Gas Flow Profile (b) Carbamate Loading Profile
70 2

60
1.9

mol/kg sorbent
50
kmol/hr

40 1.8

30
1.7
20 Base case Base case
Increased inlet solid sorbent Increased inlet solid sorbent
10 1.6
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
z/L z/L

Figure 4. Comparison of steady-state results for inlet solid-sorbent flows of 102,200 kg/hr
and 204,400 kg/hr. Figure 4a represents the outlet sorbent loading and Figure 4b
represents the profile of the molar flowrate of the gas.
 (a) Evolved Gas Flow Profile (b) Carbamate Loading Profile
100 2

80 1.9

mol/kg sorbent
60 1.8
kmol/hr

40 1.7
Base case Base case
20 Increased inlet gas 1.6 Increased inlet gas
Redesigned stage Redesigned stage
0 1.5
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
z/L z/L
(c) Temperature Profile
130
Base case
Increased inlet gas
125 Redesigned stage
T( ° C)

120

115

110
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
z/L

Figure 5. Comparison of steady-state results with low (base case) and increased flowrates of
inlet gas with the original design and redesigned stage to maintain gas residence time and
temperature profile.
 Figure 6. Results of a dynamic simulation of a disturbance in the inlet solid sorbent
temperature (6a) (Note: the time and spatial axis are switched for temperature profiles for
clear viewing of the transient response). Figures 6b and 6c show the effect of the simulated
ramp on the molar solid component flow through the moving bed stage profile, while
Figure 6d shows the molar flowrate of the evolved gas through the stage.
Figure 7. Results of a simulated ramp in the inlet solid sorbent loading. Figures 7a and 7b
show the molar flow rates of the solid carbamate and physisorbed water through the bed,
while Figures 7c and 7d show the flow profiles of the CO2 and water through the bed.
Figure 8. Results of a simulated disturbance to the inlet compositions of the gas. Figures 8a
and represent the solid sorbent loading exiting the stage and Figures 8c and 8d represent
the profile of the gaseous molar flowrate profile of CO2 and water.
Figure 9. Gas velocity (multicolored grid) and vmf (wire mesh grid) response to ramped inputs. The
left plot shows an increase in the solid carbamate loading. The right plot shows the response to an
increase and a decrease in the inlet solid’s temperature.
 Figure 10.Gas velocity (multicolored grid) and vmf (wire mesh grid) response to ramped
inputs. The left plot shows an increase of the CO2 in the inlet gas. The right plot shows the
response to a decrease in the inlet steam temperature to the integrated heat exchanger.
 Table 1. Equilibrium, Kinetic, and Diffusivity parameters

Parameter Value Reference

∆𝐻𝐻𝐻𝐻2 𝑂𝑂 [𝐽𝐽⁄𝑚𝑚𝑚𝑚𝑚𝑚] -52,100
∆𝑆𝑆𝐻𝐻2 𝑂𝑂 [𝐽𝐽⁄𝐾𝐾/𝑚𝑚𝑚𝑚𝑚𝑚] -78.5
𝐸𝐸𝐻𝐻2 𝑂𝑂 [𝐽𝐽⁄𝑚𝑚𝑚𝑚𝑚𝑚 ] 28,200
𝐴𝐴𝐻𝐻2 𝑂𝑂 5.62 × 104
∆𝐻𝐻𝐵𝐵𝐵𝐵𝐵𝐵 [𝐽𝐽⁄𝑚𝑚𝑚𝑚𝑚𝑚 ] -70,300
∆𝑆𝑆𝐵𝐵𝐵𝐵𝐵𝐵 [𝐽𝐽⁄𝐾𝐾/𝑚𝑚𝑚𝑚𝑚𝑚 ] -274.5
𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 [𝐽𝐽⁄𝑚𝑚𝑚𝑚𝑚𝑚 ] 61,850 Updated parameters24

𝐴𝐴𝐵𝐵𝐵𝐵𝐵𝐵 1.91 × 10−4

∆𝐻𝐻𝐶𝐶𝐶𝐶𝐶𝐶 [𝐽𝐽⁄𝑚𝑚𝑚𝑚𝑚𝑚] -99,630
∆𝑆𝑆𝐶𝐶𝐶𝐶𝐶𝐶 [𝐽𝐽⁄𝐾𝐾/𝑚𝑚𝑚𝑚𝑚𝑚] -265.3
𝐸𝐸𝐶𝐶𝐶𝐶𝐶𝐶 [𝐽𝐽⁄𝑚𝑚𝑚𝑚𝑚𝑚 ] 64,380
𝐴𝐴𝐶𝐶𝐶𝐶𝐶𝐶 58.88
𝑚𝑚 1.17

𝑎𝑎𝐶𝐶𝐶𝐶2 −𝐻𝐻2 0 9.24 × 10−5

𝑏𝑏𝐶𝐶𝐶𝐶2 −𝐻𝐻2 0 1.57
𝑐𝑐𝐶𝐶𝐶𝐶2 −𝐻𝐻2 0 −307.9
𝑎𝑎𝐶𝐶𝐶𝐶2 −𝑁𝑁2 3.15 × 10−5
𝑏𝑏𝐶𝐶𝐶𝐶2 −𝑁𝑁2 1.5 Marrero and Mason 25

𝑐𝑐𝐶𝐶𝐶𝐶2 −𝑁𝑁2 -113.6

𝑎𝑎𝑁𝑁2 −𝐻𝐻2 𝑂𝑂 0.187 × 10−5
𝑏𝑏𝑁𝑁2 −𝐻𝐻2 𝑂𝑂 2.072
𝑐𝑐𝑁𝑁2 −𝐻𝐻2 𝑂𝑂 0
 Table 2. Reactor Boundary Conditions

Variables Bottom of the Bed (z=0) Top of the Bed (z=L)

𝜕𝜕𝐶𝐶𝑖𝑖 𝜕𝜕𝐶𝐶𝑖𝑖
𝐶𝐶𝑖𝑖 𝑣𝑣𝑔𝑔 (0)𝐶𝐶𝑖𝑖 (0) = 𝑣𝑣𝑔𝑔,0 𝐶𝐶0,𝑖𝑖 − 𝜀𝜀𝑏𝑏 𝐷𝐷𝑧𝑧 � 𝐷𝐷𝑧𝑧 � =0
𝜕𝜕𝜕𝜕 𝑧𝑧=0 𝜕𝜕𝜕𝜕 𝑧𝑧=𝐿𝐿

𝑤𝑤𝑗𝑗 - 𝑤𝑤𝑗𝑗 (𝐿𝐿) = 𝑤𝑤𝑗𝑗,𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼

𝐶𝐶𝑝𝑝,𝑔𝑔 (0)𝑣𝑣𝑔𝑔 (0)𝜌𝜌𝑔𝑔 (0)𝑇𝑇𝑔𝑔 (0)

𝜕𝜕𝑇𝑇𝑔𝑔
𝑇𝑇𝑔𝑔 𝜕𝜕𝑇𝑇𝑔𝑔 𝑘𝑘𝑔𝑔 � =0
= 𝐶𝐶𝑝𝑝,𝑔𝑔,0 𝑣𝑣𝑔𝑔,0 𝜌𝜌𝑔𝑔,0 𝑇𝑇𝑔𝑔,0 − 𝜀𝜀𝑏𝑏 𝑘𝑘𝑔𝑔 � 𝜕𝜕𝜕𝜕 𝑧𝑧=𝐿𝐿
𝜕𝜕𝜕𝜕 𝑧𝑧=0

𝑇𝑇𝑠𝑠 - 𝑇𝑇𝑠𝑠 (𝐿𝐿) = 𝑇𝑇𝑠𝑠,𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼

Table 3. Moving Bed Regenerator Stage Parameters and Operating Conditions

Variable Base Value Units

Reactor Diameter 9.00 𝑚𝑚

Reactor Height 1.00 𝑚𝑚

Average Voidage 0.6

Flow Rate of Inlet Gas 12.73 kmol/hr

CO2 Molar Fraction in Inlet Gas 0.600

H2O Molar Fraction in Inlet Gas 0.400

Pressure of Inlet Gas 1.206 bar

Temperature of Inlet Gas 125.1 ⁰C

Flow Rate for Inlet Solid Sorbent 102,200 kg/hr

Temperature for Inlet Solid Sorbent 112.6 ⁰C

Initial Loading of Bicarbonate 2.6x10-6 mol⁄kg sorbent

Initial Loading of Carbamate 1.998 mol⁄kg sorbent

Initial Loading of Physisorbed Water 0.620 mol⁄kg sorbent

Steam Flow Rate for Heat Exchanger 150.0 kmol/hr

Steam Temperature for Heat Exchanger 138.1 ⁰C

Diameter of Heat Exchanger Tube 0.0381 m

 Table 4. Comparison of moving bed results in ACM and gPROMS

Variable Platform At Solid Inlet Side At Solid Outlet Side

Carbamate Loading ACM - 1.642
(mol/kg) gPROMS - 1.641
ACM - 0.492
Water Loading (mol/kg)
gPROMS - 0.507
ACM 112.04 (z/L=0.01) 125.10 (z/L=0.99)
Solids Temperature (ᵒC)*
gPROMS 112.09 (z/L=0.01) 125.26 (z/L=0.99)
Molar Flow of Gas ACM 62.13 -
(kmol/hr) gPROMS 62.50 -
ACM 1.2067 1.2000
Pressure (bar)
gPROMS 1.2064 1.2000
*The solids temperatures are compared at the first calculated values around the boundaries (z/L=0.01 and z/L=0.99)
as the values at the boundary depend heavily on the specified boundary conditions. However, each platform may
handle the large jumps in reactions rates near the boundary differently, which would be reflected in a change in
temperature.