Chemical Engineering Communications

ISSN: 0098-6445 (Print) 1563-5201 (Online) Journal homepage: https://www.tandfonline.com/loi/gcec20

Simulation study of fixed-bed CO2 adsorption from

CO2/N2 mixture using activated carbon

Mohammed K. Al Mesfer, Abdelfattah Amari, Mohd Danish, Basem Abdullah

Al Alwan & Mumtaj Shah

To cite this article: Mohammed K. Al Mesfer, Abdelfattah Amari, Mohd Danish, Basem
Abdullah Al Alwan & Mumtaj Shah (2020): Simulation study of fixed-bed CO2 adsorption
from CO2/N2 mixture using activated carbon, Chemical Engineering Communications, DOI:
10.1080/00986445.2020.1777111

To link to this article: https://doi.org/10.1080/00986445.2020.1777111

Published online: 16 Jun 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=gcec20
CHEMICAL ENGINEERING COMMUNICATIONS
https://doi.org/10.1080/00986445.2020.1777111

Simulation study of fixed-bed CO2 adsorption from CO2/N2 mixture

using activated carbon
Mohammed K. Al Mesfera, Abdelfattah Amaria,b, Mohd Danisha, Basem Abdullah Al Alwana, and
Mumtaj Shahc
a
Chemical Engineering Department, King Khalid University, Abha, Saudi Arabia; bResearch Laboratory: Energy and Environment,
National School of Engineers, Gabes University, Gabes, Tunisia; cChemical Engineering Department, Indian Institute of Technology,
Roorkee, India

ABSTRACT KEYWORDS
An experimental and simulation study using a packed column for CO2 sorption was con- Adsorption; breakthrough;
ducted. The adsorption data of CO2 sorption from CO2/N2 mixture using activated carbon LDF model; mass transfer;
(NORIT RB1) were obtained at different operating conditions. The sorption equilibrium and packed column; simulation
breakthrough profiles were produced with a total pressure of 1.25 bars. The breakthrough
and saturation time decreased with increasing temperature, feed rate, and inlet CO2 concen-
tration. A linear driving force (LDF) model for mass transfer was used to calculate the mass
transfer coefficient and reproduce the breakthrough curves. The findings of the LDF model
were compared with experimental data produced at various operating conditions to predict
the breakthrough curves. The experimental and predicted findings were analyzed by ascer-
taining the determination coefficient and mean error. The model consistently reproduced all
the breakthrough curves and may be considered adequate for predicting the CO2 separation
from a CO2/N2 mixture with R2 > 0.97 and mean error less than 5% under all the operat-
ing conditions.

Introduction model was adopted for CO2 and CO2–N2 mixture

The combustion of fossil fuels accounts for adsorption onto 13 X zeolite. This model repro-
approximately 81% of commercial energy con- duced the breakthrough curves excellently
sumption, and it annually releases 3.0
1013 kg (Dantas et al. 2009). The parameter estimation
of carbon dioxide. The predominant environmen- and modeling for CO2 capture onto 13 X were
tal issue pertains to the alarming rate at which performed, and the dynamic behavior of the vari-
the carbon dioxide concentration is increasing in ables was predicted with high accuracy (Won
the atmosphere. The adsorption kinetics of break- et al. 2012) using a column with an inner diam-
through curves was explained with acceptable eter of 2.54 cm.
accuracy using a linear driving force (LDF) A developed model (Mulgundmath et al. 2012)
model for adsorption on MOF-508b crystals effectively represented the concentration and
(Barcia et al. 2008). The CO2 adsorption model- temperature breakthrough curves. The investiga-
ing on amine-based mesoporous silica was under- tors (Monazam et al. 2013) analyzed the experi-
taken to describe CO2 adsorption isotherms, and mental breakthrough curves using the model, and
acceptability of the model was indicated by the characteristics parameters were deduced from the
standard deviation of less than 8% (Serna- breakthrough curves. A detailed study on math-
Guerrero et al. 2010a, 2010b). The kinetic and ematical modeling based on fixed-bed CO2
breakthrough modeling were studied for CO2 adsorption was conducted (Shafeeyan et al. 2014).
adsorption, and the best fit was attained using A dynamic model (Park and Lee 2016) was intro-
Avrami’s model (Serna-Guerrero et al. 2010a, duced to simulate the fixed-bed adsorption pro-
2010b). A LDF-approximated mathematical cess with 13 X zeolite; it was concluded that a

CONTACT Mohd Danish [email protected] Chemical Engineering Department, King Khalid University, Abha, Saudi Arabia.
ß 2020 Taylor & Francis Group, LLC
2 M. K. A. MESFER ET AL.

mass transfer model can determine the adsorp-

tion rate. The simulation and experimental study
of CO2/N2 was performed using a dual-reflux
pressure swing process with silica gel as the
adsorbent, and adsorption isotherms were fitted
by the Langmuir model (Li et al. 2016). A K-
based adsorbent was used to establish equilibrium
and kinetic breakthrough curve models for CO2
adsorption (Zhang et al. 2016). The internal mass
transfer coefficient was found to be more sensi-
tive. The CO2 adsorption modeling performed on
the commercial adsorbent relied on the tempera-
ture-dependent Sips and Toth models to predict
the adsorption equilibrium of CO2 (Bedel et al.
2017); this model fits well with the experimental
data. The adsorption response of polyaspartamide Figure 1. Packed bed sorption column.
was predicted using models based on kinetics,
and adsorption was credited to external mass the adsorption of CO2/H2 mixture using activated
transfer (Yoro et al. 2017). carbon was performed with a detailed one-
A model based on LDF approximations for dimensional model (Casas et al. 2012).
CO2 adsorption on activated carbon and N2 Microporous biochar was used for the experi-
enriched activated carbon adequately predicted mental simulation studies for CO2 and N2
the mass transfer coefficient and breakthrough adsorption, and the mathematical model was vali-
curves (Dantas et al. 2009). Ammonia-modified dated (Plaza et al. 2016). The objective of the
activated carbon was used to model the CO2 current study is to carry out the experimental
adsorption while focusing on breakthrough and simulation study of CO2 sorption from a
behavior and kinetic analysis, and this model effi- mixture of CO2/N2 using a fixed-bed column
ciently predicts the experimental data (Shafeeyan with activated carbon as adsorbent. The effect of
et al. 2015). Activated carbon and MOF-77 were temperature, feed rate, and CO2 concentration on
used to study two-dimensional modeling for the breakthrough curve was analyzed. The break-
adsorption of CO2/N2 and CO2/H2 mixtures, and through curves were reproduced very well using
the influence of the aspect ratio and geometric a LDF mass transfer model. The validity of the
shapes were analyzed (Ben-Mansour et al. 2017). model was judged in terms of the determination
The modeling of multicomponent breakthrough coefficient and mean error.
curves on BPL-activated carbons was performed
at a temperature range of 298–338 K and com- Experimental
pared with good accuracy (Brea et al. 2017).
Adsorption of CO2 using various adsorbents A granular activated carbon (NORIT-RB1) with
was examined and the maximal capacity was an average diameter of 1.83 mm was used as the
reported for activated carbons (Lu et al. 2008). sorbent for data collection. The experimental
The parametric study was conducted and break- setup used for conducting the experiments is
through behavior was observed, with specific depicted in Figure 1. The SS packed column is
emphasis on the mass transfer zone by using two jacketed to control the desired sorption tempera-
grades of activated carbons using a fixed-bed col- ture. The type-K 6 thermocouples are situated
umn (Al Mesfer et al. 2018; Al Mesfer and along the axial direction of the column to analyze
Danish 2018). The breakthrough curves were the temperature profiles during the sorption–de-
reproduced through LDF approximation, includ- sorption cycle. A PID controller equipped with a
ing momentum and energy balance (Dantas et al. main unit controls the set point temperature. The
2011). The experimental and modeling study for flow rates of N2 and CO2 are adjusted using the
 CHEMICAL ENGINEERING COMMUNICATIONS 3

Table 1. Physical property of an adsorbent and the bed. axial dispersion coefficient (m2/s), qp: adsorbent
Bed length, Lb 0.25 m density (kg/m3), e: bed void fraction, q: the cap-
Bed internal diameter, Di 0.045 m
Bed external diameter, De 0.058 m acity (kg/kg), and u: gas velocity (m/s) through
Adsorbent source Norit RB2 the packed bed. It is approximated that equilib-
Form of adsorbent Extruded
Average length of adsorbent, Lad 3.86 mm rium can be adequately explained by the
Average diameter of adsorbent, Dad 1.83 mm
Apparent density, q 820 kg/m3
Freundlich isotherm. The isotherms adequately
Mass of adsorbent, M 0.230 kg agreed with the Freundlich model for 0–80 kPa
Bed porosity, E 0.266
Bed volume, V 0.306
103 m3 q ¼ aPb (2)
where a and b are the Freundlich model’s param-
flow controllers F1 and F2. The carbon dioxide con-
eters. The initial and boundary conditions can be
centration in the column outlet stream is measured appropriately written as follows:
by the IR sensor. The maximal value of 5% CO2
C ¼ C0 , for z ¼ 0 and t > 0 (3)
was chosen as IR detector works well in the CO2
concentration range of 0–5% and recommended @C
¼ 0 for z ¼ L and t > 0 (4)
for best results. The characteristics of the column @z
and sorbent used for performing the experiments where C0 is the initial CO2 concentration (kg/m3)
have been summarized and depicted in Table 1. and q0 is the initial concentration-adsorbed phase
(kg/kg).
Model description Perfect plug flow is assumed and, as reported
(Raghavan and Ruthven 1983), this assumption is
Predictions of the response curves rely on the valid if the ratio of bed length (L) and mean par-
mathematical model. The adsorption bed was ticle diameter (dp) is >20. Thus, the axial disper-
filled with porous adsorbent of activated carbon. sion can be ignored and Equation (1) reduces to
The adsorption bed undergoes a mass transfer the following equation:
resistance, solid diffusion resistance, and axial @C @C ð1  eÞ @q
dispersion. The following approximations were u þ þ qp ¼0 (5)
@z @t e @t
used to develop the model:
The above equation is the function of two varia-
i. The bed operates under isothermal conditions bles. The bed is approximated as n-series or simi-
with single-component adsorption. lar units to eliminate the influence of the space
ii. No chemical reaction occurs. variable. The gas volume in each unit is assumed
iii. Constant conditions of temperature, flow rate, to be equal to t (m3)
feed concentration, and pressure are achieved. V sL
t¼ ¼ (6)
iv. The gas-phase behaves ideally. n n
v. Radial temperature, velocity gradient, and Identical chemical and physical characteristics are
concentration are negligible. assumed within each unit. The number of units
vi. No build-up of N2 occurs in the pores of of constant volume can be determined as:
the adsorbent. peax
n¼ þ1 (7)
vii. The bed dynamics were presented by the 2
LDF model. where Peax stands for the axial Peclet number.
Taking accumulation into account, the equation
With the help of the above approximations, reduces to
the following one-dimensional Equation (1) dC dq
describes the mass balance with axial dispersion: QðC0  CÞ ¼ eV þ qp Vð1  eÞ (8)
dt dt
@2C @C @C ð1  eÞ @q The LDF model is often applied for fixed-bed
Dax þu þ þ qp ¼0 (1)
@z 2 @z @t e @t carbon dioxide (CO2) adsorption. Based on the
where C: CO2 concentration (kg/m3), z: axial LDF model, the adsorbed rate is determined by
coordinate (m), t: the adsorption time (min), Dax: the difference between the adsorbed phase
 4 M. K. A. MESFER ET AL.

concentrations and equilibrium (Gkueckauf Results and discussion

1955). The rate of mass transfer is determined by Sorption temperature
Equation (9)
The breakthrough response curves obtained
dq

¼ kLDF q  q (9) experimentally are reproduced in Figure 2(a).
dt
The experimental breakthrough curves have been
Considering the 1st and nth cells, the mass bal- generated at different temperatures of 298, 308,
ance can be expressed as follows:$60#?accolade 318, and 328 K with a flow rate of 3 L/min
"acc357"?$62# (Co ¼ 5%). The breakthrough time varies consid-
V dC1 V dq1 erably with temperature. It has been reported
QðC0 C1 Þ ¼ e þqp ð1eÞ (10)
n dt n dt that the breakpoint times declined remarkably
Unit 1 dq
 with increased temperatures. The breakthrough
¼ kLDF q1 q1
1
(11) time of 27.50 min was obtained at a temperature
dt
of 298 K and reduced to 22.62 min upon increas-
V dCn V dqn
QðCn1 Cn Þ ¼ e þqp ð1eÞ (12) ing temperature to 308 K. The declined break-
n dt n dt
Unit n dq through period of 16.60 min was realized with an


¼ kLDF qn qn
n increased temperature at 328 K. The prolonged
(13)
dt breakpoint time at the reduced temperature was
The numerical solution was obtained for the PDE attributed to an increased sorption capacity. The
by applying dimensionless coordinates longest breakpoint was observed for the response
C qðCÞ curve obtained at a temperature of 298 K. The
X¼ and Y ¼ (14) breakthrough occurs early since the temperature
C0 q0  ðC0 Þ
 b
of the sorption increases. The adsorption capacity
q C of 0.796 mmol/g was achieved at 298 K and a

q ¼ aC ;b
q0 ¼ aC0b 
Y ¼ ¼ ¼ X b (15)
q0 C0 feed rate of 3 L/min. The extremely steep break-
 through curves obtained at different temperatures
dX1 q ðC0 Þ dY1
Unit1 Qð1X1 Þ¼et þqb tð1eÞ indicate that most of the capacity is used at the
dt C0 dt
breakpoint. For the economical separation of
(16)
dY1 CO2, utilization of most of the capacity at break-
¼kLDF ðY  ðX1 ÞY1 Þ Y  ðXÞ¼X b (17)
dt point is desirable.
For nth unit$60#?accolade "acc359"?$62# The comparative experimental and simulated
curves using the LDF model obtained at different
dXn q ðC0 Þ dYn
QðXn1 Xn Þ ¼ et þqb tð1eÞ sorption temperatures of 298, 308, 318, and
dt C0 dt
Unit n 328 K and keeping constant parameters are
(18) depicted in Figure 2(b)–(e). The feed flow rate
dYn
¼ kLDF ðY  ðXn ÞYn Þ Y  ðXÞ ¼ X b (19) was controlled at 3 L/min. The LDF model was
dt used for predicting the breakthrough response.
Relying on the model represented by Equations Significantly good agreement was observed
(18) and (19), initial and boundary conditions between the experimental and predicted results.
were used for developing the mathematical The LDF model reproduced the breakthrough
model. The numerical solution was obtained for curves very well. Insignificant deviations were
the developed model using an explicit finite-dif- observed between the predicted and experimental
ference scheme. The computation was performed results near the bed-saturation point. In general,
in MATLAB R2018a software. By minimizing the the model accurately predicted breakthrough
P
least-squares coefficient d ¼ ðC exp  Cmod Þ2 , behavior. The validity of the simulated results
the mass transfer constant (kLDF ) was calculated was judged in terms of R2 and r for each case.
with the best fitting application of the experimen- The determination coefficient and mean error
tal data to the mathematical model. equal R2 ¼ 0.9785 and r ¼ ±0.0345 were obtained
 CHEMICAL ENGINEERING COMMUNICATIONS 5

Figure 2. (a) Experimental breakthrough curves at different temperatures. (b)–(e) Simulated and experimental breakthrough curves
with F ¼ 3 L/min. (b) T ¼ 298 K, (c) T ¼ 308 K, (d) T ¼ 318 K, (e) T ¼ 328 K.

for the breakthrough curve determined at a tem- ±0.0266 and ±0.0256 were estimated at tempera-
perature of 298 K. tures of 308 and 318 K, respectively. Nearly the
On the other hand, R2 ¼ 0.9835 was determined same coefficient of determination R2 ¼ 0.9826 was
at a temperature of 308 K. The mean errors of obtained at an increased temperature of 328 K. The
6 M. K. A. MESFER ET AL.

Table 2. R2 and r values determined at different (vol.%), respectively. The coefficient of determin-
temperatures. ation was estimated to be equal to 0.9748 at an
Temperature, Feed rate, CO2 concentration, Coefficient of Mean error,
T (K) Q (L/min) Co (vol.%) determination, R2 r (vol.%) increased flow rate of 4 L/min, whereas the mean
298 3 5 0.9785 ±0.0345 error was ±0.0321 (vol.%). The determined R2 was
308
318
3
3
5
5
0.9835
0.9839
±0.0266
±0.0256
nearly the same (R2 ¼ 0.9751) at 5 L/min. A sum-
328 3 5 0.9826 ±0.0369 mary of the R2 values of the experimental and
simulated results is depicted in Table 3. The deter-
mination coefficient (R2) was greater than 97% in
summary of R2 and r determined at different tem- all cases under different operating conditions,
peratures is depicted in Table 2. The determination whereas the mean error in the measurement of exit
coefficient value (R2) is higher than 97% in all the CO2 concentration was <4%. It can be claimed
cases, thus signifying the strong correlation that the breakthrough curves were accurately
between the experimental and predicted results. reproduced using LDF model.
The maximal R2 value of 0.9839 was determined at
a temperature of 308 K. The minimal mean error Initial CO2 concentration
was obtained at the same sorption temperature of
308 K. The estimated value of mean error was less The experimental breakthrough curves generated at
than 5%. The agreement between the experimental different CO2 initial concentrations are depicted in
and simulated breakthrough curves is satisfactory. Figure 4(a) under a fixed temperature of 328 K. The
Complete overlapping is observed between the feed rate was fixed at 5 L/min. The increased CO2
experimental data points and the simulated concentration leads to a reduced breakthrough time.
approach. It is evident that the model predicts the The initial CO2 concentration corresponds to the
breakthrough response curves satisfactorily. concentration of CO2 in the CO2 þ N2 feed mixture.
CO2 concentration fixed at 1%, 1.5%, 2%, and 5% of
total feed rate (5 L/min) with balance N2 flow. The
Feed flow fates maximal value of 5% CO2 was chosen as the IR
The experimental breakthrough curves obtained at detector used to measure the exit column CO2 con-
different feed rates are depicted in Figure 3(a). The centration works effectively in the 0–5% CO2 range
temperature was fixed at 328 K to collect the data. and is recommended for best results. The maximal
The initial CO2 concentration in the feed was con- breakthrough time was attained at a concentration
trolled at 5% (vol.%). The three different feed rates of 1%. The minimal breakthrough time was
of 3, 4, and 5 L/min were selected to study the effect observed at a CO2 concentration of 5%. The break-
on the breakthrough behavior. The breakthrough through time decreased remarkably with a raised
time declined with increased feed rates from 3 to inlet CO2 level in feed. The adsorption capacity
5 L/min. The breakthrough time of 16.70 min increases with increased initial CO2 concentration
observed at a feed of 3 L/min and reduced to in feed. The minimal adsorption capacity of
12.63 min at 5 L/min. The adsorption capacity 0.165 mmol/g was determined at Co ¼ 1% with a
becomes enhanced with an increased flow rate. The temperature of 328 K. The increased concentration
increased feed rate of 5 L/min contributes to an at 2% contributed to an increased capacity of
adsorption capacity of 0.553 mmol CO2/g at a tem- 0.276 mmol/g. The adsorption capacity determined
perature of 328 K, whereas capacity declines to under various operating conditions have been sum-
0.488 mmol/g with reduced feed rate to 4 L/min. marized and depicted in Table 4. The very steep
The simulated results of the mathematical model adsorption curves illustrate that a significant bed
based on LDF approximations are depicted in capacity is used at the breakpoint.
Figure 3(b)–(d) and are in close agreement with The simulated and experimental breakthrough
the obtained experimental findings. In Figure 3(b) profiles obtained at different CO2 inlet concentra-
at 3 L/min, a strong correlation exists between the tions are depicted in Figure 4(b)–(e). The feed
experimental and simulated data. R2 and r values flow and temperature were fixed at 5 L/min and
are determined as equal to 0.9825 and 0.0369 328 K, respectively. The R2 and r values of
 CHEMICAL ENGINEERING COMMUNICATIONS 7

Figure 3. Simulated and experimental breakthrough curves at T ¼ 328 K. (a) Experimental, (b) simulation at Q ¼ 3 L/min, (c) simula-
tion at Q ¼ 4 L/min, (d) simulation at Q ¼ 5 L/min.

Table 3. R2 and r determined at different flow rates.

a maximal concentration of 5% were found to be
Temperature, Feed rate, CO2 concentration, Coefficient of Mean error,
equal to 0.9751 and ±0.0314 (CO2 vol.%).
T (K) Q (L/min) Co (vol.%) determination, R2 r (vol.%) A summary of R2 and r values obtained under
328 3 5 0.9825 ±0.0369 different operating conditions is depicted in Table 5.
328 4 5 0.9748 ±0.0321
328 5 5 0.9751 ±0.0314 In all cases, the reported R2 values of more than 97%
demonstrate the strong correlation between the
experimental and simulated data. It can be con-
0.9794 and 0.0307 (vol.%) were determined at a cluded that the LDF model under the stated
concentration of 1%. The CO2 concentrations of assumptions accurately reproduced the break-
1.5% and 2% resulted in nearly the same values through curves. A mean error of less than 5% signi-
of R2 at 0.9756 and 0.9709, respectively. The fies the strong correlation between the experimental
mean errors at 1.5% and 2% concentrations were and predicted results. Table 6 depicts the operating
±0.0447 (vol.%) and ±0.0496 (vol.%), respectively. conditions and estimated model parameters. The
The determination coefficient and mean error at mass transfer coefficients were determined for the
8 M. K. A. MESFER ET AL.

Figure 4. (a) Experimental breakthrough curves at different Co (T ¼ 328 K, Q ¼ 5 L/min). (b)–(e) Simulated and experimental break-
through curves at T ¼ 328 K and Q ¼ 5 L/min. (b) 1% CO2, (c) 1.5% CO2, (d) 2% CO2, (e) 5% CO2.

LDF model Equation (11) under different operation model parameters a and b were estimated for the
conditions. Fifty unit cells were selected for per- Freundlich model and were very well applicable in a
forming the numerical simulation of the model. The low-pressure range.
 CHEMICAL ENGINEERING COMMUNICATIONS 9

Table 4. Summary of adsorption capacity determined under operating conditions.

Feed rate, Initial CO2 Adsorption
Temperature, T (K) F (L/min) concentration, Co (vol.%) capacity, q (mmol/g)
298 3 5 0.796
308 3 5 0.615
318 3 5 0.492
328 3 5 0.439
328 4 5 0.488
328 5 5 0.553
328 5 1 0.165
328 5 1.5 0.237
328 5 2.0 0.276

Table 5. R2 and r values determined at initial CO2 concentrations.

Feed rate, CO2 concentration, Coefficient of Mean error,
Temperature, T (K) Q (L/min) Co (vol.%) determination R2 r (vol.%)
328 5 1 0.9794 ±0.0307
328 5 1.5 0.9756 ±0.0447
328 5 2 0.9709 ±0.0496
328 5 5 0.9751 ±0.0314

Table 6. A summary of operating conditions and estimated model parameters.

Temperature, T (K) 298 308 318 328 328 328 328 328 328
Flow rate, Q (L/min) 3 3 3 3 4 5 5 5 5
Initial concentration, Co (g/L) 0.071 0.068 0.066 0.064 0.078 0.09 0.018 0.029 0.035
Mass transfer coefficient, K (1/min) 5 7 9 12 14 16 10 12 13
Number of cell, n 50 50 50 50 50 50 50 50 50
Freundlich model parameter, a 0.2 0.18 0.17 0.17 0.17 0.17 0.17 0.17 0.17
Freundlich model parameter, b 0.64 0.66 0.68 0.74 0.74 0.74 0.74 0.74 0.74
Coefficient of determination, R2 0.9785 0.9835 0.9839 0.9825 0.9748 0.9751 0.9794 0.9709 0.9751

Conclusions Funding
The fixed-bed sorption using activated carbon for This work was supported by the Deanship of Scientific
CO2 separation from CO2/N2 feed was studied. Research [Grant No. 310/40], King Khalid University,
Abha, K.S.A.
The experimental breakpoint time decreased with
increased temperature, feed rate, and inlet con-
centration of CO2. A LDF model was applied for
simulating the breakthrough curves, and excellent References
conformity was observed. The close agreement of Al Mesfer MK, Danish M. 2018. Breakthrough adsorption
the simulated and experimental results was ana- study of activated carbons for CO2 separation from flue
gas. J Environ Chem Eng. 6(4):4514–4525. doi:10.1016/j.
lyzed in terms of R2 and r. A R2 of more than
jece.2018.06.042
0.97 was determined under different conditions Al Mesfer MK, Danish M, Fahmy MY, Rashid MM. 2018.
of operating parameters. The acceptability of the Post combustion CO2 capture with activated carbons
proposed model was contemplated by the mean using fixed bed adsorption. Heat Mass Transfer. 54(9):
error of less than 5%. In all cases, the predicted 2715–2723. doi:10.1007/s00231-018-2319-1.
Barcia PS, Bastin L, Hurtado EJ, Silva J, Rodrigues AE,
and experimental breakthrough curves were in
Chen B. 2008. Single and multicomponent sorption of
close agreement with the experimental data. The CO2, CH4 and N2 in a microporous metal-organic frame-
LDF model with the stated assumptions satisfac- work. Sep Sci Technol. 43(13):3494–3521. doi:10.1080/
torily explains the carbon dioxide break- 01496390802282347.
through curve. Bedel A, Vallieres C, Latifi MA. 2017. Parameter estimabil-
ity analysis and identification for adsorption equilibrium
models of carbon dioxide. Adsorption. 23(2–3):373–380.
Declaration of interest statement doi:10.1007/s10450-017-9864-7.
Ben-Mansour R, Basha M, Qasem N. 2017.
There is no conflict of interest to declare. Multicomponent and multi-dimensional modeling and
10 M. K. A. MESFER ET AL.

simulation of adsorption-based carbon dioxide separ- mass transfer. Sep Puri Technol. 85:17–27. doi:10.1016/j.
ation. Comput Chem Eng. 99:255–260. doi:10.1016/j. seppur.2011.07.038.
compchemeng.2017.01.040. Park J, Lee JW. 2016. Dynamic modeling of fixed-bed
Brea P, Delgado JA, Agueda VI, Uguina MA. 2017. adsorption of flue gas using a variable mass transfer
Modeling of breakthrough curves of N2, CH4, CO, CO2 model. Korean J Chem Eng. 33(2):438–447. doi:10.1007/
and a SMR type off-gas mixture on a fixed bed of BPL s11814-015-0180-1
activated carbon. Sep Pur Technol. 179:61–71. doi:10. Plaza MG, Duran I, Querejeta N, Rubiera F, Pevida C.
1016/j.seppur.2017.01.054. 2016. Experimental and simulation study of adsorption in
Casas N, Schell J, Pini R, Mazzotti M. 2012. Fixed bed postcombustion conditions using microporous biochar.1.
adsorption of CO2/H2 mixture on activated carbon: CO2 and N2 adsorption. Ind Eng Chem Res. 55(11):
experiments and modeling. Adsorption. 18(2):143–161. 3097–3112. doi:10.1021/acs.iecr.5b04856
doi:10.1007/s10450-012-9389-z. Raghavan NS, Ruthven DM. 1983. Numerical simulation of
Dantas T, Amorim SM, Luna F, Silva IJ Jr, Azevedo Diana fixed bed adsorption column by the method of orthog-
CS, Rodrigues AE, Moreira R. 2009. Adsorption of car- onal collection. AIChE J. 29(6):922–925. doi:10.1002/aic.
bon dioxide onto activated carbon and nitrogen-enriched 690290608
activated carbon: surfaces changes, equilibrium, and mod- Serna-Guerrero R, Belmabkhout Y, Sayari A. 2010a.
eling of fixed-bed adsorption. Sep Sci Technol. 45(1): Modeling CO2 adsorption on amine-functionalized meso-
73–84. doi:10.1080/01496390903401762. porous silica: 1. A semi-empirical equilibrium model.
Dantas T, Luna F, Silva IJ Jr, Azevedo Diana CS, Grande
Chem Eng J. 161(1–2):182–190. doi:10.1016/j.cej.2010.04.
CA, Rodrigues AE, Moreira R. 2011. Carbon dioxide-
042.
nitrogen separation through adsorption on activated car-
Serna-Guerrero R, Belmabkhout Y, Sayari A. 2010b.
bon in a fixed bed. Chem Eng J. 169(1–3):11–19. doi:10.
Modeling CO2 adsorption on amine-functionalized meso-
1016/j.cej.2010.08.026
porous silica: 2. Kinetic and breakthrough curve. Chem
Dantas T, Luna F, Silva IJ Jr, Torres A, Azevedo D,
Eng J. 161(1–2):173–181. doi:10.1016/j.cej.2010.04.024.
Rodrigues AE, Moreira R. 2011. Modeling of the fixed-
Shafeeyan MS, Daud W, Shamiri A. 2014. A review of
bed adsorption of carbon dioxide and a carbon dioxide
mathematical modeling of fixed bed columns for carbon
and nitrogen mixture on zeolite 13X. Braz J Chem Eng.
dioxide adsorption. Chem Eng Res Des. 92(5):961–988.
28(3):533–535. doi:10.1590/S0104-6632000300018
doi:10.1016/j.cherd.2013.08.018.
Gkueckauf E. 1955. Theory of chromatography: formulas
Shafeeyan MS, Daud W, Shamiri A, Aghamohammadi N.
for diffusion into spheres and their application to chro-
matography. Trans Faraday Soc. 51:1540–1551. doi:10. 2015. Modeling of carbon dioxide adsorption onto
1039/TF9555101540 ammonia-modified activated carbon: kinetic analysis and
Li D, Zhou Y, Shen Y, Sun W, Fu Q, Yan H, Zhang D. breakthrough behaviour. Energy Fuels. 29(10):6565–6577.
2016. Experiments and simulation for separating CO2/N2 doi:10.1021/acs.energyfuels.5b00653.
by dual-reflux pressure swing adsorption process. Chem Won W, Lee S, Lee KS. 2012. Modeling and parameter esti-
Eng J. 297:315–324. doi:10.1016/j.cej.2016.03.075 mation for a fixed-bed adsorption process for CO2 cap-
Lu C, Bai B, Wu B, Su F, Hwang JF. 2008. Comparative ture using zeolite 13X. Sep Puri Technol. 85:120–129.
study of CO2 capture by carbon nanotubes, activated car- doi:10.1016/j.seppur.2011.09.056.
bons, and zeolites. Energy Fuels. 22(5):3050–3056. doi:10. Yoro KO, Singo M, Mulopo JL, Daramola MO. 2017.
1021/ef8000086 Modeling and experimental study of the CO2 adsorption
Monazam ER, Spenik J, Shadle LJ. 2013. Fluid bed adsorp- behavior of polyaspartamide as an adsorbent during post-
tion of carbon dioxide on immobilized polyethylenimine: combustion CO2 capture. Energy Procedia. 114:
kinetic analysis and breakthrough behaviour. Chem Eng 1643–1649. doi:10.1016/j.egypro.2017.03.1294.
J. 223:795–805. doi:10.1016/j.cej.2013.02.041. Zhang L, Yin Y, Li L, Wang F, Song Q, Zhao N, Xiao F,
Mulgundmath VP, Jones RA, Tezel FH, Thibault J. 2012. Wei W. 2016. Numerical simulation of CO2 adsorption
Fixed bed adsorption for the removal of carbon dioxide on K-based sorbent. Energy Fuels. 30(5):4283–4291. doi:
from nitrogen: breakthrough and modeling of heat and 10.1021/acs.energyfuels.5b02588.