JID: JTICE

ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

Journal of the Taiwan Institute of Chemical Engineers 000 (2021) 1
17

Contents lists available at ScienceDirect

Journal of the Taiwan Institute of Chemical Engineers

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

Energy-saving performance of the process modifications for carbon

capture by diluted aqueous ammonia
Jialin Liua,*, David Shan-Hill Wongb, Ding-Sou Chenc
a
Research Center for Smart Sustainable Circular Economy, Tunghai University, No. 1727, Section 4, Boulevard, Taichung, Taiwan
b
Department of Chemical Engineering, National Tsing Hua University, No. 101, Section 2, Kuang-Fu Road, Hsinchu, Taiwan
c
Green Energy and System Integration Research and Development Department, China Steel Corporation, No. 1, Chung Kang Road, Kaohsiung, Hsiao Kang Taiwan

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

Article History: The regeneration energies of CO2-lean solvent and NH3-washing water comprise the main fractions of the
Received 30 January 2021 total energy consumption in an NH3-based post-combustion carbon capture (PCC) process. However, the
Revised 7 June 2021 reconfiguration of the CO2 absorption process may significantly reduce the requirement for NH3-washing
Accepted 26 June 2021
water to relieve the energy burden of regenerating it. Additionally, the utilization of a pressurized CO2 strip-
Available online xxx
per to enhance the purity of CO2 in the overhead vapor is an alternative strategy for reducing the energy
requirement for regenerating the CO2-lean solvent. Moreover, further energy reductions can be achieved by
Keywords:
modifying the stripper configurations that have been reported in the literature. In this study, the energy-sav-
Post-combustion carbon capture
Aqueous ammonia solvent
ing performance of the modified configurations, including the utilization of a two-stage absorber, a cold-split
Two-stage absorber bypass (CSB) and an advanced flash stripper (AFS) for PCC by an NH3-based solvent is investigated. Our
Cold-split bypass results indicate that the concentration of NH3 slip can be reduced to a quarter of the emission by the two-
Advanced flash stripper stage absorber. Moreover, this study also demonstrates that the configuration of AFS is a limiting case of the
CSB modification. The energy requirement of AFS is higher than that of CSB because the heat recovery by the
main heat exchanger in AFS must be sacrificed to avoid solid precipitation in the reflux.
© 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction For example, Zhang and Guo [7] reported that the diameter and
packed height of the absorber must be 40 and 72 m, respectively, to
The increasing amounts of CO2 in the atmosphere, which is caused achieve 90% CO2 removal from the flue gas of a 500 MW coal-fired
by the burning of fossil fuels, contribute greatly to global warming. power plant. In subsequent work, they [8] suggested an NH3-abate-
Although amine-based solvents exhibit good capacities and high ment system in which the NH3 and CO2 strippers were separately
absorption rates, they are expensive, degrade in the presence of oxy- designed for the water-wash and CO2 capture loops. The overall
gen, and corrode equipment. Alternatively, NH3-based post-combus- regeneration energy was reported as 8.47 GJ/t-CO2 utilizing a lean
tion carbon capture (PCC) processes have recently attracted solvent with an NH3 concentration of 3 mol/L and CO2-lean loading
tremendous attention owing to their technical and economic advan- of 0.23 mol-CO2/mol-NH3. Liu et al. [9] reconfigured the design that
tages over the conventional amine-based processes. Several pilot was proposed by Zhang and Guo [7,8] and reported that the overall
plants, including the chilled NH3 process [1], Powerspan’s ECO2 pro- energy consumption, including the reboiler duties for the CO2 and
cess [2], and the Munmorah pilot plants [3,4], as well as that of NH3 strippers, could be reduced to 4.0 GJ/t-CO2 of which 3.46 GJ/t-
Pohang Iron and Steel Company (POSCO) [5], have been constructed CO2 would be utilized to regenerate the NH3-washing water.
to investigate the feasibilities of capturing CO2 by aqueous NH3. In the aforementioned configurations [7,8], since the CO2-lean sol-
Han et al. [6] reported that the NH3-based CO2 capture technology vents were fed into the CO2 absorber from the top, the free NH3 mole-
outperformed the amine-based systems particularly because of its cules in the solvents were easily slipped into the vapor, thereby
higher stability, lower corrosiveness, higher CO2 loading capacity, leaving the absorber. Jilvero et al. [10] adapted the two-stage absorp-
lower cost of the chemicals, and lower regeneration energy duty. tion design that was proposed by Darde et al. [11]. Therein, the CO2-
However, the major drawbacks, e.g., the slow absorption rate with lean solvent was fed into the middle of the absorber. The semi-rich
CO2 and the high volatility of NH3, necessitate the utilization of a solvent was drawn from the bottom of the absorber and fed from the
larger CO2 absorber and a washing column to control the NH3 slip. top after the intercooling process. Since most of the NH3 molecules
reacted with CO2 in the first stage of the absorption, the vapor pres-
sure of NH3 at the top of the absorber only depended on the tempera-
* Corresponding author.
E-mail address: jialin@thu.edu.tw (J. Liu). ture of the solvent after intercooling [10]. By implementing the two-

https://doi.org/10.1016/j.jtice.2021.06.060
1876-1070/© 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

2 J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17

stage absorption design, Liu [12] reported that the minimum overall LVC and the integration of LVC with CSB, respectively. Therefore, only
energy was 4.91 GJ/t-CO2 for both strippers that were operated at 0.4% additional energy saving was achieved utilizing the additional
atmospheric pressure in which the reboiler duty for the NH3 stripper flash drum and compressor in modified LVC. Liu et al. [15] also
was 0.32 GJ/t-CO2. Therefore, the reboiler duty for the CO2 stripper reported that the modification of RVC might not exhibit energy-sav-
dominated the overall energy consumption. ing potentials since the top temperature would not be significantly
Since the high temperature in the reboiler of the stripper facili- reduced by the flashed rich solvent as did the flashed lean solvent in
tates CO2 desorption from the rich solvent, the utilization of a pres- LVC. Furthermore, the increases in the electric powers of the com-
surized CO2 stripper is a common alternative for saving energy. pressor and pump cannot be compensated by reducing the reboiler
Li et al. [13] reported that the regeneration energy could be reduced duty in modified RVC.
to 3.27 GJ/t-CO2 if the stripper was operated at 10 atm. Furthermore, Recently, Jiang et al. [16] applied the AFS configuration to an NH3-
Li et al. [13] proposed three types of advanced stripper configura- based PCC process and reported that a 42% energy reduction was
tions: a cold-split bypass (CSB) process, an interheating (IH) process, achieved compared with the requirement of a traditional configura-
and a combination of both processes. In the CSB process, the regener- tion. The AFS configuration was developed for amine-based PCC pro-
ation energy was reduced to 2.89 GJ/t-CO2 at 10 atm. Compared with cesses [17,18]. The reboiler in a standard stripper was replaced by a
the energy consumption of the standard CO2 stripper, which was steam heater and flash vessel, which exhibited the same function as
reported as 3.27 GJ/t-CO2, the modification of CSB achieved a 12% the reboiler, except that the solvent was heated convectively. Since
energy reduction, and the reboiler duty of the IH process was reduced amine-based solvents tend to be thermally degraded at high temper-
to 3.0 GJ/t-CO2, which represented an 8.3% energy reduction com- atures, the convective steam heater with less solvent hold-up and
pared with that of a standard stripper that was operated at 10 atm. residence time would minimize the thermal degradation. However,
Furthermore, the combined process reduced the regeneration energy the thermal degradation of NH3-based solvents could be negligible
to 2.46 GJ/t-CO2. Li et al. [13] concluded that the combined process [6]. The benefit of applying AFS to NH3-based PCC was investigated
could reduce energy consumption by 25% more than the requirement here.
of a standard stripper. Put differently, the energy-saving performance More recently, Ishaq et al. [19] investigated the energy-saving
of the CSB process can be enhanced by 13% by integrating an inter- performance for NH3-based PCC by combining a CSB and a split flow
heater with the stripper. arrangement (SFA). In the SFA configuration, a portion of the semi-
Liu [12] demonstrated that the energy-saving performance of CSB lean solvent is drawn off from an intermediate stage of the CO2 strip-
was underestimated in the work of Li et al. [13] since the energy-sav- per and introduced to the absorber at the lower location respective
ing effects of adjusting the feed location of the warm rich solvent to the CO2-lean solvent. The semi-lean solvent is cooled before being
were neglected. Liu [12] reported that the reboiler duty of a standard fed into the absorber, which acts as a partial intercooler, for improv-
CO2 stripper that was operated at 10 atm is 3.60 GJ/t-CO2, which is ing the absorption capability and reducing the overall solvent
within ~10% of the relative errors of the results that were obtained by requirement. Ishaq et al. [19] reported that the minimum energy
Li et al. [13]. Further, the IH process required 2.82 GJ/t-CO2 of the requirement for the CSB can be reduced 15.8% by applying a split
reboiler duty, which is within ~6.0% of the relative errors of the fraction of 0.05, compared to that of the standard stripper. The energy
results that were obtained by Li et al. [13]. However, only 2.37 GJ/t- reduction could be further achieved 32.8% when combined the CSB
CO2, which is 18% lower than the values that were reported by Li and SFA modifications. Ishaq et al. [19] claimed that the less CO2-lean
et al. [13], was required for the CSB process. From an energy-saving solvent will be needed by implementing SFA due to the higher CO2-
perspective, Li et al. [13] also reported that there was an optimal loading in the rich solvent. Therefore, the reboiler duty can be
cold-split fraction in the CSB process. Rather than increase the reduced by combining the both configurations. In addition, Nguyen
amount of CSB, the temperature of the top of the stripper could be and Wong [20] proposed to integrate the rich and lean vapor recom-
reduced by moving the feed location downward of the warm rich sol- pression configurations, in which pressures of the rich and lean sol-
vent. Thus, Liu [12] reported that a reboiler duty of 2.37 GJ/t-CO2, vents are released to generate vapors. The vapors were recompressed
which represents a 34% energy reduction compared with that of a into the bottom section of the CO2 stripper for reducing the reboiler
standard stripper that was operated at 10 atm, was required for the duty. The authors [20] reported that the reboiler can be eliminated
CSB process. Furthermore, Liu [12] reported that the energy require- when the pressures of the stripper, rich and lean solvent flash drums
ment of the combined process is 2.35 GJ/t-CO2, which is a 35% energy are maintained at 10.5, 4.62, and 6.93 bar, respectively. The total
reduction compared with that of the standard process. Therefore, equivalent work, including a pump and two compressors, was
only a 1% energy-saving improvement was achieved by integrating reported as 0.069 kWh/kg-CO2. In contrast to the minimum energy
CSB with an interheater. requirement reported by Refs. [13,14,16], the converted total works
Ullah et al. [14] modified the stripper configurations, including were 0.166, 0.205, and 0.128 kWh/kg-CO2, respectively; 46%
66% of
lean vapor compression (LVC) and rich vapor compression (RVC), of energy reduction can be achieved by the work of Nguyen and Wong
the NH3-based processes for PCC. They reported that the energy-sav- [20].
ing performances of LVC and the integration of LVC and CSB were Al-Hamed and Dincer [21] comparatively reviewed the NH3-based
6.4% and 18.1%, respectively. Additionally, the energy reductions that carbon capture systems, which were published between 2015 and
were obtained by RVC and a combination of RVC and CSB were 3.4% 2020. These recent NH3-based carbon capture systems were classified
and 15.0%, respectively. Further, they reported that the minimum into three categories, namely renewable energy-based systems,
energy that was required by CSB was 13.7% lower than that, which energy savings-focused systems, and integrated gasification com-
was required by a standard stripper. However, the energy-saving per- bined cycle (IGCC)-based systems. Compared to conventional carbon
formance of CSB might have been underestimated since only the capture systems, the renewable energy-based systems are powered
cold-split fraction was adjusted during the minimization procedure by renewable energy sources, such as solar cells and wind turbines.
in their approach [14]. Liu et al. [15] reported that the purity of the The authors [21] indicated that the energy requirement by the solar-
top CO2 was enhanced by reducing the pressure of the flash drum in assisted and wind turbine-based systems are 2.29 [22] and 5.62 [23]
the LVC configuration, thereby reducing the warm rich temperature. GJ/t-CO2, respectively. The energy consumption of the solar-assisted
However, the electric power burden due to the compression of the systems seems to be feasible; however, the price of vacuum tube
flashed vapor into the stripper offset the energy savings. Compared solar collectors must be reduced for commercialization [22]. The tur-
with the energy requirement of the standard process, 10.0% and bine-based systems suffer the high energy penalty, because the
34.4% of the energy reduction were achieved by the modification of ammonia is not recycled and is produced by the system using the
Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17 3

Haber-Bosch synthesis process [23]. In addition, the authors [21] regenerated by the CO2 stripper in which CO2 is desorbed by heating
pointed out that 2.27 GJ/t-CO2 of the energy requirement is needed the rich solvent, after which the hot lean solvent (Stream 11) is
for the IGCC-based systems [24], which is comparable with that of recycled to HX to recover the sensible heat. Zhang and Guo [7]
the solar-assisted systems. Nonetheless, the IGCC belongs to the pre- reported that a reboiler duty of 5.75 GJ/t-CO2 was required to return
combustion carbon capture system that cannot be directly applied to the CO2 loading of the lean solvent to 0.23 mol-CO2/mol-NH3.
the existing coal fired power plants. In the category of energy sav- Since ~3.8% of the gas exiting the CO2 absorber (Stream 5) is NH3
ings-focused systems, the ASF possesses with the lowest energy con- (Table 1), which is much higher than the emission regulations
sumption, which is 2.02 GJ/t-CO2 reported by Jiang et al. [16]. (20
50 ppm) [8], an NH3 washing tower (the NH3 absorber) is
The remainder of this paper is organized, as follows: Section 2 required to control the NH3 slip. As shown in Fig. 1, the washing
describes the large-scale CO2 capture processes by diluted aqueous water (Stream 6) in which the lean solvent contained 0.02 M NH3
NH3; Section 3 validates the rate-based NH3
CO2
H2O model, which with a CO2 loading of 1.2 mol-CO2/mol-NH3, was fed into the top of
was developed in Aspen Plus version 10, by the pilot plant data that the NH3 absorber at 25 °C [8]. A small amount of pure CO2, which
were reported by Qi et al. [25]. Section 4 describes the modification was obtained from the product of the CO2 stripper, was injected into
of two-stage absorption with strippers to regenerate CO2- and NH3- the lean NH3 solvent to maintain the CO2 loading of the lean solvent,
lean solvents, in addition to establishing the energy requirement thereby enhancing the NH3 absorption reactions in the absorber. Fur-
baseline for the modifications of the stripper; Section 5 presents the ther, Zhang and Guo [8] reported that the NH3 concentration of the
details of the energy consumption in the advanced stripper configu- vent gas (Stream 7) could be controlled at 1.6 ppm, which is much
rations, which were reported by Jiang et al. [16]; and Section 6 lower than the emission regulations reported in Ref. [8]. The NH3-
presents the conclusions. rich solvent (Stream 8) recovers the sensible heat of the NH3-lean sol-
vent, and the warm rich solvent (Stream 9) reaches 85 °C and is fed
2. Large-scale CO2 capture process into the top of the NH3 stripper to regenerate the washing water by
vaporizing NH3 from the solvent. The concentrated NH3 vapor
Zhang and Guo [7,8] proposed a large-scale CO2 capture process (Stream 10) is passed through a total condenser where part of the
by an aqueous NH3 solvent for a 500 MW coal-fired power plant. condensed liquid flows back into the NH3 stripper, whereas the
Fig. 1 shows the process flow diagram in which the flue gas (Stream others are cooled by the warm rich solvent (Stream 4) and combined
1) was fed into the bottom of the CO2 absorber, and Table 1 lists the with the hot lean solvent (Stream 11) to maintain the NH3 concentra-
results reported in Ref. [7]. The lean solvent (Stream 2) was fed from tion of the CO2-lean solvent. Zhang and Guo [8] further suggested
the top of the absorber in which 3 M NH3 and a CO2-lean loading of that the condenser of the NH3 stripper and the reboiler of the CO2
0.23 mol-CO2/mol-NH3 were employed to capture 52% of CO2 from stripper must be operated at 91 °C. Thus, the condenser duty for the
the flue gas at 40 °C, for which the diameter and packed height of the NH3 stripper could be applied as the reboiler duty for the CO2 strip-
absorber were set to 12 and 20 m, respectively. Zhang and Guo [7] per. The reboiler duty for the CO2 stripper can be reduced to 3.05 GJ/
mentioned that if 90% of CO2 removal efficiency would like to be t-CO2 after the heat integration. The overall energy consumption of
achieved, the diameter and packed height of the absorber must be the CO2 capture process, including that of the NH3 abatement system,
set to 40 and 72 m, respectively. However, this size of absorber is not was reported to be 8.47 GJ/t-CO2 [8] of which 5.42 GJ/t-CO2 was
feasible in reality. The cold rich solvent (Stream 3) exiting the bottom applied to control the NH3 slip. The high energy consumption of the
of the absorber recovers the sensible heat of the hot lean solvent via NH3 stripper might originate from the following: (1) the extremely
the main heat exchanger (HX), and the temperature of the warm rich low NH3 concentration of the vent gas (Stream 7), which was
solvent (Stream 4) could reach 60 °C [7]. The rich solvent is reported to be 1.6 ppm, and (2) the massive amount of washing water
that is required because of the high NH3 concentration (3.8%) of the
vapor exiting the top of the CO2 absorber (Stream 5).

3. Simulation model validation

A rate-based NH3
CO2
H2O model, which was developed by the

Aspen Plus version 10 software based on the technical report of
Aspen Technology [26], was applied to simulate CO2 capture by
diluted aqueous NH3. The following kinetic reactions were applied to
the rate-based model:
CO2 þ OH
@HCO

3 ð1Þ

NH3 þ CO2 þ H2 O@NH2 CO2
þ H3 Oþ ð2Þ

The kinetic parameters, which were employed in the absorption
model, were reported by Pinsent et al. [27]. Additionally, the follow-
ing equilibrium reactions must be implemented in the rate-based
model:

HCO3
þ H2 O $ H3 Oþ þ CO3
2 ð3Þ

NH3 þ H2 O $ NH4 þ þ OH
ð4Þ

2H2 O $ H3 Oþ þ OH
ð5Þ

NH4 HCO3ðsÞ $ NHþ
4 þ HCO3 ð6Þ
Fig. 1. Flow diagram of the large-scale NH3-based PCC processes.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

4 J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17

Table 1
Properties of the flue gas and gas exiting the CO2 absorber [7,8].

Stream No. Pressure (atm) Temperature (°C) Flow rate (t/h) N2 (vol%) CO2 (vol%) H2O (vol%) NH3 (vol%)

1 1 50.0 2124.6 75.73 12.43 11.84

5 1 38.0 1875.6 83.34 6.50 6.34 3.82

Notably, the ammonium salts, ammonium bicarbonate that the stripper can be assumed to operate at equilibrium due to the
(NH4HCO3), would be precipitated at low temperatures (Eq. (6)). Yu high operating temperatures in the stripper. Therefore, an equilib-
et al. [4] reported that the precipitation of ammonium bicarbonate rium model was applied to simulate strippers in this study.
blocked the stripper condenser and reflux lines during the pilot tests.
The Redlich
Kwong equation of state and electrolyte Non-Random 4. Absorption process modification
Two-Liquid (NRTL) model were applied to compute the properties of
the vapor and liquid phases, respectively. As aforementioned, high energy is required to control the NH3
Following, the pilot plant data, which were presented by Qi et al. slip, which was reported by Zhang and Guo [8], probably because of
[25], were employed to validate the simulation model. Fig. 2 shows the the requirement of a very large amount of washing water. A two-
process flow diagram of the pilot plant. The flue gas containing CO2 and stage absorption design [10,11] could significantly reduce the wash-
the lean solvent was introduced into the bottom of the first absorber ing water requirement. Fig. 4 shows the process flow diagram of the
(ABS1) and the top of the second absorber (ABS2), respectively. The two-stage absorption design in which the lean solvent was fed into
inner diameters of the two absorbers were 0.6 m each, and Pall rings the top of ABS1 instead of that of ABS2 (Fig. 2). The two-stage absorp-
with diameters of 25 mm were utilized as the packing materials. There- tion configuration is advantageous because most of the free NH3 mol-
fore, the Onda correlations [28] of the mass transfer coefficients were ecules in the lean solvent would react with CO2 in ABS1, after which
applied, and the liquid hold-up in the packings was predicted by the the semi-rich solvent would be cooled to 15 °C and fed into the top of
correlations of Stichlmair et al. [29]. The packing height of each absorber ABS2. Therefore, the partial pressure of NH3 at the top of ABS2 solely
is 3.8 m. Qi et al. [25] suggested that each absorber was divided into 25 depended on the feeding temperature of the semi-rich solvent [32].
segments to simulate the pilot plant tests. Furthermore, the film resis- The properties of flue gas (Table 1) were applied to the simulation.
tance was considered for the vapor and liquid films, and liquid-film Therein, the gas flow was parallelized into three CO2 absorbers with
reactions without discretization points were applied. adequate inner diameters that were determined by the simulator.
Fig. 3 shows the comparison of the simulation results with the The diameter and height of the column were 12 and 24 m, respec-
experimental data [25]. The overall CO2 absorption rate was overesti- tively, and Pall rings (25 mm) were utilized. Furthermore, the average
mated by the rate-based model employing the liquid hold-up, which pressure drop in the column was reported as 15.7 mm-water/m by
was predicted in Ref. [29] (Fig. 3(a)). The relative deviations of the the simulator. The flow rate of lean solvent was adjusted to achieve
overall CO2 absorption rate could be adjusted to §15% by reducing 90% of CO2 absorbed by the solvent. Fig. 5 shows the comparison of
the scale factor to 0.6 for the liquid hold-up. The predictions of the the simulation results of the 90% CO2 removal rate by the conven-
NH3 slip from the top of ABS1 were slightly affected by the adjust- tional (Fig. 2) and modified (Fig. 4) configurations, in which 10% of
ment of the hold-up scale factor (Fig. 3(b)). However, the deviations CO2 in the flue gas was vented from the top of ABS2. The liquid-to-
in the predicted NH3 concentration were §20% in most cases. The gas flow ratios of both configurations by weight (L/G) were similar
acceptable predictions were compared with the experimental data (Fig. 5(a)). However, Fig. 5(b) shows the comparison between the
[25], which correlated well with the rate-based model, to evaluate concentrations of the NH3 slips from the top of ABS2 of the two-stage
the energy-saving performance of the modified process. In addition, and conventional absorber configurations. Therein, the NH3 concen-
Fig. 3(c) compares the experimental data of regeneration energy in tration of the two-stage absorber was only a quarter of that of its con-
the stripper [30] with the simulation results, in which the rate-based ventional configuration counterpart. The results indicate that the
and equilibrium models were applied. The simulation parameters for two-stage modification might significantly reduce the amount of
the rate-based model can be found in Ref. [30] and the equilibrium washing water. In addition, Fig. 5(b) shows that the high ammonia
model consisting of 10 stages was applied in Fig. 3(c). The results concentrations were slipped for the conventional absorber in the low
show that the prediction errors by the both models were within CO2 lean loadings that led to the higher L/G needed compared with
§20% in most cases. In addition, Garđarsdo  ttir et al. [31] reported the counterpart of the two-stage absorber, as shown in Fig. 5(a). On
the other hand, in the high lean loading region, the two-stage
absorber needed the higher L/G than that of the conventional
absorber, since the driving force of CO2 absorption at the top was
lower in the two-stage absorber. Therefore, the absorption perfor-
mance (L/G) was affected by the two competitive factors, the slipped
ammonia amount and the absorption driving force.
Liu [12] integrated the two-stage absorption design [10,11] and
proposed a flow diagram for a large-scale PCC process (Fig. 6). In this
system, the flue gas (Stream 1) was fed into the bottom of ABS1 (CO2
ABS1), and the lean solvent (Stream 2) was fed into the column from
the top. To reduce the NH3 slip concentration, the semi-rich solvent
(Stream 3) was cooled to 15 °C, as suggested by Ref. [32], before it
was introduced to the top of the ABS2 (CO2 ABS2), where the gas
flow from the top of ABS1 (Stream 4) was fed from the bottom to fur-
ther remove CO2.
Next, the CO2-reduced flow (Stream 5) was scrubbed in the wash-
ing tower (NH3 absorber) by recycled washing water (Stream 6) to
Fig. 2. Flow diagram of the pilot plant that was reported in Ref. [25]. maintain the concentration of the NH3 slip at ˂10 ppm before the
Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17 5

Fig. 3. Comparison of the simulation results with the pilot plant data. (a) Overall CO2 absorption rate and (b) NH3 slip from the top of ABS1, and (c) regeneration energy.

cleaned gas (Stream 7) was discharged to the atmosphere. The CO2- a cold rich solvent (Stream 8) and mixed with the CO2-lean solvent
rich solvent (Stream 8) and NH3-rich water (Stream 9) were regener- (Stream 2), which was fed into the first CO2 absorber. Since a slight
ated by the CO2 and NH3 strippers, respectively. At the top of the NH3 amount of CO2 was absorbed by the NH3 absorber, the trace of CO2
stripper, the concentrated NH3 vapor (Stream 10) was condensed by was mixed into the lean solvent and fed into the CO2 absorber. How-
ever, the CO2 stripper was designed to bring the CO2 loading in
Stream 2 back to the initial setting. The trace of CO2 from the NH3
absorber would be compensated by the boilup ratio of the CO2 strip-
per. The detailed information about the material streams in Fig. 6 can
be found in Appendix A.
The sensible heat of the hot lean solvent (Stream 11) was recovered
from the HX to preheat the CO2-rich solvent (Stream 12), which was
pumped up to the operating pressure of the CO2 stripper. The mini-
mum temperature difference of HXs was set to 5 °C. Considering the
formation of ammonium bicarbonate in the reflux of the condenser,
the reflux (Stream 13) was heated by the warm rich solvent (Stream
14) to liquefy the solid precipitate before it entered the CO2 stripper.
The overhead vapor (Stream 15) was condensed at 35 °C, after which
high-purity CO2 (>99%) was collected and compressed for storage.
Table 2 summarizes the design specifications for the PCC process in
Fig. 6, whereas the respective adjusted parameters are described.
Since the desorption of CO2 in the stripper is an endothermic reac-
tion, a high temperature in the reboiler would favor the desorption
reactions, i.e., the high operating pressure in the CO2 stripper would
Fig. 4. Flow diagram of the two-stage absorption design that was reported in Refs. reduce the energy consumption. Moreover, since the mass transfer
[10,11]. rates might not affect CO2 and NH3, which were desorbed from the
Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

6 J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17

Fig. 5. Comparison of the simulation results of the two-stage and conventional absorbers. (a) L/G and (b) NH3 slip concentration.

Fig. 6. Flow diagram of the two-stage absorption design for PCC [13].

solvents at the high temperature that was reported by Garđarsdo ttir pressure of the CO2 stripper was assumed to be 10 atm [12,13]. Con-
et al. [31], an equilibrium model consisting of 10 stages was applied sidering that a lean solvent with a high NH3 concentration might
to simulate the two strippers. However, the electrical power of the favor the formation of ammonium bicarbonate in the reflux of the
pump might offset the energy reduction by increasing the operating stripper, the diluted NH3 with a concentration of 3 mol/L was applied
pressure of the stripper. Furthermore, the reboiler temperature was to the lean solvent that was reported in Refs. [7,8,12]. Fig. 7 shows
limited to 145 °C because low-pressure steam with a temperature of the energy that is required to achieve the design targets of 90% CO2
~160 °C was employed as the heat source. Therefore, the operating removal rate and 10 ppm NH3 concentration in the cleaned gas

Table 2
The design specifications and adjustive variables for the two-stage absorption design.

Design specification Adjusted variable

CO2 absorbers Capture 90% of CO2 from the flue gas (Stream 1) CO2-lean solvent flow rate (Stream 2)
NH3 absorber 10 ppm NH3 in the cleaned gas (Stream 7) NH3 washing water flow rate (Stream 6)
CO2 stripper Bring the CO2 loading in the rich solvent (Stream 8) back to the value of the CO2-lean solvent (Stream 2) Boilup ratio in the stripper
NH3 stripper Bring the NH3 concentration in the rich flow (Stream 9) back to the value of the NH3 washing water (Stream 6) Boilup ratio in the stripper

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17 7

the top of the CO2 stripper, while the remaining cold rich solvent was
utilized to recover the sensible heat of the hot lean solvent. Thereaf-
ter, the warm rich solvent was fed into the stripper at a location that
was lower than that of the bypassed cold rich solvent. As reported in
Refs. [12,15], the feed stage of the warm rich solvent significantly
affects the energy requirement of the CSB process.
When capturing a fixed amount of CO2, the overhead vapor
requirement is inversely proportional to the purity of CO2. Therefore,
increasing the purity of the top of CO2 might effectively reduce the
amount of vaporized solvent by the reboiler and achieve the energy-
saving purpose. Fig. 9 shows that the CO2-lean loading and CSB frac-
tion were set to 0.24 and 0.05, respectively. As aforementioned, ten
equilibrium stages were employed to simulate the stripper, and the
reboiler was the tenth stage. Additionally, the condenser was simu-
lated by a flash vessel, in which the constant temperature at 35 °C
and zero pressure drop were set. Meanwhile, the boilup ratio in the
reboiler was adjusted to bring the CO2-rich loading back to the lean
Fig. 7. Energy requirement for a large-scale PCC process reported in Ref. [12].
loading, which is 0.24 in this case. Fig. 9(a) shows the top tempera-
ture and corresponding purity of CO2. Therein, the top temperature
employing different CO2-lean loading (mol-CO2/mol-NH3). The mini- could be reduced as the feed location was moved downward from
mum energy requirement was obtained at a CO2-lean loading of 0.26, the top to the eighth stage. Consequently, the CO2 purity at the top
where the required energies for the CO2 and NH3 strippers were 3.60 was enhanced to the maximum limit, which was ~0.99. Fig. 9(b)
and 0.31 GJ/t-CO2, respectively. Additionally, the equivalent thermal shows the reboiler duty and vaporized ratio which is the mass flow
energy of the pump, which was converted by a 35.6% thermal effi- ratio of the overhead vapor to the captured CO2. As shown in Fig. 9
ciency [13], was 0.11 GJ/t-CO2. The simulation results at CO2-lean (b), the reboiler duty was reducible since only a smaller amount of
loading of 0.26 are listed in Appendix A. Since the required energy of the solvent was vaporized. However, Fig. 9(b) indicates that the
the CO2 stripper dominated the overall energy consumption, as reboiler duty bounced back when the warm rich solvent was fed
shown in Fig. 7, the reduction of energy by modifying the CO2 strip- above the reboiler (the last stage), thus reducing the purity of the top
per configuration was investigated under the optimal operating con- CO2, as shown in Fig. 9(a). Therefore, a minimum energy requirement
ditions of the standard process (Fig. 6). existed once the warm rich solvent was fed above a proper stage,
even though the CO2-lean loading and CSB fraction remained the
5. Desorption process modification same.
Fig. 9(c) shows the mole fraction of NH3 at the top and the solid
Liu et al. [15] reported that the energy reduction that was fraction of the reflux. As shown in the figure, the precipitation of
achieved by CSB was underestimated in the literature [13,14] since ammonium bicarbonate in the reflux was suppressible, as the top
the effect of the warm rich feed stage was ignored. In addition, CSB NH3 concentration approached zero. However, the NH3 concentration
can reduce 34% of the energy requirement compared with a standard at the top also bounced back once the warm rich solvent was fed
stripper when the proper feed stage of a warm rich solvent is selected above the last stage (the reboiler). Consequently, solid precipitate
[15]. Recently, Jiang et al. [16] reported that a 42% energy reduction could be found in the reflux, as shown in Fig. 9(c). Yu et al. [33]
could be achieved by the AFS configuration compared with the obtained similar results and reported that the accumulation rate of
requirement for a standard stripper. In this study, the energy-saving the solid in the reflux notably increased when the warm rich solvent
performances of CSB and AFS were investigated. was fed near the bottom of the stripper. Additionally, the solid pre-
cipitation could be avoided once the proper feed stage of the warm
5.1. CSB rich solvent was located [33]. As shown in Fig. 9(c), the solid fraction
of the reflux approached zero once the feed location of warm rich sol-
Fig. 8 shows the process flow diagram of the CSB modification, vent was between the fourth and the ninth stages. Since the mole
where a fraction of the cold rich solvent was bypassed and fed into fraction of NH3 in the overhead vapor was extremely low in these
cases, the precipitation reaction in (6) might not proceed.
Fig. 10 shows the column profiles of the simulations in Fig. 9. As
shown in Fig. 10(a), the top temperature could be reduced as the feed
stage moved downward, except where the warm rich solvent was
fed above the last stage. Since the temperature of the feed stage was
higher than that of the warm rich solvent, NH3 in the warm rich sol-
vent was easily vaporized at a high temperature. Therefore, the total
NH3 concentration of the liquid phase began to soar above the feed
stage, as shown in Fig. 10(b), where the total NH3 concentration was
the sum of the mole fractions of NH4+, NH2CO2
, and NH3. Further-
more, the reboiler possessed the highest temperature in the stripper.
The excess NH3 was vaporized when the warm rich solvent was fed
above the reboiler. Therefore, as shown in Fig. 10(c), the NH3 concen-
tration profile in the vapor phase was much higher than those in the
other cases once the warm rich solvent was fed above the reboiler. In
this case, the corresponding CO2 concentration profile was sup-
pressed, as shown in Fig. 10(d). Fig. 10(c) and (d) illustrate the
decrease in the purity of CO2 and the increase in the NH3 mole frac-
Fig. 8. Process flow diagram for CSB. tion at the top, as shown in Fig. 9(a) and (c) when the warm rich
Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

8 J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17

Fig. 9. Simulation results at different feed locations of the warm rich solvent. (a) Top temperature and CO2 mole fraction, (b) reboiler duty and vaporized ratio, and (c) top NH3 mole
fraction and solid fraction of the reflux.

solvent was fed above the reboiler, respectively. Therefore, to reduce generated by the reboiler, which could fulfill the energy-saving goal.
the reboiler duty and prevent solid precipitation in the reflux, i.e., to Fig. 12(a) shows that the vaporized ratio could be effectively reduced
improve the purity of CO2 purity and suppress the mole fraction of as the CSB fraction was increased from 0.01 to 0.05. Besides, the
NH3 at the top, the warm rich solvent must not be fed near the energy reductions in the reboiler and condenser, which were com-
reboiler. pared with the energy requirements of the standard stripper, were
The CSB fraction is another essential factor, which affects the enhanced owing to the low vapor requirement. Generally, the heat
energy-saving performance of the modified CSB. Fig. 11 shows the duty of the reboiler is the sum of the solvent heat duty, CO2 desorp-
simulation results at different CSB fractions. As Fig. 9(b) shown, the tion energy, and condenser duty [15]. At a constant CO2-lean loading,
reboiler duty could be reduced by moving the feed stage of the warm the desorption energy might not be altered by varying the cold-
rich solvent downward. The similar situations can be found at differ- split fraction. Moreover, the lower the amount of vapor that was
ent CSB fractions in Fig. 11(a), in which the minimum energy require- generated by the reboiler, the lesser the condenser duty that was
ment for each CSB fraction is around at the seventh stage. Fig. 11(b) required to condense the vapor. Therefore, the equivalent energy
shows the top temperature and corresponding purity of CO2 with dif- reductions by the reboiler and condenser could be observed in
ferent CSB fractions. The CO2-lean loading was 0.24, and the warm the CSB fraction range of 0.01
0.05, as shown in Fig. 12(a). How-
rich solvent was fed above the seventh stage. As shown in the figure, ever, when the cold-split fractions were larger than 0.05, the val-
the top temperature was significantly reduced by increasing the CSB ues of the vaporized ratio approached their limit, which is close
fraction from 0.01 to 0.05. Consequently, the corresponding purity of to 1.0. Consequently, the energy reductions in the condenser
CO2 at the top was enhanced to its limit, which is ~0.99. The excess were also maintained at a constant value in these cases. Thus, the
cold bypass could neither reduce the top temperature nor improve energy reductions in the reboiler were reduced when the CSB
the purity of CO2 once the CSB fractions were higher than 0.05, as fractions were higher than 0.05, as shown in Fig. 12(a). As afore-
shown in Fig. 11. mentioned, regarding the constant CO2-lean loading, the reboiler
As aforementioned, to capture a fixed amount of CO2, a higher CO2 duty was applied to heat and vaporize the solvent. Fig. 12(b)
purity in the overhead vapor indicated that less vapor must be shows that the heat recovery by HX reduced as the CSB fraction
Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17 9

Fig. 10. Results of varying the feed locations of the warm rich solvent. (a) Column temperature profiles, (b) total NH3 concentration profiles in the liquid phase, (c) NH3 concentra-
tion profiles, and (d) CO2 concentration profiles in the vapor phase.

Fig. 11. Simulation results at different CSB fractions. (a) Energy reduction on varying feed stage of warm rich solvent and (b) top temperature and CO2 purity.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

10 J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17

Fig. 12. Compromise between the latent heat reduction and sensible heat recovery. (a) Energy reductions in the reboiler and condenser and the vaporized ratio, (b) reboiler duty
and heat recovery by HX.

increased beyond 0.05. The deficiency of the recovered energy proper energy-saving configuration for CSB involves the simulta-
must be compensated for by increasing the reboiler duty, as neous minimization of the overhead vapor flow rate and mainte-
shown in Fig. 12(b). Therefore, the guideline for designing a nance of the heat recovery by HX.

Fig. 13. Energy reductions and vaporized ratio at various CO2-lean loadings, (a) at 0.22 of lean loading, (b) at 0.26 of lean loading, and (c) energy requirement for the modification of
CSB at 0.22, 0.24, and 0.26 CO2-lean loadings.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17 11

Fig. 14. Comparison of the driving force for mass transfer in the cases of the standard and CSB configurations, (a) the equilibrium and operating lines, (b) stage temperature, and (c)
the CO2 concentration gradient in the liquid phase for each stage.

However, Fig. 13 compares the results of applying various CO2 in the vapor phase is between 0.4 and 0.6 in the standard stripper.
lean-loadings. As shown in Fig. 13(a) and (b), the energy reductions Since the temperature profile in the standard stripper is confined
in the condenser were maintained at a constant value, once the from 135 to 147 °C, as shown in Fig. 14(b), the high temperatures
vaporized ratio was unchanged, which is identical to the results of restrict the CO2 mole fractions in the vapor phase. On the other hand,
Fig. 12(a). In addition, Fig. 13(a) and (b) also show the inconsistence
of energy reductions by the reboiler and condenser, once the cold-
split fractions were larger than 0.05. The reboiler duty was increased
to compensate the shortage of heat recovery by the excess of cold
rich bypass. Fig. 13(c) shows the energy requirements for modifying
CSB at various CO2-lean loadings. The minimum reboiler duty was
2.37 GJ/t-CO2 at a CO2-lean loading of 0.24 and a CSB fraction of 0.05,
where the energy reduction was 34% compared with that of the stan-
dard stripper under the optimal operating conditions in which the
minimum energy requirement for the CO2 stripper was 3.60 GJ/t-CO2
at a lean loading of 0.26 mol-CO2/mol-NH3. The simulation results for
the CSB with the minimum energy requirement are listed in
Appendix A. Fig. 13 displays that the excess cold split flow may not
only deteriorate the heat recovery by HX, but also reabsorb the
stripped CO2 that results in more energy needed to regenerate the
reabsorbed CO2.
Fig. 14 compares the driving force for CO2 mass transfer in the
cases of the standard stripper and the CSB configuration, in which the
CSB fraction was set to 0.05 and the warm rich solvent was fed above
the seventh stage. The equilibrium and operating lines are shown in
Fig. 14(a) for the both configurations, in which the CO2 mole fraction Fig. 15. Process flow diagram for AFS.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

12 J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17

Fig. 16. Comparison of the concentration profiles in the AFS and CSB configurations. (a) NH3 and (b) CO2 in the vapor phase.

the low temperature on the top for the CSB, which is around 50 °C in reduction of the residence time for vaporizing the solvents might
Fig. 14(b), enhances the top CO2 mole fraction to 0.99 in the vapor minimize the thermal degradation of the amine-based solvents.
phase, as shown in Fig. 14(a). Therefore, the high concentration of However, the thermal degradations of NH3-based solvents are negli-
CO2 on the top resulted in the reabsorption of the stripped CO2 that gible [6]. Contrarily, the high temperature of the feed stage favors the
can be observed in Fig. 14(a), in which the CO2 mole fraction in the vaporization of NH3 from the warm rich solvent, as shown in Fig. 10
cold rich solvent fed into the stripper is lower than that of the liquid (b) and (c).
phase on the corresponding stage in equilibrium. Fig. 14(c) shows the Fig. 15 shows the process flow diagram for ammonia-based AFS
driving force for CO2 mass transfer in the liquid phase for the stan- proposed by Jiang et al. [16], in which the reboiler of a standard strip-
dard and CSB configurations, in which the CO2 concentration gra- per was replaced by a steam heater and a flash vessel. Fig. 16 shows
dients are obvious in most stages for the CSB. Thus, Fig. 14(c) the comparison of the NH3 and CO2 concentration profiles for the AFS
indicates that the mass transfer for CO2 is easier in the CSB configura- and CSB in the vapor phase. The CO2-lean loading and CSB fraction
tion than that in the standard stripper. Since the equilibrium model were set to 0.24 and 0.05, respectively. Furthermore, regarding CSB,
was applied to simulate the standard stripper in Refs. the warm rich solvent was fed above the reboiler, which possessed
[4,9,12,15,18,20], the model is also suitable to describe the process the highest NH3 concentration profile, as shown in Fig. 10(c). Fig. 16
behavior in the CSB configuration, as the CSB fraction is 0.05. (a) indicates that the NH3 profile of AFS was higher than that of CSB,
i.e., the mole fractions of NH3 in ASF were higher than those of CSB at
5.2. AFS the same CO2-lean loading and CSB fraction. Since the warm rich sol-
vent was directly fed into the flash vessel in the AFS configuration, a
Jiang et al. [16] adapted the AFS configuration for the NH3-based larger amount of NH3 was vaporized from the solvent because of the
PCC process. However, the AFS configuration was proposed by Lin high temperature. Consequently, the top NH3 concentration of AFS
and Rochelle [17] for amine-based solvents. They reported that the was higher than that of CSB. Therefore, the purity of the top CO2 of

Fig. 17. Comparison of the energy consumption in the AFS and CSB modifications, (a) reboiler duties and (b) heat recovery by HX and vaporized ratio.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17 13

Fig. 18. Results of the AFS modification utilizing different CSB fractions at a CO2-lean loading of 0.24. (a) Top temperature and CO2 purity, (b) energy reduction and vaporized ratio,
(c) reboiler duty and heat recovery by HX, and (d) concentration of the top NH3 and solid fraction of the reflux.

AFS was lower than that of CSB, as shown in Fig. 16(b). Compared fractions are higher than 0.05 in the both configurations, whereas the
with CSB, a larger amount of vapor must be generated to achieve the reboiler duties are rising only for the CSB, as shown in Fig. 17(a).
expected removal rate of CO2 in AFS. However, Fig. 17(a) shows that the reboiler duty for the AFS can be
Fig. 17 shows the comparison between the energy requirements reduced, as the CSB fractions are between 0.05 and 0.07 whereas the
for the AFS and CSB with different CSB fractions. The CO2-lean loading heat recovery by the HX is deteriorated. The reduction of the reboiler
was set to 0.24, and the warm rich solvent was fed above the seventh duty in the AFS comes from the lower vaporized ratio, as shown in
stage for CSB, where the minimum reboiler duty was located, as Fig. 17(b). Therefore, the less vaporization energy (the lower vapor-
shown in Fig. 9(b). The results indicated that the minimum energy ized ratio) compensated the deteriorated heat recovery by the HX
requirement for AFS was 2.63 GJ/t-CO2 at a CSB fraction of 0.07. How- when increasing the CSB fractions in the AFS. Once the vaporized
ever, the CSB configuration could achieve lower regeneration energy ratio reached its limit, the reboiler duties were increased at the high
at a CSB fraction of 0.05, as shown in Fig. 17(a). As previously shown CSB fractions (>0.07) to balance the shortage of the heat recovery for
(Fig. 15), at a CSB fraction of 0.05, the purity of the top CO2 in the AFS the AFS configuration. On the other hand, Fig. 17(b) shows that the
configuration was suppressed by the high NH3 concentration. There- vaporized ratio reaches its limit at CSB fraction of 0.05 for the CSB.
fore, regarding the capture of the same amount of CO2, the flash ves- Consequently, the vaporization energy may not be further reduced as
sel of AFS must vaporize an increased amount of the solvent the CSB fractions are higher than 0.05. Therefore, the reboiler duty
compared with that of CSB. Consequently, as shown in Fig. 17(a), an needs to be increased to compensate the deterioration of the heat
increased reboiler duty is required by the AFS configuration with a recovery for the CSB, once the CSB fraction is higher than 0.05, as
CSB fraction of 0.05. shown in Fig. 17(a).
In addition, Fig. 17(a) shows that the reboiler duties of CSB and AFS could be considered as a limiting case of the CSB modification
AFS are identical at CSB fractions of 0.08 and 0.09; meanwhile, the in which the warm rich solvent was fed into the last equilibrium
reboiler duties rise as the CSB fractions are increasing. As shown in stage. The characteristics of CSB were evident in the results of the
Fig. 17(b), the heat recoveries by the HX are declining once the CSB AFS modification. Fig. 18 displays the results by varying the CSB
Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

14 J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17

Fig. 19. NH3 concentration profiles for the AFS modification with different CSB fractions. (a) NH3 mole fractions in the vapor phase and (b) total NH3 mole fractions in the liquid
phase.

fraction at a CO2-lean loading of 0.24. As shown in Fig. 18(a), the concentrations disappeared in the AFS configuration because the
top temperature decreased as the CSB fraction increased, whereas warm rich solvent was fed into the last equilibrium stage. There-
the purity of CO2 at the top was enhanced. However, the purity fore, the NH3 concentration profiles in the AFS configuration
of CO2 approached its limit when the CSB fractions were ˃0.07. could only be reduced by decreasing the flow rate of the warm
Fig. 18(b) shows that the vaporized ratio could be reduced when rich solvent, and this is similar to increasing the CSB fraction in
the purity of CO2 is improved. Consequently, the energy reduction Fig. 19.
of the condenser, which compares with that of the standard strip- Fig. 20 shows the comparison of the energy requirements for the
per, increased because of the low vapor flow rate. When the AFS at lean loadings of 0.24, 0.26, and 0.28. The obtained minimum
vaporized ratio remained unchanged as the CSB fractions overall energy was 2.51 GJ/t-CO2, and the CO2 loading of the lean sol-
increased beyond 0.07 (Fig. 18(b)), the energy reduction of the vent was 0.26, while the CSB fraction was 0.06. The simulation results
condenser remained constant. However, as shown in Fig. 18(b), for the AFS with the minimum energy requirement are listed in
the energy reduction of the reboiler was not balanced with that Appendix A. A 30% energy reduction was achieved by the AFS config-
of the condenser counterpart at a CSB fraction > 0.05. These uration compared with the minimum regeneration energy of the
unbalanced parts of the energy reductions indicated that the heat standard process. As discussed in the previous section, the CSB modi-
recovery by HX began to deteriorate as the CSB fractions fication can achieve 34% of energy reduction compared with that of
increased to ˃ 0.05, as already discussed (Fig. 12(a)). Therefore, the standard stripper. In terms of the energy-saving perspective, the
the shortage of recovered energy was compensated by the CSB outperforms the AFS because only trace of ammonia is vaporized
reboiler. As shown in Fig. 18(c), the heat recovery by HX reduced on the top in the CSB with a proper feed location of the warm rich
as the CSB fraction increased. However, the reboiler duty was solvent. Although Jiang et al. [16] reported an outstanding energy-
relieved when the CSB fractions were ˂ 0.07 because of the saving performance by the AFS configuration, which is 42% of energy
reduction in the vaporized ratio (Fig. 18(b)). When the purity of reduction, there existed a lower energy requirement for their system
the top CO2 approached its limit, the vaporized ratio could not be if the CSB was applied.
further reduced by increasing the CSB fraction, i.e., the CSB frac-
tions were larger than 0.07 (Fig. 18(b)). In these cases, the latent
heat for vaporization could not be further reduced, although the
reboiler was expected to afford the sensible heat that must be
recovered by HX. As shown in Fig. 18(c), the reboiler duty began
to increase once the CSB fraction was larger than 0.07. Fig. 18(d)
shows that the concentration of the top NH3 and the solid frac-
tion of the reflux disappeared once a reduced amount of the
warm rich solvent was fed into the stripper in which the CSB
fraction was ˃ 0.07. Since the warm rich solvent was fed into the
last equilibrium stage in the AFS configuration, excess NH3 at the
top could only be reduced by reducing the rate of the warm rich
flow, which might sacrifice the heat recovery performance of HX.
Fig. 19 shows the NH3 concentration profiles of the simulations
in Fig. 18. Fig. 19(a) shows that excess NH3 was vaporized at a
CSB fraction of ˂ 0.07. Consequently, the total NH3 concentrations
in the liquid phase increased (Fig. 19(b)) compared with the
results of CSB (Fig. 10(b) and (c)) where the NH3 concentrations
could be adjusted by moving the feed location of the warm rich Fig. 20. Energy requirement for ASF modification at CO2-lean loadings of 0.24, 0.26,
solvent. However, the degree of freedom for minimizing the NH3 and 0.28.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17 15

6. Conclusion vaporize an increased amount of the solvent to maintain a fixed CO2

removal rate. Moreover, solid precipitates, which were due to the
In this study, the energy-saving performances of modified absorp- high concentration of the top NH3, could be found in the reflux. How-
tion and desorption processes were investigated. Contrary to the con- ever, the AFS configuration, where the warm rich solvent was fed
ventional absorber, the results indicated that a quarter of the NH3 slip into the last equilibrium stage, was the limiting case of CSB. An excess
could be achieved by the two-stage absorber. Therefore, the regener- amount of NH3 was released from the warm rich solvent because of
ation energy for the NH3 washing water, which compares with the the high temperature of the stage. To improve or suppress the puri-
counterpart for reviving the CO2-lean solvent, could be neglected. ties of the top CO2 and NH3, respectively, the flow rate of the warm
This study also demonstrated that the energy consumption of the rich solvent must be reduced, and this may deteriorate the heat
CSB process was affected by the CSB amount and the feed location of recovery by HX. Therefore, the energy-saving performance of AFS
the warm rich solvent. The energy-saving effects of modifying CSB was inferior to that of CSB with a proper feed location of the warm
was due to the increased purity of CO2 that was caused by the differ- rich solvent.
ent amounts of CSB and feed stage of the warm rich solvent. How-
ever, the excess bypass reduced the heat recovery by HX, and the
Declaration of Competing Interest
shortage of the recovered energy must be compensated by the
reboiler duty. Therefore, two factors (the bypassed amount and
The authors declare that they have no known competing financial
warm rich feed stage) must be simultaneously considered when min-
interests or personal relationships that could have appeared to influ-
imizing the regeneration energy by modifying CSB.
ence the work reported in this paper.
Additionally, this study demonstrated that the concentration of
the top NH3 increased upon feeding the warm rich solvent near the
reboiler in the modification of CSB. Therefore, the purity of the top Appendix A. Material stream tables for simulation cases
CO2 was suppressed, thus increasing the energy requirement to
Tables A1
A3

Table A1
Stream table for the two-stage absorption design at CO2-lean loading of 0.26.

Stream 1 2 3 4 5 6 7 8

Temperature (°C) 30.0 30.0 15.0 30.0 14.1 15.0 19.3 18.9
Pressure (atm) 1 1 1 1 1 1 1 1
Mass Flow (t/h) 2124.6 15,340.7 15,520.3 1776.9 1804.8 345.0 1759.9 15,662.4
Mole Fraction
H2O 0.118 0.936 0.937 0.040 0.015 0.999 0.022 0.932
NH3 0.03 0.015 0.039 0.009 676 PPM 0.000 0.015
CO2 0.124 812 PPB 932 PPB 0.045 0.019 0.016 6 PPM
N2 0.757 337 PPB 7 PPM 0.877 0.957 0.961 11 PPM
NH4+ 0.019 0.026 44 PPM 0.029
H3O+ 63 PPB 0.007
NH4HCO3(s)
NH2COO
0.008 0.014 944 PPB 0.013
HCO3
63 PPB 0.003 0.005 6 PPM 0.008
OH
2 PPM 488 PPB 5 PPM 432 PPB
CO3
2 0.004 0.003 16 PPM 0.004

Stream 9 10 11 12 13 14 15 CO2

Temperature (°C) 15.8 91.6 145.5 20.8 35 135.4 134.1 35

Pressure (atm) 1 1 10 10 10 10 10 10
Mass Flow (t/h) 389.9 44.9 15,297.7 15,662.4 172.0 15,662.4 536.7 364.7
Mole Fraction
H2O 0.972 0.715 0.932 0.931 0.648 0.923 0.295 0.006
NH3 0.012 0.22 0.04 0.015 301 PPM 0.034 0.09 26 PPM
CO2 385 PPB 0.065 0.001 7 PPM 0.008 0.006 0.614 0.993
N2 13 PPM 121 PPM 11 PPM 234 PPB 11 PPM 560 PPM 0.001
NH4+ 0.009 0.014 0.029 0.056 0.019
H3O+ 2 PPB 7 PPB 3 PPB
NH4HCO3(s) 0.234
NH2COO
0.003 0.003 0.013 618 PPM 0.004
HCO3
0.002 0.01 0.008 0.054 0.014
OH
1 PPM 4 PPM 455 PPB 10 PPB 3 PPM
CO3
2 0.002 300 PPM 0.004 628 PPM 424 PPM

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

16 J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17

Table A2
Stream table for the CSB at CO2-lean loading of 0.24 and cold split fraction of 0.05.

Stream CO2 Hot Lean Overhead Reflux Cold Rich Warm Rich Cold Split

Temperature (°C) 35.0 147.3 48.0 35.0 19.7 139.8 20.0

Pressure (atm) 10 10 10 10 10 10 10
Mass Flow (t/h) 363.6 13,987.9 364.5 0.9 14,351.5 13,633.9 717.6
Mole Fraction
H2O 0.006 0.933 0.011 0.962 0.932 0.923 0.932
NH3 5 PPM 0.041 93 PPM 41 PPM 0.015 0.036 0.015
CO2 0.993 0.001 0.988 0.006 6 PPM 0.008 6 PPM
N2 0.001 0.001 132 PPB 11 PPM 11 PPM 11 PPM
NH4+ 0.013 0.016 0.029 0.016 0.029
H3O+ 2 PPB 12 PPB 3 PPB
NH4HCO3(s)
NH2COO
0.003 54 PPM 0.013 0.003 0.013
HCO3
0.009 0.016 0.008 0.012 0.008
OH
5 PPM 3 PPB 451 PPB 3 PPM 451 PPB
CO3
2 274 PPM 27 PPM 0.004 352 PPM 0.004

Table A3
Stream table for the AFS at CO2-lean loading of 0.26.

Stream CO2 Hot Lean Overhead Reflux Cold Rich Warm Rich Cold Split

Temperature (°C) 35.0 145.4 43.0 35.0 19.2 139.1 19.2

Pressure (atm) 10 10 10 10 10 10 10
Mass Flow (t/h) 364.0 15,298.4 364.4 0.5 15,662.4 14,722.6 939.7
Mole Fraction
H2O 0.006 0.932 0.009 0.961 0.932 0.923 0.932
NH3 0.000 0.040 53 PPM 44 PPM 0.015 0.036 0.015
CO2 0.993 0.001 0.990 0.006 6 PPM 0.008 6 PPM
N2 0.001 50 PPB 0.001 146 PPB 11 PPM 11 PPM 11 PPM
NH4+ 0.014 0.017 0.029 0.017 0.029
H3O+ 2 PPB 12 PPB 3 PPB
NH4HCO3(s)
NH2COO
0.003 59 PPM 0.013 0.003 0.013
HCO3
0.017 0.008 0.013 0.008
OH
4 PPM 3 PPB 435 PPB 3 PPM 435 PPB
CO3
2 301 PPM 30 PPM 0.004 365 PPM 0.004

References [14] Ullah A, Soomro MI, Kim WS. Ammonia-based CO2 capture parameters optimiza-
tion and analysis of lean and rich vapor compression processes. Sep Purif Technol
[1] Gal E. Ultra cleaning of combustion gas including the removal of CO2. United 2019;217:8–16.
States, Patent US 20080072762A1 2008. https://patents.google.com/patent/ [15] Liu J, Wong DSH, Chen DS. Energy-saving performance of advanced stripper con-
WO2006022885A1/en figurations for CO2 capture by ammonia-based solvents. J Taiwan Inst Chem Eng
[2] McLarnon CR, Duncan JL. Testing of ammonia based CO2 capture with multi-pol- 2020;113:273–84.
lutant control technology. Energy Proced 2009;1:1027–34. [16] Jiang K, Yu H, Yu J, Li K. Advancement of ammonia-based post-combustion CO2
[3] Yu H, Morgan S, Allport A, Cottrell A, Do T, Mcgregor J, Wardhaugh L, Feron P. capture technology: process modifications. Fuel Process Technol
Results from trialing aqueous NH3 based post-combustion capture in a pilot plant 2020;210:106544.
at Munmorah power station: absorption. Chem Eng Res Des 2011;89:1204–15. [17] Lin YJ, Rochelle GT. Optimum heat of absorption for CO2 capture using the
[4] Yu H, Qi G, Wang SJ, Morgan S, Allport A, Cottrell A, Do T, Mcgregor J, Wardhaugh advanced flash stripper. Int J Greenh Gas Control 2016;53:169–77.
L, Feron P. Results from trialing aqueous ammonia-based post-combustion cap- [18] Rezazadeh F, Gale WF, Lin YJ, Rochelle GT. Energy performance of advanced
ture in a pilot at Munmorah power station: gas purity and solid precipitation in reboiled and flash stripper configurations for CO2 capture using monoethanol-
the stripper. Int J Greenh Gas Control 2012;10:15–25. amine. Ind Eng Chem Res 2016;55:4622–31.
[5] Rhee CH, Kim JY, Han K, Ahn CK, Chun HD. Process analysis for ammonia-based [19] Ishaq H, Ali U, Sher F, Anus M, Imran M. Process analysis of improved process
CO2 capture in ironmaking industry. Energy Proced 2011;4:1486–93. modifications for ammonia-based post-combustion CO2 capture. J Environ Chem
[6] Han K, Ahn CK, Lee MS, Rhee CH, Kim JY, Chun HD. Current status and challenges Eng 2021;9:104928.
of the ammonia-based CO2 capture technologies toward commercialization. Int J [20] Nguyen HLQ, Wong DSH. Integration of rich and lean vapor recompression con-
Greenh Gas Control 2013;14:270–81. figurations for aqueous ammonia-based CO2 capture process. Chem Eng Res Des
[7] Zhang M, Guo Y. Process simulations of large-scale CO2 capture in coal-fired 2021;169:86–96.
power plants using aqueous ammonia solution. Int J Greenh Gas Control [21] Al-Hamed KHM, Dincer I. A comparative review of potential ammonia-based car-
2013;16:61–71. bon capture systems. J Environ Manag 2021;287:112357.
[8] Zhang M, Guo Y. Process simulations of NH3 abatement system for large-scale CO2 [22] Liu L, Zhao J, Deng S, An Q. A technical and economic study on solar-assisted
capture using aqueous ammonia solution. Int J Greenh Gas Control 2013;18:114– ammonia-based post-combustion CO2 capture of power plant. Appl Therm Eng
27. 2016;102:412–22.
[9] Liu J, Wong DSH, Jang SS, Shen YT. Energy-saving design for regeneration process [23] Siddiqui O, Ishaq H, Chehade G, Dincer I. Performance investigation of a new
in large-scale CO2 capture using aqueous ammonia. J Taiwan Inst Chem Eng renewable energy-based carbon dioxide capturing system with aqueous ammo-
2017;73:12–9. nia. Int J Energy Res 2020;44:2252–63.
[10] Jilvero H, Normann F, Andersson K, Johnsson F. The rate of CO2 absorption in [24] Bonalumi D, Giuffrida A. Investigations of an air-blown integrated gasification
ammonia implications on absorber design. Ind Eng Chem Res 2014;53:6750–8. combined cycle fired with high-sulphur coal with post-combustion carbon cap-
[11] Darde V, Maribo-Mogensena B, van Well WJM, Stenby EH, Thomsen K. Process ture by aqueous ammonia. Energy 2016;117:439–49.
simulation of CO2 capture with aqueous ammonia using the extended UNIQUAC [25] Qi G, Wang S, Yu H, Wardhaugh L, Feron P, Chen C. Development of a rate-based
model. Int J Greenh Gas Control 2012;10:74–87. model for CO2 absorption using aqueous NH3 in a packed column. Int J Greenh
[12] Liu J. Investigation of energy-saving designs for an aqueous ammonia-based car- Gas Con 2013;17:450–61.
bon capture process. Ind Eng Chem Res 2018;57:15460–72. [26] Aspen Technology. Aspen Plus: rate-based model of the CO2 capture process by
[13] Li K, Yu H, Ferona P, Tade M, Wardhaugha L. Technical and energy performance of NH3 using Aspen Plus. MA, USA: Cambridge; 2011.
an advanced, aqueous ammonia-based CO2 capture technology for a 500MW [27] Pinsent BRW, Pearson L, Roughton FJW. The kinetics of combination of carbon
coal-fired power station. Environ Sci Technol 2015;49:10243–52. dioxide with ammonia. Trans Faraday Soc 1956;52:1594–8.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060
JID: JTICE
ARTICLE IN PRESS [m5G;July 12, 2021;9:12]

J. Liu et al. / Journal of the Taiwan Institute of Chemical Engineers 00 (2021) 1
17 17

[28] Onda K, Takeuchi H, Okumoto Y. Mass transfer coefficients between gas and liq- [31] Garđarsdo  Normann F, Andersson K, Johnsson F. Postcombustion CO2 cap-
 ttir SO,
uid phased in packed columns. J Chem Eng Jpn 1968;1:56–62. ture using monoethanolamine and ammonia solvents: the influence of CO2 con-
[29] Stichlmair J, Bravo JL, Fair JR. General model for prediction of pressure drop and centration on technical performance. Ind Eng Chem Res 2015;54:681–90.
capacity of countercurrent gas/liquid packed columns. Gas Sep Purif 1989;3:19– [32] Jilvero H, Normann F, Andersson K, Johnsson F. Ammonia-based post combustion
28.
the techno-economics of controlling ammonia emissions. Int J Greenh Gas Con-
[30] Yu J, Wang S, Yu H, Wardhaugh L, Feron P. Rate-based modelling of CO2 regenera- trol 2015;37:441–50.
tion in ammonia based CO2 capture process. Int J Greenh Gas Control [33] Yu J, Wang S, Yu H. Modelling analysis of solid precipitation in an ammonia-based
2014;28:203–15. CO2 capture process. Int J Greenh Gas Control 2014;30:133–9.

Please cite this article as: J. Liu et al., Energy-saving performance of the process modifications for carbon capture by diluted aqueous ammonia,
Journal of the Taiwan Institute of Chemical Engineers (2021), https://doi.org/10.1016/j.jtice.2021.06.060