You are on page 1of 18

Fluid Phase Equilibria 550 (2021) 113235

Contents lists available at ScienceDirect

Fluid Phase Equilibria


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

New solubility and heat of absorption data for CO2 in blends of


2-amino-2-methyl-1-propanol (AMP) and Piperazine (PZ) and a new
eNRTL model representation
Ardi Hartono∗, Rafiq Ahmad1, Hallvard F. Svendsen, Hanna K. Knuutila
Department of Chemical Engineering, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway

a r t i c l e i n f o a b s t r a c t

Article history: New data comprising CO2 partial pressure, total pressure, and the heat of absorption of CO2 for over
Received 14 June 2021 aqueous solution of 3 mol/dm3 AMP and 1.5 mol/dm3 PZ and total pressure and heat of absorption for
Revised 20 September 2021
different mol ratios of AMP/PZ (3.0/0.0, 0.0/1.5, 0.5/4.0, 1.5/3.0, 2.25/2.25, 3.0/1.5, 4.0/0.5) are presented
Accepted 22 September 2021
as functions of CO2 loading and temperature.
Available online 24 September 2021
Measured solubility data and selected data reported in the literature were used in the regression
Keywords: of binary interaction energy parameters in the quaternary system (AMP/PZ/H2 O/CO2 ) using the eNRTL
AMP model. The equilibrium constants and binary interaction energy parameters from our previous work on
CO2 solubility the ternary systems of AMP/H2 O/CO2 and PZ/H2 O/CO2 were used without refitting.
eNRTL Good agreement with the literature was observed for the total pressure data over aqueous solutions
Heat of absorption of 3.0 mol/dm3 AMP and 1.5 mol/dm3 PZ, and the CO2 partial pressure and heat of reaction data over
PZ
aqueous solutions of 3.0 mol/dm3 AMP+1.5 mol/dm3 PZ at different loadings and temperatures. The re-
Volatility
sults reflect that the new data are consistent with reported data from the literature.
The modeling results show that the eNRTL model represents the data well with AARD values of 13.1%
for total pressure and 20.9% for CO2 partial pressure. The volatility of AMP and PZ as function of load-
ing and temperature and published speciation data for AMP/AMP H + , P Z/P Z H + /P ZH22+ , P ZCO− 2
/P Z H +CO−
2
,
PZ (CO− ) and HCO−3 /CO23− were also well predicted.
2 2
Only the heat of absorption data over an aqueous solution of 3.0M AMP+1.5M PZ from this work can
be compared with literature data, and good agreement is observed. The eNRTL model also predicts satis-
factorily the experimentally obtained heat of absorption data for all different ratios of AMP/PZ. Individual
reaction contributions show that when the AMP/PZ ratio increases, the predicted total heat of absorption
goes toward the single solvent AMP system where only protonated AMP is the main contributor. When
the AMP/PZ ratio decreases, the predicted total heat of absorption also approaches the single solvent PZ
system where the protonated PZ, PZ-monocarbamate and zwitterion reactions play important roles. In
systems with high PZ-concentration, the contribution of the zwitterion, protonated PZ and PZ-carbamate
cancel each other at high loadings.
© 2021 Published by Elsevier B.V.

1. Introduction cant point sources, the latter two are attractive for CO2 emissions
reductions.
Carbon dioxide emitted to the atmosphere through human ac- Carbon Capture and Storage is considered one of the most
tivity is responsible for a large part of the observed warming of promising technologies to reduce CO2 emissions and absorption
our planet [1]. In addition to the transport and domestic sectors, into amine solutions is today the most mature and applied tech-
fossil fuel-based power generation and CO2 -emitting industries are nology for CO2 capture [2]. The basic concept of this technology is
the main contributors to the total CO2 emissions. Offering signifi- a chemical solvent that reacts reversibly with CO2 . When CO2 is
absorbed, the solvent lean in CO2 approaches its equilibrium and
becomes rich in CO2 . To recycle the solvent, a regeneration process

Corresponding author at: NTNU, Sem Saeland Vei 4, 7491 Trondheim, Norway based on temperature-swing is usually applied for low-pressure,
E-mail addresses: ardi.hartono@chemeng.ntnu.no, ardi.hartono@ntnu.no (A. post combustions, releasing CO2 at elevated temperature in a des-
Hartono), rafiq.ahmad@umt.edu.pk (R. Ahmad).
1 orber. To reach regeneration conditions, energy must be added to
Present address: School of Food and Agricultural Science, University of Manage-
ment and Technology, Lahore 540 0 0 Pakistan increase the solvent temperature, provide the heat for reversing

https://doi.org/10.1016/j.fluid.2021.113235
0378-3812/© 2021 Published by Elsevier B.V.
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 1
Experimental data available from literature and used for the AMP/PZ/CO2 /H2 O system∗ .

No Experimental type and Experiment conditions Number of Source Remarks


Collected data points

1 VLE T = 313, 333, 353K 144 [29]


PC O2 , T, [AMP/PZ ], α [AMP/PZ ]=2.0/0.5, 2.0/1.0, 2.0/1.5,
• Vapor recirculation equilibrium + Shimada GC
3.0/0.5, 3.0/1.0, 3.0/1.5M
α = 0.292 to 0.876 (<200kPa)
• Kent-Eisenberg model
PC O2 = 1.03 to 139.9 kPa

T = 303, 313, 323, 318, 328K 34 [4]


[AMP/PZ ]=3.0/1.5M
• Vapor recirculation equilibrium +NDIR CO2
α = 0.06 to 0.83
• An empirical model
PC O2 = 0.03 to 20 kPa

T = 293 to 433K 89 [6]


[mAMP /mPZ ]=5.0/2.3; 4.0/2.0
• A static VLE and Wetted wall column
α = 0.0 to 0.64
• An empirical model
PC O2 = 0.05 to 2231 kPa

2 VLE T = 313, 333, 353K 175 [30]


PT , T, [AMP/PZ ], α [AMP/PZ ] = 38/2, 35/5, 32/8, 48/2,
• SS equilibrium cell
45/4, 42/8 mass%
• An eNRTL model
α = 0.38 to 1.02
PT = 0.16 to 1418kPa
T = 313, 333, 373, 393K 91 [32]
[AMP/PZ ]=25/5, 20/10 mass%
• Static-analytic apparatus + GC
α = 0.06 to 0.83
• Kent-Eisenberg model
PT = 0.03 to 20 kPa

T = 303, 333K 112 [33]


[AMP/PZ ]=8.91/0, 8.91/1.74, 8.91/6.88,
• SS equilibrium cell
8.91/10.35; 17.82/0, 17.82/1.74,
• No modeling work
17.82/6.88, 17.82/10.35 mass%
α = >1.00 to 2.40
PT = 5 to 60 bars
T = 293, 303, 313, 323K 30 [31]
[AMP/PZ ]=13.5/0.87, 13.5/2.6,
• A glass spiral cell
8.91/1.74 mass%
• Deshmukh–Mather model
α = 0.15 to 0.94
PT = 90.66 bar
PC O2 = 18.13, 27.19, 72.52 kPa
3 Calorimetric T = 313, 333, 353K 87 [34]
HABS. , T, [AMP/PZ ], α [AMP/PZ ]=3/1.5, 3/0.5, 4/1 M
• a Differential Reaction Calorimeter (SETARAM
α = 0.1 to 0.84
PT = Not available DRC-Evolution)
• No modeling work

4 NMR speciation T = 303, 313K 25 [35]


mi. , T, [AMP/PZ ], α [mAMP /mPZ ]==4/2, 2.3/5m
• A Varian INOVA 500MHz 13C-NMR
α = 0.43 to 0.65
• An eNRTL model

α = C O2 loading = mol C O2 /(mol AMP + mol PZ )



All concentrations are given as in the original publications.

the reactions and the water vapor pressure needed. The regenera- the heat of absorption. An electrolyte NRTL (eNRTL) model, as
tion thus becomes an energy-intensive part of this technology [3]. described by Chen and Evans (1986) [11], was implemented
The search for low energy penalty solvents for CO2 capture is es- and the binary interaction parameters in the quaternary system
sential. (AMP/PZ/H2 O/CO2 ) were regressed based on all available solubil-
The EU project CESAR investigated a blend of AMP/Piperazine ity data. The binary interaction parameters for the ternary systems
at laboratory conditions [4] and demonstrated an optimized con- of AMP/H2 O/CO2 and PZ/H2 O/CO2 were taken from our previous
centration of 3.5M AMP+1.5M PZ at pilot scale [5]. Li, et al. work [10,12].
(2013) [6] also recognized aqueous AMP/PZ as an interesting sys-
tem due to its high reactivity, caused by PZ, and the relatively low
heat requirement of stripping because of AMP. As a sterically hin- 2. Theoretical background
dered amine [7], AMP has in many ways similar thermal prop-
erties as tertiary amines, but unlike tertiary amines, it can react 2.1. Chemical reaction system
with CO2 to produce both carbamate (minor product) and bicar-
bonate/carbonate (major products) [8,9]. In aqueous solution, AMP The ten equilibrium reactions describing the AMP/PZ/H2 O/CO2
has shown high CO2 equilibrium temperature sensitivity [10], and system are given below.
therefore it is a good substitute for a tertiary amine. KW
In this work, new experimental solubility, and heat of absorp- 2H2 O ↔ H3 O+ + OH − (R1)
tion data for CO2 in the quaternary system of AMP/PZ/H2 O/CO2
were generated and used together with available literature data, KHCO−
shown in Table 1, to model the vapor-liquid equilibrium and 2H2 O + C O2 ↔ H3 O+ + HCO−
3
3 (R2)

2
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 2
Number of fitted parameters in the NRTL/eNRTL models.

No. System Species NP Remark


Molecule Cation Anion∗ Total Eq. 2 ai j /bi j

1 AMP/H2 O 2 0 0 2 2 4 NRTL
2 PZ/H2 O 2 0 0 2 2 4 NRTL
3 AMP/PZ 2 0 0 2 2 4 NRTL
4 AMP/PZ/H2 O 3 0 0 3 6 12 NRTL
5 AMP/H2 O/CO2 3 2 4 9 54 108 eNRTL
6 PZ/H2 O/CO2 3 3 6 12 114 228
Zwitterion = an-
ion
eNRTL
7 4 4 7 15 236 472 eNRTL
AMP/PZ/H2 O/CO2

Zwitterion is considered as an anion with zero charge.

regressed. As those parameters were kept, there remained 82 pa-


KCO2−
rameters (NP) for the parameter fitting of the quaternary system.
H2 O + HCO− ↔ H3 O+ + CO23−
3
(R3)
3 It must be noted that 12 of these parameters were set to their de-
fault values in Aspen plus® (see Table 25). The non-randomness
KPZ H + values for H2 O-salt, molecule-molecule and amine/CO2 -salt pairs
H2 O + P Z H + ↔ P Z + H3 O+ (R4)
interactions were also fixed to literature values [13].

KPZH 2+
H2 O + P ZH22+ ↔ P Z H + + H3 O+
2
(R5) 2.3. Parameter fitting

KPZCO− All the equilibrium constants needed for the quaternary system
H2 O + C O2 + P Z ↔ P ZCO− +
2
were determined in previous work (see Table 23). A gamma-phi
2 + H3 O (R6)
equation was used to solve the phase equilibrium between vapor
KPZ H + CO−
and liquid phase:
H2 O + P Z H +CO− ↔ P ZCO− +
2
2 2 + H3 O (R7) yi · P · ϕi = γi · xi · i · ψi (3)

The liquid phase activity coefficient parameters were estimated


K
( )2
PZ CO−
2  
H2 O + C O2 + P ZCO−
2
↔ P Z CO−
2 + H3 O+ (R8) using the eNRTL model [11] and in the gas phase, the Peng Robin-
2
son EoS [15] equation of state was used to calculate the fugacity
coefficients of the components: ϕi (− ). The Poynting factor, i (− ),
KAMPH +
H2 O + AMP H + ↔ AMP + H3 O+ (R9) was also considered. The reference state of species was taken as
follows:
KAMPCO− 
H2 O + C O2 + AMP ↔ AMPCO− + 2
pSi φiS , pure component re f erence state f or water and amines .
2 + H3 O (R10) i =
Hi∞ in f inite dilution re f erence state f or C O2
Altogether fifteen species are present in the system: Four
(4)
molecules (H2 O, C O2 , AMP, PZ), and cations (H3 O+ , P Z H + , AMP H + ,
P Z22+ ), six anions (OH − , HCO−
3
, CO23− , P ZCO−
2
, AMPCO−
2
, P Z (C O2 )22− ) A PSO algorithm was implemented for fitting the binary inter-
and one zwitterion (P Z H +CO− 2
) . action parameters [14] in the quaternary system whereas the pa-
rameters for the binary interaction parameters in the ternary sys-
2.2. Thermodynamic modeling tems are taken from previous work, (see [10,12]).

The local interaction energy parameters in the NRTL/eNRTL


models [11] are expressed as: 3. Materials and experimental setups

bi j 3.1. Materials
τi j = a i j + (1)
T
In any specific system, the number of interaction energy param- The purchased chemicals, as shown in Table 3, were used with-
eters needed (in eNRTL) can be calculated according to Eq. 2 (see out additional purification. Aqueous 3.0 M AMP, 1.5M PZ and the
[13,14]): five solutions with different AMP/PZ ratios, were prepared volu-
metrically and gravimetrically in 1 dm3 flasks over a scale (Met-
M!
NP = + 2 ·M ·C ·A (2) tler Toledo PM1200) at ambient temperature. All amine concentra-
( M − 2 )! tions given are for unloaded solutions. For the loaded solutions,
Where NP, M, C and A represent the number of parameters, the amine concentrations were determined by titration whereas
molecules, cations, and anions respectively. Even when only a sim- the CO2 contents were analyzed with the wet chemistry BaCl2
ple interaction parameter equation (Eq. 1) is used, this doubles the method. These two techniques are described in detail in previ-
number of parameters to be estimated. The number of interaction ous work [10,12,16–18]. The unit conversions of the prepared so-
parameters needed from the binary to the quaternary system, is lutions to molality and mass % are provided in Table 4. For conve-
given in Table 2. nience, the solution name is used in this manuscript for own work,
In the AMP/H2 O/CO2 system [10], 39 parameters (NP) were fit- whereas for the reported literature data, the original concentration
ted and in the PZ/H2 O/CO2 system[12], 100 parameters (NP) were units are used as presented in Table 1.

3
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 3
Chemicals used in this work.

Amine Abbreviation CAS number Purity Supplier

2-Amino-2-Methyl-1-propanol AMP 124-68-5 ≥0.985a Sigma Aldrich


Piperazine PZ 110-85-0 ≥0.992a Sigma Aldrich
Carbon Dioxide CO2 124-38-9 ≥0.99999 AGA Gas GmbH
a
in a mass fraction basis and taken from the Certificate of Analysis CoA from the supplier.

Table 4
The prepared solutions for experiments.

Solution Molarity (Mol per dm3 solution) Molality (Mol per kg water) Mass % (100%∗ kg/ kg total)
No. name
AMP PZ AMP PZ AMP PZ

1 3.0M AMP 3.0 0.0 4.1 0.0 26.8 0.0


2 1.5M PZ 0.0 1.5 0.0 2.1 0.0 12.8
3 3.0M AMP+1.5M PZ 3.0 1.5 5.0 2.5 26.7 12.9
4 0.5M AMP+4.0M PZ 0.5 4.0 0.8 6.1 4.2 32.8
5 1.5M AMP+3.0MPZ 1.5 3.0 2.3 4.7 12.9 25.0
6 2.25M AMP+2.25M PZ 2.25 2.25 3.6 3.6 19.8 19.2
7 4.0M AMP+0.5M PZ 4.0 0.5 6.7 0.8 35.7 4.3

Fig. 1. The CPA202 reaction calorimeter (ChemiSens AB, Sweden).

3.2. Experimental setups similar to one described by Kim and Svendsen (2007) [22] was
used in this work: the CO2 was added in several steps to the
3.2.1. Atmospheric pressure VLE apparatus solution which was kept at isothermal conditions. The amount of
An atmospheric pressure VLE apparatus was used to generate heat removed by the system to keep the solution at constant tem-
one set of partial pressures of CO2 over 3M AMP+1.5M PZ solu- perature was then assumed to be equal to the heat of absorption.
tion as function of CO2 loading and temperature. The detailed tech- The total pressure was measured and the partial pressure of CO2
niques and procedures can be found in our previous work [18,19]. above the solvent and the loading at a given temperature were
calculated according to Hartono, et al. (2008) [23].
3.2.2. Medium pressure VLE apparatus
A medium pressure VLE apparatus was also used to perform 3.2.4. Experimental uncertainties
equilibrium experiments for total pressure data over 3M AMP and For the atmospheric and medium pressure VLE measurements,
1.5M PZ solutions at different CO2 loadings and temperatures. De- the uncertainties were estimated as explained in previous works
tails about the techniques and procedures can be found in previous [16,18,24], whereas for the heat of absorption experiments, the es-
work [18,19]. timation of uncertainties is given in the supporting information
(see Part A1). In the tables showing the data, some experimental
3.2.3. Reaction calorimetry experiments points are marked and have high uncertainty because of the small
A commercially available reaction calorimeter, CPA 202 amount of CO2 added to the reactor. This is further explained in
(ChemiSens AB, Sweden) (see Fig. 1), was used as described the supporting information. This should be considered when using
in previous work [20,21]. An experimental procedure (see Fig. 2) the data.

4
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 5
Measured CO2 solubility data over an aqueous solution of 3.0M AMP+1.5M PZ at different temperatures T, pressures P and
loadings α a .

313K 333K 353K

PC O2 /kPa u(PC O2 ) α u (α ) PC O2 /kPa u(PC O2 ) α u (α ) PC O2 /kPa u(PC O2 ) α u (α )

0.026 0.002 0.19 0.01 0.044 0.002 0.09 0.01 0.036 0.002 0.037 0.01
0.034 0.002 0.21 0.01 0.075 0.002 0.12 0.01 0.059 0.002 0.035 0.01
0.090 0.002 0.29 0.01 0.16 0.002 0.17 0.01 0.11 0.002 0.05 0.01
0.11 0.002 0.31 0.01 0.27 0.003 0.21 0.01 0.24 0.002 0.08 0.01
0.20 0.002 0.38 0.01 0.40 0.03 0.24 0.01 0.36 0.03 0.10 0.01
0.38 0.03 0.44 0.01 0.54 0.05 0.27 0.01 0.48 0.04 0.11 0.01
0.53 0.05 0.47 0.01 0.81 0.07 0.30 0.01 0.71 0.06 0.14 0.01
0.95 0.09 0.52 0.02 0.97 0.09 0.33 0.01 1.36 0.1 0.19 0.01
1.60 0.1 0.58 0.02 2.35 0.2 0.42 0.01 2.21 0.2 0.23 0.01
2.31 0.2 0.60 0.02 4.62 0.4 0.48 0.01 3.24 0.3 0.26 0.01
3.04 0.3 0.61 0.02 6.78 0.6 0.53 0.02 4.25 0.3 0.30 0.01
3.06 0.3 0.62 0.02 8.16 0.7 0.54 0.02 5.24 0.4 0.33 0.01
5.18 0.5 0.66 0.02 11.46 1.0 0.59 0.02 13.06 1.2 0.42 0.01
10.12 1.0 0.70 0.02 19.66 1.8 0.61 0.02 - - - -
13.32 1.2 0.729 0.02 - - - - - - - -
22.14 2.0 0.734 0.02 - - - - - - - -

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

Table 6
Measured total pressure data over an aqueous solution of 3.0M
AMP+1.5M PZ as function of temperature and loading α a .

373K 393K

PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α )

212.2 0.5 0.52 0.02 221.1 0.7 0.25 0.02


220.6 0.5 0.53 0.02 227.9 0.5 0.29 0.02
237.9 0.5 0.54 0.02 244.6 0.8 0.30 0.02
249.9 0.5 0.55 0.02 267.3 1.1 0.31 0.02
391.0 0.7 0.64 0.02 334.2 1.3 0.40 0.02
570.6 1.1 0.69 0.02 427.1 1.8 0.49 0.02
714.5 1.5 0.74 0.03 445.4 1.6 0.51 0.02
894.9 1.7 0.78 0.03 599.3 1.2 0.56 0.02
- - - - 795.4 1.6 0.60 0.02
- - - - 837.4 1.7 0.62 0.02
- - - - 934.9 1.9 0.65 0.02

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

the system go up to 934kPa and the loading covers the range from
0.25 to 0.78 mol CO2 per mol total amine.

4.3. Reaction calorimetry

4.3.1. VLE data


New total pressure data for 3.0M of AMP are reported in Table 7
and for 1.5M PZ in Table 8. The data were obtained at five temper-
Fig. 2. Experimental procedure. atures (313, 333, 353, 373, 393K) with total pressures up to 733
kPa and with CO2 loadings from 0.07 to 0.96 mol of CO2 per mol
of AMP and up to 843 kPa with CO2 loadings from 0.24 to 1.18 mol
4. Experimental results of CO2 per mol of PZ.
Five blends with different mol ratios of AMP/PZ (0.5M/4M,
4.1. Atmospheric pressures VLE 1.5M/3.0M, 2.25M/2.25M, 3.0M/1.5M, 4M/0.5M) were also studied
in the same temperature range. Results are shown in Tables 9-13.
The measured solubilities for CO2 in 3.0M AMP+1.5M PZ aque- The pressures cover up to 693 kPa and loadings up to 1.05 mol
ous solution is given in Table 5. Partial pressures of CO2 ranging CO2 / mol amine.
from 0.026 to 22.14 kPa are reported for three temperatures (313,
333 and 353K) and loadings from 0.035 to 0.73 mol CO2 / mol total 4.3.2. Heat of absorption data
amine. Table 14 shows experimental results for the heat of absorption
of CO2 into 3M AMP where the values vary from 67 to 33 J/mol/K
when going from a CO2 loading of 0.2 to 0.78 mol CO2 / mol AMP.
4.2. Medium pressures VLE The heat of absorption is seen to decrease slightly with increasing
temperature from 313 to 353K. As a sterically hindered amine, the
In Table 6, the measured total pressures for an aqueous solution heat of absorption is seen to be lower than that of MEA [25].
of 3.0M AMP and 1.5M PZ are reported as function of loading at Table 15 shows experimental results for the heat of absorption
two temperatures (373 and 393K). The measured total pressures of of CO2 into 1.5M PZ where the values vary from 79 to 28 J/mol/K

5
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 7
Experimental total pressure data over an aqueous solution of 3.0M AMP at different temperatures T, pressures P and loadings α a .

313K 333K 353K 373K 393K

PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α )

6.8 0.15 0.07 0.01 18.8 0.15 0.07 0.01 45.7 0.20 0.07 0.01 97.9 0.20 0.07 0.01 202.1 0.20 0.07 0.01
6.9 0.15 0.13 0.01 19.3 0.15 0.14 0.01 47.9 0.20 0.13 0.01 108.6 0.20 0.14 0.01 238.0 0.20 0.14 0.01
7.0 0.15 0.20 0.01 20.2 0.15 0.21 0.01 51.5 0.20 0.20 0.01 124.6 0.20 0.21 0.01 288.6 0.20 0.20 0.01
7.2 0.15 0.26 0.01 21.4 0.15 0.28 0.01 56.4 0.20 0.26 0.01 146.4 0.20 0.27 0.01 352.8 0.20 0.26 0.01
7.6 0.15 0.33 0.01 23.2 0.15 0.35 0.01 63.2 0.20 0.33 0.01 174.9 0.20 0.34 0.01 432.4 0.20 0.32 0.01
8.1 0.15 0.39 0.01 25.8 0.15 0.42 0.01 72.6 0.20 0.40 0.01 212.5 0.20 0.40 0.01 528.1 0.20 0.38 0.01
8.9 0.15 0.46 0.01 29.7 0.15 0.49 0.01 85.3 0.20 0.46 0.01 261.1 0.20 0.47 0.01 642.1 0.20 0.43 0.01
10.2 0.15 0.52 0.02 35.6 0.15 0.56 0.02 103.0 0.20 0.52 0.02 325.0 0.20 0.53 0.02 - - - -
12.6 0.15 0.59 0.02 44.8 0.20 0.62 0.02 128.1 0.20 0.59 0.02 406.5 0.20 0.58 0.02 - - - -
16.7 0.15 0.65 0.02 59.4 0.20 0.69 0.02 163.9 0.20 0.65 0.02 510.0 0.20 0.64 0.02 - - - -
24.2 0.15 0.71 0.02 83.9 0.20 0.76 0.02 215.4 0.20 0.71 0.02 640.5 0.20 0.69 0.02 - - - -
38.7 0.15 0.78 0.02 126.8 0.20 0.82 0.02 289.9 0.20 0.76 0.02 - - - - - - - -
81.1 0.20 0.85 0.03 205.1 0.20 0.87 0.03 396.5 0.20 0.81 0.02 - - - - - - - -
174.8 0.20 0.90 0.03 341.6 0.20 0.92 0.03 542.4 0.20 0.86 0.03 - - - - - - - -
609.1 0.20 0.95 0.03 463.0 0.20 0.94 0.03 733.2 0.20 0.90 0.03 - - - - - - - -
- - - - 468.4 0.20 0.94 0.03 - - - - - - - - - - - -
- - - - 665.0 0.20 0.96 0.03 - - - - - - - - - - - -

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

Table 8
Experimental total pressure data over an aqueous solution of 1.5M PZ at different temperatures T, pressures P and loadings α a .

313K 333K 353K 373K 393K

PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α )

7.5 0.15 0.41 0.01 22.1 0.15 0.25 0.01 49.5 0.20 0.27 0.01 102.9 0.20 0.25 0.01 201.7 0.20 0.24 0.01
7.7 0.15 0.55 0.02 22.3 0.15 0.37 0.01 50.4 0.20 0.40 0.01 105.5 0.20 0.37 0.01 211.7 0.20 0.36 0.01
8.4 0.15 0.68 0.02 22.7 0.15 0.49 0.01 53.5 0.20 0.53 0.02 112.0 0.20 0.49 0.01 233.7 0.20 0.47 0.01
12.0 0.15 0.82 0.02 24.4 0.15 0.62 0.02 62.8 0.20 0.66 0.02 128.3 0.20 0.61 0.02 279.4 0.20 0.58 0.02
38.5 0.15 0.94 0.03 30.3 0.15 0.74 0.02 94.3 0.20 0.78 0.02 169.2 0.20 0.72 0.02 368.9 0.20 0.68 0.02
171.9 0.20 1.03 0.03 56.6 0.20 0.85 0.03 189.0 0.20 0.88 0.03 264.4 0.20 0.82 0.02 522.9 0.20 0.76 0.02
380.4 0.20 1.09 0.03 179.8 0.20 0.95 0.03 370.9 0.20 0.96 0.03 443.4 0.20 0.90 0.03 - - - -
607.7 0.20 1.14 0.03 400.8 0.20 1.01 0.03 610.8 0.20 1.01 0.03 699.8 0.20 0.96 0.03 - - - -
843.4 0.20 1.18 0.04 663.9 0.20 1.05 0.03 - - - - - - - - - - - -

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

Table 9
Experimental total pressure data over an aqueous solution of 0.5M AMP+4.0M PZ at different temperatures T, pressures P and loadings α a .

313K 333K 353K 373K 393K

PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α )

6.3 0.15 0.34 0.01 18.5 0.15 0.21 0.01 40.2 0.20 0.22 0.01 92.5 0.20 0.19 0.01 182.9 0.20 0.24 0.01
6.3 0.15 0.42 0.01 18.6 0.15 0.26 0.01 40.5 0.20 0.30 0.01 93.5 0.20 0.26 0.01 189.8 0.20 0.32 0.01
6.4 0.15 0.50 0.02 18.8 0.15 0.37 0.01 41.7 0.20 0.39 0.01 95.2 0.20 0.33 0.01 201.9 0.20 0.40 0.01
6.5 0.15 0.59 0.02 19.0 0.15 0.45 0.01 43.4 0.20 0.47 0.01 98.0 0.20 0.39 0.01 224.9 0.20 0.48 0.01
7.0 0.15 0.67 0.02 19.6 0.15 0.53 0.02 46.0 0.20 0.54 0.02 102.9 0.20 0.46 0.01 264.4 0.20 0.56 0.02
8.4 0.15 0.75 0.02 20.7 0.15 0.61 0.02 51.5 0.20 0.62 0.02 111.7 0.20 0.53 0.02 339.7 0.20 0.63 0.02
14.1 0.15 0.83 0.02 23.8 0.15 0.69 0.02 64.2 0.20 0.70 0.02 128.5 0.20 0.60 0.02 469.8 0.20 0.70 0.02
40.0 0.20 0.91 0.03 33.2 0.15 0.77 0.02 96.0 0.20 0.77 0.02 160.8 0.20 0.67 0.02 680.5 0.20 0.75 0.02
143.5 0.20 0.98 0.03 65.4 0.20 0.85 0.03 172.0 0.20 0.84 0.03 223.9 0.20 0.73 0.02 - - - -
373.1 0.20 1.03 0.03 166.0 0.20 0.92 0.03 337.4 0.20 0.91 0.03 - - - - - - - -
601.6 0.20 1.05 0.03 388.5 0.20 0.97 0.03 472.2 0.20 0.93 0.03 - - - - - - - -
- - - - 629.0 0.20 1.00 0.03 590.3 0.20 0.95 0.03 - - - - - - - -

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

with CO2 loadings varying from 0.4 to 1.2 mol CO2 / mol PZ. As for AMP+ 1.5M PZ). However, the total amine concentration is higher
AMP, the heat of absorption decreases somewhat with increasing which may indicate that solvation, and therefore water, plays an
the temperature from 313 to 353K. As expected for a secondary important role.
diamine, the heat of absorption is lower than that of MEA [25].
For the five blends of AMP/PZ, the heats of absorption are given
in Tables 16-20. The heat of absorption of the blends is more simi- 5. Modeling results
lar to MEA than the AMP or PZ systems. The values vary from 88 to
36 J/mol/K for CO2 loadings from 0.21 to 1.05 mol CO2 /mol amine. The pure component properties used in this work were taken
The heat of absorption is seen to be relatively insensitive to both from the DIPPR database [26] and are given in Table 21. Corre-
temperature and to the blend ratio. The heats of absorption of the lations for various properties, with source reference, are given in
individual solvents (3.0 M AMP and 1.5 MPZ) are seen to be lower Table 22. The equilibrium constants for reactions R1 to R8 are
than for the blends with the same amount of each amine (3.0M shown in Table 23.

6
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 10
Experimental total pressure data over an aqueous solution of 1.5M AMP+3.0M PZ at different temperatures T, pressures P and loadings α a .

313K 353K 373K 393K

PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α )

6.5 0.15 0.45 0.01 44.4 0.20 0.21 0.01 94.1 0.20 0.20 0.01 185.6 0.20 0.18 0.01
6.6 0.15 0.49 0.01 44.7 0.20 0.25 0.01 95.2 0.20 0.25 0.01 189.3 0.20 0.22 0.01
6.7 0.15 0.54 0.02 45.1 0.20 0.29 0.01 96.9 0.20 0.30 0.01 194.4 0.20 0.27 0.01
6.8 0.15 0.58 0.02 45.7 0.20 0.34 0.01 99.2 0.20 0.35 0.01 201.4 0.20 0.31 0.01
7.2 0.15 0.63 0.02 46.5 0.20 0.38 0.01 102.5 0.20 0.40 0.01 211.0 0.20 0.35 0.01
7.9 0.15 0.67 0.02 47.6 0.20 0.42 0.01 107.2 0.20 0.44 0.01 224.0 0.20 0.39 0.01
9.3 0.15 0.72 0.02 49.2 0.20 0.46 0.01 114.2 0.20 0.49 0.01 241.7 0.20 0.44 0.01
12.5 0.15 0.76 0.02 51.6 0.20 0.51 0.02 124.2 0.20 0.54 0.02 265.7 0.20 0.48 0.01
19.6 0.15 0.80 0.02 55.2 0.20 0.55 0.02 139.2 0.20 0.58 0.02 297.9 0.20 0.52 0.02
36.3 0.15 0.85 0.03 60.7 0.20 0.59 0.02 161.4 0.20 0.63 0.02 341.7 0.20 0.56 0.02
72.6 0.20 0.89 0.03 69.5 0.20 0.63 0.02 194.9 0.20 0.67 0.02 400.3 0.20 0.60 0.02
143.4 0.20 0.93 0.03 83.9 0.20 0.68 0.02 234.8 0.20 0.71 0.02 476.9 0.20 0.64 0.02
259.7 0.20 0.96 0.03 107.4 0.20 0.72 0.02 304.1 0.20 0.75 0.02 580.9 0.20 0.67 0.02
410.9 0.20 0.99 0.03 146.3 0.20 0.76 0.02 395.9 0.20 0.79 0.02 - - - -
- - - - 209.4 0.20 0.79 0.02 477.4 0.20 0.81 0.02 - - - -
- - - - 306.9 0.20 0.83 0.02 - - - - - - - -
- - - - 445.4 0.20 0.86 0.03 - - - - - - - -
- - - - 558.2 0.20 0.88 0.03 - - - - - - - -

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

Table 11
Experimental total pressure data over an aqueous solution of 2.25M AMP+2.25M PZ at different temperatures T, pressures P and loadings α a .

313K 333K 353K 373K 393K

PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α )

6.5 0.15 0.37 0.01 19.4 0.15 0.20 0.01 44.2 0.20 0.18 0.01 96.5 0.20 0.19 0.01 209.3 0.20 0.25 0.01
6.6 0.15 0.41 0.01 19.5 0.15 0.25 0.01 45.1 0.20 0.24 0.01 100.0 0.20 0.25 0.01 245.1 0.20 0.34 0.01
6.7 0.15 0.45 0.01 19.6 0.15 0.29 0.01 46.6 0.20 0.30 0.01 105.8 0.20 0.31 0.01 350.1 0.20 0.45 0.01
6.8 0.15 0.49 0.01 19.8 0.15 0.33 0.01 48.6 0.20 0.35 0.01 115.4 0.20 0.37 0.01 545.6 0.20 0.55 0.02
7.1 0.15 0.53 0.02 20.1 0.15 0.37 0.01 52.4 0.20 0.41 0.01 131.3 0.20 0.43 0.01 - - - -
7.6 0.15 0.57 0.02 20.5 0.15 0.41 0.01 59.2 0.20 0.47 0.01 155.4 0.20 0.49 0.01 - - - -
8.6 0.15 0.61 0.02 21.1 0.15 0.45 0.01 72.3 0.20 0.53 0.02 196.3 0.20 0.55 0.02 - - - -
10.6 0.15 0.66 0.02 22.1 0.15 0.49 0.01 98.1 0.20 0.59 0.02 267.0 0.20 0.60 0.02 - - - -
14.9 0.15 0.70 0.02 23.8 0.15 0.53 0.02 149.7 0.20 0.64 0.02 382.4 0.20 0.65 0.02 - - - -
24.4 0.15 0.74 0.02 26.5 0.15 0.57 0.02 250.4 0.20 0.69 0.02 557.2 0.20 0.70 0.02 - - - -
44.1 0.20 0.78 0.02 31.3 0.15 0.61 0.02 419.1 0.20 0.74 0.02 - - - - - - - -
60.5 0.20 0.80 0.02 40.2 0.20 0.65 0.02 - - - - - - - - - - - -
151.0 0.20 0.85 0.03 56.4 0.20 0.69 0.02 - - - - - - - - - - - -
210.4 0.20 0.87 0.03 137.3 0.20 0.76 0.02 - - - - - - - - - - - -
347.8 0.20 0.90 0.03 345.7 0.20 0.82 0.02 - - - - - - - - - - - -

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

Table 12
Experimental total pressure data over an aqueous solution of 3.0M AMP+1.5M PZ at different temperatures T, pressures P and loadings α a .

313K 333K 353K 373K 393K

PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α )

6.6 0.15 0.33 0.01 19.8 0.15 0.20 0.01 45.4 0.20 0.17 0.01 100.3 0.20 0.20 0.01 212.0 0.20 0.20 0.01
6.7 0.15 0.40 0.01 20.0 0.15 0.25 0.01 46.1 0.20 0.22 0.01 107.0 0.20 0.26 0.01 238.4 0.20 0.26 0.01
6.9 0.15 0.46 0.01 20.3 0.15 0.29 0.01 47.2 0.20 0.26 0.01 118.0 0.20 0.33 0.01 279.3 0.20 0.32 0.01
7.4 0.15 0.53 0.02 20.6 0.15 0.33 0.01 48.7 0.20 0.30 0.01 135.8 0.20 0.39 0.01 338.3 0.20 0.39 0.01
8.7 0.15 0.59 0.02 21.1 0.15 0.37 0.01 50.9 0.20 0.35 0.01 164.4 0.20 0.46 0.01 423.9 0.20 0.44 0.01
11.8 0.15 0.66 0.02 22.0 0.15 0.41 0.01 54.1 0.20 0.39 0.01 209.7 0.20 0.52 0.02 545.5 0.20 0.50 0.02
20.1 0.15 0.73 0.02 23.1 0.15 0.46 0.01 58.6 0.20 0.43 0.01 280.6 0.20 0.58 0.02 708.7 0.20 0.55 0.02
41.8 0.20 0.79 0.02 24.8 0.15 0.50 0.01 65.1 0.20 0.47 0.01 388.6 0.20 0.63 0.02 - - - -
94.4 0.20 0.85 0.03 27.5 0.15 0.54 0.02 74.4 0.20 0.52 0.02 548.4 0.20 0.69 0.02 - - - -
206.6 0.20 0.90 0.03 31.9 0.15 0.58 0.02 88.4 0.20 0.56 0.02 - - - - - - - -
315.3 0.20 0.93 0.03 38.7 0.15 0.62 0.02 109.4 0.20 0.60 0.02 - - - - - - - -
338.8 0.20 0.94 0.03 50.1 0.20 0.66 0.02 140.7 0.20 0.64 0.02 - - - - - - - -
- - - - 68.7 0.20 0.70 0.02 186.7 0.20 0.68 0.02 - - - - - - - -
- - - - 99.6 0.20 0.74 0.02 253.4 0.20 0.72 0.02 - - - - - - - -
- - - - 148.2 0.20 0.78 0.02 - - - - - - - - - - - -
- - - - 224.7 0.20 0.82 0.02 - - - - - - - - - - - -
- - - - 332.7 0.20 0.85 0.03 - - - - - - - - - - - -
- - - - 408.1 0.20 0.87 0.03 - - - - - - - - - - - -

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

7
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 13
Experimental total pressure data over an aqueous solution of 4.0M AMP +0.5M PZ at different temperatures T, pressures P and loadings α a .

313K 333K 353K 393K

PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α ) PT /kPa u(PT ) α u (α )

6.7 0.15 0.20 0.01 16.0 0.15 0.23 0.01 49.7 0.20 0.19 0.01 254.2 0.20 0.15 0.01
6.8 0.15 0.26 0.01 17.3 0.15 0.31 0.01 54.2 0.20 0.26 0.01 266.2 0.20 0.17 0.01
7.0 0.15 0.33 0.01 19.6 0.15 0.38 0.01 60.7 0.20 0.32 0.01 308.3 0.20 0.21 0.01
7.4 0.15 0.39 0.01 23.3 0.15 0.46 0.01 70.6 0.20 0.39 0.01 310.4 0.20 0.21 0.01
8.1 0.15 0.46 0.01 29.2 0.15 0.53 0.02 85.2 0.20 0.45 0.01 316.0 0.20 0.21 0.01
9.2 0.15 0.52 0.02 35.9 0.15 0.59 0.02 105.5 0.20 0.51 0.02 321.5 0.20 0.22 0.01
15.3 0.15 0.65 0.02 42.3 0.20 0.62 0.02 - - - - 322.5 0.20 0.22 0.01
22.5 0.15 0.71 0.02 - - - - - - - - - - - -
36.8 0.15 0.77 0.02 - - - - - - - - - - -
38.6 0.15 0.78 0.02 - - - - - - - - - - - -

a)α = C O2 loading = mol C O2 /(mol AMP + mol PZ )

Table 14
Measured heat of absorption data over an aqueous solution of 3.0M AMP at different temperatures T, pressures P and loadings α a a .

313K 333K 353K

−HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α )


(kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−]

67.1 4 0.20 0.01 58.3 7 0.21 0.01 53.2 4 0.20 0.01


66.4 4 0.26 0.01 57.6 7 0.28 0.01 49.1 3 0.26 0.01
63.5 4 0.33 0.01 52.6 7 0.35 0.01 43.8 3 0.33 0.01
59.8 4 0.39 0.01 51.5 7 0.42 0.01 41.6 3 0.40 0.01
56.1 4 0.46 0.01 48.0 7 0.49 0.01 37.7 3 0.46 0.01
53.7 4 0.52 0.02 48.3 7 0.56 0.02 37.6 3 0.52 0.02
48.0 3 0.59 0.02 44.7 7 0.62 0.02 34.2 2 0.59 0.02
48.7 3 0.65 0.02 39.6 7 0.69 0.02 39.1 2 0.65 0.02
41.3 3 0.71 0.02 42.1 7 0.76 0.02 55.3 2 0.71 0.02
43.1 3 0.78 0.02 50.9 8 0.82 0.02 53.6 3 0.76 0.02
33.7 ≥10b 0.78 0.02 48.1 8 0.87 0.03 50.8 4 0.81 0.02
- - - - 45.2 ≥10b 0.92 0.03 48.5 3 0.86 0.03
- - - - 45.1 ≥10b 0.94 0.03 50.0 3 0.90 0.03
a
α = C O2 loading = mol C O2 /(mol AMP + mol PZ )
b
because of small amount of added CO2 , the calculated uncertainty becomes large.

Table 15
Measured heat of absorption data over an aqueous solution of 1.5M PZ at different temperatures T, pressures P and loadings α a a .

313K 333K 353K

−HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α )


(kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−]

78.8 6 0.41 0.01 68.7 7 0.25 0.01 53.1 6 0.27 0.01


75.6 6 0.55 0.02 66.5 6 0.37 0.01 56.1 6 0.40 0.01
66.9 5 0.68 0.02 64.3 6 0.49 0.01 54.8 6 0.53 0.02
56.4 5 0.82 0.02 61.7 6 0.62 0.02 52.9 6 0.66 0.02
48.0 5 0.94 0.03 54.6 6 0.74 0.02 41.6 6 0.78 0.02
38.0 6 1.03 0.03 42.6 6 0.85 0.03 49.2 7 0.88 0.03
30.9 9 1.09 0.03 35.6 7 0.95 0.03 43.7 9 0.96 0.03
28.7 ≥10b 1.14 0.03 36.4 ≥10b 1.01 0.03 35.5 ≥10b 1.01 0.03
28.6 ≥10b 1.18 0.04 30.4 ≥10b 1.05 0.03 53.1 - - -
a
α = C O2 loading = mol C O2 /(mol AMP + mol PZ )
b
because of small amount of added CO2 , the calculated uncertainty becomes large.

5.1. The unloaded AMP/PZ and AMP/PZ/H2O systems depression data remained the same as with the original parame-
ters (τAMP−PZ = τAMP−PZ = 0 ), (see Fig. 4) where the solid line (the
The binary interaction parameters for AMP/PZ should ideally original) and the improved one (dashed line) coincide.
be estimated from data for the binary system. Unfortunately, no Li, et al., [6] reported volatility data for the unloaded system of
data for the pure AMP/PZ blends were available and the inter- 2.3m AMP/5m PZ as a function of temperature as shown in Fig. 5.
action parameters for AMP/PZ were therefore indirectly obtained These volatility data were not used in the parameter fitting. Still,
from fitting of ternary data for the AMP/PZ/H2 O system while the model predicts well the AMP volatility but overpredicts the PZ
keeping the optimum values for the binary interaction energies of volatility somewhat. This may be because the PZ volatility is very
AMP/H2 O (see [10]) and PZ/H2 O (see [12]). The result for the opti- low, and therefore, difficult to measure with high accuracy. We see
mized binary interaction parameters for AMP/PZ were close to zero that the tuning of the AMP/PZ energy interaction parameter does
(τAMP−PZ = τAMP−PZ = 0). A manual tuning of the obtained results, not influence on the volatility predictions. No effort to improve the
setting (τAMP−PZ = −0.5), slightly improved the fit of the activity volatility fit was made.
coefficient model as seen in Fig. 3. The fit to the freezing point

8
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 16
Measured heat of absorption data over an aqueous solution of 0.5M AMP+4.0M PZ at different temperatures T, pressures P and loadings
αa a .
313K 333K 353K

−HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α )


(kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−]

86.8 5 0.34 0.01 58.1 ≥10b 0.21 0.01 72.5 4 0.22 0.01
84.7 4 0.42 0.01 88.0 ≥10b 0.26 0.01 61.1 3 0.30 0.01
84.9 4 0.50 0.02 88.3 ≥10b 0.37 0.01 63.4 3 0.39 0.01
84.0 4 0.59 0.02 82.7 5 0.45 0.01 64.6 4 0.47 0.01
81.3 4 0.67 0.02 86.2 5 0.53 0.02 63.7 4 0.54 0.01
77.7 4 0.75 0.02 79.8 5 0.61 0.02 62.3 3 0.62 0.02
68.9 3 0.83 0.02 75.2 5 0.69 0.02 58.9 3 0.70 0.02
58.1 3 0.91 0.03 65.8 4 0.77 0.02 53.7 3 0.77 0.02
48.6 2 0.98 0.03 53.3 4 0.85 0.03 46.0 3 0.84 0.02
43.1 3 1.03 0.03 50.2 4 0.92 0.03 37.7 3 0.91 0.02
39.3 3 1.05 0.03 46.9 4 0.97 0.03 43.9 7 0.93 0.02
- - - - - - - - 36.4 8 0.95 0.03
- - - - - - - - 35.6 ≥10b 0.96 0.03
a
α = C O2 loading = mol C O2 /(mol AMP + mol PZ )
b
because of small amount of added CO2 , the calculated uncertainty becomes large.

Table 17
Measured heat of absorption data over an aqueous solution of 1.5M AMP+3.0M PZ at different temperatures T, pressures P and loadings
αa a .
313K 333K 353K

−HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α )


(kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−]

78.4 6 0.45 0.01 81.2 7 0.22 0.01 79.7 7 0.21 0.01


84.1 7 0.49 0.01 77.4 7 0.27 0.01 71.8 6 0.25 0.01
82.4 7 0.54 0.02 76.0 7 0.31 0.01 75.5 7 0.29 0.01
82.1 7 0.58 0.02 75.2 7 0.36 0.01 75.9 7 0.34 0.01
82.2 7 0.63 0.02 76.8 7 0.40 0.01 81.3 7 0.38 0.01
81.4 7 0.67 0.02 78.0 7 0.45 0.01 76.0 7 0.42 0.01
78.6 7 0.72 0.02 77.6 7 0.49 0.01 75.4 7 0.46 0.01
79.1 7 0.76 0.02 78.4 7 0.54 0.02 73.5 6 0.51 0.02
82.2 7 0.80 0.02 76.1 7 0.58 0.02 72.7 6 0.55 0.02
79.3 7 0.85 0.03 75.6 7 0.63 0.02 74.9 7 0.59 0.02
77.1 7 0.89 0.03 72.6 7 0.67 0.02 68.8 6 0.63 0.02
73.7 6 0.93 0.03 69.9 7 0.72 0.02 66.7 6 0.68 0.02
68.7 6 0.96 0.03 64.3 6 0.76 0.02 61.9 6 0.72 0.02
65.8 6 0.99 0.03 58.4 6 0.80 0.02 56.3 5 0.76 0.02
- - - - 50.4 6 0.85 0.03 50.4 5 0.79 0.02
- - - - 44.7 8 0.89 0.03 61.5 7 0.83 0.02
- - - - 48.0 7 0.93 0.03 - - - -
- - - - 45.4 7 0.96 0.03 - - - -
- - - - 36.8 ≥10b 0.99 0.03 - - - -
a
α = C O2 loading = mol C O2 /(mol AMP + mol PZ )
b
because of small amount of added CO2 , the calculated uncertainty becomes large.

Table 18
Measured heat of absorption data over an aqueous solution of 2.25M AMP+2.25M PZ at different temperatures T, pressures P and loadings
αa a .
313K 333K 353K

−HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α )


(kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−]

85.7 8 0.37 0.01 87.5 7 0.20 0.01 82.2 5 0.18 0.01


87.9 8 0.41 0.01 79.2 7 0.25 0.01 87.5 5 0.24 0.01
79.0 7 0.45 0.01 85.6 7 0.29 0.01 91.9 6 0.30 0.01
81.1 7 0.49 0.01 86.0 7 0.33 0.01 85.3 5 0.35 0.01
80.7 8 0.53 0.02 85.8 7 0.37 0.01 82.2 5 0.41 0.01
76.4 7 0.57 0.02 87.6 7 0.41 0.01 83.0 5 0.47 0.01
72.2 7 0.61 0.02 85.6 7 0.45 0.01 80.9 5 0.53 0.02
65.9 6 0.66 0.02 82.4 7 0.49 0.01 74.6 5 0.59 0.02
60.1 6 0.70 0.02 78.0 6 0.53 0.02 66.1 5 0.64 0.02
52.9 6 0.74 0.02 77.1 6 0.57 0.02 56.6 4 0.69 0.02
46.9 5 0.78 0.02 72.8 6 0.61 0.02 55.4 5 0.74 0.02
41.4 ≥10b 0.80 0.02 59.6 5 0.65 0.02 - - - -
42.5 ≥10b 0.85 0.03 60.8 5 0.69 0.02 - - - -
36.7 ≥10b 0.87 0.03 55.7 3 0.76 0.02 - - - -
33.7 6 0.90 0.03 49.4 3 0.82 0.02 - - - -
a
α = C O2 loading = mol C O2 /(mol AMP + mol PZ )
b
because of small amount of added CO2 , the calculated uncertainty becomes large.

9
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 19
Experimental values for the heat of absorption over an aqueous solution of 3.0M AMP+1.5M PZ at different temperatures T, pressures P
and loadings α a a .

313K 333K 353K

−HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α )


(kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−]

81.8 5 0.33 0.01 78.9 7 0.20 0.01 82.3 7 0.17 0.01


80.2 5 0.40 0.01 79.5 7 0.25 0.01 84.4 8 0.22 0.01
78.8 4 0.46 0.01 79.3 7 0.29 0.01 83.1 7 0.26 0.01
75.4 4 0.53 0.02 79.8 7 0.33 0.01 83.3 7 0.30 0.01
70.7 4 0.59 0.02 78.8 7 0.37 0.01 83.0 7 0.35 0.01
65.0 4 0.66 0.02 74.9 6 0.41 0.01 80.1 7 0.39 0.01
55.4 3 0.73 0.02 74.4 6 0.46 0.01 78.6 7 0.43 0.01
47.6 3 0.79 0.02 74.0 6 0.50 0.01 74.5 7 0.47 0.01
42.5 3 0.85 0.03 70.8 6 0.54 0.02 70.8 6 0.52 0.02
39.4 3 0.90 0.03 67.8 6 0.58 0.02 67.9 6 0.56 0.02
39.1 6 0.93 0.03 62.7 5 0.62 0.02 63.8 6 0.60 0.02
18.7 ≥10b 0.94 0.03 58.3 5 0.66 0.02 59.4 6 0.64 0.02
- - - - 53.3 5 0.70 0.02 55.3 5 0.68 0.02
- - - - 49.1 4 0.74 0.02 51.5 5 0.72 0.02
- - - - 44.8 4 0.78 0.02 - - - -
- - - - 42.0 4 0.82 0.02 - - - -
- - - - 43.8 5 0.85 0.03 - - - -
- - - - 40.5 9 0.87 0.03 - - - -
a
α = C O2 loading = mol C O2 /(mol AMP + mol PZ )
b
because of small amount of added CO2 , the calculated uncertainty becomes large.

Table 20
Measured heat of absorption data over an aqueous solution of 4.0M AMP+0.5M PZ at different temperatures T, pressures P and loadings
αa a .
313K 333K 353K

−HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α ) −HAbs. u(−HAbs. ) α u (α )


(kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−] (kJ/mol C O2 ) [−]

85.3 5 0.20 0.01 70.8 4 0.23 0.01 77.4 5 0.19 0.01


81.2 4 0.26 0.01 77.1 4 0.31 0.01 80.6 5 0.26 0.01
78.1 4 0.33 0.01 73.3 4 0.38 0.01 77.9 5 0.32 0.01
75.5 4 0.39 0.01 66.6 4 0.46 0.01 76.0 5 0.39 0.01
69.3 4 0.46 0.01 68.3 4 0.53 0.02 75.0 5 0.45 0.01
21.5 ≥10b 0.46 0.01 86.2 7 0.59 0.02 73.7 5 0.51 0.02
72.2 5 0.52 0.02 84.9 9 0.62 0.02 77.4 - - -
48.6 2 0.52 0.02 - - - - - - - -
31.8 1 0.65 0.02 - - - - - - - -
58.6 3 0.71 0.02 - - - - - - - -
53.1 3 0.77 0.02 - - - - - - - -
10.2 ≥10b 0.78 0.02 - - - - - - - -
a
α = C O2 loading = mol C O2 /(mol AMP + mol PZ )
b
because of small amount of added CO2 , the calculated uncertainty becomes large.

Table 21 in different apparatuses, but the results agree very well, both with
Pure component physical properties [26].
the model and earlier data, as seen in Fig. 6 for 3M AMP and Fig. 7
Species for 1.5M PZ. For 3M AMP, it is observed that the data at 313K over-
Properties lap with the data at 323K at the highest loadings, whereas for 1.5M
H2 O CO2 AMP PZ
PZ, the data at 333K are approaching the data at 353K at the high-
TC / K 647.096 304.21 619.818 638.0
est loadings but the difference between the two temperatures is
PC / kPa 22064 7383.0 4862.97 5530
VC / m3 · kmol −1 0.055947 0.0940 0.29650 0.31 still larger than the reported uncertainties (see Table 8).
ω /− 0.344861 0.223621 0.74259 0.41376

5.2.2. Heat of absorption


A comparison between experimental and modelled heats of ab-
5.2. The ternary AMP/H2 O/CO2 and PZ/H2 O/CO2 systems sorption for 3M AMP and 1.5M PZ at three different temperatures
(313, 333 and 353K) is shown in Figs. 8 and 9. It should be noted
The reported parameters for both ternary systems [10,12] were that the heat of absorption data where not used in the fitting. For
directly used in this work. It was not considered to try to im- both systems, it is observed that the heat of absorption seems to
prove the quality of the fit with the new sets of AMP/H2 O/CO2 and decrease slightly with increasing temperature. This is in line with
PZ/H2 O/ CO2 data for total pressure and heat of absorption from earlier findings [10]. The heat of absorption results for 3M AMP at
this work. 333K and 353K are more scattered than those for 313K. The model
overpredicts the heat of absorption at 313K but predicts better at
5.2.1. Total pressure 333/353K especially at higher loadings. The heat of absorption is
The measured total pressures of 3M AMP and 1.5M PZ were seen to decrease rapidly with increasing loading. This is reason-
compared to our previous models [10,12]. The data were measured able, as for low loadings, carbamate formation, with the high heat

10
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 22
Correlations used in the eNRTL model.

No. Correlation Source


4 6 8
1 ln (HCO∞
2
/ MPa ) = −6.8346 + 1.2817 T
·10
− 3.7668
T2
·10
+ 2.997
T3
·10
[36]
2 ln (PHS2 O / Pa ) = 73.649 − 7258
T
.2
− 7.3037 · ln(T ) + 4.1653 · 10−6 · T 2 [26]
3 log(PAMP
S
/ kPa ) = 7.1408 − T1858
−74.9
.4
[10]a
4 log(PPZS
/ kPa ) = 5.9832 − T1215 .77
−113.56
[12]

5 VAMP /cm3 · mol −1 = 84.15 − 5.887 · 10−3 · T + 1.516 · 10−4 · T 2 [37]

6 VPZ /cm3 · mol −1 = 108.84 − 1.5538 · 10−2 · T + 1.8385 · 10−4 · T 2 [26]
7 εAMP = 21.0466 + 8451.2 · ( T1 − 2981.15 ) [10]
8 εPZ = 35.76 + 14836 · ( T1 − 2731.15 ) [12]
Dielectric
constantb
a
There was a typo in the original manuscript written as a logarithmic in a natural basis (ln) but it should
be in a 10 basis (log10)
b
For CO2 and H2 O are taken from Aspen plus.

Fig. 3. Activity coefficients of AMP (1) and PZ(2) in ternary AMP/PZ/H2 O at a). 313K
Fig. 5. Volatility for an unloaded aqueous solution of 2.3 m AMP and 5m PZ at
b). 333K c). 353K d). 363K and e). 373K (Points, [27]; Lines, NRTL; ✷, AMP; ✷, PZ;
different temperatures (✷, AMP; ✷, PZ; ✷, H2 O; [6]; Lines, eNRTL; —, PAMP ; —, PPZ ; —,
Solid lines, τAMP−PZ = τAMP−PZ = 0; Dashed lines, τAMP−PZ = −0.5 and τPZ−AMP = 0).
PH2 O ; Solid lines, τAMP−PZ = τAMP−PZ = 0; Dashed lines, τAMP−PZ = −0.5 and τPZ−AMP =
0).

Fig. 4. Freezing point depression in the ternary AMP/PZ/H2 O system (Points, [28];
Lines, NRTL; Solid line, τAMP−PZ = τAMP−PZ = 0; Dashed line, τAMP−PZ = −0.5 and
τPZ−AMP = 0).

Fig. 6. Representation of the e-NRTL model for the total pressure over an aqueous
of absorption, will affect the values, whereas at higher loadings solution of 3.0M AMP at different temperatures (Points; Exp.; Lines, eNRTL; Squares,
only bicarbonate/carbonate formation takes place. The contribution [10]; , 353K; , 373K; , 393K; Stars, This work; ✷, 313K; ✷, 333K; ✷, 353K, ✷,
of each reaction in the system is shown in different colors. The 373K; ✷, 393K).
heat of protonation of AMP is the main contributor to the total

11
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Fig. 7. Representation of the e-NRTL model for the total pressure over an aqueous
Fig. 9. Representation of the eNRTL model for the total heat of absorption and
solution of 1.5M PZ at different temperatures (Points; Exp; Lines, eNRTL; Squares,
heats of each individual reactions as a function of loading over an aqueous so-
[12]; , 373K; , 393K; Stars, This work; ✷, 313K; ✷, 333K; ✷, 353K, ✷, 373K; ✷,
lution of 1.5 M PZ at different temperatures a).313K b). 333K c). 353K (✷; Exp.;
393K).
Lines, eNRTL; ✷/—, Total; •–•, PZ H + diss.; •–•, PZCO−2
f orm.; —, PZ H +CO−2
diss.; —,
PZ (CO− ) f orm.; —, HCO−3 f orm.; —, CO23− f orm.; —, PZH22+ diss.; —, C O2 dissol.; —,
2 2
H2 O diss.).

5.3. The quaternary AMP/PZ/CO2 /water system

Fifteen species can describe the quaternary system of


AMP/PZ/H2 O/CO2 system, i.e., the sum of species of each ternary
AMP/H2 O/CO2 and PZ/H2 O/CO2 system. No new reaction was
introduced. Although the parameters of each ternary system from
our previous work were kept, the remaining number of parame-
ters to be fitted for the quaternary system is still large (82 binary
interaction energy parameters).
Part of the VLE data in Table 1, as indicated in Table 24, and
the VLE results from this work were used in the regression of the
model parameters for the quaternary system. The heat of absorp-
tion data was not used. The model results for CO2 partial pressure,
total pressure, and heat of absorption are presented and compared
with experimental data.
The goodness of fit is presented in Table 24 as absolute average
relative deviations. We see that the fit is good with an AARD value
Fig. 8. Representation of the eNRTL model predictions for the total heat of absorp- of 13.2 % for the total pressure measurements and 20.4% for the
tion and heats of each individual reaction as a function of loading over an aque- CO2 partial pressure measurements. The obtained optimized pa-
ous solution of 3.0M AMP at different temperatures a).313K b). 333K c). 353K. (✷; rameters for the quaternary system are collected in
Exp.; Lines, eNRTL; ✷/—, Total; —, AMP H + diss.; —, AMPCO− f orm.; —, HCO− f orm.; —,
CO23− f orm.; —, C O2 dissol.; —, H2 O diss.)
2 3
Table 25 where five different color backgrounds show the dif-
ferent systems. In our previous work on the ternary AMP/H2 O/CO2
system [12], unfortunately there were 6 typos in the subscript
heat of absorption up toward a loading of 1 where the dissolution indexes referring to the species. The corrections are shown in
of CO2 takes over. Table B1 (in the supporting information). Readers should fol-
The experimental results for 1.5M PZ look more consistent than low this version and all the parameter values are the same as
those for 3.M AMP as seen in Fig. 9. The heat of absorption of reported in Hartono et al. [12]. The developed eNRTL model
1.5M PZ decreases with increasing temperature as mentioned ear- is expected to valid for 4.2 ≤ [AMP ](mass % ) ≤ 35.7; 4.3 ≤
lier. The model predictions for the total heat of absorption are also [P Z ](mass % ) ≤ 35.8; 0.04 ≤ loading ≤ 1.06; 293K ≤ T (K ) ≤ 433K;
given in Fig. 9 and show less temperature sensitivity than indi- 0.03 ≤ PC O2 (kPa ) ≤ 2100. One must be careful if using the model
cated by the experimental results. For 313K and 333K the agree- outside this range.
ment is good, but at the highest temperatures the heat of absorp-
tion seems to be over-predicted by the model. PZ protonation and 5.3.1. Partial pressure of CO2
PZ-monocarbamate formation give significant contributions up to a Fig. 10 shows three sets of experimental data for the 3M AMP/
loading of 0.6, but the zwitterion formation takes over up to load- 1.5M PZ solvent as function of loading and temperature. The re-
ing 1. After loading 1, CO2 dissolution is the most important part sults from this work were obtained in the same apparatus as used
in the total heat of absorption since the contributions from the by Brúder, et al. (2011) [4]. The repeatably is good as well as the
zwitterion, the protonated PZ and PZ-monocarbamate cancel each agreement between both sets of data. There is probably one out-
other. Carbonate/ bicarbonate and dicarbamate play a minor role. lier at the highest loading from the earlier work [4]. The developed

12
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 23
Equilibrium reaction constants on mol fraction basis used in the modeling.

ln(Ki ) = A + B
T
+ C · ln(T ) + D · T

Reaction Ki A B C D Source

1 KW 132.899 -13445.9 -22.4773 0 [38]


2 KHCO−3 231.465 -12092.1 -36.7816 0 [38]
3 KCO2− 216.049 -12431.7 -35.4819 0 [38]
3
a
4 KPZ H + -161.45575 -2.0232 25.0432 -0.0255 [12]
5 KPZH 2+ -99.65273 0.21748 14.97413 -0.00797 [12]
2
6 KPZCO−2 -2.8788 1992.1691 -2.01598 0 [12]
7 KPZ H + CO−2 -4.8766 -5864.4040 -0.2350 0 [12]
8 KPZ (CO−2 ) -5.6408 537.6375 -1.7401 0 [12]
2
a
9 KAMPH + -65.0354 -3569.787 8.8903 0 [10]
10 KAMPCO−2 -14.0057 2623.735 -1.8581 0 [10]
a
Asymmetric reference state

Table 24
Absolute average relative deviation (AARD) for the
various data sources.

PTotal

AARD/ % Source AARD/ % Source

11.5 This work 10.2 This work


23.7 [4] 24.4 [32]
22.6 [30] - -
20.9 Total 13.1 Total

Fig. 11. Representation of the eNRTL model for the CO2 partial pressure in 40 mass
% of total amine at different AMP/PZ ratio of ((a). 38/2 (b). 35/5 (c). 32/8) mass %
and at different temperatures (Points, [30]; Lines, eNRTL; ✷/—, 303K; ✷/—, 313K;
✷/—, 323K).

0.5PZ; 3M AMP/1.0M PZ) at the same three temperatures. The data


from Yang, et al. (2010) [29] were not used in the fitting, but the
Fig. 10. Representation of the eNRTL model for the CO2 partial pressure over an
model still agrees well with the data, especially at 313K and 333K,
aqueous solution of 3M AMP+1.5M PZ at different temperatures (✷, [29]; ◦, This but overpredicts the PC O2 at 353K. The representation of the model
work; , [4]; Lines, eNRTL; ✷/◦//—, 313K; ✷/◦//—, 333K; ✷/◦//—, 353K). for these different blends can be seen in the supporting informa-
tion (Figures S1(a) for 2M AMP/0.5M PZ; (b) for 2M AMP/1.0M PZ;
(c) for 2M AMP/1.5M PZ; (d) for 3M AMP/0.5M PZ; (e) for 3M
model works well for this blend in representing the CO2 partial AMP/1.0M PZ; (f) for 3M AMP/1.5M PZ).
pressures and fits better with the data from this work than the Dash, et al. (2012) [30] reported data for two different to-
data from Brúder, et al. (2011) [4] as seen from the AARD values in tal amine concentrations of 40 and 50 mass %, each concentra-
Table 24. tion with three different AMP/PZ ratios. Fig. 11(a/b/c) shows the
Yang, et al. (2010) [29] reported CO2 partial pressure data for AMP/PZ ratios (38/2, 35/5 and 32/8) with 40 mass % total amine
the same concentration. The results show that their data agree well concentrations while Fig. 12(a/b/c) shows the AMP/PZ ratios (45/5,
with our data and data from Brúder, et al. (2011) [4] only at 313K 48/2 and 42/8) with 50 mass % total amines.
and 333K at higher loadings. At 353K, the data show lower CO2 The model represents the CO2 partial pressure data reasonably
partial pressures than our data and data from Brúder et al. (2011) well at both total concentrations. For the 40 mass % blends and at
[4] and nearly coincide with the data at 333K. The data of Yang, higher loadings, the differences between the CO2 partial pressures
et al. (2010) [29] indicate a weaker temperature sensitivity than for the three blends become small and the model curves tend to
our data. coincide. This behavior is also seen for the 50 mass % blends. This
In addition to the concentration mentioned above, Yang, et al. indicates that the ratio of AMP/PZ becomes less important at very
(2010) [29], also reported PC O2 data for five blends of AMP/PZ, i.e. high loadings since both AMP and PZ are not available anymore as
2.0M AMP/ 0.5PZ; 2M AMP/1.0M PZ; 2M AMP/1.5M PZ; 3.0M AMP/ free amines.

13
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Table 25
eNRTL parametersa , b , c .
Molecular Parameters
1.65 -745.70 0.00 0.00 a3,1 - b3,1 - a4,1 - b4,1 -
a1,2 3.29 b1,2 -2526.63 a2,1 3.79 b2,1 -1634.58 0.60 698.51 0.34 546.92
2.14 212.85 -2.58 0.00c
a1,3 5.32 b1,3 -1279.99 a2,3 7.84 b2,3 1024.41 a3,2 1.20 b3,2 -3016.76 a4,2 -0.36 b4,2 0.00
a1,4 4.49 b1,4 -372.03 a2,4 0.84 b2,4 0.00 a3,4 -0.50 b3,4 0.00 a4,3 0.00 b4,3 0.00
Molecule-Salt Parameters
a1,5-9 8.00 b1,5-9 0.00 a2,5-9 15.00 b2,5-9 0.00 a3,5-9 -4.09 b3,5-9 531.72 a4,5-9 -2.52 b4,5-9 0.00
a1,5-10 8.00 b1,5-10 0.00 a2,5-10 15.00 b2,5-10 0.00 a3,5-10 -1.36 b3,5-10 366.48 a4,5-10 2.35 b4,5-10 -476.75
a1,5-11 8.00 b1,5-11 0.00 a2,5-11 15.00 b2,5-11 0.00 a3,5-11 6.79 b3,5-11 1238.55 a4,5-11 0.73 b4,5-11 0.00
a1,5-12 -2.43 b1,5-12 -944.19 a2,5-12 5.66 b2,5-12 -360.92 a3,5-12 0.25 b3,5-12 2243.74 a4,5-12 -2.06 b4,5-12 313.78
a1,5-13 5.52 b1,5-13 -1295.81 a2,5-13 -1.30 b2,5-13 1493.25 a3,5-13 1.64 b3,5-13 -248.07 a4,5-13 -3.25 b4,5-13 -523.67
a1,5-14 -1.16 b1,5-14 0.00 a2,5-14 2.38 b2,5-14 -497.24 a3,5-14 5.49 b3,5-14 1078.41 a4,5-14 0.49 b4,5-14 -1498.19
a1,5-15 -7.74 b1,5-15 -314.26 a2,5-15 2.86 b2,5-15 -2851.96 a3,5-15 1.03 b3,5-15 -409.23 a4,5-15 1.42 b4,5-15 740.26
a1,6-9 1.00 b1,6-9 -2305.53 a2,6-9 -3.62 b2,6-9 1242.77 a3,6-9 -5.19 b3,6-9 1064.44 a4,6-9 -3.04 b4,6-9 -967.67
a1,6-10 6.35 b1,6-10 -1830.97 a2,6-10 -2.15 b2,6-10 207.83 a3,6-10 4.75 b3,6-10 -705.53 a4,6-10 -0.68 b4,6-10 -723.74
a1,6-11 -2.25 b1,6-11 1831.93 a2,6-11 4.12 b2,6-11 -1268.70 a3,6-11 4.12 b3,6-11 1312.84 a4,6-11 3.35 b4,6-11 0.00c
a1,6-12 -2.31 b1,6-12 -970.08 a2,6-12 2.35 b2,6-12 -1705.02 a3,6-12 3.40 b3,6-12 -327.55 a4,6-12 -1.76 b4,6-12 495.46
a1,6-13 -1.56 b1,6-13 -2188.65 a2,6-13 3.38 b2,6-13 -1170.06 a3,6-13 0.84 b3,6-13 -1913.44 a4,6-13 2.06 b4,6-13 0.00c
a1,6-14 0.18 b1,6-14 1073.67 a2,6-14 2.84 b2,6-14 580.52 a3,6-14 4.12 b3,6-14 -1618.26 a4,6-14 -2.75 b4,6-14 -232.63
a1,6-15 -1.41 b1,6-15 -3096.83 a2,6-15 -0.98 b2,6-15 -1562.34 a3,6-15 -3.27 b3,6-15 0.00 a4,6-15 -0.15 b4,6-15 0.00c
a1,7-9 3.25 b1,7-9 0.00 a2,7-9 1.88 b2,7-9 0.00 a3,7-9 3.66 b3,7-9 -716.40 a4,7-9 0.00 b4,7-9 271.95
a1,7-10 1.89 b1,7-10 0.00 a2,7-10 0.94 b2,7-10 0.00 a3,7-10 2.58 b3,7-10 -213.77 a4,7-10 2.82 b4,7-10 0.00
a1,7-11 0.13 b1,7-11 0.00 a2,7-11 3.12 b2,7-11 1394.63 a3,7-11 -0.62 b3,7-11 1425.08 a4,7-11 -2.35 b4,7-11 0.00
a1,7-12 -2.87 b1,7-12 377.05 a2,7-12 -1.82 b2,7-12 -337.01 a3,7-12 3.28 b3,7-12 853.16 a4,7-12 0.57 b4,7-12 -361.96
a1,7-13 0.25 b1,7-13 1386.49 a2,7-13 -2.33 b2,7-13 0.00c a3,7-13 4.13 b3,7-13 -151.95 a4,7-13 1.75 b4,7-13 373.82
a1,7-14 2.13 b1,7-14 -1345.70 a2,7-14 0.00 b2,7-14 0.00 a3,7-14 1.63 b3,7-14 1304.27 a4,7-14 0.00 b4,7-14 966.72
a1,7-15 3.33 b1,7-15 604.08 a2,7-15 -4.71 b2,7-15 632.00 a3,7-15 -0.19 b3,7-15 -1055.87 a4,7-15 -1.33 b4,7-15 -464.49
a1,8-9 1.48 b1,8-9 -2687.27 a2,8-9 3.78 b2,8-9 1166.13 a3,8-9 -3.64 b3,8-9 886.85 a4,8-9 1.47 b4,8-9 -219.21
a1,8-10 0.89 b1,8-10 -201.93 a2,8-10 -2.04 b2,8-10 1037.63 a3,8-10 -4.72 b3,8-10 -1892.65 a4,8-10 -0.09 b4,8-10 -523.45
a1,8-11 -0.75 b1,8-11 -1912.27 a2,8-11 -3.56 b2,8-11 0.00 a3,8-11 2.07 b3,8-11 0.00 a4,8-11 -7.53 b4,8-11 -611.22
a1,8-12 -0.14 b1,8-12 -657.51 a2,8-12 -2.32 b2,8-12 1524.71 a3,8-12 -1.30 b3,8-12 1238.20 a4,8-12 0.00c b4,8-12 396.31
a1,8-13 2.05 b1,8-13 0.00 a2,8-13 -2.59 b2,8-13 3229.01 a3,8-13 -4.18 b3,8-13 -2975.16 a4,8-13 -0.55 b4,8-13 102.57
a1,8-14 1.94 b1,8-14 -122.56 a2,8-14 -2.50 b2,8-14 384.52 a3,8-14 -5.65 b3,8-14 535.40 a4,8-14 1.79 b4,8-14 0.00c
a1,8-15 4.82 b1,8-15 -868.78 a2,8-15 -1.08 b2,8-15 -2283.28 a3,8-15 0.46 b3,8-15 2319.17 a4,8-15 3.81 b4,8-15 -608.69
Salt-Molecules Parameters
a5-9,1 -4.00 b5-9,1 0.00 a6-9,1 2.88 b6-9,1 -1349.56 a7-9,1 0.29 b7-9,1 0.00 a8-9,1 -2.76 b8-9,1 1132.73
a5-9,2 -8.00 b5-9,2 0.00 a6-9,2 -1.44 b6-9,2 878.83 a7-9,2 2.85 b7-9,2 -727.43 a8-9,2 1.39 b8-9,2 1875.27
a5-9,3 2.32 b5-9,3 299.64 a6-9,3 7.31 b6-9,3 1061.16 a7-9,3 -2.16 b7-9,3 -327.95 a8-9,3 -5.95 b8-9,3 -379.24
a5-9,4 2.89 b5-9,4 -478.94 a6-9,4 3.34 b6-9,4 -228.20 a7-9,4 -0.43 b7-9,4 0.00 a8-9,4 -4.01 b8-9,4 444.87
a5-10,1 -4.00 b5-10,1 0.00 a6-10,1 -3.29 b6-10,1 319.39 a7-10,1 0.00 b7-10,1 0.00 a8-10,1 -2.55 b8-10,1 295.13
a5-10,2 -8.00 b5-10,2 0.00 a6-10,2 2.36 b6-10,2 -1668.43 a7-10,2 1.23 b7-10,2 0.00 a8-10,2 0.55 b8-10,2 249.46
a5-10,3 -0.48 b5-10,3 -422.88 a6-10,3 -5.81 b6-10,3 -1786.29 a7-10,3 0.00c b7-10,3 0.00c a8-10,3 5.08 b8-10,3 0.00
a5-10,4 0.56 b5-10,4 -636.61 a6-10,4 3.84 b6-10,4 0.00 a7-10,4 -0.46 b7-10,4 0.00 a8-10,4 4.25 b8-10,4 704.83
a5-11,1 -4.00 b5-11,1 0.00 a6-11,1 1.02 b6-11,1 -577.30 a7-11,1 0.00 b7-11,1 0.00 a8-11,1 8.40 b8-11,1 0.00
a5-11,2 -8.00 b5-11,2 0.00 a6-11,2 5.70 b6-11,2 -652.91 a7-11,2 0.94 b7-11,2 1641.81 a8-11,2 4.92 b8-11,2 237.86
a5-11,3 -3.15 b5-11,3 304.98 a6-11,3 2.88 b6-11,3 0.00 a7-11,3 2.17 b7-11,3 0.00 a8-11,3 -1.68 b8-11,3 -219.79
a5-11,4 0.96 b5-11,4 -504.45 a6-11,4 -5.13 b6-11,4 379.28 a7-11,4 0.96 b7-11,4 0.00 a8-11,4 1.18 b8-11,4 -506.30
a5-12,1 -2.99 b5-12,1 -292.36 a6-12,1 -0.67 b6-12,1 0.00 a7-12,1 -3.95 b7-12,1 -402.65 a8-12,1 -2.48 b8-12,1 611.80
a5-12,2 3.81 b5-12,2 349.07 a6-12,2 0.08 b6-12,2 -316.18 a7-12,2 4.50 b7-12,2 -424.77 a8-12,2 1.08 b8-12,2 772.11
a5-12,3 -1.46 b5-12,3 703.19 a6-12,3 0.00 b6-12,3 -1225.12 a7-12,3 -2.91 b7-12,3 493.94 a8-12,3 -1.30 b8-12,3 -215.19
a5-12,4 -3.32 b5-12,4 330.06 a6-12,4 2.99 b6-12,4 -481.24 a7-12,4 -1.90 b7-12,4 -268.45 a8-12,4 -5.84 b8-12,4 1253.56
a5-13,1 -1.75 b5-13,1 -1618.74 a6-13,1 0.93 b6-13,1 -548.33 a7-13,1 -5.72 b7-13,1 701.64 a8-13,1 2.51 b8-13,1 -1560.58
a5-13,2 3.53 b5-13,2 0.00 a6-13,2 4.14 b6-13,2 -2895.55 a7-13,2 -0.29 b7-13,2 -350.40 a8-13,2 -6.23 b8-13,2 -319.45
a5-13,3 -1.91 b5-13,3 213.55 a6-13,3 2.96 b6-13,3 -248.20 a7-13,3 3.75 b7-13,3 1194.48 a8-13,3 0.00 b8-13,3 -1377.65
a5-13,4 -4.11 b5-13,4 399.14 a6-13,4 5.92 b6-13,4 1790.76 a7-13,4 8.57 b7-13,4 -190.58 a8-13,4 -2.21 b8-13,4 380.59
a5-14,1 0.00 b5-14,1 -1015.89 a6-14,1 2.97 b6-14,1 -567.09 a7-14,1 0.00 b7-14,1 0.00 a8-14,1 -1.92 b8-14,1 -1169.18
a5-14,2 0.00 b5-14,2 0.00 a6-14,2 -1.01 b6-14,2 0.00 a7-14,2 1.84 b7-14,2 0.00 a8-14,2 -2.23 b8-14,2 -654.28
a5-14,3 -0.21 b5-14,3 1083.10 a6-14,3 2.54 b6-14,3 1092.90 a7-14,3 -1.09 b7-14,3 117.58 a8-14,3 -2.16 b8-14,3 -2405.48
a5-14,4 1.52 b5-14,4 0.00 a6-14,4 6.36 b6-14,4 723.82 a7-14,4 0.00 b7-14,4 -406.72 a8-14,4 -0.53 b8-14,4 -520.94
a5-15,1 6.43 b5-15,1 1104.30 a6-15,1 0.11 b6-15,1 -358.58 a7-15,1 -2.29 b7-15,1 -413.55 a8-15,1 -3.14 b8-15,1 -947.06
a5-15,2 1.25 b5-15,2 0.00 a6-15,2 -2.22 b6-15,2 -912.63 a7-15,2 0.15 b7-15,2 655.41 a8-15,2 -3.01 b8-15,2 1359.43
a5-15,3 3.83 b5-15,3 -1213.89 a6-15,3 1.79 b6-15,3 -1351.47 a7-15,3 0.83 b7-15,3 711.00 a8-15,3 8.03 b8-15,3 1191.30
a5-15,4 -2.97 b5-15,4 -246.05 a6-15,4 1.59 b6-15,4 736.94 a7-15,4 -2.24 b7-15,4 864.20 a8-15,4 -3.25 b8-15,4 629.91
a
AMP/PZ/H2 O/CO2, AMP/PZ, AMP/H2 O/CO2, PZ/H2 O/CO2, Aspen plus® default value
b
Subscript: 1 = H2 O, 2 = C O2 , 3 = PZ, 4 = AMP, 5 = H3 O+ , 6 = PZ H + , 7 = AMP H + 8 = PZ22+ , 9 = OH − , 10 = HCO−
3
, 11 = CO23− , 12 = PZCO−
2
; 13 = PZ (C O2 )22− , 14 =
AMPCO−2
; 15 = (PZ H +CO−
2
)
c
Set to Zero

Li, et al. (2013) [6] reported CO2 partial pressure data for two total pressure is underpredicted from the static cell total pressure
blends of AMP/PZ, i.e. 2.3m AMP/ 5m PZ and 4.0m AMP/ 2.0m PZ. measurements for the 2.3 m AMP and 5m PZ system. The results
The experiments were performed in two different apparatuses, i.e., are shown in Fig. 13(a) for 2.3m AMP/ 5m PZ and in Fig. 13(b) for
a wetted wall column and a static medium pressure VLE cell. The 4.0m AMP/ 2.0m PZ. The reason for the underpredictions might be
wetted wall column provided data for temperatures up to 373K that the total pressure sensor did not provide good accuracy in the
while the static medium pressure VLE cell provided data from 373 pressure ranges reported as the data seem to give the same level
to 433K. The data were not used in the modeling, but the model of total pressure for both ratios of PZ/AMP.
agrees well with the CO2 partial pressure data generated both in Li, et al. (2013) [6] also reported data on amine volatility in the
the wetted wall column and the static pressure cell. However, the loaded 2.3m AMP and 5m PZ system. These data were not used in

14
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Fig. 14. Volatility data for 2.3 m AMP and 5m PZ at different temperatures a). 303K
Fig. 12. Representation of the eNRTL model for the CO2 partial pressure in 50 mass b). 308K c). 313K d). 318K e). 323K and f). 328K (Stars, [6]; Lines, eNRTL; ✷/—, PPZ ;
% of total amine at different AMP/PZ ratio of ((a). 48/2 (b). 45/5 (c). 42/8) mass % ✷/—, PAMP ; ✷/—, PH2 O ; ✷/—, PC O2 ).
and at different temperatures (Points, [30]; Lines, eNRTL; ✷/—, 318K; ✷/—, 338K).

Fig. 13. CO2 solubility data for (a). 2.3 m AMP and 5m PZ (b). 4.0 m AMP and 2.0 Fig. 15. Representation of the eNRTL model for the total pressure over an aque-
m PZ at different temperatures (Points, [6]; Lines, eNRTL; ✷/—, 293K; ✷/—, 313K; ous solution of 3.0M AMP+1.5M PZ at different temperatures (Squares, data from
✷/—, 323K; ✷/—, 353K; ✷/—, 373K; ✷/—, 383K; /•–•, 393K; /•–•, 403K; /•–•, a medium pressure; Stars, data from a calorimeter; Stars, Exp.; Lines, eNRTL; ✷/—,
413K; /•–•, 423K; /•–•, 433K). 313K; ✷/—, 333K; ✷/—, 353K, ✷//—, 373K; ✷//—, 393K).

the fitting, but the model works well. The results are presented in PZ system at 373 and 393K. The agreement between the two data
Fig. 14 for different temperatures from 303 to 328K. The volatility sets obtained is reasonably good, as seen in Fig. 15. The model is
of both amines decreases with increasing loading as expected since also able to represent the data satisfactorily.
less and less free amine is available. The volatility increases with For other blends, the experiments were performed in the
temperature as also expected. AMP is more volatile than PZ as seen calorimeter and the results are shown in Fig. 16 (a). for 0.5M
by both data and model. AMP+4.0M PZ, (b). for 1.5M AMP+3.0M PZ, (c). for 2.0M
The reported CO2 partial pressure data of Jahangiri and AMP+2.0M PZ, and (d). for 4.0M AMP+0.5M PZ respectively. The
Nabipoor Hassankiadeh (2019)[31] are deemed inconsistent since model predicts reasonably well the total pressure of the system.
higher CO2 partial pressures were reported at lower tempera- Tong, et al. (2013) [32] reported data for 30 mass % total amine
tures and loadings as shown in Figure S2 (see supporting informa- for two different ratios of AMP/PZ as seen in Fig. 17(a) for 25/5
tion). The reason for this might be that equilibrium was not fully mass % of AMP/PZ and Fig. 17(b) for 20/10 mass % of AMP/PZ. The
achieved. model agrees very well with the data at the lowest temperature
(333K) but deviates more at higher temperatures.
5.3.2. Total pressure Murshid, et al. (2014) [33] reported total pressure data (from
Two different apparatuses, i.e., the medium pressure VLE 0.5 to 6 MPa) for two AMP concentrations (8.9 and 17.9 mass %)
(shown as squares) and the calorimeter (shown as stars) were used in blends with different PZ concentrations (1.74, 6.88 and 10.35
to perform total pressure measurements for the 3.0M AMP+1.5M mass % PZ) and at two temperatures (318 and 338K). The reported

15
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Fig. 16. Representation of the eNRTL model for the total pressure over aqueous so-
lutions of (a). 0.5M AMP+4.0M PZ (b). 1.5M AMP+3.0M PZ (c). 2.25M AMP+2.25M Fig. 18. Representation of the eNRTL model for the total heat of absorption and
PZ) (d). 4.0M AMP+0.5M PZ as a function of loading and temperature (Stars, Exp.; heats of each individual reactions as a function of loading over an aqueous solution
Lines, eNRTL; ✷/—, 313K; ✷/—, 333K; ✷/—, 353K, ✷/—, 373K; ✷/—, 393K). of 3.0M AMP+1.5M PZ at different temperatures a). 313K b). 333K c). 353K (Points,
Exp.; ✷, This work; , [34]; Lines, eNRTL; —, Total; —, AMP H + diss.; •–•, PZ H + diss.;
—, AMPCO− 2
f orm.; •–•, PZCO− 2
f orm.; —, PZ H +CO− 2
diss.; —, PZ (CO− ) f orm.; —,
2 2
HCO− 3
f orm.; —, CO23− f orm.; —, PZH22+ diss.; —, C O2 dissol.; —, H2 O diss.).

well at higher loadings but at lower loadings, but some disagree-


ment is seen for the lowest temperature at lower loadings. Xie,
et al. (2013) [34] show the same trend when presenting their 5M
MEA data with literature data. No significant effect of temperature
was found in the heat of absorption data.
The modeled total heat of absorption and the contributions of
each reaction are also given in Fig. 18. The agreement between
the model and experimental results is reasonable, but the model
shows more temperature-sensitivity than the data. Under ideal
conditions, where the excess heat of reaction can be disregarded,
a Van ’t Hoff relationship can be used to correlate the heat of ab-
sorption and the CO2 partial pressure. From Fig. 10, it is seen that
the differences in CO2 partial pressure between 333 and 353K is
smaller than between 313 and 333K at a given loading for loadings
up to 0.5 mol of CO2 / mol (AMP+PZ). This indicates that the heat
of absorption is expected to decrease with increasing temperature.
Fig. 17. Representation of the eNRTL model for the total pressure over aqueous so- At higher loadings the differences remain almost constant but are
lutions of (a). 25 mass % AMP/ 5 mass % PZ) (b). 20 mass % AMP/ 10 mass % PZ smaller than the differences at lower loading, implying little effect
and at different temperatures (Stars, [32]; Lines, eNRTL; ✷/—, 313K; ✷/—, 333K; —,
of temperature on the heat of absorption in this range. The heat
353K, ✷/—, 373K; ✷/—, 393K).
of absorption is predicted by the model to decrease slightly with
increasing temperature. It must be noted that the reported data
loadings were very high, ranging from 1.01 to 2.66 mol CO2 per were not used in the fitting. At low loadings, significant contribu-
mol amine (AMP+PZ). The data were not used in the fitting, and tions from PZ-monocarbamate, protonated PZ and protonated AMP
the model cannot represent the reported total pressure data for are observed. By increasing the loading, only protonated AMP con-
these high loadings (see Figure S3 in the supporting information). tributes significantly, but at loading approaching unity, the zwit-
The reported data indicate that concentration still affects the to- terion formation is a significant contributor while contributions
tal pressure at loadings well above 1. It is expected that very little from PZ-monocarbamate, PZ-dicarbamate and carbonate/ bicarbon-
free AMP and PZ is left in the system at this loading. When no ate become smaller.
free amine is left, the ratio of AMP/PZ becomes less important and Measured heats of absorption for different ratios of AMP/PZ
the data tend to coincide, as seen earlier. Another reason for the from this work are presented together with model comparisons
disagreement with the model could be that for high pressure mea- in the supporting information: Figure S4 for 0.5M AMP+ 4.0M PZ,
surements, the accuracy of the pressure transducer is lower, and Figure S5 for 4.0M AMP+0.5M PZ, Figure S6 for 1.5M AMP+3.0M
the uncertainty is higher. Efforts were made to improve the fit by PZ and Figure S7 for 2.0M AMP+2.0M PZ. Furthermore, model
using these data in the regression, but without success. comparisons with the results from the work of Xie, et al. [34] are
also presented in the supporting information: Figure S8 for 3.0M
5.3.3. Heat of Absorption AMP+0.5M PZ and Figure S9 for 3.0M AMP+1.0M PZ. It may be
For an aqueous solution of 3.0 M AMP+1.5 M PZ, the measured seen that the model predictions are good at 333 and 353K but ex-
heats of absorption in this work are compared with the reported ceeds the experimental values at 313K and at loadings lower than
data of Xie, et al. (2013) [34] as shown in Fig. 18. The data agree 0.5.

16
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Fig. 19. Representation of the eNRTL model for the reported speciation of (a). 4m AMP/ 2m PZ (300K) (b). 4m AMP/ 2m PZ (313K) (c). 2.3m AMP/ 5m PZ (313K) and for the
full speciation of (d). 4m AMP/ 2m PZ (300K) (e). 4m AMP/ 2m PZ (313K) (f). 2.3m AMP/ 5m PZ (313K) (Stars, [35]; Lines, eNRTL; ✷/—/•–•, AMP/AMP H + ; —, AMPCO− 2
; ✷/
—/•–•, P Z/P Z H + /P ZH22+ ; ✷/—/•–•, P ZCO−
2
/P Z H +CO−
2
; ✷/—/, PZ (CO− ) ; ✷/—/, HCO−3 /CO23− ; —, C O2 ).
2 2

The model suggests that the contribution of each reaction de- ported CO2 partial pressure data heat of absorption data from this
pends on the ratio of AMP/PZ. If the AMP concentration is higher work over an aqueous solution of 3M AMP+1.5M PZ show good
than PZ, the contribution from protonated AMP plays an important agreement with literature and indicates that both sets of data are
role. In contrast, when the PZ concentration is higher, the contri- consistent.
butions of protonated PZ, PZ-monocarbamate, and the zwitterion The measured solubility data and selected data reported
are important at different loadings, as in the PZ/H2 O/CO2 system. in the literature were used in the regression to estimate bi-
nary interaction energy parameters in the quaternary system
5.3.4. Speciation (AMP/PZ/H2 O/CO2 ). The equilibrium constants and binary interac-
Li, et al. (2014) [35] reported speciation data for blends of 4m tion energy parameters from our previous works on the ternary
AMP+2m PZ at (300K and 313K) and for 2.3m AMP+5m PZ, at systems of AMP/H2 O/CO2 and PZ/H2 O/CO2 were used directly.
313K. A 13C-NMR method was used to measure the concentra- The results show that the model represents the data very well
tions of AMP/AMP H + , P Z/P Z H + /P ZH22+ , P ZCO−
2
/P Z H +CO−2
, P Z (CO− )
2 2
and the AARD values are 13.1% for total pressure and 20.9%
− 2−
and HCO3 /CO3 , respectively. The quality of the NMR data does for CO2 partial pressure. The volatility of AMP and PZ as func-
not agree well with their analytical results, since the mass bal- tion of loading and temperature and the reported speciation data
ance for AMP deviates up to 25.1%, for PZ up to 23.4% and for for the AMP/AMP H + , P Z/P Z H + /P ZH22+ , P ZCO−
2
/P Z H +CO−
2
, P Z (CO− )
2 2
− 2−
CO2 up to 7.6 %. These high numbers show the challenges in the and HCO3 /CO3 were also well predicted. The developed eNRTL
assignments and quantifications of the NMR spectra. The eNRTL model is expected to valid for 4.2 ≤ [AMP ](mass % ) ≤ 35.7; 4.3 ≤
model, to some extent, predicts satisfactorily the speciation data of [P Z ](mass % ) ≤ 35.8; 0.04 ≤ loading ≤ 1.06; 293K ≤ T (K ) ≤ 433K;
AMP/AMP H + , P Z/P Z H + /P ZH22+ , P Z (CO− ) and HCO− /CO23− , but for 0.03 ≤ PC O2 (kPa ) ≤ 2100. One must be careful if using the model
2 2 3
the P ZCO− 2
/P Z H +CO− , the model underpredicts the data as seen in
2
outside this range.
Fig. 19 (a/b/c). The full speciation model results for the considered The heat of absorption model predicts satisfactory the heat of
conditions are presented in Fig. 19 (d/e/f). The NMR data were not absorption data for the studied different ratios AMP/PZ. Individ-
used in the regression. ual reactions contributions are also presented. It is observed that
when the AMP/PZ ratio increases, the predicted total heat absorp-
6. Conclusions tion goes toward the single amine AMP system where protonated
AMP is the main contributor. When the AMP/PZ ratio decreases,
New data comprising CO2 partial pressure for 3.0M AMP/1.5M the predicted total heat of absorption approaches the single amine
PZ and total pressure and heat of absorption for different ratios PZ system where the protonated PZ, PZ-monocarbamate and zwit-
of AMP/PZ (3.0/0.0, 0.0/1.5, 0.5/4.0, 1.5/3.0, 2.5/2.5, 3.0/1.5, 4.0/0.5) terion reactions play important roles. The contribution of the zwit-
M are provided at different loadings and temperatures from three terion, protonated PZ and PZ-carbamate cancel each other at high
different apparatuses, i.e. low and medium pressures VLE appara- loadings.
tuses and a CPA202 reaction calorimeter. The generated VLE data Finally, no data for pure AMP/PZ blends were available and
of this work from two different apparatuses (The medium pres- the binary interaction energy parameter for AMP/PZ was fit-
sure and the CPA reaction calorimetric) are in a good agreement. ted in the ternary AMP/PZ/H2 O system using literature data. A
The reported total pressure over aqueous solutions of 3M AMP and fine-tuning was needed to improve the fit for the AMP/PZ/H2 O
1.5M PZ are consistent with our previous work whereas the re- system.

17
A. Hartono, R. Ahmad, H.F. Svendsen et al. Fluid Phase Equilibria 550 (2021) 113235

Declaration of competing interest [15] D.-Y. Peng, D.B. Robinson, A new Two-constant equation of state, Ind. Eng.
Chem. Fund. 15 (1976) 59–64, doi:10.1021/i160057a011.
[16] S. Ma’mun, J.P. Jakobsen, H.F. Svendsen, O. Juliussen, Experimental and mod-
The authors declare no competing financial or other conflict of eling study of the solubility of carbon dioxide in aqueous 30 mass % 2-((2-
interest. Aminoethyl)amino)ethanol solution, Ind. Eng. Chem. Res. 45 (2006) 2505–
2512, doi:10.1021/ie0505209.
[17] S. Ma’mun, H.F. Svendsen, K.A. Hoff, O. Juliussen, Selection of new absorbents
CRediT authorship contribution statement for carbon dioxide capture, Energy Convers. Manage. 48 (2007) 251–258,
doi:10.1016/j.enconman.20 06.04.0 07.
Ardi Hartono: Conceptualization, Methodology, Validation, In- [18] U.E. Aronu, S. Gondal, E.T. Hessen, T. Haug-Warberg, A. Hartono, K.A. Hoff,
H.F. Svendsen, Solubility of CO2 in 15, 30, 45 and 60 mass% MEA from 40
vestigation, Software, Formal analysis, Resources, Data curation,
to 120°C and model representation using the extended UNIQUAC framework,
Writing – original draft, Writing – review & editing, Visualization, Chem. Eng. Sci. 66 (2011) 6393–6406, doi:10.1016/j.ces.2011.08.042.
Supervision. Rafiq Ahmad: Methodology, Validation, Investigation, [19] S. Ma’mun, R. Nilsen, H.F. Svendsen, O. Juliussen, Solubility of Carbon Dioxide
in 30 mass % Monoethanolamine and 50 mass % Methyldiethanolamine Solu-
Resources. Hallvard F. Svendsen: Conceptualization, Methodology,
tions, J. Chem. Eng. Data 50 (2005) 630–634, doi:10.1021/je0496490.
Software, Formal analysis, Writing – review & editing, Visualiza- [20] A. Hartono, S.J. Vevelstad, I. Kim, R. Rennemo, H.K. Knuutila, Promoted Strong
tion, Supervision. Hanna K. Knuutila: Writing – review & editing, Bicarbonate Forming Solvents for CO2 Capture, Energy Procedia 114 (2017)
Project administration, Funding acquisition. 1794–1802, doi:10.1016/j.egypro.2017.03.1307.
[21] A. Hartono, I. Kim, H.K. Knuutila, L.V. van der Ham, E. Goetheer, Calorimetric
Studies of Precipitating Solvent System, Energy Procedia 114 (2017) 744–755,
Acknowledgments doi:10.1016/j.egypro.2017.03.1217.
[22] I. Kim, H.F. Svendsen, Heat of absorption of carbon dioxide (CO2) in Mo-
noethanolamine (MEA) and 2-(Aminoethyl)ethanolamine (AEEA) solutions, Ind.
This work is performed under the ACT ALIGN CCUS Project No Eng. Chem. Res. 46 (2007) 5803–5809, doi:10.1021/ie0616489.
271501. This project has received funding from RVO (NL), FZJ/PtJ [23] A. Hartono, O. Juliussen, H.F. Svendsen, Solubility of N2O in aqueous solution
(DE), Gassnova (NO), UEFISCDI (RO), BEIS (UK) and is cofunded of diethylenetriamine, J. Chem. Eng. Data 53 (2008) 2696–2700, doi:10.1021/
je800409d.
by the European Commission under the Horizon 2020 programme [24] S.L.R. Ellison, A. Williams, Eurachem/CITAC guide: quantifying uncertainty in
ACT, Grant Agreement No 691712. analytical measurement, 2012.
[25] I. Kim, K.A. Hoff, T. Mejdell, Heat of absorption of CO2 with aqueous solutions
of mea: new experimental data, Energy Procedia (2014) 1446–1455.
Supplementary materials
[26] DIPPR-801, The information and data evaluation manager for design institute
for physical properties, Version 4.1.0., 2004.
Supplementary material associated with this article can be [27] A. Hartono, M. Saeed, A.F. Ciftja, H.F. Svendsen, Binary and ternary VLE of the
2-amino-2-methyl-1-propanol (AMP)/piperazine (Pz)/water system, Chem. Eng.
found, in the online version, at doi:10.1016/j.fluid.2021.113235.
Sci. 91 (2013) 151–161, doi:10.1016/j.ces.2013.01.015.
[28] P.L. Fosbøl, R. Neerup, M.W. Arshad, Z. Tecle, K. Thomsen, Aqueous solubility of
References piperazine and 2-Amino-2-methyl-1-propanol plus their mixtures using an im-
proved freezing-point depression method, J. Chem. Eng. Data 56 (2011) 5088–
[1] B. Metz, O. Davidson, H. de Coninck, M. Loos, L. Meyer, IPCC Special Report 5093, doi:10.1021/je200959m.
on Carbon Dioxide Capture and Storage. Prepared by Working Group III of the [29] Z.Y. Yang, A.N. Soriano, A.R. Caparanga, M.H. Li, Equilibrium solubility of car-
Intergovernmental Panel on Climate Change, Cambridge, UK, 2005. bon dioxide in (2-amino-2-methyl-1-propanol + piperazine + water), J. Chem.
[2] G.T. Rochelle, Amine Scrubbing for CO2 Capture, Science 325 (2009) 1652– Thermodyn. 42 (2010) 659, doi:10.1016/j.jct.20 09.12.0 06.
1654, doi:10.1126/science.1176731. [30] S.K. Dash, A.N. Samanta, S.S. Bandyopadhyay, Experimental and theoretical in-
[3] H.F. Svendsen, E.T. Hessen, T. Mejdell, Carbon dioxide capture by absorption, vestigation of solubility of carbon dioxide in concentrated aqueous solution
challenges and possibilities, Chem. Eng. J. 171 (2011) 718–724, doi:10.1016/j. of 2-amino-2-methyl-1-propanol and piperazine, J. Chem. Thermodynamics 51
cej.2011.01.014. (2012) 120–125, doi:10.1016/j.jct.2012.02.012.
[4] P. Brúder, A. Grimstvedt, T. Mejdell, H.F. Svendsen, CO2 capture into aqueous [31] A. Jahangiri, M.Nabipoor Hassankiadeh, Effects of piperazine concentration and
solutions of piperazine activated 2-amino-2-methyl-1-propanol, Chem. Eng. operating conditions on the solubility of CO2 in AMP solution at low CO2 par-
Sci. 66 (2011) 6193–6198, doi:10.1016/j.ces.2011.08.051. tial pressure, Sep. Sci. Technol. 54 (2019) 1067–1078, doi:10.1080/01496395.
[5] J.N. Knudsen, J. Andersen, J.N. Jensen, O. Biede, Evaluation of process upgrades 2018.1524907.
and novel solvents for the post combustion CO2 capture process in pilot-scale, [32] D. Tong, G.C. Maitland, M.J.P. Trusler, P.S. Fennell, Solubility of carbon dioxide
Energy Procedia 4 (2011) 1558–1565, doi:10.1016/j.egypro.2011.02.025. in aqueous blends of 2-amino-2-methyl-1-propanol and piperazine, Chem. Eng.
[6] H. Li, L. Li, T. Nguyen, G.T. Rochelle, J. Chen, Characterization of piperazine/2- Sci. 101 (2013) 851–864, doi:10.1016/j.ces.2013.05.034.
aminomethylpropanol for carbon dioxide capture, Energy Procedia 37 (2013) [33] G. Murshid, A.M. Shariff, M.A. Bustam, S. Ullah, Solubility of CO2 in Piper-
340–352, doi:10.1016/j.egypro.2013.05.120. azine (PZ) activated aqueous solutions of 2-Amino-2-methyl-1-propanol (AMP)
[7] A.K. Chakraborty, G. Astarita, K.B. Bischoff, CO2 absorption in aqueous solu- at elevated pressures, Appl. Mech. Materr. 625 (2014) 233–236 10.4028/www.
tions of hindered amines, Chem. Eng. Sci. 41 (1986) 997–1003, doi:10.1016/ scientific.net/AMM.625.233.
0 0 09- 2509(86)87185- 8. [34] Q. Xie, A. Aroonwilas, A. Veawab, Measurement of heat of CO2 absorption into
[8] A.F. Ciftja, A. Hartono, H.F. Svendsen, Experimental study on carbamate forma- 2-Amino-2-methyl-1-propanol (AMP)/Piperazine (PZ) blends using differential
tion in the AMP–CO2–H2O system at different temperatures, Chem. Eng. Sci. reaction calorimeter, Energy Procedia 37 (2013) 826–833, doi:10.1016/j.egypro.
107 (2014) 317–327, doi:10.1016/j.ces.2013.12.028. 2013.05.175.
[9] A.F. Ciftja, A. Hartono, E.F. da Silva, H.F. Svendsen, Study on carbamate stability [35] H. Li, P.T. Frailie, G.T. Rochelle, J. Chen, Thermodynamic modeling of
in the AMP/CO2 /H2 O system from 13 C-NMR spectroscopy, Energy Procedia 4 piperazine/2-aminomethylpropanol/CO2/water, Chem. Eng. Sci. 117 (2014) 331–
(2011) 614–620, doi:10.1016/j.egypro.2011.01.096. 341, doi:10.1016/j.ces.2014.06.026.
[10] A. Hartono, R. Ahmad, M. Usman, N. Asif, H.F. Svendsen, Solubility of CO2 in [36] J.J. Carroll, J.D. Slupsky, A.E. Mather, The solubility of carbon dioxide in water
0.1M, 1M and 3M of 2-amino-2-methyl-1-propanol (AMP) from 313 to 393K at low pressure, J. Phys. Chem. Ref. Data 20 (1991) 1201–1209, doi:10.1063/1.
and model representation using the eNRTL framework, Fluid Phase Equilib. 511 555900.
(2020) 112485, doi:10.1016/j.fluid.2020.112485. [37] K. Zhang, B. Hawrylak, R. Palepu, P.R. Tremaine, Thermodynamics of aqueous
[11] C.-C. Chen, L.B. Evans, A local composition model for the excess Gibbs energy amines: excess molar heat capacities, volumes, and expansibilities of {water+
of aqueous electrolyte systems, AIChE J. 32 (1986) 444–454, doi:10.1002/aic. methyldiethanolamine (MDEA)} and {water + 2-amino-2-methyl-1-propanol
690320311. (AMP)}, J. Chem. Thermodynamics 34 (2002) 679–710, doi:10.10 06/jcht.20 02.
[12] A. Hartono, A. Rafiq, S. Gondal, H.F. Svendsen, Solubility data for Nitrous Oxide 0937.
(N2O) and Carbon dioxide (CO2) in Piperazine (PZ) and a new eNRTL model, [38] T.J. Edwards, G. Maurer, J. Newman, J.M. Prausnitz, Vapor-liquid equilibria in
Fluid Phase Equilib. 538 (2021) 112992, doi:10.1016/j.fluid.2021.112992. multicomponent aqueous solutions of volatile weak electrolytes, AIChE J. 24
[13] E.T. Hessen, T. Haug-Warberg, H.F. Svendsen, The refined e-NRTL model ap- (1978) 966–976, doi:10.1002/aic.690240605.
plied to CO2–H2O–alkanolamine systems, Chem. Eng. Sci. 65 (2010) 3638–
3648, doi:10.1016/j.ces.2010.03.010.
[14] J.G.M.S. Monteiro, D.D.D. Pinto, S.A.H. Zaidy, A. Hartono, H.F. Svendsen, VLE
data and modelling of aqueous N,N-diethylethanolamine (DEEA) solutions, Int.
J. Greenhouse Gas Control 19 (2013) 432–440, doi:10.1016/j.ijggc.2013.10.001.

18

You might also like