How the Presence of CO₂ Absorption Promoters and Composition of the Choline Chloride/Amine/Molecular Solvent Mixtures Influence Its Thermophysical Properties and Ability to Absorb Carbon Dioxide

Abstract

The present research provides data on the density (ρ), viscosity (η) and ability to absorb carbon dioxide of systems containing amine, molecular solvent (MS) and choline chloride (ChCl), with the investigation of the physical properties of both neat amine/MS/ChCl mixtures and their samples after complete CO₂ saturation. The effect of the mixture composition was studied by varying amine from primary (monoethanolamine, MEA) to secondary (diethanolamine, DEA) and tertiary (triethanolamine, TEA) amine, and the degree of its substitution from a mono- (MEA, DEA and TEA) to a doubly-substituted (ethylenediamine, EDA) compound. The role of an MS was investigated via the exchange of ethylene glycol (EG) with water and dimethylsulfoxide (DMSO). In addition, the influence of the CO₂ absorption promoters present in the ternary MEA/EG/ChCl mixture at an amount of 5 wt. % was also investigated. We show that an increase in ρ and η in the amine/EG/ChCl mixture affects the properties of neat amines. This suggests that in the studied ternary mixtures, the nature of the interspecies interactions is very similar to those in the previously studied MEA/EG/ChCl system. When EG was exchanged for H₂O or DMSO, a decrease in ρ and an increase in η were observed. A comparison of the data with the corresponding properties of the systems composed of each pair of the mixture components indicates that the intensity and/or number of interspecies interactions in the present ternary mixtures were stronger than those in pure H₂O, DMSO and MEA. While in the presence of promoters no significant changes in the studied properties were found, for the corresponding CO₂-saturated samples, the ρ and η increased proportionally to the amount of absorbed gas. This was also the case for all the systems studied in the present research. The overall CO₂ absorption of the EG-based mixtures decreased when going from primary to secondary and tertiary amines; it was 21% higher for the MEA/H₂O/ChCl system compared to the mixtures containing EG and DMSO, which, in turn, showed similar absorption capacities. When the promoters were added to the MEA/EG/ChCl mixture, the highest capacity was found for the piperazine-containing system.

1. Introduction

Even though novel ecological trends have stimulated the rapid development of renewable energy sources, energy supply remains widely dependent on carbonaceous fuels, especially fossil fuels such as crude oil, coal and natural gas [1]. Their incineration in mobile power systems, power and industrial plants produces impressive amounts of greenhouse gas emissions, among which carbon dioxide is considered to be the most dangerous pollutant. An increase in its concentration in the atmosphere leads to global warming, and, as a consequence, to ocean acidification, the melting of glaciers, land desertification, and many other negative factors [2]. The Intergovernmental Panel on Climate Change has shown that 79% of carbon dioxide comes from burning fossil fuels that are used to generate electricity in coal-fired power plants, which are responsible for up to 60% of all CO₂ emissions [3].

Despite the intensive development of Carbon Capture and Storage (CCS) technologies aimed at reducing the negative impact of CO₂, its concentration in the atmosphere increased from 340 ppm in 1980 to 408 ppm in 2019 [4]. This means that the problem of the extraction of carbon dioxide from flue gas mixtures and its further processing will not lose its importance in the near future. The most common CCS techniques involve physical [5,6,7,8,9,10] and chemical [11,12,13] absorption, as well as low-temperature cryogenic [14,15] and membrane separation [14,16,17,18] processes. Among them, chemical absorption has been the most widely used and well-commercialized. It occupies a wide industry segment [19,20]; allows for the recycling of absorbents; enables temporary CO₂ fixing with further accumulation [21]; can be easily customized for favorable industrial process due to the diversity of compound structures [22]. When compared to alkali solutions, aqueous alkanolamines show lowered corrosion harm [23], and therefore have gained greater acceptance. Such primary amines as monoethanolamine (MEA) and 2-amino-2-methyl-1-propanol (AMP), secondary diethanolamine (DEA) and di-isopropanolamine (DIPA), tertiary N-methyl-diethanolamine (MDEA) and triethanolamine (TEA), and their blends, are the most widely considered systems for use in industrial CO₂ capture and storage [21,24,25]. With the advantages associated with high CO₂ solubility and a fast reaction rate, aqueous MEA still occupies the leading place in the market. However, MEA-based production lines have high energy demands during regeneration, lead to equipment corrosion caused by a high rate of solvent evaporation, and produce toxic compound emissions [26].

In whole or in part, these disadvantages can be overcome by an exchange of water via high-boiling point (HBP) solvents. Usually, the latter show a lower solvent heat capacity and evaporation enthalpy [27], as a result of which, when they are mixed with amines, HBP solvents contribute to significantly decreased amine evaporation and degradation, as well as equipment corrosion [26]. Moreover, it has been recently shown [28,29] that, because of their lower dielectric constant, a desorption process of MEA dissolved in HBP solvents can be facilitated due to the substantially reduced stability of the formed bicarbonate and carbamate. In addition, a significant improvement in the whole technology line can be achieved via the modification of amine CO₂ absorption (chemical method) in organic solvents via the physical methods of solvent regeneration, as shown, for example, in Ref. [30]. In principle, CO₂ absorption in HBP solutions has been carefully discussed for amines mixed with glycols [26,30,31,32,33,34,35] and glycol ethers [36], glycol/primary alcohol mixtures [27,30,34], propylene carbonate [31], N-methyl formamide [30], and N-methyl pyrrolidone (NMP) [34].

Due to their rather high liquid–gas transition temperatures, deep eutectic solvents (DESs) can be also considered as HBP solvents [37]. A diversity of possible combinations of starting materials, along with a hydrogen bond donor (HBD) and a hydrogen bond acceptor (HBA) in an appropriate molar ratio, provide powerful tools to control the properties of DESs and natural deep eutectic solvents (NADESs) [38] in accordance with their applications, including, among many others, pharmaceutical synthesis [39], band gap systems [40,41], and CO₂ capture [42,43,44,45]. Despite their absorption characteristics being inferior to conventional amine solutions, the systems in which DESs are mixed with amines or DESs comprising amines as HBDs are of particular interest when the optimization of the CO₂ absorption technology is considered. Surprisingly, a thorough literature search revealed only a few publications that provide a detailed study of the effect of the mixture of components, their ratio, the nature of amine, etc., on carbon dioxide capture by amine-based DESs and amine/DES mixtures [46,47,48,49,50,51].

In our previous work [52], we thoroughly investigated how the ratio of amine (MEA) to molecular solvent (EG) and electrolyte concentration (ChCl) in the MEA/EG/ChCl ternary system affects CO₂ absorption and such physical properties as density and viscosity. We showed that regardless of the components’ ratio and temperature, a decrease in MEA content always results in a decrease in the solutions’ absorption capacity ($m_{{\mathrm{CO}}_2}$), expressed in moles of CO₂ absorbed by 1 kilo of a solution. Interestingly, ChCl seems to play no significant role in the CO₂ absorption process undertaken by the studied system. To check whether this is only the case for the previously studied MEA/EG/ChCl ternary system, or if this applies to all similar ternary systems containing amines and molecular solvents of different natures, the following steps were carried out. First, the compositions of the investigated systems were varied by switching from primary (MEA) to secondary (diethanolamine, DEA) and tertiary (triethanolamine, TEA) amines, and by varying the degree of amine substitution from mono- (MEA, DEA and TEA) to double-substituted amines (ethylenediamine, EDA). Second, the role of the molecular solvent was investigated via an exchange of a H-bonded EG with more polar water (${\epsilon }_{20}^{\mathrm{EG}}=$ 38.9; ${\epsilon }_{20}^{{\mathrm{H}}_2\mathrm{O}}=$ 80.4) and dimethylsulfoxide (DMSO), a solvent with a dielectric constant comparable to that of EG (${\epsilon }_{20}^{\mathrm{DMSO}}=$ 46.7) but incapable of H-bonding. Finally, the MEA/EG/ChCl mixtures containing some substances recently found to act as CO₂ absorption promoters [53,54,55,56] were also studied. As a result, the thermophysical properties and CO₂ absorption characteristics of the amine/molecular solvent/ChCl (+additive) mixtures before and after complete CO₂ saturation have been carefully discussed.

2. Experimental Section

2.1. Materials

(2-Hydroxyethyl)-trimethylammonium chloride (choline chloride, ChCl; 99 wt. %) purchased from Acros Organics and 1,2-ethanediol (ethylene glycol, EG; >99 wt. %), propane-1,2,3-triol (glycerol, Gly; >99.6 wt. %), dimethyl sulfoxide (DMSO; >99.9 wt. %), and ammonium thiocyanate (NH₄SCN; >98 wt. %) supplied by Komponent Reaktiv (Moscow, Russia) were used without further purification. 2-Aminoethanol (monoethanolamine, MEA; >99.3 wt. %), bis-(2-hydroxyethyl)amine (diethanolamine, DEA; >99.2 wt. %), 1,2-diaminoethane (ethylenediamine, EDA; >99.5 wt. %), and tris-(2-hydroxyethyl)amine (triethanolamine, TEA; >99 wt. %) provided by «Oka-Sintez» Ltd. (Dzerzhinsk, Russia) were distilled under vacuum before use. The solvents were stored over freshly activated 3 Å molecular sieves in a desiccator filled with phosphorus pentoxide.

N,N,N′,N′-tetramethylguanidine (TMG; >99 wt. %), 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU; $\ge$99 wt. %) and 1,4-diazacyclohexane (piperazine, PP; >99 wt. %) were purchased from Sigma Aldrich.

Carbon dioxide (99.999 wt. %) and nitrogen (99.999 wt. %) were supplied by Monitoring (St. Petersburg, Russia); ammonia (99.99999 wt. %) was purchased from «Horst» Ltd. (Moscow, Russia). A brief overview of the substances used in this work is given in

Table 1. Compound’s chemical name, its CAS number, molar mass, supplier, initial mass fraction purity and water content, as well as the purification and drying procedures applied to the substances used in this work.

Chemical	CAS No.	Abbr.	Molar Mass a	Supplier	Mass Purity b	Initial Water Content a,c	Purification	Drying
2-aminoethanol	141-43-5	MEA	61.08	«Oka-Sintez» Ltd. (Dzerzhinsk, Russia)	>99.3	4000	Vacuum distillation	Activated 3 Å molecular sieves
bis-(2-hydroxyethyl)amine	111-42-2	DEA	105.14	«Oka-Sintez» Ltd. (Dzerzhinsk, Russia)	>99.2	4000	Vacuum distillation
1,2-diaminoethane	107-15-3	EDA	60.10	Sigma Aldrich	>99.5	5000
tris-(2-hydroxyethyl)-amine	102-71-6	TEA	149.19	«Oka-Sintez» Ltd. (Dzerzhinsk, Russia)	>99.0	2000	Vacuum distillation
water	7732-18-5	H2O	18.02	Millipore Milli-Q
dimethyl sulfoxide	67-68-5	DMSO	78.13	Komponent Reaktiv (Moscow, Russia)	>99.9	1000
N,N,N′,N′-tetramethyl-guanidine	80-70-6	TMG	115.18	Sigma Aldrich	>99.0	2000
1,8-diazabicyclo[5.4.0]-undec-7-ene	6674-22-2	DBU	152.24	Sigma Aldrich	$\ge$99.0	1500
1,4-diazacyclohexane	110-85-0	PP	86.14	Sigma Aldrich	>99.0	1300
(2-hydroxyethyl) trimethylammonium chloride	67-48-1	ChCl	139.62	Acros Organics	99	6000	Recrystallization
1,2-ethanediol	107-21-1	EG	62.07	Komponent Reaktiv (Moscow, Russia)	>99.6	700		Activated 3 Å molecular sieves
ammonium thiocyanate	1762-95-4	NH4SCN	76.12	Komponent Reaktiv (Moscow, Russia)	99	260	Recrystallization	Drying under reduced pressure
glycerol	56-81-5	Gly	92.09	Komponent Reaktiv (Moscow, Russia)	>99.6	600		Activated 3 Å molecular sieves
nitrogen	7727-37-9	N2	28.01	Monitoring (St. Petersburg, Russia)	99.9999
carbon dioxide	124-38-9	CO2	44.01	Monitoring (St. Petersburg, Russia)	99.9999
ammonia	7664-41-7	NH3	17.03	Ltd. «Horst»(Moscow, Russia)	99.99999

^a Units are g mol⁻¹ for molar mass and ppm for water content. ^b Provided by a supplier. ^c Mass fraction ($w_{{\mathrm{H}}_2\mathrm{O}}=m_{{\mathrm{H}}_2\mathrm{O}}/m_{\mathrm{sample}}$), wt. %.

.

2.2. Mixture Preparation

In accordance with the findings obtained by us in Ref. [52] and briefly discussed in the Introduction, two series of ternary amine/molecular solvent/ChCl mixtures were prepared. The first series (I) comprised blends composed of amine (component 1), molecular solvent (component 2) and ChCl (component 3) with mole fractions ($x_i$)

$${\mathit{x}}_{\mathit{i}}=\frac{{\mathit{m}}_{\mathit{i}}/{\mathit{M}}_{\mathit{i}}}{\sum {\mathit{m}}_{\mathit{i}}/{\mathit{M}}_{\mathit{i}}}$$

(1)

of x₁ = 0.667, x₂ = 0.222, and x₃ = 0.111, respectively. In Equation (1), $m_i$ refers to the masses and $M_i$ the molar masses of the mixture components. For the studied MEA/EG/ChCl system [52], this composition showed an optimal combination of absorbing capacity ($m_{{\mathrm{CO}}_2}$, moles of CO₂ absorbed by 1 kg of a solution) and amine efficiency ($c_{{\mathrm{CO}}_2}$, moles of CO₂ absorbed by 1 mole of amine). Due to this, the effect of the amine structure (primary, secondary, or tertiary amine), the number of its NH₂ groups, and the influence of the molecular solvent substitution (see

Table 2. Abbreviation and composition of the mixtures composed of amine (1), molecular solvent (2) and ChCl (3) with mole fractions of x₁ = 0.667, x₂ = 0.222, and x₃ = 0.111, respectively, (Series I) and with x₁ = 0.222, x₂ = 0.667, and x₃ = 0.111 with the presence of 5 wt. % of additives (Series II). Shown as well are amine and additive structures.

Series I: x1 = 0.667, x2 = 0.222, and x3 = 0.111
Mixture abbreviation	Amine	Amine structure		Molecular solvent
I-MEA/EG	MEA			EG
I-DEA/EG	DEA			EG
I-EDA/EG	EDA			EG
I-TEA/EG	TEA			EG
I-MEA/DMSO	MEA			DMSO
I-MEA/H2O	MEA			H2O
Series II: x1 = 0.222, x2 = 0.667, and x3 = 0.111
Mixture abbreviation	Amine	Molecular solvent	Additive name	Additive structure
II-MEA/EG	MEA	EG
II-MEA/EG/TMG	MEA	EG	TMG
II-MEA/EG/H2O	MEA	EG	H2O
II-MEA/EG/DBU	MEA	EG	DBU
II-MEA/EG/PP	MEA	EG	PP

for details) were investigated for mixtures with the above composition. Again, as was shown in [52], the best MEA efficiency ($c_{{\mathrm{CO}}_2}$) for the MEA/EG/ChCl system was demonstrated in the mixture with x₁ = 0.222, x₂ = 0.667, and x₃ = 0.111. Thus, this mole fraction ratio (hereinafter called Series II) was used to figure out how the presence of additives (see Table 2) influences the $c_{{\mathrm{CO}}_2}$ values.

Accordingly, ternary mixtures of both compositions were prepared by weighing appropriate amounts of neat substances on an analytical balance with an accuracy of 0.1 mg. The mixtures were placed in tightly closed vials and, after complete mixing of the components, were stored over freshly activated 3 Å molecular sieves in a desiccator with P₂O₅ at room temperature. For the mixtures of Series II, the additives with mass fractions of 5% were added directly into the mixtures with Series II composition ($x_1=0.222$, $x_2=0.667$, $x_3=0.111$).

The composition of the mixtures investigated in the present work and their abbreviations, together with the structures of amines and additives comprising the systems of Series I and Series II, are given in Table 2. In order to simplify the interpretation of the used abbreviations, the latter consist of the series number (I for $x_1=0.667$ and II for $x_1=0.222$), the name of the amine and molecular solvent the mixtures are composed of for (Series I) and, additionally, the name of the additive (for Series II). For example, mixture I-MEA/EG comprises the MEA/EG/ChCl system with $x_{1=\mathrm{MEA}}=0.667$, $x_{2=\mathrm{EG}}=0.222$ and $x_{3=\mathrm{ChCl}}=0.111$, and mixture II-MEA/EG/TMG is the MEA/EG/ChCl system with $x_{1=\mathrm{MEA}}=0.222$, $x_{2=\mathrm{EG}}=0.667$ and $x_{3=\mathrm{ChCl}}=0.111$ containing 5 wt. % of TMG. Since all the prepared mixtures contain the same amount of ChCl ($x_3=0.111$), ChCl is omitted from all used abbreviations for the sake of simplicity.

To minimize the contact of the samples with air, preparatory operations were carried out in a flow of dry nitrogen. Additionally, the density (ρ), viscosity (η) and CO₂ absorption measuring cells were purged with N₂ prior to measurements with the sample transfer carried out using the syringe technique.

Water content, $w_{{\mathrm{H}}_2\mathrm{O}}=m_{{\mathrm{H}}_2\mathrm{O}}/m_{\mathrm{sample}}$, expressed in parts per million (ppm), was estimated using the Karl Fischer titration (Metrohm 831 AG). The data on $w_{{\mathrm{H}}_2\mathrm{O}}$ for all the investigated samples are listed in Table S1.

2.3. Viscosity and Density Measurements

The Anton Paar SVM3001 instrument equipped with a Stabinger cell followed by a vibrating tube was used for dynamic viscosity (η) and density (ρ) measurements, carried out at atmospheric pressure. The instrument is characterized by a density repeatability of 0.00005 g·cm⁻³, a viscosity repeatability of 0.1%, and temperature stability of 0.005 K. Both the Stabinger viscosity and density measuring cells were calibrated using APS3, APN7.5, APN26, and APN415 calibration oils (Anton Paar, Austria) in accordance with the supplier calibration protocol. For density, a nominal measurement error of u(ρ) = 0.05 g·cm⁻³ was estimated. Accordingly, the data collected in Tables S3 and S4 are given with the same number of digits. However, due to the limited purity of the samples [57], the expanded related uncertainty was found to be U_r(ρ) = 0.001 (confidence level 0.95). For viscosity, an expanded related uncertainty of U_r(η) = 0.02 (confidence level 0.95) was estimated.

The density and viscosity measurements were validated by comparing experimental data ($Y^{\exp }$) and the fit ($Y^{\mathrm{fit}}$) for APN7.5, APN26 and APN415 viscosity standards in terms of their deviations (δ)

$$\mathbf{\delta }=\frac{{\mathit{Y}}^{exp}-{\mathit{Y}}^{fit}}{{\mathit{Y}}^{fit}}$$

(2)

In the form of $Y^{\mathrm{fit}}$ values, the data provided by the supplier for the batches of oil, fitted with Equation (2) for density and Equation (3) for viscosity, were used for comparison. As seen in Figure S1, the deviations never exceeded the −0.075 ≤ 100δ ≤ +0.08 interval for density and −1.9 ≤ 100δ ≤ +1.2 for viscosity; that is, they were satisfactorily below the calculated expanded related uncertainties of U_r(ρ) = 0.001 and U_r(η) = 0.02, respectively.

For the obtained samples, density and viscosity as a function of temperature were recorded within the temperature range T = (278.15–363.15) K with a step of 5 K. For the samples saturated with CO₂ at T = 318.15 K (see below), the measurements were carried out at atmospheric pressure and temperatures varying from 288.15 to 318.15 K. At higher temperatures, the measurements could not be performed due to partial gas release. In addition, for a number of samples, their viscosity T < 293.15 K was too high, which hampered ρ and η measurements at lower temperatures. We would here like to note that the ρ(T) and η(T) obtained for the CO₂-saturated samples (Tables S4 and S6, respectively) have been designed to show the general trend of the property change during the CO₂ absorption process, rather than to discuss and interpret their absolute values.

2.4. CO₂ Absorption Capacity

The CO₂ absorption capacity at T = 313.15 K was estimated as follows. A certain mass of solution (app.10.0 g) was placed into a glass vial with an inner diameter of 12 mm and length of 200 mm, equipped with a cap with a long tube reaching the bottom of the vial (CO₂-input) and a CO₂-output tube with an electronic pressure sensor operating with an accuracy of 100 Pa. CO₂ was bubbled through the sample at a rate of 20 cm³∙min⁻¹ at p = 0.1 MPa (monitored by an electronic pressure sensor). During the measurements, the vial was thermostated in a T-3.1 water bath equipped with a precision temperature controller MIT-8.15 “Iztech” (Russia). The actual temperature was determined with a PTSV-4.2 Kriotherm thermocouple (Russia), u(T) = 0.02 K. During the measurements, the sample was mixed with a magnetic stirring bar. The mass gain during absorption was measured at regular intervals using a Shimadzu AUW220D electronic balance without buoyancy correction. Each solution was examined three times, and the average value of mass gain was taken for calculation. In order to take into account the weight loss associated with solvent evaporation, blank experiments consisting of bubbling of dry nitrogen through a sample were also carried out. The mixture absorption capacity ($m_{{\mathrm{CO}}_2}$) and amine efficiency ($c_{{\mathrm{CO}}_2}$) defined as the moles of CO₂ absorbed by 1 kg of a solution and by 1 mole of amine, respectively, were calculated from the maximum value of the mass gain. The mixtures obtained as such were further used for estimation of their density and viscosity.

In general, the reproducibility of CO₂ absorption measurements was within 0.5%. However, when the errors associated with the weighing, temperature, and mole fraction of the mixture components were taken into account, the associated uncertainty was found to be $U_{\mathrm{r}}\left(m_{{\mathrm{CO}}_2}\right)=$ 0.05 for $m_{{\mathrm{CO}}_2}$ and $U_{\mathrm{r}}\left(c_{{\mathrm{CO}}_2}\right)=$ 0.05 for $c_{{\mathrm{CO}}_2}$.

The gas absorption measurements at p = 0.1 MPa were validated for 15 and 30 wt. % MEA solutions absorbing CO₂ at 298.15 and 313.15 K, and for a DES composed of NH₄SCN and glycerol in a molar ratio 2:3, absorbing NH₃ at 313.15 K. The reference values of their absorption capacity were taken from Refs. [58,59,60,61,62,63]. As seen in Table S2, the results obtained in the present work are in good agreement with the data derived by Huertas et al. [58], Deng et al. [59] and Li et al. [60], obtained gravimetrially, and in reasonable accordance with the volumentric measuremetns taken by Lee et al. [61], Jou et al. [62], and Shen and co-workers [63].

3. Results and Discussion

From a practical point of view, density and viscosity are fundamental properties whose temperature dependence must be considered when operating with fluids in any practical application. As mixtures containing amines (substances capable of capturing carbon dioxide) mixed with DES-like solvent are expected to have broad application in CCS technology, their physical properties are of particular interest for future use.

3.1. Density

The ρ(T) data for the mixtures in Series I and II are listed in Table S3; Figure 1 displays the experimentally obtained ρ(T) data (symbols) and their linear correlations (lines)

$$\mathbf{\rho }={\mathit{a}}_{\mathbf{0}}+{\mathit{a}}_{\mathbf{1}}\mathit{T}$$

(3)

where $a_0$ and $a_1$ are the fitting coefficients. The latter together with their standard deviations are listed in

Table 3. Coefficients ${\mathrm{a}}_0$ and ${\mathrm{a}}_1$ and the associated fit standard error (σ_fit) obtained by fitting experimental ρ(T) data with Equation (3).

System	a0 (g·cm−3)	104 a1 (g·cm−3·K−1)	104 σfit (g·cm−3)	r2
Series I: x1 = 0.667, x2 = 0.222, and x3 = 0.111
I-MEA/EG	1.2707 ± 0.0003	−6.995 ± 0.008	0.65	0.99994
I-DEA/EG	1.292 ± 0.002	−6.34 ± 0.05	1.32	0.99973
I-EDA/EG	1.2434 ± 0.0005	−8.04 ± 0.02	1.57	0.99994
I-TEA/EG	1.2892 ± 0.0009	−7.68 ± 0.03	2.56	0.99982
I-MEA/DMSO	1.2594 ± 0.0003	−7.03 ± 0.01	0.78	0.99998
I-MEA/H2O	1.2944 ± 0.0009	−5.74 ± 0.03	2.59	0.99967
Series II: x1 = 0.222, x2 = 0.667, and x3 = 0.111 + 5 wt. % of additive
II-MEA/EG	1.2927 ± 0.0002	−6.524 ± 0.007	0.57	0.99994
II-MEA/EG/TMG	1.291 ± 0.001	−6.63 ± 0.04	3.55	0.99953
II-MEA/EG/H2O	1.2903 ± 0.0002	−6.549 ± 0.007	0.68	0.99998
II-MEA/EG/DBU	1.2966 ± 0.0005	−6.73 ± 0.01	1.35	0.99993
II-MEA/EG/PP	1.2931 ± 0.0007	−6.65 ± 0.02	2.15	0.99983

. The quality of the obtained experimental data was analyzed in terms of their deviations ($\delta$, see Equation (2)) from the fit found with Equation (3). It has been shown that the $\delta$ values lie within the $-0.05\le 100\delta \le +0.04$ range for mixtures of Series I and $-0.05\le 100\delta \le +0.06$ for Series II. The fact that these deviations are substantially below the expanded related uncertainty of U_r(ρ) ≈ 0.001 confirms the good quality of the obtained density data.

As seen in Figure 1a, the exchange of amine and molecular solvent components significantly affects the density of Series I mixtures. For systems with varying component 1, this effect is in conformity with the density of neat amines. Its increase observed in the order EDA (${\rho }_{\mathrm{EDA}}^{303.15}=$ 1.25640 g·cm⁻³) < MEA (${\rho }_{\mathrm{MEA}}^{303.15}=$ 1.00809 g·cm⁻³) < DEA (${\rho }_{\mathrm{DEA}}^{303.15}=$ 1.08813 g·cm⁻³) < TEA (${\rho }_{\mathrm{TEA}}^{303.15}=$ 1.11847 g·cm⁻³) is reflected in the density of their ternary mixtures, that is, I-EDA/EG < I-MEA/EG < I-DEA/EG < I-TEA/EG. As was shown in Ref. [52], ChCl has a structure-breaking effect on the mixed MEA/EG solvent, in which packing effects dominate over H-bonding interactions. The above order of the systems’ density allows one to suggest that in ternary mixtures comprising amine, EG and ChCl, the nature of the interspecies interactions is very similar to that of those in the previously studied MEA/EG/ChCl system.

When the molecular solvent (EG) is exchanged for H₂O or DMSO, a decrease in density is observed. For the mixture containing water as a molecular solvent, the density increases from ${\rho }_{{\mathrm{H}}_2\mathrm{O}}^{298.15}=$ 0.99704 g·cm^–3 for pure water to ${\rho }_{\mathrm{I}-\mathrm{MEA}/{\mathrm{H}}_2\mathrm{O}}^{298.15}=$ 1.04982 g·cm^–3 for I-MEA/H₂O. A literature search shows that this behavior is typical for binary mixtures composed of ChCl and MEA (or H₂O). For example, aqueous solutions of both ChCl [64] and MEA [65] within an entire solubility range have densities higher than that of pure water. When dissolved in MEA, ChCl also increases the density of the obtained solutions, as shown in Ref. [66]. Thus, the MEA/ChCl/H₂O mixture displays an intuitively expected trend of density increase. However, one cannot say that each mixture component has a positive impact on the observed property change, as density only reflects an overall effect governed by the ion–ion, ion–molecule and molecule–molecule interactions present in a ternary system comprising an electrolyte (ChCl). As a result, in order to figure out the individual contribution of each of the mixture’s components and/or the above interactions, a more detailed investigation, including a property scan by component content at a constant ratio of the other two components, should be performed.

Although we did not find any support in the literature, it is very likely that, similar to many electrolyte systems, the addition of ChCl to DMSO leads to a density increase. Apparently, an assumed positive effect of ChCl on the DMSO density and the confirmed positive impact on MEA density [66] are compensated by a negative MEA-DMSO contribution in the mixture I-MEA/DMSO. Its density, equal to ${\rho }_{\mathrm{I}-\mathrm{MEA}/\mathrm{DMSO}}^{298.15}$ = 1.06345 g·cm^–3, is lower than of pure solvent, ${\rho }_{\mathrm{DMSO}}^{298.15}=$ 1.09519 g·cm^–3 [67]. As shown in Ref. [68], an increase in MEA content in the MEA/DMSO system leads to a decrease in its density, which is mainly caused by packing effects (excess molar volume increases with increasing temperature). Therefore, the density of the mixture I-MEA/DMSO is affected by the strength and flexibility of the network formed in a mixed MEA/DMSO solvent. For Series II mixtures (see Figure 1b), no significant density changes are observed. In the presence of all the additives used, the ρ slightly decreases compared to the II-MEA/EG system.

The density values of the CO₂-saturated samples collected in Table S4 are always higher than those for initial mixtures. Its increase is caused by the formation of products of the reaction between amine and carbon dioxide. When discussing Series I mixtures, the density of the CO₂-saturated samples is in very good agreement with the amount of the absorbed gas (see below), that is, it increases more when going from I-TEA/EG to I-MEA/H₂O than when going to I-MEA/DMSO ≈ I-MEA/EG and I-DEA/EG, and, finally, to I-EDA/EG. As seen in Figure 1a, for the density of CO₂-saturated samples of Series I, the sequence is I-TEA/EG < I-MEA/EG < I-DEA/EG ≈ I-MEA/DMSO < I-MEA/H₂O. We here note that during the course of CO₂ absorption by the I-EDA/EG mixture, a precipitate is formed and, as a result, the mixture completely solidifies. Due to this, no density was measured for it.

Apparently, the situation is very similar when Series II mixtures studied after complete CO₂ saturation are considered. Here, among the CO₂-saturated samples, the highest density is observed for II-MEA/EG/PP, that is, again, for the sample showing the highest absorption capacity, $m_{{\mathrm{CO}}_2}$. However, since for the mixtures containing other additives (TMG, DBU, and H₂O) their ${\rho }_{{\mathrm{CO}}_2}$ values are very close to that of the neat II-MEA/EG system, their density values appear to be very close to each other.

3.2. Viscosity

The logarithmic viscosities (η) of neat and CO₂-saturated mixtures of Series I and II estimated within the temperature range of T = (283.15 to 353.15) K are shown in Figure 2a,b; the corresponding viscosity values are summarized in Tables S5 and S6, respectively. Experimental ln η values for Series I and Series II as a function of temperature (T) were fitted using the Vogel-Fulcher-Tammann (VFT) equation

$$ln\frac{\mathit{\eta }}{{\mathit{\eta }}_0}=ln\left(\frac{{\mathit{A}}_{\mathit{\eta }}}{{\mathit{\eta }}_0}\right)+\frac{{\mathit{B}}_{\mathit{\eta }}}{\mathit{T}-{\mathit{T}}_{0,\mathit{\eta }}}$$

(4)

In Equation (3), ${\eta }_0=1$ mPa∙s, $A_{\eta }$ is the viscosity limit at an infinite temperature, $B_{\eta }$ is the coefficient linked by relation with the pseudo activation energy, and $T_{0,\eta }$ is the VFT temperature. The parameters of Equation (4) are summarized in

Table 4. Coefficients $A_{\eta }$, $B_{\eta }$, and $T_{0,\eta }$; the associated fit standard error (σ_fit) obtained by fitting experimental η(T) data with Equation (4).

x2	$\mathit{l}\mathit{n} {\mathit{A}}_{\mathit{\eta }}/{\mathit{\eta }}_0$	${\mathit{B}}_{\mathit{\eta }}\left(\mathbf{K}\right)$	${\mathit{T}}_{0,\mathit{\eta }}\left(\mathbf{K}\right)$	${10}^3{\mathit{\sigma }}_{\mathrm{fit}}$	r2
Series I: x1 = 0.667, x2 = 0.222, and x3 = 0.111 + 5 wt. %
I-MEA/EG	−3.53 ± 0.05	1012 ± 17	159 ± 1	2.96	0.99989
I-DEA/EG	−4.9 ± 0.1	1590 ±36	146± 2	6.77	0.99998
I-EDA/EG	−3.43 ± 0.07	839 ± 21	164 ± 2	5.55	0.99996
I-TEA/EG	−2.99 ± 0.06	775 ± 18	167 ± 2	5.05	0.99997
I-MEA/DMSO	−3.69 ± 0.07	1000 ± 21	160 ± 2	3.88	0.99998
I-MEA/H2O	−3.58 ± 0.04	1249 ± 13	162 ± 1	3.21	0.99999
Series II: x1 = 0.222, x2 = 0.667, and x3 = 0.111 + 5 wt. % of additive
II-MEA/EG	−3.40 ± 0.08	1057 ± 30	142 ± 3	3.87	0.99997
II-MEA/EG/TMG	−3.41 ± 0.09	1026 ± 31	149 ± 2	6.03	0.99996
II-MEA/EG/H2O	−3.64 ± 0.08	1078 ± 28	139 ± 2	4.64	0.99997
II-MEA/EG/DBU	−3.60 ± 0.04	1105 ± 14	144 ± 1	2.49	0.99999
II-MEA/EG/PP	−3.83 ± 0.05	1183 ± 19	139 ± 1	3.08	0.99999

.

For all of the experimental points, the deviation from the fit does not exceed the limits of $-0.95\le 100\delta \le +0.76$ for Series I systems, and lies within the interval of $-0.47\le 100\delta \le +0.49$ for Series II mixtures. Thus, these deviations are lower than the expanded related uncertainty estimated for viscosity, U_r(η) = 0.02.

Similar to density, the viscosity of the Series I systems containing EG as a molecular solvent correlate well with the viscosities of the neat amines used for the mixture preparation. For amines, the viscosity increases in the pattern EDA (${\eta }_{\mathrm{EDA}}^{303.15}=$ 1.26 mPa·s) < MEA (${\eta }_{\mathrm{MEA}}^{30315}=$ 14.9 mPa·s) < DEA (${\eta }_{\mathrm{DEA}}^{303.15}=$ 389 mPa·s) ≈ TEA (${\eta }_{\mathrm{TEA}}^{303.15}=$ 405 mPa·s). As seen in Figure 2a, within the entire temperature range, I-DEA/EG and I-TEA/EG have the highest viscosity in comparison with the reference I-MEA/EG system, while I-EDA/EG exhibits a viscosity lower than I-MEA/EG. Similarly, molecular solvent substitution has a very pronounced effect on the systems’ viscosity. For both I-MEA/H₂O and I-MEA/DMSO, a viscosity increase is observed compared to pure molecular solvents, i.e., from 0.890 mPa×s for H₂O to 277 mPa·s for I-MEA/H₂O, and from 2.06 mPa·s for DMSO to 34.1 mPa·s for I-MEA/DMSO at 298.15 K. Obviously, for both systems, their viscosity is higher than that of MEA, ${\eta }_{\mathrm{MEA}}^{298.15}=$ 18.7 mPa·s. This indicates that the intensity and/or number of interspecies interactions (presumably H-bonds) in the studied ternary mixtures are stronger than the MEA–MEA, H₂O–H₂O and DMSO–DMSO interactions in pure solvents. In addition, a possible positive contribution of the ChCl molecular solvent (alternatively, the Ch⁺/Cl⁻-molecular solvent) interaction to the systems’ viscosity cannot be ruled out. However, without the additional investigation of the H-bonds present in the above systems using spectroscopic techniques able to probe the discussed interactions, the above should be taken with a grain of salt.

Obviously, due to facilitated mass transport, systems exhibiting low viscosity are of particular interest when gas absorption processes are considered. Thus, an exchange of EG (${\eta }_{\mathrm{I}-\mathrm{MEA}/\mathrm{EG}}^{298.15}=$ 42.3 mPa·s) or water $({\eta }_{\mathrm{I}-\mathrm{MEA}/{\mathrm{H}}_2\mathrm{O}}^{298.15}=$ 42.3 mPa·s) in their absorbing mixtures for DMSO seems to be a solution to decreasing the system’s viscosity. However, as clearly seen in Figure 2a, a saturation of the corresponding sample by CO₂ leads to a significant viscosity increase, which definitely decreases the rate of the absorption process (see discussions below). As seen in Figure 2a and Table S5, the system I-EDA/EG has the lowest viscosity of ${\eta }_{\mathrm{I}-\mathrm{EDA}/\mathrm{EG}}^{298.15}=$ 16.8 mPa·s among the systems of Series I, which also makes it quite a promising mixture for efficient CO₂ absorption. Unfortunately, again, it can be hardly considered for this purpose as, during the course of the gas purge, a precipitate is formed, which results in the complete solidification of the sample.

The addition of TMG, DBU and PP has rather slight positive effects (viscosity increase) on II-MEA/EG; see Figure 2b. For water, the more pronounced negative effect can be explained by H₂O’s small molar mass and, as a result, higher concentration when a mass fraction of 5% is converted to a mole fraction, which means than the relation of H₂O to other mixture constituents is significantly greater than in II-MEA/EG/PP, II-MEA/EG/DBU and II-MEA/EG/TMG. Although the viscosity of Series II mixtures is rather hard to interpret, for the corresponding CO₂-saturated samples, η(T) is in full agreement with the amount of absorbed CO₂. As seen in Figure 2b, viscosity decreases in the order II-MEA/EG/PP > II-MEA/EG/DBU ≈ II-MEA/EG/TMG > II-MEA/EG > II-MEA/EG/H₂O, that is, in the exact same order as the absorption capacities of the Series II mixtures (see Table S7).

3.3. CO₂ Absorption

The potential of the chosen systems for use as CO₂ absorption agents was analyzed in terms of their absorbing capacities ($m_{{\mathrm{CO}}_2}$) and amine efficiencies ($c_{{\mathrm{CO}}_2}$), estimated from the absorption profiles shown in Figure 3a and Figure 4a. As seen, the overall CO₂ absorption of the EG-based mixtures decreases from I-MEA/EG to I-DEA/EG and I-TEA/EG, demonstrating a decrease in the CO₂ absorption when going from primary to secondary and tertiary amines, respectively. This tendency is quite common, and reflected in the literature, such as in Refs. [69,70,71]. For the above systems, $m_{{\mathrm{CO}}_2}$ achieves approximately 95% of its maximum in 30 min. The same level of saturation (but reached faster) is observed for I-MEA/DMSO and I-MEA/H₂O, i.e., the systems in which EG is exchanged for DMSO and H₂O, respectively. The rise in the absorption rate, shown in Figure 3b for systems of Series I, for I-MEA/DMSO and I-EDA/EG is definitely caused by their lower viscosity compared to I-MEA/EG (see discussions above).

While the $m_{{\mathrm{CO}}_2}$ for the I-MEA/DMSO system is close to that of I-MEA/EG, I-MEA/H₂O demonstrates a 21% higher CO₂ absorption capacity. Obviously, from the point of view of the mass/moles of the absorbed gas, the system I-MEA/H₂O is the most promising within Series I. However, as discussed in the literature [23], H₂O-based amine solutions are rather volatile. When volatility is considered as an important parameter for the choice of the absorbing solution, the I-MEA/EG system is more attractive, despite showing a 17% lower absorption capacity compared to I-MEA/H₂O. Interestingly, the amine efficiency ($c_{{\mathrm{CO}}_2}$) (Figure 5a) values of I-MEA/EG, I-DEA/EG, and I-MEA/H₂O are the same, within the error of estimation. For I-MEA/DMSO, the $c_{{\mathrm{CO}}_2}$ is approximately 10% higher, which indicates a favorable contribution of DMSO to amine efficiency. The EDA-based system (I-EDA/EG) shows $m_{{\mathrm{CO}}_2}$ and $c_{{\mathrm{CO}}_2}$ values that are almost 40% higher in comparison with I-MEA/EG. Indeed, this finding is in good agreement with the data from the literature [72,73].The absorption rate for I-EDA/EG reaches 95% of its maximum in only 10 min. Certainly, these characteristics make the I-EDA/EG mixture the most promising among the studied systems for efficient CO₂ capture. However, in the CO₂-saturated I-EDA/EG sample, a precipitate is formed, which results in full solution solidification. Thus, despite the good absorption properties of the EG-based mixture containing EDA, its practical applicability is quite doubtful.

Figure 4a,b display absorption profiles and absorption rates ($\partial m_{{\mathrm{CO}}_2}/\partial \tau$), respectively, for Series II systems, i.e., the neat mixture MEA/EG/ChCl with $x_1=0.222$, $x_2=0.667$, and $x_3=0.111$ (II-MEA/EG), and mixture II-MEA/EG containing 5 wt. % of promoters. Figure 5 shows the corresponding CO₂ absorption capacities ($m_{{\mathrm{CO}}_2}$) and amine efficiency ($c_{{\mathrm{CO}}_2}$).

As seen in Figure 4a,b, all the studied systems of Series II achieved 95% of their maximum absorption capacity in approximately 30 min. In principle, the absorption profiles can be conditionally divided into three time intervals. Within the first interval (<3 min), the highest CO₂ absorption rate is observed for the system containing water, II-MEA/EG/H₂O. Obviously, this is a viscosity-induced absorption rate increase, as the II-MEA/EG/H₂O mixture has the lowest viscosity in Series II. In approximately 3 min, the $\partial m_{{\mathrm{CO}}_2}/\partial \tau$ of II-MEA/EG/H₂O drops significantly, and in the intermediate time interval (from 3 to 20 min), the rest of the Series II mixtures show higher $\partial m_{{\mathrm{CO}}_2}/\partial \tau$ values than II-MEA/EG/H₂O. On average, within the second time interval, the neat mixture of MEA/EG/ChCl (system II-MEA/EG) shows the highest CO₂ absorption rate. The third interval (>20 min) is characterized by rather low $\partial m_{{\mathrm{CO}}_2}/\partial \tau$ values, which, within the errors of estimation, are the same for all mixtures of Series II. The comparison of the results shown in Figure 4a,b enables us to conclude that, in general, the CO₂ absorption rate has no effect on the overall absorption capacity. As seen in Figure 5a,b, in the presence of the promoters, the absorption capacities increase in the order II-MEA/EG/PP > II-MEA/EG/DBU ≈ II-MEA/EG/TMG > II-MEA/EG > II-MEA/EG/H₂O. However, on the whole, the positive contributions of TMG, DBU and PP to the CO₂ absorption capacity of II-MEA/EG are 4%, 9% and 30%, respectively, and the MEA efficiency increases by 30% for the latter system. Interestingly, when considering the overall process, H₂O has a negative effect on the CO₂ absorption of the MEA/EG/ChCl mixture. Clearly, the above order cannot be explained by the amounts of promoters added, as even when converted into mole fraction units, their concentrations decrease in the order H₂O > PP > TMG > DBU. Therefore, the most likely cause of this order is a difference in the mechanism of the reaction active in the chemical absorption of carbon dioxide by MEA, as promoted by the additives. Thus, in summing up the obtained results, it should be noted that from the point of view of the practical application of the investigated systems, PP represents the most promising promoter of CO₂ absorption, with only minor effects on the physical properties of the MEA/EG/ChCl mixtures.

4. Conclusions

In the present contribution, we outline the densities (ρ), viscosities (η) and absorption characteristics related to carbon dioxide for systems comprising amine (1), molecular solvent (MS, 2) and choline chloride (ChCl, 3), within the temperature range of T = 278.15–363.15 K for physical properties, and at 313.15 K for CO₂ absorption capacity. The first part of the research includes an investigation of the effect of the mixture’s composition, revealed by varying the amine from primary (monoethanolamine, MEA) to secondary (diethanolamine, DEA) and tertiary (triethanolamine, TEA), and by varying the degree of its substitution from mono- (MEA, DEA and TEA) to double-substituted (ethylenediamine, EDA) compounds. We also studied the role of an MS by exchanging ethylene glycol (EG) for water and dimethylsulfoxide (DMSO). As a base, a system with mole fractions of $x_1=0.667$, $x_2=0.222$, and $x_3=0.111$ was chosen. As was shown earlier [52], for this mixture composition, the CO₂ absorption parameters found for the MEA/EG/ChCl system show an optimal combination of absorbing capacity, $m_{{\mathrm{CO}}_2}$, expressed in moles of CO₂ absorbed by 1 kg of a solution, and MEA efficiency, $c_{{\mathrm{CO}}_2}$, moles of CO₂ absorbed by 1 mole of amine.

We show that when varying the amine in the amine/EG/ChCl mixtures, the ρ and η changes are determined by the properties of neat amines. This suggests that the natures of the interspecies interactions in these mixtures are very similar to that of the previously studied MEA/EG/ChCl system, namely, ChCl has a structure-breaking effect on the mixed amine/EG solvent, in which packing effects dominate over H-bonding interactions. When EG is exchanged for H₂O or DMSO, a decrease in density and an increase in viscosity are observed. A comparison of the data with the corresponding properties of pure MSs and the systems composed of each pair of the mixtures’ components indicates that the intensity and/or number of interspecies interactions in the present ternary mixtures are stronger than those in pure H₂O, DMSO and MEA. For the corresponding CO₂-saturated samples, the ρ and η increase proportionally to the amount of gas absorbed. The CO₂ absorption capacity of the EG-based mixtures decreases when going from primary to secondary and tertiary amines, and is 21% higher for the MEA/H₂O/ChCl system compared to the mixtures containing EG and DMSO, which, in turn, show similar $m_{{\mathrm{CO}}_2}$ values.

The second part comprised an investigation of the presence of CO₂ absorption promoters in 5 wt. % of the MEA/EG/ChCl system, with mole fractions of $x_1=0.222$, $x_2=0.667$, and $x_3=0.111$, with which the best MEA efficiency was obtained [52]. As the promoters, N,N,N′,N′-tetramethylguanidine, 1,8-diazabicyclo[5.4.0]undec-7-ene, 1,4-diazacyclohexane and water were used. We show that the addition of the promoters did not greatly affect the physical properties of the neat MEA/EG/ChCl mixture, while for the CO₂-saturated samples, again, the higher the absorption capacity, the greater the increase in ρ and η. As shown, the greatest increase was observed for the piperazine-containing system, which was found to exhibit the highest CO₂ absorption capacity.