Experimental Research on NO2 Viscosity and Absorption for (1-Ethyl-3-methylimidazolium Trifluoroacetate + Triethanolamine) Binary Mixtures

The viscosity (9.34–405.92 mPa·s) and absorption capacity (0.4394–1.0562 g·g−1) of (1-ethyl-3-methylidazolium trifluoroacetate + triethanolamine) binary blends atmospheric pressure in the temperature range of 303.15–343.15 K and at different mole fractions of [EMIM] [TFA] have been carried out. The molar fraction of [EMIM] [TFA] dependence of the viscosity and absorption capacity was demonstrated. The addition of a small amount of [EMIM] [TFA] into TEA led to rapidly decreased rates of binary blends’ viscosity and absorption capacity. However, the viscosity and absorption of binary blends did not decrease significantly when [EMIM] [TFA] was increased to a specific value. Compared with the molar fraction of the solution, the temperature had no obvious effect on viscosity and absorption capacity. By modeling and optimizing the ratio of viscosity and absorption capacity of ([EMIM] [TFA] + TEA), it is proven that when the mole fraction of [EMIM] [TFA] is 0.58, ([EMIM] [TFA] + TEA) has the best viscosity and absorption capacity at the same time. In addition, at 303.15 K, ([EMIM] [TFA] + TEA) was absorbed and desorbed six times, the absorption slightly decreased, and the desorption increased.


Introduction
Nitrogen oxides (NO X ) are the prominent pollutants in the atmosphere. They are mainly produced by the high-temperature combustion of industrial nitrogenous fuel in which NO X emission accounts for 50% of total emission when burning petroleum and over 60% (about 7.7 million tons) when burning coal in power plant boilers [1][2][3]. NO X are the main precursors of acid rain and the direct sources of indoor nitrous acid (HONO) and photochemical air pollution. They greatly increase the risk of lung damage or even lung cancer if exposed to these substances for a long time [4][5][6]. At present, SCR (selective catalytic reduction), SNCR (selective non-catalytic reduction), and NSCR (nonselective catalytic reduction) are the principal force of denitrification techniques. However, these technologies are often accompanied by high technology costs and complex tail gas pollution. In contrast, inexpensive solvents are becoming more attractive. The absorbents (water, acid solution, alkali solution) which target nitrogen dioxide (NO 2 ) account for the majority of these solvents [7,8]. As a mature and efficient separation system for industrial desulfurization and denitrification, alkanolamines (such as monoethanolamine, diethanolamine, methyldiethanolamine, and triethanolamine) can rapidly transfer acid gas from gas phase to liquid phase and have significant advantages of high absorption and low operation costs [9][10][11]. The mixture of alkanolamines and water is a common form of industrial solvents. The cavities of the water provide larger activity spaces for the molecules of alkanolamines and the two-phase hydrogen bonds significantly prevent the self-association between the molecules [12]. This behavior effectively reduces the viscosities of pure alkanolamines, strengthens the mass transfer, and increases the absorption capacity of gases. However, the degradation and volatilization of alkanolamine solutions-as well as the scaling and corrosion of equipment carried by the water of the compound system-are still unavoidable in the industrial process [13]. The crucial problem is that NO 2 reacts with alkanolamine solutions based on water easily and rapidly which makes the solutions become viscous and foam, at the same time causing the system to generate higher regeneration energy, which challenges the practicability and circulation of the absorption solutions in large absorption equipment [14,15]. Some researchers have carried out a lot of work on the absorption of acid gas by nonaqueous amine solutions [16][17][18]. For example, Kim et al. [17] used alcohol solvent as a substitute for water. Compared with water or short-chain alcohol, the use of (long-chain alcohol + alkanolamines) solution reduces the wastage of high volumes of solvents, but the insufficient absorption capacity in the process still needs to be further studied and solved in the future. Therefore, finding a safe and stable solvent is necessary to enhance the process efficiency further and reduce the environmental impact caused by the volatilization of absorbents.
In recent years, as one of the hotspots in the field of green chemistry, ionic liquids (ILs) have been proposed as new candidate materials for selective capture of CO 2 , SO 2 , and H 2 S due to their unique physical and chemical properties, such as adjustable structures, low vapor pressure, high thermal stability and chemical stability, high solubility for organic and inorganic compounds, and wide liquid phase range [18][19][20]. ILs have also been studied to absorb NO 2 in recent reports. Kunov Kruse et al. [21] first analyzed the feature spectra of several imidazolyl ILs for capturing NO 2 . These ILs can easily overcome the Lewis base catalytic reaction between NO 2 and ILs in an anhydrous state, which prevents the solvent from anion loss and dramatic viscosity changes at high temperatures. Yuan et al. [22] found that NO 2 exists as a dimer among the molecules of [BMIM] [OTF]. Processes similar to molecular-imprinting of NO 2 during absorption-desorption illustrate the adsorption of NO 2 in ILs is controllable and reversible, and the absorption quantity can reach 0.0322 g·g −1 at the normal state. Duan et al. [20] tested the solubility of pure NO 2 in the special amine-functionalized ILs. The best among these ionic liquids was CPL:TBAB (2:1) whose absorption quantity was 0.355 g·g −1 , about 10 times of the traditional imidazolyl ILs. It can be concluded that amine-functionalized ILs have similar characteristics to those of organic amine solutions. However, amine-functionalized ILs have fewer basic amine groups than alkanolamines, and their anions easily interact with the tail of cation (-NH 2 ) through hydrogen bonds, which makes the viscosities of pure amine-functionalized ILs relatively high, further leading to lower gas capture capacity in the industry [23,24]. In this case, ionic liquids used alone are not enough to separate NO 2 maturely and economically.
In recent years, researchers have tried to use binary mixture (ionic liquid + alkanolamine) to absorb acid gas [25][26][27][28][29][30]. It has been found that the mixing of binary systems significantly increases the absorption of acid gas, reduces the possible corrosion to equipment, and reduces the viscosity of the fluid, which is very important. The method can reduce pumping energy consumption and the loss in the flow process and improve equipment durability. However, the absorption of NO 2 by binary mixtures (ILs + alkanolamines) has yet to be studied. According to Wolf et al. [31], based on imidazole ionic liquids in 2012, and Gold et al. [32] Figure 1 shows that the viscosity measurement values of TEA are basically consistent with the values of Lopez et al. [33] and Li et al. [34,35], with an average error of 0.7985%. The viscosity values of [EMIM] [TFA] are slightly lower than those of Karim et al. [36] and Hector et al. [37], with an average error of −3.512%. This phenomenon may be due to the extreme sensitivity of [EMIM] [TFA] to chloride ions and trace moisture (>104 ppm) carried in the synthesis process [38].  Figure 1 shows that the viscosity measurement values of TEA are basically consistent with the values of Lopez et al. [33] and Li et al. [34,35], with an average error of 0.7985%. The viscosity values of [EMIM] [TFA] are slightly lower than those of Karim et al. [36] and Hector et al. [37], with an average error of −3.512%. This phenomenon may be due to the extreme sensitivity of [EMIM] [TFA] to chloride ions and trace moisture (>104 ppm) carried in the synthesis process [38].  The viscosity values of ([EMIM] [TFA] + TEA) binary mixtures at different temperatures and molar fractions are shown in Table 2 and Figure 2. It can be seen that the viscosity of binary mixture decreases with the increase in temperature, and the viscosity of the mixture is always between the two-phase pure liquid, that is,  (1), and the fitting parameters are shown in Table 3.
(1)   The viscosity deviation values were calculated from the data of viscosity by using the following Equation (2).
where η, η 1 and η 2 stand for the viscosity of (  Figure 4.    Similar to the behavior of most non-aqueous binary mixtures [41][42][43], the viscosity deviation of ([EMIM] [TFA] + TEA) binary mixtures presents an entirely negative deviation which is different from the ideal situation in the whole range of χ 1 in Figure 4. This behavior can be explained by weakening strong and specific interactions in the blends [44]. At the same time, it was observed that with the increase in [EMIM] [TFA] molar fraction, the viscosity deviation first decreased and then increased. When χ 1 is about 0.4, the absolute viscosity deviation is max, and the maximum value is |∆η min |. Meanwhile, |∆η min | decreases from 175.48 mPa·s to 12.19 mPa·s with temperature (303.15 K-43.15 K). Furthermore, the viscosity deviations have also been fitted by a Redlich-Kister [45] type Equation (3), a fourth-order polynomial was found to be optimum for viscosity deviations of binary mixtures ([EMIM] [TFA] + TEA). The standard deviation was calculated by using the following Equation (4). The fitted parameters and standard deviation σ were listed in Table 5.
where ∆Q represents ∆η is the excess property, χ 1 is the molar fraction of [EMIM] [TFA], B p is the fitting parameters, M is the degree of the polynomial expansion, which was optimized using the Marquardt algorithm. where n dat is the number of experimental data points, Z exp and Z cat are the experimental value, calculated by Equation (3), respectively.   (5) was found to satisfacto-rily correlate the change of absorption with temperature. By fitting these data, it is found that the second-order polynomial Equation (5) has a good correlation with the change of absorption with temperature. The parameters of the equation are shown in Table 7. ([EMIM] [TFA] + TEA) and NO 2 absorption is negatively correlated with temperature, as shown in Figure 5. Even though the viscosity of blends is low at high temperatures, the low physical solubility of ([EMIM] [TFA] + TEA) seems to dominate. The change of absorption with different molar fractions of [EMIM] [TFA] at specific temperatures and at one atmosphere is shown in Figures 6 and 7. With the increase in the TEA component, the absorption of NO 2 increased. It can be considered that the mole fraction of TEA is significant for the interaction between the binary mixture of ([EMIM] [TFA] + TEA) and NO 2 . The relationship between the mole fraction of [EMIM] [TFA] and the absorption is well fitted by the fourth-order polynomial Equation (6), and the fitting parameters are shown in Table 8.

Absorption Capacity
A = D 0 + D 1 x + D 2 x 2 + D 3 x 3 + D 4 x 4 (6)   With the increase in the molar fraction of TEA in the binary mixture ([EMIM] [TFA] + TEA), the absorption of NO 2 increased. However, at the same time, the viscosity of the binary system increases rapidly, which is easy to make the absorption equipment stagnate and lead to wall adhesion. For example, at 313.15 K, the viscosity of TEA is 207.24 mPa s, which is about 12 times the viscosity of pure ionic liquid under the same conditions. Therefore       [22,46]. The absorption capacity of NO 2 by the mixture of ionic liquid created in this experiment is significantly better than those of other absorbents, and the negative effect of water was eliminated.

Theoretical Analysis
Spectroscopic studies show that there is a reversible equilibrium between NO 2 and N 2 O 4 , (the partial pressures of N 2 O 4 at different temperatures are shown in the Supplementary Data [47,48]), and NO + and NO 3 − are easily produced by autoionization in liquids or organic phases [49]. The molecular adduct [(Don) n ·NO + ]NO 3 − will be formed reversibly When N 2 O 4 is absorbed by the organic solvent (Don) [49]. If n > 2, the molecular adduct is more stable. Based on the experimental phenomena and theoretical analysis, we thought that the reaction of NO 2 in pure TEA is similar to that introduced by Addison, and the reaction product is a stable ionic substance [{(HOC 2 H4) 3 N} 2 ·NO + ]NO 3 − , as shown in Figure 9. We found that the system's viscosity tends to decrease when adding ionic liquid

Reusability of ([EMIM] [TFA] + TEA)
The recovery rate and reuse effect of ([EMIM] [TFA] + TEA) can be determined by absorption and desorption experiments. Therefore, the purpose of the technical efficiency and economic feasibility can be realized. In this work, the reusability of ([EMIM] [TFA] + TEA) was studied. Usually, saturated absorption amount was observed in 60 min at 303.15 K and 1.01 × 10 5 Pa, with a stream of 30 mL·min −1 pure NO 2 , and the captured NO 2 can be removed in 40 min at 70°C under vacuum (4.24 kPa). When the molar fraction of [EMIM] [TFA] was 0.58, a recycle of six times NO 2 absorption and desorption of ([EMIM] [TFA] + TEA) at 303.15 K was observed ( Figure 11). The quantitatively recovered ([EMIM] [TFA] + TEA) was directly used in the following absorption process without further treatment. The results showed that the absorption was highly reversible, and the absorption amount remains basically unchanged after six consecutive absorption/desorption cycles. The slight decrease in absorption may be due to the loss of TEA and the increase in ILs mole fraction., but the advantage of this situation was that the desorption quantities of NO 2 were increased.  Table 9 for details of samples).
[EMIM] [TFA] was prepared similarly to the literature method [50,51]. 1-methylimidazole (30 mL) and ethyl bromide (90 mL) were fully mixed in a 250 mL round bottom flask and heated under reflux for 8 h to obtain an oily liquid. Then, the liquid was cooled to room temperature and a white solid was obtained. Then, the white solid was dissolved using acetonitrile and filtered, the ethyl acetate was added to the filtrate to obtain a white solid. The above operations were repeated, and the recrystallized product was dried under vacuum for 36 h to obtain 1-ethyl-3-methylimidazolium bromide (EMIMBr). Then, 19.1 g of EMIMBr and 22.4 g of silver trifluoroacetate were weighed out and added to 200 mL of acetonitrile. Finally, the silver bromide precipitate was filtered and the acetonitrile was removed by rotary evaporation. The final product [EMIM] [TFA] was obtained after vacuum drying at 60°C. The reaction process is shown in Figure 12

Instruments and Procedures
The binary mixture of ([EMIM] [TFA] + TEA) was prepared using a precise electronic balance (±10 −4 g, Shanghai Precision and Scientific Sky Beautiful Instrument Co. Ltd, Shanghai, China). The viscometer (Zhejiang Jiaojiang Glass Instrument Factory, Taizhou, China) was calibrated in a super constant temperature bath (Shanghai Pingxuan Scientific Instrument Co. LTD, Shanghai, China, ±10 −2 K) with glycerol, glycol, and water. The viscosity was detected by the Pinkevitch method (according to GB/T 10247-2008), and the viscosity meter was calibrated by glycerol [39,40]. H-NMR spectra were collected on a Bruker AVANCE II 400 M spectrometer using d6-DMSO as the solvent and TMS as the internal standard. FTIR spectra were collected on a Bruker Tencer 27 spectrometer.
The absorption device of NO 2 is shown in Figure 13. The device mainly includes one homemade NO 2 generator, an MFC (Tianjin Jisite Instrument Co. Ltd, Tianjin, China, with an accuracy of ±1% F.S.), two sealed flasks, one three-necked flask, one constant temperature reaction bath (DF-101SZ Gongyi Yuhua Instrument Co. Ltd, Gongyi, China, Temperature control accuracy is ±0.1 K), and one temperature sensor.
The three-necked flask was immersed into the constant temperature water bath of a constant temperature magnetic stirrer. NO 2 was continuously passed into the vessel after its flow rate v NO2 (v NO2 = 30 mL·min [TFA] + TEA) absorbing NO 2 gas and three-necked flask were weighed by an electronic balance at a constant time interval (t = 10 min). Repeat weighing was carried out until the range of mass variation was less than 1%, considered a reaction balance. The NO 2 load was defined as A NO 2 = mNO 2 /(m [EMIM] [TFA] + m TEA ). The unabsorbed residual exhaust was introduced into the sodium hydroxide aqueous solution for recovery. The whole system operated at a constant atmospheric pressure of 0.1 ± 0.005 MPa.

Conclusions
In summary, the viscosity of binary mixtures of 1-ethyl-3-methylimidazolium trifluo- [TFA] + TEA) shows great potential as an alternative absorbent in industrial denitrification due to its sound absorption and low viscosity.