Ammonium-based deep eutectic solvent as entrainer for separation of acetonitrile–water mixture by extractive distillation

Ammonium-based deep eutectic solvent as entrainer for separation of acetonitrile–water mixture by extractive distillation

Journal of Molecular Liquids 285 (2019) 185–193 Contents lists available at ScienceDirect Journal of Molecular Liquids journal homepage: www.elsevie...

1MB Sizes 0 Downloads 126 Views

Journal of Molecular Liquids 285 (2019) 185–193

Contents lists available at ScienceDirect

Journal of Molecular Liquids journal homepage: www.elsevier.com/locate/molliq

Ammonium-based deep eutectic solvent as entrainer for separation of acetonitrile–water mixture by extractive distillation Bandhana Sharma, Neetu Singh ⁎, Jai Prakash Kushwaha Department of Chemical Engineering, Thapar Institute of Engineering and Technology, Patiala 147004, India

a r t i c l e

i n f o

Article history: Received 12 August 2018 Received in revised form 15 April 2019 Accepted 16 April 2019 Available online 17 April 2019 Keywords: Deep eutectic solvents Acetonitrile–water Azeotrope Vapor–liquid equilibrium, recoverability

a b s t r a c t The increasing concern about environmental issues and the requirement to reduce the negative influence of industrial processes has directed the attention of scientific community toward the development of new green solvents. Recently, Deep Eutectic Solvents (DESs), which are ILs analogues have appeared as a promising substitutes to conventional volatile organic solvents. Unlike ILs, DESs offer inexpensive and easy synthesis, less toxicity and good biodegradability. In this study, ammonium-based deep eutectic solvent was synthesized and its feasibility was assessed as extracting agent for acetonitrile dehydration using extractive distillation. Glycolic acid and tetramethylammonium chloride (TMAC) based DES was prepared in 3:1 M ratio (GTM3:1) due to its wellknown advantages over ionic liquids or organic molecular solvents. Isobaric vapor–liquid equilibrium data for pseudo-binary mixtures (acetonitrile–GTM3:1 & water–GTM3:1) were determined at 101.32 kPa. The pseudoternary VLE data of acetonitrile–water–GTM3:1 were also determined by keeping the concentration of DES nearly constant (x3 = 0.05, 0.10, and 0.15). The experimental data for these systems, were fitted using nonrandom two-liquid (NRTL) model and good correlation between predicted values and experimental data were obtained. From the results, GTM3:1 was capable of improving the relative volatility of the studied azeotropic mixture and successfully eliminated the azeotrope. The recoverability of used GTM3:1 was also investigated and the chemical properties of recovered GTM3:1 were found stable. © 2019 Elsevier B.V. All rights reserved.

1. Introduction Acetonitrile (ACN) is a vital organic solvent and is accepted in many industrial applications [1]. Several industries produce large amount of acetonitrile–water mixture as spent stream. Separation of ACN–water mixture is of particular interest for scientific community due to the remarkable demand of acetonitrile. However, recovery of ACN from water is often problematic as ACN–water mixture exhibits minimum boiling azeotrope at 67.4 mol% (ACN) at 349.5 K temperature and 101.32 kPa pressure. Different separation techniques like; azeotropic distillation [2], pressure-swing distillation [3], extractive distillation [4], liquid-liquid extraction [5,6], pervaporation [7,8] and adsorption [9,10] have been described for dehydration of acetonitrile. Extractive distillation is proved as a favourable technique for bulk production of anhydrous ACN. In the previous literature, many conventional solvents like; dimethyl sulfoxide (DMSO) [4], ethylene glycol [11,12] and butyl acetate [13] have been used for dehydration of ACN– water mixture. There are many challenges associated with the use of these conventional solvents like; mixing problem of ethylene glycol with ACN–water mixture and high entrainer dose requirement. The ⁎ Corresponding author. E-mail address: [email protected] (N. Singh).

https://doi.org/10.1016/j.molliq.2019.04.089 0167-7322/© 2019 Elsevier B.V. All rights reserved.

high volatility of butyl acetate makes the recycling process troublesome. However, DMSO is found as an efficient solvent as it could successfully break the azeotrope without much operational difficulty. The popularity of ionic liquids (ILs) grew rapidly from last few years due to its unique attributes like minimal vapor pressure, broad liquid range, chemical & thermal stability, and designability. Three imidazolium based ILs; 1-butyl-3- methylimidazolium dibutyl phosphate ([Bmim][DBP]), 1-butyl-3-methylimidazolium chloride ([Bmim] [Cl]) and 1-butyl-3-methylimidazolium tetrafluoroborate ([Bmim] [BF4]) were investigated for dehydration of acetonitrile [14–15]. Out of these [Bmim][BF4] was found ineffective to the enhancement of relative volatility of the ACN–water system. Complex and Expensive synthesis of ILs as compared to traditional solvents hamper the application of ILs as entrainer [16]. In addition, the ILs are not intrinsically green. Depending on its chemical structure, few ILs are as venomous as organic solvents. The growing demand for green solvents has spurred the development of environmentally benign solvents. Deep eutectic solvents (DESs), have latterly appeared as an attractive replacement for traditional organic solvents and ionic liquids (ILs) [17]. Due to the multitasking nature of DESs, these have been applied to many chemical processes. DESs are combination of two (or more) components, commonly a salt and a hydrogen-bond donor (HBD), which forms a eutectic

186

B. Sharma et al. / Journal of Molecular Liquids 285 (2019) 185–193

Table 1 Chemicals used. Name

CAS no.

Source

Purity (wt%)

Purification method

Acetonitrile Glycolic acid Tetramethylammonium chloride Karl Fischer reagent Double distilled water

75-05-8 79-14-1 75-57-0 7446-09-5

Merck chemicals Spectrochem Merck chemicals Merck chemicals Our laboratory

≥99.8% N98% N98% N99%

None Vacuum drying Vacuum drying None

mixture [18]. Because of the strong hydrogen-bonding impact, these have much lower melting temperature than those of starting original substances. Due to these unusual chemical and physical properties, DESs are currently getting remarkable attention as “greener” means for a wide variety of applications. In this work, an ammonium-based deep eutectic solvent, GTM 3:1 (glycolic acid–tetramethylammonium chloride in 3:1 M ratio) was selected as entrainer for the separation of the azeotropic ACN–water mixture via extractive distillation. GTM 3:1 is proposed as entrainer due to the following advantages over conventional solvents and ILs: 1. DESs share many physicochemical properties with ILs. While retaining the useful properties of ILs, it offers some distinctive advantages such as low cost, easy preparation with high purity, less toxicity and more biodegradable components than ILs.

2. Deep eutectic solvents comprising carboxylic acids and ammoniumbased salts present the greater ability to donate and accept protons while comparing with organic molecular solvents and most of the ionic liquids [19]. 3. Most DESs are nonreactive with water and the range of possible reagents and compositions are almost unlimited. The objective of present study is to synthesize ammonium-based deep eutectic solvent, GTM 3:1 (glycolic acid–tetramethylammonium chloride in 3:1 M ratio) and investigate its effectiveness as entrainer for separation of ACN–water azeotropic mixture via extractive distillation. The Isobaric VLE data for the pseudo-binary systems (ACN– GTM3:1, water–GTM3:1) and pseudo-ternary system (ACN–water– GTM3:1) were measured at 101.32 kPa. These experiments were conducted for three different DES molar compositions; 5%, 10% and 15%.

Fig. 1. Typical FTIR spectrum of GTM3:1 before (a) and after (b) use.

B. Sharma et al. / Journal of Molecular Liquids 285 (2019) 185–193

187

Fig. 2. Schematic diagram of improved Othmer type recirculation still.

Correlation of the VLE data for these systems was done by NRTL model. The Predicted values correlated the experimental data very well. Furthermore, the synthesized DES in this study was also investigated for its recoverability. 2. Experimental 2.1. Chemical Acetonitrile (HPLC grade), glycolic acid, tetramethylammonium chloride (TMAC), distilled water and solvent for the Karl-Fischer titration were used for this study. Acetonitrile 99.8 + %, Tetramethylammonium chloride 98.0 + %, Karl Fischer reagent 99 + % were purchased from Merck chemicals. Glycolic acid 98.0 + % were purchased from Spectrochem. The sources and purities of used chemicals is summarised in Table 1. Glycolic acid and TMAC undergone vacuum drying prior to use, due to its hygroscopic nature. The water content of these chemicals was determined by using Karl–Fischer moisture analysis (type Esico 1760) and found to be b0.3 wt% in all cases.

was produced by melting, stirring was started. During the heating, the temperature of oil bath was maintained to 80 ± 0.1 °C by using a temperature controller. The heating was stopped when the mixture converted into a transparent and colorless liquid, and further it was allowed to cool to ambient temperature. Thermogravimetric analysis of prepared GTM3:1 was performed to verify its thermal stability using PerkinElmer, STA6000 TGA analyser (Fig. S1). The amount of water present in the synthesized DES was measured using Karl– Fischer titration method, (type Esico 1760). FT-IR analysis (PerkinElmer) of prepared DES was also conducted, and its spectra is represented in Fig. 1a. In the FT-IR spectra, a sharp and high intensity peak at 1731 cm−1 with comparatively narrower band at 1194 and 1085 cm−1 can be observed at the lower wave number side of the FTIR spectra. The peak at 1731 cm−1 can be allocated to the free carbonyl group stretching of glycolic acid. The C\\N stretching in the Tetramethylammonium chloride and hydrogen bond between OH group and Cl− of glycolic acid and Tri-methyl ammonium chloride, respectively, is represented by the peaks at 1194–1085 cm−1. A broad band in the middle of 3100–3700 cm−1, with peak at 3303 cm−1 demonstrates the existence of both free and hydrogen bonded OH groups of

2.2. Synthesis and characterization of DES Deep eutectic solvent was synthesized using glycolic acid as HBD and tetramethylammonium chloride (TMAC) as HBA in 3:1 M ratio (GTM3:1). The synthesis of GTM3:1 was done by following the similar procedure reported by Francisco et al. [20]. Vacuum dried glycolic acid and TMAC were weighted using a Mettler Toledo electronic balance (uncertainty 0.0001 g) for preparing the mixtures in exactly 3:1 M ratio. Further, both the chemicals were mixed in a closed glass container to get homogeneous mixture, and in subsequent process the mixture undergone heating in a thermostatic oil bath. When sufficient liquid

Table 2 GC conditions used for analysis of vapor and liquid phases. Detector type

FID Detector

Column type Detector temperature Injector temperature Injection volume Carrier gas Flow rate

TG-Bond Q (30 m; 0.53 mm; 20.0 μm) 250 °C 200 °C 1.0 μL Helium 5.0 ml min−1 (constant)

188

B. Sharma et al. / Journal of Molecular Liquids 285 (2019) 185–193

2.3. Apparatus and procedure

Fig. 3. Comparison of isobaric T − x − y data for ACN (1)–water (2) system at atmospheric pressure, ▲ this work; ● ref. [20].

The apparatus used for experimental measurements was improved Othmer type recirculation still [21]. The schematic diagram of the setup is shown in Fig. 2. The original Othmer recirculation still was modified by removing the drainage port and adding a widened loading port in the side of the still to avoid the clogging with salt and other compounds. A stirrer was also added under the still to assure uniform composition of the liquid within the body of the still. Moreover the constant volume condensate chamber was replaced with three condensers in series to increase the flexibility in operation by providing opportunity of controlling the holdup volume and sample size. The detail of setup is given in our previous study [22]. Setup is composed of a still (150 ml capacity) and a glass-sealed wire heater for boiling the mixture of the still. An air-cooled condenser is provided at the top of still for partial condensation of vapors generated in the still. The uncondensed vapors move toward the water-cooled recirculation condenser. The Condensate was circulated back to the still to provide intimate contact between both the phases. A condensate sample bulb is provided below the condenser for periodic sampling. The uncondensed vapor from first condenser, moves to the second condenser. The apparatus pressure was maintained at atmospheric pressure during the measurement process. A precision and calibrated mercury thermometer was utilized for measurement of the equilibrium temperature with an uncertainty of ±0.1 K. 45 min time was provided to attain the equilibrium after

glycolic acid–glycolic acid molecules. Because of formation of hydrogen bonds between the HBD and HBA, freezing point of the mixture decreases as a result DES forms. Table 3 Experimental VLE data and correlated results of pseudobinary subsystems, Activity Coefficient γi, deviation in activity Coefficient Δγi and deviation in equilibrium temperature ΔT at 101.32 kPa. Water (1) + GTM (3:1, mol/mol) (2) S.No.

T (K)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

373.1 373.3 373.7 374.2 374.9 376.1 377.4 378.7 380.2 382.2 384.5 387.6 390.9 394.5 398.8

1 2 3 4 5 6 7 8 9 10 11 12 13 14

354.7 355.5 356.5 357.8 359.2 360.8 362.6 364.5 366.5 368.6 370.7 372.9 376.2 381.4

Δγ1a

ΔTb (K)

Water (1) + GTM (3:1, mol/mol) (2) 0.00 1.000 1.002 1.42 0.987 1.007 2.95 0.970 1.010 4.60 0.954 1.008 6.39 0.934 1.004 8.35 0.918 0.979 10.48 0.892 0.963 12.83 0.870 0.943 15.41 0.845 0.922 18.27 0.817 0.890 21.46 0.787 0.855 25.03 0.753 0.807 29.07 0.707 0.772 33.66 0.665 0.730 38.93 0.612 0.693

0.002 0.008 0.014 0.018 0.023 0.008 0.010 0.008 0.009 0.003 −0.002 −0.017 −0.005 −0.005 0.009

0.003 −0.174 −0.346 −0.453 −0.617 −0.182 −0.242 −0.186 −0.223 −0.056 0.118 0.670 0.245 0.274 −0.366

ACN (1) + GTM (3:1, mol/mol) (2) 0.00 1.000 0.999 4.05 0.959 1.016 8.18 0.916 1.031 12.40 0.878 1.034 16.70 0.833 1.045 21.09 0.787 1.054 25.58 0.745 1.056 30.16 0.696 1.068 34.84 0.651 1.078 39.62 0.606 1.090 44.50 0.553 1.124 49.50 0.505 1.157 54.60 0.457 1.162 59.83 0.401 1.149

−0.001 0.013 0.022 0.016 0.013 0.008 −0.005 −0.008 −0.014 −0.017 0.001 0.021 0.013 −0.012

0.070 −0.374 −0.645 −0.450 −0.380 −0.205 0.199 0.321 0.506 0.582 0.010 −0.605 −0.364 0.440

Mole % of GTM (3:1)

x1

γexp 1

Standard uncertainty u(x1) = 0.003, u(T) = 0.1K, u(mole % ofGTM(3 : 1)) = 0.1, u(P) = 0.05kPa. a Δγ1 = γexp − γcal 1 1 . b ΔT = Texp − Tcal.

Fig. 4. Effect of GTM3:1 on the normal boiling point of ACN and water at 101.32 kPa. Experimental data for water (●) and ACN (▲); and solid lines, calculations based on NRTL model.

B. Sharma et al. / Journal of Molecular Liquids 285 (2019) 185–193

Table 4 Experimental Isobaric VLE data for ACN(1) + water(2) + GTM3:1 (3) system, experimental Activity Coefficient γexp and Relative Volatilities α12 at 101.32 kPa. Presence of GTM3:1 i was not detected in the vapor phase. T (K)

x'1

357.3 357.3 357.4 357.5 357.9 358.3 358.4 359.6 360.1 361.2 363.8 365.3 365.9 367.3 368.2 369.6 370.5 372.1 374.2 376.4 377.2 378.3 379.8

Glycolic acid/TMAC 3:1 = 5 mol% 1.000 1.000 0.901 0.856 0.094 0.951 1.025 0.842 0.800 0.150 0.914 1.051 0.806 0.765 0.185 0.876 1.050 0.757 0.720 0.231 0.849 1.069 0.709 0.675 0.277 0.828 1.098 0.689 0.655 0.296 0.812 1.106 0.658 0.625 0.325 0.782 1.077 0.601 0.571 0.379 0.755 1.121 0.543 0.515 0.434 0.717 1.143 0.413 0.392 0.557 0.647 1.253 0.392 0.372 0.576 0.624 1.221 0.347 0.330 0.620 0.599 1.298 0.316 0.300 0.650 0.580 1.326 0.291 0.277 0.674 0.563 1.361 0.267 0.253 0.697 0.558 1.416 0.248 0.236 0.715 0.530 1.408 0.202 0.192 0.758 0.484 1.511 0.177 0.168 0.783 0.453 1.521 0.138 0.131 0.820 0.387 1.569 0.110 0.105 0.846 0.340 1.690 0.086 0.082 0.868 0.285 1.759 0.000 0.000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

359.2 359.2 359.3 359.4 359.9 360.2 361.1 361.9 362.4 364.1 365.7 368.7 370.2 372.3 373.1 375.4 375.8 376.2 377.2 378.1 378.9 380.4 381.4

Glycolic acid/TMAC 3:1 = 10 mol% 1.000 1.000 0.879 0.791 0.109 0.977 1.076 0.838 0.755 0.146 0.943 1.085 0.829 0.754 0.156 0.932 1.070 0.786 0.707 0.193 0.908 1.095 0.754 0.679 0.221 0.893 1.112 0.738 0.664 0.236 0.883 1.094 0.713 0.644 0.259 0.872 1.089 0.671 0.605 0.296 0.853 1.117 0.627 0.567 0.337 0.821 1.091 0.559 0.503 0.397 0.803 1.147 0.465 0.419 0.481 0.734 1.156 0.419 0.378 0.524 0.711 1.189 0.329 0.297 0.606 0.651 1.304 0.286 0.257 0.643 0.623 1.409 0.230 0.207 0.693 0.580 1.529 0.217 0.195 0.705 0.562 1.554 0.195 0.176 0.725 0.539 1.641 0.175 0.159 0.751 0.493 1.610 0.140 0.126 0.774 0.431 1.735 0.097 0.087 0.815 0.350 1.989 0.045 0.041 0.860 0.185 2.185 0.000 0.000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

361.1 361.1 361.9 362.5 363.4 363.9 364.7 366.3 367.2 368.1 369.3 371.2 372.0 372.8 374.9 376.1 376.8 377.9 379.4 381.1 382.7 383.4

Glycolic acid/TMAC 3:1 = 15 mol% 1.000 1.000 0.856 0.728 0.122 0.989 1.119 0.829 0.705 0.146 0.977 1.113 0.811 0.690 0.161 0.973 1.113 0.781 0.665 0.187 0.969 1.119 0.759 0.645 0.205 0.962 1.130 0.705 0.599 0.251 0.927 1.145 0.654 0.555 0.294 0.903 1.149 0.618 0.527 0.326 0.876 1.144 0.572 0.486 0.364 0.858 1.184 0.511 0.434 0.416 0.819 1.222 0.449 0.382 0.469 0.782 1.257 0.413 0.351 0.500 0.759 1.297 0.387 0.329 0.522 0.746 1.330 0.289 0.246 0.604 0.677 1.527 0.220 0.187 0.662 0.591 1.696 0.195 0.166 0.686 0.553 1.750 0.168 0.143 0.707 0.524 1.873 0.112 0.095 0.757 0.413 2.121 0.071 0.060 0.790 0.305 2.366 0.037 0.031 0.820 0.189 2.693 0.000 0.000

S.No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

x1

x2

y1

γexp 1

γexp 2

α12

0.946 1.036 1.210 1.158 1.083 1.104 1.114 1.052 1.016 0.895 0.871 0.844 0.800 0.776 0.721 0.723 0.706 0.673 0.666 0.677 0.687

2.13 1.99 1.70 1.80 1.98 1.95 1.86 2.05 2.14 2.60 2.57 2.81 2.99 3.14 3.47 3.42 3.71 3.86 3.95 4.16 4.23

189

which condensed vapor and liquid phase samples were drawn from the condensate sampling bulb and still respectively. 2.4. Composition analysis Gas chromatograph (Bruker, Model: SCION 456-GC) equipped with a flame ionization detector (FID) was used to determine the equilibrium acetonitrile compositions in condensed vapor phase. A Thermo Scientific TG-Bond Q column (30 m long, 0.53 mm inner diameter and 20.0 μm film thickness) was used as separating media. No traces of DES were observed in the vapor phase. The temperature for the oven was set at 180 °C. Same GC method was used for analysing the acetonitrile compositions in liquid phase and water content was checked with the help of a Karl-Fischer Titrator (Esico model 1760) in triplicate. As DES is present in the liquid sample, it is essential to trap the DES before it enters in column. Glass liners (Agilent Technologies) were used for this purpose and it effectively prevented the DES from entering in the column. The entrainer's concentration of sample was checked by mass balance. The compositions of vapor and liquid phases were evaluated in mole fraction with an uncertainty of ±0.003. The GC method is described in Table 2. 3. Results and discussion 3.1. Reliability assessment of the set-up

0.356 0.655 0.729 0.782 0.783 0.777 0.748 0.740 0.741 0.652 0.649 0.614 0.593 0.588 0.560 0.565 0.571 0.585 0.617 0.651 0.735

5.85 3.20 2.83 2.69 2.72 2.68 2.75 2.84 2.73 3.22 3.18 3.41 3.80 4.13 4.62 4.63 4.83 4.58 4.65 5.02 4.83

0.141 0.240 0.249 0.238 0.261 0.397 0.424 0.473 0.469 0.501 0.499 0.503 0.493 0.502 0.556 0.573 0.569 0.623 0.666 0.709

15.12 8.76 8.40 8.77 8.04 5.31 4.93 4.37 4.52 4.33 4.40 4.48 4.65 5.16 5.12 5.11 5.45 5.58 5.74 6.07

Experimental setup and procedure authenticity was verified by measuring the ACN–water system binary VLE data at 101.32 kPa. The temperature vs composition data obtained in this study are given in Table S1 and its comparison with the literature data is presented in Fig. 3. As shown in figure, experimental VLE in this study is in good agreement with the literature data [23], thus validating the reliability of experimental setup used. 3.2. Vapor-liquid equilibrium (VLE) measurements The isobaric VLE data for pseudobinary systems of ACN (1) – GTM3:1 (2) and water (1) – GTM3:1 (2) were measured at 101.32 kPa and the experimental data are given in Table 3. The increment in boiling point (Tb) of ACN and water due to inclusion of non-volatile GTM3:1 is represented in Fig. 4. Boiling points of ACN and water both were higher after addition of GTM3:1. VLE Measurements were made for ACN (1)–water (2)–GTM3:1 (3) pseudoternary system at atmospheric pressure, by keeping the DES concentration fix for every group of experimentation. The isobaric VLE data for different DES mole percent (5%, 10%, and 15%) are given in Table 4. Where T is the equilibrium temperature, x1 is the mole fraction of ACN in liquid phase, x'1 is the mole fraction of ACN in liquid phase expressed on GTM3:1 free basis and y1 is the mole fraction of ACN in the vapor phase. To describe the salting-out effect caused by addition of GTM3:1, the T − x, y data are depicted in Figs. 5 (a), (b) and (c) for 5, 10 and 15 mol% concentrations respectively. The equilibrium temperature of system increases, due to the addition of the GTM3:1. The x − y diagram for different GTM3:1 M concentrations are also represented by Fig. 6. From the figure, it is clear that GTM3:1 creates the good salting-out effect on the azeotropic mixture of ACN (1) –water (2). The addition of a small quantity of GTM3:1 produces considerable effect, due to this the ACN concentration in vapor phase increase consistently. We can see from Table 4 that GTM3:1 effectively improves the relative Note to Table 4: Standard uncertainty u (x 1 ) = u (x '1 ) = u (y) = 0.003, u (T) = 0.1 K, u (mole % of GTM (3 : 1) ) = 0.1, u (α) = 0.025. x'1 is mole fraction of ACN in liquid phase expressed on DES free basis. The uncertainty calculations in mole fraction x1′ using triplicate data are illustrated in Table S2 (Supplementary information).

190

a)

B. Sharma et al. / Journal of Molecular Liquids 285 (2019) 185–193

volatility of acetonitrile to water and it is greater than unity in the whole concentration range. The ability of GTM3:1 to increase the relative volatility α12 of azeotropic mixture of ACN– water, has been studied. For this purpose, experiments were performed in near azeotrope composition (x'1 ≈ 0.674) of ACN–water mixture in the presence of GTM3:1 and relative volatilities were evaluated. Results for different molar compositions of GTM3:1 (0 to 17.7 mol%) are given in Table 5 and are depicted in Fig. 7. The relative volatility of ACN to water enhances with the concentration of GTM3:1 in the liquid phase. The more GTM3:1, the greater the relative volatility of ACN to water, which means, the easier the separation. At 17.7 mol% of GTM3:1, the value of α12 was 5.31, which is 5.31 times than for the GTM3:1 free system. 3.3. Comparison of GTM3:1 with other entrainers

b)

c)

The performance of an entrainer in extractive distillation process can be evaluated from the enhancement in relative volatility of the system. The proposed DES (GTM3:1) was compared with typical conventional entrainers [4,11–13] and previously used ILs [14–15] for its performance. DMSO [4] (molecular weight = 62.07 g/mol) presented high relative volatility values (≈10) for the ACN–water system but at very high entrainer dosage only; 40 and 60 mol% (≈55.2 and 73.5 mass% respectively at azeotropic composition). At 20 mol% concentration it could not break the azeotrope. On the other hand, GTM3:1 (molecular weight = 84.44 g/mol) at 5 mol% (≈11.7 mass% at azeotropic composition) concentration was found sufficient enough to break the azeotrope completely. Out of all the used ILs [14–15], only 1-butyl-3methylimidazolium dibutyl phosphate ([Bmim][DBP]) and 1-butyl-3methylimidazolium chloride ([Bmim][Cl]) were found effective to break the ACN–water azeotrope and the separation ability of [Bmim] [Cl] was found better than [Bmim][DBP]. [Bmim][Cl] (molecular weight = 174.68 g/mol) was best performing at 30 mass% (≈7.6 mol % at azeotropic composition). By comparing the performance of [Bmim][Cl] with GTM3:1, it was observed that even at the highest [Bmim][Cl] dose the relative volatility for ACN–water system was low as compared to GTM3:1. Mixing problem and high dose requirement of ethylene glycol [11,12] and recycling problem of butyl acetate [13] due to its low boiling point make these inconvenient to be used as entrainer for acetonitrile dehydration. As compared to conventional entrainers and ILs, amount of GTM3:1 requirement is low for separation of ACN–water azeotropic mixture. Moreover, easy and low cost preparation, less toxicity and more biodegradable components than ILs are additional advantages of using GTM3:1 as entrainer. 3.4. Correlation of the phase equilibrium V L By fugacity equality, the fugacity ^f and ^f of each component, i in the vapor phase and liquid phase, should be equal at VLE.

^f V ¼ ^f L

ð1Þ

By applying γ − ϕ approach, following equilibrium Equation can be achieved V

^ p ¼ xi γ ps ϕs PEi yi ϕ i i i

ð2Þ

Fig. 5. (a) Temperature–composition diagram for the acetonitrile (1) + water (2) + GTM3:1 (3) system at 101.3 kPa with GTM3:1 = 5 mol%. x'1 vs T (●) and y1 vs T (○), — GTM3:1 = 0%,. (b) Temperature–composition diagram for the acetonitrile (1) + water (2) + GTM3:1 (3) system at 101.3 kPa with GTM3:1 = 10 mol%. x'1 vs T (▲) and y1 vs T (Δ), — GTM3:1 = 0%. (c) Temperature–composition diagram for the acetonitrile (1) + water (2) + GTM3:1 (3) system at 101.3 kPa with GTM3:1 = 15 mol%. x'1 vs T (■) and y1 vs T (□), — GTM3:1 = 0%.

B. Sharma et al. / Journal of Molecular Liquids 285 (2019) 185–193

191

Fig. 7. Effect of GTM3:1 concentration on the relative volatility α12 (x'1 ≈ 0.674) of the ACN (1) + water (2) system at 101.32 kPa. Fig. 6. Experimental and calculated VLE data for ACN (1) + water (2) + GTM3:1 (3) pseudoternary system at 101.32 kPa. For GTM = 5 mol% (●), for GTM = 10 mol% (▲), for GTM = 15 mol% (■), — GTM3:1 = 0% and solid lines, calculations based on the NRTL model.

where, yi and ϕVi are the mole fraction in vapor phase and fugacity coefficient of the ith component respectively; γi and xi are the activity coefficient and mole fraction in liquid phase of the ith component respectively. P and ϕSi denote the system pressure and fugacity coefficient of saturated component i at the system temperature, PEi is the Poynting effect, and psi is the vapor pressure of ith component evaluated by the Antoine equation at equilibrium temperature using Antoine parameters (Table 6) from the literature [24]. Reduced Virial equation of state at second coefficient with Tsonopoulo's correlations [25,26] was used for Fugacity coefficients calculation. The step by step methodology of calculation for activity coefficient and fugacity coefficient estimation using the Tsonopoulos correlation is presented in Supplementary information. At lower pressure/atmospheric pressure, the poynting factor, PEi ≈ 1. Therefore, the Eq. (2) is reduced to (Eq. (3)). ^ V p ¼ xi γ ps ϕs yi ϕ i i i i

ð3Þ

Due to the non-volatility of the DES (GTM 3:1) used, it did not appear in the vapor phase. Also, DES (GTM 3:1) is considered as a single component, therefore the mixture system reduced to pseudo-binary (ACN-GTM 3:1; water-GTM) and pseudo-ternary (ACN-water-GTM 3:1) systems [27,28]. However, in calculation of ACN and water activity

Table 5 Effect of GTM3:1 concentration on relative volatility α12 (x'1 ≈ 0.674) of ACN (1) + water (2) system, Relative volatilities α12 at atmospheric pressure (101.32 kPa): GTM3:1 = (0 to 17.7) mol%. Mol% of GTM3:1

Relative volatility (α12)

0.0 2.0 3.3 6.7 10.2 13.9 17.7

1.06 1.26 1.47 1.87 2.68 3.99 5.31

Standard uncertainty u(mole % ofGTM(3 : 1)) = 0.1, u(α) = 0.025.

coefficient (γcal) in liquid phase, DES (GTM 3:1) mole fraction should be considered. For activity coefficient (γcal) determination for the systems comprising DES, NRTL model is applied [27–29]. Further, the experimental activity coefficients (using Eq. (3)) for pseudo-binary (ACN– GTM3:1; water–GTM3:1) and pseudo-ternary system (ACN–water– GTM3:1) were correlated to activity coefficients predicated by the NRTL model. For the regression analysis, optimization technique in MatLab program with relative least-squares criteria to minimize the following objective function (OF) was used [30]: " exp #2 n X γi −γ calc i OF ¼ γ exp i i¼1

ð4Þ

where, n is the number of measurements, the superscripts ‘exp’ and ‘calc’ denotes the experimental and calculated data respectively. The objective function was minimized for evaluating the pseudo-binary and pseudo-ternary adjustable parameters. Following equation (Eq. (5)) was applied to calculate the relative volatility (α12) of ACN (1) to water (2): 0

α 12 ¼

y1 =x1 0 y2 =x2

ð5Þ

where, x'1, x'2: ACN and water mole fractions in liquid phase on DES-free basis respectively. y1, y2: ACN and water mole fractions in vapor phase, respectively. The results of correlation of the average deviation in vapor mole fraction and temperature for both the pseudobinary systems are listed in Tables 3 and 7, and are represented in Fig. 4. Results indicate that good correlation between predicted values and experimental data were obtained by using NRTL model.

Table 6 Antoine constants for pure substances [24]. Component

Acetonitrile Water

Antoine constants Ai

Bi

Ci

7.5305 8.07131

1609.86 1730.63

264.7 233.426

Bi sat log psat i ¼ Ai − T−273þC i ; where pi is in mm Hg and T is in K.

192

B. Sharma et al. / Journal of Molecular Liquids 285 (2019) 185–193

Table 7 The parameters and correlation deviations of NRTL model at 101.32 kPa. Systems

Δg12

Δg21

α

ACN (1) + GTM3:1 (2) Water (1) + GTM3:1 (2) ACN (1) + water (2) + GTM3:1 (3)

5567.9 1833.7 −973.6

−2786.3 −5778.6 3123.2

0.3 0.3 0.3

DT (K)

dT (K)

0.37 0.28 0.45

0.64 0.67 0.74

Dy1

dy1

0.0068

0.015

Δg12 = (g12 − g22). P DT ¼ ð1=nÞ nk¼1 jT exp −T cal jk . dT = max (|Texp − Tcal|). P cal Dy1 ¼ ð1=nÞ nk¼1 jyexp 1 −y1 jk . cal dy1 = max (|yexp − y |). 1 1

The Pseudo-ternary VLE data for ACN–water–GTM3:1 system were also correlated by using NRTL model. Since, DESs have high viscosity, for binary subsystems (containing DES) NRTL parameters cannot be applied to predict the ternary system's VLE behaviour under identical conditions [27,28]. Hence, the NRTL model was employed for correlating the ternary systems directly. The results of data correlation (binary interaction parameters for NRTL model and the difference in vapor mole fraction and temperature) are presented in Table 7, and graphical representations are shown in Fig. 6. Figure shows that the calculated results agree well with the experimental data. The average absolute differences for the equilibrium temperature and vapor phase ACN mole fraction were 0.45 K and 0.0068 respectively. Fig. 6 reveals that the NRTL model correlated the experimental data quite well in the presence of GTM3:1. 3.5. Recoverability assessment The recoverability of the used GTM3:1 was studied by boiling the content of vessel in rota evaporator for 2 h. Recovered GTM3:1 was vacuum dried before its reuse. After every individual run, the GTM3:1 was retrieved and reused for the successive runs till four cycles. Recovered GTM3:1 was then characterized using FT-IR and its spectra is given in Fig. 1(b). When this spectra is compared with the freshly prepared GTM3:1 spectra in Fig. 1(a), no remarkable change was found. Hence, due to stable chemical properties of recovered GTM3:1, it can be reused as entrainer for distillation. 4. Conclusions Ammonium-based DES (GTM3:1) was evaluated as entrainer in this work, for dehydration of ACN–water azeotropic mixture via extractive distillation. The isobaric VLE data of pseudobinary systems acetonitrile–GTM3:1 and water–GTM3:1 were determined experimentally at 101.32 kPa using an improved Othmer type still. VLE data for pseudoternary system acetonitrile (1)–water (2)–GTM3:1 (3) were also measured for three distinct DES molar compositions (x 3= 0.05, 0.10 and 0.15). The pseudoternary VLE data represents that addition of GTM3:1 created a significant salting-out effect by increasing the relative volatility of acetonitrile to water, and could eventually remove the azeotropic point. The recoverability of the used GTM3:1 was also studied by recovering and reusing it for four boiling cycles without any significant change in its initial chemical structure. GTM3:1 was found a suitable entrainer as it is capable of breaking the ACN–water azeotrope at lower dose as compared to conventional entrainers and ILs. Moreover low cost, easy preparation, biodegradable nature and recyclability of GTM3:1, makes its application as entrainer more attractive. NRTL model was applied for correlating the experimental VLE data of the pseudobinary and pseudoternary systems to obtain binary interaction parameters. Good correlations with the experimental data show the applicability of NRTL model for DES containing systems. The average absolute difference for the equilibrium temperature and vapor phase mole fraction of ACN was determined as 0.45 K and 0.0068 respectively.

Acknowledgements The authors acknowledge Thapar Institute of Engineering & Technology, Patiala, India, for the financial support to conduct this research under the grant TU/SEED/2014/CHE/NS. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.molliq.2019.04.089.

References [1] K.X. Xu, Handbook of Fine Organic Chemical Raw Materials and Intermediates, 2nd ed. Chemical Industry Press, Beijing, 2002. [2] R.A. Ruiz, B. N. Borda, L.R. Alexander, R. Javier, L. Guevara, D. Ivan, C. Gil, Control of an azeotropic distillation process to acetonitrile production, 21st European Symposium on Computer Aided Process Engineering – ESCAPE 21. [3] J.U. Repke, F. Forner, F. Klein, A. Separation of homogeneous azeotropic mixtures by pressure-swing distillation-analysis of the operation performance, Chem. Eng. Technol. 28 (2005) 1151–1157. [4] Z. Zhigang, L. Ming, H. Donghao, J. Peng, S. Dezhang, L. Wenxiu, Isobaric vapor−liqquid equilibrium for the extractive distillation of acetonitrile + water mixtures using dimethyl sulfoxide at 101.3 kPa, J. Chem. Eng. Data 58 (2013) 3364–3369. [5] C.V.S.R. Rao, K.V. Rao, A. Ravlprasad, C. Chiranjiv, Extraction of acetonitrile from aqueous solutions 1. Ternary liquid equilibria, J. Chem. Eng. Data 23 (1978) 23–25. [6] D.S. Rao, K.V. Rao, A.R. Prasad, C. Chiranjivi, Extraction of acetonitrile from aqueous mixtures. 2. Ternary liquid equilibria, J. Chem. Eng. Data 24 (1979) 241–244. [7] M.K. Mandal, S.B. Sant, P.K. Bhattacharya, Dehydration of aqueous acetonitrile solution by pervaporation using PVA–iron oxide nanocomposite membrane, Colloids Surf. A: Physicochem. Eng. Aspects 373 (2011) 11–21. [8] M. Khayet, C. Cojocaru, G. Zakrzewska-Trznadel, Studies on pervaporation separation of acetone, acetonitrile and ethanol from aqueous solutions, Sep. Purif. Technol. 63 (2008) 303–310. [9] J. Soukup, P. Jandera, Adsorption of water from aqueous acetonitrile on silica-based stationary phases in aqueous normal-phase liquid chromatography, J. Chromatogr. A 1374 (2014) 102–111. [10] S.M. Melnikov, A. Höltzel, A. Seidel-Morgenstern, U. Tallarek, Adsorption of water −acetonitrile mixtures to model silica surfaces, J. Phys. Chem. C 117 (2013) 6620 −6631. [11] X.B. Cui, Y. Li, T.Y. Feng, G.X. Sun, L. Lin, Separation of acetonitrile and water by saline extractive distillation, Petrochem. Technol. 36 (2007) 1229−1233. [12] J.B. Zhou, X.B. Cui, B.L. Dong, Y.F. Wang, Z.K. Chen, Separation of acetonitrile and water mixture by batch extractive distillation, Chem. Ind. Eng. 26 (2009) 482−486. [13] J. Acosta, I. Rodríguez, U. Jáuregui, L. Nuevas, E. Pardillo, Recovery of acetonitrile from aqueous waste by a combined process: solvent extraction and batch distillation, Sep. Purif. Technol. 52 (2006) 95−101. [14] J. Fang, J. Liu, C. Li, Y. Liu, Isobaric vapor−liquid equilibrium for the acetonitrile + water system containing different ionic liquids at atmospheric pressure, J. Chem. Eng. Data 58 (2013) 1483−1489. [15] J. Fang, R. Zhao, H. Wang, C. Li, J. Liu, Salting-out effect of ionic liquids on isobaric vapor−liquid equilibrium of acetonitrile-water system, Chin. J. Chem. Eng. 23 (2015) 1369–1373. [16] A.P. Abbott, E.I. Ahmed, R.C. Harrisa, K.S. Rydera, Evaluating water miscible deep eutectic solvents (DESs) and ionic liquids as potential lubricants, Green Chem. 16 (2014) 4156–4161. [17] A.A. Samarov, M. A, M.P. Sokolova, E.N. Popova, A.M. Toikka, Choline chloride based deep eutectic solvents as extraction media for separation of n-hexane-ethanol mixture, Fluid Phase Equilib. 448 (2017) 123–127. [18] M. Francisco, A. Van den Bruinhorst, M.C. Kroon, New natural and renewable low transition temperature mixtures (LTTMs): screening as solvents for lignocellulosic biomass processing, Green Chem. 14 (2012) 2153–2157. [19] A.R. Teles, E.V. Capela, R.S. Carmo, J.A.P. Coutinho, A.J.D. Silvestre, M.G. Freire, Solvatochromic parameters of deep eutectic solvents formed by ammoniumbased salts and carboxylic acids, Fluid Phase Equilib. 448 (2017) 15–21.

B. Sharma et al. / Journal of Molecular Liquids 285 (2019) 185–193 [20] M. Francisco, A.V.D. Bruinhorst, L.F. Zubeir, C.J. Peters, M.C. Kroon, A new low transition temperature mixture (LTTM) formed by choline chloride+lactic acid: characterization as solvent for CO2 capture, Fluid Phase Equilib. 340 (2013) 77–84. [21] D.F. Othmer, R. Gilmont, J.J. Conti, An Adiabatic Equilibrium Still, Ind. Eng. Chem. 52 (1960) 625–628. [22] B. Sharma, N. Singh, T. Jain, J.P. Kushwaha, P. Singh, Acetonitrile dehydration via extractive distillation using low transition temperature mixtures as entrainer, J. Chem. Eng. Data 63 (2018) 2921–2930. [23] J. Acosta, A. Arce, E. Rodil, A. Soto, A thermodynamic study on binary and ternary mixtures of acetonitrile, water and butyl acetate, Fluid Phase Equilib. 203 (2002) 83–98. [24] R.H. Perry, Perrys Chemical Engineering Handbook, 7th ed. McGraw-Hill, New York, 1934. [25] C. Tsonopoulos, An empirical correlation of second virial coefficient, AICHE J. 20 (1974) 263–272.

193

[26] N. Gjineci, E. Boli, A. Tzani, A. Detsi, E. Voutsas, Separation of the ethanol/water azeotropic mixture using ionic liquids and deep eutectic solvents, Fluid Phase Equilib. 424 (2016) 1–7. [27] N.R. Rodríguez, A.S.B. Gonzalez, P.M.A. Tijssen, M.C. Kroon, Low transition temperature mixtures (LTTMs) as novel entrainers in extractive distillation, Fluid Phase Equilib. 385 (2015) 72–78. [28] N.R. Rodriguez, M.C. Kroon, Isopropanol dehydration via extractive distillation using low transition temperature mixtures as entrainers, J. Chem. Thermodynam. 85 (2015) 216–221. [29] Y. Peng, X. Lu, B. Liu, J. Zhu, Separation of azeotropic mixtures (ethanol and water) enhanced by deep eutectic solvents, Fluid Phase Equilib. 448 (2017) 128–134. [30] I. Kim, H.F. Svendsen, E. Børresen, Ebulliometric determination of vapor-liquid equilibria for pure water, monoethanolamine, N-Methyldiethanolamine, 3(methylamino)-propylamine, and their binary and ternary solutions, J. Chem. Eng. Data 53 (2008) 2521–2523.