View metadata, citation and similar papers at core.ac.uk brought to you by CORE Submitted to Journal of Chemical & Engineering Data provided by Repositorio Institucional de la Universidad de Alicante
This document is confidential and is proprietary to the American Chemical Society and its authors. Do not copy or disclose without written permission. If you have received this item in error, notify the sender and delete all copies.
Study of the LLE, VLE and VLLE of the ternary system water + 1-butanol + isoamyl alcohol at 101.3 kPa
Journal: Journal of Chemical & Engineering Data
Manuscript ID je-2018-00308r.R3
Manuscript Type: Article
Date Submitted by the Author: 27-Aug-2018
Complete List of Authors: Saquete, María Dolores; Universitat d'Alacant, Chemical Engineering Font, Alicia; Universitat d'Alacant, Chemical Engineering Garcia-Cano, Jorge; Universitat d'Alacant, Chemical Engineering Blasco, Inmaculada; Universitat d'Alacant, Chemical Engineering
ACS Paragon Plus Environment Page 1 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 Study of the LLE, VLE and VLLE of the ternary system water + 4 1-butanol + isoamyl alcohol at 101.3 kPa 5 6 María Dolores Saquete, Alicia Font, Jorge García Cano* and Inmaculada Blasco. 7 8 University of Alicante, P.O. Box 99, E 03080 Alicante, Spain 9 10 Abstract 11 12 In this work it has been determined experimentally the liquid liquid equilibrium of the 13 water + 1 butanol + isoamyl alcohol system at 303.15K and 313.15K. The UNIQUAC, 14 NRTL and UNIFAC models have been employed to correlate and predict LLE and 15 16 compare them with the experimental data. Additionally, the vapor liquid and vapor 17 liquid liquid equilibria of this system have also been determined using a modified 18 Fischer labodest still. The data obtained have been used to correlate and obtain the 19 binary interaction parameters for UNIQUAC and NRTL. It has been analysed the 20 validity of the models to reproduce the VLE and VLLE data. 21 22 1. Introduction 23 24 The search for sustainable alternative fuels destined for use in industry has, in 25 general, been on the rise due to the growing concerns regarding: the future availability 26 of oil reserves; environmental problems (global warming and climate change) [1]; 27 increasing crude oil prices; the volatility of oil supply; and existing legislation that 28 29 restricts the use of non renewable energy sources and mandates the use of fuel from 30 renewable sources [2]. Several non petroleum based liquid biofuels obtained from 31 biomass may constitute viable alternatives. In this regard, alcohols such as methanol, 32 ethanol and butanol are competitive due to their advantageous physico chemical 33 properties. 34 35 Ethanol is the most widely used biofuel today, but it suffers from several 36 disadvantages that suggest it needs to be replaced. The most important disadvantage 37 is that ethanol produces only about three quarters of the energy obtained from 38 conventional gasoline [3]. Butanol is seen as a potential alternative to ethanol because 39 40 it is less volatile and explosive, has a higher flash point, and lower vapour pressure, 41 which makes it safer to handle. It contains more energy, it is less hygroscopic and can 42 easily mix with gasoline in any proportion. In addition, butanol and gasoline have a 43 similar air to fuel ratio and energy content. Butanol can be used directly or blended with 44 gasoline or diesel without any vehicle retrofit and supplied through existing gasoline 45 pipelines [4]. 46 47 These second generation biofuels are made from the inedible parts of 48 lignocellulosic biomass. The biobutanol production process involves pre treatment and 49 hydrolysis of raw material followed by fermentation of sugars to butanol. The principal 50 51 obstacle is the high production cost which includes: the capital investment and the cost 52 of equipment, raw materials, pre treatment, the enzyme, strain development, recovery, 53 as well as the cost of R&D, sales and marketing of butanol. Nowadays, numerous 54 research groups are working on the cost contributing factors while also analysing the 55 R&D strategies and the latest technologies that will render the process practical and 56 cost effective [5]. 57 58 59 *Corresponding author. Tel.: +34 965903365. E-mail address: [email protected] (J. Garcia-Cano). 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 2 of 21
1 2 3 4 5 Biofuel purification is an important step in the overall production process as it 6 involves the separation of biobutanol from the undesirable impurities, water in 7 particular. During the production of biobutanol by fermentation, or obtained through 8 reaction with bioethanol as a feed, several other by products appear as impurities 9 (such as pentanols) [6]. Our research group has been investigating the viability of using 10 hydrocarbons as entrainers in the dehydration of butanol [7, 8, 9]. Going further into the 11 research of obtaining a biofuel for use as an additive in conventional fuel, it is 12 13 necessary to not only separate the 1 butanol from the water, but also to separate it 14 from the other impurities present in the reaction media. This, in turn, requires obtaining 15 experimental data for the liquid liquid, vapour liquid and vapour liquid liquid equilibria 16 (LLE, VLE, and VLLE, respectively) of these ternary systems. One of the impurities is 17 isoamyl alcohol. 18 19 In the present work, we report experimental VLE and VLLE data for the water + 20 1 butanol + isoamyl alcohol ternary system at a constant pressure of 101.3 kPa, as 21 well as LLE data at different temperatures (303.15 and 313.15 K) and the same 22 pressure. We also discuss the accuracy with which the activity coefficient models can 23 24 predict these various equilibria. Only LLE data for this system have been published 25 previously in the literature [14]. 26 27 Water + 1 butanol + isoamyl alcohol is a type 2 heterogeneous ternary system 28 with two partially miscible pairs: water + 1 butanol and water + isoamyl alcohol. It has 29 two binary heterogeneous azeotropes, water + 1 butanol and water + isoamyl alcohol 30 [10]. 31 32 2. Experimental 33 34 2.1 Chemicals 35 36 Specifications for all chemical compounds used in the experiments are provided 37 in Table 1. Ethanol (used as internal standard), 1 butanol and isoamyl alcohol were 38 used as supplied by the provider without further purification. Ultrapure water was 39 obtained using a Mili Q Plus system. The Karl Fisher titration method was used to 40 determine the water content of the chemicals. 41 42 Table 1. Chemicals employed in this work. 43
44 Chemical Provider Initial purity (mass fraction) water content (mass fraction) Purification method Analysis method 45 a 46 1 butanol Merck >0.995 ≤0.001 None GC 47 Isoamyl alcohol Merck >0.990 ≤0.002 None GCa 48 Ethanol Merck >0.999 <0.001 None GCa 49 a GC, Gas Chromatography 50 51 52 53 2.2 Experimental procedure 54 55 2.2.1 LLE data determination 56 57 58 59 60 ACS Paragon Plus Environment Page 3 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 Mixtures of known overall composition belonging to the heterogeneous region were 4 prepared in glass tubes and sealed with screw caps and Teflon septum. These tubes 5 were submerged partially inside a thermostatic bath maintained at a constant and 6 controlled temperature. They were also shaken vigorously to speed up equilibration. 7 Afterwards, the tubes were left in the bath at the same temperature in order to permit 8 decantation and promote phase separation. The bath temperature was 303.15 K in one 9 10 of the experiments and 313.15K in the other. 11 After the phase separation, weighed samples were extracted from both layers by 12 13 means of syringes and injected into vials. A measured amount of ethanol was added in 14 each of the vials as an internal standard for quantification purposes and to avoid phase 15 splitting when the samples began to cool down outside of the thermostatic bath. An 16 additional sample was taken from each of the organic layers and their water content 17 was checked against the Karl Fischer technique. The Water content determined by the 18 Karl Fischer titration is used to verify the value obtained from the chromatography. 19 20 Each of the vials were analysed by gas chromatography. The organic phases were 21 analysed in a Shimadzu GC14 with a Porapak Q packed column and a thermal 22 conductivity detector (TCD). The oven temperature was set to 463 K, the injector to 23 24 483 K, and the TCD to 483 K. The current was fixed at 100 mA. The carrier gas, 25 helium, was supplied at a flow rate of 30 mL/min. 26 27 The aqueous phases were analysed by means of an Agilent 7820A chromatograph. In 28 both GC systems the same chromatographic phase (Porapak Q) was used. A TCD and 29 a Flame Ionisation Detector (FID) were placed in series just after the column. The 30 helium flow rate was 20 mL/min. The temperatures were 503, 523, 523 and 523 K for 31 the oven, injector, TCD and FID. This chromatograph has been employed In order to 32 analyze the composition of the organic compounds in the aqueous phase whose 33 34 concentrations are low. It possesses not only a TCD but it also has a FID. The FID 35 being a more sensitive detector, it allows a better quantification of the organic 36 compounds. 37 38 In order to quantify the samples’ composition, standards of a known composition, and 39 with ethanol as internal standard, were prepared and analysed by the same procedure 40 used on the samples. 41 42 43 44 2.2.2 VLLE and VLE data determination 45 46 The VLE and VLLE data have been measured by means of a modified Fischer 47 Labodest unit with an ultrasonic probe fitted inside its boiling chamber. The use of 48 ultrasound to obtain equilibrium data when heterogeneous liquid mixtures are present 49 has been studied and shown in a previous paper [11] to be adequate for the intended 50 purpose. The experimental procedure is described in a previous work [7]. 51 52 A liquid mixture (homogeneous or heterogeneous) is brought to the boiling point. The 53 resulting vapour carries some of the liquid phases through the Cotrell pump. After the 54 55 Cotrell pump the liquid phases and their equilibrium vapour are separated in a 56 chamber: the liquid phases exit this chamber through a conduit and returns to the 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 4 of 21
1 2 3 mixing chamber. The vapour goes to a condenser where it is condensed and, 4 subsequently, is also returned to the mixing chamber. Just before the condenser an 5 amount of the vapour is bypassed through a heated tube (to avoid condensation) by 6 means of a peristaltic pump. This tube carries the vapour to the chromatograph where 7 a 6 way valve injects some of it into the chromatograph injector of the Shimadzu 8 GC14A. This chromatograph is connected with the distillation equipment in order to 9 10 analyze the vapor composition on line (to avoid vapor condensation and its phase 11 splitting). This vapour is analysed under the same conditions as the organic phases 12 during the LLE determination. The sample containing the equilibrium liquid can be 13 extracted by means of an electrovalve that lifts the glass rod which seals the tube that, 14 in turn, returns liquid to just before the mixing chamber. Temperatures of the vapour 15 are recorded during the course of the experiment using a Pt 100 with an accuracy of 16 0.006K according to its certificate of calibration (SCALE ITS90) [12]. The pressure is 17 measured and controlled at 101.3 kPa using a High Speed Pneumatic Pressure 18 19 Controller Model CPC3000 by Mensor. 20 21 If the collected liquid sample is heterogeneous, it is treated the same way as in the LLE 22 determination except that the temperature in the thermostatic bath is now set to the 23 boiling point of the collected liquid. 24 25 26 27 3. Results 28 29 3.1 Liquid liquid equilibrium data. 30 31 The experimental data obtained are presented in Tables 2 and 3. The binary liquid 32 liquid equilibrium of water + 1 butanol is reported in a previous work [13]. 33 34 35 36 37 Table 2. Experimental LLE data for the system water + 1-butanol + isoamyl alcohol at 303.15K 38 and at 101.3 kPa. Compositions are reported in units of mole fraction. The relative standard 39 uncertainty in composition is u (x)=u(x)/x= 4% and the standard uncertainty in T is u (T) = 0.1 K 40 r and in P is u (P) =2 kPa. 41 42 Organic phase Aqueous phase 43 44 x water x 1-butanol x isoamyl x water x 1-butanol x isoamyl [13] [13] [13] [13] [13] [13] 45 0.504 0.496 0.000 0.983 0.017 0.000 46 0.458 0.494 0.049 0.984 0.015 0.001 47 0.417 0.481 0.103 0.986 0.013 0.001 48 0.404 0.436 0.159 0.987 0.011 0.002 49 50 0.400 0.385 0.216 0.989 0.010 0.002 51 0.358 0.347 0.295 0.990 0.008 0.003 52 0.352 0.285 0.363 0.990 0.006 0.003 53 0.344 0.219 0.437 0.991 0.005 0.004 54 55 0.341 0.149 0.510 0.992 0.003 0.004 56 0.311 0.078 0.611 0.994 0.001 0.004 57 58 59 60 ACS Paragon Plus Environment Page 5 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 0.299 0.000 0.701 0.995 0.000 0.005 4 5 6 7 8 Table 3. Experimental LLE data for the system water + 1-butanol + isoamyl alcohol at 313.15K. 9 Compositions are reported in units of mole fraction. The relative standard uncertainty in 10 composition is ur(x)=u(x)/x=4% and the standard uncertainty in T is u (T) = 0.1 K and in P is u 11 (P) =2 kPa. 12 13 Organic phase Aqueous phase 14 x water x 1-butanol x isoamyl x water x 1-butanol x isoamyl 15 0.516[13] 0.484[13] 0.000[13] 0.981[13] 0.019[13] 0.000[13] 16 0.469 0.484 0.047 0.986 0.014 0.000 17 18 0.439 0.460 0.101 0.987 0.012 0.001 19 0.416 0.426 0.158 0.988 0.010 0.002 20 0.398 0.382 0.220 0.989 0.009 0.002 21 0.393 0.329 0.278 0.990 0.007 0.002 22 0.406 0.260 0.333 0.991 0.006 0.003 23 24 0.370 0.210 0.421 0.991 0.005 0.003 25 0.336 0.149 0.515 0.992 0.003 0.004 26 0.320 0.079 0.602 0.994 0.001 0.004 27 0.308 0.000 0.692 0.995 0.000 0.005 28 29 30
31 32 1-butanol 33 34 0.0 35 1.0 36 0.1 Exp. data 37 0.9 38 0.2 Binodal curve 0.8 39 Tie lines 40 0.3 0.7 41 42 0.4 0.6 43 44 0.5 0.5 45 0.6 46 0.4 47 0.7 48 0.3 49 0.8 50 0.2 51 L+L 0.9 52 0.1 53 1.0 54 0.0 55 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 56 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 6 of 21
1 2 3 4 5 Figure 1. LLE tie lines at 303.15 K for the system water + 1-butanol + isoamyl alcohol. 6 Compositions reported in units of mole fraction. 7
8 9 1-butanol 10 11 0.0 12 1.0 13 0.1 Exp. data 14 0.9 15 0.2 Binodal curve 0.8 16 Tie lines 17 0.3 0.7 18 19 0.4 0.6 20 0.5 21 0.5 22 0.6 23 0.4 24 0.7 25 0.3 26 0.8 27 0.2 28 L+L 0.9 29 0.1 30 1.0 31 0.0 32 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 33 34 Figure 2. LLE tie lines at 313.15 K for the system water + 1-butanol + isoamyl alcohol. 35 Compositions reported in units of mole fraction. 36 37 38
39 40 The data in Tables 2 and 3 are plotted in Figs. 1 and 2. Figure 3 is a plot of LLE data at 41 42 298.15, 333.15 and 368.15 K, obtained by Zhu et al. [14]. On the one hand, the LLE 43 does not seem very sensitive to the temperature in going from 303.15 to 313.15 K, or in 44 the organic phases, from 298.15 to 368.15 K, as reported by Zhu et al. On the other 45 hand, the results presented by Zhu et al. differ greatly from the ones determined 46 experimentally. Moreover, the LLE data obtained in Ref. [14] for the water + 1 butanol 47 pair do not match any of the other references found in the literature [15]. 48 49 In contrast, we have determined the LLE for the binary pair water + isoamyl alcohol at 50 303.15, 313.15 K and at the boiling temperature (around 369 K) (see Table 2, 3 and 4). 51 Several authors have determined this binary LLE at various temperatures (See Refs. 52 53 [16 – 20]). In Figure 4 we plot these LLE equilibria. It is evident that our experimental 54 points coincide with the other authors’ to within uncertainty. The LLE of water + isoamyl 55 alcohol in the temperature range considered tends to be slightly less heterogeneous 56 when the temperature is raised. 57 58 59 60 ACS Paragon Plus Environment Page 7 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 4 5 6 1-butanol 7 0.0 8 1.0 303K exp 9 0.1 313K exp 10 0.9 11 298K Zhu 0.2 12 0.8 333K Zhu 13 368K Zhu 0.3 14 0.7
15 0.4 16 0.6
17 0.5 18 0.5 19 0.6 20 0.4 21 0.7 22 0.3 23 0.8 24 L+L 0.2 25 0.9 0.1 26 27 1.0 0.0 28 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 29 30 Figure 3. LLE experimental data of water + 1-butanol + isoamyl alcohol at 303.15 and 313.15 31 K. Data obtained by Zhu et al. [14] at 298, 333 and 368 K are also shown. Compositions 32 reported in units of mole fraction. 33 34 35 36 37 38 39 380 Experimental 40 Stephenson [16] 370 41 Arnold [17] 42 360 Ginnings [18] 43 350 Kablukov [19] 44 340 Crittenden [20] 45 46 T (K) 330 47 320 48 310 49 50 300 51 290 52 0 20 40 60 80 100 53 % w isoamyl alcohol 54 55 56 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 8 of 21
1 2 3 Figure 4. Binary LLE for the pair water + isoamyl alcohol at different temperatures and 4 comparison with literature. The composition is reported in percentage by weight of isoamyl 5 alcohol. 6 7 8 3.2 Vapor liquid and vapor liquid liquid equilibrium data 9 10
11 12 Table 4 collects the VLLE data obtained at 101.3 kPa, and Table 5, the VLE 13 data of the homogeneous liquids in the system. 14 15 The thermodynamic L/W Wisniak consistency point to point test [21] has been 16 17 used to check the thermodynamic consistency of the experimental data. The value of 18 the L/W are found within the 0.96 to 1.00 range as can be checked in the 19 supplementary data. This value range shows that the experimental vapor liquid and 20 vapor liquid liquid data are indeed consistent. 21 22 Table 4. VLLE data of water + 1-butanol + isoamyl alcohol at 101.3 kPa. Compositions are 23 reported in units of mole fraction. The relative standard uncertainty in composition is 24 ur(x)=u(x)/x=4% and the standard uncertainty in T is u (T) = 0.06 K and in P is u (P) =0.1 kPa. 25 26 Organic Phase Aqueous Phase Vapour
27 T/K x water x 1-butanol x isoamyl x water x 1-butanol x isoamyl y water y 1-butanol y isoamyl 28 365.73[11] 0.638[11] 0.362[11] 0.000[11] 0.979[11] 0.021[11] 0.000[11] 0.754[11] 0.246[11] 0.000[11] 29 365.93 0.596 0.380 0.024 0.985 0.015 0.000 0.757 0.233 0.011 30 31 365.92 0.546 0.391 0.063 0.986 0.013 0.001 0.778 0.200 0.021 32 366.16 0.525 0.371 0.104 0.987 0.012 0.002 0.769 0.196 0.035 33 366.37 0.499 0.331 0.170 0.988 0.009 0.003 0.776 0.168 0.056 34 366.60 0.487 0.279 0.233 0.989 0.008 0.003 0.777 0.143 0.079 35 36 366.92 0.480 0.194 0.326 0.992 0.005 0.004 0.787 0.101 0.111 37 366.98 0.488 0.171 0.342 0.992 0.004 0.004 0.789 0.093 0.118 38 367.12 0.478 0.153 0.369 0.993 0.003 0.004 0.789 0.080 0.131 39 367.24 0.462 0.141 0.397 0.993 0.003 0.004 0.793 0.072 0.135 40 367.34 0.444 0.126 0.430 0.993 0.003 0.004 0.793 0.063 0.144 41 42 367.62 0.431 0.067 0.503 0.994 0.001 0.005 0.797 0.034 0.169 43 368.06 0.446 0.000 0.554 0.995 0.000 0.005 0.808 0.000 0.192 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Page 9 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 4 5 Table 5. VLE data of water + 1-butanol + isoamyl alcohol at 101.3 kPa. Compositions are 6 reported in units of mole fraction. The relative standard uncertainty in composition is 7 ur(x)=u(x)/x=4% and the standard uncertainty in T is u (T) = 0.06 K and in P is u (P) =0.1 kPa. 8 9 T/K x water x 1-butanol x isoamyl y water y 1-butanol y isoamyl 10 365.90 0.350 0.610 0.040 0.757 0.234 0.009 11 365.97 0.322 0.620 0.058 0.739 0.247 0.013 12 366.18 0.375 0.598 0.028 0.751 0.243 0.006 13 366.24 0.340 0.355 0.305 0.783 0.145 0.072 14 15 366.40 0.350 0.379 0.271 0.780 0.155 0.065 16 366.42 0.327 0.427 0.246 0.773 0.171 0.056 17 366.43 0.280 0.643 0.077 0.737 0.249 0.014 18 366.69 0.302 0.479 0.219 0.764 0.188 0.048 19 366.86 0.281 0.533 0.187 0.765 0.201 0.035 20 21 366.92 0.314 0.341 0.346 0.786 0.137 0.077 22 367.09 0.285 0.328 0.387 0.785 0.130 0.085 23 367.22 0.248 0.315 0.437 0.786 0.123 0.091 24 368.41 0.278 0.674 0.048 0.746 0.246 0.008 25 26 368.42 0.246 0.597 0.157 0.734 0.236 0.030 27 368.44 0.197 0.145 0.658 0.784 0.064 0.152 28 368.52 0.446 0.521 0.032 0.761 0.230 0.009 29 368.60 0.182 0.193 0.625 0.790 0.081 0.129 30 368.86 0.492 0.480 0.028 0.759 0.232 0.009 31 32 369.11 0.230 0.237 0.532 0.790 0.096 0.114 33 369.37 0.175 0.288 0.537 0.782 0.113 0.105 34 370.69 0.262 0.671 0.067 0.722 0.263 0.014 35 371.59 0.246 0.604 0.149 0.721 0.245 0.034 36 372.90 0.220 0.517 0.263 0.717 0.218 0.066 37 38 374.97 0.188 0.437 0.376 0.707 0.194 0.098 39 376.20 0.172 0.378 0.450 0.697 0.177 0.126 40 376.74 0.167 0.297 0.537 0.713 0.141 0.146 41 377.95 0.154 0.211 0.635 0.709 0.105 0.186 42 43 378.45 0.144 0.324 0.532 0.661 0.169 0.170 44 379.55 0.109 0.665 0.227 0.548 0.376 0.076 45 379.77 0.106 0.740 0.154 0.527 0.424 0.049 46 380.19 0.106 0.586 0.308 0.546 0.346 0.108 47 380.57 0.106 0.506 0.388 0.568 0.297 0.135 48 49 381.23 0.119 0.262 0.619 0.628 0.152 0.220 50 381.38 0.086 0.794 0.120 0.460 0.497 0.043 51 381.67 0.105 0.411 0.484 0.577 0.249 0.174 52 382.40 0.111 0.156 0.734 0.619 0.095 0.286 53 382.70 0.108 0.284 0.608 0.578 0.183 0.239 54 55 383.61 0.065 0.843 0.091 0.379 0.585 0.035 56 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 10 of 21
1 2 3 4 5 6 1-butanol 7 0.0 8 1.0 9 0.1 10 0.9 Exp. org. and aq. phase
11 0.2 Vapor phase 0.8 12 Tie lines 13 0.3 Binodal curve 14 0.7 15 0.4 VLLE eq. triangles 16 0.6 17 0.5 0.5 18 0.6 19 0.4 20 0.7 21 0.3 22 0.8 23 0.2 24 L+L 0.9 25 0.1
26 1.0 27 0.0 28 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 29 Figure 5. VLLE data of water + 1-butanol + isoamyl alcohol at 101.3 kPa. Compositions 30 31 reported in units of mole fraction. 32
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Page 11 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 1-butanol 4 0.0 5 1.0 6 0.1 7 0.9 Liquid phase
8 0.2 Vapor phase 9 0.8 Vapor curve 10 0.3 11 0.7 Binodal curve 12 0.4 13 0.6 14 0.5 15 0.5 16 0.6 17 0.4 18 0.7 0.3 19 20 0.8 0.2 21 0.9 22 0.1 23 1.0 24 0.0 25 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 26 Figure 6. VLE data of water + 1-butanol + isoamyl alcohol at 101.3 kPa. Compositions reported 27 in units of mole fraction. 28 29 30 31 Figure 5 is a plot of the data in Table 4. This table contains two binary 32 heterogeneous azeotropes: water+1 butanol and water+ isoamyl alcohol. The data 33 corresponding to the water + 1 butanol azeotrope are reported in a previous work [11]. 34 The experimental heterogeneous binary azeotrope of water + isoamyl alcohol in Table 35 36 4 has been reported by several authors [10]. The data for this azeotrope range from 37 368.25 to 368.35K, and the compositions from 0.810 to 0.828 in mole fraction of water. 38 Our compositions are in agreement with the other authors’ to within uncertainty. 39 Nevertheless, the temperature that we obtain is slightly lower than that reported in the 40 literature. Based on the VLLE data, it is not possible to ensure the existence of a 41 ternary azeotrope in this system. To the best of our knowledge, its existence has not 42 been reported in the literature either. The vapour phases on the vapour curve are all 43 44 richer in water and butanol than their equilibrium liquid counterparts, i.e., the vapour 45 vertex of the equilibrium triangles always points towards the water + 1 butanol side of 46 the composition triangle. 47 48 Figure 6 shows the VLE data of the liquid phases belonging to the 49 homogeneous region (see Table 5). The heterogeneous phase border is also shown. 50 With the exception of a few points, the liquid phases have an equilibrium vapour that 51 belongs to the heterogeneous region. Moreover, the vapour phase of a large fraction of 52 the homogeneous liquid phases lies on the vapour curve, or occurs within a small 53 range of compositions from it. 54 55
56 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 12 of 21
1 2 3 4. Data Correlation. 4 5 The UNIQUAC and NRTL equations have been used to correlate our 6 experimental data, while, for purposes of prediction and comparison, we have used 7 UNIFAC. The experimental data have been correlated in two different ways: on the one 8 hand, with the isothermal LLE data and, on the other, with the isobaric VLE data. The 9 objective functions of these correlations are the same as those employed in a previous 10 paper [13]. 11 12 4.1 LLE correlation 13 14 The experimental isothermal LLE data for the ternary system have been correlated at 15 two different temperatures by means of the above models. The α parameter in the 16 NRTL model has been kept constant at 0.2. This value is generally recommended by 17 18 Dechema for water+ alcohols pair in LLE. The correlation of experimental data was 19 carried out separately at each temperature. The binary interaction parameters (BIPs) 20 calculated at 303.15 K are reported in Table 6, and those calculated at 313.15 K, in 21 Table 7. 22 23 24 25 Table 6. Parameters and mean deviations from the LLE correlation at 303K for water (1) + 1- 26 butanol (2) + isoamyl alcohol (3). Bij (K): binary interaction parameters from the NRTL model. 27 Uij-Uii (K): binary interaction parameters from UNIQUAC. Mean deviations (D) of water and 1- 28 butanol in the organic (1) and aqueous phase (2). 29 α 30 i j Bij (K) Bji (K) Uij Ujj (K) Uji Uii (K) 31 1 2 1495.364 283.9353 0.2 307.9099 21.54459 32 1 3 1625.294 97.21615 0.2 374.7232 53.22150 33 2 3 317.2166 186.2358 0.2 256.8050 274.9637 34 Mean Deviation D x11 D x21 D x31 D x12 D x22 D x32 35 NRTL 0.013 0.008 0.005 0.002 0.001 0.0003 36 UNIQUAC 0.007 0.004 0.003 0.004 0.003 0.002 37 38 39 Table 7. Parameters and mean deviations from the LLE correlation at 313K for water (1) + 1- 40 butanol (2) + isoamyl alcohol (3). Bij (K): binary interaction parameters from the NRTL model. 41 Uij-Uii (K): binary interaction parameters from UNIQUAC. Mean deviations (D) of water and 1- 42 butanol in the organic (1) and aqueous phase (2). 43 44 i j Bij (K) Bji (K) α Uij Ujj (K) Uji Uii (K) 45 1 2 1482.049 299.0967 0.2 257.9664 11.49461 46 1 3 1453.324 91.64391 0.2 253.8073 113.76920 47 2 3 595.8065 920.9486 0.2 317.6709 284.3363 48 Mean Deviation D x11 D x21 D x31 D x12 D x22 D x32 49 NRTL 0.008 0.004 0.004 0.004 0.003 0.003 50 UNIQUAC 0.010 0.006 0.013 0.002 0.001 0.001 51 52 53 Figures 7 and 8 compare the experimental data and the binodal curves 54 calculated by the UNIQUAC and NRTL models, and predicted by the UNIFAC model. 55 56 57 58 59 60 ACS Paragon Plus Environment Page 13 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 4 1-butanol 5 0.0 6 1.0 7 0.1 8 0.9 Exp 9 UNIFAC 0.2 10 0.8 UNIQUAC 11 0.3 NRTL 12 0.7 UNIQUAC global 13 0.4 14 0.6 15 0.5 16 0.5
17 0.6 18 0.4 19 0.7 20 0.3 21 0.8 22 L+L 0.2 23 0.9 24 0.1 25 1.0 0.0 26 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 27 28 Figure 7. LLE data for the system water + 1-butanol + isoamyl alcohol at 303.15 K and binodal 29 curves calculated with UNIQUAC, NRTL and UNIFAC. Compositions are in mole fraction. 30 Parameter from Table 6. 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 14 of 21
1 2 3 4 1-butanol 5 0.0 6 1.0 7 0.1 8 0.9 Exp 9 UNIFAC 0.2 10 0.8 UNIQUAC 11 0.3 NRTL 12 0.7 UNIQUAC global 13 0.4 14 0.6 15 0.5 16 0.5
17 0.6 18 0.4 19 0.7 20 0.3 21 0.8 22 L+L 0.2 23 0.9 24 0.1 25 1.0 0.0 26 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 27 28 Figure 8. LLE data for the system water + 1-butanol + isoamyl alcohol at 313.15 K and binodal 29 curves calculated with UNIQUAC, NRTL and UNIFAC. Compositions are in mole fraction. 30 Parameters from Table 7. 31 32 33 34 As can be seen in Figs. 7 and 8, the UNIFAC model predicts a heterogeneous 35 region that is significantly smaller than the experimental one, especially for the water + 36 isoamyl alcohol pair (the error is 10%). In contrast, the UNIQUAC model reproduces 37 38 the experimental data quite accurately. 39
40 41 4.2 VLLE correlation 42 43 The 38 VLE experimental points from the homogeneous region, the 13 ternary 44 VLLE points obtained (treated as 26 VLE points), together with binary data from the 45 46 bibliography [15], have been correlated using the UNIQUAC and NRTL models. The 47 objective function tries to minimize the sum of the squares of the differences between yi 48 calculated and yi experimental (vapor composition) for each i component for each of 49 the experimental liquid phases. Each data of VLLE correspond to two VLE data in the 50 correlation. The alpha value for NRTL model was regressed in the liquid vapor 51 correlation to try to improve the regression results. 52 53 The fitted UNIQUAC and NRTL BIPs for this system are collected in Table 8. 54 55 Table 8. Parameters and mean deviations from the VLE correlation at 101.3 kPa for water (1) + 56 1-butanol (2) + isoamyl alcohol (3). Bij (K): binary interaction parameters from the NRTL model. 57 58 59 60 ACS Paragon Plus Environment Page 15 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 Uij-Uii (K): binary interaction parameters from UNIQUAC. Mean deviations (D) of temperature 4 (T) and water (1) and 1-butanol (2) in the vapour phase. 5 i J Bij (K) Bji (K) α Uij Ujj (K) Uji Uii (K) 6 water 1 butanol 1468.34 215.427 0.3634 52.86 258.57 7 water isoamyl alcohol 1615.87 361.475 0.3985 224.85 201.33 8 1 butanol isoamyl alcohol 35.5314 35.6741 0.3045 154.35 171.59 9 Mean Deviation D T D y D y D y 10 1 2 3 NRTL 1.45 0.0290 0.0195 0.0117 11 UNIQUAC 1.13 0.0241 0.0229 0.0070 12 13 14 15 16 17 The mean absolute deviations between experimental and calculated vapour 18 phase mole fractions and temperatures are not very high, and, as a result, the vapour 19 20 liquid equilibria are predicted reasonably well. The UNIFAC model also predicts the 21 binodal curve relatively well, except near the water + 1 butanol axis where the model is 22 not able to reproduce the experimental shape. In contrast, the parameters obtained 23 from NRTL and UNIQUAC cannot reproduce the non isothermal binodal curve 24 accurately, as can be seen in Fig. 9. For this reason, the LLE correlation of the 13 25 ternary heterogeneous tie lines (each at a different temperature) has been performed 26 by means of the UNIQUAC model (see Table 9). 27 28 Table 9. Parameters and mean deviations from the LLE correlation at boiling temperatures for 29 water (1) + 1-butanol (2) + isoamyl alcohol (3). Uij-Uii (K): binary interaction parameters from 30 UNIQUAC. Mean deviations (D) of water and 1-butanol in the organic (1) and aqueous phase 31 (2). 32 33 I J Uij Ujj (K) Uji Uii (K) 34 1 2 353.84 73.53 35 1 3 556.32 74.02 36 2 3 210.85 233.11 37 Mean D x D x D x D x D x D x 38 11 21 31 12 22 32 Deviation 0.010 0.005 0.006 0.004 0.003 0.002 39 40 41 42 In the latter case, the heterogeneous region can be predicted with enough 43 accuracy, as illustrated in Fig. 9, which also shows the organic phase of the binodal 44 curve, plotted with the parameters from Table 9. Nevertheless, neither the equilibrium 45 vapour composition nor temperature are well reproduced. In fact, the residue curve 46 map calculated from these binary interaction parameters (see Fig. 10) predicts a binary 47 azeotrope for the 1 butanol + isoamyl alcohol pair as well as a ternary azeotrope that 48 does not exist. 49 50 Additionally a global LLE correlation, using all the experimental LLE data 51 (303.13 K, 313.15 K and at boiling temperature) as input data and UNIQUAC as model. 52 53 The parameters from this correlation are given in Table 10. The mean deviations 54 obtained are higher than with the other parameter set. The organic phases calculated 55 with this parameter set have been included in Fig 7, 8 and 9. The binodal region 56 calculated with Table 10 parameters is similar to the experimental one for 303.15 and 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 16 of 21
1 2 3 313K and at boiling temperature. The vapor curve calculated with these parameters is 4 the farthest from the experimental one. 5 6 7 8 9 10 Table 10. Parameters and mean deviations from the LLE correlation at all the studied 11 temperatures for water (1) + 1-butanol (2) + isoamyl alcohol (3). Uij-Uii (K): binary interaction 12 parameters from UNIQUAC. Mean deviations (D) of water and 1-butanol in the organic (1) and 13 aqueous phase (2). 14 15 I J Uij Ujj (K) Uji Uii (K) 16 1 2 178.42 53.67 17 1 3 171.78 156.06 18 2 3 415.08 415.58 19 Mean D x11 D x21 D x31 D x12 D x22 D x32 20 Deviation 0.041 0.030 0.030 0.0068 0.0055 0.0045 21 22 23 1 butanol 24 0.0 25 1.0 26 0.1 27 0.9 Exp. 28 Exp. vapor curve 0.2 29 0.8 UNIFAC 30 0.3 UNIQUAC (Table 8) 31 0.7 NRTL (Table 8) 32 0.4 UNIQUAC (Table 9) 33 0.6 UNIQUAC (Table 10) 34 0.5 35 0.5 36 0.6 37 0.4 38 0.7 39 0.3 40 0.8 41 L+L 0.2 42 0.9 0.1 43 44 1.0 0.0 45 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 46 Figure 9. Correlation of vapour-liquid-liquid equilibrium data for the system water (1) + 1- 47 butanol (2) + isoamyl alcohol (3) at 101.3 kPa. Heterogeneous region and vapour curve. 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Page 17 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 1 butanol 3 4 0.0 5 1.0
6 0.1 7 0.9 8 0.2 Ternary azeotrope 9 0.8 Binary azeotropes 10 0.3 0.7 11 12 0.4 0.6 13 0.5 14 0.5 15 0.6 16 0.4 17 0.7 18 0.3 19 0.8 20 0.2 21 0.9 22 0.1
23 1.0 24 0.0 25 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 26 Figure 10. Residue curve map at 101.3 kPa, calculated via the UNIQUAC model with the 27 parameters from Table 9. 28 In addition to verifying how the models perform in calculating the heterogeneous 29 30 region, it is convenient to verify how accurately they and their corresponding 31 parameters reproduce the vapor that is in equilibrium with a liquid phase in this system. 32 The K value of 1 butanol (y but/x but) has been calculated for each of the liquid phases 33 in tables 4 and 5. From each model, a vapor composition can be calculated that 34 corresponds to an experimental liquid phase. With the calculated vapor, the calculated 35 K value can be obtained (one in each model). In this way, the error between the 36 experimental K value for 1 butanol and the calculated K value for 1 butanol has been 37 38 calculated. Figure 11 shows the error in K value in each model as a function of the 39 composition of butanol in the experimental liquid phase. As can be seen, none of the 40 models shows an error in K value that is less than 5% for the whole composition range. 41 That means that none of the models can accurately calculate the butanol content in the 42 vapor phase. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 18 of 21
1 2 3 4 error in Kbutanol 100 5 6 80 7 8 60 9 40 10 UNIQUAC (Table 8) 11 20 UNIFAC 12 NRTL(Table 8) 13 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 UNIQUAC (table 9)
14 error % Kbutanol -20 UNIQUAC (Table 10) 15 16 -40 17 18 -60 19 -80 20 x 1-butanol 21 22 Figure 11. % error in K values of 1 butanol in function of the liquid composition (mole fraction) 23 in 1 butanol between experimental data and calculated with the thermodynamic models studied 24 25 Moreover, as can be deduced from figure 11, the parameters in table 8 for the 26 27 NRTL model generally give errors in K values that are lower than the rest of the 28 models. It can also be seen that the parameters in table 9 for UNIQUAC give a larger 29 error for the vapors than the UNIQUAC parameters in table 8, as mentioned earlier. 30 The UNIQUAC parameters from the global LLE correlation in table 10 are the 31 parameters that reproduce the worse the K value. 32 33 5. Conclusions 34 35 The liquid liquid, vapour liquid and vapour liquid liquid equilibria of the system 36 water + 1 butanol + isoamyl alcohol at 101.3 kPa, have been determined at 303.15 and 37 313.15K for the LLE, and at the boiling temperature for the VLE and VLLE. These data 38 show that this system is little affected by temperature within the studied range for the 39 LLE. The VLLE data do not reveal the existence of a ternary heterogeneous azeotrope. 40 41 The VLE of the majority of the studied homogeneous mixtures involve an 42 43 equilibrium vapour that belongs to the heterogeneous region. In addition, many of the 44 determined vapour points lie near the vapour curve. 45 46 The NRTL and UNIQUAC thermodynamic models adequately reproduce the 47 LLE whereas the UNIFAC model predicts a smaller heterogeneous region. 48 49 In contrast, these models could not simultaneously, or adequately, reproduce 50 the binodal curve at the boiling temperature and the vapour curve. 51 52 Regarding the use of butanol as a biofuel, the equilibrium data presented in this 53 paper show that it could be possible to do a separation of the water from the alcohols 54 via azeotropic distillation and liquid liquid extraction. Nevertheless an economic study 55 of these processes should be done in order to assess the viability of this butanol 56 separation step. 57 58 59 60 ACS Paragon Plus Environment Page 19 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 4 5 Supporting Information 6 7 8 “Associated Content: [Results of Wisniak consistency Test, Antoine constants and 9 10 normal boiling points, Properties of pure compounds used, experimental data used as 11 imput, . Internally calculated activity coefficients, Wisniak L W point to point 12 consistency test values] 13 14 15 16 Acknowledgement 17 18 The authors wish to thank the Conselleria d´Educació, Investigació, Cultura i Esport 19 (Generalitat Valenciana) of Spain for the financial support of project AICO/2015/052. 20 21 22 23 References 24 25 [1] Stoeberl, M., Werkmeister, R., Faulstich, M., Russ, W.: Biobutanol from food wastes 26 – fermentative production, use as biofuel and the influence on the emissions. Procedia 27 Food Science 2011, 1, 1867 1874. 28 29 [2] European Parliament, Directive 2009/28/EC of the European Parliament and of the 30 Council of 23 April 2009 on the promotion of the use of energy from renewable sources 31 2009 http://eur lex.europa.eu/eli/dir/2009/28/oj. 32 33 [3] No, S.: Application of biobutanol in advanced CI engines – A review. Fuel 2016, 34 183, 641 658. 35 36 [4] García, V., Päkkilä, J., Ojamo, H., Muurinen, E., Keiski, R.: Challenges in biobutanol 37 production: How to improve the efficiency?. Renewable and Sustainable Energy 38 Reviews 2011,15, 964–980. 39 40 [5] Morone, A., Pandey, R.: Lignocellulosic biobutanol production: Gridlocks and 41 potential remedies. Renewable and Suitable Energy Reviews 2014, 37, 21 35. 42 43 [6] EP 2 679 304 A1. Process for obtaining higher alcohols in the presence of a gallium 44 containing mixed oxide. 01.01.2014. 45 46 [7] Gomis, V., Font, A., Saquete, M., Garcia Cano, J.: LLE, VLE and VLLE data for the 47 water–n butanol–n hexane system at atmospheric pressure. Fluid Phase Equilib. 2012, 48 316, 135 140. 49 50 [8] Gomis, V., Font, A., Saquete, M., Garcia Cano, J.: Liquid–Liquid, Vapor–Liquid, and 51 52 Vapor–Liquid–Liquid Equilibrium Data for the Water–n Butanol–Cyclohexane System 53 at Atmospheric Pressure: Experimental Determination and Correlation. J. Chem. Eng. 54 Data 2013, 58, 3320 3326. 55 56 57 58 59 60 ACS Paragon Plus Environment Submitted to Journal of Chemical & Engineering Data Page 20 of 21
1 2 3 [9] Gomis, V., Font, A., Saquete, M., Garcia Cano, J.: Isothermal (liquid + liquid) 4 equilibrium data at T = 313.15 K and isobaric (vapor + liquid + liquid) equilibrium data 5 at 101.3 kPa for the ternary system (water + 1 butanol + p xylene). J. Chem. 6 Thermodynamics 2014, 79, 242 247. 7 8 [10] Gmehling, J., Menke, J., Krafczyk, J., Fischer, K.: Azeotropic Data. VCH 9 Publishers, Inc. New York (USA) 1994. 10 11 [11] Gomis, V., Ruiz, F., Asensi, J.: The application of ultrasound in the determination 12 of isobaric vapour liquid liquid equilibrium data. Fluid Phase Equilib. 2000, 172, 245– 13 259. 14 15 [12] Magnum, B., Furukawa, G.: U.S. Department of Commerce, National Institute of 16 Standards and Technology: Springfield, 1990. 17 18 [13] Gomis, V., Saquete, M., Font, A., Garcia Cano, J.; Martínez Castellanos, I.: Phase 19 equilibria of the water + 1 butanol + 2 pentanol ternary system at 101.3 kPa. J. Chem. 20 21 Thermodynamics 2018, 123, 38 45. 22 23 [14] Zhu, Z., Liu, Y., Wang, Y.: Liquid liquid equilibrium for the ternary system of 1 24 butanol + 3 methyl 1 butanol + water at different temperatures. Fluid Phase Equilib. 25 2012, 335, 14 19. 26 27 [15] Gmehling, J., Onken, U.: Vapor Liquid Equilibria Data Collection DECHEMA 28 Chemistry Data Series, vol. I, Part 1a, DECHEMA, Dortmund, 1998. 29 30 [16] Stephenson, R., Stuart, J.: Mutual solubility of water and aliphatic alcohols. J. 31 Chem. Eng. Data 1984, 29, 287 290. 32 33 [17] Arnold, V., Washburn, E.: Ternary system isoamyl alcohol + isopropyl alcohol + 34 water at 10, 25 and 40°. J. Phys. Chem. 1958, 62, 1088. 35 36 [18] Ginnings, P., Baum, R.: Aqueous solubilities of the isomeric pentanols. J. Am. 37 Chem. Soc. 1937, 59, 1111 1113. 38 39 [19] Kablukov, I., Malischeva, V.: The volumetric method of measurement of the mutual 40 solubility of liquids the mutual solubility of the systems ethyl ether + water and iso amyl 41 alcohol + water. J. Am. Chem. Soc. 1925, 47, 1553 1561. 42 43 [20] Crittenden, E., Hixson, A.: Extraction of hydrogen chloride from aqueous solutions. 44 Ind. Eng. Chem. 1954, 46, 265 274. 45 46 [21] Wisniak, J.: A new test for the thermodynamic consistency of vapor liquid 47 equilibrium. Ind. Eng. Chem. Res., 1993, 32, 1531 1533. 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Page 21 of 21 Submitted to Journal of Chemical & Engineering Data
1 2 3 4 5 6 7 8 9 10 11 12 13 14 1-butanol 15 16 0.0 17 1.0 18 19 0.1 20 0.9 Exp. org. and aq. phase 21 Vapor phase 22 0.2 0.8 23 Tie lines 24 0.3 25 0.7 Binodal curve 26 VLLE eq. triangles 27 0.4 0.6 28 VLE eq. 29 0.5 30 0.5 31 32 0.6 33 0.4 34 0.7 35 0.3 36 37 0.8 38 0.2 39 0.9 40 0.1 41 42 1.0 43 0.0 44 water 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 isoamyl alcohol 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Paragon Plus Environment