INTRODUCTION Separation operation is the most important part in every chemical industry in which the liquid mixtures are separated into pure component

INTRODUCTION
Separation operation is the most important part in every chemical industry in which the liquid mixtures are separated into pure component. Design of separation system which are used for separation process, are depends on the equilibrium data of the system. Distillation is one of the most important unit operations in chemical industries and vapor–liquid equilibrium data is useful in designing columns for distillation, especially fractional distillation. This requires knowledge of either P-x-y or T-x-y data which is generated experimentally.
VLE data can be generated in two ways namely partial VLE data and complete VLE data for binary mixtures and multicomponent in general. And the setup which is used to generate partial VLE data is called ‘Ebulliometer’. In the partial VLE data only liquid composition (x) and equilibrium temperature or pressure are obtained. The vapor composition (y) is obtained by Bubble P and Bubble T calculation. In the complete VLE data both the composition, liquid and vapor are obtained experimentally. VLE data can be generated under isothermal or isobaric condition at sub and above atmospheric pressures. In the isothermal condition P-xy data and in isobaric condition T-xy data is obtained. The VLE data is correlated with thermodynamic models, Like Wilson, Margules, NRTL, UNIQUAC, Van Laar, etc.
Esters and alcohols are mostly used in industry as solvents, extracting agent, and are also used as a raw material. For example, Biodiesel consisting of fatty acid methyl ester produced by esterification of vegetable oil with alcohol. When they are used as a solvent to recover and purify them for future application it is essential to study about substances and their phase behaviours.

The esters of alpha-hydroxy acids synthesized using butanol and amyl alcohol are of industrial importance. Butyl glycolate is produced by direct esterification of Butanol with glycolic acid. Butyl glycolate is mainly used as a paint additive to glossy shine and used in manufacturing of printing inks. It is also used as a flow agent for nitrocellulose lacquers and paints and as a levelling agent for varnishes. Butyl Glycolate is used in resinous and polymeric coating in food coating. Since these reactions are potential candidates for reactive distillation knowledge about the VLE of binary pairs is absolutely necessary.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

The present work focuses on the vapor liquid equilibrium data generation for Butyl Glycolate – n-Butanol system.

The outline of the thesis is given here under:
Chapter 1 includes brief introduction about vapor liquid equilibrium, its importance and about Butanol-Butyl Glycolate system for which data was generated. Chapter 2 includes the literature survey for Butanol-Ester and Butyl Glycolate-Alcohol based systems. Chapter 3 includes the thermodynamic models which are used for VLE modeling. Chapter 4 is the experimental section which includes details of chemicals, experimental setup and experimental procedure and also includes the procedure for modeling based on Excess Gibb’s free energy. Chapter 5 deal with the obtained results and discussions on it. Chapter 6 includes the conclusions drawn from the present study. Chapter 7 proposes the path for future work.

2. LITERATURE SURVEY
Butyl glycolate is produced by direct esterification of Butanol with Glycolic acid. The given reaction shows that synthesis of alpha-hydroxy based esters is done using esterification.

n-Butanol + Glycolic Acid ? Butyl Glycolate + Water
The reaction becomes challenging because the alpha-hydroxy acids are sensitive to temperature and tend to decompose above 348 K. Hence there is either a need to carry out the reaction under vacuum or to use an entrainer that ensures that the boiling conditions are maintained below 348 K (Lewis, 2004). VLE data for Butanol-Ester system has been explained in the literature.

2.1 n-Butanol – Ester based system
Jiang et.al. (2016) generated isobaric vapor liquid equilibrium data for n-Butyl Alcohol and Sec-Butyl Acetate system at 101.3 KPa by using Ellis vapor liquid equilibrium still. The data is reported in Table 2.1 and the T-x-y plot is shown in Fig. 2.1. Wilson, NRTL and UNIQUAC models were used to correlate the generated VLE data. The models showed a good match with the experimental data. This is reflected by the RMSD values NRTL: 0.0652, Wilson: 0.0649 and UNIQUAC: 0.0597. It was observed that the system forms a minimum boiling azeotrope at 383.55 K and 26% n-butanol and 74% of Sec-Butyl acetate (Jiang, 2016). Generated data is given in Table 2.1 and T-X-Y diagram is shown in figure 2.1.

Table 2.1: VLE data for Sec-Butyl Acetate (1) + n-Butyl Alcohol (2) system (Jiang, 2016).

P = 101.3 KPa
T (K) X1 Y1 T (K) X1 Y1
384.27 0.945 0.90 385.15 0.378 0.455
383.71 0.833 0.817 385.57 0.333 0.415
383.55 0.742 0.741 386.02 0.292 0.379
383.61 0.656 0.672 386.54 0.249 0.335
383.80 0.587 0.620 387.04 0.212 0.296
384.05 0.529 0.576 387.61 0.173 0.250
384.36 0.474 0.534 388.26 0.133 0.200
384.71 0.425 0.495 388.97 0.094 0.147

Figure 2.1: T-x-y diagram for Sec- Butyl Acetate (1) and n-Butyl Alcohol (2) (Jiang, 2016).

Susial. et.al. (2015) generated isobaric vapor liquid equilibrium data for 1-Butanol and n- Propyl Acetate system at 100, 150 and 600 KPa by using Dynamic stainless steel Ebulliometer. Van laar, Margules, Wilson, NRTL(?=0.47), UNIQUAC and Redlich- Kister models were used to model the data. Additionally, UNIFAC and ASOG model were used for the prediction VLE data. It was reported that the system forms a minimum boiling azeotrope at 374.65 K (Xaz = 0.981) for 100 KPa, at 387.61 K (Xaz = 0.954) for 150 KPa and at 445.68 K (Xaz = 0.797) for 600 KPa (Susial, 2015). The generated data is reported in Table 2.2 and T-X-Y diagram is shown in Figures 2.2, 2.3 and 2.4.

Table 2.2: VLE data for n-Propyl Acetate (1) + 1-Butanol (2) system (Susial, 2015).

P = 100 KPa
T (K) X1 Y1 T (K) X1 Y1 T (K) X1 Y1
390.83 0 0 379.48 0.368 0.56 375.31 0.712 0.792
389.71 0.024 0.072 379.05 0.395 0.584 375.0 0.766 0.829
389.07 0.042 0.106 378.7 0.413 0.599 374.9 0.817 0.863
388.34 0.061 0.145 378.37 0.433 0.615 374.81 0.859 0.894
387.36 0.083 0.205 378.14 0.462 0.621 374.74 0.9 0.923
385.98 0.122 0.273 377.09 0.538 0.671 374.7 0.931 0.945
383.9 0.177 0.369 376.93 0.549 0.678 374.67 0.967 0.972
381.69 0.264 0.475 376.4 0.598 0.714 374.65 0.981 0.984
380.94 0.295 0.507 376.27 0.610 0.724 374.62 0.991 0.993
380.45 0.322 0.530 376.15 0.617 0.727 374.57 1 1
Table 2.2 (Continued)
379.70 0.361 0.556 375.87 0.652 0.748 – – –
P = 150 KPa
T (K) X1 Y1 T (K) X1 Y1 T (K) X1 Y1
402.59 0 0 392.96 0.436 0.558 388.10 0.869 0.887
401.52 0.025 0.063 392.87 0.448 0.566 387.99 0.885 0.898
400.63 0.054 0.121 392.73 0.453 0.569 387.86 0.901 0.911
399.87 0.073 0.162 392.40 0.48 0.591 387.75 0.917 0.923
399.32 0.095 0.204 392.03 0.508 0.612 387.68 0.930 0.935
398.9 0.112 0.233 391.62 0.540 0.637 387.62 0.942 0.944
398.13 0.140 0.273 391.35 0.572 0.66 387.61 0.954 0.953
396.55 0.215 0.37 390.99 0.603 0.681 387.64 0.966 0.963
395.43 0.272 0.427 390.29 0.662 0.724 387.67 0.978 0.975
394.80 0.316 0.466 389.75 0.714 0.766 387.74 0.990 0.987
393.44 0.398 0.532 388.89 0.783 0.82 387.82 0.996 0.994
393.11 0.423 0.549 388.6 0.820 0.849 387.93 1 1
P = 600 KPa
T (K) X1 Y1 T (K) X1 Y1 T (K) X1 Y1
452.24 0 0 448.33 0.281 0.333 445.95 0.66 0.671
451.38 0.032 0.05 448.09 0.305 0.357 445.82 0.711 0.717
450.94 0.057 0.084 447.77 0.332 0.382 445.73 0.748 0.751
450.62 0.08 0.113 447.58 0.356 0.404 445.68 0.797 0.796
450.19 0.099 0.137 447.42 0.376 0.424 445.68 0.808 0.806
449.89 0.125 0.166 447.25 0.401 0.447 445.69 0.837 0.832
449.71 0.147 0.191 447.04 0.43 0.469 445.7 0.862 0.855
449.37 0.17 0.218 446.9 0.453 0.487 445.73 0.897 0.889
449.09 0.192 0.241 446.75 0.479 0.511 445.76 0.923 0.916
448.8 0.216 0.266 446.57 0.518 0.546 445.8 0.955 0.949
448.63 0.236 0.287 446.3 0.576 0.598 445.84 0.97 0.966
448.48 0.257 0.308 446.01 0.633 0.647 445.89 1 1

Figure 2.2: T-x-y diagram for Propyl Acetate(1) and n-Butanol (2) at 100 KPa (Susial, 2015).

Figure 2.3: T-x-y diagram for Propyl Acetate(1) and n-Butanol (2) at 150 KPa (Susial, 2015).

Figure 2.4: T-x-y diagram for Propyl Acetate(1) and n-Butanol (2) at 600 KPa (Susial, 2015).

Juan. et.al. (2000) generated isobaric vapor liquid equilibrium data for Propyl Ethanoate and Butan-1-ol system at 160 KPa. The data is reported in Table 2.3and T-X-Y diagram is shown in Figure 2.5. UNIFAC and ASOG model were used to correlate the liquid phase activity coefficient. They discovered that UNIFAC model gave the best prediction. It was found that the system forms an azeotrope at 390.13 K (X1=0.969) for 160 KPa (Juan, 2000).

Table 2.3: VLE data for Propyl Ethanoate (1) + Butan-1-ol (2) system (Juan, 2000).

P = 160 KPa
T (K) X1 Y1 T (K) X1 Y1 T (K) X1 Y1
404.10 0 0 397.52 0.267 0.381 392.96 0.593 0.671
402.53 0.055 0.098 397.03 0.302 0.415 392.51 0.645 0.706
401.65 0.084 0.145 396.48 0.334 0.447 391.86 0.710 0.762
401.42 0.095 0.159 395.91 0.367 0.479 391.28 0.782 0.817
400.62 0.127 0.206 395.40 0.399 0.510 390.73 0.863 0.879
399.70 0.163 0.260 394.82 0.439 0.546 390.25 0.948 0.952
398.90 0.201 0.306 394.24 0.483 0.584 390.13 0.969 0.970
398.12 0.237 0.349 393.67 0.527 0.619 390.12 1 1

Figure 2.5: T-x-y diagram for Propyl Ethanoate (1) and Butan-1-ol (2) at 160 KPa (Juan, 2000).

Juan. et.al. (1996) generated isobaric vapor liquid equilibrium data for Propyl Methanoate and Butan-1-ol system at 160 KPa. Margules, Wilson, NRTL, UNIQUAC, Van Laar and Redlich- Kister model were used to correlate the liquid phase activity coefficient. All models gave satisfactory results with same RMSD = 0.004. The result shows that the system forms an azeotrope at 390.13 K (X1=0.969) for 160 KPa (Juan, 1996). The T-X-Y data is given in Table 2.4 and shown in Figure 2.6.

Table 2.4: VLE data for Propyl Methanoate (1) + Butan-1-ol (2) system (Juan, 1996).

P = 160 KPa
T (K) X1 Y1 T (K) X1 Y1 T (K) X1 Y1
403.63 0 0 388.66 0.2605 0.5323 378.06 0.5827 0.7962
401.85 0.0221 0.0846 387.90 0.2799 0.5560 376.93 0.6268 0.8172
400.23 0.0429 0.1485 386.67 0.3134 0.5902 374.15 0.7404 0.8697
398.63 0.0610 0.1981 385.92 0.3339 0.6074 372.93 0.7996 0.8976
396.59 0.0998 0.2835 385.17 0.3519 0.6266 371.49 0.8623 0.9260
393.01 0.1653 0.3978 383.99 0.3870 0.6589 370.15 0.9309 0.9616
391.79 0.1929 0.4382 382.94 0.4181 0.6848 369.40 0.9675 0.9813
391.33 0.2015 0.4480 381.55 0.4584 0.7178 368.46 1 1
390.55 0.2173 0.4717 380.66 0.4897 0.7390 – – –
389.82 0.2339 0.4967 379.07 0.5482 0.7743 – – –

Figure 2.6: T-x-y plot for Propyl Methanoate (1) and Butan-1-ol (2) at 160 KPa (Juan,1996).

Juan. et.al. (1994) generated isobaric vapor liquid equilibrium data for Propyl Propanoate and 1-Butanol system at 101.32 KPa. The data is reported in Table 2.5 and shown in Figure 2.7. The experimental data was modelled using Margules, Wilson, NRTL, UNIQUAC, UNIFAC and ASOG models. The result shows that the system forms an azeotrope at 389.3 K (X1=0.328) for 101.32 KPa (Juan, 1994).

Table 2.5: VLE data for Propyl Propanoate (1) + 1-Butanol (2) system (Juan, 1994).

P = 101.32 KPa
T (K) X1 Y1 T (K) X1 Y1 T (K) X1 Y1
390.85 0 0 389.26 0.3290 0.3265 390.78 0.6791 0.6129
390.74 0.0116 0.0166 389.33 0.3646 0.3553 391.11 0.7161 0.6483
390.60 0.0221 0.0290 389.35 0.3943 0.3792 391.41 0.7410 0.6726
390.37 0.0498 0.0629 389.40 0.4190 0.3946 391.69 0.7620 0.6939
390.22 0.0670 0.0826 389.47 0.4529 0.4207 391.87 0.7786 0.7123
390.12 0.0844 0.1020 389.56 0.4794 0.4455 392.01 0.7894 0.7249
390.11 0.1051 0.1253 389.67 0.5105 0.4703 392.83 0.8470 0.7894
389.81 0.1290 0.1507 389.79 0.5397 0.4936 393.32 0.8758 0.8246
389.63 0.1572 0.1789 389.93 0.5615 0.5142 393.61 0.8994 0.8514
389.35 0.2489 0.2647 390.05 0.5821 0.5282 394.19 0.9652 0.9448
389.27 0.2804 0.2905 390.37 0.6315 0.5698 395.65 1 1
389.27 0.3042 0.3113 390.56 0.6521 0.5878 – – –

Figure 2.7: T-x-y diagram for Propyl Propanoate (1) and 1-Butanol (2) at 101.32 KPa (Juan, 1994).

Casimiro et.al. (2015) generated isobaric vapor liquid equilibrium data for Isopropyl Acetate and 1-Butanol system at 101.32 KPa by using Dynamic recirculating still. NRTL and UNIQUAC model were used to correlate liquid phase activity coefficient and both were showed similar accuracy. The system did not exhibit an azeotrope (Casimiro, 2015). The generated data is given in Table 2.6 and T-X-Y diagram is shown in Figure 2.8.

Table 2.6: VLE data for Isopropyl Acetate (1) + 1-Butanol (2) system (Casimiro, 2015).

P = 101.32 KPa
T (K) X1 Y1 T (K) X1 Y1 T (K) X1 Y1
390.69 0 0 387.19 0.0533 0.2000 369.20 0.6240 0.8128
390.49 0.006 0.0207 385.09 0.1051 0.3149 367.40 0.6791 0.8545
390.39 0.0082 0.0281 383.19 0.1479 0.3783 365.90 0.7577 0.9005
390.09 0.0114 0.0393 381.90 0.1813 0.4224 364.50 0.8353 0.9254
389.89 0.0155 0.0552 380.30 0.2183 0.4764 362.81 0.9239 0.9700
389.49 0.0208 0.0786 378.10 0.2802 0.5488 361.91 0.9766 0.9889
389.09 0.0292 0.1025 376.30 0.3235 0.6016 361.41 1 1
388.59 0.0390 0.1292 375.5 0.3542 0.6218 – – –
388.09 0.0504 0.1584 373.50 0.4542 0.7026 – – –

Figure 2.8: T-x-y diagram for Isopropyl Acetate (1) and 1-Butanol (2) at 101.32 KPa (Casimiro, 2015).

3. THERMODYNAMIC MODELS
The modeling of pure component (P-T) data is done using various models. Moreover, for partial VLE data for binary mixtures is done using GE/RT models. The description of these models is given here under.
3.1 Pure Component Models
Vapor pressure data is important for VLE modeling and design of separation system. Several equations have been proposed for calculating the vapor pressure. They can be majorly divided into two groups. In the first, the parameters are determined by the regression of experimental vapor pressure and in the second, parameters are determined from critical temperature and pressure together with one or two further basic data (McGarry, 1983).

3.1.1 Antoine Equation
The Antoine equation is a class of semi-empirical correlations which describes the relation between vapor pressure and temperature for pure components. The Antoine equation is derived from the Clausius – Clapeyron relation. Each set of constants is valid for a specified temperature range and should not be used much outside of that range (McGarry, 1983). Equation 3.1 is the expression for the Antoine correlation.
Vapor pressure is calculated by using Antoine Equation.

lnP=A-BT+C ………………………… (3.1)
Where; P = Vapor pressure in KPa,
T = Temperature, K and A, B, C = Antoine constants.

3.1.2 Wagner Equation
It was initially derived to describe the vapour pressure of argon and nitrogen from the triple point to the critical temperature. Statistical considerations were used in the development of this equation (McGarry, 1983). Equation 3.2 is the expression for the Wagner correlation.

lnPr= 1TrA 1- Tr+B 1- Tr1.5+C 1- Tr3+D 1- Tr6 …………… (3.2)
Where; P = Vapor pressure in KPa,
T = Temperature, K and A, B, C, D = Wagner constants.

3.1.3 Thomas Equation
This equation is based on the observation that “The ratio of the value of RTdlnPdT for any non-associated compound to the value of the function for any other such compound at the same vapor pressure is constant over a range from a few millimeters of mercury to the critical pressure” (McGarry, 1983). Equation 3.3 is the expression for the Thomas correlation.

P= 253312eX-C, lnX= A-BlnT ………………………… (3.3)
Where; P = Vapor pressure in KPa,
T = Temperature, K and
A, B, C = Thomas constants.

GE Based Models
The behavior of several liquid solutions cannot be adequately expressed in terms of an equation of state. For such mixtures several empirical expressions were developed which are generally termed as Excess Gibbs free energy model.

Excess Gibbs free energy models are commonly used for addressing the non-ideality in the liquid phase. These are based on Wohl’s approach & local composition theory approach. Margules & Van Laar model are included in the former. The latter includes Wilson and NRTL, models.

Margules 2-Suffix Equation
Margules 2-suffix expression is the simplest expression among all model expressions. The relative expressions for the Margules 2-suffix is given in Equations 3.4 and 3.5.

ln?1=Ax12 ………………………… (3.4)
ln?2=Ax22 ………………………… (3.5)
The Parameter A having either positive value or negative value. For the positive value of A the value of Gibbs free energy (GE/RT) is positive and ln?1 and ln?2 values are also positive which shows the positive deviation from ideal solution. For the negative value of A the value of Gibbs free energy (GE/RT) is negative and ln?1 and ln?2 values are also negative which shows the negative deviation from ideal solution behaviour (Rao, 2011).

3.2.2 Margules 3-Suffix Equation
The Margules 3-Suffix equation is extensively used to correlate the activity coefficients, since it is simple and containing two adjustable parameters. The Margules 3-suffix equation is generally used for symmetrical systems (Rao, 2011), where the parameter A12 and A21 are approximately same. The expressions for the Margules 3-suffix is given in Equations 3.6 and 3.7. The Margules parameter can be determined from a single measurement of VLE composition together with the knowledge of pure component saturation pressure. However, the application of this model is restricted to binary systems.
ln?1=x22A12+ 2A21-A12x1 ………………………… (3.6)
ln?2=x12A21+ 2A12-A21x2 ………………………… (3.7)
3.2.3 Wilson Equation
For mixtures in which the components differ from each other in molecular size and the interactions between the like and unlike molecules are different. The expressions are given in Equations 3.8 and 3.9. The advantage of Wilson model is that it predicts the temperature dependency of the activity coefficients. Wilson model can be used for completely miscible or partially miscible systems in the region where only one liquid phase exists (Rao, 2011).
ln?1=-lnx1+?12×2+x2?12×1+?12×2-?21?21×1+x2 ………………………… (3.8)
ln?2=-lnx2+?21×1-x1?12×1+?12×2-?21?21×1+x2 ………………………… (3.9)
3.2.4 Non-Random Two Liquid (NRTL) Equation
NRTL equation is applicable to partially miscible as well as completely miscible systems (Rao, 2011) (Narayanan, 2013). Equations 3.10 and 3.11 represent the excess Gibbs free energy of strongly non ideal and partially miscible systems quite satisfactorily.

ln?1= x22?21G21x1+x2G212+ ?12G12x2+x1G122, G12=exp-? ?12, ?12=b12RT ………. (3.10)
ln?2= x12?12G12x2+x1G122+ ?12G12x1+x2G212, G21=exp-? ?21, ?21=b21RT ………. (3.11)
3.2.5 Van Laar Equation
The Van Laar equation is simple and contains two adjustable parameters and used to correlate activity coefficient. The expression for the Van Laar model is given in Equations 3.12 and 3.13.
ln?1=A1+ABx1x22 ………………………… (3.12)
ln?2=B1+BAx2x12 ………………………… (3.13)
The Van Laar equation is not applicable to systems in which ln? exhibits a maxima or minima (Rao, 2011) and for nonpolar systems (Narayanan, 2013).

4. EXPERIMENTATION AND MODELING SECTION
The present work focuses on VLE data generation for Butanol – Butyl Glycolate system. The detailing of chemicals used for the present work is given in Table 4.1.

Table 4.1: Details of Chemicals
Chemicals Source Purity Analysis Method
n-Butanol Suvidhinath Laboratories 99 % GC
Butyl Glycolate Sigma Aldirich >90 % GC
Cyclopentanol Chemika Reagents >98.5 % GC
GC – Gas Liquid Chromatography
The GC analysis does not show any additional peak other than of all these chemicals hence they were used directly without any purification.

4.1 Experimental methods
The direct experimental determination of vapour liquid equilibrium that means analysing the concentration of both liquid and vapour phase which are in equilibrium with each other. There are several methods available which are classified to find equilibrium relation experimentally (Hala, 1967).

Distillation method,
Static method,
Flow method,
Circulation method.

The distillation method is the oldest and simple method in which small amount of liquid is distilled off from a boiling flask. The major disadvantage of this method is that large amount of liquid is required and major error can be obtaining due to condensation of boiling liquid on the cold surface of the distillation flask at the beginning of the experiment (Hala, 1967).

In static method, binary mixture is charged to a closed vessel, and heated under proper mixing until equilibrium is established. Major disadvantage of this method is that, small change in the pressure and volume can have significant effect on the system due to which it is very difficult to take a sample for analysis without disturbing the equilibrium. This method is generally used for high pressure VLE (Scholten, 1997).

In flow method, mixture of known composition is continuously feed to an equilibrium chamber where it is heated to boiling. After the vapour and liquid reach steady state and equilibrium is confirmed by constant temperature and pressure in the system, samples are analyzed and analyzed. This method may give very accurate results, but requires fairly complicated equipment. Further drawback includes the possibility of long time to achieve equilibrium and large quantity of liquid sample.

The circulation method is the most common and widely accepted method. In circulation method vapors coming off from the boiling mixture are condensed in condenser and this condensate is returned to liquid chamber, creating a continuous cycle. Constant pressure and temperature in the system indicate that pseudo – equilibrium steady state is achieved.

4.2 Equipments and Setup details
The system under study was quite challenging due to butyl glycolate being thermolabile. Hence conventional techniques with the respective experimental setups did not work. Hence several permutation and combinations here invented in VLE data generation for this system.

4.2.1 Differential Ebulliometer made of Glass
Differential Ebulliometer used to generate vapour – liquid equilibrium data is shown in Figure 4.1. The name Ebulliometer comes from Latin word “EBULLIO” – meaning boil or bubble up. The primary components in the Ebulliometer are the boiler, Cottrell pump, thermowell and condenser. The overall experimental setup is shown in Figure 4.2.

Figure 4.1: Detail diagram of differential Ebulliometer (Walas, 1985).

For the VLE data generation chemicals (pure component as well as binary mixture) were fed into the modified Ebulliometer. A drop counter is provided to ensure equilibrium conditions. More over data can be generated with minimum quantity of chemicals requirement. The drop counter facility is very unique and provides additional resource to ensure equilibrium condition along with the temperature. The bottom section of the Ebulliometer was wound with external heating element. The details of the experimental setup are given in Table 4.2.

Figure 4.2: Schematic of Glass Ebulliometer setup.

1. Ebulliometer, 1 a. Feed outlet, 1 b. Boiling chamber, 1 c. Cottrell Pump, 1 d. Drop counter, 1 e. Thermowell, 1 f. Feed inlet 2. Mercury in glass thermometer, 3. External belt heater, 4. U – Tube Mercury manometer, 5. Jacketed internal coil condenser, 5 a. Cooling water inlet, 5 b. Cooling water outlet, 6. SS Ballast tank, 7. Bypass line, 8. Belt Driven oil ring Vacuum pump, 9. Nitrogen cylinder, 10. Dimmer stat.

4.2.1.1 Experimental procedure
Pure components or binary mixtures were fed to the Ebulliometer. First of all, cooling water circulation was started to avoid thermal shock. Then heating rate was gradually increased.

After maintaining of water temperature which was circulating in condenser and heating rate, an equilibrium was confirmed by constant temperature of system and uniform drop rate. Typically, equilibrium temperature was considered after 20-30 min when temperature remains constant at given pressure. After completion of the experiment the Ebulliometer was flushed with acetone followed by nitrogen purging.

Table 4.2: Details of Differential Ebulliometer of Glass Setup.

Temperature Measurement Mercury Thermometer (0 – 250 0C), Accuracy:- 0.1/0.1 0C
Pressure Measurement Mercury manometer (400 to 760 mmHg), Accuracy:- 1 mmHg
Vacuum Pump Belt Driven oil ring Vacuum pump
Cooling media water (40 – 60 0C)
Condenser 24 inch Jacketed internal coil condenser
Ballast 10 lit. SS – 316 setup
Heater 0.5 KW External Belt Heater
4.2.2 Differential Ebulliometer of SS-316:
Owing to the high boiling point of butyl glycolate the differential Ebulliometer of Glass cracked due to high temperature gradient between condensed vapor and boiling chamber. So; a new setup of stainless steel – 316 was developed as shown in Figure 4.3. Design parameters of this Ebulliometer is unchanged and same as that of glass. The details of the experimental setup are given in Table 4.3.
Figure 4.3: Schematic of SS-316 Ebulliometer setup.

The experimental procedure for this setup is also identical to the previous one. A small but a thick glass tube is provided as drop counter.
Table 4.3: Details of Differential Ebulliometer of SS-316 setup.

Temperature Measurement Mercury Thermometer (0 – 250 0C), Accuracy:- 0.1/0.1 0C
Pressure Measurement Mercury manometer (400 to 760 mmHg), Accuracy:- 1 mmHg
Vacuum Pump Belt Driven oil ring Vacuum pump
Cooling media Chilled water (15 – 25 0C) for Secondary condenser,
Water (30 – 50 0C) for Primary Condenser
Condenser 4 inch secondary Annulus condenser, SS-316
12 inch Primary Annulus Condenser with 5 Baffles, SS-316
Ballast 10 lit. SS – 316 setup
Heater 4 External bend heater of 1.5 KW each
Pure component data for n-butanol was successfully generated in Differential Ebulliometer of SS-316. The data is reported in Table 4.4 and shown in Fig. 4.4.

Table 4.4: P-T data for n-Butanol.

P (KPa) T (K)
101.30 390.75
86.62 386.55
80.00 384.15
73.32 382.15
66.61 379.75
59.77 377.35
46.60 371.55
a (P) = 0.5 KPa, u(T) = 0.5 K.

Figure 4.4: P-T Diagram for n-Butanol.

However, P-T data for butyl Glycolate could not be generated in this setup due to decomposition. To resolve this issue, a third setup was designed based on indirect heating concept.

4.2.3 Glass setup with Indirect Heating for Isobaric data
In this a cylindrical glass vessel with tilted thermowell was connected to double coiled glass condenser with O ring vacuum trap joint. System assembly is shown in Figure 4.5. Details of the experimental setup are given in Table 4.5.
Steps for experimental procedure are mention below:
Pure or mixture solution and magnetic stirrer was initially introduced in the glass vessel
Vessel was dipped in oil bath
Condenser was then connected using O ring connector
Cooling water supply and magnetic stirrer were started
Oil bath temperature was set at specific temperature
Sufficient time was given for the system to attain a constant temperature.

Figure 4.5: Schematic Experimental Setup 3.

Table 4.5: Details of experimental Setup 3.

Cooling media Water (40 – 60 0C)
Condenser Double coil condenser
Heater 6 KW External bend heater
Temperature Measurement Mercury Thermometer (0 – 250 0C), Accuracy:- 0.1/0.1 0C
Pressure Measurement Mercury manometer (400 to 760 mmHg), Accuracy:- 1 mmHg
However, a drastic drop in the temperature due to sub-cooling of condensed vapour was observed. To address this problem temperature of cooling water supplied was increased but this lead to vapor losses. Due to all these problems it was decided to generate this data under isothermal condition.

4.2.4 Glass setup with Indirect Heating for Isothermal data
The schematic diagram of experimental set up is shown in Figure 4.6 and the details are given in Table 4.6.
Table 4.6: Details of experimental Setup 4.

Heater 6 KW External bend heater
Temperature Measurement Mercury Thermometer (0 – 250 0C), Accuracy:- 0.1/0.1 0C
Pressure Measurement Mercury manometer (400 to 760 mmHg), Accuracy:- 1 mmHg

Figure 4.6: Schematic Experimental Setup 4.

The experimental procedure is outlined below:
Pure or binary mixture and magnetic stirrer were initially charged in the glass vessel
Sonication was done to make the sample free of dissolved gases
After sonication, vessel was instantly connected to the manometer
It was ensured that the assembly is totally leak proof
The vessel was dipped in oil bath
Oil bath temperature was set at the desired temperature to maintain isothermal condition
Wait till thermometer shows constant pressure at specific temperature
The experimental data is shown in Table 4.7 for pure Butyl Glycolate. P-T diagram is plotted by using the values given in Table 4.7. the P-T plot is shown in Figure 4.7.

Table 4.7: P-T data for Butyl Glycolate.

P (KPa) T (K)
100.20 461.35
95.60 452.65
90.13 443.15
84.62 435.15
78.35 428.35
a (P) = 0.5 KPa, u(T) = 1.0 K.

Figure 4.7: P-T diagram for Butyl Glycolate.

Isothermal VLE data was generated for binary system Butyl Glycolate (1) + n-Butanol (2). The P-x data is given in Table 4.8, 4.9 and 4.10 at different temperature. P-x plots are shown in Fig. 4.8, 4.9 and 4.10.
Table 4.8: Isothermal VLE data for (1) Butyl Glycolate + (2) n-Butanol at 373.15 K.

x1 x2 P (KPa)
0 1 50.04
0.2 0.8 96.80
0.4 0.6 98.40
0.6 0.4 93.93
0.8 0.2 88.43

Figure 4.8: P-x2 diagram for (1) Butyl Glycolate + (2) n-Butanol at 373.15 K.

Table 4.9: Isothermal VLE data for (1) Butyl Glycolate + (2) n-Butanol at 403.15 K.

x1 x2 P (KPa)
0 1 152.72
0.2 0.8 131.28
0.4 0.6 105.43
0.6 0.4 97.16
0.8 0.2 99.5
1 0 45.64

Figure 4.9: P-x2 diagram for (1) Butyl Glycolate + (2) n-Butanol at 403.15 K.

Table 4.10: Isothermal VLE data for (1) Butyl Glycolate + (2) n-Butanol at 424.15 K.

x1 x2 P (KPa)
0 1 263.94
0.4 0.6 159.37
0.6 0.4 107.37
0.8 0.2 105.57
1 0 73.20

Figure 4.10: P-x2 diagram for (1) Butyl Glycolate + (2) n-Butanol at 424.15 K.

Thermodynamic modeling
The Antoine, Wagner and Thomas parameters were determined for pure components. The binary system was modelled using Margules 2-suffix, Margules 3-Suffix, Wilson, NRTL, Van Laar and UNIQUAC. The model parameters were obtained by regression using SOLVER of M.S. EXCEL.

Pure Component Modeling
In this, pure component vapor pressure data was regressed to obtain the Antoine, Wagner and Thomas parameters.

VLE data for n-Butanol was determined at a given pressure. Hence ideally what is determined is the saturation temperature at a set pressure. To model this data to obtain the Antoine parameters, the Equation 3.1 was rearranged (Equation 4.1).

T=BA-lnP-C ……………….. (4.1)
The objective function used for regression was ? (Tpre – Texp)2. The parameter A, B and C were assumed and the objective function was minimized to obtain the final values.

In the case of Thomas and Wagner model, it was difficult to rearrange the equation as Tsat = f(P). Hence regression was done by minimizing the objective function ? (Ppre-Pexp)2.

The deviation of experimental and predicted values were determined by root mean square deviation (RMSD) and average absolute deviation (%AAD) which is shown in Eq. 4.2 and 4.3.

RMSD (T) = i=1n(Tiexp-TiPred)2n ………………………(4.2)
% AAD (T) = i=1n100Tiexp-TipredTiexpn …………………………(4.3)
Where; n = Number of data points,
Tiexp, TiPred = Experimental and Predicted temperature,
VLE data for Butyl Glycolate was determined at a given temperature. Hence ideally what is determined is the saturation pressure at a set temperature. To model this data to obtain the Antoine parameters, the Equation 3.1 was rearranged (Equation 4.4).

PSat=eA-BC+T ……………….. (4.4)
The objective function used for regression was ? (Ppre-Pexp)2. The parameter A, B and C were assumed and the objective function was minimized to obtain the final values.

In the case of Thomas and Wagner model, regression was done by minimizing the objective function ? (Ppre-Pexp)2.

The deviation of experimental and predicted values were determined by root mean square deviation (RMSD) and average absolute deviation (%AAD) which is shown in Eq. 4.2 and 4.3.

RMSD (P) = i=1n(Piexp-PiPred)2n ………………………(4.5)
% AAD (P) = i=1n100 Piexp-PiPred?Piexpn …………………………(4.6)
Where; n = Number of data points,
Piexp, PiPred = Experimental and Predicted pressure,
Modeling of Binary VLE data
Modeling for VLE data was done using Excess Gibbs free energy models like, Margules 2-Suffix, Margules 3-Suffix, Wilson, NRTL and Van Laar. BUBL P calculation was carried out to determine model parameter data reduction. Experimental data comprises of pressure, temperature and composition (x). The degrees of freedom of vapor liquid equilibrium is 2 so two variables must be specified either (liquid or vapor composition) or (temperature or pressure). BUBL P calculation was carried out by minimizing objective function (OF) (Eq. 4.7).

OF = ? (Ppre-Pexp)2 ……………………. (4.7)
The deviation of experimental and predicted values were determined by root mean square deviation (RMSD) and average absolute deviation (%AAD) which is shown in Eq. 4.5 and 4.6.

Modeling Procedure for BUBBLE P: –
The procedure followed for BUPPLE P calculation is as follow:
Given – T, P, xi
Tabulate all the terms with respective model equations in MS EXCEL for Margules 2-Suffix, Margules 3-suffix, Wilson, NRTL and Van Laar.

Guess initial (Margules 2 – Suffix (A), Margules 3 – suffix (A12 , A21), Wilson model (?12 , ? 21), NRTL (b12 , b21, = any value(0.1 to 0.6)), Van Laar (A, B). (First try manually for best guess).

So PTcal , Psi , yi, yi , ?i would be calculated automatically by background equations,
Now, using SOLVER feature
– Target cell : (P)2 = 0
– Variables : Margules 2-Suffix (A),
Margules 3-suffix (A12, A21),
Wilson model (?12, ? 21),
NRTL (b12, b21, = any value (0.1 to 0.6)) and
Van laar (A, B).

– Solve to get parameters.

5. RESULTS AND DISCUSSION
Pure component boiling points were determined for n-Butanol under isobaric conditions and vapor pressures for Butyl Glycolate were determined under isothermal conditions. Binary VLE data for these species was generated under isothermal conditions for selected composition.

Vapour Pressure Modeling
Experimental P – T data for pure species Butyl Glycolate and n-Butanol was regressed to obtain the Antoine, Wagner and Thomas model parameter. These are reported in Table 5.1. Deviation plot with respect to literature data for n-Butanol is shown in Figure 5.1. Deviation plots for Antoine, Wagner and Thomas constants with respect to experimental data are shown in Figure 5.2 and 5.3. Root Mean Square Deviation (RMSD) (Eq. 4.2 and Eq. 4.5) and % Absolute Average Deviation (%AAD) (Eq. 4.3 and Eq. 4.6) of Predicted data with respect to experimental data are reported in Table 5.2.
Table 5.1: Antoine, Wagner and Thomas parameter for Butyl Glycolate and n-Butanol.

Model Parameter Butyl Glycolate n- Butanol
Antoine A 2.1738 4.4057
B 15.0436 372.6236
C -374.5297 -235.4688
Wagner A -40.4425 -28.1914
B 64.5840 47.8539
C -3.1412 -3.7446
D -170.2728 -185.9162
Thomas A 13.2278 15.1724
B 2.0263 2.2020
C -2111.0968 -459.8298
Table 5.2: RMSD and % AAD for pure component, Butyl Glycolate and n-Butanol.

Model Butyl Glycolate n- Butanol
RMSD (P) % AAD (P) RMSD (T) % AAD (T)
Antoine 0.0931 0.0875 0.1093 0.0217
Wagner 0.1196 0.1275 0.0886 0.0628
Thomas 0.1310 0.1288 0.1060 0.0211

Figure 5.1: Comparison of boiling points for n-Butanol with the literature data.

Figure 5.2: Comparison of the model prediction for n-Butanol with the experimental data.

Figure 5.3: Comparison of the model prediction for Butyl Glycolate with the exp. data.

It is observed from the plots that for n-Butanol, Antoine shows ± 0.06%, Wagner shows ± 0.04% and Thomas shows ± 0.05% deviation from the experimental data. This indicates that they are comparable. For the Butyl Glycolate component Antoine shows ± 0.17%, Wagner shows ± 0.18% and Thomas shows ± 0.24% deviation from the experimental data.

5.2 Binary VLE Modeling
Experimental P – T data for Butyl Glycolate + n-Butanol system was regressed to obtain the GE based model parameters. These are reported in Table 5.3. Root Mean Square Deviation (RMSD) (Eq. 4.5) and % Absolute Average Deviation (%AAD) (Eq. 4.6) of predicted data with respect to experimental data is reported in Table 5.4.

Table 5.3: GE based model parameters for the binary system: Butyl Glycolate (1) + n- Butanol (2).

Model Parameter Values at 403.15 K Values at 424.15 K
Margules 2-suffix A 0.3803 -1.0978
Margules 3-suffix A12 2.2831 -4.9567
A21 -12.0168 -7.8629
Wilson ? 12 (J/mol) 1299.2658 -2113.4471
? 21 (J/mol) -245.1306 893.1222
NRTL b12 (J/mol) 19.8102 0.0028
b21 (J/mol) -19.7705 3.3E-05
? 0.11 0.11
Van Laar A 0.76 -1.005
B 0.186 -2.088
Table 5.4: RMSD (P) and % AAD (P) for the binary system: Butyl Glycolate (1) + n- Butanol (2).

Temp. (K) Margules 2-suffix Margules 3-suffix Wilson NRTL Van Laar
RMSD (P) % AAD (P) RMSD (P) % AAD (P) RMSD (P) % AAD (P) RMSD (P) % AAD (P) RMSD (P) % AAD (P)
403.15 1.7387 0.9427 1.1664 0.6780 0.720 0.0230 5.9405 3.8247 3.4E-8 1.6E-8
424.15 2.7797 1.5962 1.5E-5 8E-6 0.760 0.1422 28.073 15.899 0.7621 0.1421
Isothermal P-x-y diagrams
Isothermal P-x data along with Predicted P-y data from Excess Gibbs free energy model along with y are plotted in the form of isothermal P-x-y diagrams at two temperatures which are shown in Figures 5.4 to 5.13. The vapour mole fraction y is obtained from the BUBL P calculation using modified Raoult’s law based on Margules 2-Suffix, Margules 3-Suffix, Wilson, NRTL and Van Laar models. P – x – y diagrams for GE based models along with experimental data are shown in Figures 5.4 to 5.13. Since ri and qi values for Butyl Glycolate were not known, modeling using UNIQUAC could not be done. Experimental P – x data with predicted y values for all the models at all two temperatures is given in Appendix A. The x-y plots for all the models at constant temperature are shown in Figure 5.14 to 5.23. The deviation plots for model prediction and experimental data are shown in Figure 5.24 to 5.33. the plots are based on the mole fraction of n-Butanol.

Figure 5.4: P-x-y diagrams for Margules 2-Suffix along with experimental data at 403.15 K.
Figure 5.5: P-x-y diagrams for Margules 2-Suffix along with experimental data at 424.15 K.

Figure 5.6: P-x-y diagrams for Margules 3-Suffix along with experimental data at 403.15 K.

Figure 5.7: P-x-y diagrams for Margules 3-Suffix along with experimental data at 424.15 K.

Figure 5.8: P-x-y diagrams for Wilson along with experimental data at 403.15 K.

Figure 5.9: P-x-y diagrams for Wilson along with experimental data at 424.15 K.

Figure 5.10: P-x-y diagrams for NRTL along with experimental data at 403.15 K.

Figure 5.11: P-x-y diagrams for NRTL along with experimental data at 424.15 K.

Figure 5.12: P-x-y diagrams for Van Laar along with experimental data at 403.15 K.

Figure 5.13: P-x-y diagrams for Van Laar along with experimental data at 424.15 K.

Figure 5.14: x-y diagrams for Margules 2-Suffix based on n-Butanol at 403.15 K.

Figure 5.15: x-y diagrams for Margules 2-Suffix based on n-Butanol at 424.15 K.

Figure 5.16: x-y diagrams for Margules 3-Suffix based on n-Butanol at 403.15 K.

Figure 5.17: x-y diagrams for Margules 3-Suffix based on n-Butanol at 424.15 K.

Figure 5.18: x-y diagrams for Wilson based on n-Butanol at 403.15 K.

Figure 5.19: x-y diagrams for Wilson based on n-Butanol at 424.15 K.

Figure 5.20: x-y diagrams for NRTL based on n-Butanol at 403.15 K.

Figure 5.21: x-y diagrams for NRTL based on n-Butanol at 424.15 K.

Figure 5.22: x-y diagrams for Van Laar based on n-Butanol at 403.15 K.

Figure 5.23: x-y diagrams for Van Laar based on n-Butanol at 424.15 K.

Figure 5.24: Relative differences of experimental Pressure based on Margules 2-Suffix model for the Butyl Glycolate + n-Butanol binary System at 403.15 K.

Figure 5.25: Relative differences of experimental Pressure based on Margules 2-Suffix model for the Butyl Glycolate + n-Butanol binary System at 424.15 K.

Figure 5.26: Relative differences of experimental Pressure based on Margules 3-Suffix model for the Butyl Glycolate + n-Butanol binary System at 403.15 K.

Figure 5.27: Relative differences of experimental Pressure based on Margules 3-Suffix model for the Butyl Glycolate + n-Butanol binary System at 424.15 K.

Figure 5.28: Relative differences of experimental Pressure based on Wilson model for the Butyl Glycolate + n-Butanol binary System at 403.15 K.

Figure 5.29: Relative differences of experimental Pressure based on Wilson model for the Butyl Glycolate + n-Butanol binary System at 424.15 K.

Figure 5.30: Relative differences of experimental Pressure based on NRTL model for the Butyl Glycolate + n-Butanol binary System at 403.15 K.

Figure 5.31: Relative differences of experimental Pressure based on Van Laar model for the Butyl Glycolate + n-Butanol binary System at 403.15 K.

Figure 5.32: Relative differences of experimental Pressure based on Van Laar model for the Butyl Glycolate + n-Butanol binary System at 424.15 K.

From the plots (Fig. 5.24 to 5.32) Margules 2-Suffix shows less deviation at 403.15 K than at 424.15 K, Margules 3-suffix shows more deviation at 403.15 K than at 424.15 K, Wilson shows less deviation at 403.15 K as compared to deviation at 424.15 K, NRTL shows high deviation and Van Laar shows less deviation at 403.15 K as compared to deviation at 424.15 K.