Separation of Tetrahydrofuran-Ethanol Azeotropic Mixture by Extractive Distillation

Asian Journal of Chemistry; Vol. 25, No. 5 (2013), 2774-2778

http://dx.doi.org/10.14233/ajchem.2013.13870

* KE TANG , PENG BAI, CHONGSHUN HUANG and WEIMING LIU

School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, P.R. China

*Corresponding author: E-mail: [email protected]

(Received: 9 April 2012; Accepted: 23 November 2012) AJC-12452

Tetrahydrofuran and ethanol formed a minimum azeotrope and extractive distillation should be a suitable method to separate their mixture. In this work, a systematic study of the separation of tetrahydrofuran-ethanol mixture with an entrainer by extractive distillation was performed. In order to break the original binary azeotrope, ethylene glycol was chosen as the entrainer. Vapour-liquid equilibrium of tetrahydrofuran-ethanol-ethylene glycol system was determined experimentally and NRTL model was the best one to predict this system. The feasibility analysis of this process was validated by simulation through chemical process simulation software-Aspen Plus and the optimal operating conditions of the extractive distillation process were valuable for industrial applications. Simulation results were then corroborated in a batch experimental packing column. In both the simulation and experiments, tetrahydrofuran, whose purity is more than 99.5 %, was obtained respectively.

Key Words: Tetrahydrofuran, Ethanol, Extractive distillation.

INTRODUCTION EXPERIMENTAL Binary mixtures are commonly used in the pharmaceutical The component mass fraction of raw material is 0.935 and fine chemical industries, form azeotropes or close-boiling tetrahydrofuran and 0.065 ethanol. The purity of ethylene mixtures and they are characterized as behaving like a pure glycol is 0.998 (mass fraction). component when submitted to a distillation process. It is Batch distillation experiments were carried out in a small impossible to achieve the separation of azeotropic mixtures laboratory column to evaluate the performance of the solvent. through conventional methods. Therefore, extractive disti- The column has a total height of 1.2 m and an internal diameter llation is a common process used to separate these kinds of of 25 mm. It is packed with stainless steel θ-rings of Φ 3 mm mixtures with a third component, a salt solution or a solvent × 3 mm. The total packing height is about 30 theoretical plates as the entrainer1. The function of the entrainer is to increase and the total liquid hold up is 30 mL. The bottom is a flask of the relative volatility of the mixture. In this way, the com- 4000 mL with three openings. The solvent is fed to the column ponents will boil separately, which allows the collection of near the top section. The flow rate is controlled by a flow meter. heavier components (solvent and heavy key) at the bottom The reflux ratio is provided at the top of the column by a solenoid of the distillation column and the light key component at the valve. The sketch map is showed in Fig. 1. top2,3. The tetrahydrofuran, ethanol, ethylene glycol and other high Tetrahydrofuran-ethanol mixtures are commonly encoun- boiling organic impurity composition are determined by a GC- tered in the fine chemical and pharmaceutical industries. The SP-6890 gas chromatograph with a FID detector and with a GDX- tetrahydrofuran-ethanol mixture forms a minimum boiling 102 column. The injector and detector temperature are held 210 ºC point azeotrope at a mole composition of 0.905 tetrahydrofu- while the column temperature is set equal to 125 ºC. ran and at a constant pressure of 101.3 kPa4,5. Experiment procedure: 540 mL original tetrahydrofuran- The aim of this work is to study the vapour-liquid equili- ethanol mixture was charged into the bottom. Then it was brium (VLE) of tetrahydrofuran-ethanol-ethylene glycol heated and kept under total reflux for 0.5 h. The condenser system and furthermore to establish industrial operating condi- was subcooled to and the power for boiler heating is 400 W. tions for the extractive distillation process of tetrahydrofuran- When the temperature of the bottom and the condenser was ethanol by using ethylene glycol as the entrainer. steady, ethylene glycol was charged into the column near the Vol. 25, No. 5 (2013) Separation of Tetrahydrofuran-Ethanol Azeotropic Mixture by Extractive Distillation 2775 top section of the column at a speed of 20 mL/min and then could not be thrown away but recycled with the mixture in the kept total reflux 0.5 h again. The purpose of this step was to next process. After 80 min, the concentration of methanol in bring down the concentration of ethanol at the top product. the distillate decreased sharply while the toluene mass fraction After the total reflux, the distillate was withdrawn at a reflux increased. The tetrahydrofuran concentration in transition ratio of 3. The analysis of the distillate composition was distillate cut is high which lead to decreased yield. In the repeated about every 5 min. The temperatures of the top and experiment, the temperature of the top is stable at the qualified the bottom of the column were recorded every 5 min. When tetrahydrofuran withdrawn step. The top temperature is nearly the top temperature exceeded 66 ºC, it showed that the tetra- the temperature of pure tetrahydrofuran. While the solvent is hydrofuran in the raw material had been withdrawn completely. charged into the column, ethanol is carried to the bottom by At this moment, the solvent should be stopped to charge. The ethylene glycol. There is more and more solvent at the bottom, top product will be transition distillate in turn. When ethylene so the bottom temperature rises all the time. glycol began to appear in the top product, this solvent at the bottom was collected for recycle. 1.0 e t a l l i t s i

8 3 d 0.8 n i

9 t n

e tetrahydrofuran n 0.6 o

p ethanol

7 m ethylene glycol o c

e 0.4 h t

6 f o n o

i 0.2 t c a r f s

5 s 0.0 10 a M

0 20 40 60 80 100 Time(min) Fig. 2. Variation of the distillate composition in the experiment 11 4

12 200

180 condenser 3 reboiler 160 ) C

13 º (

2 r 140 u t a r

e 120 1 p m e T 100

1) Heating jacket; 2) Flask; 3) 6-Thermometer; 4) Extractive rectifying 0 20 40 60 80 100 section with insulation; 5) Rectifying section with insulation; 7) Sample Time (min) connection; 8) Condenser; 9) Storage tank of solvent; 10) Distillate receiver; 11) Flowmeter; 12) Solvent inlet; 13) U-type pressure meter Fig. 3. Temperature of reboiler and condenser in the experiment Fig. 1. Experiment equipment RESULTS AND DISCUSSION Experiment results: The experimental results are shown Vapour-liquid equilibrium of tetrahydrofuran- in Figs. 2 and 3. Time 0 started from the qualified distillate ethanol-ethylene glycol system: The ternary vapour-liquid withdrawn step. It can be seen that, at the beginning, the equilibrium of tetrahydrofuran-ethanol-ethylene glycol concentration of tetrahydrofuran in distillate maintained a high system was determined experimentally. Briefly, mixtures of value that is over 99.9 % wt averaged and the yield is about tetrahydrofuran-ethanol with ethylene glycol at 101.3 kPa were 80 %. When processed for 80 min, the tetrahydrofuran mass prepared in a vapour-liquid equilibrium still. After the vapour- fraction at the top decreased while the concentration of ethanol liquid equilibrium was achieved (constant temperature increased gradually. But the solvent in the distillate was close throughout the system), samples of condensed vapour and to zero all the time. This distillation cut was transition distillate. liquid were taken for GC analysis. Each assay was made in As there was lots of tetrahydrofuran in this part of distillate, it duplicate. Fig. 4 shows the pseudo-binary diagram obtained 2776 Tang et al. Asian J. Chem. for a solvent to feed molar ratio (E/F) of 5.Three thermo- Effect of azeotropic mixture feed stage: In Fig. 6, the dynamic models (Wilson, NRTL and UNIQUAC) were effect of azeotrope feed stage on the molar composition of analyzed to determine which was the best model to predict distillate in the extractive column and reboiler energy the system performance. The obtained results show that the consumption is presented. It can be observed that, as the NRTL model presented the highest accuracy. azeotropic feed stage was closer to the reboiler, the mole fraction of tetrahydrofuran in distillate increased; so did the 1.0 reboiler energy consumption. The purity and energy consumption reached a maximum when feed stage is 19. This meant n

a that the biggest separation took place in the rectifying section. r 0.8 u f

o In order to obtain the tetrahydrofuran composition required in r d

y the distillate and decrease the reboiler energy consumption, h a r

t 0.6 the stage near the reboiler should be chosen as the suitable e t

f azeotrope feed stage. o n o i t 0.988 57325

c 0.4 a r f p e l o 0.986 t o

e 57320 m h t

r 0.2 t o a 0.984 p n a a r V u

f 57315

o 0.982 r )

0.0 d W y ( 0.0 0.2 0.4 0.6 0.8 1.0 h Q a

r 0.980 t

e 57310 Liquid mole fraction of tetrahydrofuran t f o

Fig. 4. Pseudo-binary equilibrium curve for tetrahydrofuran-ethanol with n 0.978 o i t ethylene glycol at 101.3 kPa. Thick and thin lines predicted by c 57305 a r NRTL model for the pseudo-binary and binary systems, respectively. f 0.976 e l

experimental data o M 0.974 57300 Simulation of the process: The separation of tetra- 10 12 14 16 18 20 22 24 26 Azeotrope feed stage hydrofuran-ethanol mixture was carried out with the help of Fig. 6. Effect of azeotrope feed stage in the extractive distillation column an entrainer, ethylene glycol. The simulation of the process with ethylene glycol on distillate composition and reboiler energy was performed with the chemical process simulation software- consumption Aspen Plus V7.0. The NRTL model was used to predict the activity coefficients and ideal gas model to predict the fugacity Effect of reflux ratio: Fig. 7 shows that the change of coefficient. The binary interaction parameters of tetrahydro- molar reflux ratio exerted a great infuence on the distillate furan-ethanol, tetrahydrofuran-ethylene glycol and ethanol- composition and reboiler energy consumption. The results ethylene glycol were from "APV70VLE-IG" databank in showed that the purity reached a maximum with the increasing Aspen. molar reflux, because as the reflux rate increased, the solvent The flow sheet of the extractive distillation process is was diluted in the column. The results also showed that the presented in Fig. 5. The process has two columns, one for reboiler energy consumption increased sharply as the molar extractive distillation and the other for solvent recuperation. reflux ratio increased. The liquid brought by reflux should be The azeotropic mixture and the solvent are fed to the first vapourized; therefore, the molar reflux ratio could be as low column, in which the components boil separately which allows as possible to bring down the reboiler energy consumption. the less volatile components to be collected at the bottom and the light key component at the top. The bottom product is fed 0.988 80000

to the second column, in which the solvent is recovered and p o t 0.986 recycled to the extractive distillation column. e 70000 h t t a

n 0.984 a r u

f 60000 o ) r d W y

0.982 ( h Q a r t

e 50000 t f

o 0.980 n o i t c 40000 a r

f 0.978 e l o M Fig. 5. Extractive distillation flow sheet 0.976 30000 1 2 3 4 5 6 Reflux ratio Sensitivity analysis: To establish the operating conditions Fig. 7. Effect of molar reflux ratio in the extractive distillation column for the extractive distillation process, a sensitivity analysis was with ethylene glycol on distillate composition and reboiler energy conducted. consumption Vol. 25, No. 5 (2013) Separation of Tetrahydrofuran-Ethanol Azeotropic Mixture by Extractive Distillation 2777

Effect of solvent to feed ratio (S/F): Solvent to feed ratio TABLE-1 (S/F) exerts a direct effect on the distillate purity. In Fig. 8, OPTIMAL DATA FOR THE DISTILLATION COLUMNS before S/F molar ratio reached to 5, tetrahydrofuran content Parameter Extractive Recovery kept increasing, but after that point, it remained stable. This column column Azeotropic feed flowrate (kmol/h) 1 - occurred because more solvent may enlarge the relative Number of stages 30 10 volatility of the tetrahydrofuran-ethanol mixture; however, the Solvent feed stage 3 - purity of tetrahydrofuran was so high that it is not necessary Azeotrope feed stage 19 5 to increase S/F sequentially. Rising S/F increased the liquid Solvent feed temperature (ºC) 25 - flow in the reboiler. Although the liquid brought by the solvent Azeotrope feed temperature (ºC) 25 - remained in liquid phase in column, it may absorb heat there. Mole fraction of tetrahydrofuran in azeotropic 0.905 - Taking into consideration the situation that high S/F ratios feed Solvent feed flow rate (kmol/h) 5 - increased the energy consumption in reboiler, S/F should not Mole fraction of ethylene glycol in solvent feed 1 - be too large. Reflux ratio 2 3

1.01 120000 Simulation results: The simulation results were presented

p 1.00

o in Table-2. It was shown that the obtained tetrahydrofuran mole t 105000 e

h 0.99

t composition at the top of the extractive column was 0.9963 t a and the ethanol mole composition at the recovery column top n 0.98 90000 a r u

f was 0.9644. Meanwhile, the ethylene glycol which was 1 in o r 0.97 ) d W y 75000 mole composition was obtained at the bottom of the recovery ( h Q a 0.96 r t column. e t f

o 0.95 60000 n o i

t TABLE-2

c 0.94 a

r SIMULATION RESULTS FOR THE PROCESS FLOWSHEET

f 45000 e l 0.93 o C4H8O C2H5OH SOLVREC M 0.92 30000 T (ºC) 65.86 77.1 197.8 0 2 4 6 8 10 Mole flow kmol/h 0.905 0.095 5 Solvent to feed molar ratio Mole fraction Fig. 8. Effect of S/F molar ratio in the extractive distillation column with Tetrahydrofuran 0.9963 0.0356 0 ethylene glycol on distillate composition and reboiler energy Ethanol 0.0036 0.9644 0 consumption Ethylene glycol 0.0001 0 1.000 Effect of entrainer feed stage: The results shown in Fig. 9 allow establishing a maximum in the distillate molar fraction In Fig. 10, the temperature profile in the extractive disti- for all the tested conditions when the solvent is fed on stage 3. llation column showed a temperature change at stage 3 due to As the solvent feed stage approached to the condenser tetra- the entrainer inlet, whereas at stage 19 the temperature had a hydrofuran content went up to a maximum, after which point minimal decrease due to the azeotropic mixture feeding. A it decreased. The reason why this decrease occurred was that significant temperature increase was observed at stages 29 and the vapourization of the ethylene glycol entered the column, 30 due to the reboiler proximity. which became part of the vapour that flowed up to the condenser 180 and that was withdrawn as distillate.

Based on the sensitivity analysis above, the best configu- 160 ration and operating conditions of the columns was established and presented in Table-1. 140 ) C º ( 0.98750 58000 e r 120 u t p a r o t 0.98625 e e p h t

m 100

t 57800 e a T

n 0.98500 a r u f 80 o r d )

y 0.98375 57600 W h ( a r Q t

e 60 t f

o 0.98250 n o i 57400 t 0 5 10 15 20 25 30 c a r f 0.98125 Extrctive distillation column stage e l o

M Fig. 10. Extractive distillation column temperature profile 0.98000 57200 2 3 4 5 6 Entrainer feed stage Fig. 11. presents the liquid and vapour molar flow rate Fig. 9. Effect of solvent feed stage of the distillation column with ethylene profiles in the extractive column. The liquid molar flow glycol on distillate composition and reboiler energy consumption exposed two significant changes due to the solvent and the 2778 Tang et al. Asian J. Chem. azeotropic mixture feed at stages 3 and 19, respectively. Mean- while, the vapour molar flow rate remained constant along 1.0 tetrahydrofuran the column, except at stage 3 where the solvent was fed. This ethanol change was caused by the liquid-phase vapourization due to 0.8 ethylene glycol the ethylene glycol inlet temperature.

n 0.6 o i

10 t c a r 9 f e

l 0.4 8 o M

7 0.2 6 ) h / l o 5 0.0 m k ( 4 w o l 0 5 10 15 20 25 30 F 3 Extractive distillation column stage 2 Liquid Fig. 13. Liquid-phase concentration profiles of tetrahydrofuran, ethanol and Vapor 1 ethylene glycol in the extractive distillation column. 0

0 5 10 15 20 25 30 Conclusion Extrctive Column Stage Ethylene glycol was chosen as the entrainer to achieve Fig. 11. Liquid and vapour molar flow rate profiles in the extractive the separation of the azeotropic mixture tetrahydrofuran- distillation column ethanol by extractive distillation. NRTL model was the best model to predict vapour-liquid equilibrium of tetrahydrofuran- Fig. 12 showed that tetrahydrofuran concentration kept ethanol-ethylene glycol system. The feasibility of the process decreasing, from the top of the column to the bottom. Ethanol was confirmed via simulation and experiments. Sensitivity composition kept increasing, from the top of the column to analysis was done to obtain the best conditions and configu- the bottom. The ethylene glycol composition was approaching ration for extractive and recovery columns. The temperature, to zero, from the top of the column to the bottom. The main composition and molar flow profiles obtained for the extractive product at the top of the column was tetrahydrofuran and at distillation column were consistent. In both the simulation and the bottom the main components were ethanol and ethylene experiments, high purified tetrahydrofuran and ethanol were glycol. Meanwhile, in Fig. 13, for the liquid molar composition obtained. profiles, tetrahydrofuran and ethylene glycol had an important change at stage 3, in which the tetrahydrofuran composition REFERENCES decreased rapidly and ethylene glycol increased due to the 1. B. Kotai, P. Lang and G. Modla, Chem. Eng. Sci., 62, 6816 (2007). entrainer inlet. The feed of azeotropic mixture at stage 19 2. E. Batista and A. Meirelles, J. Chem. Eng. Jpn., 3, 45 (1997). caused an increase in tetrahydrofuran composition, whereas a 3. P. Langston, N. Hilal, S. Shingfield and S. Webb, Chem. Eng. Process., decrease in ethylene glycol composition was observed. The 44, 345 (2005). tetrahydrofuran composition decreased along the column and 4. Y. Yoshikawa, A. Takagi and M. Kato, J. Chem. Eng. Data, 25, 344 (1980). at the bottom only ethylene glycol and ethanol were present. 5. B. Erwln and G.R.S. Andreas, J. Chem. Eng. Data, 29, 28 (1984). Ethanol exposed only one significant change at stage 19 where the binary mixture was fed.

1.0

0.8

n 0.6 tetrahydrofuran o i

t ethanol c a

r ethylene glycol f e

l 0.4 o M

0.2

0.0

0 5 10 15 20 25 30 Extrctive distillation column stage Fig. 12. Vapour-phase concentration profiles of tetrahydrofuran, ethanol and ethylene glycol in the extractive distillation column