methods of multicomponent distillation...

Azeotropic Distillation

Dr. Jungho Cho, Professor

Department of Chemical Engineering

Dong Yang University

Azeotropic Distillation Slide 2

Limit of Distillation by Azeotrope

Liquid Mole Fraction of Ethanol

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Va

po

r M

ole

Fra

ctio

n o

f E

tha

no

l

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

� Distillation range is restricted by

the azeotropic point.

� Binary azeotropic mixtures, such

as ethanol/water and IPA/water,

can be separated into their pure

components by distillation by the

addition of a third component, so

called the entrainer, which forms

a ternary azeotrope with a lower

boiling point than any binary

azeotrope

Ethanol / Water System


Separation of Azeotropic Mixture

� Shift the Azeotropic Point by Changing Pressure.

� Add the Third Component. � Azeotropic Distillation: Volatile Addition

� Extractive Distillation: Non-volatile Addition


Azeotropic vs. Extractive Distillation (1 of 2)

� Azeotropic distillation � By forming a ternary heterogeneous azeotrope lower than any

other binary azeotropic temperatures, nearly pure ethanol can be obtained as a bottom product in an azeotropic distillation column.

� Ethanol is obtained as a bottom product from an azeotropic distillation column using an entrainer such as benzene or normal pentane.

� Extractive distillation

� By adding a solvent which is exclusively familiar with a wanted component in a feed mixture, a desired component can be obtained in an extractive distillation column overhead.

� Ethanol is obtained as a top product from an extractive distillation with ethylene glycol solvent.


Azeotropic vs. Extractive Distillation (2 of 2)

FEEDFEEDFEEDFEED

RECYCLERECYCLERECYCLERECYCLE UPPERUPPERUPPERUPPER PHASEPHASEPHASEPHASE

LOWERLOWERLOWERLOWER PHASEPHASEPHASEPHASE

PURE ETHANOLPURE ETHANOLPURE ETHANOLPURE ETHANOL

FEEDFEEDFEEDFEED

SOLVENTSOLVENTSOLVENTSOLVENT

PURE ETHANOLPURE ETHANOLPURE ETHANOLPURE ETHANOL

Azeotropic Distillation Extractive Distillation


Principle of Azeotropic Distillation

0.0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.00.0

.1

.2

.3

.4

.5

.6

.7

.8

.9

1.0

WaterBenzene

EthanolP

A

F

W

R D

B

C

V

G

IIII

II

A : 78.07 oC B : 67.99 oC

C : 69.31 oC D : 63.88 oC

Only mixtures in region II will give the desired products of pure ethanol as a bottom product.

Aqueous ethanol can be separated into their pure components by

distillation by the addition of a third component, so called the

entrainer, which forms a ternary heterogeneous azeotrope

with a lower than any other binary azeotropes.


How to Select Entrainer

In this case, pure ethanol cannot be obtained since

G is in region III.

0.0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0

0.0

.1

.2

.3

.4

.5

.6

.7

.8

.9

1.0

WaterWater

Ethanol

Ether

A

F

W

B

C

RD

V

P

G

A : 78.07 oC B : 75.08 oC

C : 73.44 oC D : 69.94 oCII

III

I


Homogeneous Azeotrope (I): EtOH-H2O

Mole Fraction of Ethanol

0.0 0.2 0.4 0.6 0.8 1.0

Tem

per

ature

( oC )

75

80

85

90

95

100

105


Homogeneous Azeotrope (II): EtOH-Benzene

Mole Fraction of Ethanol

0.0 0.2 0.4 0.6 0.8 1.0

Tem

pera

ture

( oC

)

66

68

70

72

74

76

78

80

82


Homogeneous Azeotrope (III): Benzene-H2O


Theory of Azeotropic Distillation (I)

Ethanol / Benzene Benzene / Water Ethanol / Water


Theory of Azeotropic Distillation (II)

Ethanol

H2O Benzene

Ethanol

Benzene

H2O

Highly

interaction

Water is

concentrated on

vapor side.

Dew point

temperature

valley


Theory of Extractive Distillation

Ethanol

EG

H2O

Highly

interaction

Ethanol is

concentrated on

vapor side.

Ethanol

H2O EG


Configuration of Azeotropic Distillation

Feed

Waste water

Concentrated ethanol

Recycle stream

Upper

phase

Lower phase

Pure ethanol Waste water

Concentrator

Azeo Column Stripper

Decanter


Example 1: Ethanol Dehydration

� Feed � Feed Composition

1) Ethanol : 60mole%

2) Water : 40mole%

� Feed Condition

1) Inlet Temperature : 25oC

2) Inlet Pressure : 3.5 bar

� Flow Rate

1) 100 Kg-mole/hr

� Entrainer: Benzene

� Cooling Medium: Mono-ethylene Glycol

1) In/Out Temperature : 10oC /18oC

2) Decanter Operating Temperature : 45oC


Simulation Procedure

� Step 1: Selection of Proper Thermo Model & Construction

of Binodal Curve

� Step 2: Finding Binary & Ternary Azeotropic Points

� Step 3: Construction of Binary & Ternary Residual Curves

� Step 4: Concentrator Simulation

� Step 5: Azeotropic Column Simulation

� Step 6: Azeo + Dryer(Stripper) Column Simulation

� Step 7: Simulation of Overall Azeotropic Distillation Unit

� Step 8: Optimization of Ethanol Dehydration Process


Step 1a: Selection of Proper Thermo Model

� NRTL. This model has up to 8 adjustable binary

parameters that can be fitted to data.

( )[ ]ijijijij

TG τβα +−= exp

−+=

∑∑

∑∑∑

∑

k

kkj

l

ljljl

ij

j

k

kkj

ijj

k

kki

j

jjiji

ixG

Gx

xG

Gx

xG

xG ττ

τγln

2T

c

T

ba

ijij

ijij ++=τ


Step 1b: Construction of Binodal Curve

� Binodal Curve Construction

using NRTL

1) Water

2) Ethanol

3) Benzene

� Temperature = 45oC

� Binary Interaction Parameters

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

WaterBenzene

Ethanol

I J A(I,J) A(J,I) αααα (i,j)

1 2 483.324 195.495 0.291

1 3 2990.025 758.494 0.200

2 3 -60.000 660.00 0.300


Step 2a: Finding Azeotrope (Ethanol + H2O)

M1 F1 M2

GUESS

VAP

L1

L2

LIQ

Stream Name Stream Description

Phase

Temperature Pressure

Flowrate

Composition H2O ETHANOL

C KG/CM2

KG-MOL/HR

VAP

Vapor

78.07 1.033

1.000

0.12 0.88

L1

Liquid

78.07 1.033

1.000

0.12 0.88

L2

Unknown

n/a n/a

n/a

n/a n/a

LIQ

Liquid

78.07 1.033

1.000

0.12 0.88


Result for Step 2

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

WaterBenzene

Ethanol

A

B

C

D

A: 78.07oC, B: 69.77oC

C: 69.31oC, D: 63.88oC

Entrainer Selection

Criteria 1


Step 3: Construction of Residual Curves

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

WaterBenzene

Ethanol

A

B

C

D

I

II

III

Only mixtures in region II will give the desired products of pure ethanol as a bottom product.

Entrainer Selection Criteria 2


Step 3: Feed having composition in region I

Feedstock

Composition Mole %

Water 10.00

Ethanol 10.00

Benzene 80.00

Total Flow 100.00

Top Product

Composition Mole %

Water 50.63

Ethanol 46.57

Benzene 2.80

Total Flow 19.76

Bottom Product

Composition Mole %

Water 0.00

Ethanol 1.00

Benzene 99.00

Total Flow 19.76


Step 3: Feed having composition in region II

Feedstock

Composition Mole %

Water 50.00

Ethanol 20.00

Benzene 30.00

Total Flow 100.00

Top Product

Composition Mole %

Water 20.10

Ethanol 31.59

Benzene 48.31

Total Flow 62.11

Bottom Product

Composition Mole %

Water 99.00

Ethanol 1.00

Benzene 0.00

Total Flow 37.89


Step 3: Feed having composition in region III

Feedstock

Composition Mole %

Water 10.00

Ethanol 70.00

Benzene 20.00

Total Flow 100.00

Top Product

Composition Mole %

Water 24.74

Ethanol 27.23

Benzene 48.03

Total Flow 40.41

Bottom Product

Composition Mole %

Water 0.00

Ethanol 99.00

Benzene 1.00

Total Flow 59.59


Step 4: Concentrator Simulation

Feed

Waste water

Concentrated ethanol

Concentrator

Basis: Feed = 100 Kg-mole/hr

xF = 0.60

Ethanol Mole Balance

F xF = D xD (Nearly Pure Water at

Column Bottom)

( ) ( )( )

18.6888.0

6.0100=

⋅=

⋅=

⋅=

azeo

F

D

F

x

xF

x

xFD


PRO/II Keyword Input for Step 4

UNIT OPERATIONS

COLUMN UID=CONC, NAME=Concentrator

PARA TRAY=25,CHEMDIST=150,DAMP=0.4

FEED 1,8

PRODUCT OVHD(M)=2,68.18, BTMS(M)=3

COND TYPE=TFIX, PRES=1.033, TEMP=45

DUTY 1,1/2,25

PSPEC PTOP=1.379, DPCOLUMN=0.21

ESTIMATE MODEL=CHEM

CESTIMATE(L) 1,0.1152,0.8848/25,1,0

SPEC RRATIO, PHASE=L, VALUE=3.0

SPEC STREAM=3,FRACTION, COMP=1,WET, VALUE=1

VARY DUTY=1,2


Distillation Algorithm Selection (I)

� Inside Out (I/O)

� Relatively Ideal Thermodynamics including Hydrocarbon with Water Decant

� Incorporates sidestrippers into column: No recycle!

� Thermosiphon Reboilers

� Flash Zone Model

� Very forgiving of bad initial estimates

� Fast!

� No VLLE


I/I Column Features

2 phase condenser +

water decant

N-1

2 1

Side Columns

Side Streams

Heater/Cooler

Heat Source/Sink

Multiple Feeds

Kettle and Thermosiphon

Reboilers

Pumparounds

N


Distillation Algorithm Selection (II)

� Chemdist

� Mechanically simple columns, complex thermo

� True VLLE

� Azeotropic and Reactive distillation

� Sidestrippers solved by recycle

� No Pumparounds or Thermosiphons

� More sensitive to bad initial estimates


Distillation Algorithm Selection (III)

� Sure

� Very general: Complex column and thermo

� Use when I/O and Chemdist do not apply

� Newton Method

� Very sensitive to bad initial estimates


Distillation Algorithm Selection (IV)

• Generality

• Total Pumparounds

• VLWE on any tray

• Water draw any tray

• Slow

• Sensitive to initial guesses

• Free water or water draw on trays other than condenser

• Total pumparounds or vapor bypass

Inside/Out (I/O) CHEMDIST

• Very fast

• Insensitive to initial estimates

• Thermo non-ideality

• NO VLLE capability (VLWE at condenser)

• Hydrocarbons

• EOS & Slightly non-ideal LACT Thermo

• Interlinked columns

• Side & main columns solved simultaneously

• Reactive Distillation

• VLLE on any tray

• Highly Non-Ideal Systems

• No Pumparounds

• Side columns solved as recycles

• Non-Ideal Systems

• Mechanically simple columns

• VLLE within column

Unique Features

Strengths

Limitations

Applicability

SURE


Specifications and Variables

� Specifications are constraints to be met by the column.

� Variables are calculated to meet specifications.

� Column always balances equations and unknowns.

� To impose a specification, you must add a variable, otherwise equations and unknowns don’t balance.

� Example: Impose two product specifications by declaring reboiler & condenser duties as variables.


Column Status at Initialization

� Fixed quantities remain at their current values unless you declare them as variables.

� If no specs/variables provided, default status used:

QUANTITY STATUS

Overhead and Bottoms Rates

Side Draw Rates

Duties

Feed Rates

Tray Temperatures

Tray Pressures

Vapor and Liquid Rates

Product Properties (e.g. Viscosity)

Tray Vapor or Liquid Properties

Calculated

Fixed

Fixed

Fixed

Calculated

Fixed

Calculated

Calculated

Calculated


Improper Specification

� 0% methane in crude column bottoms � Infinitely many solutions

� 300 lb-mole/hr propylene in overhead

� No solutions if column feed only 250 mol/hr propylene

� 98% ethanol product

� No solutions if Water-Ethanol Azeotrope present


You Can Help I/O or Chemdist using DAMP

� DAMP reduces iteration step and suppresses oscillation.

� Conventional Columns: DAMP = 1.0 (default)

� Columns with Steam: DAMP = 0.6 - 0.8 � Crude, Vacuum, FCC Main Fractionator

� Less-ideal: DAMP = 0.2 - 0.6 � N Increase number of allowed iterations

� If oscillations persist, use Chemdist


Condenser Options (I)

Overhead � Available in all algorithms

� Partial

� Vapor overhead product

� Liquid product is totally

refluxed to the column.

� Overhead vapor product is

a dew point state at the

overhead reflux operating

temperature.


Condenser Options (II)

Overhead

� Bubble

� Liquid product at bubble point

� Fixed temperature

� Sub-cooled at a specified

temperature

� Delta T below bubble point

� Specify degree of sub-cooling


Condenser Options (III)

Overhead

Side Draw

� Mixed

� Vapor product at dew point

� Liquid product at bubble point


Condenser Options (IV)

� Mixed with Decanter

� Vapor product at dew point

� Overhead vapor sub-cooled

and liquid phase split

� Upper phase refluxed or

produced as a column top

product

� Lower phase decanted

Overhead

Liquid 1

Liquid 2


Reboiler Models

� Most reboilers can be simulated as: � Kettle

� Thermosiphon with Baffles

� Thermosiphon without Baffles


Kettle Reboilers

N-1

Bottom Tray

N

Reboiler BTMS Q

VN-1 LN-2

VN LN-1

BTMS

LN-1

VN

Vapor in Equilibrium with Bottoms


Single Pass (Once Through) Thermosiphon

BTMS

LN-1

VN

Bottom Sump Reboiler

Sump

Baffle

LN-1

VN

BTMS

Bottom

Sump

� Equivalent to a Kettle Reboiler because Bottoms

is in Equilibrium with VN


Circulating Thermosiphon Adds 2 Stages

� Simulate as TS without Baffles.

N-2

Bottom Tray

N-1

Combined Sump

N

Reboiler

BTMS

Q

VN-1 LN-2

RL RF

R V

LN-2

VN-1 R V

RL

RF BTMS

Combined Sump


Circulating Thermosiphon

BTMS

VN-1

Bottom Sump

RF

RV

RL

LN-2

Reboiler Sump

N-2

Bottom Tray

N-1

Reboiler Sump

N

Reboiler

BTMS

Q

VN-1 LN-2

RL RF

R V

� Simulate as TS without Baffle.


Preferential Thermosiphon

BTMS R L

R F

L o

R V

N-2 Bottom Tray

N-1 Reboiler Sump

N Reboiler

Bottom Sump

Q

V N-1 L N-2

BTMS

L N-2

V N-1

Bottom Sump Reboiler

Sump

R F

R V

R L

L O

� Simulate as TS with Baffle.


Step 5a: Azeotropic Column Simulation

Azeo Column

F

F2 R

W

P

V

Composition of Stream F

Component Mole %

Ethanol 81.15

Water 18.85

Flow Rate 73.9 K-mol/hr


Step 5b: Decanter Simulation

V

R

W

Assume OVHD Vapor Composition, V around ternary azeotrope

Mole %

Benzene 53.00

Ethanol 31.00

Water 16.00


Step 5b: Decanter Simulation (Continued)

TITLE PROJ=AZEOTROPE, PROB=FLASH,USER=J.H.CHO

PRINT INPUT=ALL, RATE=M, FRACTION=M, PERCENT=M

DIMENSION METRIC

COMPONENT DATA

LIBID 1,BENZENE/2,ETHANOL/3,WATER

THERMODYNAMIC DATA

METHOD SYSTEM(VLLE)=NRTL, SET=NRTL01

STREAM DATA

PROP STRM=V, TEMP=70, PRES=1.033, RATE(M)=100, & COMPOSITION(M)=1,53/2,31/3,16

UNIT OPERATIONS

FLASH UID=COND, NAME=Condenser, KPRINT

FEED V

PRODUCT L=R, W=W

ISO TEMPERATURE=45, PRESSURE=1.033

END

V (Mole %) R (Mole %) W (Mole %)

Benzene 53.00 73.3072 3.1511

Ethanol 31.00 24.0964 47.9467

Water 16.00 2.5965 48.9022

Flow Rate 100 % 71.05 % 28.95 %


Step 5c: Component Mass Balance around

Azeo Column • Water balance around Azeotropic Column

( ) ( )8115.018115.019.73_ 2 −×+−×= FfeedAzeo

21885.093.13 F+=

( ) ( )2895.0489022.0 ××=VW

Azeo Column

F

F2 R

W

P

V

V⋅= 141572.0

21885.093.13141572.0 FV ⋅+=⋅ - (1)

• Ethanol balance around Azeotropic Column

( ) ( ) ( ) 28115.0479467.02895.0 FV ×=××VF ⋅= 171048.02 - (2)

• Combining Eq. (1) & (2):

4.127=V 79.212 =FK-mol/hr , K-mol/hr

• Benzene flow from the decanter

( ) ( ) ( ) 162.1031511.02895.04.127 =⋅⋅=K-mol/hr


Step 7: Simulation of Overall Azeotropic Unit

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

1

25

CONC

DCNT

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

AZEO

CTRL

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

1

20

STRP

1

2

3

V

R

W

F2

P 7


Step 8: Optimization of Overall Process

Mole fraction Ethanol in Conc Top

0.70 0.75 0.80 0.85 0.90

MM

KJ / K

g-m

ole

140

160

180

200

220

Mole fraction Ethanol in Conc Top

0 .7 0 0.75 0 .8 0 0.85 0 .9 0

MM

KJ /

Kg-m

ole

35

40

45

50

Concentrator, azeotropic tower and stripper reboiler duties vs. conc. EtOH in feed

Concentrator

Azeotropic tower

Stripper

Mo le frac tion E tha no l in C o nc T op

0.70 0 .7 5 0.80 0.85 0.9 0

MM

KJ / K

g-m

ole

200

210

220

230

240

250

Azeotropic tower


Step 8: Objective Function

M o le fr a c t io n E th a n o l in C o n c T o p

0 .7 0 0 . 7 5 0 .8 0 0 .8 5 0 .9 0

To

tal re

bo

iler

du

ties M

M K

J /

Kg-m

ole

EtO

H p

rodu

ct

4 1 0

4 1 5

4 2 0

4 2 5

4 3 0

4 3 5

4 4 0

4 4 5

4 5 0

4 5 5


Conclusion

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

WaterBenzene

Ethanol

A

B

C

D

I

II

III

F

P

R V

W

G

Optimal concentration of ethanol in the feed

to the azeotropic tower is approximately

80.6 mole percent.


Future Works

� Comparison of Two-columns and Three-columns Configuration in Ethanol Dehydration Using Benzene as an Entrainer

� Application of an Environmentally Friendly Entrainer, NC5 to Obtain an Absolute Ethanol in Azeotropic Distillation

� Experimental Works for Ternary LLE Rather Than Each

Binary VLE’s & LLE

� Replacing a New Equation of State Approach with a LACT

Model


Two-Columns Configuration


Performance Comparison between Three-

column & Two-columns Configuration

Three-columns Scheme Two-columns Scheme

Column Diameter (Valve Tray)

Concentrator: 1,370mm

Azeo Tower: 1,680mm

Stripper: 800mm

Azeo Tower: 1,500mm

Stripper: 1,670mm

Reboiler Duties 8.2865 MM Kcal/Hr 9.2203MM Kcal/Hr

Condenser Duties 8.4462 MM Kcal/Hr 9.0236MM Kcal/Hr

Steam Consumption 16.4 Ton/Hr 18.0 Ton/Hr

Cold Utility Consumption 1,027 Ton/Hr 1,127 Ton/Hr

Ethanol Purity 99.9493 wt% 99.9665 wt%

Other Impurities Benzene: 500 ppm (wt)

Water: 0.04 ppm (wt)

Benzene: 335 ppm (wt)

Water: 0.04 ppm (wt)


Performance Comparison between Benzene

& NC5 as an Entrainer

Benzene NC5

Top Press. 1.38 Kg/cm2 3.5 Kg/cm2

Entrainer Benzene NC5

RRATIO 651 Kgmol/hr 439 Kgmol/hr

Theo. Trays 25 25

Feed Tray 5 5

Reboiler Heat Duty

5.2268x106 Kcal/hr

2.6964x106 Kcal/hr

Condenser Heat Duty

3.3290x106 Kcal/hr

2.8552x106 Kcal/hr

Column

Diameter

1600 mm

(valve tray)

120 mm

(valve tray)

Impurities Water : 16ppm

Benz. : 12ppm

Water < 3ppm

NC5 < 1ppm

Ethanol

Recovery

99.99 mole % > 99.99 mole%

Stripping Column

Theo. Trays 25 25

Reboiler Heat Duty 2.2512x106 Kcal/hr

0.1543x106 Kcal/hr

Condenser

Heat Duties

2.0437x106 Kcal/hr

0.1369x106 Kcal/hr

Total Reboiler Duties

7.7580x106 Kcal/hr

0.1369x106 Kcal/hr

Total Condenser

Duties

5.3727x106 Kcal/hr

2.9912x106 Kcal/hr


Azeotropic Distillation

� Although the azeotropic distillation scheme, using NC5, operates at a higher pressure, comparative calculation results through this study show that azeotropic distillation using NC5 as an entrainer is better than that using benzene as an entrainer.


Experimental Works for Ternary LLE

Mole Percent of Benzene

0 10 20 30 40 50 60 70 80 90 100

Mo

le P

erc

ent

of E

tha

no

l

0

10

20

30

40

50

60

70

80

90

100


0 10 20 30 40 50 60 70 80 90 100

Mole

Pe

rcen

t of

Eth

an

ol

0

10

20

30

40

50

60

70

80

90

100


In Renon’s Original Paper

� The following sentence is partially wrong:

“The NRTL equation is applicable to calculate multi-component

VLE and LLE mixtures with only binary parameters.”

“We have to obtain the NRTL constants from the ternary LLE data

for the systems containing ternary components such as ethanol +

water + entrainer system.”


Comment for Azeotropic Simulation

� “The NRTL parameters obtained (regressed) from the

ternary LLE experimental data give a better result than the

NRTL parameters obtained from the binary experimental

data only. However, the above state holds true when you

simulate a system containing only three components in the ternary system. When you simulate a system containing

only ternary components system like ethanol-water-

benzene azeotropic unit, a better approach should be the

ternary LLE interaction parameters.”

- Recommended by Dr. Twu


Vapor Pressures using Improved Alpha Form

600 500

Vapor Pressure, Bar

400 300

ETHANOL

Temperature, K

WATER

BENZENE


Soave’s Original Alpha form

� Good representation of

liquid vapor pressure:

“Proper alpha form”

� Soave’s original alpha

form is wrong since it increases again as

reduced temperature of

hydrogen, Tr approaches

to infinity. Reduced Temperature of Hydrogen

0 20 40 60 80 100A

lpha V

alu

e for

Hyd

rog

en

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

( ) ( )( )[ ]22126992.054226.137464.01

riiTT −−++= ωωα (11)


Requirements for Alpha form

� Requirements for alpha form:

� The α function must be finite and positive for all

temperature. � The α function must equal unity at the critical point.

� The α function must approach a zero value as the temperature approaches infinity.

� The trend from now is to set the coefficients of

alpha function component dependently by

regressing the experimental vapor pressure data

vs. temperatures.


Several Alpha functions

Soave (1972)

Peng-Robinson (1980)

Soave (1979)

Boston-Mathias (1980)

Twu (1988)

Twu-Bluck-Cunningham (1990)

(Recommended by SimSci)

( )[ ]α = + −1 110 5

2

C Tt.

( )[ ]α = + −C C TtC

1 2

2

1 3

( )α = + − +

1 1 12T CC

Tr

r

( )[ ]α = −expC TrC

112

( ) ( )[ ]α = −−

T C Tr

C

rC2 1

1

22 21exp

( ) ( )[ ]α = −−

T C Tr

C C

rC C3 2 2 3

1

11exp


New Alpha Form

� Since 1972, many alpha forms have been

proposed, some better than others.

� PRO/II allows input of parameters for 11 different

forms, including the SIMSCI (TBCC) alpha form.

� This 3 parameter form eliminates the 2 problems

with the Soave alpha for defined components

( ) ( ) ( )[ ]iiii MN

ri

MN

ri TLTT −= − 1exp1α


Mixing Rules

� The accuracy of correlating vapor-liquid equilibrium data

using a cubic equation of state can be improved by

choosing an appropriate mixing rule for calculating a and b

for mixture.

� Expressions for mixing rules a and b are:

∑∑=i j

ijji axxa ∑=i

iibxb


New Mixing Rules

� The advanced mixing rules in Aspen Plus or PRO/II are:

� Quadratic (original SRK & PR)

� PR with Panagiotopolous mixing rule

� RKS with Panagiotopoulos mixing rule

� PR with Wong-Sandler mixing rule

� RKS with Won-Sandler mixing rule

� RKS with modified Huron-Vidal-2 mixing rule

� Kabadi and Danner (SRKKD)

� Modified Panagiotopolous & Reid (SRKM & PRM), 1990.

� Twu, Bluck, Cunningham & Coon (SRKS), 1992.

� These mixing rules are very important whenever the

system is highly non ideal


van der Waals mixing rule (1)

∑∑=i j

ijji axxa ∑=i

iibxb

( )ijjiijkaaa −= 1


Acetone - Water Using SRK

3000

2500

2000

Pressure, mm Hg

1500

1000

500

0.0 0.2 0.4

Mole Fraction Acetone

0.6 0.8 1.0


Panagiotopoulos mixing rule (2)

� In 1986 Panagiotopoulos and Reid proposed an

asymmetric mixing rule containing two parameters for SRK

and PR equations of state (denoted as SRKP and PRP).

� The Panagiotopoulos mixing rule

∑∑=i j

ijji axxa ∑=i

iibxb

( ) ( )[ ]ijiijijjiijxkkkaaa −+−= 1


Fatal Weaknesses of Panagiotopoulos

mixing rule

� The dilution of the mixture with additional components

(reducing all the mole fractions, xi) multiplies the effect of

the second binary parameter kji. In the limit of an infinite

number of components so that all the xi approach zero, the

mixing rules reduces to the original van der Waals mixing rule.

� The mixing rule is not invariant to dividing a component

into a number of identical pseudo-components. For

example, if methane in a mixture is divided arbitrarily into

“alpha” and “beta” methane, the calculated properties of the mixture will be slightly different.


Modified Panagiotopoulos mixing rule (3)

� In 1990, SimSci has modified Panagiotopoulos mixing rule

in a way that eliminate the first of the two flaws. This

mixing rule has four adjustable binary parameters such as

kij, kji, cij and cji.

� The modified Panagiotopoulos mixing rule

( ) ( )

+−+−=

ijc

ji

ijiijijjiij

xx

xkkkaaa 1


Huron-Vidal Mixing Rule (4)

� In 1991, Twu et al. proposed another modified mixing rule

that eliminated both of the inconsistencies of the

Panagiotopoulos-Reid mixing rule noted above slide. For a

binary system, the mixing rule can be expressed in the

following form for a12:

( )

++−=

2121

2

212121xGx

xGHkaaa ijjiij

( )122112 kkH −= ( )121212 exp HG β−=


Requirements for the New Mixing Rule

� Composition dependent mixing rule.

� Does not suffer from the invariant problem associated with

the Panagiotopoulos – Reid mixing rule:

� It is capable of correlating highly nonideal systems (i.e.

hydrocarbon / alcohol)

� It can be used for LLE systems as well.


Twu-Bluck-Cunningham-Coon Mixing Rule

++

−=

jiji

jijijij

jiijxGx

xGH

T

kaaa

2

1

T

kkH

ijji

ij

−=

( )ijijij HG β−= exp

� Binary form:

where


3000

2500

2000

Pressure, mm Hg

1500

1000

500

0.0 0.2 0.4

Mole Fraction Acetone

0.6 0.8 1.0

Acetone - Water Using SRKH


Solubility of Water in Benzene

20 16

Mole Percent of Water

12 8 4 0 0

40

80

120

160

200

120

Temperature, C


Solubility of Benzene in Water

0.06 0.40


0.20 0.00 0

40

80

160

200

120

Temperature, C


Ethanol - Benzene VLE at 45 C

1.0 0.8

Mole Percent of Ethanol

0.6 0.4 150

200

250

350

300

Pressure mmHg

0.2 0.0


Methanol - Heptane at 1 atm.

1.0 0.8

Mole Percent of Methanol

0.6 0.4 50

60

70

90

80

Temperature, C

0.2 0.0

100

110


Vapor Pressures using Improved Alpha Form

600 500

Vapor Pressure, Bar

400 300

ETHANOL

Temperature, K

WATER

BENZENE


Ethanol / Water System


Vapor

Mole

Fra

ction o

f E

thanol

Figure:

Ethanol-benzene system: comparison of predicted x-y

curve (line) with experimental

data (symbols) at 760 mmHg from Ellis, S.R.M. and Clark,

M.B. Chem. Age, India, 12, 377 (1961)


Ethanol / Benzene System


Vapor

Mole

Fra

ction o

f E

thanol

Figure:

Ethanol-benzene system: comparison of predicted x-y

curve (line) with experimental

data (symbols) at 760 mmHg from Ellis, S.R.M. and Clark,

M.B. Chem. Age, India, 12, 377 (1961)


Solubility of Water in Benzene

20 16

Mole Percent of Water

12 8 4 0 0

40

80

120

160

200

120

Temperature, C

Figure:

Benzene-water system: comparison of predicted water

solubility in benzene (curve) with

experimental data (symbols) from DECHEMA Vol. V. Part 1

(1979)


Solubility of Benzene in Water

Figure:

Benzene-water system: comparison of predicted

benzene solubility in water

(curve) with experimental data (symbols) from

DECHEMA Vol. V. Part 1 (1979)

0.06 0.40


0.20 0.00 0

40

80

160

200

120

Temperature, C


Prediction of Ternary System

� The predictions for the ternary system at 1 atm are given

on the next slide.

� Note that only pure and binary data fitting have been done.

No ternary data have been regressed.

� The same results, as predicted using the NRTL equation, are presented on the following slide.

� Note that the results are between EOS with a new mixing

rule and NRTL extremely similar.



Mole

Fra

ction o

f E

thanol

Figure:

Binodal curve, azoetropes, tie lines and residue curve

according to NRTL.

temperatures are reported at 101.325 kPa. Dashed

lines are experimental tie lines at 64oC.

Prediction of Ternary LLE (NRTL, I)



Mole

Fra

ction o

f E

thanol Figure:

Binodal curve, azoetropes, tie lines and residue curve

according to a new mixing

rule (binary parameters regressed from VLE data for

miscible binaries and LLE data for benzene-water).

Temperatures are reported at

101.325 kPa. Dashed lines Are experimental tie lines at

64oC.

Prediction of Ternary LLE (A New EOS, I)



Mole

Fra

ction o

f E

thanol

Figure:

Binodal curve, azeotropes, tie lines and residue curve

according to a new mixing

rule (ethanol-benzene binary parameters has been

adjusted to fit ternary LLE data). Temperatures are

reported at 101.325 kPa.

Dashed lines are experimental tie lines at

64oC.

Prediction of Ternary LLE (A New EOS, II)


Simulation Results

� The predictions of our simulation of this system are given on the

next three slides.

� The NRTL results are given, in addition to the results from a new

mixing rule with binary fitting only and also with a new mixing rule

after adjusting the ethanol / benzene binary interaction

parameters.

� The main difference between the results is that the EOS predicts

somewhat higher slopes for the tie-lines than NRTL predicts when

only binary fitted parameters are used. In actually, the tie0lines

are essentially horizontal.

� We have purposely simulated the main column to operate only in

the VLE range. This is faster from a simulation perspective but

should also be the more stable way to operate these columns.


Azeotropic Distillation Scheme for Test


Simulation Results (I)

(FEED COMPOSITION, MOLE %)

NRTL EOS (binary) EOS (adjusted)

Ethanol 85.00 85.00 85.00

Benzene 0.00 0.00 0.00

Water 15.00 15.00 15.00

(2nd COLUMN RETURN FEED, MOLE %)


Ethanol 77.77 68.54 81.09

Benzene 7.23 16.46 3.91

Water 15.00 15.00 15.00


Simulation Results (II)

(BOTTOM COMPOSITION, MOLE %)


Ethanol 99.87720 99.897656 99.96782

Benzene 0.00007 0.102340 0.03176

Water 0.12273 0.000004 0.00042

(OVERHEAD VAPOR COMPOSITION, MOLE %)


Ethanol 38.53 37.26 39.91

Benzene 49.45 50.98 46.89

Water 12.02 11.76 13.20


Simulation Results (III)

(REFLUX COMPOSITION, MOLE %)


Ethanol 36.32 28.22 39.31

Benzene 57.83 67.57 54.13

Water 5.85 4.21 6.56

(DISTILLATE COMPOSITION, MOLE %)


Ethanol 50.32 57.51 43.58

Benzene 4.68 13.81 2.10

Water 45.00 28.68 54.32


Conclusions

� Advanced alpha functions and advanced mixing rules have

greatly expanded the applicability of the EOS methods to highly

non-ideal systems.

� The results of our simulation of this system using NRTL or the

equation of state approach with a new mixing rule and alpha

function are reasonably close.

� The results of either method could be improved by fitting VLLE

ternary data. We have demonstrated this with a new EOS.

� The EOS approach should be more versatile and should be

superior in multicomponent systems involving light gases and

wide ranges of T and P.


The End of Azeotropic Distillation

The EndQ.

methods of multicomponent distillation...

Documents