<<

Solutions for R&D to Design PreFEED Effect of on Separation Operation

April 30, 2012 PreFEED Corporation Hiromasa Taguchi

Solutions for R&D to Design 1

PreFEED Introduction

Generally, it is known that, for distillation, separation tends to be enhanced by lower . For two components, the ease of distillation separation can be judged from the value of the relative (α).

Here, we will take typical substances as examples and examine the effect of pressure changes on their relative volatilities.

Solutions for R&D to Design 2 PreFEED Ideal Solution System

The (α) is a ratio of - equilibrium ratios and can be expressed by the following equation:

K1 y1 y2   K1  , K2  K2 x1 x2 In the case of an ideal solution system , using Raoult’s law, the relative volatility can be expressed as a ratio of vapor pressures. Py  x Po 1 1 1 Raoult’s law Py  x Po o 2 2 2 K1 P1    y Po o K  1  1 K2 P2 1 x1 P o y2 P2 K2   x2 P Solutions for R&D to Design 3

PreFEED - System

As an example of an ideal solution system , let’s consider a methanol-ethanol binary system. α is calculated by obtaining vapor pressures from the Antoine constants in the table below.

0 B 2.8 ln(Pi [Pa])  A  2.6 C T[K] 2.4 ABC 2.2 Methanol 23.4803 3626.55 -34.29 2 Ethanol 23.8047 3803.98 -41.68 1.8 As the (saturated 1.6 pressure) decreases, the value of 1.4 α increases. Relative α : [-] Volatility 1.2 From this, it can be seen that the 1 0 20406080100lower the operation pressure, the Temperature : [℃ ] easier the separation. Next, we will examine the reasons for which Antoine Constants: “Kagaku Kougaku Binran revised 7th ed.” separation becomes easier. (in Japanese) (P12-13)

Solutions for R&D to Design 4 PreFEED Clausius-Clapeyron Equation

Clausius-Clapeyron equation (*) is a formula that is the theoretical basis of the Antoine equation. A linear relationship can be obtained by plotting the logarithm of the ln(P0) versus the inverse of the absolute temperature 1/T. At this time, the slope of the straight line is obtained by dividing the of by the constant. 14 0 dP PH v 1/T vs ln(P ) 13  2 dT RT 12

dP H v dT 11  2 ) : [Pa] : ) P R T 0 10 dP H dT ln(P 9  v   2 8 P R T Methanol 7 Ethanol H v 1 ln(P)    const 6 0.0025 0.003 0.0035 R T 1/T : [K] H 1 ln(P)  const  v When plotting the vapor pressure of the two R T methanol/ethanol components against 1/T, the Note: For the derivation of the equations, refer to “On difference in vapor pressure widens as the the B Value of the Antoine Equation” (Tips #0811). temperature (saturation pressure) decreases. Solutions for R&D to Design 5

Relationship between Vapor Pressure

PreFEED and Latent Heat of Vaporization

In general, the latent heat of vaporization increases in proportion to the point (Hildebrand rule). In the case of methanol and ethanol, ethanol has a higher boiling (a larger molecular weight) and its latent heat of vaporization is larger. Therefore, if ln(P0) is plotted against 1/T, the gradient is larger for ethanol which has a larger latent heat of vaporization. H 1 ln(P)  const  v Latent Heat of R T Vaporization @ Normal (kcal/mol) 14 1.14 Methanol 8.3932 Ethanol 9.2994 13 1.12

1.1 12 A system showing ideality contains 1.08 substances of homologous series having close

) : [Pa] : ) 11

0 molecular sizes. For homologous substances,

1.06 α : [ - ]

ln(P the latent heat of vaporization is almost 10 1.04 proportional to the difference in boiling Methanol points (molecular weights). For this reason, it 9 Ethanol 1.02 can be seen that for an ideal system α(2nd Axis) 8 1 composition, α increases (separation becomes 0.0025 0.003 0.0035 easier) as pressure decreases. 1/T : [K]

Solutions for R&D to Design Latent heat of vaporization: “DIPPR Databank” 6 PreFEED Non-Ideal Solution System (*)

For a non-ideal solution system, the relative volatility ( ) can be expressed using vapor pressures and activity coefficients ( ).

o K1  1P1    o K2  2 P2

y  Po o K  1  1 1 Py1   1x1P1 1 Modified Raoult’s Law x1 P o Py   x P o 2 2 2 2 y2  2 P2 K2   x2 P

Note: For simplification, it is assumed that the non-ideality of the vapor can be ignored.

Solutions for R&D to Design 7

Methanol-Water System (Vapor Pressure Ratio) PreFEED

As an example of a non-ideal system, consider a methanol- water binary system. Regarding the latent heat of vaporization, since methanol < water, the vapor pressure ratio increases as pressure (temperature) decreases.

13 1.15 Latent Heat of Vaporization @ Normal 1.14 Boiling Point (kcal/mol) Methanol 8.3932

12 [‐] : Water 9.7479

1.13 ) 2 0 ) : [Pa] 0 )/ln(P 1

ln(P 1.12 0 11 ln(P

Methanol 1.11 Water ln(P0) Ratio (2nd Axis) 10 1.1 0.0026 0.0027 0.0028 0.0029 0.003 1/T : [K]

Solutions for R&D to Design Latent heat of vaporization: “DIPPR Databank” 8 Methanol-Water System (Activity Coefficient Ratio) PreFEED

Considering a constant pressure vapor-liquid equilibrium, the activity coefficient ratio (*) is verified for an equimolar liquid phase composition (mole fraction = 0.5) by reducing the pressure starting from the atmospheric pressure (left figure). The lower the pressure, the larger the activity coefficient ratio becomes. From this, it can be seen that the relative volatility increases as pressure decreases thus making separation easier (right figure).

Activity Coefficient vs Pressure 1 Constant Pressure 1.21 0.98 0.9 VLE Diagram 1.2 1.19 0.8 1.18 0.97 0.7 Methanol 1.17 Water 0.6 1.16 Activity Coefficient Ratio 0.96 0.5 1.15 0.4 1.14 y (Methanol) 0.3 1.13 0.95 0.2 ATM 1.12 0.2

Activity Coefficient Ratio : -[ : Ratio Coefficient ] Activity 0.6 ATM 0.1 1.11 1 ATM Activity Coefficient γ @mole fraction=0.5 1.1 0.94 0 00.20.40.60.811.2 0 0.10.20.30.40.50.60.70.80.91 Pressure P : [atm] x (Methanol) Next, let us consider a system in which the relationship between Activity coefficient equation: Wilson Equation the boiling points and the latent heats of vaporization is inversed. Wilson parameters are fitted from VLE data

Solutions for R&D to Design 9

PreFEED Methanol-MEK(Methyl-Ethyl-Ketone) System

In a homologous system, the boiling point is proportional to the latent heat of vaporization, but in a heterogeneous system the situation may differ due to the influence of hydrogen bonds and other factors. For instance, consider a methanol / MEK binary system. Methanol is the low boiling point component, but has a greater latent heat of vaporization than MEK. For this reason, the vapor pressure ratio decreases with decreasing pressure (temperature).

13 1.05

Latent Heat of Normal Boiling Vaporization @ Normal 12 [‐] ˚

: Point ( C)

) Boiling Point (kcal/mol) 2 0 Methanol 8.3932 64.7 ) : [Pa] 1.04 0

)/ln(P MEK 7.5364 79.6 1 ln(P 0 11 Methanol ln(P MEK ln(P0) Ratio (2nd Axis)

10 1.03 0.0028 0.00285 0.0029 0.00295 0.003 1/T : [K]

Solutions for R&D to Design Latent heat of vaporization: “DIPPR Databank” 10 Methanol-MEK System (Activity Coefficient Ratio) PreFEED As in the methanol/water system, considering a constant pressure vapor-liquid equilibrium, the activity coefficient ratio (*) is verified by reducing the pressure starting from the atmospheric pressure (left figure). The activity coefficient ratio is increased due to pressure decrease, but the vapor pressure ratio shows the opposite trend. As a result, it can be seen that, unlike the other system, the relative volatility decreases as pressure decreases (right figure).

1.25 Activity Coefficient 1.02 1.8 vs Pressure Relative Volatility 1.24 1.7 vs Pressure

1.6 1.23 1.5 1.22 1.01 1.4 1.21 1.3 1.2 1.2 Methanol

1.19 MEK -[ : Ratio Coefficient ] Activity 1.1 α Activity Coefficient Ratio 1.18 1 1 Activity Coefficient γ @mole fraction=0.5 Relative fraction=0.5mole α @ Volatility 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 Pressure P : [atm] Pressure P : [atm]

Activity coefficient equation: Wilson Equation

Solutions for R&D to Design Wilson parameters are fitted from VLE data 11

PreFEED Methanol-MEK System (VLE)

The methanol-MEK binary system has a minimum , thus it cannot be separated into pure substances by a simple distillation operation. However, it can be seen that the higher the pressure, the better the distillation separation operation will work.

Methanol-MEK System Methanol-MEK System Constant Pressure VLE Constant Pressure VLE Diagram (Calculated Values) Diagram (Measured Values)

Solutions for R&D to Design VLE data : Dechema Data Series 12 Industrial Case Study

PreFEED (Cyclohexanone-Cyclohexanol System ) Typical production methods of caprolactam include the method and method of direct oxidation of cyclohexane. In these processes, cyclohexanone obtained by the oxidation of cyclohexane or cyclohexanol is separated by distillation and used as a raw material for caprolactam production. At this time, since the cyclohexanone / cyclohexanol system has a low relative volatility, a very large number of stages is required for distillation separation at atmospheric pressure. This is why has been adopted industrially. Sulzer also introduced an example where a packed tower is used to suppress the pressure loss in the column. Cyclohexanone-Cyclohexanol System Constant Pressure VLE Diagram Relative Volatility 1.0 4 (Measured Values) vs Temperature

0.8 3

0.6

2 0.4

30 torr 1 (Cyclohexanone)y α @mole fraction=0.5 0.2 100 torr α (Measured Value) 300 torr Vapor Pressure Ratio 750 torr 0 0.0 0 50 100 150 200 0.0 0.2 0.4 0.6 0.8 1.0 Temperature : [℃ ] x (Cyclohexanone)

Solutions for R&D to Design 13

PreFEED Conclusion

• In an ideal solution system, the latent heats of vaporization are proportional to the differences in boiling points (molecular weights). For this reason, in an ideal system, the relative volatility always increases as pressure decreases, making distillation separation easier. • On the other hand, in the case of a non-ideal solution system, it is not possible to determine as a rule whether distillation separation will be enhanced by decreasing pressure. • Generally, if there is a proportional relationship between the latent heat of vaporization and the boiling point difference, the relative volatility will tend to increase by decreasing the pressure of the system. However, specific verifications are required for systems with a large change ratio of activity coefficients.

Solutions for R&D to Design 14