Distillation.

Separation of water and ethanol

in Raschig-Ring Packed Column

Ole Håvik Bjørkedal

olehb@stud.ntnu.no
Therese Bache
theresba@stud.ntnu.no
October 29, 2013

Abstract

This experiment was performed as part of the Felles lab, in the

course TKP4105 Separation Technology. A mixture of water and
ethanol was distilled in a Raschig Ring Packed Column. The pur-
pose of the experiment was to understand operation of a distillation
column, and the parameters which determine it's condition. It was
found that the column reached steady state after 25 minutes, and that
column eciency increases with reboiler power duty.
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 2

Contents
1 Theory 3
1.1 Packed Column . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Column eciency . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 McCabe-Thiele . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Flooding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Reux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.6 GC-Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Experimental 6
2.1 Startup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Exercise 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Exercise 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 Shutdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3 Results 7
3.1 Time to reach Steady State . . . . . . . . . . . . . . . . . . . 7
3.2 Varying reboiler power levels . . . . . . . . . . . . . . . . . . . 7
3.3 Eciency vs Vapor Velocity . . . . . . . . . . . . . . . . . . . 8
3.4 Flooding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4 Discussion 14
5 Conclusion 14
List of Symbols 15
A Measurements & Calculations 17
A.1 Calculating required amount of ethanol . . . . . . . . . . . . . 17
A.2 Vapor velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
B Risk Assessment 20
C Copy of journal 23
D McCabe-Thiele Matlab-script 25
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 3

1 Theory
Theory for this experiment is mainly found from DistInstruct.pdf.1

1.1 Packed Column

A packed distillation does not, contrary to i.e. a sieve-tray column or an
Oldershow column, have a set number of trays. Instead it is lled with a
packing that gives a high surface area for vapor and liquid to react, thus
increasing eciency.
In a Raschig-Ring packed column, which was used in this experiment, the
packing consists of hollow glass cylinders.

1.2 Column eciency

Overall column eciency for a packed distillation column is dened as the
ratio of total number of ideal equillibrium stages to the height equivalent of
a theoretical plate (HETP). HETP is the height of packing which does the
same separation as a theoretical step. HETP can be calculated by (1.1),
Packing Height
HETP = (1.1)
Nt − 1
where Nt is the theoretical number of steps, calculated from a McCabe-Thiele
diagram. A lower HETP indicates more theoretical "steps" in the column,
thus higher eciency.
Column eciency can be determined by plotting HETP against vapor
velocity, which is dened as (1.2)1
˙
Vgas
ν= (1.2)
A
where Vgas
˙ is volume rate of gas, and A is the cross-sectional area of the
column. The diameter of the packed column is given as D = 0.05 m.

1.3 McCabe-Thiele
The McCabe-Thiele method is a method to graphically determine the ideal
amount of steps for a binary destillation process.2 The method assumes
constant molar overow. This implies constant molar ow rates of both
vapor and liquid leaving every stage. To construct a McCabe-Thiele diagram
one needs to determine an operating line and a vapor-liquid equilibrium line.
To make the operating line, one needs the enriched operating line and the
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 4

stripping line. Both the enriched operating line and the stripping operating
line can be derived from the mass balance of light component. To make the
vapor-liquid equilibrium line one construct a plot by plotting the vapor-liquid
equilibrium data.
Equation for nding the enriched operating line is given by (1.3);
 
R xD
yn+1 = xn + (1.3)
R+1 R+1

where yn+1 is the mole fraction of light component in the gas phase in tray
n + 1. xn is the mole fraction of light component in the liquid phase in tray
n, and xD is the mole fraction of light component in the distillate. R is the
reux ratio.
Equation for the stripping operating line is given by (1.4);
 
Lm W xw
ym+1 = xm + (1.4)
Vm+1 Vm+1

where ym + 14 is the mole fraction of light component in the gas phase in

tray m+1 with vapor ow Vm+1 , xm is the mole fraction of light component
in the liquid phase in tray m with liquid ow Lm . xW is the mole fraction of
light component in the bottom ow with bottom ow W. When reux ratio
increases the slope of the enriched operating line increases and the slope of
the stripping operating line decreases.
The condition of the feed, q, is dened as (1.5);
heat needed to vaporize 1 mole of feed at entering conditions
q= (1.5)
molar latent heat of vaporization of feed
Equation (1.5) is used to plot the q-line ine the diagram. The q-line has
slope q−1
q
, which is crossing the crossing-point of the operating lines and the
x=y line.
The number of theoretical stages required in the column is determined by
plotting the vapor-liquid equilibrium line end the operating line in a diagram
with the mole fraction of light component. The mole fraction of the gas phase
should be on the y-axis and the mole fraction of the liquid phase should be
on the x-axis. Steps are being drawn between the operating line and the
equilibrium line from the top tray(distillate); x = xD to the bottom x = xB
and the number of steps drawn is equivalent to the number of theoretical
steps needed.
At total reux the reux ratio goes to innity large. Because of this the
operating line can be found by taking the limit of equation (1.3) when R
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 5

approaches innity:
R xD
y = lim x+
R→∞ R+1 R+1 (1.6)
y=x

The result of this can be used as a new operating line, and the operating line
at total reux will be equal to the line y = x, where x and y are the mole
fraction of light component in liquid and vapor phase.
Using a matlab-script, gure 1.1 was produced. From the plot, it is found
that ve steps are needed. The script is included in the appendix.

Figure 1.1: The McCabe-Thiele diagram used for calculating the number of
theoretical steps. The y-axis shows the mole fraction of ethanol in the vapor phase,
the x-axis shows the mole fraction in liquid phase. The upper curved line is the
equllibrium line, while the straight line y=x is the operating line. The horizontal
and vertical lines in between count the theoretical steps

1.4 Flooding
The ooding point is a condition caused by high vapor velocity. At the
ooding point, the vapor velocity is so high that liquid accumulates in the
top of the column. This causes a sudden increase in the pressure drop
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 6

1.5 Reux
Reux is the portion of the vapor that condenses and returns to the destil-
lation column. This particular experiment will work with total reux, that
is, all vapor is condensed back in to the column.

1.6 GC-Analysis
Gas chromatography is a common type of chromatography used in analytical
chemistry. It is used for separation and detection of compounds that can be
vaporized without decomposition.3
In gas chromatography there is a moving phase, also called the mobile
phase, and a stationary phase. The moving phase is usually an inert gas, and
the stationary phase, called column, usually consist of a polymer or glass.
The compounds that are being analysed will interact with the walls of the
stationary state. Each compound will start to elute at dierent times. The
retention time of each compound will vary, and comparing retention times
can identify compounds in the sample. The quantity of a compound in the
sample can be found by plotting a chromatogram, which is a plot of the
measured signal against time. By integrating the area under this curve, the
quantities can be calculated.
The samples were analysed with Gas Chromatography. The output of
this analysis are given in mass fractions. Equation (1.7) was then used to
convert the data to mole fractions.
wEtOH
MEtOH
xEtOH = wEtOH EtOH )
(1.7)
MEtOH
+ (1−w
MH O
2

2 Experimental
2.1 Startup
The PC and cooling water was turned on, and it was ensured that the taps for
top and bottom samples were properly closed. The column was charged with
5800 mL of water-ethanol mixture, with a mole fraction of ethanol (xEtOH )
of 0.1. The experiment required 1596 mL of ethanol and 4202 mL of water.
Calculation can be found in appendix. (A.5)
The reboiler was set to 50 % power, and the column was monitored until
the rst drop of distillate was made. This was marked as Time Zero. The
column was now run on the same power level for an hour, while excercise 1
was performed.
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 7

2.2 Exercise 1
In the rst exercise, the time required for reaching Steady-State was de-
termined. Samples of the distillate was taken every ve minutes, and the
composition of these were measured with GC. 12 samples were collected,
plus one nal sample of the bottom for use later in the experiment. The
samples' composition can be found in A.1.

2.3 Exercise 2
In the second exercise, the column eciency as a function of vapor velocity
was determined.
Reboiler power was set to 35%, and the column was left to reach steady
state. As the column had reached steady-state, samples of the top and bot-
tom of the column was taken and analyzed with GC. The reux rate at the
time of sampling was noted. The reboiler was set to a new power level, and
the procedure was repeated for a total of ve dierent power levels. The
power levels used in this experiment was 35 %, 40 %, 45 %, 50 % and 55 %.

2.4 Shutdown
When all samples were taken, the heater was turned o and the column was
set to cool down. When the column was cold, it was emptied, and cooling
water and the computer were turned o.

3 Results
3.1 Time to reach Steady State
Figure 3.1 shows top samples taken during the rst hour of the experiment.
From this plot, it can be seen that the column reaches a steady state at
approximately 25 minutes. It was assumed that the time required to reach
steady state was constant throughout the experiment.

3.2 Varying reboiler power levels

The reboiler power levels used, and the corresponding molar fractions of
the top and bottom samples can be found in Table A.2. After the reboiler
power was adjusted, the column was run for 25 minutes and allowed to reach
steady-state before samples from top and bottom were taken.
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 8

Figure 3.1: The graph shows molar fraction of ethanol against time. 12 samples
were taken during one hour, in order to determine when the column reaches steady
state. The reboiler level were 50 % for all samples. Data for the diagram are given
in table A.1.

Figure 3.2 to 3.6 show McCabe-Thiele plots for the ve dierent reboiler
levels. The plots were made in Matlab, using the script that can be found in
appendix D.

3.3 Eciency vs Vapor Velocity

By plotting HETP against vapor velocity, it can be seen how column e-
ciency varies with dierent power levels. A plot of this for the experiment
can be found in gure 3.7. A low HETP indicates higher eciency, as this
implies more theoretical steps of equllibrium in the column.1 Values for vapor
velocity can be found in table A.3.

3.4 Flooding
The reboiler power was increased gradually up to 60%. Accumulation of
liquid in the top of the column was observed, but the actual ooding point
was not reached in the experiment.
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 9

Figure 3.2: McCabe-Thiele plot for the column at 35 % reboiler level. The y-axis
shows the mole fraction of ethanol in the vapor phase, the x-axis shows the mole
fraction in liquid phase. The upper curved line is the equllibrium line, while the
straight line y=x is the operating line. The horizontal and vertical lines in between
count the theoretical steps
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 10

Figure 3.3: McCabe-Thiele plot for the column at 40 % reboiler level. The y-axis
shows the mole fraction of ethanol in the vapor phase, the x-axis shows the mole
fraction in liquid phase. The upper curved line is the equllibrium line, while the
straight line y=x is the operating line. The horizontal and vertical lines in between
count the theoretical steps
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 11

Figure 3.4: McCabe-Thiele plot for the column at 45 % reboiler level. The y-axis
shows the mole fraction of ethanol in the vapor phase, the x-axis shows the mole
fraction in liquid phase. The upper curved line is the equllibrium line, while the
straight line y=x is the operating line. The horizontal and vertical lines in between
count the theoretical steps
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 12

Figure 3.5: McCabe-Thiele plot for the column at 50 % reboiler level. The y-axis
shows the mole fraction of ethanol in the vapor phase, the x-axis shows the mole
fraction in liquid phase. The upper curved line is the equllibrium line, while the
straight line y=x is the operating line. The horizontal and vertical lines in between
count the theoretical steps
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 13

Figure 3.6: McCabe-Thiele plot for the column at 55 % reboiler level. The y-axis
shows the mole fraction of ethanol in the vapor phase, the x-axis shows the mole
fraction in liquid phase. The upper curved line is the equllibrium line, while the
straight line y=x is the operating line. The horizontal and vertical lines in between
count the theoretical steps

Figure 3.7: The gure shows a plot of HETP against vapor velocity for ve
dierent power duties.
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 14

4 Discussion
From gure 3.1, it can be seen that the molar fractions of ethanol are stable,
with some variation. This variation may come from uncertainty in the GC-
analysis. It was assumed that the time needed to reach steady state was
constant for all power duties, and the measurements from exercise 2 seem to
support this assumption.
For the calculation of vapor velocity assumptions of ideal gas and a col-
umn pressure of 1 bar was made. As the vapor is condensated continously
during steady state, it is very unlikely that gas will accumulute in the top
and increase column pressure signicantly.
Water and ethanol are rather small molecules and in the scale of this
experiment it is very unlikely that intermolecular forces will have any signif-
icance.
The ooding point was never reached, but at 60% power duty it was
observed accumulation of liquid at the very top of the column. If power duty
were to be increased further, the ooding point would most likely be reached.
From gure 3.7, it can be seen that HETP decreases with increasing vapor
velocity. This indicates a higher column eciency with higher power duties.

5 Conclusion
From the measurements of the rst hour of the experiment, it was found that
the column reaches steady state after 25 minutes. These measurements was
made with a power duty of 50 %.
At 60 % power duty tendencies of ooding was observed, as liquid accu-
mulated in the top of the column. The actual ooding was not reached.
By plotting HETP against vapor velocity it was found that column e-
ciency increases with higher vapor velocity and thus power level.

Ole Håvik Bjørkedal

Trondheim, October 29, 2013

Therese Bache
Trondheim, October 29, 2013
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 15

List of symbols
Symbol Dimension Description
A m2 cross-sectional area of column
D m Diameter of packed column
Lm V liquid ow in tray m
MEtOH g/mol Molar mass ethanol
MH2 O g mol−1 Molar mass H2O
ni mol moles of component i
ntot mol mole total of ethanol and water
dotn mol s−1 molar ow
pi bar Partial pressure of component i
q no dimension condition of feed
Ri mL s−1 Reux rate
R J mol−1 K Universal gas constant
T K Temperature
ν m s−1 vapor velocity
V mL Volume
Vi L Volume of component i
Vtot L total volume of water and ethanol
VEtOH,96% L Volume of 96% ethanol
Vi L Measured pump ow rate
Vm+1 L Vapor ow
dotVgas L volume rate gas
W L s−1 Bottom ow
WEtOH
x no dimension mole fraction of ethanol in column top
xD Volume
xEtOH L Volume
xn L Volume
xm L Volume
xW L Volume
yn+1 L Volume
ym+1 L Volume
ρi g cm−3 Density of component i
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 16

References
[1] Felles Lab: Distillation Columns , September, 2012, read 7th Oct. 2013.
Available at http://www.nt.ntnu.no/users/preisig/Repository/
TKP_4110_Felles_Lab/experiment%20descriptions/DistInstruct.
pdf
[2] Geankoplis, C.J Transport Processes and Separation Process Principles,
4th ed.; Pearson Education, Inc, 2003
[3] Rebecca Carrier and Julie Bordonaro Intro to Gas Chromatography, 1994,
read 29th Oct. 2013. Available at http://www.rpi.edu/dept/chem-eng/
Biotech-Environ/CHROMO/chromgram.html
[4] Aylward, G. Findlay, T. SI Chemical Data , 6th ed.; John Wiley & Sons
Ltd., 2008
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 17

A Measurements & Calculations

Table A.1 shows mass fractions and corresponding mole fractions of EtOH
from the twelve top samples taken during the rst hour of the experiment.
Mass fractions of ethanol were measured using GC-analysis, the correspond-
ing mole fractions were calculated using (A.4). Table A.2 show measured
Table A.1: The table show mass fractions and corresponding mole fractions mea-
sured as the column approached steady state.

Time [min] Mass Fraction EtOH Mole Fraction EtOH

0 0,77 0,57
5 0,83 0,66
10 0,87 0,72
15 0,89 0,76
20 0,92 0,81
25 0,89 0,77
30 0,91 0,81
35 0,92 0,83
40 0,91 0,79
45 0,91 0,80
50 0,93 0,83
55 0,91 0,79
60 0,92 0,82

mass fractions and corresponding mole fractions of ethanol from top and
bottom samples taken for dierent reboiler power levels.

A.1 Calculating required amount of ethanol

The mole fraction of ethanol, xEtOH , is given by (A.1);
nEtOH nEtOH
xEtOH = = (A.1)
ntot nEtOH + nH2 O
Where nEtOH is moles of ethanol and ntot is the total numbers of moles in
the mixture and nH2 O is moles of water. Moles of ethanol is given by (A.2);
VEtOH ρEtOH
nEtOH = (A.2)
MEtOH
Where VEtOH is the volume of pure ethanol, ρEtOH is the density of pure
ethanol and MEtOH is the molecular weight of ethanol. In a mixture of only
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 18

Table A.2: The table shows measured mole fractions from the top and bottom
of the column for dierent reboiler power levels. The column had reached steady
state for all measurements.
Power level Mass Fraction EtOH Mole Fraction EtOH
35% top 0,90 0,78
35% btm 0,15 0,06
40% top 0,90 0,78
40% btm 0,13 0,06
45% top 0,91 0,80
45% btm 0,11 0,05
50% top 0,92 0,82
50 % btm 0,14 0,06
55% top 0,93 0,83
55 % btm 0,09 0,04
55% top2 0,91 0,80

water and ethanol the moles of water, nH2 O , is given by (A.3);

ρH2 O (Vtot − VEtOH )
nH 2 O = (A.3)
MH2 O

Where ρH2 O is the density of water, Vtot is the total volum in the mixture and
MH2 O is the molecular weight of water. By rearanging the equations (A.1),
(A.2) and (A.3):
VEtOH ρEtOH
xEtOH =
MEtOH
ρH O (Vtot −VEtOH )
(A.4)
VEtOH ρEtOH
MEtOH
+ 2
MH
2O

Solving this for values xEtOH = 0, 1, ρEtOH = 0, 791g/cm3 , MEtOH = 46, 0g/mole,
MH2 O = 18, 02g/mole4 and Vtot = 5800mL, the volume of pure ethanol is
VEtOH = 1532mL. In 96 % ethanol, the volume of ethanol needed is;
VEtOH 1532mL
VEtOH,96% = = = 1596ml (A.5)
0.96 0.96
The amount of water needed is then:
vH2 O = Vtot − VEtOH,96% = 5800mL − 1596ml = 4202ml (A.6)
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 19

A.2 Vapor velocity

The vapor velocity was calculated based on the measured pump ow rates.
When the column has reached steady-state, the ow of liquid reux will be
equal to the vapor ow. By calculating the total number of moles in the
vapor ow, the volume ow rate can be determined by assuming ideal gas.
Vapor velocity is then given as
Total number of moles in the vapor was calculated by (A.7),
ρEtOH ρ H2 O
ṅ = xV + (1 − x) V (A.7)
MEtOH MH2 O

where V is the measured pump ow rate, x the molar fraction of ethanol in
the top of the column and ρ and M are density and Molar weights of water
and ethanol.4 In the instructions, the vapor velocity, ν , is dened as
V̇
ν= (A.8)
A
Further the ideal gas law is given in (A.9),
nRT
V = (A.9)
P
where P is pressure, R the universal gas constant n the number of moles
and T the temperature in Kelvin. A pressure of 1 bar was assumed for the
calculations.
By combining (A.7), (A.8) and (A.9), an expression for the vapor velocity
is found as (A.10).
ṅRT
ν= (A.10)
AP
The vapor velocities for the dierent power levels are given in table A.3.
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 20

Table A.3: The table shows calculated vapor velocities for ve dierent reboiler
power levels. Ideal gas and a pressure of 1 bar is assumed. For the calculations
made, a costant temperature of 78,3 ◦C was assumed based on measurements in
the column.

Boiler level Mole frac- Pump ow Total mo- Vapor

[%] tion Top rate [ml/s] lar ow velocity
[mol/s] [m/s]
35 0,78 0,700 0,018 0,267
40 0,78 0,844 0,021 0,318
45 0,80 1,033 0,026 0,380
50 0,82 1,100 0,026 0,392
55 0,83 1,317 0,031 0,460

B Risk Assessment
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 23

C Copy of journal
 Group B20, Ole H. Bjørkedal & Therese Bache, Page 25

D McCabe-Thiele Matlab-script