You are on page 1of 10

Rcv. Per. Qum. Ing. Qum. Vol. 16 N. 0 1, 2013. Pgs.

75-84

MODELLING ETHANOL-WATER PRESSURE SWING


DISTILLATION IN AN STRUCTURED PACKED BED
COLUMN
Alfredo Palomino l. 1 r>, Deyvi A. Parientes S.2 ,Hctor Gmez R.1 , Pedro Pucar C. 3

ABSTRACT

Applying the method of pressure swing distillation (PSD) we obtained (97 % v/v) ethanol starting from 95
% v/v ethanol, which shows that the ethanol-water azeotrope has been broken in this study. Experimental
worl< to achieve the final product was carried out operating the column in two stages or modes, these
being: (1) Low pressure distillation (LPD), and (2) High pressure distillation (HPD). In the first operating
mode, a low degree industrial alcohol (80 % v/v ethanol) is fed to a bench scale distillation column packed
with Mellapack structure, where (95 % v/v) ethanol is obtained, working at atmospheric pressure. Then
product collected in !he first stage is fed to the reboiler, and the HPD process takes place to gel 97 %
(v/v) ethanol. Here we have shown that it is possible to perform the pressure swing using an auxiliary
engineered device, instead of using a second distillation column, in contras! to the opinion of Doherty and
coworkers (2001 ). As part of this work we also have shown how to exploit the pressure sensitive property
of an azeotropic mixture, as it is the case of the ethanol-water mixture. The ratio of the volume of product
obtained to !he energy consumption is 0.7 L. et-OH (97 %)/kWh.
Keywords: Pressure Swing Oistillation , structured bed, intensification, azeotrope, non equilibrium
thermodynamics.

RESUMEN
Aplicando el mtodo de destilacin por oscilacin de presin (PSO) se obtuvo (97% v I v) de etanol a
partir de etanol al 95% v / v, lo que demuestra que el azetropo etanol-agua se rompe en este estudio.
El trabajo experimental para lograr el producto final se llev a cabo operando la columna en dos fases o
modos, siendo estos: (1) la destilacin a baja presin (LPD), y (2) destilacin de alta presin {HPO). En
el primer modo de funcionamiento, un alcohol industrial de bajo grado (etanol 80% v/v) se alimenta a una
columna de destilacin a escala de banco, empacado con estructura Mellapack, donde se obtiene etanol al
95% v I v trabajando a presin atmosfrica. Luego, el producto colectado en la primera etapa se alimenta
al hervidor, y el proceso de HPD tiene lugar para obtener 97% (v I v) de etanol. Aqui hemos demostrado
que es posible llevar a cabo la oscilacin de presin usando un dispositivo auxiliar de ingeniera, en lugar
de usar una segunda columna de destilacin, en contraste con la opinin de Doherty y colaboradores
(2001). Como parte de este trabajo tambin hemos demostrado cmo explotar la propiedad sensible a
la presin de una mezcla azeotrpica, como es el caso de la mezcla de etanol-agua. La relacin entre
el volumen del producto obtenido con respecto al consumo de energa es de 0.7 L. et-OH (97%) I kWh.
Palabras clave: destilacin, cambio de presin, lecho estructurado, intensificacin, azetropo, termodinmica
de no equilibrio.

Facultad de Qumica e Ingeniera Qumica, Universidad Nacional Mayor de San Marcos - Lima, Per
2 Facultad de Ingeniera Qumica y Metalrgica, Universidad Nacional Jos Faustino Snchez Carrin - Huacho, Per
3 Facultad de Ciencias Fsicas, Universidad Nacional Mayor de San Marcos - Lima. Per

75
Rev. Per. Qum. Ing. Qum. Vol. 16 N. 0 1, 2013. Pgs . 75-84

1. INTRODUCTION Concerning PSD, in the present case we


have found that it is necessary to take into
Separation processes in chemical enginee-
account the following considerations, before
ring are used weather to purify raw materials,
taking a decision:
intermediate species, or final products to
satisfy the market requirements. The concentration of the feed should be near
to its azeotropic point.
Distillation is one of the most used separation
process in the chemical industry, despite of The azeotrope should be pressure sensitive
its high energy consumption and low effi- type.
ciency compared to other separation pro-
cesses. Currently research works worldwide Pressure swing should be conducted from
is aimed at intensifying processes, whether low to high pressure, or viceversa, aimed at
by designing compact units, or engineering getting pseudo steady state operation.
accessories that enable an optima! use of
the energy available. 2.1 Sorne applications of distillation

In this study we conduct pressure swing Distillation is widely applied in petroleum


distillation (PSD) in an structured packed bed processing, for example fo r separation of
column at bench scale, with no use of any propylene / propane. Oil refineries rely on
mass separation agent. The final product distillation to obtain the prod ucts used in
obtained is ethanol over its azeotropic value modern society, such as fuels , so lven ts,
(97 % v/v), after operating the same column etc. Distillation processes equipped with
at two different pressures. heat pumps increase process efficiency. In
aeronautics, distillation is used to recycle
astronauts' urine.
2. SOME CONSIDERATION FOR
CHOOSING PRESSURE SWING Distillation strength is based on the fact that
DISTILLATION it is a process whose flowchart is quite sim-
ple, it requires low cost of investment, it is
Chemical engineering has devised a series feasible to scale up by the power law 0.6 or
of separation processes and techniques to less. lts weakness is that it can not be used
make products for the market. Among them, to separate temperature-sensitive mixtures,
distillation is considered one of the most ar mixtures whose components have clase
important separation processes. However, it boiling points. In the same way it does not
is important to note that it is not a panacea
work to separate azeotropic mixtures. Whe-
to solve ali separation problems. As a result, never distillation is not the ch osen method of
it is necessary to take into account certain purification, other separation methods such
aspects relevant to its selection as a sepa- as extraction , adsorption, crystallization,
ration process.
chromatographic separations, membran e
Sorne factors to be considered for a prelimi- separations, as well as hybrid systems come
nary decision making as far as selection of a to play fo r a variety of industrial applications.
separation method is concerned are: 3. LITERATURE REVIEW
Minimum relative volatility 1.3.
Lewis (1928) was the first to suggest disti~
Thermally stable products. lling azeotropic mixtures by pressure swing
distillation , PSD. This method is in principie
Production rate of 5,000 to 10,000 lb Id . similar to vacuum distillation, where the boi-
ling point of the mixture is manipulated by the
High rates of corrosion and explosive reac- change in pressure of the system.
tions should not take place under conditions
of distillation.

76
Rev. Pcr. Qum. Ing. Qum. Vol. 16 N. 0 1, 2013. Pgs. 75-84

As it is well known, azeotropic mixtures cannot of separation . In practice it depends upan a


be separated into its components by traditio- series of factors and boundary conditions.
nal distillation, however, there exists sorne
lf we think on distillation under non equili-
mixtures which may be named as pressure
brium thermodynamics domain , we should
sensitive azeotropes, that may be separated
be able to develop certain kind of auxiliary
by PSD (Sattler et al. , 2008, Seider et al. ,
device, and to match properly the sub do-
1999, Frank T. , 1997). lf the pressure increa-
mains to transfer energy from side to another.
ses, the azeotropic point is shifted to favor
the separation of the mixture, in such a way Now to theorize the statement, let us put
that the azeotrope and the high boiling point forward the theory of entropy production from
component may be obtained as the distillate, Onsager (1931).
otherwise the low boiling point component
may be found as the major species in the Ta = l,JX, (1)
i
distillate. This behavior depends upon the
Where o is the rate of entropy production
thermodynamic nature of the azeotrope.
across the frontier of the domain, T is
Pressure swing distillation is used as an the absolute tempera tu re, Ji is the flow at
alternative process to azeotropic distillation, steady state, and Xi represent the gene-
reactive distillation or extractive distillation. ralized forces.
Thus in PSD the pressure of the system
is used to break the azeotrope, and can Now using the relationship due to Tun Sac
be used in batch or continuous processes (2004 ), J is described by,
(Phismister et al., 2000) . Few results for
continuous processes have been found in (2)
the literature (Repke et al. , 2005). Studies of
Where are the phenomenological coeffi-
PSD far the separation of multi component
cients . These coefficients satisfy the
mixtures involving azeotropes may be found
Onsager's relationships. Thus,
in the current literature (Modla et al. , 2009).
(3)
Variations such as reactive-PSD has been
applied to a methanol/methyl acetate azeo- Recalling the work of Palomino (2012), the
tropic mixture produced as a waste stream previous described relationships generate,
in the polyvinyl alcohol industry. The ethyl
acetate (EtAc) is a high value product and the Ta = - l;J'. [RTLl(ln y)k] (4)
k
methanol is a raw material for the polyvinyl Here R is the gas universal constant.
alcohol industry which may be recycled in
the process (Bonet et al. , 2006, Phimister This equation shows that if we are able to
et al., 2000; Modla et al. , 2008; Klein , 2008). harness the entropy produced in a sub do-
main , we may use itas an separation agent,
which is precisely what we have been able
to perform in this study.
4. NON EQUILIBRIUM THERMODYNA-
MICS Non equilibrium thermodynamics 5. MATERIALS AND METHODS
states the existen ce of systems that produce
entropy, in such a way that if we were able Adistillation column equipped with structured
to put in contact two entropy producing sys- Mellapack type packing randomly ordered,
tems, we might transfer part of the available constructed in stainless steel 304, has been
energy from one system to another, ali in ac- used to carry out the experimental part of
cord with the second law of thermodynamics. this study. The diagram of the distillation
As a result, we might be able to transform onset is shown in Figure N 1. The process
the entropy produced in the system, into work is conducted in batch mode. In addition to

77
Rev. Per. Qum. Ing. Qum. Vol. 16 N.0 1 , 2013. Pgs . 75-84

the distillation facility, standard laboratory ratio; that on the whole configure a process
glassware has been used for sampling and of high energetic efficiency, that is to say, an
measurement purposes. intensified process.

6. PROCESS DESCRIPTION Let us refer to the work of Wolley et. al.


(1999), who stated that by conventional
The diagram of the PSD is shown in Figure distillation it is possible to achieve around
No. 2, which in spite of showing two columns 90 to 92 % v/v, which is lower than our result
of distillation, it rather represents two opera- (95 %).
ting modes of the same distillation column;
these being: (1) Low pressure distillation , 6.2 High pressure distillation (HPD)
and (2) High pressure distillation. In fact the
second mode is much more relevant to the Once the first stage is completed and the
ethanol-water mixture has increased it con-
issue of removing the desired product, out
centration close to its azeotropic point. The
of feeding an ethanol-water mixture near to
its azeotrope composition. high pressure distillation starts by feeding the
product obtained in the previous stage. This
The manipulation of the distillation pressure operation req uires a departure condition,
and the recycling ratio are key factors to carry which starts at atmospheric condition, then
out the PSD at pseudo steady state. the outlet valve located in the engineered
device is closed. Thus the distillation column
starts swinging its pressure up to the desired
value, that is 2.26 bar in our case. Stabiliza-
tion of the distillation procedure requires to
match the rate of heating with the recycling
ratio, to be able to get pseudo steady state
operating condition. As a result, starting with
95 % v/v ethanol-water mixture, the sepa-
ration of the components of the azeotropic
mixture is achieved, since the new azeotrope
is intrinsically connected to the new pres-
sure of the distillation column. The product
obtained at this condition is a mixture of 97
Feed
% (v/v), which is above its azeotropic point.

Step 1 S1ep2
Fig. N 1: Distillatio~
Bottom J 1

?~'" te~.~,.
6.1 Low pressure distillation (LPD)

This operation is rather similar to a conven-


tional rectifying distillation, and is conducted
...
mixture __ mixture

at atmospheric pressure (1.013 bar or 1 atm). 1 HP


Thus feeding 80 % v/v industrial alcohol , a
~--
product containing a concentration close to
the azeotrope is obtained (95 % v/v). ,--
A number of factors are combined to achie-
ve the enrichment of the industrial alcohol,
which are the nature of the packing , its
ft:
tl
Feed Feed
contact area and the hydrodynamic of the
Fig. N 2: PSD Process diagram
column , the rate of heating and the recycling
(Adapted from Repke et al., 2007)

78
Rev. Per. Qum. Ing. Qum. Vol. 16 N. 0 1, 2013. Pgs. 75-84

7. MODELLING OF THE PSD DISTILLA-


TION Mi.j = l j-1xi.j-1+ Vj-1Y;,j+1+ 0zi.j
5
PSD may be mathematically described ba- -(lj +U)x.j -(Vj +Wj )Y;, =O ( )
sed upon the assumption that an intimate
contact between two phases (vapor-liquid) Energy balance:
coexist through the equilibrium stages, where
J stages are idealized being in equilibrium. H ) =lJ-1hL1- I +V; -h\' +I +F.h
J J
- (6)
However, in the real world, such contact does un+Uj \ - V+W \-Q =
not take place, weather due to the nature of
the own packing , the velocity of the ascen-
7.2 Solution
ding vapor, ar sorne anomalies concerning
the rate of recycling of the liquid, as well as Equations (5) and (6)) generate a set of non
the distribution ofvapor and liquid inside the linear algebraic equations that are solved
separating column. using a highly efficient algorithm developed
by Wang and Henke (1966).
Modelling a real system of distillation is very
cumbersome to achieve , due to a series of A set of constitutive equations of Soave-
uncertainties involved , which goes from the Redlich-Kwong , UNIQUAC , Wagner, and
proper knowledge of the parameters , as well the fugacity coefficient are used to perform
as the operating conditions. the complete solution of the PSD problem.

One way to face the modeling procedure NTU is calculated from the following equa-
of a real world distillation makes use of the tion:
theory of non equilibrium thermodynamics,
z=NTU * HETP (7)
which enables the possibility of using part
of the entropy generated in the system as Where HETP is obtained through a system
a separating energy, depending upan the of equations developed by Rocha-Bravo-Fair
engineered device used to claim such energy for Mellapack type structured packings. This
(Palomino, 2012). model is based on the interaction between
7.1 Mass and energy balance in the pro- the descending liquid film and the ascending
vapor occupying the space available. Also,
cess
the physical properties and the geometry of
Mass and energy balance are derived based the packing are taken into account (Bravo
on the MESH model. In arder to facilitate et al. , 1996). This model is used to predict
the formulation of the model Figure N. 3 is the pressure drop , liquid retention, and the
presented, where an idealized jth-element of effective area of mass transfer.
volume is shown.Mass balance equations for
Hydraulic effects
each compound :
5
v,
!:J' !:J',cco )
w, L -; & = & [ 1-K.h ]
(8)
y .,
hv 1,, 1 Xhh1-1
<r l
T, r .,
P1 P, , Liquid retention

-(4 J
F eed
Stage j
F,
2
3
z '
hj.. ,
Tf,
P/1
Y .. r1
hll 1. r 1
Tr hli. t
r,
h1
-
F,
S
(
p1 (
p,
senB) Egeff
J: (9)

p ,. 1
p,
v ,.~ U

Fig . N 3: General scheme fer heat and mass bal-


ance (Adapted from Seader et al., 1998)

79
Re\'. Per. Qum. Ing. Qum. Vol. 16 ~- t , 2013. Pgs. 75-84

Effective area
0.92
15
a 29.12r:. (we,F1 ) S 0359 . .f
1O)
__!!_= -
a Re~ &
-
2
---''-'--'---'--'-----
06
o-3
(1-0.93cos19) (sena) ..
0.88
Mass transfer coefficient in gas phase
"'
>- .'
.
K;S [(U -'-UlefJ ) pgSlo.s (~ Jo.JJ 11 )
0.84
'
.
=0.54 xeff
g x Dgpc
Mass transfer coefficient in liquid phase 0.80
0.80 0.84 Xe 0 .88 0.92

(12) .. Distillate Feed


- - Bottom - - Xaz
Fig. N 5: Vapor-liquid equilibrium during HPD .
8. RESULTS Source: Elaborated by the authors.

The following figures show the limits of ope- Table N 1: Mass and energy balance for LPD
ration for each case, observing that stage LP LiP=3.2mmHg
(figure 4) works as a conventional distillation, Mass Energy Exergy
below the azeotrope in terms of the more Stage
(kmol/s)
(kW) (kW)
volatile component, and the stage HP (Figure
Feed 7.16E-05 -2.854 -0.5 12
N 5) operates according to the least volatile
component and above the azeotropic point. Top -3.27E-05 1.164 0.206
Bottom -3.89E-05 1.364 0.265
Tables No. 1 and 2 show the values of the
Q condenser -5.059 -0.766
mass and energy balance of each operation.
Q reboiler 5.384 0.868
Clearly the low pressure distillation is less Summatory O.OOE-+-00 o 0.061
energetically efficient than the high pressure
Recycle
distillation, thus it may be observed that the
2.976
HPD is 40 % more efficient than the LPD. Calculated !her-
ma! efficiency 0.400
Simulation of the process show that the
structured packed bed is equivalent to 15 Experimental 0.500
theoretical stages. thermal efficiency
15
1.0 NTU
. Source: Elaborated by the authors
0.8

Table N 2: Mass and energy balance for HPD


0.6 ... 1
., LiP=50mmHg
>- 1
0.4 :
1 Mass Exergy
1 Stage Energy(kW)
0.2 1 (kmol/s) (kW)
1 Feed 2.61E-04 -64.474 -81 .843
00 l
Top -1.80E-04 43.883 56.309
o.o 0.2 0.4 0.6 0.8 lo
Xe Bottom -8.11E-05 19.750 25.403
Distillate Feed O condenser -5.579 -1.118
- - Botton - - - Xaz
O reboiler 6.420 1.299
Fig. N 4: Vapor -liquid equilibrium during LPD.
Source: Elaborated by the authors.

80
Rev. Per. Qum. Ing. Qum. Vol. 16 N. 0 1, 2013. Pgs. 75-84

g: vapor phase viscosity ~ This process may be qualified as intensi-


fied, since the energy efficiency is high
I: liquid phase viscosity
(73% ). This value is far overa conven-
pi: liquid density tional distillation efficiency, which is 57%.

pv: vapor density ~ Distillation under conditions of non equi-


librium thermodynamics, produce entropy
a: tension energy in the system. This energy is harnessed
to improve the efficiency of the PSD,
8: vacuum fraction
which explains why the energy effic iency
is really high.
Dimensionless Numbers

Re g : Reynolds number of vapor phase


11. ACKNOWLEDGMENT
Re1: reynolds number of liquid phase
This work was supported by Consejo Supe-
Fr,: Fraude number of liquid phase rior de Investigaciones, and FQIQ, UNMSM.
We thank Paul Sharratt far critica! read ing of
We 1: weber Number of liquid phase manuscript. Thanks also go to Jorge Dvila,
Leighton Estrada and Javier Valeriana far
10. CONCLUSIONS helping during experimental part of this work.

~ Pressure swing distillation at bench scale


was performed, by switching the pressure 12. LITERATURE
from low (1.013 bar) to high (2.26 bar), [1]Abrams, D. S., & Prausnitz, J. M. (1975).
thus distilling at 99.9C. The final product Statistical thermodynamics of liqu id
was obtained in the second stage getting mixtures: a new express ion far the
ethanol with 97 % v/v concentration, from excess Gibbs energy of partly or
a feed with 95 % v/v ethanol. This fact in- completely miscible systems. AIChE
dicates that the ethanol-water azeotrope Journal, 21(1) , pp.116-128.
has been broken in the process.
[2] Bonet, J., Galan , M. l., Costa, J., Thery,
~ The model produced in this work is vali- R., Meyer, X., Meyer, M., & Reneaume,
dated by simulation, with an error of 0.546 J. M. (2006). Pressure optimization of
%, as being contrasted with experimental an original system of pressure swing
results. with a reactive column. lnstitu tion of
Chemical Engineers Symposium Series
~ The topology of the structured packed (Vol. 152, p. 344). lnstitution ofChemi-
bed, favored the process of contacting cal Engineers; 1999.
descending liquid with the ascending
[3] Bravo, J. L., Rocha, J. A., & Fair, J. R.
vapor inside the column. This is one of
(1996). Distillation colu mns containing
the most important components in the
structured packing's: a comprehensive
process of separating the mixture by the
model for their performance. 2. Mass-
PSD method; the other one being the
transfer model. Industrial & Engineering
rate of vapor production and the velocity
Chemistry Research , 35(5), 1660-1667.
of recycling.
[4] Fowler, R., & Guggenheim, E. A., (1939).
~ The efficiency ofthe fractionating column Statistical thermodynam ics: a ve rsion
in the LPD mode may be considered to of statistical mechanics far students of
be high , since the concentration jumps physics and chemistry (Vol. 2). CUP
from 80 % v/v to 95 % v/v, which is very Archive .
clase to its azeotropic point.

82

J
Rev. Per. Qum. Ing. Qum. Vol. 16 N. l , 2013. Pgs. 75-84

(5) Frank, T.C. , (1997) . Break azeotropes [17) Phimister, J. R. , & Seider, W. D. (2000).
with pressure sensitive distillation , Apr. Semicontinuous, pressure-swing distil-
52-63, CEP. lation . Industrial & Engineering Chem-
istry Research, 39(1), 122-130 ..
[6) Gmehling, J. , Onken , U. &Arlt, W, (1981).
Vapour-Liquid equilibrium data collec- [18] Prigogine, llya. (1967). Estructuras disi-
tion, Part 1a Aqueous-Organic Systems pativas. Premio Nobel de Qumica en
(Supplement 1), DECHEMA. el ao 1977 - (http://es.wikipedia.org).

[7] Klein , A. (2008). Azeotropic pressure [19) Renon, H., & Prausnitz, J. M. (1968).
swing distillation. (Doctoral dissertation, Local compositions in thermodynamic
Universitatsbibliothek). excess functions far liquid mixtures.
AIChE Journal, 14(1), 135-144.
(8] Lebon, G., Jau, D ., & Vzquez, J . C.
(2008). Understanding non-equilibrium [20] Repke, J . U ., Forner, F. , & Klein, A.
thermodynamics: foundations, applica- (2005). Separation of homogeneous
tions, frontiers. Springer, A lemania. azeotropic mixtures by pressure swing
distillation-analysis of the operation
(9] Lewis, W. K. (1928) . Dehydrating alcohol
. . g of performance. Chemical Engineering &
and the Like. U.S. Patent, 1 676 700.
Technology, 28(10), 1151-1157.
o far (10] Modla, G., & Lang P. (2008). Separation
[21) Repke , J. U ., Klein, A., Bogle, D., &
,york. of ternary homoazeotropic mixture by
Wozny, G . (2007) . Pressure swing
pressure swing bath distillation. (Vol.
batch distillation far homogeneous
36(1-2), pp. 89-94) Hungarian Journal
azeotropic separation . Chemical Engi-
of Industrial Chemestry Veszprm.
. 975). neering Research and Design, 85(4),
[11) Modla, G., Kopasz,A., & Lang, P. (2009). 492-501 .
quid
Feasibility of separation of ternary mix-
J th e [22] Santiago, R. S., Santos, G. R., & Aznar,
tures by pressure swing batch distilla-
-: '/ or M . (2009) . UNIQUAC correlation of
tion . PSE2009.
:.. ChE liquid-liquid equilibrium in systems
[ 12] Onsager, L. (1930). Reciproca! relations involving ionic liquids: The DFT-PCM
in irreversible processes l. Department app roach. Flu id Phase Equi libria ,
ery,
of Chemistry, Brown University. 278( 1), 54-6 1.
~..im e ,

[13] Onsager, L. (1 93 1). Reciproca! re lations [23] Sattler, K., Feindt, H. J. (2008). Thermal
in irreversible processes 11 . Departmen t separation processes: principies and
of chemistry, Brown University. design. Wiley-VCH ..
[14) Palomino, A. (2012). Etanol anhidro [24] Seader, J., & Henley, E. (1998) . Separa-
va ingeniera de destilacin sin uso tion Process Principies. John Wiley &
de agentes msicos de separacin ni Sons, New York.
" R. vaco, Premio Graa y Montero a la
[25) Seider W.D., Seader, J.D. & Lewin D.R.
Investigacin en la Ingeniera Peruana;
(1999). Process Design Principies.
Mencin Honrosa, categora de inves-
Synthesis Ana lysis and Evaluation .
tigacin profesional.
John Wiley & Sons, New York.
[15] Perry, R. H., & Chilton, C. H. (1982).
[26] Smith , J.M ., Van Ness, H. C., & Ab-
Manual del ingeniero qumico. Mc-
bott, M . M. (2008). Introduccin a la
Graw-Hill.
termodinmica en ingeniera qumica.
[16] Perry, R. H., & Green, D. W. (1997). Stima edicin, McGraw Hill.
Perry' s chemical engineering hand-
"" CuP (27] Soave, G. (1972). Equilibrium constants
book, seventh edition , United States of
from a modified Redlich-Kwong equa-
America , McGraw-Hill.