2

Gas*LiquidContactors for Distillation:

Packed Columns

Jean-Paul Gourlia
Pierre Trambouze

Following the discussion of columns with plates in Chapter 1, this chapter

deals with packed columns. Packings are in fact an alternative technology for
mass and heat transfers between a liquid and a vapor phase circulating
counter-currently. Whereas a plate column is made up of a series of distinct
contact stages between which the two phases circulate separately, a packed
column achieves a counter-current continuous contact. In other words, the
phases are in contact throughout the whole column and are separated just
before leaving it. The exchange of components between phases is even more
effective as the contact surface area and turbulence are considerable. This is
why the column is filled with solid elements (the packing material) of carefully
selected shapes and sizes. Compared with plate columns, packed columns
exhibit a number of advantages and drawbacks as will be seen in the course of
this chapter.

2.1 Different m e s of Packing

The simplest type of contact apparatus is an empty column: a sprayed column
if the liquid is dispersed by sprinkling in the gas phase, or a bubble column if
the gas is injected in the liquid constituting the continuous phase. These sys-
tems are not particularly efficient. In order to improve the contact between
phases and the agitation in each one, the column is filled with packing mate-
rial in one or more sections, with liquid redistribution devices at the top of
each packed section. The liquid trickles down over the packing elements form-
ing a film which is licked by the gas phase rising counter-currently.
Theoretically, solids of any shape could be used. However, an efficient
packing element must have the highest possible area per unit of volume and
the resistance to flow (pressure drop) must be low. In addition, it must not
44 Chapter 2 GAS-LIOUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS

hold back stagnant liquid. For a given packing shape, the essential character-
istics are the specific area (area per unit of packed volume) and the proportion
of voids or porosity. The type of material is also important, as the liquid must
wet the surface of the packing, forming a film that is as continuous as possible,
i.e. it must have a small contact angle. Good packing efficiency is, however,
dependent on the distribution of phases over the column cross-section, and in
particular the liquid phase distribution at the top of the column.
The packing types proposed by manufacturers have substantially evolved
with time and a wide variety is available today. There are three main cate-
gories:
dumped, randomly arranged packings,
arranged or structured packings,
stacks of grids, similar to the preceding type.

2.1.1 Random Packings

These were the first packings used (since the turn of the century). Developed
from two basic shapes (rings and saddles), they have since evolved apprecia-
bly. Some authors (Kister, 1992) classify them into three generations:
First generation: represented by Raschig or Lessing rings and Berl sad-
dles (Fig. 2.1). They are still used due to their low cost.

iigure
2.1 Examples of first generation random packings.
__ A. Raschig rings (ceramic). B. Lessing rings (ceramic). C. Spiraled rings
Chapter 2 GAS-LlouiD CONTACTORS FOR DISTILLATION: PACKED COLUMNS 45

A B

1I
Figure
2.2
1
Examples of second generation random packings.
A. “Intalox”saddles by Norton (ceramic). B. Pall rings (metal). C. “lntalox”
saddles by Norton (plastic).

Third generation, comprising a very wide variety of packings, usually made

of metal or plastic and characterized by greater porosity than the preced-
ing ones (Fig. 2.3). They can not all be listed here, but some of the charac-
teristic types are: Intalox, Cascade Miniring, Chempak, Nutter ring, etc.
There is a choice of materials: ceramic, porcelain (cheap and corrosion
resistant but brittle), steel (thinner, thus providing a larger fraction of void, but
more expensive) and plastic (low cost, light but temperature limited, for exam-
ple polypropylene). In addition, plastics exhibit surface properties that d o not
facilitate wetting by the liquid.
Lessing rings have a central wall which increases the surface area by
around 20% compared to Raschig rings. However, the increase involves the
inside surface, which is less efficient. Other rings are divided into three or four
compartments by walls. The commonly used Pall rings are slotted and have a
tongue-shaped extension toward the inside. They are more difficult to manu-
facture than Raschig rings, but perform much better. For the same dimension
and the same wall thickness, the flow rate is 50 to 100% higher with the same
pressure drop. Alternatively, if the flow rate remains the same, the pressure
drop is 50 to 70% lower. Additionally, the efficiency at optimum flow rate is
from 10 to 50% higher.
46 Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS

C D

Figure
2.3 Examples of third generation random packings.
I I A. “Cascade Minirig” rings by Glitsch cmetalj B. “Cascade Miniring” (plastic).
C. “Tellerette” packing by Ceilcote Co. (plastic). D. “Chempak” packing by
Chem-Pro (metal). E. “Fleximax” packing by Koch (metal). E Nutter rings
(metal). G. “htalox Metal ’’ packings by Norton (metal).
Chapter 2. GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS 47

Other types of packing are sometimes advantageous, even though they cost
more than rings: saddle-shaped packings. The most well known are Berl and
Intalox saddles. Their streamlined shape gives a lower pressure drop for a rela-
tively large contact area. Intalox saddles allow flow rates 20 to 25% higher than
Raschig rings. Berl saddles perform a little less well, but their shape gives better
wetting than rings. Furthermore, there is little tendency for channeling to occur.
Manufacturers provide the characteristics of the various types of packing,
with several examples for random packings in Table 2.1. Note that from one
manufacturer to another variations can be quite significant, in particular in
ring thickness. Among the characteristics, particularly important points are
the specific area (I,,and the fraction of void E. The packing factor Fp = is
often used to compare packings. Table 2.1 lists a number of these values.
Those marked with a * have been adjusted to improve results from conven-
tional correlations. Note that the column filling mode is extremely important
for all random packings. It conditions the hydrodynamics of the bed and there-
fore its efficiency.

2.1.2 Structured Packings

Other packing shapes are employed in industrial columns, for example wire
mesh or grids assembled in a variety of geometric patterns. Different types
were proposed, but their efficiency was not fully recognized until considerable
progress was made in irrigation at the top of the column. With structured pack-
ings it is in fact essential to distribute the liquid properly at the top of the bed
of packings, which in itself provides only mediocre radial redistribution of the
phases. For instance Sulzer, Goodloe, Koch, Glitsch and Nutter propose pack-
ings, usually manufactured from thin metal sheet perforated with small holes,
corrugated and/or folded, and stacked in layers with a specified spacing due
t o the relief of the corrugations or folds (Fig. 2.4). The flow of vapor and liquid
is thus held up by the succession of layers and channeled in the alternatingly
oriented interstices. The packings are assembled in cylindrical or paral-
lelepiped-shaped blocks so that the column can be completely filled by stack-
ing them up (Fig. 2.5). The main characteristics of a number of these packings
are given in Tables 2.2. and 2.3.

2.1.3 Grids
Grids, usually made of metal, were developed from the principle of the wooden
grids that used to be employed in cooling towers. The alternating layers of
more or less elaborate metal grids are manufactured from metal sheet that is
corrugated or folded and amply perforated the same as for structured pack-
ings. Their chief characteristic is the high proportion of void, leading to
reduced pressure drop, high capacity and a relative tolerance for any solids
that might be suspended in the liquid. The various packing manufacturers
(Glitsch, Koch, Sulzer) propose this type of grid, which finds applications
mainly in vacuum distillation towers. A number of characteristics of grids pro-
posed by Glitsch (1986) with the designation EF 25A are given in Figure 2.6.
48 Chapter 2. GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS

Dimensions Number Specific Specific Fraction

of element weight area of void
Type
per m3 CS/m3) @-I) (3)
Ceramic 6000 650 95 75 210 *
Raschig 15000 700 130 68 310 *
rings 48000 700 200 73 525 *
2.5 100000 700 240 72 840 *
2.5 370 000 800 370 64 2 100 *
9.5 1.5 900000 800 500 65 1820
6.3 0.8 3000000 800 ~-
800 70 3300 *
Steel 51 1.2 6 000 460 100 94 190 *
Raschig 38 0.9 15000 480 140 94 270
rings 25
19 1 0.7
0.6
48000
100000
370000
560
580
700
210
270
400
93
92
91
380 *
730
530
900 000 930 600 88 880
3 000000 1400 __800 82 1450
Ceramic 8500 800 110 68 350
Lessing 14000 900 150 60 690
rings 46 000 800 220 66 765
- -
Steel 1.2 8500 580 120 93
Lessing 38 0.8 14000 610 170 92
rings 25 0.7 46 000 690 250 91

"'
0.6 370000 880 500 89
6.3 0.5 3000000 1600 -1000
- 80
Ceramic 51 8800 640 110 77 150 *
Berl
saddles 3,
25 I 22 000
80000
620000
610
720
900
150
250
480
75
70
65
210 *
360 *
1250 *
4600000 900 1000 62 3000 *
Ceramic 9400 760 118 79 115*
lntalox 19 25000 670 195 80 170 *
saddles 84 000 705 255 77 320 *
210000 710 335 77 430 *
13 720
730000 _625
_ 78 800 *
Steel 6 040 385 102 96 55 *
Pall 38 13000 415 130 95 80 *
rings 49 600 480 205 94 150 *
210000 590 _340
_ 93 230
Plastic 6 360 72 100 92 82
Pall 13600 76 130 91 104
rings 50 100 88 205 90 170
214000 116 340 87 318
Steel 14 700 97.8 60
lntalox 40 46960 97.3 88
Metal 25 157800 96.7 135
Ceramic 6350 620 77 98
Sup-lntalo 52 600 570 __ 75 197
Tellerette 35 300 119 _180
_ 90 245
Steel 46
Cascade 72
Miniring 95
130
Plastic 36
Cascade 50
Miniring 95
- Table 1 --
2.1 Geometric characteristics of a number of random packings.
Chapter 2. GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS 49

I<
B '- Cross-sectional view

Figure
2.4 Diagram showing the geometry and characteristic dimensions of structured
packings.

Geometry of the structure Assembly

Type
a(degr) h@m) B(cm) LCCm) H(m) B(degr) O(desr)
Mellapak 45 1.19 2.41 1.7 21 90 45
Flexipac 45 1.24 2.59 1.8 26.7 90 45
Gernpak 45 0.61-2.54 1.09-5.08 0.91-3.66 24.8 90 45
Intalox 28 1.04 3.89 2.23 27.9 45
Montz 37 1.5 4 2.5 20.3 90 45
Max-Pak 45 1.59 2.87 2.13 30.5 90 45
Sulzer BX 45 0.63 1.27 0.89 17.8 90 60
-
Table
2.2 Geometric characteristics of a number of structured packings (see Fig. 2.4).
50 Chapter 2 GAS-LloUlo CONTACTORS FOR DISTILLATION: PACKED COLUMNS

Specific area Fraction of void

Type
cm-’1 €

Mellapak 125-500 0.95

Flexipac 220 0.91-0.98
Gempak 1 15-450 0.91-0.96
Intalox 220 0.97
Montz 98-300 0.94
Max-Pak 250 0.975
Sulzer BX 490 0.90

w
Table
Characteristic parameters of a number of structured packing.

Sulzer (1994) proposes the “Mellagrid” MG64 and MG90 types, with the spe-
cific area in m-l indicated by the designation number and an element height of
approximately 13 cm.
Chapter 2 GAS-LIQUID
CONTACTORS FOR DISTILLATION:
PACKED COLUMNS 51

Thickness Weight/vol Dimensions of a panel (mm)

% of void -
(mm) 0 Height Width

97 1.2 180
97 1.5 225
97 2.0 300 388 1534

Figure

A. Element 1.8 m long and 38 cm wide. B. Stack ofgrids constituting a packing.

2.2 Fluid Flow in Packings

2.2.1 Pressure Drop
The counter-current flow of two fluids through a bed of solid particles has
been the subject of a good deal of research. If the gas flow rate is increased
gradually in a packed column with the liquid flow rate remaining constant, the
pressure drop AP of the gas rises. For a zero liquid flow rate, the variation in
pressure drop bp versus the gas flow rate C can be represented by a straight
line in logarithmic coordinates (Fig. 2.7). For a given liquid flow rate, the vari-
ation exhibits the same pattern for low gas flow rates. Although the pressure
drop is higher than that obtained without any liquid, it is roughly proportional
to the square of the gas velocity. It has been observed that under these con-
ditions, the liquid hold up in the packings is practically independent of C and
depends on the liquid flow rate L alone.
A change in the curve representing the variation in AP (point A in Fig. 2.7)
can be observed for a certain value of C. The slope of the straight line becomes
steeper. The friction of the gas is then sufficient to hinder the downward move-
ment of the liquid, causing an increase in liquid hold up which in turn
decreases the open area available for gas flow. This discontinuity, which is
sometimes difficult to detect in the variation of AP,corresponds to what is
52 Chapter 2 GAS-L~QU~O
CONTACTORS FOR DlSTILLATION: PACKEO COLUMNS

L =O

/
Loading
point

-
Figure
2.7 Variation in pressure drop versus gas and liquid flow rates in a packed column.

G2
A p = a x l0pLx -
PC

where:
AP in Pa/m of packed height
L and G are the mass flow rates of liquid and gas per unit of column cross-
sectional area
PC is the gas density
a and p are constants characteristic of the packing
Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATIOM: PACKED COLUMNS 53

Dimension
% void (Y P

Ceramic 133 0.03

Raschig 292 0.045
rings 1328 0.1
Steel 51 94 126 0.03
Raschig 25 94 159 0.04
rings 13 93 226 0.043
Ceramic 51 78 73 0.025
Pall 25 73 193 0.033
Steel 51 95 33 0.025
Pall 25 94 80 0.03
Berl 38 75 80 0.03
saddles 26 70 199 0.035
13 65 664 0.04
lntalox 38 74 73 0.03
saddles 25 75 226 0.033
19 73 266 0.035
13 73 571 0.04

Parameters for calculating pressure drop

Maximum pressure drop value

Type of separation
(Pam

Atmospheric distillation 400-800

Low or medium pressure distillation 400-800
Vacuum distillation 8-400
Foaming system 80-200
Absorbers 160-570
Amine absorbers 205
Amine regenerators 245
Gas scrubbers 200-490
___
-
Table
2.5 Maximum recommended pressure drop in various columns.

Table 2.4 gives values of a and p for some of the packings listed in Table 2.1.
54 Chapler 2 GAS-LlQUlO CONTACTORS FOR DlSTlLLATlON. PACKEG COLUMNS

The Leva correlation was established from test data for a system with
water. It can however be applied to other liquids provided operation is suffi-
ciently below the loading point. Between the loading and the flood point, the
correlation logically gives values that are too low. Other methods of calculat-
ing pressure drop will be suggested later on. In particular, packing manufac-
turers give curves that allow pressure drop to be found versus liquid and gas
flow rates. Maximum recommended pressure drop per unit of packing height
for a specific application can also be found in the literature (Kister, 1992)
(Table 2.5).

2.2.2 Capacity
A number of empirical correlations have been suggested for use in predicting
flooding limits. The most widely used are derived from the one established by
Sherwood et al. in 1938 for random ring packings and revised by Eckert (1970).
The correlation allows the gas velocity to be calculated at flooding by means
of a graphic relation between two groups:

where the terms have the following meaning:

U,, gas velocity at flooding in the column assumed t o be empty (m/s)
a,, specific area of packing (m-I)
E packing fraction of void
L mass flow rate of liquid (kg/m2-s)
C mass flow rate of gas (kg/m2.s)
pL liquid viscosity (cP)
PH2O
and v= -
PL
The packing factor F,, = already seen and given in Table 2.1, can be
seen to appear in A. This term is subject to significant error in assessment and
Eckert (1970) indicates the F,, values that should be used for a number of com-
mon randomly stacked wetted packings. Some of these values are found in
Table 2.1 marked with an asterisk. The relation between groups A and B is
expressed by the flooding curve in Figure 2.8 along with other curves giving
pressure drop per meter of packed height. The shaded area between the two
broken lines corresponds to the loading point region defined earlier.
These curves can be represented as equations:
Flooding relation: In (A3 = 0.11 1 7 - 4.012 x B0.25 (2.2)

Pressure drop:
H
with: Kl = 21.79 - 36.19 x B0.25+ 16.60 x
K2 = 7.06 + 10.30 x B0.25- 10.36 x B0.5
Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS 55

0.2

0.1
0.08
0.06

0.04

N
0 2
5 0.02

-
812
0.01
0.008

0.004
Flooding and pressure drop
in packed columns
0.002

0.001
0.01 0

-
Figure
2.8 Basic chart showing flooding and pressure drop in packing.

Kister (1992) recommends using the correlation above only for a packing
factor Fp in excess of 200 m-l.
Billet et al. (1987) developed an interesting correlation applicable with both
random and structured packings. It takes the liquid hold up E~ and the group
B into account.

B is the parameter defined above

for B 0.4, i = 1 and n 1 = 0.388
for B > 0.4, i = 2 and n2 = 1.416
C, and C2 are constants characteristic of the packing (Table 2.6).
Moreover, hold up at flooding (ELf) is given by the following equation:

E~~ is the solution of this fourth degree equation, but only the solution rang-
ing between ~ f and3 E is acceptable.
56 Chapter 2 GAS-LIQUID
CONTACTORS FOR DISTILLATION: PACKED COLUMNS

I
Type
of packing

Raschig rings
Dimension (mm)
and material

50 ceramic
CI

1.574
1 c2
1.499
25 ceramic 1.899 1.950
Pall rings 50 metal 1.580
35 metal 1.679
25 metal 2.083
15 metal 2.081
I
50 ceramic 1.913 1.830
Intalox saddles 50 plastic 1.548
35 plastic 1.600
25 plastic 2.132
T-304 Gempak A2 metal 2.099
100 metal 1.911
200 metal 2.339
300 metal 2.464 2.034
C2-Montz 200 plastic
Mellapak Y 250 plastic 2.464

Values o f constants C, and C, for a number o f packings (After Billet, 1987).

The Reynolds number is calculated at flooding:

Re,= - PL ULf
a p PL

and the Galileo number is conventionally:

More recently, Kister (1992) suggested a simple relation between the pres-
sure drop at flooding and the packing factor:
APr = 41 x F:7 (Palm) (2.8)
The flooding limits for different types of packing can be calculated from
other correlations that have been developed. For example, the brochures pub-
lished by the various packing manufacturers (Hydronyl, Koch, Mass Transfer
Ltd., Raschig, Norton, Glitsch) give calculation methods suited to their prod-
ucts. The data are often presented in the form of diagrams similar to those
used for plate columns, i.e. with the group involving the flow rates of the two
phases, B = L/C vx,
on the abscissa and the capacity parameter giving the
Chapler 2 GAS-LlQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS 57

linear velocity of the vapor at flooding on the ordinate (Fig. 2.9). Strigle (1993)
recently suggested plotting these diagrams in semi-logarithmic form in order
to yield an equation of the following type:
C, = A - 0.06 x log (B) (2.9)
with C, being the maximum operational capacity expressed in superficial
velocity of the gas phase (m/s).

1.o I I I I I I I I I I I I I - 0.20
0.9 -
- 0.18
0.8 -
- 0.16
0.7 -
0.6 - - 0.14

Figure
2.9 Example o f a graphic representation of maximum acceptable flow rates for
packing (After Norton, 1987).

Size (mm) A

15 0.0174
25 0.0285
40 0.039 0
50 0.048 2
70 0.0649

The same procedure can be used for structured packing: for example, the
flooding equation is written as follows for Intalox #2T packing:
C, = 0.059 - 0.06 x log (B) (2.10)
58 Chapter 2 GAs-LloUlD CONTACTORS FOR DISTILLATION: PACKED COLUMNS

A number of authors have adapted this sort of diagram to predict flooding

or pressure drop by interpolation from a few experimental points. Kister and
Gill (1992) for instance have collated most of the known data involving random
and structured packings and grids in the form of curves similar to Eckert's,
with a slightly modified ordinate:

These various correlations can be used to calculate the column diameter

required for each specific case. Manufacturers also provide maximum capac-
ity curves of the C, = f ( B ) type. It is often recommended to correct C, in order
to take the physical properties of the system into account for a specific appli-
cation (Norton, 1993).
The effective capacity C, can then be determined:

c,-- c, -(
;j0.16( &)"" (2.1 1)

2.2.3 Hold Up
Empirical correlations have also been employed to predict the liquid hold up
in a packed column. Even though this characteristic is less important than
flooding limits, it is sometimes of interest, especially for batch distillation.
Here it is a significant factor in the precision with which the mixture compo-
nents are successively separated.
Hold up is the volume of liquid per unit of column volume. Static hold up
corresponds to the liquid remaining trapped in packing when the column is
drained. During operation, dynamic or operating hold up is added and the sum
of the two gives total hold up.
Mackowiak (1989) suggests the following simplified formula:

(2.12)

with h, = the dynamic hold up in volume of liquid per unit of volume occupied
by the packing.
This correlation can be used below 65% of flooding with the following limi-
tations:
0.57 < E < 0.99
56 < aP < 425 m-l
930 < pL < 1120 kg/m3
0.6 < pL < 14 CP
0.15 < Re, < 200
Chapter 2 GAS-LIQUID CONJACTORS FOR DISTILLATION: PACKED COLUMNS 59

2.2.4 Minimum Liquid Irrigation Flow Rate

A certain liquid flow rate is required to form a film on the packing surface. The
minimum flow rate, QLm, can be calculated from various expressions that take
the interfacial properties of the gas-liquid-solid system into account.
Interfacial properties, in particular the angle of contact, are not always avail-
able and the proposed correlations should be used with care even though they
are based on a theoretical analysis as is the case for the Schmidt relation:
Ct/9
QLm= 0.24 (2.13)
(1 - TL)".5x a;5

with: c,= -
P."L
and TL= 0.9 ( u,,U,) 2.8

QLmis in m3/h.m2of column cross-section, p, in CP and rs in dyn/cm.

This correlation is applicable only under the following conditions:
P < 1 atm
L / G 0.5 to 1.25
Galileo number: 2.7 x lo4 t o 3.6 x lo7
Angle of contact: 8-20 degrees
Packing material: ceramic, steel.
Glitsch suggests values for Miniring type packing. They correspond to a
specific area greater than 140 m-l for various materials:

Material

Unglazed ceramic 0.5

Glazed ceramic 1.9
Oxidized metal 0.7
Stainless steel 2.9
PVC 3.4
PTFE 4.9
Polypropylene 3.9
Glass 2.44

Kister (1992) suggests using relation 2.13 and the figures from the table
above in the following way:

QLm = QLm(from table) ( 3 5 (2.14)

Other authors suggest taking values proportional t o ap:

Q~~ 3
(
3.5 x 10-3 -
)'Qp
(2.15)
60 Chapter 2 GAS-LIQUID
CONTACTORS FOR DISTILLATION: PACKED COLUMNS

is the critical interfacial tension which depends on the material making up

the packing:
71 dyn/cm for steel
61 dyn/cm for ceramic
73 dyn/cm for glass
40 dyn/cm for PVC
33 dyn/cm for polypropylene
Structured packings exhibit better wetting characteristics: QLmvalues of
0.25 for corrugated sheet metal and 0.12 for metal gauze seem acceptable in
actual practice.

2.3 Packing Efficiency

The efficiency of a packed column is an indication of the separation quality
achieved in the column. The study of efficiency for plate columns leads to a
correspondence between the number of theoretical and real plates. Here in
contrast, the height of packing required for a given separation will be deduced.
Consider the example of absorption of a single component in a gaseous
mixture by a solvent. If the two phases were in equilibrium, the concentrations
x and y of the component in the liquid and gas phases would be those of a
point on the equilibrium curve. Furthermore, according to the mass balance
relations, the concentrations in the two phases which are in contact at any
point in the column are the coordinates of a point on an operating curve
whose extremities represent the conditions at the top and bottom of the col-
umn. For a column with plates that are assumed to be theoretical, the
MacCabe and Thiele construction gives the number of theoretical plates
required to achieve the separation, which corresponds to the variation from
one extremity to the other of the operating curve (see Vol. 2).
In a packed column the phases in contact are at the same time those that
flow in the same section. Nevertheless, the separation difficulty can still be
expressed by the number of theoretical plates required. The number is the
same as that of the steps in the MacCabe and Thiele construction between the
operating and equilibrium curves and will be designated by N. If a packing
height H i s required to accomplish the operation in actual practice, the quo-
tient H/N is called the height equivalent to a theoretical plate (abbreviated
HETP).
A packed section with a height equal to the HETP accomplishes the same
separation as a theoretical stage. The concentrations of the two phases that
leave it are the coordinates of a point on the equilibrium curve, but the two
phases are not in contact since they are separated by the height of packing.
Accordingly, using the concept of theoretical stage to express the efficiency of
a packed column is somewhat artificial. More directly deduced from the col-
umn's real operation, the concept of transfer unit is used by many authors.
Since the concept was discussed in Volume 2, the main formulas will simply be
Chapter 2 GAS-LIQU~O
CONTACTORS FOR DISTILLATION: PACKED COLUMNS 61

recapitulated here (N and HTU, with the appropriate subscripts designate the
number of transfer units and the height equivalent to a transfer unit).
lndividual or film transfer units:

Liquid : (2.16a)

Gas : (2.16b)

Overall transfer units

Liquid : (2.16~)

Gas : (2.16d)

x and y are the mole fractions of the component in the liquid and vapor phases
that contact in a column cross-section, xi and yi are the mole fractions at the
interface.
x* is the mole fraction in the liquid in equilibrium with the vapor of composi-
tion y (and likewise, y* corresponds to the equilibrium with the liquid of com-
position x).
kL, k,, KL and K, are the individual and overall transfer coefficients. aGL
is the
interfacial area of contact between the two phases per unit of column volume.
L M and GM are the molar flow rates of liquid and of gas per unit of column
cross-sectional area.
The formulas above are relative to the equimolecular diffusion of two com-
ponents (distillation of a binary mixture). Nevertheless, they are used for com-
plex mixtures, by assuming that the HTUs defined in relation to any one of the
components are equal, to the extent that the transfer coefficients are inde-
pendent of the composition of each phase.
In actual practice, the formulas below are used to calculate the packing
height H required to accomplish a given separation:
H =Nx HETP in theoretical stages (2.17)

H = No, x HTUoc or H = NOL x HTUoL (2.18)

in transfer units.
As the concentrations at the interface need to be known in order to utilize
the formulas involving N, and NL,they are less convenient t o employ.
62 Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKEL COLUMNS

There is a relationship between HETP and HTU:

HETP = HTU (with h = - (2.19)

h- 1
where m is the slope of the equilibrium curve.
The group h is the ratio of the slope of the equilibrium curve and that of the
operating curve. Generally speaking, it can be assumed that L,/CM is constant
(the operating curve can be likened to a straight line). In contrast, m can vary
significantly in which case an average value is taken. This is true in particular
in distillation where it is preferable to consider the rectifying and stripping
sections separately.
The numbers N , No, or NoL can be calculated algebraically or graphically
from equilibrium data and operating conditions (ratio of flow rates and opera-
tion yield or product purity). They are independent of column architecture.
In contrast, the height equivalent to a theoretical stage or to a transfer unit
depends not only on the nature of the components, the temperature and pres-
sure conditions and the flow rates, but also on the packing used (material,
shape and dimensions) and the characteristics of phase flow in this packing.
As a result, the means of determining it are essentially empirical, whether by
comparison with published data or by correlations. Generally HETP and HTU
have a minimum value (corresponding to the best efficiency) for a gas veloc-
ity slightly lower than the flooding velocity. In addition, efficiency decreases
when the packing size or the column diameter is increased.
Note that the concept of HTU is being phased out increasingly in favor of
transfer models involving the interfacial area and mass transfer coefficients
directly (see Vol. 2, Chapter 4). For absorption with a chemical reaction, only
the transfer models give satisfactory results.

2.3.1 Estimating HETPs

There are a number of simple empirical relations allowing the HETP of random
packing to be estimated, e.g.:
HETP (m) = K d,, (cm) for small diameter columns and packings of the Pall
ring type, with K varying according to packing size:
K = 0.14 to 0.20 for dp= 2.5 cm
K = 0.12 t o 0.18 for dp = 3.8 cm
K = 0.10 to 0.18 for dp= 5.1 cm.
Kister (1992) suggests the formula below for structured packings:

1200
HETP = 1.22 + - (2.20)
UP

These correlations should be used with the greatest care, only an approxi-
mate HETP can be expected. Given the wide diversity of packing types, it is
Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS 63

safer to use manufacturer recommendations. For example, for Intalox packings

by Norton, the manufacturer’s booklet proposes the following correlations for
estimating HETP:

(2.2 1a)

HETP = B [a]-’”’(fi)”” for pL > 0.4 CP (2.21b)

with the values for A and B given in the table below according to packing size:

Size 15 25 40 50 70

A (mm> 272 351 412 550 758

B (mm> 296 383 452 599 827

The validity domains of these relations are also indicated:

relative volatility < 3
2.0 s CY s 27 ; if o > 27, take o = 27
0.06 G pL s 0.83.
For structured packings it is even more necessary to resort to the data spe-
cific to the packing that is used. For example, mention can be made of the data
given by Sulzer concerning BX and CY metallic gauze packings, and Mellapak
and Kerapak packings.
An initial graph (similar to Figure 2.9) allows the maximum loading factor
(C,,.J to be set for a reference packing (Mellapak 250Y) versus the flow rate
parameter B = L/C
C,,,
mL. A table with correction factors to be applied to
is given for the various Mellapak, Kerapak, BX and CY packings. Usually
7580% of this maximum value is used to calculate the column cross-section:
QG
A, =

PL - Pc

Series of curves give the number of theoretical stages per meter (N) as a
function of F = USc 6, as well as the pressure drop (Fig. 2.10) for each type
of packing. The curves giving N are often parameterized versus the pressure at
the top of the column while those giving pressure drop are parameterized ver-
sus the liquid load. As an indication, the values of N vary from 1 to 12 depend-
ing on the packings but more often range between 1 and 6. In small diameter
columns (less than 8 cm), high efficiencies can be achieved ( N = 20 to 50).
64 Chapler 2 GAS-LIQUlD CONTACTORS FOR DISTILLATION: PACKED COLUMNS

Curve parameter = pressure at top of column (mbar)

15
100
--.
E
10
8 -
400 / 25

2
n 6 -
E
v

c 4 -
E
&
0.
Q 2 -
9
U
?
3
2 1
0.8 -
0.6 -

0.1 0.2 0.4 0.6 0.8 1 2 3 4 5

F factor (m /s \/ks/m3)

Example of curves giving the number oftheoreticul stages per meter along with
the pressure drop for a given packing (CY by Sulzer, 1991).
Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS 65

2.3.2 Estimating HTUs

A number of correlations have been established to calculate the film values
HTU, and HTUL. The following formulas are used to determine overall HTUs:

HTU,, = HTUG + h x HTUL (2.22a)

HTU,L= HTUL + h x HTUG (2.22b)
If the resistance to transfer is essentially in one of the phases, it is recom-
mended to use the corresponding overall HTU, as it is very close to the HTU
of the film. The influence of the group h is thus reduced. To give a few exam-
ples, some cases can be mentioned where the transfer is controlled by the gas
phase alone: absorption of ammonia or hydrochloric acid by water, evapora-
tion and condensation in the presence of a non-condensable gas. In contrast,
the resistance is essentially in the liquid phase for absorption of hydrocarbons
by an oil and for absorption of CO, by water, alkaline solutions or amines. The
resistances of the two phases are approximately the same for absorption of
acetone or SO, in water.
In order to calculate HTU, the concepts of mass transfer coefficient and
interfacial area can directly be used for which there are a few correlations in
the literature.
The Onda correlation (1972) was established particularly for the random
packings used in gas-liquid absorption:
kc = 5.23 x Ret7 x S C ; / ~ ( Q ~ ~ ~ ) - ~ (2.23)
Qp9G

(2.24)

with:

Q, is the area of packing wetted by the liquid phase and therefore assumed to
be the interfacial area of gas-liquid contact. ocis the critical interfacial tension
used earlier to calculate the minimum irrigation flow rate (Eq. 2.15).
Additionally, the value of the product a&,,, characteristic of the packing,
was determined by Onda (1972) as indicated below:

I Packing
I apdpfactor
I
Raschig rings 4.7
Berl saddles 5.6
Spheres 3.4
Pall rings 5.8
Intalox saddles 7.1
66 Chapter 2 GAS-LlOUlD CONTACTORS FOR DISTILLATION: PACKED COLUMNS

The dimensionless numbers have the following definitions:

(2.26)

Scj = -PI (i = G or L) (2.27)

ai.Pi
a,, G L
Fr,= - (2.28)
g

(2.29)

Bravo and Fair (1982) used this correlation for distillation and suggested
the following formula for calculating the effective contact area:
a, = 0.62 x a,, x 0°.5x H-0.4 (Ca, x Re,)0.392 (2.30)
USLPL
with: Ca,= - o
(2.31)

and H being the height of packing.

The HTUs are then calculated as follows:

(2.32)

(2.33)

According to the authors of this correlation, who made a thorough com-

parison of the results obtained and experimental data, it is advisable t o use a
safety factor of 1.6 to set the heights of transfer unit. Although the authors sug-
gested a correlation for structured packings, it seems preferable to follow con-
structors' recommendations in this case.

2.4 Phase Distribution

The devices designed to introduce fluids into the column need to be carefully
studied in order to avoid what may be a considerable loss of efficiency.
Feeding the vapor phase into the column does not pose any particular prob-
lems, provided a few simple rules are observed (Kister, 1990) (Fig. 2.11):
Use a perforated pipe of sufficiently large diameter to avoid excessive
velocities and overly high pressure drop.
The pipe can be submerged in the liquid if the feed is totally vaporized,
or it can be placed above the liquid level at the bottom of the column
when mixed phases are let in.
Chapter 2 GAS-LIOUID
CONTACTORS FOR DISTILLATION. PACKED COLUMNS 67

~
1
D

t
c
-
n +
0 0 0 0 0 0 0 0

HL
t
30 crn
or D

> D+50 crn

A
1
Cross-
sectionAA

orifices
diam. c 5 cm

Figure
Diagram showing examples of devices for introducing vapor feed into the
bottom of a column (Kister, 1990).

Proper distribution of the gas in the packing is accomplished by means of

perforated plates equipped with risers. The risers penetrate the packing and
the gas is injected through them (Fig. 2.12).
Some devices used for introducing the phases into plate columns were
mentioned. For packed columns, the effect of liquid phase distribution is of
particular importance and special care is required in choosing and manufac-
turing distributors.
A packing element performs a certain amount of redistribution related to
the radial dispersion specific to the packing itself. On this point random pack-
ings offer better redistribution than structured packings, which generate dis-
persion oriented solely along the planes of the structure. lnitial distribution
must therefore be adapted to the type of packing and in particular to its size,
which governs the specific distribution of the packing. If initial distribution is
better than the specific packing distribution, the initial advantage will prevail
68 Chapter 2 GAS-LIOUID
CONTACTORS FOR DISTILLATION: PACKEC~
COLUMNS

Figure
Diagram showing an example of a packing support grid that allows the vapor
phase to be let in properly distributed (After Kock, 1989).

over a certain height before specific distribution takes over. In contrast, if ini-
tial distribution is not as good as the specific distribution, it can only get
worse.
With the development toward packing structures giving ever lower pres-
sure drops, this problem of distribution has been considerably accentuated. In
particular, the distribution of phases outside the parallel layers of structured
packings is severely curbed if not totally eliminated. These packings are highly
sensitive to both liquid and vapor distribution. As a result, large-scale studies
have been carried out by the companies that market them in order to design
suitable distributors for large diameters.
There is a relatively wide variety of liquid distributors, among which four
main types can be differentiated (Bonilla, 1993):
1. Systems with a trough t o hold up the liquid and liquid flow through ori-
fices perforated over the whole surface. Circular or rectangular risers
allow the gas to pass through (Fig. 2.13).
Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS 69

Figure
2.13 Diagram o f a liquid distributor of the trough and orifice tjpe at the top ofpack-
ing (After Koch, 1989).

2. Systems with parallel channels and flow through orifices perforated

along the vertical walls (Fig. 2.14).
3. Perforated pipes set out in a parallel or concentric pattern (Fig. 2.15).
4. Systems with spray nozzles (Fig. 2.16).
The first two types used gravity flow, whereas the others require a certain
amount of pressure.
1. Retention troughs are used for relatively small sized columns (diameter
< 3 m) and when the liquid phase is free of solids and any products that might
settle out by degradation or polymerization. The system is not recommended
when the vapor flow rate is high. Flexibility with respect to liquid flow rate is
limited, with a maximum/minimum ratio of around 2:l unless risers are higher.
The number of distribution orifices can be up to 220 per m2, with risers 6 cm
in diameter.
2. In the second type of system, the parallel channels are usually rectan-
gular and range between 2 and 8 cm in width. Distribution boxes placed above
the channels distribute the liquid into them through holes at the intersection
of the two stacked elements. The channels distribute the liquid at the top of
the packing by lateral orifices which may be circular, rectangular (slot) or tri-
angular, or through holes in the bottom of the channels. When the orifice is lat-
eral, its outlet should be through a small vertical pipe or against a vertical wall.
One of the advantages of this type of distributor is its flexibility, which can
reach a ratio of 5:l. In addition, high vapor flow rates can be accommodated
with no trouble. With lateral orifices there is no risk of solids being deposited.
70 Chapter 2 GAS-LIOUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS

However, distribution around the circumference of the packing can in some

cases prove to be more difficult.Triangular orifices are more sensitive to the
liquid level and this may lead to non-uniform distribution in large diameter
columns.

Figure
Diagram o f a liquid distributor with parallel channels at the top o f packing
(After Koch, 1989).
A. With lateral v-shaped weirs. B. With lateral orifices extended by elbow pipes.
Chapter 2 GAS-LlQU/D CONTACTORS FOR DISTILLATION: PACKED COLUMNS 71

Figure
Diagram o f a perforated pipe liquid distributor at the top o f packing (After
Koch, 1989).

1 I
2*16
Diagram o f a spray nozzle liquid distributor at the top o f packing (After Koch,
1989).

3. Perforated pipe is one of the best means of distributing t h e liquid at the

top of packing when flow rates are low (lower than 25 m3/h.m2) and column
diameters are under 3 m. Flexibility is limited (around 2:l) and given the high
sensitivity t o solids, it is imperative t o filter t h e liquid feed.
72 Chapter 2. GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKED COLUMNS

4. Systems with spray nozzles are similar t o type 3 but the number of ori-
fices may be considerably reduced. The angle of the dispersion cone must be
less than 90". Although this type of distributor is low cost, it is not often used
a s the resulting distribution is not very satisfactory and may generate liquid
entrainment by the vapor.
The quality of liquid distribution can initially be assessed on the basis of
the number of distribution points. This value must take the packing size and
liquid flow rate into account. The smaller the column diameter, the more liq-
uid distribution points there must be per unit of area. On this point random
packings are more favorable than structured packings. Recommended injec-
tion point densities vary between 60 per m2 for coarse packings and high liq-
uid flow rates and 140 per m2 for fine packings and low liquid flow rates.
Redistribution between packed beds is necessary and the same care must be
taken here as for distribution at the top. There are a large number of redis-
tributor models, with the simplest being a truncated cone welded t o the wall
where the liquid is collected and sent back toward the center of the column.
Generally speaking, plates with risers are used to collect the liquid which then
flows through a downcomer to another distributor.

2.5 Calculating a Packed Column

2.5.1 Principle of the Method
Whatever the type of packing, its dimension must be suited to the column
diameter. The usually accepted uppermost limit is 1/10 of the tower diameter.
According to Eckert (1961), it is preferable to use 1/30 of the tower diameter
a s the limit for Raschig rings, 1/10 for Pall rings and 1/15 for Berl and Intalox
saddles, otherwise fluids may not be uniformly distributed. In fact, since the
fraction of void near the wall is greater in small column diameters than the
fraction in the center of the packing, channeling occurs along the wall.
However, if an excessively small packing is used, the liquid flow rate may
become insufficient for the whole surface area to be wetted and pressure drop
will tend to increase.
It is usually advantageous to have several packed sections in the column
separated by redistributors. This results in lower channeling risks. The height
to be selected for a packed section depends on the packing and its dimen-
sions: around three times the diameter of the tower for Raschig rings, 5 to 10
times the diameter for Pall rings and Berl or Intalox saddles. In any case the
height should be no more than 6 m according to Eckert (1961).
Each packed section is supported by a grid that must not cause greater
resistance to fluid flow than the packing. The grid often has relatively large
corrugations that facilitate the gas phase penetration into the packing
(Fig. 2.12). It is often necessary t o keep the packing in place by a grid at the
top of each section to prevent any lifting or entrainment by the gas which
might cause channeling.
Chapter 2 G A S - L i o u i o CONTACTORS FOR DISTILLATION: PACKED COLUMNS 73

The same as for plate columns, the packing diameter calculation is based
on the flooding correlations that were mentioned earlier. The maximum vapor
velocity for a selected packing can be obtained in meters per unit of empty col-
umn cross-section. A factor (Cd is then applied to find the point below flood-
ing in the packing loading zone where efficiency is at a maximum. In actual
practice, a gas velocity of approximately 50 to 70% of the flooding velocity is
accepted for Raschig rings, and up to 80% for Pall rings and saddles. The same
as for plate columns, a supplementary reduction coefficient (SF) is employed,
which is related to the nature of the system and particularly to the foamabil-
ity.
This finally results in an operational linear velocity:
(2.34)
The column cross section is then calculated from the maximum vapor flow
rate in the column.
The height of packing required can be determined in two different ways:
(1) From the number of theoretical plates required for the separation. The
HETP or HTU is determined for the selected packing, based on specific corre-
lations such as Eq. 2.21 or by comparison with published data and the average
height is obtained by Eq. 2.17. The mass transfer coefficients k , and k , can also
be used, as determined by correlations (for example Eqs. 2.23-2.25) or pro-
vided by packing manufacturers. Then the HTUs can be found by Eqs. 2.16a, b,
c and d.
(2) For chemical absorption of a gas by a liquid, the problem is most often
specified in the degree of gas purification as far as one or more components
are concerned. Calculation proceeds the same as for a chemical reactor by
writing molar balances of the components in each of the phases and using the
mass transfer coefficients given by the same correlations as above (see
Chapter 7 on reactors).
Once the packed height has been determined, a few points still remain to
be checked to make sure the proposed solution is acceptable. If the packed
height exceeds 6 m or corresponds to over 10 theoretical plates, it is recom-
mended to incorporate a redistributor. In addition, it is important to calculate
the pressure drop to be sure it is acceptable for the relevant application.
Lastly, the packing must be properly wetted to insure good efficiency. If any of
these points poses a problem, the calculation should be done again with a dif-
ferent type of packing. In any case, it is advisable to choose packings for which
sufficient data are available. It is always recommended to resort to manufac-
turers for verification before settling on the final solution.
Last of all, one or more types of distributors are selected and inlet and with-
drawal piping for the phases is designed and sized. Figure 2.17 gives an exam-
ple of the final diagram of a complete packed column.
74 Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION. PACKED COLUMNS

Distribution system

-Structured packing

Support grid

U U U U U U L l o g o rn
\Trough distributor

\Retaining grid

\ /
\ /
/ \
Support grid
/ \
, \
f
,~ Draw-off plate

Y.
.Y

9 u - 4 r -Perforated pipe distributor

,,o.. Perforated spray Pipe

Chapler 2 GAS-LlOUlD CONTACJORS FOR DISTILLATION: PACKED COLUMNS 75

2.5.2 Example of Sizing a Packed Column

The same example will be used here as for plate columns in Chapter 1. It is a
depropanizer handling 1000 kmol/h of feed, 60% of which is withdrawn at the
top and 40% at the bottom. A plate is chosen at the bottom of the column
where the liquid and vapor flow rates are the highest. The phase flow rates and
properties at this point are listed below:

I Liquid I Vapor

Mass flow rate 84 111 58 564

Density 435.6 57.9
Volume flow rate 193.1 1011.5
Viscosity 0.113
Interfacial
tension 2.84 mN/m

For this type of distillation, a system factor SF = 0.9 is used. Then the type
of packing must be chosen, for example metal IMTP random packing, size
25 mm, for which manufacturer data are available (Norton, 1987).
(1) Calculating vapor flooding. The flow rate parameter B is calculated:

B= LC = 0.5236

Figure 2.9 corresponding t o this type of packing can be referred to.

Equation 2.9 can also be employed with the appropriate coefficients to obtain
C, = 0.045 m/s. The manufacturer recommends correcting C,, by taking the
physical properties into account a s in Eq. 2.11:

c -- c [;I"'
-
21 -O.I1

At 75% of flooding, the maximum vapor velocity is obtained:

U,, = 0.75 x S F x 0.061

=r
(IG m/s
The column cross-section is then determined a s follows:

Cross-section = -Qv = 4.6 m2

UIll,,
and the diameter is easy to calculate: D, = 2.42 m.
In comparison, the diameter calculated for the plate column was slightly
smaller. 1.90 m.
76 Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION' PACKED COLUMNS

(2) The packed height required is calculated from the HETP, correspond-
ing t o 11 theoretical plates. The HETP of the selected packing can be calcu-
lated using Eq. 2.21a with A = 0.351. HETP = 0.51 m. The manufacturer advises
applying a safety coefficient of 1.13 to take any non-uniform fluid distribution
into account. The packed height required for the separation is therefore:
H = 11 x 0.51 x 1.13 = 6.36 m
This will be rounded off at 6.4 m, the maximum recommended height limit
for a packed bed. The packed column will therefore be smaller in height than
the plate column, since the HETP is lower than the plate spacing.
(3) Now the pressure drop through the packing needs t o be calculated.
The chart provided by the manufacturer can be utilized (Fig. 2.18). It allows
the pressure drop to be read versus B on the abscissa and versus the product
Y = F x C;(lOOO x kL/pJo.l on the ordinate. The coefficient F depends on the
packing size and is equal to 441 for 25 mm. Y = 0.22 and the corresponding
pressure drop AP/H = 6.5 mm H,O per meter can be read on the chart. This

General pressure drop correlation AP in inches of H,O / ft (mm HO

, /m)
10

1.o
r
0

X
N

0"
X
LL
2
0.1

0.01

___
Figure
2.18 Chart giving the pressure drop with a Metal-Intaloxpacking (Norton, 1987)
Chapter 2 GAS-LIQUID
CONTACTORS FOR DISTILLATION:
PACKED COLUMNS 77

value, compared to 1:he recommended figures for maximum pressure drop in

columns (Table 2.5.) is quite acceptable. Total pressure drop will therefore be
41.6 mm H,O.

(4) A liquid distributor is then selected, for example the channel type with
lateral orifices extended by elbow pipe (Fig. 2.14B).
This type of calculation lends itself well to a spreadsheet such as Excel so
that the operating conditions can be varied and repeated, thereby giving the
best solution.

2.6 Packing or Plates?

Is there a “best” technology? As is often the case, no definite answer can be
expected. Each of the technologies has it own qualities and only case-by-case
studies can allow the best solution to be defined, bearing in mind that as a gen-
eral rule plates are less expensive than structured packings.
The chief advantages of packings are as follows:

Low pressure drop, of particular interest for vacuum distillation of com-

ponents with low thermal stability.
Small liquid hold up.
N o difficulty installing packings in small diameter columns.
Possibility of increasing capacity, especially with structured packings.
Often lower in price for the same column volume when corrosive mix-
tures are processed.
Simple construction.
In contrast, the drawbacks are listed below:

Little flexibility. The efficiency drops quickly the farther the packed col-
umn works from optimum operating conditions. Difficulties are encoun-
tered in wetting the packing properly, particularly at low liquid flow
rates.
Fouling is possible with products that can settle out or with solid incrus-
tations. The only remedy is sometimes t o unload the packing completely,
whereas a plate column is accessible for maintenance.
Installing draw offs, circulating refluxes and multiple feed inlets poses
more problems than with a plate column.
Lower efficiency of large diameter columns due to the difficulty in achiev-
ing uniform fluid distribution in the packing.
78 Chapter 2 GAS-LIOUID CONTACTORS FOR DISTILLATION: PACKEL) COLUMNS

0.20

0.15

UGf JZ
0.10

0.05

0
0.01 0.1 Lp
G PL
1

1 1 2'19
Comparison o f capacity curves for plates and for structured packings (Bravo,
1997).

In addition to these technical criteria, column sizing arguments can also be

important. Bravo (1997) showed (Fig. 2.19) that for low flow rate factors, pack-
ings allow higher rising vapor velocities. Meanwhile, for high factors, plates
proved to be superior. As a result, the author suggested the selection flow
chart on the next page.
Chapler 2 GAS-LIOUID
CONTACJORS FOR DISTILLATION: PACKED COLUMNS 79

Calculate the flow rate factor

L
B= c [P,/P,,l ‘I5

B > 0.2 B < 0.1

I Preselect plates -1 Preselect packings

I

Consider the two

alternatives

Plugging
problem
NO YES

NO
I
YES

J.
problem

1 YES
I Weight
the risks -I NO
NO

d Confirm I
be taken into packing and
plates consider the packings
account during
effect of B
design?
80 Chapter 2 GAS-LIQUID CONTACTORS FOR DISTILLATION: PACKEL1 COLUMNS

References
Billet R., Schultes M. (1987) Capacity studies of gas-liquid two-phase counter-
current-flow columns. I. Chem. E. Symposium Series 104, B255-266.
Billet R. et al. (1989) Druckverlust in berieselten Schiittungen. Chem.-Zng.-Tech.
61, 2, 157-159.
Bonilla J.A. (1993) Don’t neglect liquid distributors, Chemical Engineering
Progress 89, 47-61.
Bravo J.L. (1997) Select Structured Packing or Trays, Chemical Engineering
Progress 93, 7, 36-41.
Bravo J.L., Fair J.R. (1982) Ind. Eng. Chem. Proc. Des. Dev 21, 162.
Eckert J.S. (1961) Design techniques for sizing packed towers, Chem. Eng.
Progress 57, 9, 54-58.
Kister H.Z. (1990) Distillation-Design.McGraw-Hill Publishing Co., New-York.
Kister H.Z. (1992) Distillation-Operation. McGraw-Hill Publishing Co., New-York.
Leva M. (1953) Tower packings and packed tower design. US Stoneware Co.,
Akron, USA.
Onda K. (1972) Gas absorption in packed column, Memoirs o f the Faculty of
Engineering, Nagoya University, 24, 2, Nov. 1972, 165-215.
Perry J.H. (1984) Chemical Engineers’ Handbook, 6th Edn. McGraw-Hill, New
York.
Sherwood T.K., Pigford R.L. (1952) Absorption and Extraction. McGraw-Hill, New
York.
Strigle R.F. (1993) Understand flow phenomena in packed columns. Chemical
Engineering Progress 89, 79-83.
Wuithier P. (1972) Pgtrole, Raffinage et Ggnie Chimique, 2nd Edn. Editions
Technip, Paris.
Glitsch (1986) Cascade mini-rings for high efficiency mass transfer. Bulletin 345,
published by Glitsch (UK) Ltd.
Glitsch (1986), Glitsch grid Bulletin 423 and GFO15, published by Glitsch (UK)
Ltd.
Koch (1989) Packed tower internals. Bulletin KFG-3, published by Koch
Engineering Co., USA.
Koch (1994) Flexigrid. Bulletin KI-6, published by Koch Engineering Co., USA.
Norton (1987) Intalox high-performance separation systems. Brochure pub-
lished by Norton, USA.
Norton (1993) lntalox high-performancestructured packing. Brochure published
by Norton, USA.
Sulzer (199 1) Colonnes de sgparation pour distillation et absorption. Brochure
published by Sulzer Chemtech, Switzerland.
Sulzer (1994) Mellagrid. Brochure published by Sulzer Chemtech, Switzerland,