1

The Basics of Separation Operations

Alexandre Rojey

1.1 Function of Separation Operations in Refining

A crude oil is made up of a complex mixture of a very large number. of compo-
nents, most of which are hydrocarbons (see Vol. 1, Chapter 1). In order to get
petroleum products that meet given specifications, it is first necessary to sep-
arate out the different fractions or cuts, which then need to be purified or con-
verted, especially when they are designed to be used by the petrochemical
industry. The basic principle of a separation operation is illustrated in
Figure 1.1.

- Components
or “fractions”

- I

- Figure
1.1 Simplified flowsheet of a separation operation.
2 Chapter 1. THEBASICS OF SEPARATION OPERATIONS

The initial mixture (crude oil or any other mixture, A, B, C) is separated into
different components or into different "fractions", characterized by a distilla-
tion range for example. This type of operation generally requires an input of
energy or of a selective separation agent, under conditions that will be dis-
cussed later.
More specifically in oil refining, separation processes have three main func-
tions as defined below.
Fractionation
Atmospheric distillation, which is designed to fractionate crude oil into differ-
ent cuts such as liquefied petroleum gases, gasoline, kerosine, diesel oil, fuel
oil, etc., is a good example (Fig. 1.2).

Catalytic reforming Gasoline

Petrochemical
feedstocks
Hydrodesulfurization Kerosine
Crude Diesel oil,
7
,
space heating oil
---

-
Vacuum distillate

E+
0 :z
3=
-
- Catalytic cracking

~
Hydrocracking
~

Lube oil plant

b Gasoline
b Diesel oil
Lube oils, waxes
Atm. >o ___) Gas
residue
' Visbreaking
Coking
b Gasoline
Vacuum b Diesel oil

Gas
I ' Hydrodesulfurization
(with conversion)
w Diesel oil

3-b Desulfurized fuel oil

I
-I
Figure
1.2 Fractionating crude oil by atmospheric distillation (refining flowsheet).

Additionally, most catalytic processes include a fractionation stage

designed to yield products that will ultimately meet the desired specifications.
Recycling
Most chemical transformations are characterized by incomplete conversion
for a single pass through the reactor, and a separation stage is required to s e p
arate and recycle the nonconverted components (Fig. 1.3).
For instance, the need to increase the octane number of gasoline stocks to
produce unleaded gasoline means that normal C , and C , paraffins have to be
Chapter 1. THEBASICS OF SEPARATION OPERATIONS 3

Non-converted components

Figure
1.3 Recycling nonconoerted components.

isomerized. The normal paraffins that remain unconverted when they exit
from the isomerization process are separated by adsorption on 5A molecular
sieves and recycled to achieve complete isomerization.
Purification
Certain toxic impurities that are detrimental to the environment or incompat-
ible with the product’s end use can be eliminated by separation processes.
A purification stage may be necessary in anticipation of a later operation,
for example a catalytic stage when the catalyst is sensitive to the impurities in
the feed.
Lastly, some products must meet high purity requirements for later use in
the petrochemical industry baraxylene, for example).

1.2 Main Separation Operations

Generally speaking, a mixture is homogeneous when it forms only one phase
or heterogeneous when it includes two or more phases.
A heterogeneous mixture is usually separated by decanting in an appropri-
ate apparatus.
Volume 2 mainly covers the separation of homogeneous mixtures, which
often requires implementation of complex processes. Separation processes
use an operation or series of operations that are designed to concentrate cer-
tain components selectively in a given fraction.
The required selectivity may be thermodynamic or kinetic in origin.
Thermodynamic selectivity is more commonly utilized.
4 Chapter 1. THEBASICS OF SEPARATION OPERATIONS

As an illustration of the principle of this type of operation, let us consider

the case of a mixture of two components, A and B. Beginning with a homoge-
neous mixture of A and B, we introduce a new phase into the system. The new
phase is the result either of adding a selective external phase (solvent, adsor-
bent) or energy (heating, cooling, compression, expansion, etc.).
The separation factor can be defined by the following equation (King,
1980):

which designates the mole fractions of A and B respectively in phases 1 and 2

by xA1, xBl,xA2and xB2.Designating by convention component A as the one that
is concentrated in phase 1, aAB is greater than 1 and the higher the separation
factor, the more complete the separation (see also Chapter 2: equilibria
between phases).
If the two phases are thoroughly mixed in the contact stage outlined in
Figure 1.4, the separation factor a, is usually lower than its value at equilib-
rium ah. This boundary value is reached in the case of a "theoretical stage",
but for a real stage the value depends on mass transfer kinetics: the slower the
kinetics, the lower the value.

Phase 2 * B(A)
Feed
............................
(mixture of A and B)

Phase 1 * A(B)

Addition of energy andlor

--
Figure
1.4 Separation by phase equilibrium. Simplified flowsheet. The minor component
is between parentheses.

Separation operations based on equilibrium selectivity can be classified

according to the separating agent that is used (matter or energy) and the
nature of the phases in contact:
Chapter 1. THEBASICS OF SEPARATION OPERATIONS 5

Distillation is the most commonly used basic operation. The feed is nor-
mally liquid, but refrigerated distillation can also be used to fractionate a
gaseous mixture. The second phase is generated at the bottom of the col-
umn by heating (vapor reflux) and at the top by cooling (liquid reflux).
The other operations, necessarily more complex, are generally used only
when the difference in volatility between the components to be sepa-
rated is not sufficient to achieve separation by distillation in satisfactory
conditions.
It is also possible in some cases to generate a solid phase by cooling. In
it a component or fraction is concentrated that needs to be separated.
The main refinery application of crystallizationprocesses is in dewaxing.
Adding a solvent can separate components whose volatilities are close
but which belong to different chemical families. For instance, liquid-
liquid extraction is used to separate aromatics from paraffins in a liquid
feed. The same principle is implemented in the case of a solvent absorp-
tion operation that is used for a gaseous feed.
Extractive distillation combines the effect of a phase change by energy
input with the effect of solvent addition to separate a component on the
basis of its volatility and its chemical nature.
Difficult separation operations such as isomer separation (paraxy-
lene/metaxylene, n-lisoparaffins), and deep purification can be carried
out by using an adsorbent. The feed may be gaseous or liquid.
Separation can also be based on kinetic selectivity, i.e. on a difference in
migration velocity of the relevant components. This principle is used in a
membrane separation operation as represented schematically in Figure 1.5 for
a binary mixture of two components A and B.

-
Figure
1.5 Membrane separation. SimpliFied flowsheet. The minor component is between
parentheses.

Separation is based on the difference in diffusion velocity of components A

and B through the selective membrane. Assuming that the transfer of compo-
nent B through the selective membrane is faster, the mixture that goes through
the membrane is enriched in component B, while the residual mixture is
enriched in component A.
6 Chapter 1. THE BASICS OF SEPARATION
OPERATiONS

Membrane separation does not involve a change in phase and the two
existing phases can be either gaseous (gaseous permeation) or liquid (ultrafil-
tration, reverse osmosis). However, a phase change along with a selective
transfer through a membrane can in some cases enhance separation (e.g. in
pervaporation).
The different configurations of the main separation operations in use are
summarized in Table 1.1.

Separation agent Type of phase Operation

L/v Distillation
Energy
L/S Crystallization
G/L Absorption
Solvent L/L Liquid-liquid
Equilibrium selectivity extraction
Eluent L/V or L/G Stripping
Extractive
Energy + solvent L/V distillation
Adsorbent L/S or G/S Adsorption
GIG Gaseous
permeation
Kinetic selectivity Membrane L/L Ultrafiltration/
Reverse osmosis
L/V Pervaporation

Main separation operations.

L: liquid, F oapor, G: gas, S: solid.

The general principles of separation operations are described in a number

of manuals and reference works, more specifically by Schweitzer (1988),
Rousseau (1989, Perry (1984) and King (1980).
In refining, distillation is the most commonly used separation operation
and is presented in detail in Chapter 4. Atmospheric distillation, or topping, is
in particular the basic operation in refining (see Chapter 5).
The other separation processes are used in a much more specific way:
The main application of crystallization is in dewaxing lube oils. In the pro-
duction of petrochemical feedstocks, crystallization is also used in pro-
ducing paraxylene, but separation by selective adsorption is becoming
increasingly important.
Gas-liquid absorption is used in treating refinery gases, mainly to elimi-
nate sour gases.
Chapter 1. THEBASICS OF SEPARATION OPERATIONS 7

Liquid-liquid extraction is mainly used to separate aromatic fractions,

and more especially in processing lube oils and extracting aromatics that
will serve as petrochemical feeds (benzene, toluene and xylenes).
Membrane separations have not been used much on a large scale in refin-
ing, but could become increasingly important in the future.

1.3 Implementation of Separation Processes

1.3.1 Thermodynamic Equilibrium and Kinetic Factors
Table 1.1 shows that the most common separation operations are based on
equilibrium selectivity, also termed thermodynamic selectivity.
This selectivity is due to the fact that at thermodynamic equilibrium, the
distribution of the component(s) to be separated out is different in the exist-
ing phases. The difference is related either to the thermodynamic properties
of the relevant pure substances (boiling, or possibly crystallization, point) or
to a difference in the components' affinity with a selective separating agent
such as a solvent or an adsorbent that has been added to the system. This is
how a polar solvent can be used to separate aromatics from paraffins for exam-
ple by mechanisms that will be discussed in Chapter 2.
In a diagram with coordinates C, concentration in phase 1 and C', concen-
tration in phase 2, the equilibrium equation is represented by a curve going
through the origin (Fig. 1.6). The concentrations are expressed either in kg/m3
or in mol/m3, or expression can be considered in terms of mass fractions or
mole fractions. Thermodynamic models generally use equations expressed in
mole fractions. In contrast, when concentrations are used, it is easier to take
kinetic factors (molecular diffusion) and hydrodynamic factors (such as axial
mixing) into account. When the properties of the two phases are known, espe-
cially the composition and the molar volumes of the components, it is possi-
ble to go from mole fractions to concentrations and vice-versa.
For low concentrations, the equilibrium curve can be identified with its tan-
gent at the origin and thermodynamic equilibrium is characterized by a parti-
tion coefficient, defined as the ratio of the concentrations at equilibrium.
For higher values of concentrations C and C', the equilibrium curve is gen-
erally not linear. In a narrow area of variation in concentrations Cand C', it can,
however, be identified with its tangent and the equilibrium equation is
expressed in the approximate linear form:
C*=mC+q (1.2)
by designating the concentration in phase 2 in thermodynamic equilibrium
with phase 1 as C".
Predicting the respective compositions in the relevant phases at thermo-
dynamic equilibrium has improved greatly during the past twenty years.
Predictions can be made by means of models that lend themselves to use with
a computer. The methods are presented in Chapter 2.
8 Chapter 1. THEBASICSOF SEPARATION OPERATIONS

Figure
1.6 Equation of equilibrium between phases.

In order that the component to be separated can be transferred from phase

1 to phase 2, concentration C in phase 1 must be higher than the equilibrium
concentration C* with phase 2. Concentration C" is the one that, in phase 1,
corresponds at equilibrium to concentration C' in phase 2 (Fig. 1.7).
Chapter 1. THEBASICS OF SEPARATION OPERATIONS 9

The difference between concentrations (C - C*) represents the mass trans-

fer driving force. The mass flux through the interface is directly related to the
driving force by the following equation:
N = K (C- C*) (1.3)
In this equation, N represents the flow transferred per unit of interface area
and K the overall transfer coefficient with respect to phase 1.
Kinetics of mass transfer needs to be known to size the contacting equip-
ment. The main methods used in sizing are presented in Chapter 3.

1.3.2 Contact between Phases

The contact that allows separation to be made between phases is usually
made in a vertical column where the heavier phase flows by gravity counter-
current to the light phase.
Different modes of contact are possible.
Two basic cases, shown schematically in Figure 1.8, can be considered:
a) Contact can be made in a series of stages - this is the case for example
in the tray columns that are widely used in distillation. If contact is made
by means of R stages, each one of them can be indexed in terms of i rang-
ing from 1 to n. In Figure 1.8A, stage i and its adjacent stages i - 1 and i + 1
are shown.
b) Contact can also be made continuously, and this corresponds for exam-
ple to a packed column where one of the phases trickles along the pack-
ing and circulates counter-flow to the other phase. A contact section is in
this case identified by a z coordinate from the entry point of one of the
phases. In order to analyze the concentration profile in each phase, a
material balance equation is derived for a differential column height with
length d z .
In Figure 1.8 the two phases in contact are assumed to circulate in counter-
flow. Other types of contact are possible, particularly cocurrent and cross-
current. These types are shown in Figure 1.9.
Figure 1.9A shows a simplified case of continuous cocurrent contact. As for
continuous counter-flow contact, a given contact section is identified by a z
coordinate from the entry point of the two phases in contact and the concen-
tration profile of each phase can be analyzed by setting up a material balance
equation on a differential column height with length dz.
For continuous crosscurrent contact, such as the one schematically
shown in Figure 1.9B, it is necessary to consider two coordinates u and u, the
two phases circulating perpendicular to each other.
The contact on a tray is roughly crosscurrent, with one of the phases cir-
culating horizontally along the tray and the other phase vertically. In this case,
the concentrations of the two phases can vary according to the two coordi-
nates u and v and the changing concentrations along these two axes are ana-
lyzed by considering an elementary crosssection du du.
10 Chapter 1. THEBASICS OF SEPARATION OPERATIONS

Z ,

A F !
i+l B
dz

Figure
1.8 Contact modes between phases. Counter-flow.
A. Contact with stages. B. Continuous contact.

4
tttttt

Figure
1.9 Cocurrent contact (A) and cross current (B) contact.
Chapter 1. THEBASICS OF SEPARATION OPERATIONS 11

corresponding to the exiting concentration in phase 2. The more effective the

contact between the two phases, the less difference there is between the two
concentrations. Looking at the ideal case of two phases exiting at equilibrium
(theoretical stage), it can be seen that the exiting concentration in phase 1 of
the component transferred from phase 1 to phase 2 approaches zero only if
the partition coefficient (or the flow rate of phase 2) approaches infinity.
In contrast, for the counter-flow contact column in Figure 1.10, it is possi-
ble to transfer the component to be separated from phase 1 to phase 2 com-
pletely, provided there is sufficient flow rate of phase 2.
In this case, by increasing the height h of the column and the residence
time of the two phases, it is possible to get a concentration value in the exit-
ing phase 1 that is close to that which corresponds to equilibrium with the
inlet (rather than the exiting) concentration in phase 2, which may be very
low. In this way, the lowest desirable value for the exiting concentration of
phase 1 can be achieved, provided the column has been sized properly.
Let us consider the example of a gas purification operation by means of a
solvent, that consists in transferring a contaminant (e.g. a sour gas) into the
solvent. It is possible to achieve deep purification in a counter-flow column
and get only gas with an extremely reduced contaminant level at the outlet,
provided that a sufficiently pure solvent has been sent into the top of the
column.

- ~

Figure
1.10 Counter-flowcontact column.
12 Chaoter 1. THEBASICSOF SEPARATION OPERATIONS

The same thinking is applied in a liquid-liquid extraction or distillation

operation. The basic equations that can help determine the number of stages
or the height of the column to achieve a given separation for the simplest
cases are presented in Chapter 3.

1.3.3 Fixed Bed Adsorption Operations

In an adsorption operation the movement of the solid phase gives rise to seri-
ous problems (solid phase displacement devices, mechanical resistance of the
adsorbent, etc.). As a result, the adsorbent is generally implemented in a fixed
bed.
The adsorptiondesorption operation is carried out batch-wise as shown in
the simplified flowsheet in Figure 1.11.

U U
a, al
D D
c c
c c
gc5 gc
5: 5:
2 2

--
Figure
1.1 1 Separation by fixed bed adsorption.

The unit includes at least two beds, with one running in the adsorption
phase, while the other is in desorption and switchover by a set of valves. For
example if bed C, is operating in the adsorption phase, valves V,, and V13are
open whereas valves VI2 and V,, are closed. In the case of a mixture of two
Chapter 1. THEBASICS
OF SEPARATION OPERATIONS 13

components A and B, where component B is retained selectively by the adsor-

bent, it is possible during the adsorption phase to get component A pure at the
exit of bed C, with component B fixed on the bed. During the desorption stage
which is carried out at the same time on bed C, a desorption agent D is
pumped in by keeping valves V24 and V2, open whereas valves V14 and V,, are
closed. During the desorption phase, component B with desorption agent D is
obtained at the exit of bed C,.
In practice, the operating procedure is often more complex and may
include a purge step designed to keep from polluting the effluent from the des-
orption stage by a large amount of fluid phase remaining in the fixed bed after
the adsorption stage. As a result, the unit may have three or four beds oper-
ating simultaneously.
This type of process is well suited whenever the adsorbent has high selec-
tivity versus the component to be separated out. Fixed bed adsorption is used
in particular to dehydrate refinery gases and to separate normal paraffins from
isoparaffins in light gasolines in order to improve the octane number and send
only isoparaffins to the gasoline pool.
When selectivity is low, fixed bed adsorption is not so well suited and it is
important to approach counter-flow conditions. Here a counter-flow contact
can be simulated without moving the solid phase by changing the inlet and
exit points of the feed, the desorption agent and the effluents. This is the solu-
tion used in particular to separate paraxylene by means of X or Y zeolites. The
applications of adsorption separation operations in refining are described in
Chapters 10 and 11.

1.3.4 Membrane Separations

Membrane separation operations are still seldom used in refining today. The
potential applications of the different ultrafiltration, pervaporation and
gaseous permeation technologies are described in Chapter 12. Since selectiv-
ity is kinetic, a membrane separation module separates a mixture into prod-
ucts whose composition may be very far from thermodynamic equilibrium.
Provided the membrane is sufficiently permeable and selective, a single mod-
ule can achieve the desired separation. In practice, in the present state of the
technology, membranes have proven to be ill suited to purification operations,
at least in the case of homogeneous mixtures. In the main applications that are
currently in use, they are combined with more conventional separation tech-
nologies based on a contact between phases.

1.4 Recent Progress in Separation Processes

Separation operations have undergone constant and relatively continuous
development in recent years. New separation processes have appeared to
meet the present-day needs of the refining industry: environmental protection,
14 Chapter 1 THEBASICS OF SEPARATION OPERATIONS

motor fuel reformulation, development of conversion units, changing outlets

for petrochemicals (Blytas, 1986).
Reduction of investment and operating costs is also a strong incentive to
innovate and introduce better performing processes. This means more thor-
ough integration in particular, with two different functions being performed by
the same equipment. The use of reactive distillation to produce ethers is a
case in point.
A lot of progress has additionally been made in the design of equipment
and processes: real time optimization and automation of units, more efficient
phase contact equipment (structured packing, low pressure drop trays, etc.).
The basic design of the process itself has often evolved, particularly
regarding the regeneration stage of the separating agent (solvent or adsor-
bent).
For instance in deasphalting, the solvent may be regenerated in conditions
close to the critical point to significantly reduce the amount of energy
required.
The most fundamental progress is in the implementation of new separating
agents (solvents, adsorbents, membranes). Accordingly, membrane separa-
tion processes, which have had only limited use in refining in the past, should
become increasingly important in the future. The latest technical advances
will be discussed in Chapter 12.