Membrane Separation

Processes
 retentate permeate

FEED

driving force

membrane
MEMBRANE:
selective barrier
barrier between two phases
a phase that forms barrier hinders mass transport but enables limited and controlled passage of
certain components
liquid, solid, gas character (or combination of all)
Membrane separates feed into:
Retentate (concentrate)
= enriched with substances retained by the membrane

Permeate (filtrate)
= a stream passing through the membrane, devoid of
substances retained by the membrane
Basic mechanisms of substance separation on
Membrane
➢ Different particle size (sieving effect)
➢ Different charge of mixture components
➢ Different diffusivity
➢ Different solubility of components in the membrane

Aspects of membrane process's classification

➢ Driving force
➢ Type of the membrane
➢ Properties of separated particles
Driving force of membrane processes
➢ Pressure
Gradient of vapor pressure
➢ Concentration
pervaporation (feed is liquid)
➢ Chemical potential vapor permeation (feed is vapor)
➢ Electric potential
Pressure gradient
gas permeation (feed & permeant are gases)
reverse osmosis (feed & permeant are liquids)

Temperature gradient
thermoosmosis

Concentration gradient
dialysis (osmosis, liquid permeation)
pertraction

Gradient in electric potential

electrodialysis (ion-selective membrane)
Properties of separated particles
➢ Size
➢ Shape
➢ Charge
Principle:
➢ Liquid or gaseous mixture passes
– permeates
➢ Through a thin semi-permeable solid barrier
– membrane
➢ Membrane separation processes are
– rate controlled processes
Permeation rates differ due to:

➢ Different particle/solute sizes or molar masses

(sieving effect), different shapes.

➢ Different interactions of components with membrane matrix material.

➢ Different electrical charges of compoonents.
➢ Combinations of factors and other factors.
Classification of membrane separation processes
Filtration set-up (dead-end, cross-flow)
Classification according to the Membrane

(1) microfiltration (Retain particles as small as 0.1 mm),

(2) ultrafiltration, and
(3) reverse osmosis

 Nanofiltration !!
 Membrane process's characteristics:
Driving force Retentate Permeate
Chemical Substances dissolved
Osmosis in a solution, water Water
potential
Concentration Small molecules,
Dialysis Big molecules, water
gradient water
Suspended solids Substances dissolved
Microfiltration (MF) Pressure (e.g. yeasts), water in a solution, water
Big particles,
molecules (bacteria), Small molecules,
Ultrafiltration (UF) Pressure
water water

Small molecules, Monovalent ions, non-

Nanofiltration (NF) Pressure water, bivalent ions, dissociated acids,
dissociated acids water
Reverse osmosis
Pressure Substances, water water
(RO)
Dissolved non- Ionised substances
Elektrodialysis ionogenic
Potential dissolved in a solution,
(ED) substances, water water
Non-volatile Volatile small
Pervaporation (PV) Partial pressure molecules, water molecules, water
 Membranes .
Separation properties of the membrane are given by:
permeability, selectivity and cut-off

A) Permeability
Permeation = a transport of atoms, molecules and ions in a permeable media due
to the gradient (of concentration, temperature, pressure, electric potential)

b) Porosity
Porosity or void fraction is a measure of the void (i.e. "empty") spaces in a material,
and is a fraction of the volume of voids over the total volume, between 0 and 1, or as a
percentage between 0 and 100%.
  k  P
Darcy's law: v =  P  v m ( R + R ) = Lp P
j =
 m  m p

where  = thickness the membrane

m = viscosity of the permeate
jv = the volume of solvent per area per time (or the
solvent velocity)
Rm = membrane resistance
Rp = resistance of the polarized boundary layer
Lp = the permeability
A flow through ideal semipermeable membrane can
expressed as

J = Lp • (PT - F)

J = permeate flow, expresses a velocity of component passage through the membrane

Lp = permeation coefficient (reciprocal resistance)
PT = trans-membrane pressure
F = osmotic pressure of solvent

➢ The pressure that is needed to press water through a membrane is called Trans
Membrane Pressure (TMP)
➢ The TMP is defined as the pressure gradient of the membrane, or the average feed
pressure minus the permeate pressure
➢ Osmosis is simply the flow of solvent molecules through a semipermeable
membrane (one that blocks the passage of dissolved substances—i.e., solutes)
from higher solvent concentration to lower solvent concentration.

➢ The osmotic pressure of a solution is the minimum amount of pressure needed

to prevent water from flowing into it across a semipermeable membrane

➢ The osmotic pressure of a solution is the pressure difference needed to stop the
flow of solvent across a semipermeable membrane

➢ Reverse osmosis is a process that is used to take water molecules from an area
where they are in low concentration to an area where they are in high
concentration . It is the exact opposite of osmosis.
 ANALYSIS OF ULTRAFILTRATION (4/4)

jv = LP ( P −  )

where  = the osmotic pressure

 = a reflection coefficient

*=1
 The membrane rejects all solutes.
*=0
 The membrane freely passes both solvent and
solute.
Estimation of Osmotic Pressure—the van’t Hoff
equation (1885)
 = c1RT
where c1 = molar concentration of solute (mol/L)
R = gas law constant (0.082 L atm mol−1 K −1)
T = absolute temperature

• The van’t Hoff equation should only be used at low

molar concentrations.
 At higher concentrations, the van’t Hoff equation
significantly underpredicts the osmotic pressure.
C) Selectivity

Frequency (number of pores)

Pore size

The membrane selectivity, α, is the ratio of most permeant to least

permeant component permeabilities, i.e. α > 1.
Cutt-off:
 Flow patterns in membrane contactors
(flows define the performance)

differential
balances
Membrane classification:

➢ Membrane origin (natural, synthetic)

➢ Membrane structure (porous, non-porous, morphology)surface
 Membrane structure:

Microporous membrane

➢ isotropic
➢ anisotropic

isotropic membrane
Asymmetric membrane (with active layer)

A thin active layer is deposed on a support

 Composite (sandwich) membrane

Several layers – every layer produced by different technique

Membrane’s material

Organic
cellulose acetate (low costs, low price, wide range of pore sizes, hydrophilic character
– reduced fouling)
polyamide (low resistance to Cl2, biofouling)
polysulphone (high temperature, pH and chemical Cl2
resistance)
other polymers: nylon, PVDF, PTFE, PP, polycarbonate)

Inorganic (= mineral, ceramic)

also metallic (stainless steel)
support layer: ceramic, Al2O3, TiO2
separation layer: TiO2, zirconium, carbon-titanium, carbon-zirconium
Ceramic membranes:

positives:

➢ inert toward most of chemical solutions (exceptions: HF, H3PO4

– Al membranes)
➢ high temperature resistance
➢ wide pH range(1 – 13)
➢ high pressure resistance (1 MPa)
➢ high lifetime
➢ back-flushing possible

drawbacks:
➢ high pore size (UF, MF, opened NF membranes)
➢ high pump discharge needed (2 – 6 m/s tangential velocity)
➢ high producingcosts – high prices (high investment costs)
Membrane´s shapes/ Modules

➢ Plate
➢ Tubular (channel inside > 4 mm)
➢ Capillary (low diameter)
➢ Hollow fibres (inner diameter 0.2-3 mm)
➢ Spiral-wound
➢ Pleated sheet cartridges (dead-end filtration)
Tubular membrane
Membralox ®

Cheryan M.: Ultrafiltration and

Microfiltration Handbook, Technomic Pub.
Co., 1998
Pleated-sheet membrane

Cheryan M.: Ultrafiltration and Microfiltration Handbook, Technomic Pub. Co., 1998
Hollow fibres

 Asymmetric structure

 Microporous structure

Cheryan M.: Ultrafiltration and Microfiltration Handbook, Technomic Pub. Co., 1998
Biomembranes

Enzyme immobilization on the membrane´s surface

- adsorption, chemical bonds

Liquid membranes

a) ELM = Emulsion Liquid Membrane – immiscible liquids

b) ILM = Immobilized Liquid Membrane
Filtration theory:
Pressure-driven processes: driving force = pressure gradient

Feed Řez membránou

Cross section of the membrane

Active
layer
Support
layer
Permeate
solvatace

PT = PF − PP
Pressure drop :
Osmotic pressure within membrane:  = F − P

Real driving force: = PT − 

Driving force of the process: = PT
Concentrationpolarization:

Higher concentration of components in boundary layer close to the

membrane comparing to bulk
Forming of concentration profile – under extreme conditions – gel
or precipitate formation, secondary membrane
Increased membrane resistance
Pressure-driven membrane processes

Driving force = pressure gradient

Real driving force: ➔ reduced by pressure drop in polarization

layer, secondary membrane or gel layer

Convective flow Js JS = J  c1
dc
Js = D 
Diffusion velocity
(neglected concentration
gradient c) dx
dc
Filtration equation:
G = specific membrane − G  c2 + D  + G   c1 = 0
permeability dx
Permeate flow Diffusion Convection
 Discharge, fouling, membrane cleaning

Discharge = permeate flow (l.h-1.m-2)

Permeate flow rate drop within the filtration – problem!

80

The drop is caused by: 70

concentration 60
(l/hm2)

polarization
silting of membrane 50
Permeatevýflkuoxn

pores called 40
fouling effect 30

20
0 20 40 60 80 100 120 140 160 180 200
čas (min)
time
 Factors influencing fouling effect
a) Membrane properties
Pore size (permeability)
Active layer thickness
An affinity of dissolved components for the membrane

b) Character of filtered medium

Viscosity
Ionic strength
pH
Density
Concentration
Reactivity of molecules with the membrane
Shape and size of separated molecules
Sample pre-treatment – precipitation, addition of ballast material,
pre-filtration (to improve the size ratio between molecules and pores)
c) Conditions of process operation
Pressure (increasing during filtration, consider irreversible
fouling! – limit pressure)
Temperature
Hydrodynamics (character of a flow along membrane) –
maintain turbulent flow:

Membrane purification
Back-flushing
CIP (Cleaning in place) chemicals, detergents, high temperature
Steam sterilization (ceramic)
Chemically: ozone, Cl2 – corrosive action, NaClO, HNO3, NaOH –
ceramic membrane)
Mechanically (after removal, sheet membranes)
Enzymatically
Dialysis
A transport of small molecules
(dissolved in liquid) across the
membrane (from hypertonic to
hypotonic media); kidneys

driving force: concentration

gradient

Osmosis
A transport of solvent across the
membrane, which retains dissolved
molecules (from hypotonic to
hypertonic media)
driving force: chemical potential
gradient
Reverse osmosis(RO)

A process reversal to osmosis.

Driving force: pressure gradient

through the membrane, needs to be
higher than osmotic pressure

high working
pressure (3 – 10 MPa).
Peffective > Posmotic
Microfiltration (MF)
Separates particles within th e range of 10 – 10-1 mm (i.e.
bacteria, yeasts, suspended solids, high molecular substances;
M>106)
Particles bigger than 5 - 10 mm are better to separate by
conventional filtration (dead-end filtration).
Pressure difference needed: 0.02 do 0.5 MPa
Refining technique, suspended solids are separated from
dissolved substances.

Ultrafiltration (UF)
Separates particles within the range of 10-2 – 10-3 mm (i.e. bacteria,
viruses, colloids, macromolecular substances, MWCO 5000 – 500
000)
Pressure difference needed: 0.1 do 1 MPa
a technique for mutual concentration and fractionating of
molecules or fine colloid suspensions
A sieving effect is a separation mechanism of both processes (MF
and UF).
Nanofiltration (NF)
Relatively new medium pressure membrane process for
separation of compounds within a size of 10-3 – 10- 4 mm.
Separation mechanism:
1. sieving effect (large molecules, e.g. sucrose)
2. electrostatic forces between membrane and particles
present in solution (ion separation).
Most of commercially produced NF membranes are negatively
charged.
NF membranes separate dissociated compounds from non
dissociated (e.g. organic acids pass the membrane more easy at
low pH, but are retained at high pH in a form of their salts,
MWCO < 500).
Membrane cut-off expressed as molecular weight or in Daltons
Operation pressure 2 - 4 MPa
 Pervaporation (PV)
The term pervaporation is derived from the two steps of the process: (a) permeation
through the membrane by the permeate, then (b) its evaporation into the vapor phase.

➔ Liquid feed – wet side of the membrane – swelling (atmospheric pressure)

➔ Gas permeate – dry side of the membrane – almost dry (low pressure of
vapours)

Principle – phase change

1. sorption
2. diffusion
3. desorption and evaporation

Solution
ROZTOK MEMBRANE
MEMBRÁNA Permeate
PERMEÁT
Sorp c e PiP
CiP
PiF
c iF Desorp c e

Wetstrana
Vlhká side Dry side
Suchá strana
Driving síla: m m
Hnací force: i jako p as pi
i expressed
vyjádřen i
 Pervaporation (PV)

Driving force: chemical potential difference (expressed as a partial

pressure), the pressure behind the membrane is lower, which causes
evaporation followed by condensation of vapors.

Separation mechanisms: different diffusivity

Separation of volatile compounds (hexane, toluene, ethanol) from
liquid mixtures (dehydration of organic solvents), separation of
azeotropic mixtures, pollutants and impurities removal, solution
concentration
PV membrane:
Composite membranes (active layer)
Hydrophobic membranes – separation of organic solvents
(polysulfone, polydimethylsiloxane, polyamide)
Hydrophilic membrane – polar solutions (water, water
vapours) – glass, crystalline polymers of hydrophilic nature)
 Pervaporation set-up
Modules - membranes: Capillary, hollow fibre, planar, spiral wound, tubular
Minimal piP:
1. Permeate side – maintain a vacuum (20 – 30 mbar)
2. Permeate side purified by sweeping gas – removal of desorbed compounds)
3. Temperature reduction on permeate side – lower pi of component i (-20 °C,
opt. T = 50 °C)
R R R
ad 1. Vacuum
V ad 2. ad 3.
pump
pump

F K F K F P
condenser P condenser P
membrane The effect of temperature
Feed, l ➢Permeate side - latent heat of vaporization causes
cooling
➢Isothermic separation (tF = tP) the flow profile
Permeate, g
through the membrane is shorter – faster process.
Gas permeation
An analogue of pervaporation,
but uses non-porous
membranes Feed (higher pressure)

The same phase on both size

of the membrane (gas)
Concentration gradient is
provided - sweeping gas
Polymeric
Separation mechanism: membrane

different velocity of gas

permeation (sorption,
diffusion, sieving effect,
desorption)

Permeate (lower pressure)

 Membrane distillation (MD)
Uses both distillation and membrane separation – porous hydrophobic membrane
(permeable for water vapours but not for liquid water
Driving force: temperature and pressure gradient

Principle:
One side of the membrane:
liquid is heated
– evaporation, vapors pass
through
Other side: cooling –
vapor condensation – water
removing
The separation is influenced only by the equilibrium liquid-vapours (the
membrane has no effect):
Limit – unsuitable for azeotropic mixture separation (azeotropic point =
identical composition of gas and liquid phase)
Necessary requirement – the membrane must not be sodden – pore
blocking, application for hydrophobic solutions
Velocity of permeation is given by t – high t ensures speed and selectivity
Membranes for the membrane distillation
➢ Hydrophobic (non-polar)
➢ Microporous
➢ Material: PTFE (polytetrafluorethylene = teflon)
PP (polypropylene)

Applications
Separations
➔ Mixtures EtOH – water (up to 30 - 40 vol. %)
➔ Water solutions of salts – desalination (e.g. water for
heating systems)
➔ Sea water desalination
Drawbacks
➔ Low selectivity
➔ Limited application
[Example] Ultrafiltration of a well-stirred suspension containing 0.1 vol% yeast
suspension gives a flux of 36 gal/ft2-day under a pressure difference of 130 psi.
(a) What is the value of Lp? (b) What is the water velocity through the
membrane?
Solution:

(a) The yeast cells have a very high molecular weight, so that their molar
concentration and the resulting osmotic pressure will be small.

jv = LP ( P −  ) = L p P
gal
36 2
= L p (130 psi)  Lp = 0.28 gal/ft2-day-psi
ft - day
[Example] Ultrafiltration of a well-stirred suspension containing 0.1
vol% yeast suspension gives a flux of 36 gal/ft2-day under a pressure
difference of 130 psi. (a) What is the value of Lp? (b) What is the water
velocity through the membrane?
Solution (cont’d):

(b) The water velocity through the membrane = jv

gal  3785 cm 3   ft 2  day 

= 36 2
ft - day  gal   (30.48 cm) 2   24  3600 s 

= 0.0017 cm/s
* Molecular Weight of Yeast Cell
d = 8 mm
r = 1.05 g/cm3

d 3
V= = 2.7  10−10 cm3
6
 MW = (6.02  1023)(2.7  10−10)(1.05)
= 1.7  1014
Estimation of Surface Concentration
Material balance:
Rate of solute accumulation
= (rate of solute flow in)
− (rate of solute diffuse out)

dc
 0 = cj v + D
dx
B.C. 1: x = 0, c = c10
B.C. 2: x = , c = c1

D c10
 j v = ln
 c1
[Example] We are carrying out the ultrafiltration of chymotrypsin in a spiral
wound module at a rate of 1.3  10−5 cm/s. The solution concentration is 0.44 wt%,
the protein’s diffusion coefficient is 9.5  10−7 cm2/s, and the boundary layer is
about 0.018 cm thick. How high is the surface concentration?
Solution:

D c10
j v = ln
 c1

9.5  10 −7 c10 c10

1.3  10 −5
= ln  = 1.3
0.018 c1 c1
A Simplified Case of Batch Concentration
Assumptions:
(1) There is little concentration polarization.  c10 = c1
 When (c10 – c1)/c1 < 0.1, the effect of concentration polarization can
be neglected.
(2) The membrane rejects all the solute.   = 1

dV  c1 RT 
= − Ajv = − AL p ( P −  ) = − AL p P1 − 
dt  P 
(V = liquid volume in the system)

dV  n1 RT / P 
 = − AL p P1 − 
dt  V 
 A Simplified Case of Batch Concentration (2/2)

dV  n1 RT / P 
= − AL p P1 − 
dt  V 
I. C.: t = 0, V = V0

1   n1 RT   V0 − n1 RT / P 
t= (V0 − V ) +   ln  
AL p P   P   V − n1 RT / P 
[Example] We want to ultrafilter 840 L of a solution containing 0.061 wt% of a
protein used as a vaccine for herpes. This protein has a diffusion coefficient of
1.1  10−6 cm2/s and a molecular weight of 16,900. We would like to get the
concentration up to about 2% by weight. The ultrafilter, which we hope to use,
has eight hollow fiber cartridges, each of which has a surface area of 1.20 m2.
It is cooled to 4C during the operation. The membrane in these cartridges
gives an initial flux of 5.7  10−5 cm/s under a pressure drop of 31 atm.
Assuming negligible concentration polarization, estimate the time to complete
this filtration.
Solution:

1   n1 RT   V0 − n1 RT / P 
t= (V0 − V ) +   ln  
AL p P   P   V − n1 RT / P 
[Example] We want to ultrafilter 840 L of a solution containing 0.061 wt% of a
protein used as a vaccine for herpes. This protein has a diffusion coefficient of 1.1
 10-6 cm2/s and a molecular weight of 16,900. We would like to get the
concentration up to about 2% by weight. The ultrafilter, which we hope to use, has
eight hollow fiber cartridges, each of which has a surface area of 1.20 m2. It is
cooled to 4C during the operation. The membrane in these cartridges gives an
initial flux of 5.7  10-5 cm/s under a pressure drop of 31 atm. Assuming negligible
concentration polarization, estimate the time to complete this filtration.
Solution (cont’d):

1   n1 RT   V0 − n1 RT / P 
t= (V0 − V ) +   ln  
AL p P   P   V − n1 RT / P 

 0.061 % 
V0 = 840 L ; V = 840 L    = 25.62 L
 2% 
0.061%  (840 103 ) g
n1 = = 0.03 mol
16900 g/mol
n1 RT (0.03 mol)(0.082 L - atm/mol-  K)(277  K)
= = 0.02 L
P 31 atm
(To be continued)
[Example] We want to ultrafilter 840 L of a solution containing 0.061 wt% of a
protein used as a vaccine for herpes. This protein has a diffusion coefficient of 1.1
 10-6 cm2/s and a molecular weight of 16,900. We would like to get the
concentration up to about 2% by weight. The ultrafilter, which we hope to use, has
eight hollow fiber cartridges, each of which has a surface area of 1.20 m2. It is
cooled to 4C during the operation. The membrane in these cartridges gives an
initial flux of 5.7  10-5 cm/s under a pressure drop of 31 atm. Assuming negligible
concentration polarization, estimate the time to complete this filtration.

Solution (cont’d):
1   n1 RT   V0 − n1 RT / P 
t=  0
(V − V ) +   ln  
AL p P    P   V − n1 RT / P 

Initial flux: jv = LP ( P −  )
0.03 mol  L - atm 
 = c10 RT =  0 . 082   ( 277 
K ) = 8.11  10−4 atm
840 L  mol- K 
5.7  10-5 = Lp (31 − 8.11  10-4)
 Lp = 1.84  10-6 cm
s - atm
(To be continued)
[Example] Ultrafiltration of a vaccine for herpes.
The ultrafilter, which we hope to use, has eight hollow fiber cartridges, each of
which has a surface area of 1.20 m2.
Assuming negligible concentration polarization, estimate the time to complete
this filtration.
Solution (cont’d):

ALpP = (8  1.2  104)(1.84  10−6)(31) = 5.48 cm3/s = 5.48  10−3 L/s

1   n1RT   V0 − n1RT / P 
t=  0
(V − V ) +   ln  
ALp P   P   V − n1RT / P 

1   840 − 0.02 
= (840 − 25.62) + 0.02 ln  25.62 − 0.02 
5.48  10 − 3   
= 1.486  105 s = 41.3 h

 840 − 0.02 
Note : 0.02 ln   = 0.07
 25.62 − 0.02  #
CONCENTRATION POLARIZATION IN
ULTRAFILTRATION
Material balance for solute in the
boundary layer:
(Flux of solute in due to convection)
= (flux of solute out due to
diffusion)

c1 
dc jv
 j c = − D dc 
v
dx
c c = − D 0 dx
10

c10 j v  j v
 ln = =
c1 D kc
 c10 j v  j v
ln = =
c1 D kc
Correlation for flow inside pipes (for turbulent flow):

NSh = 0.0096N Re NSh = 0.082N Re

0.913 0.346 or 0.69 0.33
NSc NSc

k d
c
NSh (Sherwood number) = (d: pipe diameter)
D
NRe (Reynolds number) = rvd
m
NSc (Schmidt number) =
m
rD
 cross - sectional area 
* Equivalent diameter of the flow channel = 4  
 wetted perimeter 
c10
* Concentration polarization becomes severe when  10.
c1
 CONCENTRATION POLARIZATION IN ULTRAFILTRATION (3/3)

* Prediction of liquid diffusivities—Stokes-Einstein equation (for particles or

large spherical molecules, VA > 500 cm3/mol)

7.32  10 −16 T
D=
r0 m

T = absolute temperature, K
r0 = radius of the particles, cm
m = viscosity of solution, cP
VA = molar volume of solute as liquid at its normal boiling point,
cm3/g mol
[Example] Equipment is available for ultrafiltration of a protein solution at
constant volume to remove low molecular weight species (achieved by the addition
of water or buffer to the feed in an operation called diafiltration). The flow
channels for this system are tubes 0.1 cm in diameter and 100 cm long. The
protein has a diffusion coefficient of 9  10−7 cm2/s. The solution has a viscosity of
1.2 cp and a density of 1.1 g/cm3. The system is capable of operating at bulk
stream velocity of 300 cm/s. At this velocity, determine the polarization modulus
(c10/c1) for a transmembrane flux of 45 L m−2 h−1.

c10 jv kc d
Solution: ln =  NSh = 0.082N Re
0.69
NSc
0.33
 N Sh =
c1 k c D
 g  cm 
1.1 3  300 (0.1 cm )
rvd  cm  s 
N Re = = = 2750
m 1.2  10 - 2
g
cm - s (To be continued)
 Example: Determination of the polarization modulus (c 10/c1).

Solution (cont’d):

NSh = 0.082N Re N Re = 2750

0.69 0.33
NSc ;

g
1.2  10 − 2
m cm - s
N Sc = = = 1.21  10 4

rD  g  − 7 cm 
2

1.1 3  9  10 
 cm  s 
kc d
N Sh = = 0.082( 2750) 0.69 (1.21  10 4 ) 0.33 = 431
D
cm 2 −7
kc d 9  10
N Sh =  D s = 3.88  10 − 3 cm
k c = N Sh  = 431 
D d 0.1 cm s

(To be continued)
 Example: Determination of the polarization modulus (c 10/c1).

Solution (cont’d):

cm
kc = 3.88  10−3
s

c10 jv
ln =
c1 k c
 L  1000 cm 3 m2 h 
 45 2   4   
c10  jv   m -h  L 2
10 cm 3600 s  
  
= exp   = exp = 1.38
c1  kc   cm 
 3.88  10 -3

s
 
#
[Example] A tubular membrane with a diameter of 2 cm and a water
permeability of 250 L/m2-h-atm is used for ultrafiltration of cheese whey. The
solution velocity is 1.5 m/s and the protein concentration is 10 g/L. The whey
proteins have an average diffusivity of 4  10−7 cm2/s, and the osmotic pressure in
atm is given by Jonsson’s equation:
 = 4.4  10−3c − 1.7  10−6c2 + 7.9  10−8c3
where c is the protein concentration in g/L. Calculate the applied pressure if the
permeate flux is 10−3 cm/s. Assume the protein rejection is 100 percent and the
bulk solution has the same density and viscosity as water.

Solution:

jv = LP ( P −  )  Need c10 to estimate  .

c10 j v
ln =  Need kc.
c1 k c
(To be continued)
[Example] d = 2 cm; Lp = 250 L/m2-h-atm; v = 1.5 m/s; c1 = 10 g/L; D =
4  10-7 cm2/s;  = 4.4  10-3c10 − 1.7  10-6c102 + 7.9  10-8c103 ;
r = 1 g/cm3; m = 1 cp; jv = 10-3 cm/s; P = ?
Solution (cont’d):

rvd (1 g/cm 3 )(150 cm/s)(2 cm)

N Re = = = 30,000
m 0.01 g/cm - s

m 0.01 g/cm - s
N Sc = = = 25,000
rD (1 g/cm )( 4  10 cm / s)
3 −7 2

kc d
N Sh = = 0.0096NRe 0.913NSc0.346 = 3.9  103
D

 N Sh D (3.9  103 )( 4  10 −7 )
kc = = = 7.8  10 − 4 cm/s
d 2

(To be continued)
[Example] d = 2 cm; Lp = 250 L/m2-h-atm; v = 1.5 m/s; c1 = 10 g/L; D =
4  10-7 cm2/s;  = 4.4  10-3c10 − 1.7  10-6c102 + 7.9  10-8c103 ;
r = 1 g/cm3; m = 1 cp; jv = 10-3 cm/s; P = ?
Solution (cont’d):
kc = 7.8  10−4 cm/s
−3
c10 j v c 10
ln =  ln 10
= −4
 c10 = 36.04 g/L
c1 k c 10 7.8  10
 = 4.4  10-3c − 1.7  10−6c2 + 7.9  10−8c3 = 0.16 atm

L 1000 cm 3   m2  h 
Permeability, Lp = 250 2
m - h - atm  L  10000 cm 2   3600 s 
  
= 6.94  10-3 cm/s-atm

Permeate flux, jv = LP ( P −  )
 0.001 = 6.94  10−3 (P − 0.16)  P = 0.304 atm
#
CONCENTRATION POLARIZATION IN
ULTRAFILTRATION WITH PARTIAL
REJECTION OF SOLUTES
dc
jv c + D = jv c2
dx
B.C.1: c = c10 at x = 0
B.C.2: c = c1 at x =

The solution is:

c10 − c 2 j v  j v
ln = =
c1 − c 2 D kc
Recall:
dc D c10
cjv + D =0  j v = ln
dx  c1
CONCENTRATION POLARIZATION WITH PARTIAL REJECTION OF SOLUTES (2/4)

c10 − c 2 j v  j v
ln = =
c1 − c 2 D kc

 c10 − c2 = (c1 − c2)exp

jv
kc
jv
 c10 = c2 + (c1 − c2)exp
kc

 c2  c2  jv   j 
 c10 = c1  + 1 −  exp  = c1 1 − R + Rexp v 
 c1  c1  kc   kc 

c2
where R = 1 − , the fraction rejected
c1
 CONCENTRATION POLARIZATION WITH PARTIAL REJECTION OF SOLUTES (3/4)

 jv 
c10 = c1 1 − R + Rexp 
 kc 
c2
R = 1−
c1

Assume the permeate concentration, c2, is in equilibrium with c10, i.e., c2 = Kc10.

c10 c j 1 j
= 2 = 1 − R + Rexp v  (1 − R ) = 1 − R + Rexp v
c1 Kc1 kc K kc

 1 R j 1 1− K R j
=1+ exp v  −1 = = exp v
K 1− R kc K K 1− R kc

\ 1− R K j
= exp v
R 1− K kc
 CONCENTRATION POLARIZATION WITH PARTIAL REJECTION OF SOLUTES (4/4)

1− R K jv
= exp
R 1− K kc
c2 = Kc10

 K is a constant, and R varies

with jv.

 (1) The rejection, R, approaches (1 − K) as the permeate flux, jv, approaches

zero.
(2) The rejection decreases with increasing the permeate flux.
[Example] Ultrafiltration tests with a 1.5-cm tubular membrane at NRe = 25,000 gave
a permeate flux of 40 L m−2 h−1 and 75 percent rejection for a 5 percent polymer
solution. The polymer has an estimated diffusivity of 5  10−7 cm2/s. The bulk
solution has the same density and viscosity as water. Predict the fraction rejected for
a flux of 20 L m−2 h−1. What is the maximum rejection?
Solution:

1− R K j
= exp v  Need K and kc.
R 1− K kc
m 0.01 g/cm - s
N Sc = = = 20,000
rD (1 g/cm )(5  10 cm /s)
3 −7 2

kc d
N Sh = = 0.0096NRe 0.913NSc0.346 = 3060
D
N Sh D (3060)(5  10 −7 )
 kc = = = 1.02  10 −3 cm/s
d 1.5 (To be continued)
[Example] d = 1.5 cm; NRe = 25,000; jv = 40 L m-2 h-1; R = 0.75; D = 5  10-7
cm2/s. Predict the fraction rejected for a flux of 20 L m−2 h−1. What is the
maximum rejection?
Solution (cont’d):

1− R j
=
K
exp v ; kc = 1.02 10 −3 cm/s
R 1− K kc

Permeate flux, jv = 40 L m−2 h−1 = 1.11  10−3 cm/s

1− R K j 1 − 0.75 K 1.11  10 −3
= exp v  = exp
R 1− K kc 0.75 1− K 1.02  10 −3
 K = 0.101

(To be continued)
[Example] d = 1.5 cm; NRe = 25,000; jv = 40 L m-2 h-1; R = 0.75; D = 5  10-7
cm2/s. Predict the fraction rejected for a flux of 20 L m−2 h−1. What is the
maximum rejection?
Solution (cont’d):

1− R K jv
= exp ; kc = 1.02  10−3 cm/s ; K = 0.101
R 1− K kc

If the permeate flux is 20 L m−2 h−1 or 0.556  10−3 cm/s,

1− R 0.101 0.556  10 −3
= exp  R = 0.84
R 1 − 0.101 1.02  10 −3
The maximum rejection occurs as the flux approaches zero.

 Rmax = 1 − K = 0.899
#
REVERSE OSMOSIS
• Performance variables of reverse osmosis (2/5):
The rate of water passage through a semipermeable membrane is defined by:

Qw = kwA(P −  )/

Qw = water flow rate through the membrane

kw = membrane permeability coefficient for water
A = membrane area
P = hydraulic pressure differential across the membrane
 = osmotic pressure differential across the membrane
 = membrane thickness
 REVERSE OSMOSIS (6/19)

• Performance variables of reverse osmosis (3/5):

The rate of salt flow through the membrane is given by:

Qs = ksAc/

Qs = flow rate of salt through the membrane

c = salt concentration differential across the membrane
ks = membrane permeability coefficient for salt
 REVERSE OSMOSIS (7/19)

• Performance variables of reverse osmosis (4/5):

Qw = kwA(P −  )/ ; Qs = ksAc/

  The rate of water flow through the membrane is proportional to the

pressure differential across the membrane.
 The rate of salt flow is proportional to the concentration differential,
and is independent of the applied pressure.
 An increase in operating pressure will increase the water flow without
changing the salt flow.
 REVERSE OSMOSIS (8/19)

• Performance variables of reverse osmosis (5/5):

Recovery (or conversion) is defined by:
Y = 100Qp/Qf
Y = % recovery
Qp = product water flow rate
Qf = feed water flow rate

Salt passage is defined by:

SP = 100cp/cf
SP = % salt passage
cp = salt concentration in the product stream
cf = salt concentration in the feed stream
CONCENTRATION POLARIZATION IN
REVERSE OSMOSIS
• Concentration polarization has two deleterious effects:
(1) The increase in c10 increases  .
Qw = kwA(P −  )/
 It necessitates a greater applied total pressure to produce a given
water flux across the membrane.
(2) The increase in c10 serves to increase the driving
force for salt transport through the membrane.
Qs = ksAc/
 It engenders more salt leakage into the product water.
 CONCENTRATION POLARIZATION IN REVERSE OSMOSIS (2/3)

• When the concentration of retained solutes at the membrane surface, c10,

exceeds the solubility limit, it forms a thixotropic gel.
 It is referred to the phenomenon as “fouling” of the
membrane.
[Example] A reverse osmosis process is used for desalination of seawater. The
volumetric flux of water through the membrane is 3  10−5 m/s (or m3 s−1 m−2),
and the applied feed pressure is 8.0 MPa greater than the product-water
pressure. For seawater, the osmotic pressure is 2.5 MPa. What is the water
velocity through the membrane if the polarization modulus (c10/c1) rises to 1.2-
fold of the original?

Solution:

Osmotic pressure of seawater,  = 1.12(T + 273)Smi

 2 (Smi ) 2
 = = 1.2   2= 1.2 1 = 1.2  2.5 = 3.0 MPa
 1 (Smi )1

(To be continued)
[Example] Reverse osmosis process for desalination of seawater.
Flux of water = 3  10-5 m/s (or m3 s-1 m-2); P = 8.0 MPa;  = 2.5 MPa.
What is the water velocity through the membrane if the polarization
modulus (c10/c1) rises to 1.2-fold of the original?

Solution (cont’d):

 2 = 3.0 MPa

Rate of water passage through the membrane:

Qw = kwA(P −  )/ or Qw/A = jv = Lp(P −  )

jv 2 ( P −  ) 2 8.0 − 3.0
= = = 0.91
jv1 ( P −  )1 8.0 − 2.5
\ jv2 = 0.91jv1 = 0.91  (3  10−5) = 2.73  10−5 m/s
#
 CONCENTRATION POLARIZATION IN REVERSE OSMOSIS (3/3)

• Correlation of Dittus and Boelter (1930) for concentration polarization in

turbulent flow:

c10 jv d h0.2 ( m / r ) 0.47

= exp
c1 0.023v 0.8 D 0.67
where jv = the volume of solvent per area per time
dh = equivalent hydraulic diameter
m = viscosity of fluid
r = density of fluid
v = average velocity of fluid
D = local solute diffusivity in solution
[Example] A reverse osmosis desalting process is carried out using turbulent
flow through a tubular 1.0-cm-diameter membrane with a system temperature
of 18.5C. Which of the following factors would be effective in reducing the
degree of concentration polarization if the water flux is held constant? (a)
Reduced temperature; (b) reduced tube diameter with the same mass flow rate
of seawater; and (c) recirculation of the seawater with the same size and length.

Solution:
c10 jv d h0.2 ( m / r ) 0.47
= exp
c1 0.023v 0.8 D 0.67
(a) Lower temperature increases the viscosity and lowers the diffusivity.

 Lower temperature makes concentration polarization more severe.

(To be continued)
[Example (cont’d)] Which of the following factors would be effective in
reducing the degree of concentration polarization if the water flux is
held constant? (b) Reduced tube diameter with the same mass flow rate
of seawater; and (c) recirculation of the seawater with the same size and
length.

Solution: c10 jv d h0.2 ( m / r ) 0.47

= exp
c1 0.023v 0.8 D 0.67
(b) Reducing tube diameter with the same mass flow rate of water will raise v.

 Concentration polarization will be reduced.

(c) Recirculation of the seawater will increase v.

 It alleviates concentration polarization but does so at the expense of much

more pumping power (more flow and more pressure drop).

#
ANALYSIS OF DIALYZER
A shell-and-tube type of hollow-fiber dialyzer:

Q = flow rate, A = area of membrane, K = dialysis coefficient

 ANALYSIS OF DIALYZER (2/2)

For a counterflow dialyzer,

The rate of salt removed

= QF (CF 1 − CF 2 ) = QD (CD 2 − CD1 ) = KAClm

( C F 1 − C D 2 ) − ( C F 2 − C D1 )
Clm =
ln[( C F 1 − C D 2 ) /( C F 2 − C D1 )]
[Example] A solution of raffinose containing 100 g/L of NaCl is to be dialysed in a
shell-and-tube type of hollow-fiber dialyzer operating countercurrently. With a
dialyzer having 1000 cm2 area of membranes the dialysis coefficient for NaCl was
determined to be 0.0415 cm/min, when the feed rate was 200 cm3/min, and the flow
rate of pure water was 500 cm3/min. If 90% of the salt is to be removed, what area
of the hollow-fiber membranes will be needed, if the same flow rates for feed and
water are used?
Solution:

QF (CF 1 − CF 2 ) = KAClm

The rate of salt removed = QF (CF1 − CF 2 ) = QD (CD2 − CD1 )

200(100 − 0.1 100) = 500(CD2 − 0)
 CD 2 = 36 g/L
(To be continued)
[Example] Shell-and-tube type of hollow-fiber dialyzer
CF1 = 100 g/L; K = 0.0415 cm/min; QF = 200 cm3/min; QD = 500
cm3/min. If 90% of the salt is to be removed, what area of the hollow-
fiber membranes will be needed?
Solution (cont’d):

CD 2 = 36 g/L

(C F 1 − C D 2 ) − (C F 2 − C D1 )
Clm =
ln[( C F 1 − C D 2 ) /( C F 2 − C D1 )]
(100 − 36) − (10 − 0)
= = 29.1 g/L
ln[( 100 − 36) /(10 − 0)]
QF (CF 1 − CF 2 ) = KAClm
QF (C F 1 − C F 2 ) 200(100 − 10)
 A= = = 14,905 cm 2
KClm (0.0415)( 29.1)
#