CHE449 - Merged

Download as pdf or txt
Download as pdf or txt
You are on page 1of 246

CHE 449

MEMBRANE PROCESSES and


SEPARATION of MIXTURES

CHAPTER 1
FUNDAMENTALS Of MEMBRANE
SEPARATION PROCESSES
DIFFERENT TYPES OF MEMBRANES
DEFINITION of a MEMBRANE

 Selective or nonselective barrier.


 Seperates and/or contacts two
adjacent phases.
 Allows or promotes the exchange
of matter, energy, and information
between the phases in a specific or
non-specific manner.
APPLICATIONS OF MEMBRANE
TECHNOLOGY IN FOOD INDUSTRY

 Removal of bacteria from liquid food as an


efficient alternative to heat treatment (cold
pasteurization).
 Concentrate milk for continuous production
of cheese.
 Treatment of whey for production of whey
protein concentrate and whey protein
isolate.
 Wastewater treatment (The dairy industry
is considered as one of the most polluting
food industries in term of water
consumption and discharged volume of
wastewater.
 Clarification of fruit juice
 Partial demineralization of whey
 Defatting of skim milk

APPLICATION OF MEMBRANE
TECHNOLOGY IN PHARMACEUTICAL
INDUSTRY

 Sterilization of pharmaceuticals
 Purification of antibiotics
 Recovery of vaccines and antibiotics
from fermentation broth
APPLICATION OF MEMBRANE
TECHNOLOGY IN PETROCHEMICAL
INDUSTRY

 Monomer recovery in polyolefin


production.
 Ethylene recovery in ethylene oxide
production.
 H2 recovery from refinery fuel.
 CO2 and N2 removal.
 Natural gas treatment to meet
pipeline specifications.
APPLICATION OF MEMBRANE
TECHNOLOGY IN BIOMEDICAL
INDUSTRY

 Artificial kidney/liver/pancreas
 Separation of cells from tissue and/or cell
fractionation
 Blood oxygenation with membranes
 Controlled drug delivery
 One Step Separation of Plasma from Whole
Blood Without Centrifugation (Diagnostic
applications).
For separation, we need difference in properties of
the components in the mixture

Which properties ???


MEMBRANES
How can a separation process be
selected to solve a given problem?

 The separation must be feasible


technically.

 The separation must be feasible


economically.
Objectives of Separation
 Concentration: Desired component is present in a
low concentration and solvent needs to be removed.
Example: Whey comprises 80 – 90 % of the total
volume of milk entering the process and contains
about 50 % of the nutrients in the original milk:
soluble protein, lactose, vitamins and minerals.
Proteins should be concentrated to have commercial
value.

 Purification:Undesirable impurities/compounds have


to be removed.
Example:Removal of viruses , bacteria in water
Objectives of Separation
 Fractionation:A mixture must be separated into 2 or
more desired components.
Objectives of Separation
 Reaction mediation:Continuous removal of
products will increase the rate of reaction.
Critical Question

Compare the energy consumption for distillation and


membrane based separation (qualitatively).

Assume that the driving force for separation is pressure difference


Advantages of Membrane
Separation Processes
 Separation can be carried out continuously.
 Energy consumption is generally low.
 They can be easily combined with other
separation processes.
 Separation can be carried out under mild
conditions.
 Up-scaling is easy.
 Membrane properties are variable and can
be adjusted.
 No additives are required.
Distillation
Condenser Duty: Mass of vapor condensed x Latent Heat of vaporization

Reboiler Duty: Mass of steam x Latent heat of condensation

2260 kJ/kg
Feed: Water Basis: 1 kg Density : 1000 kg/3 Volume of feed: 10 -3 m3
Condenser and Reboiler Duties≈ 2260 kJ

Membrane Separation
Separation at ambient conditions: Generally no need for preheating

Energy consumption required for pressurizing the feed :


Volumetric flowrate of feed x Pressure of feed

( m3/sec) x (N/m2 ) = N.m/sec= J/sec


Pressure of feed : 40 bar = 4x106 N/m2
Energy consumption required for pressurizing the feed ≈ 10-3 x 4 x10-6
≈4x103 J = 4 kJ
https://cen.acs.org/articles/95/i25/Putting-distillation-business-
chemical-industry.html
Drawbacks of Membrane
Separation Processes

 Concentration polarisation,
membrane fouling.
 Low membrane life time.
 Low selectivity or flux.
 Up-scaling factor is more or less
linear.
Schematic Representation of a
Membrane Process
Single module: Smallest unit in which membrane area is packed.

Feed
Retentate

Membrane

Permeate

Retentate : The stream containing species which do not pass through the
membrane

Permeate : The stream containing species which passed through the


membrane
MEMBRANE MATERIALS and
STRUCTURES
memb rane materials

polymers ceramics glass metals liquids

memb rane structures


symmetric asymmetric

integral asymmetric composite structures

homogeneous cylindrical sponge-type porous skin homogeneous homogeneous


films pores structures layer skin layer skin layer

Membrane configuration
Hollow fiber
Flat-sheet Spiral wound Tubular Capillary
DIFFERENT TYPES of MEMBRANES
Asymmetric membrane Dense membrane

Electrically charged membrane Supported liquid membrane


SYMMETRIC and ASYMMETRIC
MEMBRANES
 Symmetric membrane:
 The structure and the transport properties are identical
over the entire cross-section.
 Flux is determined by the thickness of the membrane.
 Used mainly in dialysis and electrodialysis.

 Asymmetric membrane:
 The structural and transport properties vary over the
cross-section.
 Consists of skin layer (0.1 to 1 micron) and highly
porous layer (100-200 micron).
 Skin represents actual selective barrier of the
membrane and its thickness controls the flux.
 Porous sublayer serves only as a support and has little
effect on the separation characteristics.
Asymmetric membrane
Electron Microscope Pictures
Of Asymmetric Membranes
Symmetric membrane
Morphology Symmetric

Cylindrical Porous Porous Homogeneous


(resistance to mass transfer is determined by total membrane thickness)

Asymmetric – Single Material


top layer

Porous Porous with Top Layer


Asymmetric – Composite-Thin Film Composite Membrane

dense skin layer (0.1 to 0.5 µm)


porous membrane (50 to 150 µm)

(resistance to mass transfer determined by skin layer thickness)


Structure of Membranes
Critical Questions:

When do you prefer asymmetric membrane ? Application areas ?

When do you prefer symmetric membrane ?

When do you prefer thin film composite membrane ?


POROUS MEMBRANES
Microporous membrane: Achieve separation if molecules differ
considerably in size

Particles smaller than the


pores in the membrane
pass through the membrane

Particles larger than the pores in the membrane are rejected


DENSE MEMBRANES

 Yellow molecules have


higher transport rate than
gray molecules

The relative transport rates of components within the membrane


should be different.

Transport rate: Determined by the diffusivity and solubility of


components in the membrane material.
DEFINITION OF VARIOUS MEMBRANE
PROCESSES
(According to the applied driving forces)

 Hydrostatic pressure driven processes


(reverse osmosis, nano-, ultra,
microfiltration, gas separation,
pervaporation)
 Concentration gradient or chemical
potential driven processes (dialysis,
Donnan dialysis, pervaporation,
membrane contactors)
 Electrical potential driven processes
(electrodialysis)
 Temperature difference driven processes
(membrane distillation)
https://encyclopedia.che.engin.umich.edu/Pages/SeparationsC
hemical/Membranes/Membranes.html
DRIVING FORCE: PRESSURE DIFFERENCE
DRIVING FORCE: ELECTRICAL POTENTIAL DIFFERENCE

• The negatively charged anions migrate towards the anode and


the positively charged cations towards the cathode.
ELECTRODIALYSIS

https://www.youtube.com/watch?v=wYkkLUckmg4

https://www.youtube.com/watch?v=QNX150DtnMc
DRIVING FORCE: CONCENTRATION DIFFERENCE

Problem : Purification of protein mixture. Remove small molecular


weight substances such as reducing agents and salts.

Driving Force: Concentration Difference ???

https://en.wikipedia.org/wiki/Dialysis_tubing
DIALYSIS
https://www.youtube.com/watch?v=kp9EvMeTMUs
DRIVING FORCE: TEMPERATURE DIFFERENCE

VAPOR PRESSURE DIFFERENCE


Critical Question
• Large scale separations require hundreds to thousands of square
meters of membrane to perform the separation

• How can we efficiently pack the membranes to provide a large surface


area for effective separation.

• Cost of membrane based separation:


• Membrane cost + cost of the containment vessel into which they are
mounted.

• Problem is to pack the most membrane area into least volume

• Minimise the cost of the containment vessel


• Provide acceptable flow ydrodynamics in the vessel
MEMBRANE MODULES

• A membrane module is a complete unit composed of :


Membranes
A housing
Feed inlet
Concentrate outlet
Permeate outlet

• Membrane modules have contain different configurations:


Hollow fiber
Plate and frame
Tubular
Spiral wound
HOLLOW FIBER
https://www.youtube.com/watch?v=B11JuAjOF-M

SPIRAL WOUND
https://www.youtube.com/watch?v=SdEut88ySg8

https://www.youtube.com/watch?v=VmIlAySHhkQ

PLATE AND FRAME TYPE


https://www.youtube.com/watch?v=lQ7al48dElM

TUBULAR MEMBRANE MODULE


https://www.youtube.com/watch?v=5LYq032iB8c

Module & Housing Assembly

https://www.youtube.com/watch?v=7hXCEFWZ6O4
CHE 449
MEMBRANE PROCESSES and
SEPARATION of MIXTURES

CHAPTER 2
MICROFILTRATION
http://www.osmotechmembranes.com/micro-filtration.php
http://www.katmaj.com/tubular-membranes-mbr-juice-apple-filtration.html
• Membrane pore size < Solute size • Membrane pore size >Solute size
• Solute particles are physically
• Solute particles can pass through
blocked from passing into or
the membrane.
through the membrane.
• Solute particles adsorb on the pore
• This sieving mechanism is called
wall.
surface filtration.
TYPES OF MICROFILTRATION CONFIGURATION
Selection of a Membrane

 Suitable pore size


 Narrow pore size distribution
 Low fouling tendency : Life time
 Sufficient mechanical strength
 Sufficient chemical stability
 High solution permeability
 High economic efficiency
Membrane Morphology
Microfilters are typically homogeneous. The structural and
transport properties across their cross section are similar.
MEMBRANE PRODUCTION METHODS

Phase Inversion Track Etching


PHASE INVERSION

HOW

Initially Later on
Homogeneous Solution Heterogeneous Solution
Single phase Two phases:
Transparent solution Phase 1: Polymer rich phase
Polymer is completely dissolved (Matrix)
Phase 2: Polymer lean phase
(Pores)
Turbid solution

There are 4 methods to change homogeneous solution to a


heterogeneous solution. As a resut, porous (symmetric or asymmetric)
membranes are obtained.
Phase inversion techniques

 Precipitation from the vapour phase


 Vapour phase: nonsolvent + saturatedsolvent
 Prepare porous without top layer
 Precipitation by controlled evaporation
 Polymer dissolved in mixture of solvent and nonsolvent
 Prepare membranes with skinned layer
 Thermal precipitation
 Polymer solution is cooled to enable phase separation
 Prepare membranes with skinned layer
 Immersion precipitation

Solvent: Dissolve polymer completely


Nonsolvent: Cannot completely dissolve the polymer
Polymer Precipitation by Water (Nonsolvent Induced
Phase Inversion(NIPS))

Step 1: Polymer is dissolved in a solvent Step 2:Homogeneous solution is cast Step 3: Homogeneous solution
And homogeneous solution is formed. on a suitable support with a knife is transformed into a
heterogeneous solution

Watch the videos!


https://www.youtube.com/watch?v=WTrjqXbzhjE
https://www.youtube.com/watch?v=8yLTYltbC64
 Polymer Precipitation by Water (Nonsolvent Induced
Phase Inversion(NIPS))

Polymer Solution: Polymer +Solvent


Coagulation Bath: Coagulant (Nonsolvent (e.g. water)

Solvent and nonsolvent should


be
miscible

Polymer should be
immiscible
in the precipitation bath
Phase Inversion Time: Precipitation Time

Instantaneous Delayed

Does the precipitation time influence membrane morphology?

YES !
Critical Question

How can we control the precipitation time?

Control mass transfer rate in

Casting solution Coagulation bath


Parameters That Control Delayed versus
Instantaneous Phase Inversion

 Casting solution :
 Add additives into casting solution.
 Change the temperature of the casting solution.
 Change the solvent

 Coagulation bath :
 Add additives into coagulation bath.
 Change the temperature of the coagulation bath.
 Change the nonsolvent type
Objective: Manufacture MF Membrane with Wet Phase
Inversion Technique

Desired Structural Features

Very very thin dense skin layer (Almost no skin layer is desired)

Open cell structure

Large pore size


SCHEMATIC DRAWING OF MEMBRANE PREPARATION BY
TRACK-ETCHING

Polymer film exposed


to radiation.
Exposure time:
Determines the
number of pores.

The film is etched along the damaged regions in


a chemical bath
Etching time (time in chemical solution bath):
Determines pore diameter
Critical Question

How will you test if the membrane supplied to your company


is a MF membrane or not ?

Measure the pore size of the membrane


Bubble Point Measurement

• A non-toxic liquid is allowed to spontaneously fill the pores in the


membrane.

• Non-reacting gas, usually air, is allowed to displace the liquid from


pores by increasing the gas pressure.

• First the larger pores will get emptied, as the pressure increases more
and more smaller pores are progressively emptied.
2
rp  Cos
P
γ: Surface tension of the liquid.
rp: Radius of a capillary shaped pore.
θ: Contact angle between solid (membrane) and liquid (=0)
ΔP:Pressure difference applied through the membrane
Based on pore size /pore size distribution, you can
check if the membrane is suitable for your application
Suppose you have different membrane samples supplied by different
companies.
You want to select the one which has the least fouling tendency

How can you select the right membrane ?

Filter the solution that will be treated by the membrane for a


while and follow the flux change as a function of time (or volume
filtered)
Mode of Operation: Constant Transmembrane Pressure

With fouling, flux decreases with time

Best membrane
which showed
the lowest flux
decline
Mode of Operation: Constant Flux

With fouling, transmembrane pressure increases with time


TYPICAL FOULANTS

https://www.mdpi.com/2073-4360/13/6/846
https://www.mdpi.com/2073-4360/13/6/846
https://www.mdpi.com/2073-4360/13/6/846
FOULING DURING MICROFILTRATION

http://www.msrjournal.com/article_34772.html
https://www.eeer.org/journal/view.php?number=793
Which factors control degree of fouling ?

Membrane properties
Feed conditions • Narrow pore size distribution
• Heat treatment • Use of hydrophilic polymers
• PH adjustment
• Addition of complexation agents,
• Chlorination
Module design
• Adsorption onto
activated carbon, • Use of various kinds of turbulence
chemical promoters may reduce fouling
clarification
Watch the Videos!

https://www.youtube.com/watch?v=u7IWZQqMX9k
Critical Questions

What are the effective cleaning methods ?

Which factors should be considered while selecting a


suitable cleaning method ?
Case Study
Milk contains many food products , however, these products
have functionality and commercial value if they are separated.

Whole milk contains 2.6% (w/w) casein (on average), and the
separation of casein micelles from whey protein is important.
Concentrated casein can be used in the following applications:
• Mixed with cream for production of cheese
• Used for standardization of milk
• Used for production of dried native casein which is used
as food additive.
• Used for development of edible coatings and bio-active peptides.

Whey proteins recovered from milk have pharmaceutical and food


applications.

Can we design a membrane separation unit to concentrate casein


from milk ?
ALTERNATIVE METHODS
Traditional Methods Membrane Based Method
 Laborious  Separated casein remains

 Require chemical functional


addition  No chemical usage

 Liquid waste stream is  Reduced environmental


generated impact
 Nutrional and
functional properties of
coagulated casein is
reduced, thus, product
usage becomes limited
Determine the composition of milk. Define the feed stream

G. Brans et al. / Journal of Membrane Science 243 (2004) 263–272


Determine Membrane Seperation Process

G. Brans et al. / Journal of Membrane Science 243 (2004) 263–272


References

1) Microfiltration of skim milk using polymeric membranes for casein


concentrate manufacture. N.D. Lawrence et al. / Separation and Purification
Technology 60 (2008) 237–244

2) V.S. Espina et al. / Journal of Membrane Science 325 (2008) 872–879

3) Microfiltration of skim milk using polymeric membranes for casein concentrate


manufacture Separation and Purification Technology 60 (2008) 237–244

4)Journal of Dairy Science Volume 92, Issue 11, November 2009, Pages 5357-
5362

5) Comparison of alternative processes.


(https://hal.archives-ouvertes.fr/hal-00929030/document)
Propose a FlowSheet for the Separation

Is this a suitable flowsheet ?

Can you see any problem with this design ?

Feed Retentate ???


Whole milk

Permeate ???
DIAFILTRATION

• Diafiltration: An operation mode of a pressure-driven


membrane filtration process.

• A diluant (i.e. water or any other solvent or buffer) is added


to the process liquor in order to enhance the degree of
separation of species retained by the membrane.

• Single Pass Diafiltration


• Multistage Cocurrent Diafiltration
• Multistage Countercurrent Diafiltration
Batch Diafiltration

It is preferably used in small-scale operations

Diluant: Microsolute-free solvent, most commonly pure water, sometimes


a buffer solution to exchange microsolutes with buffer components
that stabilize the macrosolutes
Continuous Diafiltration

Preferably employed when large throughputs are


required.
Cocurrent Diafiltration

Continuous DF, as opposed to batch DF: The product (i.e., the final
retentate) is being continuously withdrawn from the system during the
entire course of filtration.
Suppose N:1 (Single stage )

Overall mass balance around the unit

q f  qd  q p  qr q: Volumetric flowrates

Retentate
Feed Diluant Permeate

Assume that : Permeate loss is replaced by the diluant qd  q p


q f  qr
Solute balance around the unit

q f C f  qd C d  q p C p  qr C r
Diluant is pure : No solute in it
Cp Cr 1
Rejection of solute : R R  1 Y 
Cr C f  qp 
1  1  R 
 q f 
N:1 (Single stage )

Cr 1
Y 
C f  qp 
1  1  R 
 q f 

For N stages

Cr 1
Y  N Cf: Fresh feed
Cf  q  concentration
1  1  R 
p

 q f  Cr:Retentate
concentration
at the end of Nth stage
Calculation of Required Membrane Area

What type of equation do we need to calculate membrane


area ?

Feed Retentate ???


Whole milk

Permeate ???

Amount permeated=qp=Solution flux through the membrane x Area

m3/s m3/m2.s
If we know amount permeated (qp ) and solution flux
through the membrane, we can calculate the area

Amount permeated (qp ): We can calculate from the mass balance

Solution flux through the membrane: Pressure driven process


The pressure-driven permeate flux through this cake layer and the
membrane can be described by Darcy's law:

qp P
J 
Area  Rm  Rc 
Rm: Mass transfer resistance of the membrane
Rc: Mass transfer resistance of the cake layer
ΔP: Transmembrane pressure difference through the membrane
Which factors determine the solution flux ?

Transmembrane Pressure Membrane Morphological


Difference Properties

• Pore size
• Porosity
• Pore connectivity
• Pore shape
How can we find the resistances (needed for flux calculation ) ?

Membrane Resistance

P
J pure water 
Rm
 Measure pure water flux through the membrane at the constant
ΔP

P
Rm 
J pure water
Cake Layer Resistance
 Measure solution flux (e.g. skim milk) through the membrane at the
constant ΔP

P
RTotal 
J

Rc  RTotal  Rm

Flux reached steady state


This flux should be used
to determine total
resistance.
Cr 1
Y 
C f  qp 
1  1  R 
 q f 

What is your design criteria ?

Reduction in the protein content of skim milk in retentate ???

Cr
Y Design Criteria  Depending on the product
Cf requirement, you can set
this value

 Lower the Y value, higher the purity of casein

 After setting Y value, we can calculate qp value if R is known.


Cr 1
Y 
C f  qp 
1  1  R  R: Rejection of solute by the
 q f  Membrane. Should be found after
selecting the membrane
Amount of skim milk to be fractionated.
Design Criteria Known

qp P Set by the designer


J 
Area  Rm  Rc 

Total resistance. Should be found after selecting


the membrane.

We should select the membrane and find the experimental data


to find solute rejections and solution flux versus time data
Experimental Data

N.D. Lawrence et al. / Separation and Purification Technology 60 (2008) 237–244

Membrane : Polyvinylidine fluoride (PVDF) (0.5 µm)

Compound ΔP:50 ΔP:100 ΔP:150 ΔP:258


Rejection (%) kPa kPa kPa kPa

Casein 96 97 98 100
ß-lactoglobulin 78 88 92 99

ΔP:100 kPa ΔP:150 kPa


Solution Flux 25 40
(L/m2h)

Pure Water Flux (L/m2h):800 ΔP:52 kPa


Remarks:

 Solute rejection changes with the transmembrane pressure

 Total resistance observed during milk filtration depends on


the transmembrane pressure

 Experimental data for solute rejection and solution flux


should be collected at the same transmembrane pressure
Countercurrent Diafiltration
Assume : Every stage has the same rejection and qd  q p
Cr 1
YN   i
Cf i N  qp 
  1  R 
i 0  q f 
Single Stage Continuous Diafiltration Unit

Transmembrane Pressure: 100 kPa

The Change of Membrane Area with Lactoglobulin Content


in Retentate
1500

1200
Area (m2)

900

600

300

0
0 0.1 0.2 0.3 0.4 0.5 0.6
Lactoglobulin concentration in retentate/Feed

 As the purity of casein increases (lactoglobulin content in the


retentate is decreased) then required membrane area increases
Multistage Countercurrent Diafiltration Unit

Transmembrane Pressure: 100 kPa

The Effect of Stage Number on the Membrane Area


1600

1400

1200

1000
Area (m2)

800
N=1
N=2
600

400

200

0
0 0.1 0.2 0.3 0.4 0.5 0.6
Lactoglobulin concentration in retentate/Feed

 By increasing the stage number to 2, the area requirement was


significantly reduced.
Multistage Continuous Diafiltration Unit

Transmembrane Pressure: 100 kPa

The Effect of Stage Number on the Membrane Area


450

400

350

300
Area (m2)

250
N=2
200 N=5

150 N=10

100

50

0
0 0.1 0.2 0.3 0.4 0.5 0.6
Lactoglobulin concentration in retentate/Feed
Multistage Cocurrent Diafiltration Unit

Transmembrane Pressure: 100 kPa

Effect of Stage Number on the Area Requirement


1600.00

1400.00

1200.00

1000.00
Area (m2)

800.00
N=1
N=2
600.00

400.00

200.00

0.00
0 0.1 0.2 0.3 0.4 0.5 0.6
Lactoglobulin Concentration in Retentate /Feed
Multistage Cocurrent Diafiltration Unit

Transmembrane Pressure: 100 kPa

Effect of Stage Number on the Area Requirement


350.00

300.00

250.00
Area (m2)

200.00
N=2
150.00 N=5
N=10
100.00

50.00

0.00
0 0.1 0.2 0.3 0.4 0.5 0.6
Lactoglobulin Concentration in Retentate /Feed
Multistage Cocurrent Diafiltration Unit

Transmembrane Pressure: 100 kPa

Effect of Stage Number on the Diluant Requirement


9.00

8.00
Diluant Requirement (m3/h)

7.00

6.00

5.00
N=2
4.00 N=5
N=10
3.00

2.00

1.00

0.00
0 0.1 0.2 0.3 0.4 0.5 0.6
Lactoglobulin Concentration in Retentate /Feed
Multistage Countercurrent Diafiltration Unit

Transmembrane Pressure: 100 kPa

The Effect of Stage Number on the Diluant Requirement


12.00

10.00
Diluant Requirement (m3/h)

8.00

6.00 N=2
N=5
N=10
4.00

2.00

0.00
0 0.1 0.2 0.3 0.4 0.5 0.6
Lactoglobulin concentration in retentate/Feed
The Comparison of Area Requirement for Cocurrent and
Countercurrent
Flow Configurations

Cr
Y Design Criteria
Cf
The Comparison of Diluant Requirement for Cocurrent and
Countercurrent
Flow Configurations

Cr
Y Design Criteria
Cf
CASE STUDY 1

Select a microfiltration-based separation problem. Below,


you can find what should be included in your first case
study.
Define the importance of the problem and clearly explain the
aim of the separation.
Discuss the alternative methods for the separation you have
chosen (advantages and disadvantages)
Select the membrane and report its properties and explain
the reasons for your selection
Draw the flowsheet of the process.
Apply necessary mass balances
Report the experimental data you used for the calculations.
Calculate the area required for the separation
Calculate the main operating costs for the separation
Suggest parameters for optimizing the separation
CHE 449
MEMBRANE PROCESSES and
SEPARATION of MIXTURES

CHAPTER 3
ULTRAFILTRATION
https://synderfiltration.com/ultrafiltration/membranes/
Some Applications of Ultrafiltration
Commonly Used Production Technique: Phase Inversion

Membrane Properties

Morphological parameters Membrane properties that


• Pore size relate directly to membrane function
• Pore size distribution • Permeability
• Surface porosity • Solute rejection
• Skin thickness • Surface chemistry, including
• Surface roughness surface charge.
Theory and Mechanistic Concepts
Assume:
A uniform pore size
Laminar flow through these pores

The volumetric flux Jv, can be expressed in terms of


the Hagen-Poiseuille equation:

Pe
J v  P
l

np: an aggregate of cylindrical pores per unit area , np ' with


dp:Average pore diameter
l: Length of the pores: Thickness of the membrane
η : Viscosity of the solution
ΔP: Transmembrane pressure difference
Osmotic Pressure Model

Explains the flux reduction by the increased osmotic pressure


difference from the feed/membrane surface to the permeate phase that
has to be overcome.
Solvent permeability

J v  P   
Pe
l Osmotic pressure difference
Membrane
thickness Transmembrane
pressure
CONCENTRATION POLARIZATION

Observed when Cf≠ Cw


Mass Balance:
Rate of mass input =Rate of mass output

cio x 
J v c i cip   Di i
dc dci J
dx 
Cib
 v
ci cip  Di 
x 0
dx

cwall  cio cbulk  cib


J v Jv Convective flux
  Peclet Number 
Di Di /  Diffusive flux
J v  Di /  Concentration polarization is small

J v  Di /  Concentration polarization is large

k  Di / 
Mass transfer coefficient of the solute

What should we do to reduce concentration polarization ?


 In liquid separation processes,
concentration polarization is
controlled by controlling the stirring
rate (fluid velocity).
 Increasing velocity reduces boundary
layer thickness.
 Modules are arranged in series (high
fluid velocity) rather than in parallel
(low fluid velocity) to control
polarization.
Calculation of Mass Transfer Coefficient
Gel Polarization Model

Rc: The resistance due to the cake or gel layer


Rm:Membrane resistance

• To calculate the membrane area, we need to find amount


permeated through the membrane and flux through the membrane

qp:Total amount permeated through the membrane (L/h)

JTotal  J solvent   J solute


If the feed is dilute in solute(s) Csolvent  Csolute

J Total  J solvent
Specification of Commercial Ultrafiltration Membranes

1. Pure Water Permeability of the Membrane

ΔP ΔP3
Amount of water ΔP2
permeated (L)
ΔP1
Slope1
Membrane
Time (h)
ΔP3 >ΔP2 > ΔP1

Water permeated Water Flux =Slope1/Membrane Area


Measure it (L) (L/m2.h)

Measure amount permeated at


different ΔP values
Slope 2=Water
Water Flux
permeability
(L/m2.h.bar)
ΔP
2.Molecular Weight Cutoff (MWCO) of the Membrane

Molecular weight cutoff of a membrane: The molecular weight (MW)


of a neutral solute that is rejected to 90% by the membrane.
Experimental Protocol for Determining
MWCO

 Filter neutral solutes with different


molecular weights.
 Measure the concentration of the
solute in feed, permeate and
retentate.
 Calculate the rejection of each
solute .
 The solutes should be neutral and
their concentration should be low.
 The pore radius of the membrane
can be estimated by using a model
of pore transport and an
appropriate correlation between the
hydrodynamic radius and MW of the
specific type of solute.
PROCESS ECONOMICS

 Fixed cost for installing the unit


 Replacement cost for the modules
 Cleaning chemicals
 Energy/power usage
 Labor
 Depreciation
Case Study- Clarification of Fruit Juice

Fruit juices have turbidity caused by the presence of polysaccharides


(cellulose, hemicellulose, lignin, pectin and starch), protein in
colloidal suspension, tannins and metal. However, turbidity has a negative impact
on the final product quality, thus, fruit juice should be clarified.
Clarification of fruit juice mainly removes pectin and other carbohydrates
which are present in the juice.

The selection of a method for clarification is important. Following issues


should be considered for the selection.
• Safety and quality improvement
• Nutritional value
• Consumer convenience and acceptance
• Product and process

We want to evaluate the feasilibity of using membrane filtration for


pineapple juice clarification. How should we start ?
Comparison of Alternative Processes

Alternative processes are centrifugation, enzymatic treatment or


applying clarifying agents such as gelatin bentonite, silica sol, and
polyvinyl pyrrolidone

However, these processes can be labor-intensive, time-consuming and


discontinuously operated. Besides, the use of additives may change
taste in the juice .

Clarification with membranes does require additives, it is a continuous


Process. The taste of juice does not change
References

https://www.scielo.br/j/cta/a/hkrc7G6Z5pHQVTZbc5Xm5S
N/?format=pdf&lang=en

Journal of Membrane Science 215 (2003) 213–224


DESIGN -1
Overall mass balance around the unit

q f  q p  qr

Solute balance around the unit

q f C f  q p C p  qr C r

Feed: Fruit juice Retentate


Galacturonic acid
(Pectin)

Permeate

Design objective: The concentration of pectin in the clarified product should


be reduced 10 times of its concentration in the feed.
Cp=0.1Cf or Cp/Cf=0.1
q f C f  q p C p  q f  q p Cr

Cp Cp
Rejection of pectine : R R  1 Cr 
Cr 1 R
Cp

Cf

     
q f 1    q p   
 1  R   1  R 

To calculate qp from here we need to know:

Ɵ: Design parameter R: Solute rejection. Membrane property, should


q :Design parameter be found from experimental data
 Galacturonic acid concentration was measured in retentate and permeate
streams at different transmembrane pressure and temperature.

 Experimentally measured permeate flux was reported. This data included


fouling of the membrane.

Journal of Membrane Science 215 (2003) 213–224


Should be used in design
calculations
     
q f 1    q p   
 1  R   1  R 

Rearrange this equation

 1  R    
qp   
  R  

 To obtain a positive qp value, Ɵ should be greater than 1-R


Case 1:   0.76 For each stage !

Suppose N: 2

C p1  C f  N=2
qf qR1 Cp2  C f  2
C p 2  C p1

qp1 qR2

C pN qp2
Design objective:  0 .1 Number of stages?
Cf
 C pN 
C pN  C f  N
ln    N ln   N 9
C 
f
Case 1:   0.76 For each stage !

Area Requirement for Different Operating Conditions

TMP
T ( C) (Bar) A1(m2) A2(m2) A3(m2) A4(m2) A5(m2) A6(m2) A7(m2) A8(m2) A9(m2) Atotal(m2)
40 0.2 18.63 16.63 14.85 13.26 11.84 10.57 9.43 8.42 7.52 111.14
40 0.8 16.50 13.25 10.64 8.55 6.87 5.51 4.43 3.56 2.86 72.17
40 2 19.82 18.92 18.05 17.23 16.44 15.69 14.98 14.29 13.64 149.07
30 0.2 18.51 13.77 10.25 7.63 5.68 4.22 3.14 2.34 1.74 67.28
20 0.2 24.84 22.40 20.20 18.22 16.43 14.81 13.36 12.04 10.86 153.17

When reduction in pectin concentration is Cp/Cfeed=0.1 (in overall)

The lowest total area requirement is obtained when T=30 C and


TMP=0.2 bar
Case 2:   0.5 For each stage !

 C pN 
C pN  C f  N
ln    N ln   N 4
C 
 f 

Area Requirement for Different Operating Conditions

T ( C) TMP (Bar) A1(m2) A2(m2) A3(m2) A4(m2) Atotal(m2)


40 0.2 13.78 9.10 6.01 3.97 32.88
40 0.8 7.74 2.92 1.10 0.41 12.17
40 2 17.77 15.20 13.00 11.12 57.10
30 0.2 4.72 0.90 0.17 0.03 5.82
20 0.2 18.98 13.07 9.01 6.20 47.26

Increasing the reduction of pectin concentration in each stage ends up


with lower area requirement. But think about the how the energy
cost will change with this strategy ! Think about the increased pectin
concentration in retentate
Design-2

qf qR1 qR2

qp2 qR3

qp1

qp3

Permeate
Overall mass balance around the unit-1

q f  qr 3  q p1  qr1 (1)
Solute balance around the unit-1

q f C f  qr 3Cr 3  q p1C p1  qr1Cr1 (2)


Overall mass balance around the unit-1

qr 1  q p 2  q r 2 (3)
Solute balance around the unit-1

qr1Cr1  q p 2C p 2  qr 2Cr 2 (4)


Overall mass balance around the unit-3

q p 2  q p 3  qr 3 (5)
Solute balance around the unit-3

q p 2C p 2  q p 3C p 3  qr 3Cr 3 (6)
Degrees of Freedom (df) Analysis

Unknowns : qr1, qp1, Cp1, Cr1, qr2, qp2, Cp2, Cr2, qr3, qp3, Cp3, Cr3

Number of Equations=NE=6 df=12-6=6 !


Number of Unknowns =NU=12
What should we do to obtain df=0?

Can we add additional equations ? YES!

Cp Cp
Rejection of pectine : R R  1 Cr 
Cr 1 R

We use the same membrane in each unit ! Thus, rejection values


are fixed
C p1 C p2 C p3
Cr1  (7) Cr 2  (8) Cr 3  (9)
1 R 1 R 1 R

df=12-9=3 Can we add additional equations ? NO!


We have set 3 parameters as a designer to make df=0

C p1 C p2 C p3
1  2  3 
Cf C r1 C p2
• Don’t forget , design parameters (or decision variables) change
depending on the nature of the separation
Procedure:
Set Ɵ1 Ɵ2 Ɵ3 values
Cf: Known, find Cp1 and from rejection values, you can
calculate Cr1.
Using set values, you can calculate
Cp1, Cr1, Cp2, Cr2, Cp3, Cr3

Now we have solve 6 equations simultaneously to find

qr1, qp1, qr2, qp2, qr3, qp3


Overall mass balance around the unit-1

q f  qr 3  q p1  qr1 (1)
Solute balance around the unit-1

q f C f  qr 3Cr 3  q p1C p1  qr1Cr1 (2) Solve

Ax=b
Overall mass balance around the unit-1

qr 1  q p 2  q r 2 (3) Linear system


of equations
Solute balance around the unit-1

qr1Cr1  q p 2C p 2  qr 2Cr 2 (4)


Overall mass balance around the unit-3

q p 2  q p 3  qr 3 (5)
Solute balance around the unit-3

q p 2C p 2  q p 3C p 3  qr 3Cr 3 (6)
qr1 qp1 qr2 qp2 qr3 qp3

1 1 0 0 -1 0
Cr1 Cp1 0 0 -Cr3 0
-1 0 1 1 0 0
A= -Cr1 0 Cr2 Cp2 0 0
0 0 0 -1 1 1
0 0 0 -Cp2 Cr3 Cp3

qf
qr1
qfCf
qp1 x=A-1b
b= 0 x= qr2
0
qp2 Inverse of the
0
qr3 matrix
0
qp3
Solving linear set of equations in Excel

https://www.youtube.com/watch?v=vLsouen-G8M
C p1 C p2 C p3
1  2  3  Cpfinal
Cf C r1 C p2 (g/L) %Reduction

0.6 0.5 0.1 767.7 41.4

0.3 0.3 0.1 344.6 73.7

0.4 0.3 0.2 518.3 60.4

C p1 C p2 C p3
1  2  3  Area1 Area2 Area3 Atotal
Cf C r1 C p2 (m2) (m2) (m2) (m2)

0.6 0.5 0.1 16.77 4.11 1.97 22.86

0.3 0.3 0.1 5.66 3.55 1.70 10.91

0.4 0.3 0.2 10.26 2.25 2.04 14.55


C p1 C p2 C p3
1  2  3  qp,total qr,total Cpfinal Atotal
Cf C r1 C p2 (L/h) (L/h) (g/L) (m2)

0.6 0.5 0.1 808.80 91.20 767.7 22.86

0.3 0.3 0.1 317.60 582.40 344.6 10.91

0.4 0.3 0.2 530.80 369.10 518.3 14.55


Comparison of 2 Different Designs
C pFinal
Objective:  0.26
Cf

Operating conditions: T=40 C ; Transmembrane pressure:0.2 bar


Pectin rejection: 75 %

  0.5 For each stage

 C pN 
C pN  C f  N
ln    N ln   N 2
C 
 f 
Design-1

qR1
qf=900 L/h
Cf=1310 g/L A1=13.8 m2

qp1=594.6 L/h qR2


Cf=1310 g/L A2=9.1 m2

qp2=392.8 L/h
Cp2=327.5 g/L
Design-2

qR1 qR2
qf=900L/h A1=5.7 m2 A2=3.6 m2
Cf=1310g/L qR3
qp2 =153.1 L/h
Cp2=465.3 g/L A3=1.7 m2

qp1=244.2 L/h
Cp1=393 g/L qp3 =73.4 L/h
Cp3=183.6 g/L

Permeate
qp=317.6L/h
Cp=344.6 g/L
Design-1: Design -2

qp=317.6 L/h
qp=392.8 L/h
Cp=344.6 g/L
Cp=327.5 g/L
N=3
N=2
A1=5.7 m2
A1=13.8 m2
A2=3.6 m2
A2=9.1 m2
A3=1.7 m2
Atotal=22.9 m2
Atotal=10.9 m2

Fixed Cost Fixed Cost

Think about other cost items !


Critical Question

Total flux decreases due to fouling. We used experimental flux data


to calculate the area.

Is it possible to predict the cake layer resistance so that we can


predict flux without requiring experimental data ?
Alternative Flowsheets
CHE 449
MEMBRANE PROCESSES and
SEPARATION of MIXTURES

CHAPTER 4
NANOFILTRATION
https://synderfiltration.com/ultrafiltration/membranes/
NANOFILTRATION
 Developed after 1970’s. Why ?
 Reverse osmosis (RO) membranes: Have
low flux at relatively low pressures,
resulted in a considerable energy cost
 The quality of the obtained permeate with
RO is often too
good.
 Thus, membranes with lower rejections of
dissolved components, but with a higher
water permeability were desired .
 Such low-pressure RO membranes became
known as nanofiltration membranes.
Some Applications for NF systems are:

•Water softening

•Specific removal of heavy metals from process streams


for reuse of water

•Reduction of salt contents of slightly brackish water

•Recovery of organic solvents


FEATURES OF NANOFILTRATION MEMBRANES

 High rejections for multivalent ions (99%)

 Low to moderate rejections for monovalent ions (0–


70%)

 High rejection (90%) for organic compounds with a


molecular weight above the molecular weight of the
membrane, which is usually in the range of 150–300.
https://www.safewater.org/fact-sheets-1/2017/1/23/ultrafiltrationnanoandro
Nanofiltration Membrane Preparation Methods

Phase Inversion Thin Film Composite


One step process (At least 2 step process)
• Interfacial polymerization
• Coating

 NF membranes contain functional groups that can be


charged, depending on the pH of the solution in contact with
the membrane.
MEMBRANE MATERIALS AND SYNTHESIS PROCEDURES

Tight NF Membrane Loose NF Membrane


Solvent Resistant NF Membranes

 Membranes show important performance loss if they don’t have


resistance to organic solvents.

 Zero flux due to membrane collapse

 “infinite” flux due to membrane swelling

 Membrane deterioration

 Poor separation quality


Application of SRNF in the vegetable oil industry

• 2 million tons of extraction solvent is used in the United


States alone.

• Starting from seed from plants as soybean or sunflower, oil


is obtained by solvent extraction, eventually in combination
with mechanical extraction.

• Hexane is the most common extraction solvent.

• Currently, evaporation is used to recover these solvents and


reuse them in the process.

• Amount of energy consumed for evaporation is


approximately 530 kJ/kg oil.

• The elevated temperatures hold a risk on thermal damage,


and explosive vapors may create safety problems

• These limitations can be partially overcome by membrane


technology.
Application of SRNF in the pharmaceutical industry

• Many fine chemicals or pharmaceutically active ingredients are


synthesized with complex reaction paths.

• Different steps require different solvents.

• Traditional method: “put-and-take distillation,” in which the origina


solvent is stepwise boiled out and replaced by similar
volumes of the second solvent.

• Limitations/Drawbacks of the traditional technique:


Only effective if the first solvent has a much lower boiling
temperature than the second solvent
Azeotropic mixtures may cause additional problems.
Application of SRNF in the concentration of oils
from oil-micelle mixtures

• Oil–micelle mixtures are formed during hexane extraction


and consist of triglycerides (oils), phospholipids, and
solvent.

• A mixture containing 20% of oil could be concentrated to


45% with a commercial SRNF membrane resulting in a 50%
reduction of the energy consumption for the evaporation
unit.
The comparison of the energy consumption for OSN
and distillation

Basis: Recovery of 451 tons


of methanol by distillation

Sustainability Assessment of Organic Solvent Nanofiltration: from Fabrication to Application.


July 2014. Green Chemistry 16(10)
• The comparison was made based on a pressure pump for the nanofiltration
system, and a total reboiler for the distillation

• The energy requirement of the pressure pump required to provide a


pressure difference ΔP for a feed flow Ff

• The energy required in a throughput distillation


Application of SRNF in the petrochemical industry

• Wax is used to modify the properties of lube oil and


must therefore be removed.

• Classical method: Cooling of the feed to precipitate wax and


filtration of wax.

Multistage flash /distillation to recover solvent


Cooling solvent before recycling in the dewaxing process
Membrane-Based Separation Process
• Waxy feed is dissolved in methyl ethyl ketone (MEK) and toluene to
precipitate the wax components
• Chilled wax is filtered. Wax is recovered in retentate. Permeate is
lube oil in MEK and toluene.

• A membrane process can recover the solvents for recycle to the


wax precipitation step.

• Lube oil molecular weight: In the range of 300 g/mol and


higher. Solvents MWs: 84 and 96 g /mol. Therefore this
can be considered as mostly a sized based separation
suitable for OSN.

Economic benefits of membrane-based separation process:


• Reduced energy consumption (20% per unit volume of product)
• Increased yield of lube and wax products from a barrel of oil
(3–5%)
• Increased product quality , the overall solvent consumption is
kept low.
• The membrane-based solvent recovery unit of ExxonMobil’s
refinery allows the annual recovery of 300 000 m3 of clean
solvent in the lube processing area.
Application of SRNT in the recovery of homogeneous
catalysts
Inorganic synthesis of complex molecules,homogeneous catalysis has
become increasingly important. Reaction medium includes solvent

Challenge: Separation of product and catalyst. Possible with OSNR


membranes.

Catalyst recycle
Performance of NF Membranes

Three parameters are crucial for the operation of a


(nano)filtration unit:

• Solvent permeability or flux through the membrane

• Rejection of solutes

• Recovery
Solvent Flux
J v L p P   
Solute Flux

J s C 1   J v  C
Lp: Hydraulic permeability of solvent

ΔP: Transmembrane pressure difference

ΔП: Osmotic pressure difference

ω : The solute permeability Cf:Feed concentration


σ : The reflection coefficient Cp: Permeate concentration

C f  Cp
C
Cf
ln
Cp
For pure water
J v L p P J s C 1   J v  C

Js C
1   J v  
Jv
Slope=Lp

C C

ΔP

Js
C
Slope= 1   

C
Jv
C
Rejection mechanisms in NF

Steric (size) exclusion Donnan (charge) exclusion


• Species with hydrated size larger • A membrane with fixed charged
than the membrane pore size are groups repels ions with the same
sterically retained charge (co-ions) and attracts ions
• Species with size similar to that of with the opposite charge
the membrane pores are hindered. (counterions)
Case Study-Demineralization of Lactose

• In the concentration of whey protein, the permeate stream


mainly contains salty lactose solution.

• If salt is removed from lactose, edible and high quality lactose


can be produced.

• If salt is removed, fouling in the evaporator due to precipitation


of salts is also reduced.

• Demineralized lactose has applications in the food,


pharmaceutical, and brewing industries.

Desired Features of the Nanofiltration Membrane

• High lactose rejection


• High monovalent salt passage
• High flux
Food Sci Tech Int 2004; 10(4):255–262
© 2004 Sage Publications
ISSN: 1082-0132
DOI: 10.1177/1082013204045883
For Na+
600.00
Js

C
1   J v  
y = 2.682x - 357.13
R² = 0.7088
C C
500.00

400.00
For partially rejected
Js 300.00
 1 solutes
C 200.00

100.00 But data shows that


0.00
0 50 100 150 200 250 300 350
 0
C
Jv
C Not possible
Js
Assume that  1 J s C Psolute   
C

Cp
C  C f  C p R  1 C  RC f
Cf
Membrane : NF-200

NaCl PNaCl=107 L/m2.h


KCl PKCl=73.6 L/m2.h
MgSO4 PMgSO4=1.3 L/m2.h
Design Criteria: Concentration of salts in retentate stream (Cr) (where
lactose is mainly concentrated)

Procedure:

• Assume Cp value
• Calculate osmotic pressure difference

Calculation of Osmotic Pressures

On the upper side of the membrane, salt concentration changes


from Cf to Cr. Thus, we should take the average of these concentrations

C f  Cr
 Feed  iC f RT Cf 
2    Feed   Permeate
 Permeate  iC p RT
We should check the assumed Cp values for each salt

Js J s C
J s J wC p Cp 
Jw J v L p P   
• If CpAssumed –Cpcalculated > Tolerance , iteration should be
continued with new Cp values!
• Once iteration is calculated, then we can apply total and
component mass balances

q f  q p  qr q f C f  q p C p  qr C r

 C f  Cr  qp
qp   q f Area 
 C C  Jw
 p r 
CHE 449
MEMBRANE PROCESSES and
SEPARATION of MIXTURES

CHAPTER 5
REVERSE OSMOSIS
Applications of RO Process

 Desalinating brackish water or


seawater (50 %)
 Producing ultrapure water for the
electronics, pharmaceutical, and
power generation industries (40 %).
 Small niche applications such as
pollution control and food
processing (10 %).
Theoretical Background

.
When does rising
stop?

When the tube is immersed into water


what happens ?

Membrane can only permeate water but


not salt
OSMOTIC EQUILIBRIUM
❑ Membrane allows the solvent, but not the solute, to pass.
❑ Water diffuses through the membrane toward the solution of higher
concentration, until equilibrium is attained.
❑ During this process the volume of the side that initially contained the
higher concentration increases.
❑ The pressure that must be applied to this side to reverse the process of
osmosis is called the OSMOTIC PRESSURE
OSMOSIS AND REVERSE OSMOSIS

Driving forces: Concentration difference + Pressure difference

Feed: I w:Water
Membrane
Salt +H2O Pure Water: II S: Salt

 w I −  w II =  w (P I − P II ) −  w ( I −  II )

a) Case 1: P I = P II Side II: Pure water  II = 0  wI −  wII  0  wII   wI

This is OSMOSIS
Diffusion of water from pure water side (II) to salt solution side (I)
 Case 2: (P I
−P II
) = ( I
− II
)  wI =  wII

This is OSMOTIC EQUILIBRIUM

 Case 3:
(P I
−P II
)  ( I
− II
)  wI   wII

This is REVERSE OSMOSIS


Diffusion of water from salt solution side (I) to pure water side (II)

 Reverse osmosis : Very important process to


produce pure water from salt solutions.
• The RO membrane units in large-scale desalination plants
consist of a large number of membrane elements connected
together by certain patterns.

• An RO membrane element is the basic membrane device that


is engineered in such a way that a large area of membrane
can be packed into as small as possible a volume.

• Certain configurations are used to build membrane elements


so that the raw water can be fed to all the membrane areas
and the permeate and retentate can be easily removed from
the elements without mixing.
• Multistage RO systems: Often used to further increase energy
utilization and permeate production

• The concentrate of a stage is used as feed stream for the next stage.

• The main objective is to increase the amount of permeate produced


from the same amount of raw water
• Multiple-pass RO system: Used when a higher permeate quality is
needed.

• Permeate of one pass serves as the feed for the next pass in a two-
pass RO system.

• Much lower salt concentration in the permeate can be obtained.

• Multiple-pass RO systems are not as commonly employed as


the multiple-stage RO systems.

• Some cases where multiple-pass RO systems can play a role are:


Desalination of seawater with very high salt concentration and
The production of ultrapure water where very low salt
concentration is required.
Performance Indicators

• The permeate flux

• The salt concentration in the permeate

• The recovery of raw water, that is, the ratio of


permeate volume to the raw feed water volume is
also an important concern of an RO process.
Case Study

A red pepper farmer in Italy wants to prepare process water from a


brackish water source with a single reverse osmosis (RO)-unit.
The source contains 3000 ppm salt (NaCl) and the required water
Quality must be at least 200 ppm. In other words, the water must
Contain less than 200 ppm of NaCl.
The required capacity is 10 m3/hour.
Requirements for the Water Treatment Unit

1. Installation should be single-pass.


2. The maximum applied pressure should be 42 bar
3. Water recovery should be at least 75%

4 different companies provided following the technical data sheet for the
membrane modules. Which one do you suggest the farmer buy ?

Note: 1) Each module is made from the same material.


2) Retention and flux data were determined with a solution of
3000 ppm NaCl and a pressure of 28 bar.
➢ This is a preliminary evaluation. Based on given data , we should
find a reliable way to choose the best module .

➢ What should be the best criteria for the selection ?

COST ????

Yes. The cost of the modules is proportional to area of each module

Larger the area of the module, the higher the cost. Thus, we should
calculate the area of each module to meet the requirements!
Single-Pass System

Given Data:
Capacity:qf: 10 m3/hour. Recovery:qp/qf:0.75
Cf: Salt concentration in brackish water: 3000 ppm
Cp: Desired salt concentration in treated water (permeate): 200 ppm

We can find qp=0.75 (10) =7.5 m3/h=7500 L/h

Permeate stream mainly consists of water, thus, qp=qwater

We should find an equation which will define rate of permeation of


water through the membrane
Rate of permeation (L/h) =Flux x Area
The rate of permeation of water through the membrane (Jw)
is given by the following equation

J w = Pew (P −  )
Pew: Permeability constant of water
ΔP: Pressure difference applied through the membrane
Δπ:Osmotic pressure difference applied through the membrane

Osmotic pressure can be calculated by the following equation

 = iCRT
Cs: Solute (NaCl) concentration (mole/L)
R: Gas constant (0.08314 L.Bar/mole. K
T: Temperature (K)
+ −
i: Number of ions dissociated in water. NaCl → Na + Cl
For NaCl i:2
Rate of permeation (L/h) =Flux x Area

Area=qp /Jw qp=0.75 (10) =7.5 m3/h=7500 L/h

J w = Pew (P −  )
To calculate flux under desired operating conditions, we
Should know Pew.

The maximum applied pressure should be 42 bar ΔP=42 bar

Cf: Salt concentration in brackish water: 3000 ppm


Cp: Desired salt concentration in treated water (permeate): 200 ppm

 =  Feed −  Permeate
 Feed = iC f RT
 =  Feed −  Permeate
 Permeate = iC p RT

J w = Pew (P −  )
P = 28 bar  = 6.1 − 0.17 = 5.93 bar Pew = ?

We can use the technical data provided by the company to calculate


the permeability of the membranes.
Retention (R ) and flux (Jw) data were determined with a solution of
3000 ppm NaCl and a pressure of 28 bar.

Cp
R = 1− C p = (1 − R )C f
Cf
For module A
Jw
Pew = Pew =
480
= 21.7 L / m 2 h.bar
(P −  )
Case 1:
➢ Recovery: 75 %
➢ Desired salt concentration in the permeate will be
100 ppm, 200 ppm and 500 ppm

Case 2:
➢ Desired salt concentration in the permeate: 200 ppm
➢ Recovery of water will be 50 %, 75 % and 90 %

All other design conditions will the same as reported in the problem.

Plot area required for MODULE A

1) As a function of salt concentration in the permeate

2) As a function of recovery of water

Comment on the results


We wrote that
C f = Cr + C p Is this balance correct ???? NO !

We should apply component (for salt) mass balance around the unit
qp
q f C f = qr C r + q p C p We know the recovery S = Re cov ery =
qf
C f − SC p Cr = 11400 ppm
Cr =
1− S
Cr = C f − C p C f − SC p
Cr =
1− S
Correct result
Cr = 2800 ppm Cr = 11400 ppm

Huge difference between the results


How is the area calculation affected from this common mistake

qp
A=
Pew (P −  )

Since osmotic pressure is a function of dissolved salt concentration,


the change in osmotic pressure difference will influence the area calculation!

C f + Cr
 Feed = iC f RT Cf =
2
C f + Cr
 Feed = iC f RT Cf =
2
Cr = C f − C p Cr = 2800 ppm  Feed = 2.46 bar

 = 2.46 − 0.17 = 2.29 bar

C f − SC p  Feed = 6.1 bar


Cr = Cr = 11400 ppm
1− S
 = 6.1 − 0.17 = 5.93 bar
Correct value
qp
A=
Pew (P −  )
Lets know compare the areas calculated using wrong Cr and correct
Cr values
Calculated with wrong Cr value Cr = 2800 ppm
Module Retention Flux Membrane
Area (m2)
Module A 0.9 480 10.2
Module B 0.95 320 15.2
Module C 0.97 200 24.2
Module D 0.98 80 60.4

Calculated with correct Cr value Cr = 11400 ppm


Module Retention Flux Membrane
Area (m2)
Module A 0.9 480 11.14
Module B 0.95 320 16.62
Module C 0.97 200 26.55
Module D 0.98 80 66.30

The error in area calculation becomes larger as the salt rejection increases
ENERGY CONSUMPTION IN RO PROCESSES

 Turbine: Extracts energy from a fluid. As the


pressure of the fluid is decreased from P1 to P2
(expansion of liquid from high pressure to low
pressure) work is produced.
 Energy produced bya a turbine=-ηqΔP

η:Turbine efficiency (η changes between 0.5 and 0.8).


q: Volumetric flowrate (m3/s)

 Pump: Adds energy to a fluid (Convert mechanical


energy to hydraulic energy). Pressure of the fluid is
increased from P1 to P2 and for this energy is
consumed.
 Energy consumption by a pump=qΔP/η
 η:Pump efficiency (η changes between 0.5 and 0.8).
ENERGY CONSUMPTION IN RO PROCESSES

 A pressure exchanger transfers pressure energy


from a high pressure fluid stream to a low pressure
fluid stream.

 In RO process, exit retentate stream has a high


pressure. This high pressure fluid can be sent to a
turbine to produce work or it can be sent to a
pressure exchanger to increase pressure of the feed
stream.

 Specific energy consumption (kWh/m3) required to


to produce permeate:Se

Se =
(q P )/
f f p − T qr Pr
Pf, Pr:Feed and retentate pressures
qp
Schematics of a reverse osmosis system (desalination) using a
pressure exchanger.
1:Sea water inflow
2: Fresh water flow (40%)
3:Concentrate Flow (60%)
4:Sea water flow (60%)
5: Concentrate (drain)
A: High pressure pump flow (40%), B: Circulation pump,
C:Osmosis unit with membrane, D: Pressure exchanger
Case Study- In recent years your company has been experiencing water
shortage problem. Therefore, they would like to install a desalination
plant to process sea water. The salt content of sea water was measured
as 35000 ppm. The desalination unit will have a capacity of 20000 m3/day and
recovery rate of water will be 50 %. The plant will be operated 7884 h/year.
You have been asked to evaluate three possible designs.
1) Membrane module integrated with a turbine for energy recovery
2) Membrane module integrated with a pressure exchange unit for energy
recovery
3) Standalone system (No energy recovery unit).

The design should allow maximum 1 bar pressure drop through the
membrane module.
Prepare a feasibility report on your choice. In your report you should compare
these three alternatives and explain the reasons for your choice. To convince
your boss, you should compare the feasibility of your design with others
available in the literature.
Additional Data Supplied :
Efficiency of high pressure pump: 84 %; Cost of pump: 25000 $
Efficiency of turbine: 80 % ; Cost of turbine: 3000 $/kW
Efficiency of pressure exchange device : 98 % ; Cost of pressure
exchanger: 300000 $
Power Cost: 0.06 $/kWh
Labor Cost: 0.01 $/m3; Maintenance Cost: 0.01 $/m3; 0.01 $/m3;
Chemical Cost: 0.01 $/m3
Critical Step: Choosing a suitable membrane
Performance Criteria for RO Membrane Selection

• Water permeability (A)

• Salt permeability (B)

Selectivity= A / B
fcp: Describes the effect of concentration polarization

Water flux

Mass transfer coefficient

• Jw ≪ K then fcp=1 (Concentration polarization becomes negligible)

Journal of Membrane Science 590 (2019) 117297


Commercial Membrane Development

https://wetlab.net.technion.ac.il/files/2021/10/FregerRamon-PPS-2021.pdf
https://wetlab.net.technion.ac.il/files/2021/10/FregerRamon-PPS-2021.pdf
q f = qr + q p q f C f = qr C r + q p C p

J Total = J water + J salt

J water  J salt JTotal = J water


J water = A(P −  ) J Salt = B(C f − C p )

J salt = J TotalC p
kg salt/L
kg salt/m2.h L/m2.h

qp
Design Criteria: Set the recovery of water : S S=
qf
J salt = J W aterC p = B(C f − C p )
If we rearrange this equation, it becomes

BC f
Cp =
JW ater + B

q p = Sq f qr = q f − q p

q f C f − q pC p
Cr =
qr
Calculation Procedure
• Assume Cp value

• Calculate Cr value.

• Calculate osmotic pressures:

C f + Cr  Permeate = iC p RT
 Feed = iC f RT Cf =
2

 =  Feed −  Permeate J water = A(P −  )

• Calculate Cp value from BC f


Cp =
JW ater + B
• If Cp,Assumed –Cp,Calculated < Tolerance then calculations are done.

• You can do this iteration very easily in Excel using Solver/Goal Seek
Recovery: 50 %

Transmembrane Pressure: 50 Bar

Stand Turbine Pressure


Alone Exchange
Area (m2) 56283 56283 56283
Fixed
Cost($/year) 115,871 173,027 137,157
Operating Cost 811,238 684,218 655,638
($/year)
Total Cost 927,109 857,245 792,795
($/year)
Specific 1.98 1.34 1.19
Energy
Consumption
(kwh/m3)
Recovery: 60 %

Transmembrane Pressure: 50 Bar

Stand Turbine Pressure


Alone Exchange
Area (m2) 161120 161120 161120

Fixed 328,398 374,123 349,684


Cost($/year)
Operating Cost 1,430,015 1,328,399 1,305,536
($/year)
Total Cost 1,758,413 1,702,522 1,655,220
($/year)
Specific Energy 1.65 1.22 1.12
Consumption
(kwh/m3)

As the recovery of water increases, the total cost and specific energy
consumption increases significantly too !
RO MEMBRANES

https://hdl.handle.net/1721.1/122632
From single stage to two-stage RO
To minimize specific energy consumption
Permeate pass to achieve
higher permeate
quality

• By applying an ERD, the specific energy


consumption can be decreased by up to 25%

Seawater RO: The average water permeate flux : 12-17 L/m2.h (LMH)
Feed pump pressure: 55-80 bar,
The recovery ratio: 35-50%
Salt rejection : 99.4-99.7%.

Brackishwater: The average water permeate flux: 12-45 LMH


Feed pump pressure: 6-30 bar
The recovery ratio: 50-85%
Salt rejection: 95-99%
• Practical minimum energy: 1.56 kWh/m3
• Best reported RO energy consumption:
~ 2 kWh/m3

Elimelech, M. and Phillip, W.A., Science, 333, 2011, 712-717.


The decrease in energy consumption is attributed to

• Higher-permeability membranes
• Installation of energy recovery devices
• Use of more efficient pumps

The minimun amount of energy required to separate pure


water from seawater
• Independent of the desalination method
• Realized when the separation occurs as a reversible
thermodynamic process

Energy Environ. Sci., 2014, 7, 1134–1141
➢ Seawater desalination (35 g/L)

Minimum energy:

➢ Recovery: 0% 0.76 kWh/m3


➢ Recovery: 50% 1.06 kWh/m3

The theoretical minimum energy of desalination for


seawater at 35,000 parts per million (ppm) salt
and at a typical recovery of 50% is 1.06 kWh/m3
Elimelech, M. and Phillip, W.A., Science, 333, 2011, 712-717.
The Benefit of Using a 2-Stage RO Design

Recovery: 50 % Csalt:35000 ppm

Minimum energy requirement: 1.06 kWh/m3


Estimated energy requirement for a single stage RO: 1.56 kWh/m3

0.5 kWh/m3 energy: For irreversible losses such as frictional losses

Estimated energy requirement for a 2-stage stage RO: 1.28


kWh/m3

First Stage : Operating pressure < Second Stage

Volume of water entering to the 2nd stage is less , thus, energy


Consumption is less

Consider total cost : Operating cost +Fixed cost


How much more can we decrease energy consumption
for RO processes?

With best membranes and an ideal RO stage,


Practical energy consumption : 25 % higher than minimum energy
consumption

Overall energy consumption : 3-4 times higher than theoretical


Minimum energy consumption:

RO Process Energy Requirement:

Pretreatment + RO (Salt removal) + Postreatment

Energy consumption for these processes should be


decreased
PH > П c

Thermodynamic limit

Elimelech, M. and Phillip, W.A., Science, 333, 2011, 712-717.


Pressure drop and pump work

• Pressure loss in the feed side path is usually expressed by


the Darcy-Weisbach equation:

fD : The Darcy friction factor, ρF: Density of feed flow,


uF : The velocity of feed side bulk, L :The length of feed side path
dF :Hydraulic diameter of the feed side path
fD=4f f:Fanning friction factor

• Pressure loss in the permeate side path can be


expressed by Darcy-Weisbach equation or empirical equations
due to the complexity of permeate side spacer configuration.
https://doi.org/10.1016/j.desal.2019.07.004

You might also like