TECHNICAL MANUAL

REVERSE OSMOSIS MEMBRANE

 Table of Contents
0
1. Introduction of Toray Advanced Materials Korea Inc.

2. Introduction of Reverse Osmosis Membranes

3. Specifications of Toray Advanced Materials Korea CSM

4. Water Chemistry and Pretreatment

5. System Design

6. System Operation

7. Cleaning in Place, Cleaning and Disinfection

8. Replacing Elements, Assembly and Loading of Pressure Vessels

9. Troubleshooting
 1 Introduction to TAK

1-1 History of Toray Advanced materials Korea Inc.

 1999 Launch of Toray Saehan Industries, joint venture with Toray

 2000 Establishment of Technology Research Institute
 2001 Winner of Energy Innovation Management Awards
 2002 Development of IT Material Business
 2003 OHSAS 18001, First certification
 2004 Establishment of Advanced Material Research Center
 2005 Gumi 4-Factory Construction Co., Ltd.
 2008 Opening of Photovoltaic Power Generation at Gumi Plant 3
 2009 Establishment of New Technology Research Institute
 2010 Change of Company name Toray Advanced materials Korea Inc.
 2011 ISO 50001, certification
 2012 Establishment of TCK textile
 2013 Super engineering plastic PPS, Business
 2014 Establishment of Gunsan Plant
 2015 Establishment of TCK micro-filter
 2016 Gumi 5-Factory Construction Co., Ltd.
 2017 Opening of TAK, a Chinese subsidiary
 2018 Construction of the Korea Toray R&D Center

1-2 History of Reverse Osmosis Membrane Business

 1990 Research on the reverse osmosis membrane commences at the R&D center
 1994 Development of high salt rejection and flux reverse osmosis membranes
Completion of membrane testing pilot plant at R&D center
Membranes receives Japanese High Polymer certificate
Membranes become FDA approved
 1995 Commercialization of reverse osmosis membranes for industrial and residential use
Export to China, Taiwan and India
 1997 Completion of main manufacturing facility at Gyungsan Plant
Reverse osmosis membranes receive Industrial Research 52 Award from the Korean
Ministry of Science and Technology and the KT mark (new technology award)
 1998 Development of reverse osmosis membranes for seawater applications
 2000 Development of fouling resistant reverse osmosis membranes

 2003 Expansion of reverse osmosis membrane manufacturing capacity at Gumi Plant I

Development of reverse osmosis membranes for ultrapure water applications
 2 Introduction to Reverse Osmosis Membrane

2-1. Overview of Reverse Osmosis Membrane Applications

Rapid technological advancement brought about many benefits to our life and also
concurrently caused pollution to our environments such as contaminating water sources
from industry effluents.
Moreover, natural fresh water source could not meet the ever increasing water demand
from our growing population and industries. Additionally, uneven seasonal rain fall
makes the water shortage problem worse. Therefore, insuring enough quantity and
quality of water is one of the most serious issues in the world, especially in Europe, the
United States, China and Middle East and North Africa.
Distillation has been traditionally used to obtain pure water from contaminated water
sources. Other processes such as ion exchange and electrodialysis have been employed
for water purification since 1950. Recently, RO membrane was applied to the
desalination of sea water and brackish water.
RO has been proven to be the most economical technology not only for the desalination
of water containing salts, but also for purifying water contaminated with heavy metals,
pesticides and other contaminants. Especially, seawater desalination particularly with the
utilization of a reverse osmosis (RO) membrane, has been a rapid growth in the global
market. In many Middle Eastern countries such as Saudi Arabia, the United Arab
Emirates, and Kuwait, significiantly, desalinated seawater occupies a major portion of
their available water resources. Although thermally driven processes are still used in a
majority of their desalination plants, most plants built after 1990 adopt the RO membrane
separation process due to its lower operational/maintenance costs and smaller
environmental footprint.
And also RO can be used for recycling waste water and reclaiming useful materials
from waste stream such as the recovery of dyes from dye industry effluents. The
applications for RO also include food and beverage processing. In addition, advances in
the fields of biotechnology and pharmaceuticals, coupled with advances in new
membrane development, are making membranes an important separation step, which
offers energy savings without leading to thermal degradation of the products.
 Introduction to Reverse Osmosis Membrane
2
2-2. Theory of Reverse Osmosis Membrane

The phenomenon of osmosis is illustrated in the Figure below.

A semi-permeable membrane (RO membrane) is placed between two compartments. An
RO membrane is consisted of a supporting layer with 50㎛ in thickness and a barrier
layer with about 0.2㎛ in thickness. The phenomenon of osmosis occurs when pure water
flows from a dilute saline solution in one compartment through the RO membrane into a
higher concentrated saline solution in the other causing a rise in the height of the salt
solution in the compartment of the higher concentrated solution.
The water flow will stop when the pressure of the column of the salt solution equals to
the difference in chemical potential between the two aqueous solutions. The equilibrium
point of the water column height in terms of water pressure against the membrane is called
osmotic pressure. If a force is applied to this column of water, the direction of water flow
through the membrane can be reversed. This phenomenon is called reverse osmosis. This
reversed flow produces pure water from the salt solution, since the membrane is not
permeable to salt. By this manner, relatively pure water passes through membrane, and
the dissolved solids are retained in the other section Consequently, the water in this
section is purified or demineralized and the solids in the other section are concentrated or
dewatered
 2 Introduction to Reverse Osmosis Membrane

2-3. Types of Reverse Osmosis Membrane

◎ Asymmetric Membrane --- Cellulose Acetate(CA) Membrane

Historically, the asymmetric membrane is formed by casting a thin film

acetone-based solution of cellulose acetate(CA) polymer, which was developed by
Loeb and Sourirajan in 1962 and the first commercially viable RO membrane.
The resulting CA membrane has an asymmetric structure with a dense surface layer
of about 0.1 - 0.2㎛ which is responsible for the salt rejection property. The rest of
the membrane, which is 100-200㎛ thick and supports the thin surface layer
mechanically, is spongy and porous, and has high water permeability. Salt rejection
and water flux of a CA membrane can be controlled by variations in temperature and
duration of the annealing step.

◎ Thin Film Composite Membrane --- Polyamide(PA) Membrane

Thin film composite (TFC) polyamide membranes are consisted of a porous support
layer and a thin film dense layer which is a cross linked membrane skin and is formed
in situ on the porous support layer, usually made of polysulfone. The thin film dense
layer is a cross-linked aromatic polyamide made from interfacial polymerization
reaction of a polyfunctional amine such as m-phenylenediamine with a polyfunctional
acid chloride such as tri-mesoyl chloride. This TFC manufacturing procedure enables
independent optimization of the distinct properties of the support and salt rejecting
skin. The TFC membrane is characterized by higher specific water flux and higher salt
rejection than cellulose acetate membranes.

◎ Comparison of Polyamide TFC Membranes with Cellulose Acetate(CA)

Membranes

TFC membranes have advantages over single-material asymmetric membranes in that,

the top-separating layer is formed in situ and hence the chemistry and performance of the top
barrier layer and the bottom porous substrate can be independently studied and modified to
maximize the overall membrane performance. As mentioned above, the TFC membranes
exhibit higher water flux and higher salt rejection than CA membranes which had
been used widely until the commercial introduction of TFC membranes in 1981. TFC
membranes are stable over a wider pH range and operable at lower pressure than CA
 Introduction to Reverse Osmosis Membrane
2
membranes. Detailed comparisons between the two types of membranes are shown in
the table below.

Parameters PA Membrane CA Membrane

Operating pH range 2~12 4~8

Operating Temperature (oC) 0~45 0~35

TDS 99+ 98
Salt Rejection (%)
Silica(SiO2) 99+ < 95

Productivity (Flux) High Low

Salt Rejection Change after 3years 99%→98.7% 98%→96%

Chlorine Tolerance <0.1 ppm 1 ppm

Membrane Fouling Moderate Low

Compaction Resistance High Low

Susceptibility to Biodegradation Negligible High

 2 Introduction to Reverse Osmosis Membrane

2-4. General Property of CSM RO Membranes

◎ General Separation Property of RO Membrane

1) Inorganic solutes are rejected by RO membrane better than organic solutes.

Organic solutes with molecular weight (MW) larger than 100 are also well rejected
by the membrane
2) Ionizable solutes are rejected better than non-ionizable solutes.
3) Ionizable solutes with higher charges are rejected better than lower charges.
For examples, aluminum ion (Al3+) is rejected better than magnesium ion (Mg2+)
which is in turn rejected better than sodium ion (Na+).
4) The rejection of inorganic solutes depends also on the size of the ions and the size of
hydrated ions. The bigger the ions and the hydrated ions, the better they are rejected
5) The bigger the non-ionizable solutes (the higher the molecular weight), the better the
rejection.
6) Gases with MW lower than 100 can easily permeate through the membrane.
For example, the rejection of ammonia, chlorine, carbon dioxide, oxygen and
hydrogen sulfide is very low.
7) The rejection of weak acids is low, which also depends on the MW of the acids. The
rejection of the following acids is decreasing in the order of citric acid, tartaric acid
and acetic acid as the MW of the acids decreases.
8) The greater the dissociation, the higher rejection, for example boric acids are
rejected better at higher pH
9) The more branching, the greater rejection, for example, isopropanol is rejected better
than normal propanol

◎ Characteristic Property of CSM RO Membranes.

1) High permeate flux and high salt rejection

2) Chemically stable in a wide range of pH (pH2 - 12)
3) Long membrane life time
4) Resistant to a biological attack
5) Operable in a wide range of pressure (20 - 1000psi)
6) Operable at a wide range of temperature(4 - 45℃)
7) Economical
 Introduction to Reverse Osmosis Membrane
2
2-5. Solute Rejection Property of CSM RO Membranes

◎ Solute Rejection of BE and BN Grade of CSM

NO Solute Rejection (%) Molecular Weight

1 NaF 99 42
2 NaCN 98 49
3 NaCl 99 58
4 SiO2 99 60
5 NaHCO3 99 84
6 NaNO3 97 85
7 MgCl2 99 95
8 CaCl2 99 111
9 MgSO4 99 120
10 NiSO4 99 155
11 CuSO4 99 160
12 Formaldehyde 35 30
13 Methanol 25 32
14 Ethanol 70 46
15 Isopropanol 92 60
16 Urea 70 60
17 Lactic acid (pH2) 94 90
18 Lactic acid (pH5) 99 90
19 Glucose 98 180
20 Sucrose 99 342
21 Chlorinated Pesticides 99 -
22 BOD 95 -
23 COD 97 -
 3 Specification of CSM

3-1. CSM Element Nomenclature

RE8040-BE
① Types of Element
- RE : Reverse Osmosis Element
- NE : Nanofiltration Element
- UE : Ultrafiltration Element

② Diameter of Element
- 80 : 8.0 inch
- 40 : 4.0 inch
- 25 : 2.5 inch

③ Length of Element
- 40 : 40 inch
- 21 : 21 inch

④ Related to Specification
1) The First Order
- B : Brackish Water Membrane
- T : Tap Water Membrane
- S : Sea Water Membrane
- F : Fouling Resistant Membrane
2) The Second Order
- N : Normal grade
- E : Extended effective area
- R : High salt rejection
- L : Low pressure use
3) In case of BL grade(three orders)
- BLN : Normal grade
- BLR : High salt rejection grade
- BLF : High flux grade
4) In case of NE element
- 40 : NaCl Rejection 40%
- 90 : NaCl Rejection 85∼90%
5) In case of UE element : PF
 Specification of CSM
3
3-2. Representative CSM Installations

1. Boiler Feed Water Use

Customers Location Model No. Capacity Date Remarks

Kia Motors Asan RE8040-BN 400m3/hr Dec.1999

Dongbu Steel Dangjin RE8040-BE 400m3/hr Mar.2002

Veolia Daesan RE8040-BE 300m3/hr Jan.2003

SK Corporation Ulsan RE8040-BN 350m3/hr Oct.2002

Samsung General Chemical Daesan RE8040-FE 420m3/hr Jul.2003

SK Corporation Ulsan RE8040-BE 200m3/hr Oct.2002

POSCO Pohang RE8040-BN 60m3/hr Jan.1999

Hyundai Motors R&D

Namyang RE8040-BE 100m3/hr Mar.1998
Center
Saehan Industries Inc Gyungsan RE8040-BN 100m3/hr Apr.2000

Saehan Industries Inc Gumi RE8040-BN 150m3/hr Oct.2000

RE8040-FEN
Hyndai Steel Dangjin 300m3/hr Oct.2012
RE8040-FLR
Hyndai Hysco Dangjin RE8040-FE 210m3/hr May.2005

DWChem Pyung teak RE8040-FEN 580m3/hr Oct.2010

Hyndai Oil Bank Daesan RE8040-FD 650m3/hr Sep.2005

2. Ultra Pure Water/Pure Water Use

Customers Location Model No. Capacity Date Remarks
Samsung Electronics Onyang RE8040-BE 100m3/hr Oct.1997
Anam S&T Cheonan RE8040-BN 30m3/hr Sep.1996
Hankuk Electronics Kumi RE8040-BN 100m3/hr Jan.1998 Replace CA with PA
Hyundai Electronics Shanghai RE8040-BN 50m3/hr Oct.1997
LG Siltron Kumi RE8040-BE 120m3/hr Aug.2001 UPW 2nd pass
Samsung SDI Kachun RE8040-BN 420m3/hr Jul.1997
Samsung SDI CheonAn RE8040-HUE 40m3/hr Jul.2010
LG Dispay M5 Kumi RE8040-BE 570m3/hr May.2012 BLR 2nd pass
LG Display P7 Paju RE8040-UL 130m3/hr Feb.2010
Veolia Icheon RE8040-HUE 56m3/hr Feb.2012
 3 Specification of CSM

3. Brackwish water/ Overseas

Customers Location Model No. Capacity Date Remarks
3
Chesapeake Virginia RE8040-BE 310m /hr May.2012
Gloucester Virginia RE8040-BE 315m3/hr Aug.2009
Torrance California RE8040-BE 315m3/hr Mar.2012
3
Dumai Indonesia RE8040-BLF 93m /hr Jun.2010
Karabuk Turkey RE8040-BLF 430m3/hr May.2012
FAJR Iran RE8040-BR 3,900m3/hr Jul.2006

4. Process Water Use

Customers Location Model No. Capacity Date Remarks
Samyang Kasei Chunju RE8040-BE 200m3/hr Oct.1999
Samyang Kasei Chunju RE8040-BF 200m3/hr Aug.2001
Samsung Motor Busan RE8040-BE 220m3/hr May.2000

5. Power Plant Use

Customers Location Model No. Capacity Date Remarks
3
Busan Cogeneration Busan RE8040-BN 100m /hr Apr.1999
KEPCO Pyungtaek RE8040-BE 100m3/hr Jan.1998

6. Zero Discharge System Use

Customers Location Model No. Capacity Date Remarks

RE8040-BE
Doosan Electronics Jeungpyung 30m3/hr Jul.1997
RE4040-SN
Hyundai Motors Asan RE8040-BE 300m3/hr Sep.1998
Hyundai Motors Asan RE8040-BF 200m3/hr Apr.2001

Hyundai Motors Asan RE8040-SF 20m3/hr Jun.2001

7. Waste Water Recycle Use

Customers Location Model No. Capacity Date Remarks
LG Siltron Ichun RE8040-BE 30m3/hr Jul.1998

Samsung Electro-Mechanics Jochiwon RE8040-BE 120m3/hr May.1999

RE8040-BE
Capro Ulsan RE8040-SN 180m3/hr Aug.2002
RE8040-SR

Samnam Petrochemical Ulsan RE8040-BF 100m3/hr Aug.2002

LG Siltron Kumi RE8040-BN 120m3/hr May.1999

 Specification of CSM
3
DWchem Pyeng Taek RE8040-FEN 600m3/hr Jul.2009
Samsung Electro-Mechanics Hwaseong RE8040-FEN 60m3/hr Jun.2009

8. Waste Water Recycle Use / Overseas

Customers Location Model No. Capacity Date Remarks
Kranji Singapore RE8040-FE 286m3/hr Jun.2006

Bedok Singapore RE16040-BLR 2,318m3/hr Feb.2008

Wollongong Australia RE8040-FEN 950m3/hr Aug.2007

Brisbane Australia RE8040-FEN 500m3/hr Mar.2008

Scottsdale Arizona RE16040-FE 1,161m3/hr Jun.2012

West Basin Pillippine RE8040-FE 612m3/hr Jul.2008

EBMUD USA RE8040-FE 630m3/hr Dec.2008

Exxon Mobil USA RE8040-FE 156m3/hr Jul.2009

\
9. Sea Water Desalination Use
Customers Location Model No. Capacity Date Remarks

RE8040-SN
EAST PARS Iran 1,800m3/hr Aug.2000
RE8040-BE
Korea Navy Force Jinhae RE4040-SR 500pcs Jun.1998

RE8040-SN
KROSIS S. Korea RE4040-SN 1,000pcs/yr Plant Manufacturer
RE2540-SN
Korea Navy Force Jinhae RE4040-SR 255pcs May.2002

RE8040-SR
SOUTH PARS Iran 70m3/hr Aug.2002
RE8040-BE
Korea Navy Force Jinhae RE4040-SR 730pcs Aug.2002
3
Shinan-Gun Hongdo RE8040-SN 100m /hr May.1999
RE8040-SN
Woodo S. Korea
RE8040-SHN 15m3/hr Apr.2010

Chujado S.Korea RE8040-SHN 9m3/hr Apr.2009

Dokdo S.Korea RE4040-SR 15pcs Nov.2007

RE16040-SHF
GiJang S.Korea
RE16040-BLR 1,875m3/hr May.2014 BLR 2nd pass

Shuwaikh Kuwait RE8040-SHA 0.8m3/hr Dec.2011

 3 Specification of CSM

10. Potable Water with NF

Customers Location Model No. Capacity Date Remarks

Mapyeong Samcheok NE4040-90 3.2m3/day Feb.2010

Yeongwol Yeongwol NE4040-90 2.0m3/day Nov.2011

Cheonggye Anyang NE4040-90 6.3m3/hr Dec.2009

Siheung Ansan NE4040-90 13m3/hr Jul.2010

Hanrim Jeju NE4040-90 2.3m3/hr Jun.2012

 Specification of CSM
3
3-3. Brackish Water Elements

Brackish Water Elements

Parameter Unit
RE8040-BE RE8040-BE34 RE8040-CE RE4040-BE
GPD 11000 11000 12000 2400
Performance Permeate Flux m3/d 39.7 36 45.4 9.1
(Nominal)
Salt Rejection % 99.7 99.7 99.0 99.7
Operating Pressure psig 225 225 225 225
Temperature ℃ 25 25 25 25
Test NaCl Concentration ppm 2000 2000 2000 2000
Conditions
pH - 6.5∼7.0 6.5∼7.0 6.5~7.0 6.5∼7.0
Recovery % 15 15 15 15
Max. SDI(15min.) - 5 5 5 5
Max. Turbidity NTU 1 1 1 1
Max. Chlorine ppm 0.1 0.1 0.1 0.1
Concentration
3
Max. Feed Flow Rate m /hr 17 17 17 4.0
Operating Min. Concentration Flow m3/hr 3.6 3.6 3.6 0.91
Rate
Limits
Max. Pressure psig 15 15 15 15
Drop/Element
Max. Operating ℃ 45 45 45 45
Temperature
Normal 2 ~11 2 ~11 2 ~11 2 ~11
pH Range
Cleaning 1∼13 1∼13 1∼13 1∼13
ft2 400 365 400 85
Effective Area
m2 37.2 33.9 37.2 7.9
Element Length inch 40 40 40 40
Size Element Diameter inch 8.0 8.0 8.0 4.0
Permeate Tube Inner Dia. inch 1.125 1.125 1.125 -
Permeate Tube Outer Dia. inch - - - 0.75
Permeate Tube Protrusion inch 0 0 0 1.0
GPD 7180 6180 4078 1530
Surface Water
m3/d 27.3 23.5 15.5 5.8
Maximum GPD 9160 7890 5207 1950
Permeate Softened Water
m3/d 34.8 30.0 20.0 7.4
Flow Rate
GPD 11200 9630 6355 2370
RO/UF Permeate
m3/d 42.4 36.6 24.2 9.0
 3 Specification of CSM

3-4. Fouling Resistant Elements

Fouling Resistant Elements

Parameter Unit
RE8040-FEN RE8040-FLR RE4040-FEN RE4040-FLR
GPD 11000 10000 2400 2400
Performance Permeate Flux m3/d 41.6 37.9 9.1 9.1
(Norminal)
Salt Rejection % 99.7 99.6 99.7 99.7
Operating Pressure psig 150 150 225 225
Temperature ℃ 25 25 25 25
Test NaCl Concentration ppm 2000 1500 1500 2000
Conditions
pH - 6.5∼7.0 6.5∼7.0 6.5∼7.0 6.5∼7.0
Recovery % 15 15 15 15
Max. SDI(15min.) - 5 5 5 5
Max. Turbidity NTU 1 1 1 1
Max. Chlorine ppm 0.1 0.1 0.1 0.1
Concentration
Max. Feed Flow Rate m3/hr 17 17 4.0 4.0
Operating Min. Concentration Flow m3/hr 3.6 3.6 0.91 0.91
Rate
Limits
Max. Pressure psig 15 15 15 15
Drop/Element
Max. Operating ℃ 45 45 45 45
Temperature
Normal 2 ~11 2 ~11 2 ~11 2 ~11
pH Range
Cleaning 1∼13 1∼13 1∼13 1∼13
ft2 400 400 85 85
Effective Area
m2 37.2 37.2 7.9 7.9
Element Length inch 40 40 40 40
Size Element Diameter inch 8.0 8.0 4.0 4.0
Permeate Tube Inner Dia. inch 1.125 1.125 - -
Permeate Tube Outer Dia. inch - - 0.75 0.75
Permeate Tube Protrusion inch 0 0 1.0 1.0
GPD 6180 6180 1530 1530
Surface Water
m3/d 23.5 23.5 23.5 5.8
Maximum GPD 7890 7890 1950 1950
Permeate Softened Water
m3/d 30.0 30.0 7.4 7.4
Flow Rate
GPD 9630 9630 2370 2370
RO/UF Permeate
m3/d 36.6 36.6 9.0 9.0
 Specification of CSM
3
3-5. Low Pressure Elements

Low Pressure Elements

Parameter Unit
RE8040-BLF RE8040-BLN RE8040-BLR RE4040-BLF
GPD 11500 12000 10000 2500
Permeate Flux
Performance m3/d 43.5 45.4 37.9 9.5
(Norminal)
Salt Rejection % 99.2 99.2 99.6 99.2
Operating Pressure psig 100 150 150 100
Temperature ℃ 25 25 25 25
Test NaCl Concentration ppm 500 1500 1500 500
Conditions
pH - 6.5∼7.0 6.5∼7.0 6.5∼7.0 6.5∼7.0
Recovery % 15 15 15 15
Max. SDI(15min.) - 5 5 5 5
Max. Turbidity NTU 1 1 1 1
Max. Chlorine ppm 0.1 0.1 0.1 0.1
Concentration
Max. Feed Flow Rate m3/hr 17 17 17 4.0
Operating Min. Concentration Flow 3
m /hr 3.6 3.6 3.6 0.91
Rate
Limits
Max. Pressure psig 15 15 15 15
Drop/Element
Max. Operating ℃ 45 45 45 45
Temperature
Normal 2 ~11 2 ~11 2 ~11 2 ~11
pH Range
Cleaning 1∼13 1∼13 1∼13 1∼13
ft2 400 400 400 85
Effective Area
m2 37.2 37.2 37.2 7.9
Element Length inch 40 40 40 40
Size Element Diameter inch 8.0 8.0 8.0 4.0
Permeate Tube Inner Dia. inch 1.125 1.125 1.125 -
Permeate Tube Outer Dia. inch - - - 0.75
Permeate Tube Protrusion inch 0 0 0 1.0
GPD 8484 8484 8484 1808
Surface Water
m3/d 32.1 32.1 32.1 6.4
Maximum GPD 10822 10822 10822 2304
Permeate Softened Water
m3/d 40.9 40.9 40.9 6.7
Flow Rate
GPD 13216 13216 13216 2413
RO/UF Permeate
3
m /d 50.0 50.0 50.0 9.1
 3 Specification of CSM

3-6. Sea Water Elements

Sea Water element

Parameter Unit
RE8040-SHF RE8040-SHN RE4040-SHF RE4040-SHN

GPD 6500 1900 1200 1200

Performance Permeate Flux
m3/d 24.6 7.2 4.5 4.5
(Norminal)
Salt Rejection % 99.75 99.7 99.75 99.2
Operating Pressure psig 800 800 800 800
Temperature ℃ 25 25 25 25
Test NaCl Concentration ppm 32000 32000 32000 32000
Conditions
pH - 6.5∼7.0 6.5∼7.0 6.5∼7.0 6.5∼7.0
Recovery % 8 8 8 8
Max. SDI(15min.) - 5 5 5 5
Max. Turbidity NTU 1 1 1 1
Max. Chlorine ppm 0.1 0.1 0.1 0.1
Concentration
Max. Feed Flow Rate m3/hr 17 4.0 4.0 4.0
Operating Min. Concentration Flow m3/hr 3.6 0.91 0.91 0.91
Rate
Limits
Max. Pressure psig 15 15 15 15
Drop/Element
Max. Operating ℃ 45 45 45 45
Temperature
Normal 2 ~11 2 ~11 2 ~11 2 ~11
pH Range
Cleaning 1∼13 1∼13 1∼13 1∼13
ft2 400 75 75 75
Effective Area
2
m 34.4 6.9 6.9 6.9
Element Length inch 40 40 40 40
Size Element Diameter inch 8.0 4.0 4.0 4.0
Permeate Tube Inner inch 1.125 - - -
Dia.
Permeate Tube Outer inch 0.75 0.75
Dia. 0.75
Permeate Tube inch 0 1.0 1.0 1.0
Protrusion
GPD 5932 1100 1100 1100
Surface Water
m3/d 22.5 4.2 4.2 4.2
Maximum GPD 6790 1530 1530 1530
Permeate Softened Water
3
m /d 25.8 5.8 5.8 5.8
Flow Rate
GPD 8660 1950 1950 1950
RO/UF Permeate
3
m /d 32.9 7.4 7.4 7.4
 Specification of CSM
3
3-7. Nanofiltration(NF) Elements

Nanofiltration(NF) Elements
Parameter Unit
NE8040-90 NE8040-70 NE8040-40 NE4040-90
GPD 8000 7000 10000 1700
Permeate Flux
Performance m3/d 30.3 26.5 37.9 6.0
(Norminal)
Salt Rejection % 97(85∼95) 97(40~70) 97(40~70) 97(85∼90)
[MgSO4(NaCl)]
Operating Pressure psig 75 75 75 75
Temperature ℃ 25 25 25 25
Test MgSO4(NaCl) ppm 2000(2000) 2000(2000) 2000(2000) 2000(2000)
Concentration
Conditions
pH - 6.5∼7.0 6.5∼7.0 6.5∼7.0 6.5∼7.0
Recovery % 15 15 15 15
Max. SDI(15min.) - 5 5 5 5
Max. Turbidity NTU 1 1 1 1
Max. Chlorine ppm 0.1 0.1 0.1 0.1
Concentration
3
Max. Feed Flow Rate m /hr 15 15 15 4.0
Operating Min. Concentration Flow m3/hr 3.6 3.6 3.6 0.91
Rate
Limits
Max. Pressure psig 15 15 15 15
Drop/Element
Max. Operating ℃ 45 45 45 45
Temperature
Normal 2 ~11 2 ~11 2 ~11 2 ~11
pH Range
Cleaning 1∼13 1∼13 1∼13 1∼13
ft2 400 400 400 85
Effective Area
m2 37.2 37.2 37.2 7.9
Element Length inch 40 40 40 40
Size Element Diameter inch 8.0 8.0 8.0 4.0
Permeate Tube Inner Dia. inch 1.125 1.125 1.125 -
Permeate Tube Outer Dia. inch - - - 0.75
Permeate Tube Protrusion inch 0 0 0 1.0
GPD 6180 6180 6180 1340
Surface Water
m3/d 23.5 23.5 23.5 5.1
Maximum GPD 7890 7890 7890 1710
Permeate Softened Water
m3/d 30.0 30.0 30.0 6.5
Flow Rate
GPD 9630 9630 9630 2080
RO/UF Permeate
m3/d 36.6 36.6 36.6 7.9
 3 Specification of CSM

3-8. Tap Water Elements

Tap Water Elements

Parameter Unit
RE4040-TL RE2540-TL RE4021-TL RE2521-TL
GPD 2600 800 1050 350
Permeate Flux
Performance m3/d 9.8 3.0 4.0 1.3
(Norminal)
Salt Rejection % 99.2 99.2 99.2 99.2
Operating Pressure psig 150 150 150 150
Temperature ℃ 25 25 25 25
Test NaCl Concentration ppm 1500 1500 1500 1500
Conditions
pH - 6.5∼7.0 6.5∼7.0 6.5∼7.0 6.5∼7.0
Recovery % 15 15 15 15
Max. SDI(15min.) - 5 5 5 5
Max. Turbidity NTU 1 1 1 1
Max. Chlorine ppm 0.1 0.1 0.1 0.1
Concentration
Max. Feed Flow Rate m3/hr 4.0 1.3 4.0 1.3
Operating Min. Concentration Flow m3/hr 0.91 0.22 0.91 0.22
Rate
Limits
Max. Pressure psig 15 15 15 15
Drop/Element
Max. Operating ℃ 45 45 45 45
Temperature
Normal 2 ~11 2 ~11 2 ~11 2 ~11
pH Range
Cleaning 1∼13 1∼13 1∼13 1∼13
ft2 85 27 35 12
Effective Area
m2 7.9 2.5 3.3 1.1
Element Length inch 40 40 21 21
Size Element Diameter inch 4.0 2.5 4.0 2.5
Permeate Tube Inner Dia. inch - - - -
Permeate Tube Outer Dia. inch 0.75 0.75 0.75 0.75
Permeate Tube Protrusion inch 1.0 1.0 1.0 1.0
GPD 1808 574 744 255
Surface Water
m3/d 6.4 2.2 2.8 1.0
Maximum GPD 2304 732 949 325
Permeate Softened Water
m3/d 6.7 2.8 3.6 1.2
Flow Rate
GPD 2413 766 994 340
RO/UF Permeate
m3/d 9.1 2.9 3.8 1.3
 Specification of CSM
3
Tap Water Elements
Parameter Unit
RE4040-TE RE2540-TE RE4021-TE RE2521-TE
GPD 2400 800 1050 300
Permeate Flux
Performance m3/d 9.1 3.0 4.0 1.1
(Norminal)
Salt Rejection % 99.5 99.5 99.5 99.5
Operating Pressure psig 225 225 225 225
Temperature ℃ 25 25 25 25
Test NaCl Concentration ppm 2000 2000 2000 2000
Conditions
pH - 6.5∼7.0 6.5∼7.0 6.5∼7.0 6.5∼7.0
Recovery % 15 15 8 8
Max. SDI(15min.) - 5 5 5 5
Max. Turbidity NTU 1 1 1 1
Max. Chlorine ppm 0.1 0.1 0.1 0.1
Concentration
Max. Feed Flow Rate m3/hr 4.0 1.3 4.0 1.3
Operating Min. Concentration Flow m3/hr 0.91 0.22 0.91 0.22
Rate
Limits
Max. Pressure psig 15 15 15 15
Drop/Element
Max. Operating ℃ 45 45 45 45
Temperature
Normal 2 ~11 2 ~11 2 ~11 2 ~11
pH Range
Cleaning 1∼13 1∼13 1∼13 1∼13
ft2 85 27 35 12
Effective Area
m2 7.9 2.5 3.3 1.1
Element Length inch 40 40 21 21
Size Element Diameter inch 4.0 2.5 4.0 2.5
Permeate Tube Inner Dia. inch - - - -
Permeate Tube Outer Dia. inch 0.75 0.75 0.75 0.75
Permeate Tube Protrusion inch 1.0 1.0 1.0 1.0
GPD 1530 470 630 210
Surface Water
m3/d 5.8 1.8 2.4 0.8
Maximum GPD 1950 630 790 260
Permeate Softened Water
3
m /d 7.4 2.4 3.0 1.0
Flow Rate
GPD 2370 760 970 340
RO/UF Permeate
m3/d 9.0 2.9 3.7 1.3
 3 Specification of CSM

9. Ultrafiltration(UF) Elements
3-9. Ultrafiltration(UF) Elements

Ultrafiltration(UF) Elements
Parameter Unit
UE8040-ES10 UE8040-CEP UE4040-CEP
GPD 7200 6500 1400
Permeate Flux
Performance m3/d 27.2 24.6 5.3
(Norminal)
Salt Rejection 10k
Operating Pressure psig 20 14.5 14.5
Temperature ℃ 25 25 25
Test Feed Water - Pure water Pure water Pure water
Conditions
pH - 6.5∼7.0 6.5∼7.0 6.5∼7.0
Recovery %
Max. SDI(15min.) -
Max. Turbidity NTU
Max. Chlorine ppm
Concentration
Max. Feed Flow Rate m3/hr 17 17 4.0
Operating Min. Concentration Flow m3/hr 3.6 3.6 0.91
Rate
Limits
Max. Pressure psig 15 15 15
Drop/Element
Max. Operating ℃ 45 45 45
Temperature
Normal
pH Range
Cleaning
ft2 400 400 85
Effective Area
m2 37.2 37.2 7.9
Element Length inch 40 40 40
Size Element Diameter inch 8.0 8.0 4.0
Permeate Tube Inner Dia. inch 1.125 1.125 -
Permeate Tube Outer Dia. inch - - 0.75
Permeate Tube Protrusion inch 0 0 1.0
GPD 12653 12653 2600
Surface Water
m3/d 47.9 47.9 9.8
Maximum GPD 17700 17700 3000
Permeate Softened Water
m3/d 67.0 67.0 11.4
Flow Rate
GPD 19900 19900 3360
RO/UF Permeate
m3/d 75.3 75.3 12.7
 Specification of CSM
3
◎ 가정용 역삼투 분리막
Test Conditions Size(mm)
Flux Salt Rej'
Model No. Pressure Temp. Feed Conc. Recovery
(GPD) (%) A B C D E
(psig) (℃) (ppm) (%)
RE1810-30 30 98 60 25 200 15 45.0±0.5 256 25 14 17

RE1810-50 50 98 60 25 200 15 45.0±0.5 256 25 14 17

RE1810-80 80 98 60 25 200 15 45.0±0.5 298 22 22 17

RE1812-50 50 98 60 25 200 15 48.5±0.5 298 23 12 17

RE1812-60 60 98 60 25 200 15 48.5±0.5 298 23 12 17

RE1812-80 80 98 60 25 200 15 48.5±0.5 298 23 12 17

RE2012-LP 50 93 20 25 100 15 50.0±0.5 298 23 12 17

RE2012-LPF 60 93 20 25 100 15 50.0±0.5 298 23 12 17

■ Operating Limits
· Maximum Operating Pressure 125psig
· Maximum Feed Flow Rate 2gpm
· Maximum Operating Temperature 45℃
· Feed Water pH Range 3.0~10.0
· Maximum Turbidity 1.0NTU
· Maximum Silt Density Index 5.0
· Free Chlorine Tolerance <0.1ppm

※ Consult CSM for other operating conditions.

Brine Seal O-Ring

Outer Taping

Concentrate

Feed E Permeate

Concentrate

A C
D
B
 Water Chemistry and Pretreatment
4
4-1. Introduction

The efficiency and lifetime of a reverse osmosis (RO) process depends on effective
pretreatment of the feed water. The pretreatment includes any process which can minimize fouling,
scaling and membrane degradation to maintain product flow, salt rejection, product recovery and
economical operating costs.
Fouling is the entrapment of particulates such as inorganic and organic colloids. Examples of
the inorganic colloids are iron floc, silica, clay and silt. The organic colloids are mostly consisted
of organic polymers and microorganisms. Scaling is the precipitation and deposition within the
RO system of sparingly soluble salts such as calcium carbonate (CaCO3), calcium sulfate (CaSO4)
and barium sulfate (BaSO4).
A proper pretreatment scheme for the feed water mainly depend on feed water source, feed
water composition and application. There are three types of feed water. Well water contains a
low Silt Density Index (SDI) (typically < 2) and low bacteria count, thus requiring a simple
pretreatment scheme. Surface water, on the other hand, is characterized by a high SDI and can
have a high bacteria count. Pretreatment for surface water is more elaborate than for well water,
requiring the addition of a coagulant, clarification, and multimedia filtration.
Once the feed water source has been determined, an accurate analysis of the feed water
composition including bacteria count should be conducted to determine the dosage size of a
coagulant, a scale inhibitor (antiscalant) and a biocide for the disinfection.
 4 Water Chemistry and Pretreatment

4-2. Feed Water Analysis

A complete and accurate feed water analysis must be performed prior to RO system design.
The water analysis report should contain the type and the concentration of all constituents in
the water. The constituents are consisted of dissolved ions, silica, colloids, and organic (TOC)
■ Typical dissolved anions are as follows :
bicarbonate(HCO3-), carbonate(CO32-), hydroxide(OH-), sulfate(SO42-), chloride(Cl-),
fluoride(F-), nitrate(NO3-), sulfide(S2-) and phosphate(PO42-)
■ Typical dissolved cations are shown below :
calcium(Ca2+), magnesium(Mg2+), sodium(Na+), potassium(K+), iron(Fe2+ or Fe3+),
manganese(Mn2+), aluminum(Al3+), barium(Ba2+), strontium(Sr2+), copper(Cu2+)
and zinc (Zn2+).
Certain combinations of cations and anions form sparingly soluble salts in water and scaling of
RO membrane may occur when the salts are concentrated within the RO element beyond their
solubility limit. Typical sparingly soluble salts and their solubility product limit are shown in
Table 1.

Substance Formula Temp.(℃) Solubility Product

Aluminum Hydroxide Al(OH)3 20 1.9×10-33
Barium Carbonate BaCO3 16 7×10-9
Barium Sulfate BaSO4 25 1.08×10-10
Calcium Carbonate CaCO3 25 8.7×10-9
Calcium Fluoride CaF2 26 3.95×10-11
Calcium Sulfate CaSO4 10 6.1×10-5
Cupric Sulfide CuS 18 3.5×10-45
Ferric Hydroxide Fe(OH)3 18 1.1×10-36
Ferrous Hydroxide Fe(OH)2 18 1.64×10-14
Magnesium Ammonium MgNH4PO4 25 2.5×10-13
Phosphate
Magnesium Carbonate MgCO3 12 2.6×10-5
Magnesium Hydroxide Mg(OH)2 18 1.2×10-11
Manganese Hydroxide Mn(OH)2 18 4×10-14
Strontium Carbonate SrCO3 25 1.6×10-9
Strontium Sulfate SrSO4 17.4 2.81×10-7
Zinc Hydroxide Zn(OH)2 20 1.8×10-14
Table 1 : Solubility Products of Sparingly Soluble Inorganic Compounds

In the RO system the most common sparingly soluble salts encountered are CaSO4, CaCO3 and
silica. Other salts creating a potential scaling problem are CaF2, BaSO4 and SrSO4, though less

Global Initiative 21
 Water Chemistry and Pretreatment
4
prevalent. Other ions causing problems are described below. Sulfates are present in relatively large
concentrations in most raw waters. Their concentration can be artificially increased when sulfuric
acid is added to water to adjust pH. In this case, Ba2+ and Sr2+ must be analyzed accurately at
1㎍/ℓ(ppb) and 1㎎/ℓ(ppm) level of detection, respectively, since BaSO4 and SrSO4 are much less
soluble in water than CaSO4 and moreover, barium and strontium sulfate scales are extremely
difficult to redissolve.

Alkalinity consists of negative ions which include bicarbonate, carbonate and hydroxide. Most
of the alkalinity in naturally occurring water sources is in the form of bicarbonate alkalinity
(HCO3-). Below a pH of 8.3, the bicarbonate alkalinity will be in equilibrium with a certain
concentration of dissolved carbon dioxide. At a pH greater than 8.3, HCO3- will be converted to
the carbonate form (CO32-). With water sources of pH above 11.3, hydroxide (OH-) will be
present.
Water can dissolve carbon dioxide from the air, forming carbonic acid (H2CO3). The acidic
water will tend to dissolve calcium carbonate from the ground as it passes over or through the
calcium carbonate rock. Most naturally occurring water sources are close to saturation in calcium
carbonate which is in equilibrium with calcium bicarbonate, depending on the pH of the water.
Calcium bicarbonate is much more soluble in water than calcium carbonate. If the water is
concentrated in the RO system, calcium carbonate salt is likely to precipitate in the system.
Thus the use of a scale inhibitor or lowering the pH below 8 by an acid injection is required in
most RO systems.

Nitrates are very soluble in water and thus will not precipitate in an RO system, Nitrates are a
health concern since, when ingested by mammals including humans, they are converted to nitrites
which interfere with hemoglobins to exchange oxygen in blood. This can cause serious problems
especially for fetus and children. For this reason, it is desirable to maintain a nitrate concentration
below 40 ㎎/ℓ in drinking water. Typical nitrate removal rate by RO is in the range of 90-96%.

Iron and manganese are present in water either in a divalent state, which is soluble in water,
or in a trivalent state, which forms insoluble hydroxides. The soluble iron (Fe2+) can come from
either a well water or the rust of pump, piping and tanks, especially if acid is injected upstream of
the equipment. If the iron or manganese concentration is greater than 0.05 ㎎/ℓ in RO feed
water and they are oxidized by air or an oxidizing agent to the trivalent state, then the insoluble
hydroxides Fe(OH)3 and Mn(OH)3 will precipitate in the system, when the water pH is neutral or
higher. They can also catalyze the oxidative effects of residual oxidizing agents, possibly
accelerating the membrane degradation. Thus iron and manganese must be removed at the
pretreatment step (see the pretreatment section).
 4 Water Chemistry and Pretreatment

Aluminum is usually not noticeably present in naturally occurring water sources. With its
valence of 3+ like iron (Fe3+), aluminum will form very insoluble hydroxide [Al(OH)3] at the
normal operating pH range of 5.3 to 8.5 in an RO system. Because of the high charge
characteristics, alum [Al2(SO4)3] or sodium aluminate (NaAlO2) is used to coagulate negatively
charged colloids in the pretreatment of surface waters. Care must be taken not to employ the
coagulant excessively to result in carrying the residual aluminum over to the membranes.
A concentration of aluminum greater than 0.01mg/l in the dialysis water is a health concern for
kidney dialysis patients. In this regard, iron [FeCl3 or Fe2(SO4)3] may be preferred as a coagulant.

Copper and Zinc are not appreciably detected in natural water sources. Sometimes, it is
possible to pick up trace amounts from piping materials. Their hydroxides Cu(OH) 2 and Zn(OH)2)
will drop out of solution over the operating pH range of 5.3 to 8.5. Because of the low
concentrations of copper and zinc, their precipitants will foul the RO system only if allowed to
precipitate over an extended period of time without cleaning the system. However, more serious
situation may develop, when an oxidizing agent such as hydrogen peroxide is present together
with copper or zinc, to degrade the membranes rapidly.

Sulfides are present as a dissolved gas, hydrogen sulfide (H2S). Hydrogen sulfides oxidizes and
releases elemental sulfur which can cause irreversible fouling of RO membranes, Hydrogen
sulfide gas can be removed by running the water through a degasifier. Oxidizing sulfides by
chlorine or air to insoluble elemental sulfur followed by media filtration also can be used to
address the sulfides issue.

Phosphates have a strongly negative charge (3-) and a tendency to react with multivalent
cations such as Ca2+, Mg2+ ,Fe2+ , Fe3+ to give insoluble salts. Calcium phosphate has a very
limited solubility at neutral pH and an even lower solubility at higher pH. The use of a scale
inhibitor or lowering the pH of the feed water below 7 is a measure to control the phosphate
precipitate.

Silica is naturally present in most feed waters in the range of 1-100㎎/ℓ and exists mostly in the
silicic acid form [Si(OH)4] below a pH of 9. At low pH, Silicic acid can polymerize to form a
colloid (colloidal silica). At high pH above 9, it dissociates into the silicate anion (SiO 32-) and can
precipitate as a salt with calcium, magnesium, iron or aluminum. Silica and silicates are difficult
to redissolve. Ammonium bifluoride solutions are somewhat successful at cleaning silica scaling.
However, ammonium bifluoride is considered a hazardous chemical posing problems for disposal.
Silica present in an RO feed water at a concentration greater than 20㎎/ℓ may pose a potential for

Global Initiative 21
 Water Chemistry and Pretreatment
4
silica scaling.
Silt density index (SDI), also known as the Fouling Index (FI), is a good indicative to
determine the colloidal fouling potential of RO feed water. The source of colloids in RO feed
waters is varied and often includes bacteria, clay, colloidal silica and iron corrosion products.
Pretreatment chemicals used in a clarifier such as alum, ferric chloride or cationic
polyelectrolytes can also cause colloidal fouling if not removed in the clarifier or through proper
media filtration.
SDI measurement should be done prior to designing the RO pretreatment system and on a regular
basis during RO operation (at least once a day for surface waters). The test measures the rate of
fouling of a 0.45㎛ membrane filter by the following procedure.
Place the membrane filter (47㎜, 0.45㎛) on its support, bleed water pressure on carefully, tighten
the O-ring seal and fix the support vertically.
Adjust feed pressure to 2.1bar (30psi) and measure initial time t0, necessary to filter 500㎖ of
sample water (feed pressure to be kept constant by continuous adjustment).
Continue filtering the water for 15 min at 2.1bar (30psi). If the filter is plugged up in 15min, use
the 10 or 5min measurement. After 15 minutes, measure again time, t 1, necessary to filter
additional 500㎖ sample water.

SDI is calculated as follows ;

(1-t0/t1)
SDI = 100 ×
T

T is the time before the second flow measurement (5,10 or 15 minutes).

The guideline is to maintain SDI at less than or equal to 5 for RO feed water.

Turbidity is another guideline as an indicator for the rate of RO membrane fouling.

Turbidometers (also called nephelometers) measure the scattering of light caused by various
suspended solids in the water sample. Water samples having turbidity value greater than 1NTU
(Nephelometric Turbidity Units) regarded as high fouling potential feed water for RO
membranes. Like the SDI test, turbidity is only an indicator of fouling potential. High turbidity
does not necessarily mean that the suspended solid is going to deposit on the RO membrane.
In fact, there are some membrane foulants such as surfactants and soluble polymers that do not
scatter light to register a turbidity reading. Although turbidity and SDI are not perfect in
predicting the behavior of the colloidal foulants, they are useful for characterizing the RO feed
water. For examples, SDI greater than 5 and NTU greater than 1 strongly suggest that some
coagulants should be used in the clarifier step of the feed water pretreatment, followed by media
 4 Water Chemistry and Pretreatment

filtration. The feed water with SDI less than 5 and NTU less than 1 may just require media
filtration or cartridge filters without the necessity of coagulation of the colloids.
It is also necessary to have some guidelines for controlling the amount of the coagulant added
to the clarifier, since excessive addition of the coagulant should be avoided. Two feasible
guidelines are zeta potential measurement and streaming current detector.

Zeta Potential is a measurement of the overall charge characteristics of the suspended solids in
the water, when the water containing the charged colloids flows in one direction between the
oppositely charged electrodes. Most colloids in natural water sources have negative charges
which help to repel each other to keep them suspended in solution. In overall, the colloidal water
will show negative zeta potential. These negative charges can be neutralized by the addition of
cationic coagulants such as aluminum sulfate (alum) and ferric chloride. The coagulants are
portionwise added until zeta potential reached to zero (neutral). The colloids without charges do
not repel each other and are more likely to coagulate into larger particulate groups which can be
easily filtered out by media filters.

Streaming Current Detector uses a mechanical plunger to create a high water velocity, which
causes movement of the ions surrounding negatively charged suspended colloids. It measures the
eletrical current generated by the moving ions. If the charge of the colloids has been neutralized
by the addition of a coagulant, there will be very little current generated by the streaming current
detector.

The last constituents in the feed water to be analyzed are bacteria count and organic compounds.
There are two ways to count bacteria in the water. One method is to collect bacteria by filtering a
measured quantity of water through a membrane filter followed by culturing the retained
organisms and counting the developed colonies under low power magnification. The second
method is to count the retained microorganisms on the filter plate directly under a fluorescent
microscope after staining the microorganisms with dyes such as acridine orange. Direct count
methods should be preferred, because they are much faster and more accurate than culture
techniques.
Examples of organic compounds in the feed water are oils, surfactants, water soluble polymers,
and humic acid. The organic compounds are collectively analyzed by Total Organic Carbon
(TOC), Biological Oxygen Demand (BOD) and Chemical Oxygen Demand (COD). Identification
of individual organic substance may require more elaborate analytical tools such as
chromatography (HPLC) and GC-Mass spectrometry. Removal of the organic compounds from
the feed water at the pretreatment step should be considered when TOC exceeds 3㎎/ℓ.

Global Initiative 21
 Water Chemistry and Pretreatment
4
4-3. Prevention of Scale Formation

Scaling of an RO membrane may occur when sparingly soluble salts are concentrated in the RO
element beyond their solubility limit. Sparingly soluble salts are listed below in the order of
decreasing scaling problem:

CaCO3 > CaSO4 > Silica > SrCO3 > BaSO4 > SrSO4 > CaF2 > CaSiO3 > MgSiO3 > MgSiO3 >
Ca3(PO4)2 > Fe(OH)2

Calcium sulfate (CaSO4) is more soluble than BaSO4 and SrSO4. However, calcium ion (Ca2+)
is present in natural water sources more abundantly than Ba2+ and Sr2+ and thus CaSO4 will cause
more scaling problem than BaSO4 and SrSO4. On the other hand, BaSO4 and SrSO4 are difficult
to redissolve once precipitated. Hence, scaling of the two salts should be avoided.
The most frequent scaling problems come from calcium carbonate (CaCO3) because it
precipitates fast, once concentrated beyond its solubility limit and also most natural waters are
almost saturated with respect to CaCO3. CaCO3 scaling including SrCO3 and BaCO3 can be
prevented by acid addition, a scale inhibitor, softening of the feed water, preventive cleaning and
low system recovery.
CaSO4 scaling including BaSO4, SrSO4 and CaF2 is preventable by the same methods as
CaCO3 scaling except the acid addition. In fact, using sulfuric acid to lower pH for the prevention
of CaCO3 scaling would increase the probability of the sulfate scaling.

Acid Addition
The solubility of CaCO3 depends on the pH as shown in the following equation.

CaCO3 + H Ca2+ + HCO3-

The equilibrium can be shifted to the right side to convert CaCO3 to soluble Ca(HCO3)2 by
adding an acid (lowering pH). The acid used should be of food grade quality. Sulfuric acid is
commonly employed, but hydrochloric acid is preferred in the case of high scaling potential due
to CaSO4, SrSO4 and BaSO4 were also considered.
In order to avoid calcium carbonate scaling, the pH of the concentrate stream in an RO system
should be lower than the pH of saturation (pHs) where the concentrate stream is in equilibrium
with CaCO3. This relationship is expressed by the Langelier Saturation Index (LSI) for brackish
waters and Stiff & Davis Saturation Index (S&DSI) for sea waters.

LSI = pH - pHS (TDS < 10,000㎎/ℓ)

 4 Water Chemistry and Pretreatment

pHs = (9.3 + A + B) - (C + D)

where ; A = (Log10 [TDS] - 1) / 10

B = -13.12 × Log10 (℃ + 273) + 34.55
C = Log10 [Ca2+] – 0.4, where [Ca2+] is in ppm as CaCO3
D = Log10 [alkalinity], where [alkalinity] is in ppm as CaCO3

(Reference : ASTM D3739-88 Calculation and Adjustment of the Langelier

Saturation Index for Reverse Osmosis)

S & DSI = pH - pCa - pAlk - K

Where : pCa is the negative log of the calcium concentration in moles/ℓ.

pAlk is the negative log of the total alkalinity concentration in moles/ℓ.
K is a constant which is a function of water temperature and ionic strength.

(Reference : ASTM D4582-86 Calculation and Adjustment of the Stiff and Davis Saturation
Index for Reverse Osmosis)

Concentration factor is used to calculate the concentration of the constituents in the

concentrate stream from that in the feed water.

1
Concentration factor =
1 - recovery

Where recovery = permeate flow rate / feed flow rate

In reality, concentration polarization should be taken into account to get more accurate scaling
potential as shown below.
concentration polarization
Concentration factor =
1 - recovery
The concentration polarization value depends on the turbulence of the bulk stream in the RO
element and varies from 1.13 to 1.2, meaning that the concentration of salts at the membrane
surface is 13% to 20% higher than in the bulk stream.
To control calcium carbonate scaling by acid addition alone, the LSI or S & DSI in the
Global Initiative 21
 Water Chemistry and Pretreatment
4
concentrate stream must be negative. A rule of thumb is to lower the feed water pH to 6.0. If a
high quality scale inhibitor is used, the LSI in the concentrate stream can be as high as 1.8 (refer
to the inhibitor manufacturer's literature for reference points). This will reduce or eliminate the
acid consumption, and also could decrease the potential for corrosion due to the acid addition.

Scale Inhibitor Addition

Scale inhibitors (antiscalants) slow the precipitation process of sparingly soluble salts by being
absorbed on the forming salt crystals to prevent the attraction of the supersaturated salt to the
crystal surfaces. In this situation the crystals never grow to a size or concentration sufficient to
fall out of suspension. Furthermore, many scale inhibitors have some dispersive qualities which
involve surrounding particles of suspended salt or organic solids with the anionically charged
scale inhibitor. Now the anionically charged particles will repel each other to prevent the
agglomeration of the particles to larger particles that may precipitate.
Scale inhibitors effective in controlling carbonate scaling, sulfate scaling and calcium fluoride
scaling are listed below.

Sodiumhexametaphostphate(SHMP) is most widely used because it offered good inhibition

at a low cost. However, care must be taken in order to avoid hydrolysis of SHMP in the dosing
feed tank (a fresh solution should be made every 3 days). Hydrolysis would not only decrease the
scale inhibition efficiency, but also create a calcium phosphate scaling risk.
SHMP should be dosed to give a concentration in the concentrate stream of 20㎎/ℓ. The dosage
into the feed stream can be calculated by the equation :
SHMP dosing rate = 20mg × (1 - Y).
Where Y is a fraction of recovery

Organophosphonates are more resistant to hydrolysis compared to SHMP even though they
are more expensive. They offer scale inhibition and dispersion ability similar to SHMP.

Polyacrylic acids (PAA) are good at both scale inhibition and dispersion. The usual molecular
weight of PAA is 2000 to 5000. PAA with higher molecular weight distribution in the range of
6000 to 25000 showed the best dispersion ability at the sacrifice of scale inhibition ability. In
general, PAA are more effective than SHMP. However, precipitation reactions may occur with
cationic polyelectrolytes or multivalent cations such as aluminum or iron to foul the membrane.

Blend Inhibitors are a combination of low and high molecular weight of PAA or a blend of
low molecular weight PAA and organophosphonates for excellent dispersive and inhibitor
performance.
 4 Water Chemistry and Pretreatment

There are many manufacturers of scale inhibitors such as BF Goodrich, Arrowhead, Betz,
Grace Dearborn, Calgon, FMC, Degremont, Nalco, King-Lee and Maxwell Chemicals, etc.
(antiscalant manufacturer list re-check needed) Please, consult with them for the chemical
identity of the scale inhibitor brands and their compatibility with RO membranes. They can
usually be considered compatible with polyamide RO membranes up to 50 ppm concentration in
the concentrate.
Additionally, RO permeate should be used when diluting the scale inhibitors, since calcium
present in the feed water may form a precipitate with the scale inhibitors at high concentrations.
Precautions must also be taken so that there is no microbial growth in the inhibitor dilution tank.
Make certain that no significant amounts of cationic polymers are present when adding an anionic
scale inhibitor.

Softening with a Strong Acid Cation Exchange Resin

The scale forming cations such as Ca2+, Ba2+, Sr2+, Mg2+, and Fe2+ can be removed by a cation
exchange resin which could be a good alternative way to prevent scale formation for small or
medium size brackish water RO systems. The resin has to be regenerated with NaCl at the
saturation point. A drawback of this process is its relatively high sodium chloride consumption,
which might cause an environmental (disposing the brine) or an economic problem.

Softening with a Weak Acid Cation Exchange Resin

The weak acid cation exchange resin can remove only Ca2+, Ba2+ and Sr2+ linked to bicarbonate
and release H+, thus lowering the pH to a minimum value of 4.2 where the carboxylic acid groups
of the resin are no longer dissociated. Therefore, it is only a partial softening and ideal for waters
with high bicarbonate content.
The advantages of softening with a weak acid cation exchange resin are :
· Almost stoichiometrical consumption (105%) of acid for regeneration to minimize the
operating costs and the environmental impact
· Reduction of the TDS value of the water by the removal of bicarbonate salts to result in
the lower permeate TDS.
The disadvantages are :
· Residual hardness due to incomplete softening
· Variable pH of the treated water from 3.5 to 6.5 depending on the degree of
exhaustion of the resin
At pH < 4.2, the passage of mineral acid may increase the permeate TDS. It is therefore
recommended to use more than one resin column in parallel and to regenerate them at different
times in order to level out the pH. Other possibilities to avoid extremely low pH values are CO2
removal or pH adjustment by NaOH.

Global Initiative 21
 Water Chemistry and Pretreatment
4
Lime Softening
Lime [Ca(OH)2] reacts with soluble calcium or magnesium bicarbonate to remove carbonate
hardness as shown in the following equations ;

Ca(HCO3)2 + Ca(OH)2 → 2CaCO3↓ + 2H2O

Mg(HCO3)2 + 2Ca(OH)2 → Mg(OH)2 + 2CaCO3↓ + 2H2O

The non-carbonate calcium hardness can be further reduced by adding sodium carbonate (soda
ash) :

CaCl2 + Na2CO3 → 2NaCl + CaCO3

The lime-soda ash process can also reduce the silica concentration. When sodium aluminate
and ferric chloride are added as coagulants, the lime-soda ash will precipitate calcium carbonate
and a complex consisted of calcium, aluminum and iron silicate. Silica can be reduced to 1㎎/ℓ
by adding a mixture of lime and porous magnesium oxide. The lime softening can also reduce
barium, strontium and organic substances such as humic acid significantly. The effluent from this
process needs media filtration and pH adjustment prior to the RO elements. Lime softening
should be considered for brackish water plants larger than 200m3/h (1.2 million gpd).

Preventive Cleaning
In some applications such as small systems, preventive membrane cleaning allows the system
to run without dosage of acids or scale inhibitors or softening. Typically, those systems operate at
low recovery in the range of 25% and the membrane elements are replaced after 1-2 years. The
simplest way of cleaning is a forward flush at low pressure by opening the concentrate valve.
Short cleaning intervals are more effective than long cleaning times, e.g. 30seconds every 30
minutes. Cleaning can also be performed with cleaning chemicals, which may be done
occasionally between the short cleaning intervals.

Adjustment of System Recovery, pH and Temperature

The precipitation of dissolved salts can be avoided by keeping their concentration in the
concentrate stream below the solubility limit, which can be achieved by reducing the system
recovery, raising temperature, and adjusting pH.
To be more quantitative for the above operation, solubility product (limit) for each sparingly
soluble solute should be calculated under conditions in the concentrate stream as discussed in the
following sections. Silica is usually the only reason for adjusting the above three operating
 4 Water Chemistry and Pretreatment

variables as a scale control method, since these adjustments have high energy consumption due to
low system recovery or other scaling risks such as CaCO3 scaling at high pH. For small systems,
a low recovery combined with a preventive cleaning program might be a convenient way to
control scaling.

Global Initiative 21
 Water Chemistry and Pretreatment
4
4-4. Solubility Product Calculations

To determine the scaling potential, the ion product IPc of a sparingly soluble salt in the
concentrate stream should be compared with the solubility product Ksp of the salt under conditions in
the concentrate stream (Ksp is a function of temperature and ionic strength). If IPc < Ksp, no scale
control is necessary.
The concentration of ion species in the concentrate stream is usually not known unless
measured experimentally, but can easily be estimated from the concentration in the feed stream
by multiplication with the concentration factor CF= 1 / ( 1 - r ). Where r is recovery ratio
(expressed as a decimal).

The ionic strength of the feed water is :

If = ½Σ(mi ×Zi2 )
Where mi = molar concentration of ion i (mol/kg)
Zi = ionic charge of ion i

Where the water analysis is not given in molar concentrations, the conversion is as follows :

1000Ci
mi =
MWi
Where Ci = concentration of ion i in ㎎/ℓ
MWi = molecular weight of ion i

The ionic strength Ic of the concentrate stream is obtained from :

1
Ic = If ×
(1 - r)
With the ionic strength of the concentrate stream, the solubility product K sp of the sparingly
soluble salt can be obtained.
To make sure that scaling will not occur, the IPc for CaSO4, BaSO4, SrSO4 and CaF2 should be
less than 0.8 Ksp of the corresponding salts, respectively. If IPc > 0.8Ksp, one of the scale
preventing methods discussed in the previous section must be applied.

If proper scale inhibitors are used, IPc could be greater than Ksp as shown in the following
equation.
 4 Water Chemistry and Pretreatment

IPc ≤ 2.0 Ksp for CaSO4 if PAA or organophosphonates are employed

IPc ≤ 1.5 Ksp for CaSO4 if SHMP is used
IPc ≤ 50 Ksp for BaSO4
IPc ≤ 10 Ksp for SrSO4
IPc ≤ 100 Ksp for CaF2

Barium sulfate is the most insoluble of all alkaline-earth sulfates. In most natural waters,
barium is present at a level close to precipitation in the concentrate stream. The critical feed
concentration of BaSO4 may be as low as 15㎍/ℓ in sea waters, 5㎍/ℓin brackish waters or even
2㎍/ℓ if sulfuric acid is added to brackish waters.

Global Initiative 21
 Water Chemistry and Pretreatment
4
4-5. Silica Scale Prevention

In addition to BaSO4 scaling, silica scale is also difficult to redissolve. Thus silica scaling has
to be prevented. The presence of Al3+ and Fe3+ complicates the silica scaling via formation of
insoluble aluminum and iron silicates. Therefore, if a silica scaling potential exists, aluminum and
iron must be removed by 1㎛ cartridge filtration and preventive acid cleanings.
The calculation of the silica scaling potential requires the following data of the feed solution :
SiO2 concentration, temperature, pH and total alkalinity.
The SiO2 concentration in the concentrate stream is calculated from the SiO2 concentration in
the feed solution and the recovery of the RO system :

1
SiO2c = SiO2f ×
1 r

Where : SiO2c = silica concentration in concentrate as SiO2 in ㎎/ℓ

SiO2f = silica concentration in feed as SiO2 in ㎎/ℓ
r = recovery of the RO system expressed as a decimal

Calculate the pH of the concentrate stream from the pH of the feed stream using the
following equation.

pH = Log10 ([alkalinity as CaCO3] / [CO2]) + 6.3

Alkalinity of feed
Alkalinity of concentrate =
1-r
(need to revised same as SiO2C calculation equation

Obtain the solubility of SiO2 as a function of temperature. Temperature of the concentrate is

assumed equal to temperature of feed solution.

Example are shown in the following table :

 4 Water Chemistry and Pretreatment

T(°C) Solubility of SiO2 (mg/ℓ)

5 85
10 96
15 106
20 118
25 128
30 138
35 148

Obtain the pH correction factor for the concentrate pH

Since the solubility of silica increases below a pH of about 7.0 and above a pH of about 7.8, the
actual solubility of SiO2 in the concentrate stream can be further affected by the pH of the
concentrate stream and thus is obtained by multiplying the solubility of SiO2 at a specific
temperature by the pH correction factor to give the corrected solubility (SiO2cor).
For examples, pH correction factor are 1.0 at pH 7.8 and 1.5 at pH 8.5, respectively. See
ASTM D4993-89 for more details. Compare the silica concentration in the concentrate (SiO2c) of
the RO system with the pH corrected silica solubility (SiO2cor). If SiO2c is greater than SiO2cor,
silica scaling can occur and adjustment is required.
The easiest way to prevent the silica scaling is to lower recovery. Reiteration of the calculations
can be used to optimize recovery with respect to silica scaling, once a reverse osmosis system is
operating.
Lime plus soda ash softening can be used in the pretreatment system to decease the SiO2
concentration in the feed stream. The maximum allowable recovery against silica scaling can be
increased significantly by increasing the water temperature using a heat exchanger.

A dispersant such as high molecular weight polyacrylate scale inhibitor is helpful in silica scale
control by slowing agglomeration of scale particulate.

Silica solubility as a function of pH (pH correction factor) need to be included..!

Global Initiative 21
 Water Chemistry and Pretreatment
4
4-6. Colloidal Fouling Control

The removal of suspended and colloidal particles can be done by media filtration, crossflow
microfiltration and cartridge microfiltration only for the raw waters with an SDI slightly above 5.
For raw waters containing high concentrations of colloidal matter showing SDI well above 5, the
coagulation and flocculation process is necessary before media filtration.
Ferric sulfate or ferric chloride is usually used as a coagulant to destabilize the negative surface
charge of the colloids to result in coagulation and further to entrap them into the freshly formed
ferric hydroxide microflocs. Aluminum coagulants are also effective, but not recommended
because of possible fouling problems with residual aluminum. Cationic polymers may be used as
both coagulants and flocculants.
To further agglomerate the hydroxide microflocs for better filterability, flocculants can be used
in combination with coagulants. Flocculants are soluble high molecular weight polymers such as
polyacrylamides which may contain cationic, anionic, or neutral active groups.
The hydroxide flocs are allowed to grow and to settle in specifically designed reaction
chambers or clarifiers. The hydroxide sludge is removed, and the supernatant water is further
treated by media filtration.
Care must be taken not to allow coagulants and flocculants or the hydroxide flocs to escape
from media filters to reach the RO membranes. Furthermore, reaction of the residual coagulants
and flocculants with a scale inhibitor added after the media filter can cause a precipitate to form,
which is very hard to be cleaned. Several RO plants have been heavily fouled by a gel formed
from cationic polyelectrolytes and scale inhibitors.

Media filtration
Media filters use a filtration bed consisting of one or more layers of media granules which are
sand and anthracite. The grain size for fine sand filter is in the range of 0.35 to 0.5㎜ or 8×12
mesh to 60 mesh and for anthracite filter 0.6 to 0.9 ㎜.
A multimedia filter is designed to make better use of the bed depth in the removal of a greater
volume of suspended solids. This is achieved by loading larger (irregular shaped) media granules
of lower density such as anthracite over smaller media granules of higher density such as sand.
The larger granules at the top remove the larger suspended particles, leaving the smaller particles
to be filtered by the finer media, thus to result in more efficient filtration and longer runs between
cleaning.
The design depth of the filter media is normally about 0.8m (31 inches) minimum with a 50%
freeboard for the media expansion during backwashing. The multimedia filters are usually filled
with 0.5m (20 inches) of sand covered with 0.3m (12 inches) of anthracite.
 4 Water Chemistry and Pretreatment

The multimedia filters can be operated by either gravity or a pressure. A higher pressure drop
can be applied for higher filter beds or smaller filter grains or higher filtration velocities. The
design filtration flow rates are usually 10-20 m/h, and the backwash rates are in the range of
40-50 m/h.
For feed waters with a high fouling potential, flow rates of less than 10m/h and / or second pass
media filtration is preferred.
The available pressure is usually about 5m of head for gravity filter and 2 bar (30 psi) to 4 bar
(60 psi) for pressure filters. Periodically, the filter is backwashed and rinsed to remove the
deposited matter, when the differential pressure increase between the inlet and outlet of the filter
is 0.3 to 0.6 bar (4 to 9 psi) for the pressure filter and about 1.4m for the gravity filter. Backwash
time is normally about 10min.
Frequent shut-downs and start-ups should be avoided, because a velocity shock will release
previously deposited particulate matter.

Oxidation Filtration
Some well waters, usually brackish waters, contain divalent iron (Fe2+), manganese (Mn2+) and
sometimes sulfide in the absence of oxygen. If such a water is exposed to air or is chlorinated,
Fe2+, Mn2+ and sulfide are oxydized to Fe3+, Mn3+ and elemental sulfur, respectively, which form
insoluble colloidal hydroxides and elemental sulfur as shown in the following equations :

4Fe(HCO3)2 + O2 + 2H2O → 4Fe(OH)3 + 8CO2

4Mn(HCO3)2 + O2 + 2H2O → 4Mn(OH)3 + 8CO2
2H2S + O2 → 2S + 2H2O

Iron fouling occurs more frequently than manganese fouling, since iron is present in the raw
waters more abundantly than manganese and the oxidation of iron occurs at a much lower pH.
Thus iron fouling is still possible even if the SDI is below 5 and the level of iron in the RO feed
water is below 0.1㎎/ℓ.
An RO system can treat the well water containing iron (Fe2+) in a closed system without
exposure to air or to any oxidizing agent, e.g. chlorine. A low pH is favorable to retard Fe 2+
oxidation. At pH <6 and oxygen <0.5㎎/ℓ, the maximum permissible Fe2+ concentration is 4㎎/ℓ.
Oxidation and filtration can be done in one step by using a filter media greensand coated with
MnO2 to oxidize Fe2+. Greensand is a green (when dry) mineral glauconite and can be regenerated
with KMnO4 when its oxidizing capability is exhausted. After the regeneration, the residual
KMnO4 has to be thoroughly rinsed out in order to avoid an oxidation damage of the membranes.
This technique is used when Fe2+ is present in the raw water in less than 2㎎/㎖.
An alternative to the greensand is a product called Birm (a registered trademark of Clack

Global Initiative 21
 Water Chemistry and Pretreatment
4
Corporation) which is a light silica coated with manganese dioxide. It catalyzes the oxidation of
iron and manganese by dissolved oxygen. It does not lose the catalytic activity and thus does not
need to be regenerated using KMnO4. However, Birm requires a certain concentration of
dissolved oxygen in the feed water at an alkaline pH. The alkaline condition may increase a
possibility of CaCO3 precipitation. Birm is less expensive than greensand.

Crossflow Microfiltration and Ultrafiltration

Crossflow microfiltration can effectively remove most suspended matter, depending on the
pore size of the membranes. Normally microfiltration MF membranes with a pore size in the
range of 1 to 10㎛ are used, depending on the feed water quality. When the silica concentration
in the concentrate stream exceeds the calculated solubility, MF membranes with 1㎛ pore size
are recommended to minimize the interaction with other colloids such as iron and aluminum
colloids.
Now the fouling problem due to colloid deposit is transferred from the RO membrane to the
MF membrane. In fact, the fouling of MF membranes is more severe and more often than RO
membranes because of a high specific permeate flow. However, it is easier to clean a fouled MF
membrane than an RO membrane, using periodic backflush cleanings which have been proven
very effective for cleaning MF membranes, not for RO membranes. Chlorine can be added to the
wash water in order to prevent biological fouling. Continuous microfiltration membranes (CMF)
equipped with an automated backflush washing mechanism are available in the market . It is
usually economical to employ multimedia filters prior to MF membranes since the cheaper media
filters will take a majority of suspended matter to reduce the costly cleaning frequency of MF
membranes and thus to extend the membrane life time.
Cheaper disposable cartridge filters may be an alternative to the MF membranes and replaced
before the pressure drop has increased to the permitted limit, but latest after 3 months. Replacing
cartridge filters more often than every 1 to 3 months usually indicates a problem with the
pretreatment. If it is desirable to remove soluble organic or inorganic polymers, then
ultrafiltration (UF) membranes could be used.
 4 Water Chemistry and Pretreatment

4-7. Biological Fouling Prevention

All raw waters contain microorganisms such as bacteria, algae, fungi, viruses and various
organisms. Microorganisms can be regarded as colloidal matter and removed by the pretreatment.
However, it is very difficult to remove all the microorganisms and a few of them may escape the
pretreatment to reproduce and form a biofilm in the downstream processes. MF membranes could
remove most of the microorganisms, but still some will escape through the pores or defects of the
membranes to reach the RO membranes. The symptoms of the biologically fouled RO system are
an increase of the differential pressure accompanied by membrane flux decline.
The potential for biological fouling is more pronounced with surface water than well water. The
concentration of bacteria in water is directly related to the biological fouling potential.

Total Bacteria Count (TBC) is a method to determine the total number of viable
microorganisms in a water sample according to ASTM F60 by filtering a measured quantity of
water through a membrane filter. Subsequently, the organisms retained on the filter surface are
cultured on the proper nutrient medium for several days to develop colonies, which are then
observed and counted at low power magnification.
This culture technique can be applied to monitor the microbial activity from the intake through
the subsequent treatment steps up to the concentrate stream and the permeate.

Direct Bacteria Count techniques employ filtration of the water sample and counting the
retained microorganisms on the filter plate directly under a fluorescent microscope after they are
stained with acridine orange or INT stain. INT stain can tell the difference between living cells
and dead cells.
The germicidal efficiency of free residual chlorine is directly related to the concentration of
undissociated hypochlorous acid HOCl which is 100 times more effective than the hypochlorite
ion OCl-.
The fraction of undissociated HOCl increases with decreasing pH and temperture.

At pH = 7.5(25℃), 50% HOCl

At pH = 6.5(25℃), 90% HOCl
At pH = 7.5( 5℃), 62% HOCl

Global Initiative 21
 Water Chemistry and Pretreatment
4
In case of existence of any ammonia in the water being treated, chlorine can react with ammonia to
give various chloramines in a series of stepwise reactions :

HOCl +NH3 → NH2Cl + H2O

HOCl +NH2Cl → NHCl2 + H2O
HOCl +NHCl2 → NCl3 + H2O

Chloramines also have a germicidal effect, albeit lower than that of chlorine. One advantage
with the chloramines is that they do not oxidize the RO membranes. However, there will be
always some residual unreacted HOCl which can still oxidize the membranes and thus care must
still be taken when chloramines are used as a disinfectant.
To determine the optimum chlorine dosage, best point of injection, pH and contact time to
prevent biofouling, the ASTM D1291, Standard Practice for Determining Chlorine Requirement
of water should be applied to a representative water sample.

Dechlorination
The residual chlorine has to be dechlorinated before it reaches the RO membrane. CSM RO
membrane has some chlorine tolerance, but eventual degradation may occur after 200-1000 hours
of exposure to 1 ㎎/ℓ of free chlorine, depending on the pH, temperature and residual transition
metals such as iron in the feed water. Under alkaline pH conditions, chlorine degrades the
membrane faster than at neutral or acidic pH. At an acidic pH, chlorine becomes more effective
as a disinfectant. The influent of RO must be dechlorinated to maintain the residual free chlorine
concentration below 0.05ppm.

Activated carbon(AC) bed is very effective in dechlorination of the residual chlorine in RO

feed water according to the following reaction :

C + 2HOCl → CO2 + 2HCl

Sodium Metabisulfite (SMBS) or sodium Bisulfite (SBS) is most commonly used for removal
of free chlorine and as a biostatic, as shown in the following reaction.

Na2S2O5 +H2O → 2NaHSO3

2NaHSO3 + 2HOCl → H2SO4 + 2HCl + Na2SO4

Based on the equivalent weight estimation, 1.34㎎ of sodium metabisulfite will react with
1.0㎎ of free chlorine. In practice 3.0mg of sodium metabisulfite is normally used to remove
 4 Water Chemistry and Pretreatment

1.0㎎ of chlorine to make sure that all the chlorine is reduced.

The injection point of the SMBS solution is preferably upstream of the cartridge filters in order
to keep the residual chlorine up to the filters to prevent microbial growth in the filters.

The SMBS solution should be filtered through a separate cartridge and injected through static
mixers for good mixing into the RO feed line. The absence of chlorine should be monitored using
an oxidation - reduction potential (ORP) electrode downstream of the mixing line.

Disinfection by Ultraviolet (UV) Irradiation

UV at 254㎚ is known to have a germicidal effect and has been used especially for small-scale
plants. No chemicals are added and the equipment needs little attention other than periodic
cleanings or replacement of the mercury lamps. However, UV treatment is limited to relatively
clean waters, because colloids and organic matter interfere with the penetration of UV into depth
of the turbid water.

Biological Activity Control by Sodium Bisulfite

Sodium bisulfite (SBS) concentrations in the range of up to 50㎎/ℓ in the feed stream of sea
water RO plants have proven effective to control biological fouling. Colloidal fouling has also
been reduced by the method. SBS is also helpful in controlling calcium carbonate scaling by
supplying protonium (H+) ions as shown below.

2NaHSO3 + CaCO3 → Na2SO3 + Ca2+ + HCO3- + HSO3-

Global Initiative 21
 Water Chemistry and Pretreatment
4
4-8. Organic Fouling Prevention

Adsorption of organic substances on the membrane surface causesflux loss. Sometimes the
adsorption is irreversible when the organic substances are hydrophobic or positively charged
polymers with high molecular weight. Such organic substances including organics present as an
emulsion must be removed in the pretreatment.
Organics occuring in natural waters are usually humic substances in concentrations between 0.5
and 20㎎/ℓ as TOC. Pretreatment should be considered when TOC exceeds 3㎎/ℓ. Humic
substances can be removed by a coagulation process with hydroxide flocs or by ultrafiltration or
by adsorption on activated carbon.
Oils (hydrocarbon or silicon-based) and greases contaminating the RO feed water at levels
above 0.1㎎/ℓ should be removed by coagulation or activated carbon. Once the membranes are
fouled by oils and greases, they can be cleaned off with alkaline cleaning agents if the flux has
not declined by more than 15%.

Trihalomethane (THM)
THMs (e.g. chloroform) are produced from a chemical reaction between free chlorine and trace
organics. THMs are considered potentially carcinogenic, so are a concern when found in drinking
water. And also trace THMs can be a major concern in the semiconductor industry, if present in
feed water.
About 90% of THMs can be rejected by RO membranes. Activated carbon can also adsorb
THMs well. A complete removal of THMs is possible by using both activated carbon and RO
membranes.
 System Design
5
5-1. Introduction

After the pretreatment of the raw water, the treated feed water enters the RO desalination
system. The goal of an efficient RO system for a certain required permeate flow is to minimize
feed pressure and membrane costs (number of elements) while salt rejection and recovery should
be maximized. The optimum design is influenced by the relative importance of these aspects (e.g.
recovery vs. membrane costs) related to operating parameters. The desired salt rejection is usually
achievable but the recovery determined by a permeate flow is affected by many factors.
An RO system design starts with prioritizing the relationship between the desired permeate flow
and operating parameters followed by optimizing the parameters within physical limits levied by
both RO membrane elements and the feed water possessing a potential of scaling and fouling. For
examples, the recovery of brackish water systems is limited by the solubility of sparingly soluble
salts and colloidal fouling potential of the feed water up to 87.5% (achievable only by multi-array
system). The recovery by a single brackish water element is limited to 15% by the scaling and
fouling potential of the feed water with SDI of 3 to 5.
On the other hand, in seawater desalination, the limit of 30 to 45% recovery (achievable only by
multiple elements in series) is mainly imposed by the enhanced osmotic pressure of the
concentrate stream, because a typical sea water element is durable only up to 69 bar (1000psi).
The recovery by a single seawater element is limited to 10% of the seawater feed with SDI less
than 5.
While in seawater desalination systems, the permeate flux is relatively low even at maximum
allowed pressure, the permeate flux could be very high in brackish water systems without reaching
the limit of 41 bar (595 psi) for brackish water elements. Although it is tempting to increase the
permeate flux in order to minimize the costs for membrane elements, the flux has to be limited in
order to avoid fouling and scaling. From experience, the flux limit in system design depends on
the fouling tendency of the feed water.
 5 System Design

5-2. System Design Guidelines

As mentioned above, the main factors influencing seawater system designs are the osmotic
pressure and the physical durability of seawater elements, while those affecting brackish water
system designs are scaling and fouling potential of the feed water. Hence, seawater systems could
be built relatively easily by designing the permeate flow rate within the limit of the two factors
with a minimal attention to scaling and fouling potential of seawater. The usual sea water recovery
of 30 to 50% can be obtained from single-array systems.
Comparatively, brackish water systems require more elaborate designs mainly due to various
scaling and fouling potentials of the feed waters and additionally due to multi-array systems when
more than 50% recovery is desired. The system designed with high permeate flux rates is likely to
experience higher fouling rates and more frequent chemical cleaning. The silt density index (SDI)
value of the pretreated feed water correlates well with the amount of fouling material present.
Experience from the correlation of the SDI value with the membrane fouling trend can set the
limits on permeate flux and element recovery for different types of waters, which can be the
starting point of the system design guidelines shown in Table 1 below.

Table 1. System Design Guidelines

Well Water/ Softened Surface Sea

Feed Source
Softened Water Surface Water Water Water

Feed Water SDI <3 3-5 3-5 <5

Max, % Recovery per
19 17 15 10
Element

Max, 2.5" Diameter 710(0.11) 500(0.08) 500(0.08) 500(0.08)

Permeate
Flow Rate 4" Diameter 2100(0.33) 1870(0.3) 1740(0.27) 1500(0.23)
per Element,
gpd(m3/h) 8" Diameter 7400(1.16) 6600(1.04) 5800(0.92) 5800(0.92)

2.5" Diameter 5.7(1.3) 5.7(1.3) 5.7(1.3) 5.7(1.3)

Max. Feed
Flow Rate 4" Diameter 18(4.1) 18(4.1) 18(4.1) 18(4.1)
per Element,
gpm(m3/h)
8" Diameter 62(14.1) 60(13.7) 55(12.6) 60(13.7)

Min. 2.5" Diameter 1(0.22) 1(0.22) 1(0.22) 1(0.22)

Concentrate
Flow Rate 4" Diameter 4(0.91) 4(0.91) 4(0.91) 4(0.91)
per Element,
gpm(m3/h) 8" Diameter 16(3.6) 16(3.6) 16(3.6) 16(3.6)

Global Initiative 21
 System Design
5
The guidelines in Table 1 are based on a continuous process with a well-designed and operated
pretreatment system. Exceeding the limits will result in more frequent cleanings than about four
times a year, and a reduced membrane lifetime Consequently, this is not only increase capital and
operating costs, but also result in additional operating cost due to increased membrane cleaning
replacement frequencies.
 5 System Design

5-3. Batch vs. Continuous Process

The majority of RO systems are designed for continuous operation with constant permeate flow
and constant system recovery as shown in Figure 1. Variations in feed water temperature and
fouling effects are compensated by adjusting the feed pressure.

Concentrate

Feed Permeate
▶

Figure 1 : Continuous RO Process

In certain applications, when relatively small volumes (batches) of special feed waters occur
discontinuously, e.g. wastewater or industrial process solutions, the batch operation mode is
preferred. The feed water is collected in a tank and treated subsequently. The permeate is
removed and the concentrate is recycled back to the tank. At the end of the batch process, a small
volume of concentrate remains in the feed tank. After this has been drained, the membranes are
typically cleaned before the tank is filled again with a new batch. Figure 2 shows the batch
operation mode.

Feed/
Concentrate Permeate
▶

Figure 2 : Batch RO Process

Global Initiative 21
 System Design
5
A modification of the batch mode is the semi-batch mode. The feed tank is refilled with feed
water already during operation. The batch is terminated with the feed tank full of concentrate.
This allows a smaller tank to be used.
Batch systems are usually designed with a constant feed pressure and a declining permeate flow
while the feed becomes more concentrated. The guidelines given in Table 1 should be applied to
batch systems as well. The permeate flow limits however, are conservative and may be exceeded,
if an appropriate cleaning frequency is taken into account.
Some advantages of the batch process over the continuous process are :
· System recovery can be maximized batch by batch
· Cleaning is easily implemented
The disadvantages are :
· No constant permeate quality;
· Larger pump required;
· Higher total running costs.
 5 System Design

5-4. Single Pressure Vessel System

A Single Pressure Vessel System consists of a pressure vessel with up to seven membrane
elements, which are connected in series. The concentrate of the first element becomes the feed to
the second, and so on. The product tubes of all elements are coupled and connected to the vessel
permeate port. The permeate port may be located on the feed and or on the concentrate end of the
vessel.
Single Pressure Vessel Systems are chosen when only one or few membrane elements are
needed for the specified permeate flow.
Feed water enters the system through the shut-off valve and flows through the cartridge filter to
the high pressure pump.
From the high pressure pump, the feed water flows to the feed inlet connection of the vessel.
The product stream should leave the vessel at no more than 0.3 bar( 5 PSI ) over atmospheric
pressure.
In other words, in any event, the back pressure of the permeate must not be greater than the feed
pressure in order not to damage the membranes.
The concentrate leaves the concentrate outlet connection at essentially the feed pressure.
Pressure drop will usually amount to 0.3 to 2 bar (5~30 PSI) from feed inlet to concentrate outlet,
depending on the number of membrane elements, the feed flow velocity and the temperature. The
concentrate flow is controlled by the concentrate flow control value.
The system recovery is controlled by this valve and must never exceed the design value.

Global Initiative 21
 System Design
5
5-5. Single Array System

In a single array system, two or more modules are arranged in parallel. Feed product and
concentrate are connected to manifolds. Other aspects of the system are the same as in a
single-pressure vessel system. Single-array systems are typically used where the system recovery
is less than 50%. e.g. in sea water desalination.
An example of a single-array system is outlined in Figure 3. Each of the three pressure vessels
houses six CSM elements RE8040-SHN.

RE8040-SHN

Feed
▶ RE8040-SHN
High Pressure
Pump
Cartridge RE8040-SHN
Filter

Permeate

Concentrate Valve →

Concentrate

Figure 3 : Single Array System

 5 System Design

5-6. Multi Array system

Systems with more than one array or stage are used for higher system recoveries without
exceeding the single element recovery limits. Usually two arrays will suffice for recoveries up to
75%, and three must be used for higher recoveries 87.5%. These numbers are based on the
assumption that standard pressure vessels with six elements are used. Higher recoveries can be
achieved if feed water contains less suspended solids and scale potential.
Generally speaking, the higher the system recovery, the more membrane elements have to be
connected in series. In order to compensate for the permeate that is removed and to maintain a
uniform feed flow to each array, the number of pressure vessels per array decreases in the
direction of feed flow.
A typical two-array system using a staging ratio of 2:1 is shown in Figure 4. The staging ratio is
defined as the ratio of pressure vessels in two adjacent arrays, upstream vessels: downstream
vessels.

RE8040-BE

Feed
RE8040-BE RE8040-BE Concentrate

▶ Concentrate
High Pressure Valve
Pump RE8040-BE RE8040-BE

Cartridge
Filter RE8040-BE

Permeate

Figure 4 : Two Array System

Global Initiative 21
 System Design
5
5-7. Double Pass System

A double pass system is the combination of two RO systems where the permeate of the first
system (first pass) becomes the feed for the second system (second pass). Both RO system may be
of the single-array or multi-array type, either with continuous flow or with concentrate
recirculation. With this concept, a product conductivity of < 1㎲/㎝ can be achieved without any
contamination from organic matter and pyrogens including bacteria. Thus applications of double
pass systems can be found in the production of pure water for pharmaceutical, medical and
semiconductor industries.
Figure 5 shows a schematic flow diagram of a double pass system.

Concentrate Valve

Concentrate

Feed

▶ ▶
Permeate
High Pressure High Pressure
Cartridge Pump RO Ⅰ Pump
RO Ⅱ
Filter

Figure 5 : Double Pass System

The concentrate of RO II is recycled back to the feed of RO l, because its quality is usually
better than the system rawfeed water.
The high quality of RO II feed water (RO l permeate) enable the RO II can be designed for a
higher recovery than RO I and with fewer membrane elements.
Instead of having a separate high pressure pump for the second pass, the whole system can also
be operated with one single high pressure pump, provided the maximum permissible feed pressure
of the membrane element is not exceeded (41 bar (600psi) for brackish water elements).
The second pass is then operated with the permeate backpressure from RO I. Care must be
exercised that the permeate backpressure at no time exceeds the feed pressure by more than 0.3 bar
(5psi)
 5 System Design

5-8. Number of Elements and Pressure Vessels for System Design

As mentioned in the introduction, RO systems are usually designed for a specified amount of
water production per day (total permeate flow QT). Then a number of membrane elements (NE)
required to produce QT is estimated from dividing QT by average permeate flow per element (QA)

QT
NE = (1)
QA

In most standard applications, the average permeate flow per element (QA) is about 75% of the
maximum permeate flow per element (QM) which is shown in Table 1 in the section of system
design guidelines.
QA = 0.75 QM (2)

Thus equation (1) can be converted to equation 3.

QT
NE = (3)
0.75QM

And also the number of pressure vessels, NV is obtained from dividing NE by PE which is the
number of elements per pressure vessel.

NE
Nv = (4)
PE

Standard vessels contain six elements. Nv is rounded to the next highest whole number.
With QT , NE , NV and the analysis of the feed water source, a system can be selected. The
selected system must then be verified using CSM Pro5 computer simulation program. This
program calculates the feed pressure and the permeate quality of the system including the
operating data of all individual elements. Furthermore, the system design can easily be optimized
by adjusting the number and type of elements and their arrangement (e.g. single array or multi
array). However, the optimization should be conducted within the physical limit of the elements
and the empirical limit of the recovery rate according to the guidelines of the fouling potential of
the feed waters.
For examples, Two-array systems with 6-element vessels effectively employ twelve spiral
wound elements in series and are generally capable of operating at an overall recovery rate of 60

Global Initiative 21
 System Design
5
to 75%. For such systems the average individual recovery rate per element will vary from 7 to
12%. To operate a two-array system at an overall recovery much higher than 75% will cause an
individual element to exceed the maximum recovery limits (e.g. 15% for a feed water with SDI in
the range of 3 to 5). Then, a third array will have to be employed to place eighteen elements in
series for a recovery rate lower than 15% per element.
If two-array systems are operated at too low a recovery (e.g. <60%), the feed flow rates to the
first-array vessels can be too high causing excessive feed/concentrate-side pressure drops and
potentially damaging the elements. As a result, systems with lower than 60% recovery will
typically utilize single array configurations.

Table 1. Recommended average permeate flux vs various water source

Water Source Recommended permeate flux
Waste Water (SDI<5) 8 ~ 12 gfd
Waste Water pretreated by UF(SDI<3) 10 ~ 14 gfd
Seawater, open intake(SDI<5) 7 ~ 10 gfd
High walinity well water(SDI<3) 8 ~ 12 gfd
Surface water(SDI<5) 12 ~ 16 gfd
Surface water(SDI<3) 13 ~ 17 gfd
Well water(SDI<3) 13 ~ 17 gfd
RO/UF permeate(SDI<1) 21 ~ 30 gfd
 5 System Design

5-9. Important Parameters for System Design

The performance of an RO system is usually defined by its feed pressure, its permeate flow and
its salt passage (salt rejection). Two simple equations show the relationship among the
parameters.
∆𝑃𝑓𝑐
Q = A × S × (𝑃𝑓 − − 𝑃𝑝 − π𝑎𝑣𝑒 ) (1)
2

Q = permeate flow
A = membrane permeability coefficient
S = membrane surface area
Pf = feed pressure
ΔPfc = concentrate side pressure drop
Pp = permeate pressure
πave = average osmotic pressure

Equation 1 indicates that the permeate flow Q is directly proportional to the surface area S
times a net permeation driving force.

∆𝑃𝑓𝑐
(𝑃𝑓 − − 𝑃𝑝 − π𝑎𝑣𝑒 )
2

On the other hand, the salt passage is mainly governed by diffusion and thus the salt flux N is
proportional to the difference in salt concentration between both sides of the membrane as shown
in equation 2.
N = B × (Cfc - CP) (2)
B = salt diffusion coefficient
Cfc = feed - concentrate average concentration
CP = permeate concentration

The equations 1 and 2 can be used to calculate the performance of a single RO element and also
of a multi RO element system if average values for
(Pf - ΔPfc / 2 – Pp - πave) and (Cfc - Cp), temperature, and number of elements are known.

When a large number of elements are combined in a system with a complex

series-parallel-series configuration and only inlet operating variables are known, system
performance calculation becomes more complex and tedious. However, if all the operating
Global Initiative 21
 System Design
5
conditions for first element can be confirmed, then, the operation variables including flow,
pressure, concentration, etc., for following elementcan be calculated by a computer simulation
program.

The first step is to guess the applied pressure Pf needed to produce the water flow rate (Q)
required. Equation 1 can be converted to equation 3.
𝑄 ∆𝑃𝑓𝑐
𝑃𝑓 = + π𝑎𝑣𝑒 + + 𝑃𝑝 (3)
𝐴×𝑆 2

To solve equation 3 for the feed pressure, the total area of the membrane from the total number
of elements is summed to give S. The average osmotic pressure (πave) is calculated from the
average concentration of the feed and the concentrate which is obtained from the nominal
recovery and salt rejection of the element. The differential pressure ΔPfc may be estimated from
the average of ΔPfc calculated from flow to the first element in series and the last element in the
array.
The membrane permeability coefficient (A) is provided by the manufacturer or obtained from
the slope of the curve of the permeate flow (Q) versus the operating pressure (Pf - Δπ).
Now using the guessed applied pressure (Pf), the computer calculates feed flow rates, permeate
flow rates and salt concentrations on an element by element basis through the system. The total
calculated permeate flow (Q) is lower than the target flow, then the pressure is raised
proportionally and if Q is higher than the target flow, then the pressure is lowered proportionally.
The estimated water qualities of the concentrate exiting the last array is then checked for
solubility and pH limits for various scaling forming salts. If the solubility limits are exceeded, Q
or the pressure may be decreased, but not so low to drop below the minimum concentrate flow
rate. If it does, then more staging (number of arrays) is required.
Another limiting factor is that in any event, the maximum feed flow rate and permeate flow rate
of the first element in the first array must not be exceeded. If they do, then a physical damage to
the element and rapid fouling would occur. Controlling the feed pressure or adding more
elements may be necessary to avoid such problems.
To be accurate and closer to the real situation, equation 1 could have several additional
variables as shown in equation 4.

∆𝑃𝑓𝑐
Q = A × S × TCF × FF × (𝑃𝑓 − − 𝑃𝑝 − π𝑎𝑣𝑒 ) (4)
2

𝑆
𝐶𝑝 = 𝐵 × 𝐶𝑓𝑐 × 𝑃𝑓 × 𝑇𝐶𝐹 × (5)
𝑄
 5 System Design

1 1
TCF = exp [𝑈 × ( − )] (6)
298 273 + 𝑇

∆π = π − π𝑝 (7)

𝐶𝑓𝑐
π = π𝑓 × × 𝑃𝑓 (8)
𝐶𝑓
Pf = exp[0.7 × Y] (9)

Q = permeate flow
A = membrane permeability coefficient
S = membrane surface area
TCF = temperature correction factor
FF = fouling Factor, about 0.8
Pf = feed pressure
∆Pfc = feed-concentrate side pressure drop
Pp = permeate pressure
∆π = driving osmotic pressure
Cp = permeate concentration
B = salt diffusion coefficient
Cfc = concentration of the concentrate
Cf = concentration of the feed
Pf = concentration polarization
U = temperature correction factor coefficient which depends on membrane sheet
π = average feed - concentrate osmotic pressure
πf = osmotic pressure of feed
π p = permeate side osmotic pressure
πf = feed osmotic pressure
Y = recovery of the element

's CSM PRO5 computer program utilizes equations 4 and 9. Through the iterative trial and error
process, it projects the performance of given systems, and optimizes the design of the system.

Global Initiative 21
 System Design
5
5-10. Testing of System Designs for Unusual Applications

For the desalination of well characterized waters with known SDI and composition, the RO
performance can be projected with reliable accuracy by the computer program. However, testing
is recommended to support the proper system design in special situations such as :
· Unknown feed water quality
· Waste waters
· Special permeate quality requirements
· Very high system recoveries
· Very large plants
First, a screening test is demanded to select the right membrane and to obtain an idea about flux
and rejection properties of this membrane in the special situations. Usually a small piece of flat
sheet coupon membrane is used for the screening test.
Second, using a 2540 size element, scale up data such as permeate flow and salt rejection as a
function of feed pressure and system recovery are obtained. In subsequent batch mode test,
leading the permeate into a separate containment and returning the concentrate to the feed tank,
permeate flow and salt rejection are monitored until the permeate flow has declined to a very low
value such as 0.09 gpm. From the batch tests, an indication of membrane stability and fouling
propensities can be revealed.
Third, a pilot test is conducted in the field with a continuous operation mode. The pilot plant
has at least contains one 8040 size element, preferably an arrangement of elements similar to the
arrangement in a large scale system. The permeate flow of the pilot plant should be at least 1% of
the large scale plant flow. The test should be run for more than 30 days. The objective is to
confirm the system design reliability and to optimize operating parameters as well as to minimize
the risk involved in large scale plants.
 5 System Design

5-11. System Components

High Pressure Pump

The high pressure pump must provide smooth and continuous flow to the RO membrane
elements and also be sized to provide the necessary flow rate at the desired pressure. Its energy
consumption is one of the major expenses of RO system operation. There are two types of high
pressure pumps, centrifugal pumps and piston pumps. Centrifugal pumps, which operate by
spinning the fluid with the pump impellers are more energy-efficient than piston pumps, which
use various numbers of mechanical plungers to create pressure. The centrifugal pumps provide
more smooth and continuous flow than the piston pump, and are controlled by a throttling valve
on the discharge line. The piston pump cannot be throttled, so pressure is controlled by a back
pressure valve installed in a bypass line from the pump discharge to the pump suction. The
pressure from piston pumps tends to pulsate, creating surges, which could damage RO membrane
elements and therefore should be controlled using a damper (accumulator).

Pressure Vessel
Pressure vessel (membrane element housing) is designed for specific pressure applications.
Most pressure vessels are overdesigned for safety reasons to withstand a pressure at 1.5 times the
rated operating pressure. The vessel materials are usually FRP (fiber glass reinforced plastic) and
sometimes stainless steel (316L) for special applications such as very high pressure applications
(>800psi).
Pressure vessels are available with different diameters, lengths, and pressure ratings. The
smaller vessels with diameters in the range of 1.5 to 2.5 inch are usually supplied by (need to be
checked). All other vessels with diameter in the range of 2.5 to 8 inch and the pressure rating up
to 1000psi are manufactured by Advanced Structures Inc (also need to be checked).

Alarms and Shutdowns

There is always the possibility of a malfunction in the RO system or with the pretreatment.
Instruments can be used to monitor the quality of the RO feed water as well as the performance of
the RO system. The instruments such as flow meters, pressure gauges, pH meters and
conductivity meters can be connected with an audible alarm that will sound if some parameter is
not within design specifications. Furthermore, if the parameter is significantly out of the
specifications to damage the high pressure pumps and the RO elements or to produce an
unacceptable permeate water, then the alarm should warrant automatic shutdown of the RO
system. Suggested alarms and shutdowns are shown below :

Global Initiative 21
 System Design
5
· Low inlet pressure damaging the high pressure pump
· Too high feed pressure damaging the RO elements
· High feed temperature damaging the RO elements
· High permeate pressure damaging the RO elements
· Too high concentration of colloidal matter or sparingly soluble salts in the feed
damaging the RO elements
· Oxidizing agents in the feed damaging the RO elements
· Low concentrate flow fouling and/or scaling the RO membranes
· Oil in the feed fouling the RO membranes

Instruments, Valves, and Equipment

As mentioned above, instruments and valves are necessary not only to detect a malfunction in
the RO system, but also to ensure proper routine operation of the system. The necessary
instruments, valves, and other equipment are listed below in more details.
Pressure gauges to measure the pressure drop across the cartridge filter, the pressure on the
pump inlet line and discharge line, the feed pressure to the membrane element(s), the pressure
drop between feed and concentrate of each array, and eventually the pressure in the permeate line.
Liquid-filled gauges should contain membrane compatible fluids such as water or glycerine in
place of oils or other water immiscible liquids.
Flow meters to measure feed, concentrate and total permeate flow rate, also permeate flow rate
of each array. pH meter in the feed line after acidification to control carbonate scaling potential.
Conductivity meters in the feed line, in the brine line, and in the permeate line to determine
permeate quality and salt rejection. Sample ports on the feed, concentrate and permeate line (total
permeate and permeate of each array) to be able to evaluate system performance. A sample port
on each pressure vessel permeate outlet is recommended to facilitate troubleshooting.
Feed inlet valve to shut down the plant for maintenance and preservation. Valve on the pump
discharge line or pump bypass line to control feed pressure during operation and feed pressure
increase rate during start-up. Check valve on pump discharge line. Check valve and atmospheric
drain valve on permeate line to prevent the permeate pressure from exceeding the feed pressure.
Flow control valve on the concentrate line to set the recovery (Caution : back-pressure valve must
not be used). Valve in the permeate line to provide permeate drain during cleaning and start-up.
Valves in the feed and concentrate line (and between arrays) to connect a cleaning circuit.
A small draw-back tank is necessary in the permeate line to provide enough volume for
osmosis backflow when a seawater system shuts down. Without the tank, air could be sucked into
the membrane elements to dry the membrane (flux loss) and to contaminate the permeate side by
airborn bacteria and fungi.
A shut down flush system flushes the feed-concentrate line with pretreated feed water or with
 5 System Design

permeate water after shut down, especially when scale inhibitors are used, and also in the case of
a seawater system.

Materials of System Construction

The materials of system construction including the RO elements, pumps, pressure vessels,
pipes, valves and instruments should be compatible with the pressures, vibrations, and
temperatures fluctuations during the RO system operation. The materials must also be resistant to
the potential corrosion attacks caused by the high chloride content of the feed water and the
concentrate stream, and the chemicals used for membrane cleaning.
Non-metalic materials such as plastics and fiberglass are widely used not only for economic
reasons (e.g. pressure vessels and pipes), but also for preventing corrosion and chemical attacks,
usually in the low pressure (<10 bar) applications.
However, it is usually necessary to use metals for the high-pressure (10-70 bar/ 200-1,000PSI)
parts such as pumps, piping and valves. Carbon and low alloy steels do not have sufficient
corrosion resistance, and their corrosion products can foul the membranes.
Stainless steel type AISI 316 L with <0.03% C is recommended for the pipe system for RO
plants with concentrate stream TDS below 7000 ppm. For TDS higher than 7000ppm, stainless
steel type 904L is preferred for pipes and bends for welding and stainless steel type 254 SMO
should be used for flange connections, valves, and pumps where crevices occur.

Global Initiative 21
 System Operation
6
6
6-1. Introduction

Successful long term performance of the RO system mainly depends on proper operation and
maintenance of the system.
This includes well planned pretreatment and initial plant start-up. Membrane fouling and
scaling, which can reduce permeate flow rate and salt rejection significantly, are caused primarily
by a poor pretreatment and also an improper system operation. Mechanical and chemical induced
damages of the RO system including the membranes may also arise from improper operations.
Record keeping and data normalization is indispensably required to analyze the actual plant
performance and to localize the source of any problem. Data normalization with reference to the
initial start-up system performance is useful to show any performance changes during the
sequential operations.
Complete records are required in case of a system performance warranty claim including
elements.
 6 System Operation

6-2. Initial Start-up

1) Checklist before Start-up

Before starting up the RO system, it is important to make sure that the whole pretreatment
process is working according to the specifications. If the chemical characteristics of the raw water
are changed, then a full analysis of the water entering the RO unit should be conducted so that
proper measurements can be made to put the variables under control.
Factors affecting the feed water quality and thus the system design are as follows :
Flow, SDI, Turbidity, Temperature, pH, TDS, residual chlorine, and bacteria counts

And also the following mechanical inspections of the RO system are recommended for the
initial start-up.

· Operational conditions of media filters and cartridge filters

· Feed line is purged and flushed before pressure vessels are connected
· Chemical dosing lines and valves
· Proper mixing of chemicals in the feed stream
· Safety shut off of the RO system when the chemical dosage pumps are shut-down
· Complete chlorine removal prior to the membranes
· Instrumentation for proper operation and monitoring of the pretreatment and RO system
· Installation and calibration of such instrumentation
· Installation of pressure relief protection device
· Piping and securing pressure vessels for operation and cleaning mode
· Lubrication and proper rotation of pumps
· Valves for permeate line, feed flow, and reject flow control are in open position
· Initial feed flow should be limited to less than 50% of operating feed flow

2) Start-up Sequence
· Before starting the initial operation sequence, the pretreatment section should be thoroughly
rinsed to flush out debris and other contaminants without allowing the feed to enter the
elements
· Make sure that all valve settings are correct. The feed pressure control and concentrate
control valves should be fully open
· Use the feed water at a low flow rate to expel the air out of the elements and pressure
vessels at a gauge pressure of 30 to 60 psi for more than 30 minutes. All permeate and
 System Operation
6
6
concentrate flows should be directed to an approved waste collection drain during flushing.
At this point, all pipe connections and valves are checked for leaks
· After the system has been flushed, close the feed pressure control valve, but make sure that
the concentrate control valve is open. Improper concentrate valve control may cause the
recovery to be exceeded which can initiate scaling within the elements.
· Open the feed pressure control valve little by little so that feed pressure does not exceed 4.0
kg/cm2 (60 psi) and then start the high pressure pump
· Increase opening of the feed pressure valve slowly to elevate the feed pressure and feed
flow rate to the elements until the design concentrate flow is reached. Then slowly close
the concentrate control valve until the ratio of permeate flow to concentrate flow
approaches the designed recovery ratio
· Repeat opening of the feed pressure control valve and closing of the concentrate control
valve until the design permeate and concentrate flows are obtained, while checking the
system pressure to ensure that it does not exceed the upper design limit
· After adjusting the two valves, calculate the system recovery and compare it to the system
design value
· Check chemical additions of acid, scale inhibitor, and sodium metabisulfite. Measure pH,
conductivity, calcium hardness and alkalinity levels to calculate the Langelier Saturation
Index (LSI) or the Stiff & Davis Saturation Index (S & DSI) for the evaluation of scale
formation potential.
· Take the first reading of all operating parameters after allowing the system to run for one
hour. Read the permeate conductivity from each pressure vessel and identify any vessel
with any malfunction
· After 24 to 48 hours of operation, record all plant performance data such as feed pressure,
differential pressure, temperature, flows, recovery ratio and conductivity readings. And
also analyze the constituents of feed water, concentrate, and permeate water samples.
Compare system performance to design values. Use the initial system performance
information as a reference for evaluating future system performance. Measure system
performance regularly during the first week of operation
 6 System Operation

6-3. Record Keeping for Maintenance

All relevant data of the RO system is strongly recommended to be recorded properly. They are
not only necessary for monitoring the performance of the system, but also valuable tools for
trouble shooting and also needed in the cases of warranty claims.
The RO system performance depends for the most part on the proper operation of the
pretreatment and thus the operating characteristics of the pretreatment equipment for the
following items should be recorded and maintained.

1) Pre-treatment System
· Discharge pressure of any well or booster pumps and pressure drop of all filters including
sand filter, multi media filter, and activated carbon filter. The data of the pressure drop may
indicate when a backwash for the filters is needed
· Total residual chlorine concentration in the RO feed
· Regeneration period of water softener if used for the removal of hardness
· Inlet and outlet pressure of microfilters and cartridge filters. An increase in the differential
pressure between inlet and outlet pressure may indicate the time for cleaning and
backwashing of the microfilters or a replacement of cartridge filters
· Silt Density Index (SDI) and turbidity of the RO feed stream. Measure SDI and the
turbidity before and after all the filters
· Consumption of acid and any other chemicals such as coagulants and scale inhibitors. Scale
inhibition can also be accomplished by pH control (usually pH 6-7), depending on the
amount of hardness in the feed water

2) RO Operating Data
The following data must be recorded frequently, preferably once per shift.
· Complete water analysis of the feed, permeate, and concentrate water and the raw water
before pre-treatment once at start-up and every week thereafter
· pH of the feed, permeate and concentrate water and temperature of the feed water
· Langelier Saturation Index (LSI) of the concentrate water from the last array (for
concentrate water < 10,000 ppm TDS)
· Stiff and Davis Saturation Index (S & DSI) of the concentrate water from the last array (for
concentrate water > 10,000 ppm TDS)
· Feed flow pressure after high pressure pump
· Feed, permeate and concentrate flow pressure of each array. Pressure drop per cartridge
 System Operation
6
6
and per array
· Permeate and concentrate flows of each array. Calculate recovery ratio to ensure that it
does not exceedthe design limit
· Conductivity/TDS of the feed, permeate, and concentrate streams for each array. The TDS
of the RO concentrate can be used along with the feed water TDS to calculate an average
concentration which can be used to calculate an average membrane salt rejection as shown
in the following equations

()

permeate TDS
Average salt rejection(%) = 1 - ×100
(feed TDS + concentrate TDS) / 2

· Calibration of all gauges and meters based on manufacturer's recommendations at least

once every three months. Important gauges include pH meters, flow meters, pressure
gauges, and conductivity meters. It is recommended that the water pH is verified weekly
andon a weekly schedule, the pH probe should be placed into buffer solutions with
particular pH values, calibrating the probe to those values. If the values are drifting every
time the pH meter is checked, the reference probe needs to be replaced or have its KCl
solution replenished

3) Maintenance Log
· Record regular maintenance
· Record mechanical failures and replacements or additions of RO devices and pre-treatment
equipment such as cartridge filters
· Record any change of membrane element locations with element serial numbers
· Record calibration history of all gauges and meters
· Record all details about cleanings of RO membranes including date, duration of cleaning,
cleaning agents and concentration, solution pH, temperature, flow rate and pressure
 6 System Operation

6-4. Data Normalization

The performance of an RO system is influenced by changes in the feed water TDS, feed
pressure, temperature and recovery ratio. Data normalization is a process to convert the actual
performance of the RO system into a form which can be compared to a given reference
performance which may be the designed performance or the measured initial performance.
A difference between the normalized data and the initial or designed performance may indicate
there are some problems in the system as shown below.
· Membrane fouling and / or scaling
· Membrane chemical damage - poor salt rejection due to a chemical change and / or
degradation in the membrane structure by excessive exposure to chlorine or extreme pH
· Mechanical failure - a broken O-ring or element glue line
· Hydraulic plugging - the presence of foulants (large size colloids) or scale sticked to the
flow channel spacing between the membrane leaves of spiral - wound elements
The problems could be identified early when the normalized data are recorded daily. Three
representative variables such as, normalized permeate flow rate, normalized differential pressure,
and normalized salt rejection are calculated from the RO operating data. Thus, the effects of the
above four problems can be directly monitored by the three variables as shown below.
· Normalized permeate flow rate
· Normalized salt rejection
· Normalized differential pressure

1) Normalized Permeate Flow rate

Normalized permeate flow rate is the most important monitoring parameter for an RO system.
Normalizing for the effects of pressure, temperature, and solute concentration on permeate flow
rate will enable the resulting flow value to reflect changes due to characteristics of the membrane,
and the integrity of the membrane elements or vessels. Thus the normalized permeate flow rate
can be used to monitor the following problems :

① The extent of fouling and scale formation on the membrane surface, causing a decrease in
the permeate flow rate
② Membrane compaction, causing a decrease in the flow rate
③ The integrity of the membrane system such as mechanical leaks in the system, causing an
increase in the flow rate
④ The extent of membrane deterioration, causing an increase in the flow rate
 System Operation
6
6
The normalized permeate flow rate can be obtained by the following equation :

ΔPi
Pi - - Ppi - πi
2 TCFi
QN = QO × ×
ΔPo TCFo
Po - - Ppo - πo
2

QN = normalized permeate flow rate.

Qo = measured (actual) permeate flow rate.
Pi = initial operating pressure,
Po = actual operating pressure
ΔPi = initial differential pressure,
ΔPo = actual differential pressure
Ppi = initial permeate pressure,
Ppo = actual permeate pressure
πi = initial osmotic pressure of the feed-concentrate,
πo = actual osmotic pressure of the feed-concentrate
TCFi = initial temperature correction factor
TCFo = actual temperature correction factor

The temperature correction factor can be obtained by the following equation :

TCF = EXP [ 3140 × (1 / 298 – 1 / (273 + T)) ] ; T ≥ 25℃

TCF = EXP [ 2750 × (1 / 298 – 1 / (273 + T)) ] ; T < 25℃

The osmotic pressure of the feed-concentrate mixture can be obtained by the following
equation :

0.0385 × Cfc (T + 273)

π=
14.25 × (1000 – Cfc / 1000)

The concentration of feed-concentrate can be obtained by the following equation :

1
ln
Cfc = Cf × 1-Y
Y
 6 System Operation

Cfc = concentration of feed-concentrate as ion in mg/l

Cf = concentration of feed as ion in mg/l
Y = recovery ratio = permeate flow rate / feed flow rate
T = temperature of feed flow

2) Normailized Salt Rejection

Salt rejection is the most widely known method of monitoring the performance of an RO
system, though any problem in the system could have been noticed and corrected sooner by
monitoring other parameters such as normalized differential pressure and permeate flow rate,
before the problem affects the salt rejection.

The normalized salt passage can be obtained by the following equation,and then normalized
salt rejection (%) can be easily calculated by substracting normalized salt passage (%) value.from
integer 100.

Normalized Salt Cpo × NDPo × TCFi × Cfci × Cfo

passage (%)) = Cfco × NDPi × TCFo × Cfco × Cfi

Cpo = actual concentration of permeate as ion in mg/l

NDPi = initial net driving pressure
NDPo = actual net driving pressure
Cfci = initial concentration of feed-concentrate as ion in mg/l
Cfco = actual concentration of feed-concentrate as ion in mg/l
Cfi = initial concentration of feed as ion in mg/l
Cfo = actual concentration of feed as ion in mg/l

The net driving pressure can be obtained by the following equation :

NDP = P - ΔP/2 - Pp - π

P = operating pressure
ΔP = differential pressure
Pp = permeate pressure
π = osmotic pressure of the feed-concentrate
 System Operation
6
6
The rate of rejection varies for each of the particular salts in the feed water and thus a variation
in the ion composition of an RO feed water will result in a change in the overall percent rejection
of the TDS. Hence, it is suggested to record an individual ion rejection in order to have a basis for
future performance comparison, when a system starts with a new membrane.
The data on the individual ion rejections is also helpful in diagnosing some system malfunction.
For example, a rejection calculated using a divalent ion such as calcium can tell a difference
between a mechanical leak in the system and membrane deterioration. Mechanical damages in
membranes, glue lines and O-rings will result in a similar decrease in rejection for both
monovalent and divalent ions, while, in the case of membrane deterioration, the rejection decline
will be more severe for monovalent ions.

3) Normalized Differential Pressure (ΔP)

Differential pressure is the difference between the feed pressure and the concentrate or brine
pressure exiting the end of the elements. It is a measure of the pressure drop as the feed water
passes through the flow channels of all the elements in the system. At constant flow rate, an
increase in the differential pressure indicates that large colloidal particles or physical debris such
as pump shavings, inorganic scales and bio-film particulates are blocking the flow channels. The
telescoping of spiral wound elements can also cause an increase in the differential pressure,
which is a function of the permeate and concentrate flow rates. Since these rates may vary daily
due to variation in water temperature or some other changing parameters, the actual differential
pressure should be normalized according to the following equation to compare with the initial
differential pressure.

Normalized
(2×initial concentrate flow + initial permeate flow)1.5
Differential = actual ΔP ×
Pressure (2×concentrate flow + permeate flow)1.5

A percent change (e.g. 10%) in normalized differential pressure could suggest when to clean an RO
system.
 Cleaning and Disinfection
7
6
7-1. Introduction

Fouling of RO membrane is more or less a normal phenomenon of most RO systems since the
pre-treatment of the feed water prior to the RO membrane is basically designed to reduce fouling
substances as much as possible and technically could not remove all of them. Fortunately, with
correct cleaning frequency, most foulants can be removed from the membrane. The cleaning
frequency could be minimized as long as the pre-treatment is well maintained without upset
conditions such as an uncontrolled change in feed water composition and uncontrolled biological
contamination. Sometimes mistakes in the system operation such as too high recovery and failure
of chemical dosing systems could end up with fouling the membrane.
A fouled membrane exhibits two major performance issues, one is higher operating pressure and higher
pressure drop. As foulants collect on the surface of the membrane, they form an additional barrier layer to
transport through the membrane. Higher pressure drop is a result of the increased resistance to crossflow
caused by the layer of foulants. Pressure drop can be converted into axial pressure on the membrane
(especially membrane module). High axial pressure caused by membrane fouling can cause membrane and
membrane module to fail. Cleaning the fouled membranes can be accomplished by suitable
cleaning agents at alkaline (up to pH12) and acidic (pH2) conditions because CSM RO
membranes are stable at the pH conditions and at an elevated temperature (45℃). Many foulants,
particularly clay-type soils, can compact with time as the foulant layer increases in depth. As the
foulants compact, it will become more difficult to remove them during cleaning. Thus the time of
cleaning must not be delayed too long.
 7 Cleaning and Disinfection

7-2. Timing for Cleaning

Elements should be cleaned immediately when one of the following symptoms is detected :
· Loss of 10 to 15% in normalized permeate flow rate
· Decrease of 0.5% in salt rejection
The differential pressure (feed pressure - concentrate pressure) ΔP increases by 15% from the
reference conditions (initial performance established during the first24 to 48 hours of operation).
Ideally, element cleaning is planned when any performance deteriorations by 10% and should be
completed prior to the time performance changes by 15%. Missing the proper timing for elements
cleaning can lead to irreversible fouling and/or scaling of the membrane.
It should be noted that it is important to normalize flow and salt content of the permeate
according to the normalization procedure described in Section 6.
 Cleaning and Disinfection
7
6
7-3. Cleaning Tank and Other Equipments

The mixing tank for cleaning agents should be made of polypropylene or FRP which is resistant
to pH in the range of 1 to 13. The cleaning agents exhibit enhanced cleaning efficiency at an
elevated temperature. (e.g. 35-40℃). The cleaning solution temperature should not be below 15℃
at which the cleaning efficiency is very retarded. Cooling may also be required to avoid
overheating, so the heating and cooling equipments may be necessary to control the temperature
of the cleaning solution adequately.
The size of the tank needs to be large enough to contain the volume of the cleaning solution
approximately equivalent to the volume of the pressure vessel plus the volume of the feed and
return pipes. If it is difficult to calculate the exact volume of the pipes, it can be assumed to be
about 20% of the vessel volume. Appropriate pump, valves, flow meters, and pressure gauge
should be installed to control the flow. Especially, pump capacity should be determined enough to
handle the maximum recirculation flow rate for the largest stage (generally, 1st stage) in the entire
RO skid. Catridge filters should be changed prior to conducting cleaning with every new cleaning
solution.
 7 Cleaning and Disinfection

7-4. Cleaning Procedure

1. Fill the cleaning solution tank with RO permeate water. The volume of cleaning solution
should be sufficient to fill all the pressure vessels and pipe lines. Add the calculated amount
of the cleaning chemicals to the tank. Use a mixer or recirculate the solution with the
transfer pump to ensure that all chemicals are completely dissolved and well-mixed before
circulating the solution to the RO elements.
2. Drain most of the water from the RO system to prevent the dilution of the cleaning
solution by remaining process water in the RO system.
3. Heat the solution to the temperature recommended by the manufacturer to improve
cleaning effectiveness.
3. Pump the preheated cleaning solution to the vessel at conditions of low flow rate (about
half of that shown in Table 1) and low pressure to displace the process water remaining in
the vessel. Dump the displaced water until the presence of the cleaning solution is evident
in the RO concentrate system or in the return pipe, indicated by the pH and temperature of
the cleaning solution. Adjust flow rate and pressure according to the Table 1. Open the RO
concentrate throttling valve completely to minimize operating pressure during cleaning.
Use only enough pressure to recirculate the cleaning solution without permeate coming out.

Table 1 : Feed Flow Rate of Cleaning Solution per Pressure Vessel during Recirculation

Element Diameter RecommendedFeed Flow Rate

(in) (gpm) (m3/h)

2.5 3-5 0.7-1.2

4 8-10 1.8-2.3

8 30-45 6.0-10.2

5. Recycle the concentrate to the cleaning solution tank until the desired temperature is
maintained throughout the system. Observe any increase in the turbidity to judge efficiency
of the cleaning solution, especially in the case of an alkaline cleaning solution or a
detergent solution. If the cleaning solution becomes turbid or colored, drain the solution
and restart with a freshly prepared cleaning solution. Check the pH during acid cleaning.
The acid is gradually consumed when it dissolves inorganic precipitates on membrane
surface. If the pH increases more than 0.5 pH unit, add more acid.
6. Turn the pump off and allow the elements to soak. Sometimes a soak period of about 1
hour is sufficient. For intensive fouling, an extended soak period is beneficial; soak the
 Cleaning and Disinfection
7
6
elements for 10-15 hours. To maintain a high temperature during an extended soak period,
use a slow recirculation rate (about 10% of that shown in Table 1).
7. Circulate the cleaning solution at the rates shown in Table 1 for 30-60 minutes. The high
flow rate flushes out the foulants loosened from the membrane surfaces by the cleaning. If
the elements are heavily fouled, a flow rate which is 50% higher than shown in Table 1
may enhance cleaning efficiency. However, excessive pressure drop may be a problem at
higher flow rates. The maximum recommended drops are 1.4 bar (20PSI) per element or
4.1 bar (60PSI) per multi-element vessel. The direction of flow during cleaning must be the
same as during normal operation to avoid telescoping of the elements.
8. Drain the used cleaning solution out of the system. Analyze a sample of the used solution
to determine the types and the amount of substances (fouling materials) removed from the
membrane elements. The results could tell the degree of cleaning and the causes of fouling.
9. RO permeate or good quality water (filtered, SDI < 3), free of bacteria and chlorine,
conductivity < 10,000 s/cm is used for flushing out the residual cleaning solution. The
minimum flush out temperature is 20℃ to prevent precipitation.
10.The RO plant is started up again at normal operating conditions. However, the permeate at
the initial stage of start-up must be drained until conductivity and pH returns to normal.
And also the permeate side draining is necessary when another cleaning cycle with another
cleaning chemical is to follow. During the rinse out step, the operating parameters should
be noted to judge the cleaning efficiency and to decide if another cleaning is required. If the
system has to be shut-down after cleaning for longer than 24 hours, the elements should be
stored in a preservative solution such as 1% sodium bisulfite and 0.5% formaldehyde. For
multi-array systems, cleaning should be carried out separately for each array. This can be
accomplished either by using one cleaning pump and operating one array at a time, or using
separate cleaning pump for each array.
 7 Cleaning and Disinfection

7-5. Cleaning Chemicals

Choosing right cleaning chemicals is important since harsh and frequent cleaning will shorten
the membrane lifetime, and sometimes a inappropriate choice of cleaning chemicals can worsen
the fouling situation. The cleaning will be more effective if it is tailored to the specific fouling
situation. Therefore, the type of foulants should be determined prior to cleaning, there are helpful
ways to determine the type of foulants as shown below :
· Analyze the plant performance data
· Analyze the feed water characteristics to find potential fouling substances
· Check the results of previous cleanings which may indicate specific fouling substances
· Analyze the foulants collected with a membrane filter used for SDI measurement
· Analyze the deposits on the cartridge filter
· Inspect the inner surface of the feed line tubing and the feed end scroll of the RO element. If it
is reddish-brown, fouling by iron is possible. Biological fouling or an organic material
deposit is often slimy or gelatinous.

Table 2 lists suitable cleaning chemicals depending on the type of foulants. The acid cleaners
are to redissolve inorganic deposits including iron, while the alkaline cleaners are to remove
organic fouling including biological matter. Sulfuric acid should not be used for cleaning because
of the risk of calcium sulfate scaling formation.
For the preparation of the cleaning solutions, RO permeate is preferred, but prefiltered raw
water may be used. The raw water could have some buffering ability, so more acid or hydroxide
may be needed to reach the desired pH level, which is about 2 for acid cleaning and about 12 for
alkaline cleaning at 30℃, respectively. At 35℃, the pH limit is in the range of 2 to 11, and at 50℃
the operable pH range is 3 to 10.
 Cleaning and Disinfection
7
6
Table 2. Cleaning Chemicals for CSM Membrane

Foulant Cleaning Chemical Comments

0.2% Hydrochloric Acid. Best

Inorganic salts
0.5% Phosphoric Acid. O.K.
(CaCO3,CaSO4,BaSO4)
2.0% Citric Acid. O.K.

Metal Oxides 0.5% Phosphoric Acid. Good

(Iron) 1.0% Sodium Hydrosulfite. Good

Inorganic Colloids 0.1% Sodium Hydroxide (NaOH), 30℃ Good

(silt) 0.025 Sodium Dodecylsulfate/0.1% NaOH, 30℃ Good

Best
0.1% Sodium Hyudroxide, 30℃.
Best
Biofilms 1.0% Sodium Ethylene Diamine Tetra
When biofilm contains
Acetic Acid (Na4 EDTA) and 0.1% NaOH, 30℃
inorganic scaling
0.025%Sodium Dodecylsulfate/0.1% NaOH,
Good
Organics 30℃.
Good
0.1% Sodium Triphosphate/1% Na4 EDTA

0.1% Sodium Hydroxide, 30℃. O.K.

Silica 1.0% Sodium Ethylene Diamine Tetra O.K.
Acetic Acid (Na4 EDTA) and 0.1% NaOH, 30℃

Table 2 shows the working formula for cleaning solutions, but commercial cleaning chemicals
are frequently used in the field rather than self-made formulations. Most of the commercial
chemicals are compatible with CSM membranes in short term test. The long term compatibility
test including cleaning efficacy should be carried out. In the meantime, they can be used as long
as the membrane performance is carefully monitored to detect any long term effects at an early
stage. In any event, make sure that the commercial chemicals do not contain cationic and
nonionic surfactants, and the pH of the cleaning solution should not exceed the limits at the
specified temperature.
 7 Cleaning and Disinfection

7-6. Disinfection

If the RO system is suspected to be infected by bacteria or mold, e.g. slimy deposit or rotten
smell, a disinfection should be performed after the cleaning. The procedure is the same as for
cleaning, except the high flow pumping step.
Commonly used disinfectants are formaldehyde, hydrogen peroxide, peracetic acid, and
chlorine. Quaternary ammonium disinfectants, iodine, and phenolic compounds should not be
used because they cause flux losses.
The effective concentration of formaldehyde is in the range of 0.5 to 3%. Care should be taken
in handling this chemical since it is considered as a carcinogen.
A 400ppm peracetic acid solution (also containing 2,000 ppm of hydrogen peroxide) can be
used to disinfect the RO system. The biocidal efficacy of peracetic acid is much higher than
hydrogen peroxide. Care must be exercised not to exceed the 0.2% concentration as a sum of both
compounds. Only periodic use is recommended since continuous exposure at this concentration
may damage the membrane. When the peracetic acid is used, pH adjustment is usually not
required.
However, when hydrogen peroxide is employed alone up to 0.2% concentration, pH of the
solution is preferably adjusted to be 3. This will ensure optimal biocidal effect and minimum
damage to the membrane. If an alkaline cleaning has preceded disinfection, an acid rinsing is
recommended for both sides of the membrane. Additionally, hydrogen peroxide can attack the
membrane more aggressively at temperature above 25℃ and in the presence of transition metals
such as iron and manganese.
CSM membranes can withstand short term exposure to free chlorine (hypochlorite). However,
eventual degradation may occur after 200-1000 hours of exposure to one ppm chlorine, depending
on feed water characteristics, e.g. pH and the presence of heavy metals. Thus chlorine is not
recommended for disinfecting the membrane, but can be used in the pre-treatment prior to the RO
elements.
Disinfection using chloramine, chloramine-T, and N-chloroiso-cyanurate is not recommended,
since their effectiveness as disinfectants at low concentration (<3㎎/ℓ) is limited and the
compounds can also slowly damage the membranes.
 Replacing Element
8
6
8-1. Opening Pressure Vessel

STEP 1 RELIEVE PRESSURE

1. Shut off all sources of pressure and relieve residual pressure from the vessel, following the
system manufacturer's recommendations.

STEP 2 DISCONNECT PERMEATE PORT

1. Disconnect and remove permeate piping from the permeate port of the vessel.

STEP 3 EXAMINE END CLOSURE

1. Examine end closure of vessel for corrosion. If any is evident, proceed as follows :
A. Loosen any deposits with a small wire brush and/or a medium grade piece of Scotchbrite.
B. Flush away loosened deposits with clean water.

STEP 4 REMOVE INTERLOCK

1. Use a 1/4" hex key wrench to remove the three securing screws from the yellow plastic
securing ring. Place one of the screws aside for use on reassembly. Thread the other two
screws into the threaded jacking holes in the securing ring until they contact the bearing
plate. If excessive corrosion build-up prevents threading,
A. Apply penetrating fluid (such as WD-40® or LPS-1® ) to interfacing areas of securing ring.
(Securing ring may have become bonded to locking ring set and/or bearing plate.)
B. With a screwdriver handle or similar tool, tap the securing ring to release the bond.
C. Again attempt to remove ring by turning the jacking screws an additional 1/4 turn.
2. Grip the two (2) screws firmly and pull the securing ring towards you. Once ring is free,
carefully remove it from the bearing plate.
3. If ring cannot be freed by the method, then use the screws as jacking screws by turning 1/4
turn on each screw.
4. Once ring has started to move, repeat procedure in step 2. above.
5. Use a cushioned mallet or hammer in conjunction with a wood block to tap bearing plate
face. This should free locking ring segments.
6. Refer to Figure. 1 for locking ring segment identification.
 8 Replacing Element

Pressure Vessel of Advanced Structures Inc. for 8040 Elements

7. Rotate locking ring set so that segment A (key segment) is at the 12 o'clock position.
Segment A can now be removed. If segments will not rotate, repeat 1A, 1B, and 5.
8. Repeat step 7 above for segments B&C.
9. After removal of segments, remove all debris (corrosion products, dirt, etc.) from the vessel
ends.

BEVELED
B
ENDS

SQUARE
A
ENDS

BEVELED
C
ENDS

Figure 1.

STEP 5 REMOVE HEAD

STEP 5A REMOVAL BY HAND
1. Thread a short length (12 in.) of 1 in. I.D. pipe into the permeate port and pull the head
straight out. A sharp forceful tug may be required to start the head assembly moving.
2. If the head seal remains in the vessel bore, it should be removed at this time.

STEP 5B REMOVAL USING HEAD TOOL

1. Insert the tool into the shell with threaded rods in line with bearing plate holes.
2. Thread the rods into the bearing plate holes and turn until the knobs bottom out.
3. Grasp tool with both hands and pull straight out to remove the head. If the head will not
release from the shell,
A. Thread a 1" ID pipe approximately 2 feet long into the permeate port.
B. Carefully rock the head assembly back and forth to release the seal.
C. Once the head seal has been broken, complete removal as instructed in the Opening
Vessel section.
4. To remove the tool from the head, reinstall the head part-way into the shell so that the tool is
compressed, then unscrew the rods.
 Replacing Element
8
6
8-2. Element Removal and Loading

Do not proceed the steps for removing elements until all pressure has been relieved from the
vessel and both heads have been removed from the vessel.

STEP 1 REMOVE ELEMENT INTERFACE HARDWARE

1. Remove thrust ring from downstream (concentrate) end.
2. Remove adapters from elements at each end.

STEP 2 ELEMENT REMOVAL

1. Remove elements from vessel. Clean off any excess lubricant from vessel inside diameter
before removing elements. Elements must be removed in direction of feed flow.
2. Flush out the vessel with clean water to remove all dust and debris.
3. Examine membrane element surfaces for any imperfection which could scratch the vessel
bore. Pay particular attention to edges of anti-telescope device (ATD/brine seal carrier). If
any defects are found which cannot easily be corrected, contact the element manufacturer
for corrective action.
4. Using an approximate 50% mixture of glycerine in water, lubricate the inside of the vessel.
This may best be accomplished using a suitably sized swab soaked in the mixture. This
procedure will ease membrane element loading and reduce chance of scratching the vessel
bore.
5. Load the first element into the upstream end of the vessel. Leave a few inches of the element
projecting from the vessel to facilitate interconnection to the next element.
6. Apply a light film of a non-petrolium based lubricant, such as Parker Super O-Lube, to the
inter connector O-ring. (The amount of O-lube should be just enough to give a luster to the
O-ring. Excess O-lube must be removed to prevent possibility of element contamination).
7. Assemble the inter connector to the loaded element.
8. Line up the next element to be loaded and assemble it to the inter connector already
assembled on first element.
 8 Replacing Element

CAUTION
Maintain element alignment carefully during assembly
process. Do not allow element weight to be supported
by interconnector.

Misalignment can result in damage to interconnectors or

permeate tubes or to element outer surface.

9. Push both elements into the vessel until a few inches are projecting from the vessel. Repeat
loading process until all elements are installed.
10. When the final element is installed, push the element stack forward until the face of the
first(downstream) element is at dimension D as shown in figure 2.

Take care to avoid pushing elements too far as it can be difficult to push the stack in a reverse
direction.

D
Downstream element Position

Figure 2

VESSEL TYPE DIMENSION "D"

(see figure 2)

E8U/SP 8.40 in.

E8L/SP 8.65 in.
E8B/SP 9.15 in
E8B/SP 9.90 in.
 Replacing Element
8
6
Alternate To Measurement Method
1. Insert a clean thrust ring into downstream end of vessel.
2. Insert head assembly, without quad seal or adapter, into downstream end of vessel.
3. Place the two square ended sections of locking ring into locking ring groove (with squared
ends together, stepped side outwards.)
4. Load elements as described in 5 through 9.
5. Install upstream adapter per Step 4 and head assembly per section on "Closing Vessel."

STEP 4 INSTALL ELEMENT INTERFACE HARDWARE

1. Assemble adapter to element permeate tube at each end of vessel.
 8 Replacing Element

8-3. Closing Vessel

1. Do not proceed to closing vessel steps until elements and adapters have been installed in
vessel.
2. Do not proceed to closing vessel steps until head has been checked for correct component
assembly, and vessel has been shimmed to prevent movement of the elements if required

STEP 1 INSPECT SHELL INSIDE SURFACE

1. Inspect the vessel inside surface for any corrosion deposits or other foreign matter. If any are
found, clean the surface as follows :
A. Using a medium or finer grade of Scotchbirter and a mild soap solution, clean each end
of the vessel liner surface up to 8" in from each end of vessel.
B. Rinse away all loosened deposits from the shell inside surface using clean fresh water.
2. Inspect the vessel inside surface for scratches or other damage which could cause leaks.
Vessels that leak must be replaced.
3. Inspect feed and concentrate port seals and attachments for internal and external damage or
deterioration. Contact Advanced Structures, Inc. for guidance, if damage to the vessel's
internal surface or feed / concentrate port, seals or attachments are discovered during
inspection.

STEP 2 SHELL AND HEAD SEAL LUBRICATION

1. Work O-ring lubricant glycerin into the shell from half way up the bevel to approximately
1/2" in from the bevel. (See Figure 3)
2. Ensure the entire head seal is covered with a thin layer of O-ring lubricant, with no dirt or
dust contamination. Use only glycerin for lubrication.

STEP 3 INSTALL HEAD

STEP 3A INSTALLATION BY HAND
1. Hold the head assembly square to axis of the shell and slide it straight in until a slight
resistance is felt. Do not rotate the head assembly after insertion into the vessel as this may
cause the head seal to become detached.
 Replacing Element
8
6
Bevel
1/2"

Figure 3

2. Using both hands, firmly push the head in as far as it will go. (A sharp, forceful thrust may
be necessary to enter the head seal into the vessel bore). When the head is correctly
positioned, approximately 1/2" of the locking ring groove will be exposed.

STEP 3B INSTALLATION USING TOOL

1. Hold the head assembly square to axis of the shell and slide it straight in until a slight
resistance is felt.
2. Slide the head tool into the shell just behind the head. Do not engage threaded rods.
3. Give a sharp, forceful thrust on the head tool to enter the head into the vessel bore. Then
push into the shell as far as it will go. When the head is correctly positioned, approximately
1/2" of the locking ring groove will be exposed.
4. Remove the tool by pulling straight out. Do not rotate.(The tool can be obtained from
Advanced Structures, Inc.)

STEP 4 INSTALL INTERLOCK

1. Refer to Figure 1 for correct segment identification.
2. With the head assembly inserted into the shell, install segment B into the bottom of the shell
groove, with the stepped edge facing outwards.
3. Slide segment B counterclockwise making room to install segment C into the bottom of the
shell groove.
4. Slide segments B & C in the shell groove until the square ends meet at the 3 o'clock position.
Hold these in position while installing segment A (the key segment in the 9 o'clock
position).
5. Rotate the installed locking ring set counterclockwise until the square ends of segments B&C
are in the 12 o'clock position. (This will prevent the segments from falling out.) Locking
ring segments must be installed with stepped edge facing outwards.
6. Install the yellow securing ring with its ends flush. Align the three mounting holes in the ring
with their corresponding holes in the bearing plate. Insert the three securing screws and turn
them in about two turns.
 8 Replacing Element

7. Press the securing ring in until it seats securely on the bearing plate. Before inserting of
securing screws, it is advisable to lightly coat the screw threads with anti-seize compound,
to ease later disassembly.
8. Tighten all three mounting screws until snug. Over-tightening may cause disassembly
problems!
9. Visually inspect locking ring set to ensure it is correctly positioned between shell and
bearing plate.
10. Verify that securing ring is fully seated and held in place by securing screws.

WARNING
INTERLOCKING COMPONENTS MUST BE
CORRECTLY INSTALLED. INCORRECT
ASSEMBLY OR INSTALLATION CAN
RESULT IN EXPLOSIVE HEAD FAILURE.

STEP 5 RECONNECT PERMEATE PIPING

1. Reconnect manifold piping to the vessel permeate port. Using Teflon tape on all threaded
connections will help ensure a leak-free assembly.

Do not tighten a component into thermoplastic permeate port more than one turn past hand
tight.

STEP 6 PRE-PRESSURIZATION CHECKS

It is vitally important that the following checks be carried out before any attempt is made to
pressurize the vessel.

HEAD ASSEMBLY
Verify the following at each end of the vessel.
1. Head assembly is in good condition, with no evidence of damage or corrosion.
2. Locking ring set is properly in place and yellow securing ring is snugly held in place by the
securing screws.
 Replacing Element
8
6
MEMBRANE ELEMENTS
Verify belows
1. Elements are installed in the vessel.
2. Element adapters are installed at each end of vessel.
3. Thrust ring installed at downstream end of vessel.

PIPING CONNECTIONS
1. Check all piping connections to ensure that they will provide a leak-free seal.

STEP 7 PRESSURIZATION
1. After following the above pre-pressurization checks, pressurize vessel in accordance with the
element specifications.
2. Vessels should be filled slowly to assist trapped air to escape.
3. Vessels should be pressurized slowly to avoid damage to membrane elements and vessel
components.
 Trouble Shooting
9
6
9-1. Introduction

Any RO system malfunction manifests itself in a loss of salt rejection, a loss of permeate flow,
and an increase in differential pressure, respectively or collectively.
If one of the three parameters or combined ones deviates slowly from the normalized value, it
may indicate a normal fouling and scaling which can be removed by proper cleaning.
A fast or an immediate performance decline indicates a defect or mis-operation of the system. It
is essential in this case that the proper corrective measure is taken as early as possible because any
delay decreases the chance of restoring the system performance and also it may create other
problems.
A prerequisite for early detection of potential problems is consistent record keeping and
performance normalization, including proper calibration of all instruments. It may not be possible
to detect a problem at the early stage without accurate system performance readings
After the problem has been detected, the next step is to localize the problem and to identify the
causes of the problem. This can be done using the data of the record keeping log sheet or some
additional on-line measurements.
If the data are not sufficient to determine the causes, one or more membrane elements must be
taken out of the system and analyzed by either nondestructive or destructive methods.
 9 Trouble Shooting

9-2. Instrument Calibrations

Instrument calibrations are the first thing to check in any troubleshooting operation since wrong
instrumentation can miss a real decrease in salt rejection or cause a false alarm.

On-line TDS Meters

Accuracy of on-line TDS meters can be verified by measuring the feed and permeate water
TDS with a separate handheld meter.
The percent salt rejection calculated by the handheld meter is compared to the values obtained
from the on-line meter. The two values should be relatively close. If a discrepancy arises,
recalibrate the on-line meter fbased on the instructions of the manufacturer. Before recalibrating,
inspect the probe for the possible accumulation of foreign material that may interfere with the
accurate reading. Also, make sure that the probes are mounted properly according the
manufacturer's specifications. Improperly mounted probes may trap air bubbles or have
inadequate water circulation through the probe, which may cause false readings.

Flowmeters
Flow meter calibration is also important in an RO system, since the correct measurement of
permeate and concentrate flow rates is critical to the successful operation of the system. For
instance, the concentrate flow rate lower than the system specification may create a situation for
the accelerated fouling and scaling, and the normalized permeate flow rate is a critical parameter
to monitor any trend in fouling or membrane deterioration.
One method of calibration is to direct the water from the flow meter into a vessel and to
measure the time that it takes to fill the known volume. This should be repeated several times for
accuracy. If the variation is slight, the values (the volume divided by the time) can be averaged.
The other method is to calibrate a flow meter to a known flow meter which has been calibrated
previously.

Pressure Sensors
Besides the feed pressure and the permeate pressure, the measurement of the differential
pressure is necessary to monitor the build up of deposits on the membrane surface, or within the
element flow channels. Inaccuracy in the pressure readings will cause to miss the proper timing
for cleaning the system.
The accuracy of mechanical pressure gauges should be periodically verified using a calibrated
pressure gauge. Electronic pressure sensors have the potential for greater accuracy. However,
they are subject to sensor drift and damage resulting from vibration of the high pressure pumps.
 Trouble Shooting
9
6
To reduce the effects of vibration, the sensor can be mounted remotely and connected to the high
pressure piping with a length of stainless steel or high-pressure nylon tubing.
A calibration pressure gauge mounted with a quick-connect fitting is useful for calibrating
pressure sensors.

pH Meters and Temperature

The pH meters should be regularly calibrated using buffer solutions with a known pH. Small
variations in feed water temperature do not significantly affect the percent salt rejection. However,
The temperature readings are still important because they are used to determine the normalized
permeate flow rate. Therefore the accuracy of the feed water temperature readings should
therefore be regularly verified with an accurate thermometer.
 9 Trouble Shooting

9-3. Locating High Salt Passage (Low Salt Rejection)

A loss in salt rejection may be uniform throughout the entire system or it could be limited to the
front or to the tail end of the system. It could be a general system failure or it could be limited to
one or few individual vessels. The location of the high salt passage can be isolated by following
three steps :
· Check the individual vessel permeate TDS values.
· Probe the suspected vessel.
· Individually test the performance of each element in the vessel.
A well-designed system contains a sample port located in the permeate stream from each vessel.
Care must be taken during sampling to avoid mixing of the permeate sample with permeate from
other vessels. All permeate samples are then measured for their concentration of dissolved solids
with a TDS meter. Notice that from one array to the next the average permeate TDS usually
increases, because the second array is fed with the concentrate from the first array. To determine
the salt passage of all vessels from their permeate TDS, the TDS of the feed stream to each array
must also be measured. The salt passage is the ratio of the permeate TDS to the feed TDS.
If one pressure vessel shows a significantly higher permeate TDS than the other vessels of the
same array, then this vessel should be probed. Probing involves the insertion of a plastic tube
(approx. 1/4" for 8" module) into the full length of the permeate tube (see Figure 1).
While the RO system is operating at normal operating conditions, water is diverted from the
permeate stream of the vessel in question. A few minutes should be allowed to rinse out the
tubing and allow the RO system to equilibrate. The TDS of the permeate sample from the tubing
can then be measured with a hand-held meter and the data be recorded. This measurement should
reflect the TDS of the permeate being produced by the CSM element at that location.

Concentrate
Feed
Permeate

Element No. #1 #2 #3 #4 #5 #6
Marks TDS
Meter
 Trouble Shooting
9
6
9-4. Element Analysis

If high salt passage is found only in one or several elements in one or several pressure vessels,
then it is most likely that the element(s) could have mechanical damages such as punctures on the
membrane surface, glue line failure, a cracked centerfold of the elements and damaged O-rings
including brine seals. Damaged O-rings and brine seals can be verified easily by visual inspection
of the failed elements.
Damaged membranes and glue line failure can be visualized only by the autopsy of the
elements. Alternatively, those physical damages can be verified by a dye test along with the test
for salt rejection and flux using a small test line containing methylene blue or rhodamine B. If the
dye is detected visually or spectroscopically in the permeate, this proves there is a considerable
damage in the membrane or glue line. Then the element can be autopsied to find the causes of the
damages.
When there is a general system failure, a front end element or a tail end element should be
taken out of the vessels for examination, depending on where the problem is located.
When the problem cannot be located, an element from both ends of the system should be taken
for further analysis. Typical front end problems are due to fouling and typical tail-end problems
are orginated from scaling. Vessels/elements with these problems usually show low permeate
flow rate and sometimes a high salt passage from severe fouling and scaling due to enhanced
concentration build-up on membrane surface.
If the membranes are damaged by chemicals such as chlorine and concentrated acid, a high salt
passage along with a higher permeate flow rate would occur usually in all the elements of the first
array. If the accidental high dosage of the chemicals into the system is not corrected immediately,
the membranes of the second array would also be damaged.
 9 Trouble Shooting

9-5. Cleaning Test for Heavily Fouled Elements

The regular cleaning procedure, which is part of the system operation, usually restores a decline
in less than 15% of the permeate flow rate of the system back to the normal value. However, when
the decline in the flow rate is greater than 15% due to missing the proper cleaning time or an
accidental pre-treatment upset, it is usually difficult to recover the lost flow rate fully by the
normal cleaning. In this case, an element from the front end or tail end, depending on the location
of the problem, should be taken out for cleaning tests using more proper chemicals or more
aggressive chemicals. When the cleaning test has proven effective, the treatment can be applied to
the whole RO system.
However, cleaning may not be successful when the membrane is damaged, or when the
permeate flow of the element is below 50% of specification due to heavy fouling and scaling.
Then, the element is autopsied to examine the membrane surface, the glue line, and the fouling
deposits by the methods shown in the following section.
 Trouble Shooting
9
6
9-6. Analytical Methods for Heavily Fouled Elements

Visual Inspection
To a certain degree by a visual inspection of the autopsied element, the appearance of the
foulant will provide clues to its nature and to the difficulty in its removal. Large filter media
particles like activated carbon will be evident to the sight.
A biological foulant will have a different appearance than an inorganic scale, and smell
differently. Other mechanical problems with the element such as the broken glue line and, the
damaged feed and product water channel materials can be spotted visually.

Dissolution in Acid
If the deposits on the membrane appear to be crystalline and dissolve in an acidic solution (HCl)
of pH 3 to 4 with some gas evolution (carbon dioxide), then it is likely that the deposits are
consisted of carbonates such as CaCO3.
Sulfates or silica will only dissolve with difficulty in very low pH(e.g. pH1) solution. If the scale
is soluble in 0.1M hydrofluoric acid (HF) solution, it is possibly silica.

Dye Test
A dye such as methylene blue or rhodamine B can be added in a concentration of 0.001 to 0.005%
to the feed water of the test element on an individual element test stand. The dye will permeate the
membrane through areas of degradation and mechanical leaks in the element. The dye in the
permeate water can be detected visually or measured using a spectrophotometer.
After the dye test, the element can be autopsied to visually inspect the specific location of the
dye passage. Damaged areas may pick up more dye than unaffected areas. Chemical attack by
chlorine or high dosage acid induced hydrolysis of the membrane will tend to result in uniform
absorption of the dye.

Optical Microscopy
When the visual inspection of the autopsied element can not reveal enough information about
the nature of foulants, a high power light microscope can tell if a foulant is biological or inorganic
scale. It can also provide information about the crystalline structure of a scale formation. Using
polarized light, the microscope can tell the difference between calcium sulfate and calcium
carbonate scales, since calcium sulfate crystal has more than one refractive index to give a unique
appearance.
Sometimes it can be recognized under the microscope that a layer of scale sitting on the
membrane surface is covered by another layer of organic and biological foulants. In this case, it
 9 Trouble Shooting

may be a most effective cleaning to remove the organic and biological foulants first with an
alkaline cleaning solution, prior to attempting to remove the scale with an acidic solution.

Fourier Transform Infrared Spectroscopy (FTIR)

FTIR can provide additional information about foulants on the membrane surface. When there is
a thick layer of foulants on the membrane, it is preferred to scrape off a sample of foulants from
the membrane surface. The sample is dried and FTIR of the dried sample is run. If a suitable size
of sample cannot be collected from the membrane surface, an FTIR analysis may be performed
directly on the fouled membrane using a technique called attenuated total reflection (ATR). ATR
will give FTIR spectrum corresponding to the foulants after it subtracts FTIR spectrum of a fresh
membrane from that of the fouled membrane. Sometimes, the subtraction process does not work
well to result in a mixture of peaks corresponding to both the foulants and the membrane.
Organic and inorganic compounds have their own specific FTIR peaks which can in turn be
used to identify the compounds in a mixture. From the peak intensity, semiquantitative analysis is
also possible. If the FTIR spectrum shows peaks originated from Si-O-Si, CO3, and SO4, then it
indicates there are silica, calcium carbonate, and sulfates, respectively in the foulants. If the
spectrum exhibits peaks corresponding to C-H, -CO-, C-C, and C-N, then it indicates the foulant is
organic or biological substances. Peaks due to C-H, -CO-, C-C, and phenol groups strongly
suggest that the foulant is consisted of humic acid.

Scanning Electron Microscopy (SEM)

SEM can distinguish much smaller objects than the optical microscope. So SEM gives clear
photographs of particles as small as 0.1㎛ to identify small crystalline and amorphous inorganic
scaling matter, and also the cell structure of microorganisms. SEM could be very helpful in
analyzing the foulants in more detail.

Energy Dispersive X-ray (EDX)

During the process of SEM, X-ray radiation is emitted from the sample due to the electron
bombardment. The X-ray is low energy and includes the characteristic of the elements in the
sample. Thus EDX can identify elemental composition in the sample and even offer a
semi-quantitative analysis of the sample. It can detect very small amount of inorganic elements in
the sample and also identify carbon, nitrogen and oxygen, though less sensitively. It works the
best for analyzing an inorganic scale, but also is useful for organic sample analysis.
The method can also furnish the evidence of halogen damage of the membrane due to chlorine
oxidation by detecting the presence of chlorine chemically attached to the polyamide.
 Trouble Shooting
9
6
9-7. Causes of Element Failures and Corrective Measures

9-7-1. High Salt Passage and High Permeate Flow

Membrane Oxidation
A combination of a high salt passage (low salt rejection) accompanied by high permeate flow is
a typical symptom of the damaged membrane oxidized by oxidizing chemicals including chlorine,
bromine, and ozone. Other oxidizing chemicals such as peracetic acid, hydrogen peroxide, and
N-chloro compounds are less aggressive, but still can damage the membranes when they are
present in excessive amount or coexist with transition metals. A neutral to alkaline pH favors the
attack to the membrane. At the early stage of the oxidation, the front end elements are usually
more affected than the rest.
The oxidation damage can be made visible by a dye test on the element or on membrane
coupons after autopsy of the element. Once membrane oxidation occurred no corrective action is
possible. Consequently, all damaged elements must be replaced.

Leak
A leak from feed or concentrate to permeate through a mechanical damage of the element or of
the permeate tubing can cause high salt passage and high permeate flow. The contribution of the
leak to the permeate flow may depend on the magnitude of the damage usually caused by high
pressure and vibration. The types of the damages include leaking O-rings, cracked tubes,
telescoping, punctured membranes, and centerfold cracking.

9-7-2. High Salt Passage and Normal Permeate Flow

Leaking O-ring
Leaking O-rings can be detected by the probing technique (Section 9-3). Inspect O-rings of couplers,
adaptors, and end plugs for correct installation and aging condition. Replace old and damaged O-rings.
O-rings may leak after exposure to certain chemicals, or to mechanical stress, e.g. element movement
caused by water hammer. Sometimes, they have been improperly installed or moved out of their proper
location during element loading.

Telescoping
Telescoping is caused by excessive pressure drop from feed to concentrate. Eight inch elements
 9 Trouble Shooting

are more critical because of their greater feed side area. Make sure that a thrust ring is used with
eight inch elements to support the elements' outer diameters. Elements with smaller diameter are
supported by their permeate tubing. Severe telescoping can rupture the glue line or the membrane
itself. Telescoping damage can be identified by probing (section 9-3). The operating conditions
leading to excessive pressure drop are detailed in the section of High Differential Pressure. For an
example, when the pressure pump is started before a drained system has sufficient time to fill ,
the front end elements will be exposed to higher than normal water velocities. This can hammer
the elements to telescoping which can be prevented by opening the throttling valve slowly.

Membrane Surface Abrasion

The front-end elements are typically most affected by crystalline or sharp-edged metallic suspended
solids in the feed water. Check the incoming water for such particles. Microscopic inspection of the
membrane surface will also reveal the damage. No corrective action is possible, damaged elements
must be replaced. The pretreatment process also must be adequately changed to cope with this kind of
problem. Ensure that no particles and debris are released from the high pressure pump and piping.

Permeate Back-pressure
When the permeate pressure exceeds the feed/concentrate pressure by more than 0.3 bar (5 PSI) at
any time, the membrane may tear. The damage can be identified by probing and element leak test.
Upon autopsy of the damaged element, the outer membrane typically shows creases parallel to the
permeate tube, usually close to the outer glue line. The rupture of the membrane occurs mostly in the
edges between the feed-sided glue line, the outer glue line, and the concentrate- sided glue line.

Centerfold Cracking
The regular process for making a spiral wound element requires folding a leaf of membrane sheet in
the center (centerfold). The creased (folded) membrane can break at the centerfold under certain
conditions. Then the salt passage increases with or without an increase in the permeate flow.
Centerfold cracking may be caused by :
· Hydraulic shock during start-up (e.g. by air in the system)
· Too fast pressure increase
· Increased shear stress
· Abrasion by scaling and fouling
· Permeate back-pressure
Centerfold cracking typically occurs only after one year or more of improper operation, and only at
systems with a high frequency system shut down.
 Trouble Shooting
9
6
9-7-3. High Salt Passage and Low Permeate Flow
High salt passage combined with low permeate flow is the most commonly occurring system failure,
usually induced by colloidal fouling, metal oxide fouling and scaling.

Colloidal Fouling
Colloidal fouling occurs predominantly in the first array. The problem can be more easily located
when permeate flow meters have been installed in each array separately. SDI should be checked more
frequently to identify the pre-treatment upset.
Inspect SDI filters and cartridge filters for deposits. Clean the elements according to the cleaning
procedure and the types of foulants, and correct the pre-treatment process accordingly.

Metal Oxide Fouling

Metal oxide fouling also occurs predominantly in the first array. Check feed water for levels of iron
and aluminum. Check the materials of construction upstream of the membranes. Improper construction
materials may undergo corrosion to shed iron in the feed water. Inspect SDI filters and cartridge filters
for deposits. Clean the membranes with an acidic cleaning solution. Correct the pre-treatment and / or
material selection.

Scaling
Scaling is originated from the precipitation and deposition of sparingly soluble salts onto
membrane surface. Scaling generally starts in the last array, and then gradually moving to the
upstream arrays. Analyze the concentrate for levels of calcium, barium, strontium, sulfate,
fluoride, silicate, pH and LSI (S & DSI for sea water). Try to calculate the mass balance for those
salts, analyzing also feed water and permeate. Scaling occurs slowly because of the low
concentrations involved except CaCO3.
The crystalline structure of the deposits can be observed under the microscope. The type of
scaling is identified by a chemical analysis or X-ray analysis. Cleaning with acid and/or an
alkaline EDTA solution with subsequent analysis of the spent solution may also help to identify
the type of scalant. In the case of carbonate scaling, adjust the pH of the pre-treatment. For the
other salts, either use an appropriate scale inhibitor or other suitable pre-treatment techniques, or
lower the recovery. Make sure that the design recovery is not exceeded at any time of operation.

9-7-4. Low Permeate Flow and Normal Salt Passage

Biofouling
Biofouling of the membrane occurs predominantly at the front end of the system and affects
permeate flow, feed flow, feed pressure, differential pressure, and salt passage in the way as shown
 9 Trouble Shooting

below :
· Permeate flow decreases when operated at constant feed pressure and recovery.
· Feed flow decreases when operated at constant feed pressure and recovery.
· Feed pressure has to be increased if the permeate flow is maintained at constant recovery. Increasing
the feed pressure will invoke a worse situation, since it increases the fouling, making it more
difficult to clean later.
· Differential pressure increases sharply when the bacterial fouling is massive or when it is combined
with silt fouling.
· Since pressure drop across the pressure vessels is a sensitive indicator of fouling, installing pressure
monitoring devices is strongly recommended for each array.
· Salt passage is normal at the beginning, but may increase when fouling becomes massive.
· High counts of microorganisms in water samples from the feed, concentrate, or permeate stream
indicate the beginning or the presence of biofouling.
· Corrective measures require disinfection of the whole system including pre-treatment , and
optimization of the pre-treatment system to cope with the microorganism growth in the raw water.
· An incomplete cleaning and disinfection will result in rapid re-growth of the micro- organisms.

Aged Preservation Solution

Elements of RO systems preserved in a bisulfite solution can also become biologically fouled, if the
preservation solution is too old, too warm, or oxidized by oxygen. An alkaline cleaning usually helps to
restore the permeate flow.

Incomplete Wetting
Elements that have been allowed to dry out, usually give a very low permeate flow with a normal salt
passage. The lost permeate flow may be recovered by soaking the elements in a 50:50 mixture of
alcohol and water for one or two hours followed by soaking in water.

9-7-5. Low Permeate Flow and Low Salt Passage

Compaction
Membrane compaction usually results in low permeate flow and low permeate salt passage
(increased salt rejection). CSM membrane does not undergo considerable compaction at normal
operation, but significant compaction may occur at high feed pressure (see section 5 of System Design),
high water temperature (> 45℃) and water hammer.
The water hammer can occur when the high pressure pump starts with air in the system and full
opening of the throttle valve.
The compaction can also induce intrusions of the membrane into the permeate channel spacer fabric,
 Trouble Shooting
9
6
which are visible. Thus, the permeate flow is not only restricted by the compaction of the polyamide or
the polysulfone layer, but also by the reduced cross-section of the permeate spacer that is available for
permeate flow. Damaged elements should be replaced, or the flux loss of system must be compensated
by the addition of new elements. New elements should be located properly in the system to avoid
uneven flow distribution and recovery of the individual vessels.

Organic Fouling
Organic matter in the feed water can deposit on the membrane surface to cause flux loss, mainly in
the first array. The deposited organic layer could act as an additional barrier for dissolved solutes, or
plug pinholes of the membrane, to increase salt rejection.
Organics with hydrophobic characters or cationic groups can produce such an effect. Examples are
hydrocarbons, cationic polyelectrolytes, cationic surfactants, nonionic surfactants, and humic acids.
Analyze the incoming water for oil and organic matter, and check the SDI filter and the cartridge
filter for organic deposits. Conduct SDI and TOC measurements on a more frequent basis. Improve the
pre-treatment accordingly.
An oil fouling can be removed with an alkaline cleaning agent,. Cationic polyelectrolytes may be
cleaned off at an acidic pH, if it is not a precipitation product with other compounds, e.g., antiscalants.
Cleaning with alcohol has also proven to be effective in removing adsorbed organic films.

9-7-6 High Differential Pressure

High differential pressure, also called pressure drop from feed to concentrate, generates a high force
pushing the feed side of the element in flow direction. This force impacts on the permeate tubes and the
fiberglass shells of the elements in the same vessel. The stress on the last element in the vessel is the
highest since it has to bear the sum of the forces from the pressure drops of all prior elements.
The upper limit of the differential pressure per multi-element vessel is 4.1 bar (60 PSI), per single
element 1.4 bar (20 PSI). When these limits are exceeded, even for a very short time, the elements
might be mechanically damaged to result in telescoping and/or breaking the fiberglass shell. This type
of damage may not disturb the membrane performance temporarily, but eventually cause flux loss or
high salt passage.
An increase in differential pressure at constant flow rates usually arises from the accumulation of
debris, foulants and scale within the element flow channels (feed spacer). It usually decreases the
permeate flow. An excessive increase in differential pressure can occur from operating mistakes such
as exceeding the recommended feed flow (section 5, System Design), and building up the feed
pressure too fast during start-up (water hammer).
Water hammer, a hydraulic shock to the membrane element, can also happen when the system is
started up before all air has been flushed out. This could be the case at initial start-up or at restart-ups,
after the system has been allowed to drain. Ensure that the pressure vessels are not under vacuum when
 9 Trouble Shooting

the plant is shut down (e.g. by installation of a vacuum breaker). In starting up a partially empty RO
system, the pump may behave as if it had little or no back-pressure. It will suck water at great velocities,
thus hammering the elements. Also the high pressure pump can be damaged by cavitation.
The feed-to-concentrate differential pressure is a measure of the resistance to the hydraulic flow of
water through the system. It is very dependent on the flow rates through the element flow channels, and
on the water temperature. It is therefore suggested that the permeate and concentrate flow rates be
maintained as constant as possible in order to notice and monitor any element plugging which causing
an increase in differential pressure.
The knowledge of the extent and the location of the differential pressure increase provide a valuable
tool to identify the cause(s) of a problem. Therefore it is useful to monitor the differential pressure
across each array as well as the overall feed-to-concentrate differential pressure. Some of the common
causes and prevention of high differential pressure are discussed below.

Failure of Cartridge Filters

When cartridge filters are loosely installed in the housing or connected without using inter
connectors, they can shed debris and particles to block the flow channels in the front end elements.
Sometimes cartridge filters deteriorates while in operation due to hydraulic shock or the presence of
incompatible materials.

Media Filter Breakthrough

Fines from multimedia, carbon, weak acid cation exchange resin, or diatomaceous earth filters may
set loose and enter into the RO feed water. Sometimes some of the coagulated colloids can pass
through the channeling of the filters when the filters are not regularly back-washed to result in caking
and channeling of the filters. The channeling could occur in very old filters. Cartridge filters should
catch most of the larger particles. Smaller particles can pass through a five micron nominally rated
cartridge filter to plug the lead elements.
Chemical cleaning is difficult. It can be tried to rinse out the deposits with detergents. Separate
single-element cleaning is recommended to avoid the transport of removed particles into other
elements. Diatomaceous earth filters should be taken off-line when such problems are encountered.
Soft carbons made from coal should be replaced by coconut-shell-based carbons with a hardness rating
of 95 or better. New media should be sufficiently back-washed to remove fines before the bed is put
into service.

Pump Impeller Deterioration

Most of the multistage centrifugal pumps employ at least one plastic impeller. When a pump
problem such as misalignment of the pump shaft develops, the impellers have been known to
deteriorate and throw off small plastic shavings. The shavings can enter and physically plug the
 Trouble Shooting
9
6
lead-end RO elements.
Many high pressure pumps are equipped with an optional discharge screen. This screen will catch
most of the shavings. Pump discharge screens should be checked regularly for shavings or other debris.
The screen may be cleaned or replaced. As part of a routine maintenance schedule, monitoring the
discharge pressure of the pumps prior to any control valves is suggested. If not enough pressure, it may
be deteriorating.

Scaling
Scaling can cause the tail-end differential pressure to increase. Make sure that scale control is
properly taken into account (see Section 4-3), and clean the membranes with the appropriate chemicals
(see Section 7-5). Ensure that the designed system recovery will not be exceeded.

Brine Seal Damage

Brine seals can be damaged or turned over during installation or due to hydraulic impacts. A certain
amount of feed water will flow through the chasm in the damaged seals to bypass around the element,
resulting in less flow and velocity through the element. This will cause to exceed the limit for
maximum element recovery to increase the potential for fouling and scaling.
As a fouled element in the multi-element pressure vessels becomes more plugged, there is a greater
chance for the downstream elements to become fouled due to insufficient concentrate flow rates within
that vessel.
The brine seal damage causing an increase in differential pressure could happen randomly in any
pressure vessel. Early detection for the increase in differential pressure is important for an easy
correction of system malfunctions.

Biological Fouling
Biological fouling is typically associated with marked increase of the differential pressure at the lead
end of RO system. Biofilms are gelatinous and quite thick, thus creating a high flow resistance.
Corrective measures have been described in Section 9-7-4. It is important to frequently clean out the
microbial growth and disinfect the system. It is also suggested that water samplesare taken and
analyzed on a regular basis from the feed, permeate, and concentrate streams to estimate the
microorganism occurence.

Precipitated Antiscalants
When polymeric organic anti-scalants come into contact with multivalent cations like aluminium or
residual cationic polymeric flocculants which can heavily foul the lead elements, repeated applications
of an alkaline EDTA solution may clean the fouled elements with some difficulties.