Integrated Membrane System for Low Fouling RO

Desalting of Municipal Wastewater

Craig Bartels, Ph.D., Hydranautics, Oceanside, CA
Rich Franks, Hydranautics, Oceanside, CA
Ryan Furukawa, Hydranautics, Oceanside, CA
Pravin Murkute, Hydranautics, Oceanside, CA
Uri Papukchiev, Hydranautics, Oceanside, CA

August, 2004
 Legal Notice

This report was prepared as a result of work sponsored by the California Energy
Commission (Commission, Energy Commission). It does not necessarily
represent the views of the Commission, its employees, or the State of California.
The Commission, the State of California, its employees, contractors, and
subcontractors make no warranty, express or implied, and assume no legal
liability for the information in this report; nor does any party represent that the use
of this information will not infringe upon privately owned rights. This report has
not been approved or disapproved by the Commission nor has the Commission
passed upon the accuracy or adequacy of this information in this report.

2
 Acknowledgement
Hydranautics would like to express its appreciation to the City of Oceanside for their
cooperation and support in providing space, supplies and personnel to assist in the
successful completion of this project.

3
 Table of Contents

Preface................................................................................................................................. 6
Executive Summary ............................................................................................................ 7
Abstract ............................................................................................................................... 8
Introduction......................................................................................................................... 9
Background and Overview ......................................................................................... 9
Project Objectives ..................................................................................................... 10
Report Organization.................................................................................................. 10
Project Approach .............................................................................................................. 11
Source Water................................................................................................................. 11
Ultrafiltration Pilot Unit................................................................................................ 12
Pretreatment .............................................................................................................. 13
Performance Assessment .......................................................................................... 13
Backwashing ............................................................................................................. 14
Cleaning .................................................................................................................... 14
Microfiltration Pilot Unit .............................................................................................. 15
Pretreatment .............................................................................................................. 17
Performance Assessment .......................................................................................... 18
Backwashing ............................................................................................................. 18
Aeration..................................................................................................................... 19
Cleaning .................................................................................................................... 19
Reverse Osmosis Pilot Unit .......................................................................................... 19
Pretreatment .............................................................................................................. 21
Performance Assessment .......................................................................................... 21
Cleaning .................................................................................................................... 21
Project Outcomes .............................................................................................................. 23
Ultrafiltration ................................................................................................................ 23
Performance .............................................................................................................. 23
Backwash Frequency ................................................................................................ 24
Cleaning Optimization .............................................................................................. 24
Chemical Pretreatment.............................................................................................. 25
Feed/Filtrate Quality ................................................................................................. 26
Microfiltration............................................................................................................... 28
Performance .............................................................................................................. 28
System Process Optimization ................................................................................... 32
Chemical Pretreatment.............................................................................................. 33
Feed/Filtrate Water Quality ...................................................................................... 34
Off-line Chemical Cleaning Results. ........................................................................ 38
Reverse Osmosis........................................................................................................... 41
UF pretreatment ........................................................................................................ 41
MF Pretreatment ....................................................................................................... 44
Cleaning .................................................................................................................... 47
UF vs MF pretreatment............................................................................................. 48
Conclusions and Recommendations ................................................................................. 52

4
 Ultrafiltration ................................................................................................................ 52
Microfiltration............................................................................................................... 52
Reverse Osmosis........................................................................................................... 53
Commercialization Potential......................................................................................... 54
Recommendations......................................................................................................... 54
Benefits to California.................................................................................................... 55
References......................................................................................................................... 56

5
Preface
The Public Interest Energy Research (PIER) Program supports public interest
energy research and development that will help improve the quality of life in
California by bringing environmentally safe, affordable, and reliable energy
services and products to the marketplace.

The PIER Program, managed by the California Energy Commission

(Commission), annually awards up to $62 million to conduct the promising public
interest energy research by partnering with Research, Development, and
Demonstration (RD&D) organizations, including individuals, businesses, utilities,
and public or private research institutions.

PIER funding efforts are focused on the following six RD&D program areas:

• Buildings End-Use Energy Efficiency

• Industrial/Agricultural/Water End-Use Energy Efficiency
• Renewable Energy
• Environmentally-Preferred Advanced Generation
• Energy-Related Environmental Research
• Strategic Energy Research.

What follows is the final report for the Desalination Research and Innovation
Partnership, conducted by the City of Oceanside in cooperation with
Hydranautics. The report is entitled Integrated Membrane System for Low
Fouling RO Desalting of Municipal Wastewater. This project contributes to the
Industrial/Agricultural/Water End-Use Energy Efficiency program.

6
Executive Summary
In recent years, California’s rapidly growing population has increased the volume of
municipal waste as well as the demand for fresh water. An increasing number of
municipalities are considering recycling after treating to tertiary and higher levels to
minimize environmental impact and supplement fresh water supplies. But treating
wastewater for high purity industrial or indirect potable reuse requires treatment
involving membrane filtration. The availability of a range of membrane products affords
the opportunity to use two or more membrane technologies synergistically to achieve
stringent treatment objectives and enhance performance of the overall system. A system
which uses more than one type of membrane for a treatment scheme is called a Multiple
Integrated Membrane System (MIMS).

The MIMS consists of microfiltration (MF) or Ultrafiltration (UF) membrane

pretreatment followed by low fouling or energy saving reverse osmosis (RO) membranes.
MF/UF membrane processes provide an effective means of meeting increasing water
demands because of their ability to remove solids as well as microbial contaminants
(including viruses) by size exclusion. Reveres osmosis membranes, being tighter than
MF/UF membranes, reduce the concentration of dissolved solids and provide a secondary
barrier to microbial and other contaminants that may pass the membrane pretreatment.

This study demonstrates the use of a MIMS consisting of UF/MF followed by energy
saving and low fouling RO membranes for the efficient and economical treatment of
secondary effluent to higher than tertiary levels over extended operating periods.

7
Abstract
This study uses a Multiple Integrated Membrane System (MIMS) consisting of UF/MF
followed by energy saving or low fouling RO membranes to demonstrate the efficient
and economical treatment of secondary effluent under on-site operating conditions over
extended operating periods. The MIMS ran from September 2002 to September 2004 at
La Salina WWTP in Oceanside CA. The UF membrane as pretreatment to the RO
successfully ran at 32 gfd and 87% recovery with a period of 30 days between cleanings.
The MF as pretreatment to the RO ran at 12 gfd and 95% recovery with a period of 60
days between cleanings. Both UF and MF produced high quality filtrate that allowed the
RO to run at 11 gfd and 75% recovery.

The RO successfully reduced the level of dissolved solids from 1100 ppm in the feed to
20 ppm in the permeate. Energy saving membranes operated in parallel with low fouling
membrane showed no greater propensity to fouling. Elements constructed with 26 mil
brine spacer were run in parallel with elements constructed of 31 mil brine spacer. The
smaller brine spacer elements showed no greater propensity to fouling than the thicker
brine spacer elements.

8
 Introduction
Background and Overview
Secondary treated municipal effluents in Southern California have historically been
discharged to large water bodies with manageable environmental impact. At the same
time, a majority of the region’s potable water supply has been imported from Northern
California and the Colorado River. In recent years, a rapidly growing population has
increased the volume of municipal waste as well as the demand for fresh water. An
increasing number of municipalities within California are considering recycling after
treating to tertiary and higher levels to minimize environmental impact and supplement
fresh water supplies (Alexander 2003). Tertiary treated secondary effluents, involving
media filters to remove particulate matter followed by disinfection by chlorination, have
been used for agricultural and industrial purposes. But treating wastewater for high purity
industrial or indirect potable reuse requires further treatment involving membrane
filtration. The availability of a range of membrane products affords the opportunity to
use two or more membrane technologies synergistically to achieve stringent treatment
objectives and enhance performance of the overall system. A system which uses more
than one type of membrane for a treatment scheme is called a Multiple Integrated
Membrane System (MIMS).

The MIMS consists of MF or UF membrane pretreatment followed by low fouling or

energy saving RO membranes. MF/UF membrane processes provide an effective means
of meeting increasing water demands because of their ability to remove solids as well as
microbial contaminants, including viruses by size exclusion (DeCarolis, 2001). For
example, some tests on MF membranes treating secondary effluent demonstrate at least a
five log removal rate on of all coliform bacteria (Schimmoller 2001, Alonso 2001). The
removal rate depends largely on the actual concentration of viable bacteria on the feed
side of the membrane, along with some processing factors such as: degree of fouling,
vacuum pressure, pretreatment chemicals used, and actual moment of sampling. Removal
of total suspended solids is complete for UF/MF membranes. With typical feed solids
concentrations below 100ppm, the filtrate TSS is always below detectable limits
(Schimmoller 2001, Alonso 2001).

Reveres Osmosis membranes, being tighter than even MF/UF membranes, have been
successfully used to reduce the dissolved solids concentration in municipal effluents
(Abi-Samra, 2002). The RO membranes also act as a secondary barrier to viruses and
bacteria.

Presently, two large scale sites in Southern California employ MIMS to reclaim
municipal waste water for industrial and indirect potable reuse. The West Basin
Municipal Water District operates the West Basin Recycling Center in El Segundo where
MIMS are used to treat 12.5 MGD of municipal waste water from the City of Los
Angeles’ Hyperion WWTP for industrial reuse. Nearby, the Orange County Water
District’s Ground Water Replenishment System (GWRS) is currently using MIMS to
treat 5 MGD of municipal waste to recharge its groundwater supply which has been taxed

9
by years of drought and a growing population. The GWRS is expected to reach its full
capacity of 70 MGD by 2007.

Project Objectives
This study uses a MIMS consisting of UF/MF followed by energy saving or low fouling
RO membranes. The MIMS ran from September 2002 to September 2004 at La Salina
WWTP in Oceanside CA. The primary purpose of this study is to demonstrate a MIMS
for the efficient and economical treatment of secondary effluent under on-site operating
conditions over extended operating periods and assess the effect of different operating
parameters. Specifically, this study will:

1) Evaluate the performance of a capillary ultrafiltration membrane as pretreatment to

reverse osmosis by monitoring the effect of variables such as pretreatment, feed
quality/composition, feed temperature, and operating flux and determining the
optimum operating conditions, cleaning frequency, cleaning conditions, and
compatibility of the feed and membrane material;
2) Evaluate the performance of a submersible microfiltration membrane as pretreatment
to reverse osmosis by monitoring the effect of variables such as pretreatment, feed
quality/composition, feed temperature, and operating flux and determining the
optimum operating conditions, cleaning frequency, cleaning conditions, and
compatibility of the feed and membrane material;
3) Evaluate the performance of a reverse osmosis system treating secondary effluent by
monitoring the effect of variables such as MF vs UF pretreatment, feed
quality/composition, feed temperature, and operating flux and determining the
optimum operating conditions, cleaning frequency, cleaning conditions, membrane
material (low fouling vs energy saving) and element’s brine spacer thickness on
propensity to fouling;
4) Determine the ability of an integrated membrane system consisting of capillary
ultrafiltration membrane pretreatment or submersible microfiltration membrane
pretreatment followed by reverse osmosis to treat municipal secondary effluent and
produce treated water of a quality to meet the pertinent regulations for discharge or
reuse;
5) Use the data collected from the MIMS pilot to evaluate the design and economics of a
large-scale treatment system.

Report Organization
Throughout the “Project Approach” and “Project Outcomes” section of the paper, each of
the three pilots will be discussed separately. The UF will be presented first followed by
the MF, then the RO. Because the RO treated UF and MF filtrate in two different tests,
the discussion of the RO will itself be divided into two sections. The conclusion of the
paper will include a discussion of the whole system as a multiple integrated membrane
system.

10
 Project Approach
Source Water
The MIMS treated secondary effluent from the LaSalina Waste Water Treatment Plant in
Oceanside, California. LaSalina WWTP treats approximately 3.5 million gallons per day
of municipal waste from the city of Oceanside using: bar screen, grit chambers, load
equalization, primary clarification, and secondary aeration/clarification. Effluent from
the secondary clarifiers was supplied as feed to the MIMS. Feed water characteristics for
the secondary effluent, including individual ion composition, are found in Table 1.

11
Table 1. Raw water analysis, La Salina Wastewater treatment plant, Oceanside CA.

Category Parameter Units Low High

General pH 7 7.5
Temperature C 20 30
Dissolved
TDS ppm
Solids 900 1370
Conductivity us 1750 2600
Ca ppm 60 68.9
Mg ppm 36 37
Na ppm 197 226
NH4 ppm 15.2 38
PO4 ppm 3.54 21
Alk (ppm CaCO3) ppm 137 260
SO4 ppm 210 254
Cl ppm 290 350
NO3 ppm 16.2 38
SiO2 ppm 12.6 15.9
Solids TSS ppm 4.8 9.4
Turbidity (NTU) NTU 2.5 100
Organic UV(254) cm (-1) 0.16 0.16
TOC ppm 8 16
BOD5 ppm 2.5 15
COD ppm 30 110
Count/
Microbial Total Coliforms 160,000 500,000
100ml

Ultrafiltration Pilot Unit

The HYDRAcap UF pilot is a Self-Encapsulated Capillary Ultrafiltration system with

two 60” modules. Each module contains 13,000 capillary fibers. The system’s total
active membrane area is 1000 sq. ft. The membrane material, polyethersulfone, tolerates
moderate concentrations of free chlorine and can withstand a pH range of 2-13. Figure 1
below shows the configuration of the UF pilot at LaSalina.

HYDRAcap UF is an inside out hollow fiber system. Feed water enters the inside of the
fibers under pressure and is forced out through the fiber wall. The UF system operates in
dead-end mode in which all feed is converted to filtrate during the period of filtration.
Table 2 shows the minimum and maximum operating parameters tested during this
investigation. Operating parameters such as flux, transmembrane pressure (TMP), feed
and filtrate turbidity, pH, temperature, and processing time were recorded on a regular
basis using manual spreadsheets.

12
 Feed pump
UF filtrate
Secondary effluent RO feed

Backwash
Backwash pump

Figure 1. Configuration of HYDRAcap UF system treating secondary effluent for

pretreatment to RO.

Table 2. General Operating Conditions for the HYDRAcap UF pilot unit

Operating Unit Value (min-max)

Parameter
Feed flow gpm 20-40
Filtrate flow gpm 20-40
Backwash flow gpm 100-130
Filtrate flux gfd 32-46
Feed pressure psig 5-25
TMP psig 4-25
Recovery % 84-90

Pretreatment
Both ferric chloride and sodium hypochlorite were used as a chemical pretreatment to the
UF system. Ferric chloride was used as a coagulant in the feed water to enhance the
removal of colloids and organics. The sodium hypochlorite was used to prevent
biological growth in the system. Chemical injection occurred 45 ft before the UF system
(22 second contact time) to allow for sufficient chloramine formation and pin floc
formation.

Performance Assessment
The productivity of the system was monitored using temperature corrected specific flux
(TCSF) normalized to a temperature of 20ºC (equation 1).

13
 Flux( gfd )
TCSF = ⋅ (− 0.032 ⋅ exp(T − 20 )) (1)
TMP( psi )

Flux, TMP and temperature all play important roles in determining the value of the
TCSF. The flux is the amount of water flowing through each square foot of membrane
area per day (gfd). The formula for flux is given in equation 2.

Flow( gpm)
Flux = (2)
1440 ⋅ MembraneArea( ft 2 )

The TMP in equation 1 is equal to the amount of pressure required to push water through
the membrane, or the pressure difference across the membrane. The formula for TMP
appears in equation 3 where: Pf, Pc, and Pp are feed, concentrate, and filtrate stream
pressures, respectively.

 Pf + Pc 
TMP =   − Pp (3)
 2 

As the membrane fouls over time, the transmembrane pressure (TMP) begins to rise. As
TMP rises, TCSF decreases. All parameters needed to calculate TCSF were collected
twice daily. The data was collected two minutes after a backwash cycle to ensure
consistency.

The turbidity of the feed and filtrate water was recorded twice daily and used to estimate
the removal of Total Suspended Solids (TSS). TOC, UV254, and total coliforms were also
tested to determine removal of organics and microorganisms.

Backwashing
Backwashing is a regularly scheduled event intended to remove the foulant layer that
accumulates on the membrane surface during filtration. The frequency of backwashing
was adjusted between 25 and 35 minutes to optimize the performance of the system. The
multiple step backwash parameters were controlled by the programmable logic controller
(PLC). The backwashing sequence consisted of 5 steps: Air Enhanced Backwash (AEB),
backwash bottom, backwash top, backwash screen filter, and final rinse.

The AEB is a patented technique which pressurizes the inner fibers using air. When
pressurized, the fibers expand slightly to dislodge the foulant layer which accumulates
during the filtration process. Following the AEB, the backwash steps remove the
foulants from the system by pumping filtrate water from the filtrate side to the feed side
of the fiber. The five steps of the backwash sequence last a total of 60 to 90 seconds after
which the filtration process is resumed.

Cleaning
Over time, the foulant layer accumulates on the outer surface of the membrane to a point
where the backwashing sequence is no longer effective. A cleaning is then required to

14
fully restore membrane permeability. Cleanings involved recirculation of a high or low
pH solution through the system. The effects of each cleaning were studied to determine
the optimal cleaning procedure for this application.

Microfiltration Pilot Unit

The HYDRAsubTM MF is an open submerged type membrane system containing 12

membrane bundles per module. Each bundle holds 16,800 membrane fibers with 0.31mm
external diameter and a nominal pore size of 0.2 microns. The module’s total active
membrane area is 2690 sq. ft. The membrane material is polypropylene. Figure 2 below
shows the HYDRAsubTM module, the pilot unit, and the configuration of the MF system
at La Salina WWTP. The MF pilot used a single membrane module.

Filtrate
Removable headers
fiber bundles

Air
diffusers

HYDRAsubTM pilot unit HYDRAsubTM module

Figure 2. HYDRAsubTM MF system treating secondary effluent for pretreatment to

RO.

15
The HYDRAsubTM MF is an outside-in hollow fiber system. A vacuum draws feed water
from the outside of the fiber to the inside lumen of the fiber. A filtrate flow of 22gpm was
selected to provide sufficient feed flow for the downstream RO. A system recovery of
89-95% was maintained by creating a periodic bleed stream of 10.5 gpm during the
filtration cycle. When bleeding was utilized the volume of the bleed stream was close to
75% of the total waste stream. The other 25% of the total waste stream was released
during the backwash cycle in several short pulses with a high instantaneous flow of 55
gpm.

A summary of the specific process parameters for the MF operation are listed in Table 3.

16
Table 3. MF System Processing Parameters

Operating Unit Value

Parameter
Feed flow gpm 24-26
Filtrate flow gpm 22
Concentrate flow gpm 0-56
Operating flux gfd 11.7
Aeration air flow cfm@psi 10@10
Avg. daily air flow cfd@psi 420@10
TMP psig 3-10
Recovery % 89-95
CEB 1 ppm NaOH 4000
CEB 2 ppm SMBS 5000
Feed coagulant dose ppm FeCl3 0.5-0.9 as Fe
Feed oxidant dose ppm NaOCl 2.5 as free chlorine
CEB 1 frequency hour 0-24
CEB 2 frequency hour 0-4
Source water for CEB - RO permeate

The MF was equipped with an on-line data logging system, allowing key system
parameters such as flow, pressure and turbidity to be recorded. Other parameters -
temperature, pH and air flow, were recorded manually on a regular basis for redundancy.
The MF process is a multiple step process controlled by a number of parameters. A
human machine interface (HMI) panel provided detailed information on all parameters.
Full manual control on the pilot was available for maintenance and troubleshooting
purposes.

Pretreatment
The presence of ammonia in the feed was considered advantageous for the operation of
the MF because it allowed for chloramination. Chloramines act as a biocide against most
waterborne microorganisms and are compatible with the RO and MF membrane as well
as other materials in the system. By using a continuous in-line dosing pump, 2.5 to 3.0
ppm of free chlorine was injected in the feed stream. The free chlorine reacted with
ammonia in the feedwater to form chloramines.

To further improve MF filtrate quality and reduce fouling rate, inline coagulation using
ferric chloride (0.8 ppm as pure iron) was implemented in the feed stream. Stable fine
floc formation occurred in the membrane tank, enhancing solids settling and subsequent
disposal to drain via bleed or CEB. Both coagulant and oxidant addition had a positive
effect on system performance.

The feed screen filter used was 1.5 mm perforated cylinder (Amiad 2” T filter). The filter
was installed to protect membranes from large particles entering the membrane tank. The

17
filter was backwashed every 7.5 hours for 10 seconds. The pressure drop across the filter
was always less then 1.5 psi and averaged 0.5 psi.

Performance Assessment
The flow applied over fixed area of membrane surface (flux) is shown in equation 4
below.

Q
F= , [ 4]
1440 ∗ S
where
F - flux, [gfd]
Q - flow through the membrane, [gpm]
S - actve area of membrane, [ft 2 ]

Membrane resistance was calculated by measuring trans-membrane pressure (TMP).

Since HYDRAsubTM is vacuum driven system, vacuum pressure measured on the filtrate
line adjacent to membrane module provides a direct measure of membrane resistance.
Vacuum pressure was directly used for measuring trans membrane pressure.
System productivity was measured by using temperature compensated specific flux
(TCSF) as shown in equation 5 below. Temperature compensation was made to 20oC.

F 0.989
TCSF = * ; [5]
(
TMP 1.794 − (0.55 * T ) − 0.00076 * T 2 )
where
TCSF - temperature compensated specific flux, [gfd/psi]
F - flux, [gfd]
TMP - trans membrane pressure or vacuum pressure, [psi]
T - water temperature, [ o C]

Filtrate flow, filtrate pressure, turbidity, and pH where measured with online transmitters.
For redundancy purposes many parameters where recorded once a day on a paper
spreadsheet and later transferred into electronic files. Occasionally samples from feed
and filtrate where taken for laboratory analysis to measure COD, TOC, TSS and other
water quality constituents.

Backwashing
Every 29 minutes, the membrane was backwashed for one minute with filtrate water. In
addition to these filtrate backwashes, the membrane was subjected to a chemically
enhanced backwash. Chemically enhanced backwash (CEB) was used to evaluate the
cleaning efficiency on the membrane as well as the affordability of such chemical

18
cleaning regimes compared to conducting more frequent CIP cleanings. The CEB could
be conducted with caustic or acid or a reducing agent such as sodium metabisulfite. The
reducing agent would help in removing the coagulant (iron) deposited on the membrane
surface. During the high pH CEBs, RO permeate was used as the source water to avoid
dissolved solids precipitation. A CEB consists of following steps:
• Partial or complete initial tank drain (optional);
• Rinse;
• Chemical dosing;
• Soak;
• Rinse;
• Final tank drain (optional);
Backwash and rinse flux was set to 27 gfd, while chemical dosing flux was set to 6gfd.

Aeration
To further enhance mechanical cleaning, the membranes were aerated with short air
pulses at fixed intervals from multiple coarse bubble air diffusers located beneath the
membrane bundles. The air flow was set at 10 cfm with 10 psi pressure. Two types of air
pulsation modes were used – intermittent and rapid. Intermittent aeration was used only
during filtration. Rapid air pulsation was used during filtration, backwash and rinse
routines. Later during the test, rapid aeration was also used during a portion of the soak
time.

In the intermittent air pulsation, a single pulse of air is released after a long pause. A
typical setting for intermittent air pulsation is 4 seconds for the air pulse with an interval
of 135 seconds between pulses. Rapid air pulsation consists of air pulses occurring in
rapid succession. A typical setting for rapid pulsation is 4 sec pulse followed by 4 second
pause with a total of 10 pulses occurring in succession. Air diffusers were backwashed
with water once a month to prevent clogging of the diffuser pores with solids.

Cleaning
When the TMP or vacuum reaches 8-9 psi, cleaning is required to restore the membrane
permeability and remove the deposited foulants. Before cleaning, the feed tank was first
drained to remove the concentrated feed. Tap water was added to the feed tank.
Cleaning chemicals were then added to the feed tank and the membranes were allowed to
soak for 90 minutes. After cleaning with one chemical, the feed tank was drained and
filled again. Then cleaning with second chemical was conducted. The first cleaning step
used sodium hydroxide (NaOH) with 0.25-0.40% (w) at pH 12.5. The second cleaning
step used a mix of two chemicals – citric acid (C6H8O7) 0.2% (w) and sulfuric acid
(H2SO4) 0.05% (w) at pH=2.0.

Reverse Osmosis Pilot Unit

The RO pilot unit consisted of a four stage (2:2:1:1) configuration with three elements
per vessel (Figure 3). The pilot was equipped with flowmeters, pressure gauges, and
collection ports so that the performance of the six vessels could be monitored separately.

19
The unit was equipped with one feed, one concentrate, and six permeate taps to monitor
conductivity with a hand held conductivity meter.

To compare the performance of different membranes, three different types of spiral

wound polyamide elements were installed in the unit. Each element measured 4inches in
diameter by 40 inches in length. Vessels 1 and 3 were equipped with the LFC1. The LFC
technology is characterized by a neutrally charged, hydrophilic membrane surface.
Vessels 2 and 4 contained the ESPA2, a membrane designed for high production at low
pressures. Though the ESPA2 has a greater specific flux than the LFC1, the two elements
were subject to identical operating conditions (flux and recovery) by adjusting permeate
back pressure accordingly. The last two pressure vessels, 5 and 6, contained the LFC3,
a membrane similar to that of the LFC1 but with lower specific flux and higher rejection.

During the second phase of the testing, the LFC1s in pressure vessels 1 and 3 were
replaced by a specially constructed ESPA2s made with a thicker brine spacer to compare
the effect of brine spacer thickness on fouling tendency. The standard brine spacer
thickness is 26 mil; the modified ESPA2 contained a 31 mil brine spacer. Due to the
increased brine spacer thickness, feed flow to the two parallel pressure vessels was
similar, but not identical. Permeate back pressure to the vessels was adjusted to ensure
similar flux. Table 4 lists the characteristics and the standard test performance of the
elements used in the RO pilot unit.

1 3
Pretreat 5 6
2 4

Figure 3. Configuration of the RO pilot unit at LaSalina WWTP. Each vessel

contains three elements measuring 4” x 40”.

20
Table 4. Spiral Wound Composite Polyamide Membranes Installed in the RO pilot
unit. Test conditions: 150psi, 1500ppm NaCl, 25 C, pH=7, rec = 15%.

Membrane Description Flow (GPD) Rej (%) Area (sq.ft) Vessel

Low Fouling,
LFC1 1470 99.0 85 1,3
neutral charge
ESPA2 Negative charge 1900 99.5 85 2,4
Negative charge
ESPA2-31mil 1500 99.5 70 1,3
31 mil spacer
Low Fouling,
LFC3 1275 99.6 85 5,6
neutral charge

Pretreatment
During the first phase of testing, the HYDRAcap Ultra filtration unit served as pre-
treatment to the RO. During the second phase, the HYDRAsubTM Micro filtration unit
served as pretreatment. Both the MF and UF improve water quality by removing
suspended solids. To reduce the possibility of scale formation, the RO feed was dosed
with acid in the form of H2SO4 (60%), reducing pH form 7.1 to 6.7, as well as
Hypersperce MDC150 anti-scalant (3 to 5 ppm).

Performance Assessment
Performance of the RO was assessed twice on a daily basis by recording pH, temperature,
pressures, flows, and conductivity. System flux was varied from 9 to 13 gfd. Specific
membrane flux, differential pressures, and salt passage were normalized to 25C for each
of the three element types as described by the American Society for Testing and
Materials (ASTM) method D 4516-001. Other supporting data, including individual ion
analysis, SDI, and microbial counts, were collected periodically.

Cleaning
Four different types of fouling generally pose a challenge to the successful operation of
any RO and, if not controlled, will lead to frequent cleanings. These four types of fouling
include:

• ORGANIC. The membrane surface is fouled due to the absorption of dissolved

organics. Organic fouling generally leads to a loss in membrane permeability and
a gradual increase in differential pressure.

1
American Society of Testing and Materials (ASTM), Standard Practice for Standardizing Reverse
Osmosis Performance Data. ASTM Designation D 4516-00, Volume 11.02, 2001

21
 • BIOLOGICAL. Blockage of the brine spacers may occur due to the growth of
bacterial colonies. Biological fouling leads to a marked increase in differential
pressure and a decrease in permeability. A slimy film can be detected on the
membrane and on the inside of the pressure vessel. The presence of chloramines
in the feed to the RO serves to hinder biological fouling.

• COLLOIDAL. Blockage of the brine spacers and membrane surface is caused by

colloidal matter present in the feed water. Colloidal fouling generally occurs in
the lead elements of the first stage and will lead to a gradual increase in
differential pressure as well as a gradual decrease in permeability. The presence
of membrane pretreatment to the RO serves to control colloidal fouling.

• SCALING. Sparingly soluble salts reach their saturation limits in the concentrate
stream and precipitate out of solution onto the membrane surface. Scaling
generally occurs in the tail elements of the last stage and leads to a slight decrease
in permeability and rejection. The RO in this study was run at a conservative
75% recovery to avoid the possibility of scaling. Acid and anti scalant were also
dosed in the feed stream to further reduce the chances of scaling.

The RO was considered fouled when normalized performance deviated at least 10% from
startup reference performance. After fouling, high and low pH cleanings were performed
on individual elements as well as the whole system to determine cleaning effectiveness
and the nature of foulant. Low pH cleanings of pH 2 to 4 used 2.0% (w) citric acid
(C6H8O7) to target inorganic scale (e.g. calcium carbonate, calcium sulfate, barium
sulfate, strontium sulfate) and metal oxides/hydroxides (e.g. iron, manganese, nickel,
copper, zinc), and inorganic-based colloidal material. The high pH (pH = 10-11)
cleaning of 2.0% (w) of STPP (sodium tripolyphosphate) (Na5P3O10) and 0.8% (w) Na-
EDTA (sodium salt of ethylaminediaminetetraacetic acid) was used to target calcium
sulfate scale and light to moderate levels of organic foulants of natural origin. STPP
functions as an inorganic-based chelating agent and detergent. Na-EDTA is an organic-
based chelating cleaning agent that aids in the sequestering and removal of divalent and
trivalent cations and metal ions. Cleanings were done in recirculation mode for at least
one hour at temperatures between 30 and 40 C.

22
 Project Outcomes

Ultrafiltration

• The successful performance of the capillary ultrafiltration membrane treating

secondary municipal wastewater as pretreatment to reverse osmosis was
demonstrated. The optimum operating flux was found to be 32 gfd while the
optimum backwash frequency was set at 25 minute intervals. Based on these
settings, the recovery was calculated to be 87%. At these optimal settings,
chemical cleanings could be performed every 30 days (high pH) to 60 days (low
pH).Variables such as pretreatment and operating flux where shown to have a
significant impact on performance while variations in feed water quality as
measured by turbidity were shown to have little effect on filtrate quality. SDI as a
measurement of filtrate quality varied widely.

Performance
An important factor affecting the performance of an ultrafiltration system is operating
flux. One of the main objectives of this study was to determine the optimal operating
flux which would allow the system to run continuously for 30 days without chemical
cleaning. The effect of operating flux on UF membrane performance was investigated by
conducting test runs at 32, 36, and 46 gfd. All runs were done with a 25 minute
processing time between backwash cycles and used chemical pre-dosing of 3-5 ppm total
chlorine and 3 ppm FeCl3 (as 1 ppm pure iron) . During these runs, the TMP was
monitored to ascertain when significant fouling had occurred and a cleaning was
required. The membrane was considered fouled when TMP reached 10 psi.

Figure 4 displays temperature corrected TMP at the three different fluxes. At 32 gfd, the
TMP remained stable through 17 days of operation when a system failure ended the test.
The trend after 17 days suggests the goal of 30 days could easily be achieved before the
membranes became fouled. At higher fluxes, the membrane fouled more rapidly. At 46
gfd, fouling occurred in less than two days of operation. When run at 36 gfd, the system
reached 14 days before fouling. Though an improvement over the 46 gfd run, the 36 gfd
run could not be deemed a stable run. Fouling occurred primarily due to the build up of
particulate matter on the membrane surface. Another possible source of fouling was the
high content of organics in the feed water (as seen in the feed water analysis). It has been
observed that the presence of organic constituents in source water can cause significant
productivity loss by plugging membrane pores, adsorbing to the internal matrix of the
membrane, and forming a cohesive gel on the cake layer. In addition, the organic matter
can also enhance biological activities in the membrane system, which may increase
biofouling. (DeCarolis, 2001).

23
 12

10
Temperature Corrected TMP (psi)

0
0 5 10 15 20 25 30

Process Time (days)

32 GFD 36 GFD 46 GFD

Figure 4. Effect of membrane flux on temperature corrected TMP.

Backwash Frequency

Backwash frequency is anther factor which affected the performance of the UF system.
If the backwash frequency was set too low, the recovery of the system would be high, but
the modules would foul in less than 30 days. If the backwash frequency was set too high,
then the recovery of the system would be too low to be practical. It was found that a
processing time of 25 minutes between backwashes was the optimal backwash frequency.
Runs were made at processing times of 30 and 35 minutes between backwash cycles at
32 gfd, but the system failed to achieve stable performance.

The flux used during the backwash steps was 170 gfd. The flux was reduced to 110 gfd
in an attempt to increase the recovery of the system, but the lower flux was not enough to
effectively backwash foulant from the membrane surface.

Cleaning Optimization

Effectively cleaning the membranes of foulant is an important consideration when

designing a HYDRAcap UF system. A combination of high and low cleanings worked
well to restore membrane permeability. The first cleaning used a high pH solution of
sodium hypochlorite and caustic soda to remove both organics and biological growth.
First, a solution of 50 ppm NaOCl was prepared using RO permeate. The pH of this

24
solution was raised to 12 using caustic soda. This solution was then re-circulated at
elevated temperatures (32-38C) for approximately 40 minutes. The system was then
drained and rinsed with city water. At this time, a clean water flux profile was conducted
to determine the condition of the membrane. Typically, it was found that this type of
cleaning recovered the TMP to around 5.5 psi.

To further recover the TMP, a low pH cleaning was then conducted to help remove
inorganics and scaling from the membrane. Citric acid was used to lower the pH of a
sump of city water down to around 2. Typically, the TMP recovered an additional 2 psi
after this cleaning step. The result from this type of cleaning on a fouled membrane can
be seen repeatedly in Figure 5, where failure of the pretreatment chemical dosing pump
led to rapid fouling of the membrane.

A low pH cleaning by itself was sufficient for cleaning the system after fouling occurred.
The TMP was only recovered to about 5 psi after a low pH cleaning. Because of this, it
was determined that the most effective method of cleaning the modules was to clean with
a high pH solution first, followed by a low pH cleaning. The low pH cleaning also
removed any precipitants from the membrane that may have been deposited by the high
pH cleaning.

Chemical Pretreatment

Chemical pretreatment was found to be vital to the performance of the UF system. Ferric
chloride was dosed at a concentration of 3 ppm (1 ppm as pure iron). Sodium
hypochlorite was dosed at a concentration of 3-5 ppm as total chlorine. During the
course of this experiment, many dosing problems were encountered. If either the ferric
chloride or the sodium hypochlorite failed to dose properly, the modules quickly fouled.
It was found that the chemical injection valves clogged with solid debris relatively
quickly and needed to be cleaned out on a regular basis. Figure 5 shows how the
TCTMP was affected when the pretreatment chemicals failed to properly dose during an
experimental run at 36 gfd. When fouling occurred due to these problems, cleanings
were performed to restore the TCTMP back to their starting levels.

25
 TCTMP vs. Process Hours
36 gfd @ 84% Recovery
14.00 Cl2 and FeCl3 dosing pumps
experienced repeated problems.

12.00

10.00
TCTMP (psi)

8.00

6.00

4.00

2.00

0.00
0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1000.0
Process Hours (hrs)

Figure 5. Modules quickly fouled without the presence of either pretreatment

chemical.

Feed/Filtrate Quality

One of the main benefits of using this UF system over conventional wastewater
reclamation processes is that the HYDRAcap membrane removes solids and microbial
contaminants by size exclusion. This means that all particles larger than its absolute pore
size of the membrane get filtered out. This leads to a filtrate that is not sensitive to
fluctuations in feed water quality. Having a stable feed water quality for the reverse
osmosis system is very beneficial. Throughout the course of these experiments, the
filtrate turbidity produced by the UF system remained very stable. This in turn allowed
the RO system to show stable performance as well. Figure 6 shows how the filtrate
turbidity remained below about 0.05 ntu, regardless of the fluctuations seen in the feed
water turbidity which ranged anywhere from 2 to 100 NTU.

Laboratory testing of the feed water showed TSS levels of 7 ppm and TOC of 10.5 ppm.
The filtrate had TSS levels of 0.096 ppm and TOC of 8.5 ppm. This indicates that the
HYDRAcap is excellent at the removal of TSS from secondary effluent and also capable
of removing some TOC.

26
 Turbidity vs. Process Hours
40 0.09

35 0.08

0.07
30

Filtrate Turbidity (NTU)

Feed Turbidity (NTU) 0.06
25
0.05
20
0.04
15
0.03
10
0.02
5 0.01

0 0.00
0 1000 2000 3000 4000 5000
Process Hours (hrs)

Feed NTU Filtrate NTU

Figure 6. Filtrate turbidity remained stable through fluctuations in feed water

turbidity.

Another portion of this experiment involved laboratory testing of the feed and filtrate
water to determine the log removal of bacteria for this HYDRAcap system. A coliform
group bacteria test was performed on both the feed and filtrate water in order to calculate
this value. It was found that the feed water contained over 160,000 mpn/100 mL, while
the filtrate contained only 2 mpn/100 mL. From this information, it was found that the
log removal for this HYDRAcap system was 4.9.

Another aspect of the testing was the fluctuations in the SDI values found on the UF
filtrate. Figure 7 shows a number of 15 minute SDI values taken during the course of
this experiment. The SDI values ranged anywhere from 0.5 to 5 while the filtrate
turbidity remained fairly stable at around 0.03 ntu during the testing. In some cases, it
was found that the SDI varied from hour to hour. The reason for these fluctuations is
unknown, but it is important to note that despite these fluctuations in SDI values, the RO
system continued to operate quite stably.

27
 SDI 15 Filtrate Turbidity
7.00 0.2

0.18
6.00
0.16

Filtrate Turbidity (ntu)

5.00 0.14

0.12
4.00
SDI 15

0.1
3.00
0.08

2.00 0.06

0.04
1.00
0.02

0.00 0
1/ 2
29 03

16 04

6/ 4
6/ 4
9 4
10 04

/2 t
5/ 1
/6 02

In /20 2
al 3

/3 03
/3 03
7 3

20 04
22 04
22 04
23 04
23 04
23 04

al 4

10 04
In /11 04
4

4
11 s
5/ PS

2/ PS
2/ 00
2/ 00
2/ 0
23 0
st 0

1/ 00

st 0

gr 00

00
2/ y Te
9/ 20

1/ /20

20
2 / /2 0
11 /20
1/ /20

10 /20
10 /20

1/ /20
1/ /20
1/ /20
1/ /20
1/ /20
1/ /20
In /20

2/ /20
ss 2 /20
/2

2
2

te /2
l
/6

it
11

Pa
Sample Date

Figure 7. Fluctuations in filtrate SDI values evident during stable turbidity levels

Microfiltration
• Successful performance of the HYDRAsubTM submersible microfiltration
membrane treating secondary municipal effluent as pretreatment to reverse
osmosis was demonstrated. Filtrate flux of ~ 12 gfd was achievable. Recovery of
94-95% was achieved with optimum CEB and pretreatment settings. Filtration
cycle was 29 minutes followed by 1 minute backwash. Optimum CEB frequency
was found to be 24 hours using two chemicals – sodium hydroxide and sodium
metabisulfite. Off-line cleaning period was found to be 9 weeks. Cleaning period
may have been longer if accidents had been avoided. Adequate pretreatment and
proper settings on the CEB are key factors for system stability. The system was
able to tolerate high feed turbidities associated with spikes in secondary effluent
water quality. Recovery after upset in feed water quality could be achieved with
the help of CEB so that CIP chemical cleaning was not always necessary.

Performance
Run number 1:
During the first two months of operation, the system was tested at a constant flux of
11.7gfd with inline coagulation and chloramination. No regularly scheduled CEB was
implemented. The intent was to compare MF pilot performance to that of the recently
completed UF pilot study using a similar chemical regime. This portion of the test run
demonstrated that HYDRAsubTM microfilter was unable to sustain high permeability
without the implementation of CEB. Figure 8 plots flux and recovery over the course of
the 1000 processing hour (42 days) run. The pure membrane permeability at startup was
about 10.5 gfd/psig. During the first 200 hours of operation, the permeability dropped to
3gfd/psig with decline rate of about 0.84gfd/psig-day. Such a decline is very common for

28
virgin membranes and has been observed with various other types of hollow fiber
membranes. Permeability stabilized after 200 – 350 processing hours. At 335 processing
hours a single CEB with sodium hydroxide was performed. This CEB restored the
permeability to about 5 gfd/psi. However, after the CEB, the fouling rate was even faster
and permeability stabilized in the range 1.5-2.0gfd/psi. Further CEBs at 680h, 915h and
920h were ineffective. A short CIP with sodium hydroxide followed by a 90 minute soak
proved only marginally effective at restoring the permeability to around 3gfd/psi. The
test was temporarily suspended at 1000 processing hours due to maintenance at the La
Salina wastewater plant.
Run 1 MF permeability and flux vs processing time.
14.00 16.00

permeability flux

14.00
12.00

12.00
10.00
Permeability, [gfd/psi@20oC]

CEB NaOH
CEB NaOH CIP CEB NaOH 10.00

8.00

Flux, [gfd]
8.00

6.00

6.00

4.00
4.00

2.00
2.00

0.00 0.00
0 100 200 300 400 500 600 700 800 900 1000
Processing time, [h]

Figure 8. MF flux and permeability during run number one.

During run 1, several process interruptions associated with equipment malfunction where
encountered. At 150 process hours the filtrate pump failed to deliver the required flow at
vacuum pressures exceeding 6psig. The filtrate pump was upgraded with a larger
impeller to provide ample reserve in flow and pressure. At 210 process hours, the pilot
stopped due to a low feed tank level alarm. A similar event happened after 496
processing hours. Later during Run number 2, these types of system alarm shutdowns
continued. After close observation, it was found that the level transmitter in the feed
membrane tank was having inconsistent signal. The sensor was an ultrasound type which
is prone to loose the signal when foaming or solids build up occurs in the tank. In order
to resolve this issue, a new type of pressure driven level transmitter was installed which
appeared to have solved the problem. Following run number 1, a rigorous offline
cleaning was performed in preparation for Run number 2.

Run number 2

During run number two, which lasted 9 weeks, the backwash cleaning frequency was
increased to twice a day. Aeration was provided at fixed flow and pressure. Airflow was

29
kept at around 13-14 scfm. The actual air pressure to the diffuser outlet was always 10-
11psig.

Run number two was designed with two main objectives:

• finding efficient processing conditions to meet desired flux by sustaining high
permeability;
• optimize operating cost while at the same time keeping system performance
stable;

Initially the system was started without implementation of regular CEB to confirm that
feed water quality and system settings were the cause of the increased rate of fouling on
the system. After 180 processing hours, it became clear that permeability decline was
steeper than anticipated and the intended period between off-line cleaning would not be
achieved. At 184 processing hours, an automated CEB procedure was introduced.
Initially the CEB helped to stabilize permeability in the range 1.2 – 1.9gfd/psig.
However, permeability decline following CEB was large enough to exceed the maximum
recommended filtrate vacuum pressure of 8psig. Restoration of permeability following
CEB was still far from what was desired. To resolve this issue, the CEB procedure was
modified at 380 hrs.

After modification of the CEB, the effect on membrane cleaning improved significantly.
Visual observations of the membrane fibers revealed less solids residual after completion
of the CEB. During the course of next week, permeability continued to improve-
confirming the improved cleaning efficiency of the new CEB.

At this condition, the system continued to operate stably for 300 processing hours. Then
one of the two activated sludge reactors at the wastewater treatment plant experienced
wash-out conditions due to an error in laboratory analysis. This caused a dramatic
change in secondary effluent quality. Poorly treated wastewater blended with the
secondary effluent. Feed COD almost doubled and reached values of 120 ppm. The feed
suspended solids in the membrane tank increased correspondingly. Membrane feed
turbidity exceeded 50 NTU and peaked at 100NTU in the hours before CEB. Pilot
operation was continued to test the system’s ability to deal with spikes in feed water
quality.

30
 Run 2 - MF permeability and flux vs processing time. FLUX = 11.8 GFD.

5.00 14.00
MODIFIED CEB PROCEDURE STOP BLEED BAD FEED WQ LESS DRAIN
RECOVERY=93.2% RECOVERY=95%
4.50
12.00
4.00

3.50 10.00
Permability, [gfd/psi@20oC]

Start daily CEB

3.00
8.00

Flux, [gfd]
2.50

6.00
2.00

1.50 4.00

1.00
Perm Flux 2.00
0.50

0.00 0.00
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400
Processing time, [h]

Figure 9. MF run number two. Membrane flux and permeability.

As seen in Figure 9 & 10, permeability restoration after CEB remained unchanged.
However, the performance between CEBs changed substantially. Permeability decline
was steeper and reached values previously seen when no CEB was used (1.2gfd/psig). In
order to compensate for the higher organic loading on the membrane, the dosage of
sodium metabisulfite (SMBS) was increased temporary from 5.0g/l to 7.5g/l during this
period. This helped clean the membrane, providing slightly higher permeability just after
CEB compared to normal conditions. After 6 days of transition, the wastewater treatment
plant reestablished normal operation and feed water to MF pilot returned to normal
values. At that time SMBS dosage for CEB was returned to normal values of 5 g/l. At
the end of this period membrane permeability remained equal to what it was before the
feed water quality spiked (3-4gfd/psig).

31
 Run 2 - MF filtrate vacuum pressure vs processing time.
0.00

STARTCEB NEW CEB PROCEDURE FEED W.Q. UPSET NO CEB for 72 HOURS
-2.00
FILTRATE PRESSURE, [PSI]

-4.00

-6.00

-8.00

-10.00

-12.00

ul

ul
n

un

un

g
ay

l
Ju
Ju

Au
-J

-J
-M

-J

-J

2-
2-

12

22

1-
12

22
23

CALENDAR TIME, [mm/dd]

Figure 10. MF run number two. Filtrate vacuum pressure.

At 1108 processing hours the pilot unit was left without CEB for 74 hours due to operator
error. This resulted in rapid fouling of the membrane. Vacuum pressure eventually
exceeded 11psig. The first CEB after the accident restored the TMP back to 4.3psig.
However, in the weeks after, the TMP continue to increase – both before and after CEB.
It became clear that residual fouling after the accident left the membrane with severe
fouling that required off-line chemical cleaning.

It should be noted that in the last 200 processing hours of run 2, higher recovery was
tested (94-95%). This brought more solids to the membrane over the period between
CEBs. It is very unlikely that this higher solids load was the primary cause for the
membrane fouling during the last week of operation. During a third test run (not shown
in this report) 95% recovery was maintained as long as the CEB is performed promptly
with the required dosage of chemicals.

System Process Optimization

During Run 1 no optimization was attempted. However, during run 2 various parameters
were optimized.
A. CEB optimization:
• At 185 processing hours (run 2) daily CEB was introduced. Caustic soda
followed by SMBS was used in the CEB.

32
 • At 380 processing hours the CEB procedure was modified to allow membrane
tank aeration during soak step. Partial chemicals neutralization was achieved
inside the membrane tank.
• At 438 processing hours (run 2) the CEB was performed with NaOH only.
The second portion of the CEB (with SMBS) was not executed. The purpose
of this test was to examine the degree of importance of the SMBS CEB on the
overall TMP restoration. TMP was restored only 50% with NaOH CEB.
Further both NaOH and SMBS CEB were conducted so that the interval
between successive CEBs of the same type was the same but the occurrence
of the two types of CEBs was staggered by 12 hours. It was found that the
effect of CEB decreases dramatically if both CEBs are not performed
consecutively.
• At 1184 processing hours (run 2) the chemical for CEB2 was changed to
evaluate efficiency of alternative chemical. Sodium metabisulfite was
substituted with citric acid. Backwash concentration of citric acid was
adjusted to 9 g/l. The effect of citric acid was similar to what was seen with
SMBS. Reversing the order of chemicals for CEB (first SMBS followed to
NaOH) proved to be less effective.
B. Recovery optimization
• At 686 processing hours (run 2) the bleed flow was discontinued. This
increased the recovery from 88.5% to 92%.
• At 1260 processing hours (run 2) the dump volume per single backwash was
decreased from 43 gallons/BW to 28gall/BW. This increased the recovery
even further to about 94%.
C. Feed chemical optimization.
• At 970 processing hours (run 2) the feed coagulant dosage was lowered from
0.9ppm to 0.55ppm (as pure iron). This change did not affect the membrane
performance until end of the test.
• At 550 processing hours (run 2) the feed oxidant dosage was lowered from
3.0ppm to 2.5ppm. There was no visible detrimental effect on the MF pilot
performance until end of the test after that change.

Chemical Pretreatment
Run number 1
Concentration of pretreatment chemicals in the feed stream was closely monitored using
field test kits. The chlorine dosing pump was adjusted accordingly to produce the
required dosage of 3.0 ppm chlorine in the feed. Due to a large excess of free ammonia
in the feed (15-30ppm) all of the chlorine was converted into chloramines. Total chlorine
corresponds to concentration of free chlorine dosed in the system.

Feed iron concentration was maintained around 0.9ppm with maximum value not
exceeding 1.0ppm as iron.

33
Run number 2
During run number two, feed chemicals dosage initially was similar to that used during
run number one. Later both feed chemicals where optimized. Chlorine and iron where
lowered to 2.5 ppm and 0.45 ppm respectively.

Feed/Filtrate Water Quality

Run number 1
Feed turbidity in the membrane tank is measured via a submersible sensor and transmitter
type OptiQuant (HACH Company). The unit measures turbidity directly in membrane
tank thus recording actual feed water conditions that the membrane is facing during
operation time. During run number one, feed turbidity averaged 20.6 NTU, with a
median value of 19.25 NTU. The graph on Figure 11 is shows the values for feed
turbidity during this period. Due to constant bleeding and recovery of about 89%, feed
turbidity was lower than feed turbidity experienced during run number 2.
Run 1 MF feed turbidity vs processing time.

50

45

40

35
Feed Turbidity, [NTU]

30

25

20

15

10

0
0 200 400 600 800 1000 1200
Processing time, [h]

Figure 11. MF run number one. Membrane tank turbidity.

Filtrate turbidity was measured with in-line low range turbidimeter type 1720E (HACH
Company) and displayed on a SC100 universal communication module (HACH). The
average value during the test was 0.180 NTU with a median value of 0.180 NTU. Filtrate
turbidity is plotted on Figure 12.

34
 Run 1 MF filtrate turbidity vs processing time.
0.5

0.45

0.4

0.35
Filtrate turbidity, [NTU]

0.3

0.25

0.2

0.15

0.1

0.05

0
0 200 400 600 800 1000 1200

Processing time, [h]

Figure 12. MF run number one. Filtrate turbidity.

Feed and filtrate pH was nearly constant for most of the test. There were no visible
deviations on the feed or on the filtrate. The inline pH transmitter on the backwash line
was compared with the pH transmitter mounted inside the feed tank. During backwash
both values where compared. There where no differences between these pH values for
the duration of the test.

Feed temperature during run one slowly but steadily climbed from 23oC to 26oC. The
increase in temperature is contributed to changes in seasons – spring to summer.

Run number 2
Feed turbidity during run two was generally higher than feed turbidity during run one. It
reached values exceeding 100NTU on several consecutive days. Figure 13 plots online
data for feed turbidity during run two. Initially feed turbidity was quite low – less that 10
NTU. Once the bleed flow was discontinued and feed recovery increased, feed solids
concentration was elevated. Correspondingly, feed turbidity shifted to the range of 10-30
NTU. During feed water quality upset, turbidity exceeded 50 NTU for longer periods of
time and in some hours of the day reached 100 NTU. After the spike, turbidity returned
to a range of 10-30NTU. The very low values of feed turbidity seen, 3-5 NTU, are
recorded just after CEB. At the end of the CEB procedure (once a day), the membrane
tank is filled with secondary effluent. Turbidity in first several hours corresponds to raw
secondary effluent turbidity as discharged from the main water treatment plant. After
filtration continues over the course of the day, feed solids accumulate in the tank and
turbidity increases correspondingly. Finally, peak turbidity is reached just before CEB.
A second increase in recovery, made in the last week of operation, did not substantially
affected feed turbidity.

35
 100.00

90.00

80.00

70.00
FEED TURBIDITY, [NTU]

60.00

50.00

40.00

30.00

20.00

10.00

0.00
n

l
ay

ul
n

g
l

u
Ju
u
u
Ju

Au
-J

-J
-M

-J
-J

2-

12

22
2-

1-
22
12
23

CALENDAR TIME, [mm/dd]

Figure 13. MF run number two. Membrane tank turbidity.

Like run number one, the filtrate turbidity showed large oscillations during run number
two. The problem with sampling flow continued to challenge the measurement stability
and was the primary cause for the wide range in values. Figure 14 plots on-line data for
filtrate turbidity.

36
 0.5

0.45

0.4

0.35
FILTRATE TURBIDITY, [NTU]

0.3

0.25

0.2

0.15

0.1

0.05

g
un

un

ul
l
ay

u
Ju

Au
Ju

-J

-J
-M

-J

-J

2-

22
12
2-

1-
12

22
23

CALENDAR TIME, [mm/dd]

Figure 14. MF run number two. Filtrate turbidity.

Due to overlapping on the graph, it is difficult to get an idea of how many points in the
turbidity measurement fall within each sub region of the scale. To resolve graphing
problem, a normal distribution of all values is plotted on Figure15. From this graph, it
can be seen that feed turbidity remained less that than 0.1NTU for 65% of the time, and
less that 0.3NTU for 95% of the time.

37
 Run 2 - MF filtrate turbidity normal distribution.

FILTRATE TURBIDITY, [NTU]

1

0.1

0.01
0 10 20 30 40 50 60 70 80 90 100
Sample percdnatge, [% ]

Figure 15. MF run number two. Normal distribution on filtrate turbidity.

Feed and filtrate pH were similar for the whole test (pH=7.3-7.6). Occasional spikes are
attributed to readings taken immediately after CEB procedure, when feed pH in the
membrane tank is temporary elevated due to a small residue of caustic in the membrane
tank. This value decreased during the first filtration cycle and dropped below pH=8.0 at
the end of the first filtration cycle following CEB.

Water temperature continued to climb almost linearly from 25oC to 27.5oC at the end of
the test.

Off-line Chemical Cleaning Results.

During the course of the test the microfiltration membrane was subjected to two major
chemical cleanings. The first cleaning was performed after run number one and the
second cleaning was performed at the end of run number two. All chemical cleanings
were performed in the membrane tank with city water as the feed source. High pH
cleanings used RO permeate to prepare the solution for chemical cleanings prior to
contact with membrane. Water temperature during cleanings was in the range 17-20 oC.

Cleanings after Run 1

The cleaning after run number one was performed in four steps. Table 5 below provides
details for conditions at which the cleaning was performed.

38
 Table 5. Cleaning conditions for CIP after run number one.

Cleaning step 1 2 3 4
Chemical 1 NaOH C6H8O7 SMBS SDBS
Concentration in BW 4,000ppm 20,000ppm 0ppm 0ppm
line pH 13
Concentration in the 400ppm 2000ppm 2,500ppm 500ppm
tank pH 12.2 pH=2.9
Source of delivery BW line BW line Feed side Feed side
Chemical 2 NaOCl H2SO4 H2SO4 None
Concentration in BW 0ppm 0ppm 0ppm 0ppm
line
Concentration in the 50ppm 5,000ppm 1,000ppm 0ppm
tank pH=2.6
Source of delivery Feed side Feed side Feed side N/A
Contact time 80 min 120 min 90 min 900 min
Aeration regime 1min ON/ 1min ON/ 1min ON/ No aeration
30min OFF 30 min OFF 30min OFF
Notes - - - -

The result from this cleaning can be seen on Figure 16. Membrane permeability on the
fouled membrane at the end of run number one was 1.5-2.0 gfd/psi. The first cleaning
step (high pH) did little to restore membrane TMP. The second and third cleaning steps
(low pH) restored permeability and delivered a higher cleaning efficiency. This can be
contributed to presence of iron based coagulant in the feed. Coagulant bonded with
organic matter created a dense fouling layer on the membrane surface. Thus, caustic soda
alone was unable to remove the organic matter due to large amounts of iron. Once the
iron is dissolved, organic matter is much easier to bring back into solution.

39
 DRIP MF Pilot Test. Clean Water Flux Profiles after CIP - end of run 1.

3.00

2.50
Permeability, [gfd/psi@20oC]

2.00

Caustic
1.50 SMBS&H2SO4
Citric&H2SO4
SDBS

1.00

0.50

0.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00
Flux, [gfd]

Figure 16. MF clean water flux profile after CIP followed run number one.

Cleaning after Run 2

At the end of run two, permeability was in the range 1.3-1.8 gfd/psig. Chemical
cleanings were performed in two steps – alkaline followed by acid. Clean water
permeability was measured after the second cleaning step and found to be 3.6gfd/psig.

The first step consisted of caustic soda cleaning. The solution was introduced from the
filtrate side of the membrane via backwash pulses. Each pulse was 40 seconds long out
of five total pulses. The final pH concentration in the tank was 12.2-12.4. During soak
time, the membrane was aerated for longer cycles to enhance mechanical cleaning on the
membrane. Aeration was continuous for five minutes followed by a 10 minutes
pause/soak time. After alkaline cleaning, the membrane was washed with a spray hose.

The second cleaning step consisted of a mix of three chemicals – citric acid, sulfuric acid
and sodium metabisulfite. The solution was delivered to the membrane from the filtrate
side on four individual backwash steps. Each step was 40 seconds long. The final pH in
the membrane tank was about 2.2. During the 90 minute soak, the system was aerated
50% of the total time. Aeration regime was 5 min ON followed by 5 min OFF. Again,
the membrane was rinsed with a spray hose briefly. This time, large amounts of solids
where seen leaving the membrane module during the flushing procedure. Table 6
demonstrates the conditions under which this cleaning was performed.

40
 Table 6. CIP after Run 2. Detailed cleaning conditions.

Number cleaning step 1 2

Chemical 1 NaOH C6H8O7
Concentration in BW 6,000ppm 20,000ppm
line pH 13
Concentration in the 600ppm 2,000ppm
tank pH 12.2
Source of delivery BW line BW line
Chemical 2 None SMBS
Concentration in BW 0ppm 0ppm
line
Concentration in the 0ppm 5,000ppm
tank
Source of delivery Feed side
Contact time 90 min 90 min
Aeration regime 5min ON/ 5min ON/ 5min OFF
5min OFF
Notes - Add 1000ppm H2SO4 for pH
adjustment down to 2.0

It can be speculated that implementation of a daily CEB helped not only to sustain long
periods between off-line cleaning, but also to restore membrane TMP back to the original
values with only two regular cleaning steps. Long soaks were no longer needed.

Reverse Osmosis
• Both UF and MF provided adequate pretreatment to the RO when operating at a
flux of 11 gfd. The standard energy saving membrane performed as well as the
low fouling membrane and the use of the smaller (26 mil) brine spacer showed no
greater tendency to foul than elements using larger (31 mil) brine spacer. When
fouling did occur, the high pH cleaning proved more effective than the low pH
cleaning; suggesting the presence of colloidal and organic fouling.

UF pretreatment
Due to the high quality feed coming from the UF pretreatment, the permeability of each
of the three RO membranes remained stable during 2500 hours of operation at 9gfd and
then 11 gfd (Figure 17). At 1000 hours of operation, a dosing pump malfunction dropped
the feed pH below 2 for a short period. The exposure to low pH served to loosen the
membranes, after which operation returned to normal and the membranes performed
stably for the remainder of the 9.4 gfd portion of the test.

41
 0.20
LFC1

0.18 ESPA2

0.16 LFC3
Specific Flux, gfd / psi

0.14

0.12

0.10

0.08
9 gfd to 11 gfd
0.06

0.04

0.02

0.00
0

500

1000

1500

2000

2500
Operating Time, hours

Figure 17. Normalized permeability of three RO membranes treating UF filtrate.

ESPA2 and LFC1 were located in parallel in the first stage. LFC3 was located in the
second stage.

At 1300 hours, system average flux was increased from 9.4 gfd to 11.3 gfd by replacing a
single element from each pressure vessel with a dummy element. The membranes
performed well at 11.3 gfd with only a slight decrease in flux.

During the first 500 hours of operation, the LFC1 and LFC3 both demonstrated an initial
10 % drop in permeability before reaching a stable operating condition. This drop in
permeability is typical of RO systems treating waste streams and is likely caused by the
initial deposition of organics on the membrane surface. Though the initial increase in
permeability as seen in the first hours of ESPA2 operation appear to contradict this
phenomenon, factory test data of the six ESPA2 elements reveal a different trend. An
average specific flux of 0.18 gfd/psi was measured in the six ESPA2s prior to installation
and startup of the system. The initial, lower specific flux of 0.12 gfd/psi for the ESPA2
elements in Figure 17, suggest a possible data collection or instrumentation error.
During the startup of the second phase of testing, after new ESPA2s were installed, the
membranes showed the expected initial drop in permeability followed by stable
performance.

Corresponding to the initial drop in specific flux was an initial decrease in salt passage
for all three element types during the first hours of operation as seen in Figure 18.
Except for the pH excursion at 1000 hours, salt passage remained stable throughout

42
operation at 9.4 gfd. The dilution effect led to a drop in salt passage as system flux was
increased from 9.4 gfd to 11.3 gfd. Salt passage remained stable throughout 11.3 gfd
operation.

2.50 LFC1

ESPA2
Normalized Salt Passage, %

2.00 LFC3

1.50

1.00

0.50
9 gfd to 11 gfd

0.00
0

500

1000

1500

2000

2500
Operating Time, hours

Figure 18. Normalized salt passage of three membranes treating UF filtrate. ESPA2 and
LFC1 were located in parallel in the first stage. LFC3 was located in the second stage.

As expected, when utilizing membrane pretreatment, differential pressures also remained

stable for all three elements throughout the test (Figure 19). Because ESPA2 and LFC1
are loaded in parallel pressure vessels, the differential pressure of each is identical. The
drop in differential pressure starting at 1300 hours is attributed to the replacement of one
element from each vessel with a dummy element.

43
 20 LFC1
Normalized Differential Pressure, psi

18 ESPA2

16 LFC3

14
12
10
8
6
4
9 gfd to 11 gfd
2
0
0

500

1000

1500

2000

2500
Operating Time, hours

Figure 19. Normalized differential pressures of three membranes treating UF filtrate.

ESPA2 and LFC1 were located in parallel in the first stage. LFC3 was located in the
second stage.

MF Pretreatment
Figure 20 shows the permeability of the RO with MF as pretreatment. The permeability
of each of the three RO membranes remained stable during 1500 hours of operation at the
average system flux of 9 gfd and 11 gfd. When flux was increased to 13 gfd,
permeability began to decline indicating some fouling. The permeability of all three
elements dropped slightly in the first 24 hours. This drop is caused by the initial build up
of a permanent fouling layer and is characteristic of new membranes treating waste water.
The magnitude and duration of this initial dip is unique to different waste waters. The
stability of all three elements was similar regardless of membrane charge or brine spacer
thickness.

44
 0.20 ESPA2-31

0.18 ESPA2

0.16 LFC3
Specific Flux, gfd / psi

0.14

0.12

0.10

0.08
9 gfd to 11 gfd 11 gfd to 13 gfd
0.06

0.04

0.02

0.00
0

500

1000

1500

2000

2500
Operating Time, hours

Figure 20. Normalized permeability of three RO membranes treating MF filtrate.

ESPA2 and ESPA2(31 mil) were located in parallel in the first stage. LFC3 was located
in the second stage.

Corresponding to the initial drop in specific flux was an initial decrease in salt passage
for all three elements during the first hours of operation as seen in Figure 21. As with
permeability, salt passage remained stable for all three elements throughout testing at 9
gfd and 11 gfd. Despite the loss in permeability, salt passage remained stable while
operating at 13 gfd.

45
 2.50
ESPA2-31

ESPA2
Normalized Salt Passage, %

2.00 LFC3

1.50

1.00

0.50
9 gfd to 11 gfd 11 gfd to 13 gfd

0.00
0

500

1000

1500

2000

2500
Operating Time, hours

Figure 21. Normalized salt passage of three membranes treating MF filtrate. ESPA2
and ESPA2(31 mil) were located in parallel in the first stage. LFC3 was located in the
second stage.

Because of the MF pretreatment, differential pressures also remained stable for all three
elements throughout the testing at 9 gfd and 11 gfd – Figure 22. Because the ESPA2
with 26 mil brine spacer and the ESPA2 with 31 mil brine spacer were loaded in parallel
pressure vessels, the differential pressure of each vessel is identical. For this reason, feed
flow into pressure vessel 1 was monitored. The stability of the feed flow into pressure
vessel 1 at approximately one half of the total feed flow suggested that the 31 mil
elements fouled at a rate similar to that of the 26 mil elements. No correlation between
brine spacer thickness and fouling could be established. This is not surprising when
considering the thickness of the brine spacers relative to the size of the colloidal material.
The range of brine spacer thickness from 26 mils to 31 mils equals 660 microns to 790
microns respectively. The size of any colloidal particles passing through the membrane
pretreatment would be less than 10 microns – more than an order of magnitude smaller
than the thickness of the brine spacers. The difference between 26 mil brine spacers and
a 31 mil brine spacer is therefore insubstantial compared to the very small size of the
colloidal particles relative to the brine spacers should

46
 20 ESPA2-31
Normalized Differential Pressure, psi

18 ESPA2

16 LFC3

14
12
10
8
6
4
9 gfd to 11 gfd 11 gfd to 13 gfd
2
0
0

500

1000

1500

2000

2500
Operating Time, hours

Figure 22. Normalized differential pressure of three membranes treating MF filtrate.

ESPA2 and ESPA2(31 mil) were located in parallel in the first stage. LFC3 was located
in the second stage.

Cleaning
The system failed to foul during the UF pretreatment portion of the test. However, the
system was inadvertently exposed to a mixture of UF filtrate and UF feed which lead to
significant fouling within 24 hours. High and low pH cleanings were performed on lead
elements after the fouling incident. Elements were tested at their standard test conditions
before and after each cleaning to determine effectiveness.

The results as presented in Figure 23 suggest that the low pH cleaning had a negligible
impact while the high pH cleaning restored a significant portion of the membranes’
original permeability. The success of the high pH cleaning suggests that the primary
foulant was organic and/or coloidal in nature. It is therefore surprising that the negatively
charged ESPA increased in permeability more than the neutrally charged LFC as
negatively charged membranes tend to attract and become fouled from positively charged
organic material.

47
 Specific Flux

0.20

0.18

0.16

0.14

0.12
gfd / psi

ESPA2-4040
0.10
LFC1-4040

0.08

0.06

0.04

0.02

0.00
X-Factory Before Cleaning After Low pH=4 Clean After High pH=10
Clean

Figure 23. Effect of high and low pH cleaning on ESPA2 and LFC membrane permeability

UF vs MF pretreatment
A comparison of the RO performance when running on the different pretreatments can be
seen in Figure 24. The initial increase in permeability of the UF/RO is uncharacteristic
and opposite to that of the MF/RO system. However, the stable performance of the two
systems appears to be similar. The performance of the two systems is consistent with
other studies comparing the effect of MF vs UF pretreatment on RO performance
(Alexander 2003).

48
 0.20

0.18

0.16
Specific Flux, gfd / psi

0.14

0.12

0.10

0.08

0.06

0.04

0.02

0.00
0

500

1000

1500

2000

2500
Operating Time, hours

UF Pretreatment MF Pretreatment

Figure 24. Normalized permeability of two sets of ESPA2 (26 mil) membrane treating
MF filtrate and UF filtrate. Flux ranged from 9 gfd to 11 gfd.
• Feed and permeate water analysis showed that the integrated membrane system
successfully treated municipal secondary effluent and produced treated water that
met or exceeded standards for discharge or reuse.

Regardless of UF or MF pretreatment, the RO successfully reduced the total dissolved

solids concentration from 900-1300 ppm in the feed to 30 ppm in the permeate. A typical
feed and permeate analysis for the different membranes is given in Table 7 below.

49
 Table 7. Feed and permeate analysis (11/4/02).
UF LFC1 ESPA2
Filtrate Permeate Permeate
Ca 68.9 0.02 0.08
Mg 36.1 0.01 0.04
Na 226 5.835 4.20
NH4 15.2 0.9 0.87
Alk as ppm CaCO3 137 5.5 7.00
SO4 254 0.89 1.14
Cl 340 5.6 3.27
NO3 16.2 2.3 1.62
SiO2 15.9 0.3 0.18
TOC 16.3 0.72 0.70
PO4 3.54 0.001 b.d.l
Conductivity 1890 43.50 32.32
TDS 1109.3 21.4 18.4

• Data collected from the integrated membrane pilot was used to evaluate the
design and economics of a large-scale treatment system. Because performance of
the RO was similar regardless of membrane pretreatment, the economics is also
similar for the RO. The operating cost of the MF and UF are shown to be similar.

Based on the performance of the pilots, a 3.5 MGD Integrated Membrane System was
designed using commercially available design software. The overall operating cost of the
system, based on chemical and power consumption, was calculated for both UF and MF
pretreatment to the RO. One disadvantage of MF operation was its reliance on a daily
chemically enhanced backwash (CEB). This disadvantage was offset, however, by the
MF’s ability to operate for 60 days without a clean in place (CIP) compared to 30 days of
operation for the UF. Based on the pilot operating data, the choice of membrane
pretreatment would make little difference for RO performance and operating cost.
Because low fouling and energy saving membranes showed similar performance,
operating cost for the RO would be minimized by the use of energy saving membranes.
Operating cost would be further reduced by constructing elements with 26 mil brine
spacer as opposed to 31 mil brine spacer and thus increasing membrane area. A
breakdown of the operating cost for the three systems is shown in Table 8 below.
Though power and chemical cost are similar for the MF and UF, a significant difference
for the two systems appears in the membrane replacement rate. The MF shows a lower
cost for replacement rate due to its unique design which allows smaller, lower priced
membrane bundles to be replaced upon failure instead of the whole module.

50
Table 8. Projected operating cost ($/kgal) of a 3.5 MGD integrated membrane
system based on UF/MF/RO pilot performance.

$ / kgal MF UF RO
Power ($0.09/kW-hr) 0.021 0.026 0.12
Chemicals (pre-treatment and cleaning) 0.058 0.053 0.04
Membrane replacement 0.046 0.098 0.05
Maintenance and labor 0.076 0.076 0.03
TOTAL 0.20 0.25 0.24

Capital cost of a plant is influenced by a number of factors including the design of the
pre-exiting wastewater treatment plant including the space available at that plant. If a
secondary settling tank exits which can be retrofitted with submersible membranes, and if
space is limited, then the HYDRAsubTM MF system may prove more economical. If,
however, the plant is less than 1 MGD and no settling tank exists for the inclusion of the
submersible MF, then the self contained HYDRAcap UF may result in lower capital cost.

51
 Conclusions and Recommendations
Ultrafiltration
This study showed that a HYDRAcap UF system may be successfully operated on
secondary effluent as a pretreatment to reverse osmosis when proper operating
parameters are implemented. The pilot system successfully produced filtrate with low
turbidity, regardless of feedwater turbidity levels which was essential for the successful
operation of the reverse osmosis system. It was demonstrated that the system would
operate continuously for 30 days before chemical cleaning. Testing also demonstrated
the optimal cleaning procedures needed to restore membrane performance. The optimal
flux was found to be 32 gfd and recovery was found to be 88%.

Microfiltration
Pilot test with HYDRAsubTM membrane successfully demonstrated the use of
microfiltration as pretreatment to reverse osmosis. There is no apparent difference
between UF and MF pretreatment on the performance of the RO. The larger pore size of
the microfiltration membrane provides adequate pretreatment to the RO when secondary
wastewater effluent is treated.

Feed water pretreatment to the MF using coagulant and oxidant appears to be the main
factor for system stability. The failure of either one of the two pretreatment chemicals
leads to rapid fouling. Of similar importance is the implementation of daily CEB. The
proper selection of CEB chemicals, dosages, and process sequence is critical for system
stability. The CEB can significantly reduce the CIP cleaning frequency so that the
interval between CIP cleanings is 9 weeks or longer.

Using CEB once a day was sufficient to restore membrane TMP to levels close to the
original TMP. A stable pattern of TMP increase and complete restoration after CEB was
seen on the pilot unit. The period of these repetitive cycles is equal to the period between
two CEBs. Small changes in feed water quality do not affects TMP restoration after
CEB. The best chemicals for CEB where found to be sodium hydroxide and sodium
metabisulfite.

Pretreatment with coagulant and introduction of chloramination in the MIMS

significantly improved performance and substantially suppressed biogrowth. Minimum
dosage of 0.5ppm iron and 2.5ppm chlorine where found sufficient to deliver required
result.

Off-line chemical cleaning was found successful to restore membrane TMP close to
initial values. For high pH cleaning, the most useful chemical was caustic soda. For low
pH cleaning, a mix of three acids – citric acid, sulfuric acid and sodium metabisulfite,
was the most economical and efficient cleaning composition. The order of chemicals
should first be a high pH cleaning followed by low pH cleaning.

52
System recovery of 94-95% was achieved. Elimination of system bleed flow and
reduction on backwash drain flow helped to reduce overall waste flow from the system.
It appears that the system can run for more than two months at these settings.

A filtrate flux of 12 gfd is applicable for the feed water treated in this study. Aeration
during filtration and backwash was found to be very economical. The total fraction of
daily time used for aeration with intermittent and rapid air pulsation was only 3% of the
total 24h operational period each day. Continuous aeration during part of the soak time
was also found to be helpful for cleaning the membrane.

Reverse Osmosis
The reverse osmosis unit successfully treated secondary effluent while operating at 11
gfd and 75% recovery using both MF and UF as pretreatment. The RO reduced salinity
from 1100 ppm in the feed to 20 ppm in the permeate.

The side by side operation of the neutrally charged LFC and negatively charged ESPA
membranes gave similar results. Both membranes showed a similar restoration in flux
after high pH cleaning. Though previous side by side testing of the two elements treating
waste water have demonstrated the superior resistance to fouling of the LFC (Wilf,
2000), other testing has shown the ESPA to resist fouling as well as or better than the
LFC (Alexander 2003). This study and previous studies lead to the conclusion that the
ability of the membrane to resist fouling when treating secondary effluent is dependent
not just on the membrane charge, but on a combination of membrane chemistry and
specific wastewater composition. The ability of the ESPA2 to resist fouling as well as
the LFC is advantageous for the end user due to the higher permeability, and therefore
lower pressure requirement, of the ESPA2.

The side by side operation of the ESPA2 with 26 mil brine spacer and the ESPA2 with 31
mil brine spacer demonstrated similar performance. Due to the high quality of the feed
water, both elements operated successfully at an average system flux of 11 gfd.
Increasing flux to 13 gfd resulted in a loss of flux and an increase in differential pressure
which was similar for both elements. For this reason, no correlation between brine spacer
thickness and propensity to fouling could be established. The 26 mil brine spacer
provides a significant advantage over the 31 mil brine spacer since more membrane can
be rolled into an element with a smaller brine spacer. More membrane area per element
results in savings for the end user. Either fewer elements can be used in a system to
achieve the same flux and thus reduce capital cost or the same number of elements can be
used to lower the flux and thus reduce operating cost.

The RO pilot was shown to have similar performance regardless of membrane pre
treatment. Whether using MF as pretreatment or UF as pretreatment, the RO was stable
at a flux of 11 gfd. This flux is typical of existing RO systems treating secondary
wastewater with membrane pretreatment (Alexander 2003). Because the RO performs
well with either MF or UF pretreatment, any advantages of one pretreatment type are
primarily economic and a function of the size of the proposed treatment plant.
Furthermore, because the operation costs of the two pretreatments are similar, the

53
economic advantage of either pretreatment option is based on capital cost. As a general
rule, treatment plants with a capacity of less than 1 MGD in which a settling tank is not
already present, should use the self contained UF to save on capital costs. Other plants
with a capacity greater than 1 MGD should consider MF; especially if the existing
settling tanks can easily be retrofitted with the submerged MF modules.

The successful operation of the UF, MF, and RO demonstrate the ability of the overall
multiple integrated membrane system to treat municipal secondary effluent and produce
treated water of a quality that meets the pertinent regulations for discharge or reuse. Both
the UF and MF served to remove suspended solids and greatly reduce turbidity to a level
acceptable for RO feed. The RO served to reduce dissolved solids and act as a secondary
barrier to any microbial or colloidal mater that passed through the UF or MF membrane
pretreatment.

Commercialization Potential
In terms of commercialization potential, this study should be considered at Stage 5,
Product Development and Field Testing, of the Stages and Gates Process as defined by
the Public Interest Energy Research (PIER) Program. This study used a sub-scale MIMS
pilot unit to demonstrate the systems feasibility, reliability, availability, maintainability,
and durability in the field. The UF and RO membranes used in this study are currently in
production and readily available. These membranes have been used to treat a variety of
water sources including wastewater.

The study successfully demonstrated the use of an MIMS to reclaim secondary effluent
for reuse, thereby reducing the demand for other water sources such as imported water.
The MIMS successfully treated water to a quality exceeding that of Title 22 of the
California Code of Regulations.

Recommendations
Law makers and the people of California have shown themselves to be proactive in
supporting this MIMS studies and others like it. The use of membranes to reclaim
wastewater can greatly reduce the negative impact of an ever diminishing water supply
coupled with a rapidly increasing population. The next step is to go from pilot studies to
full scale systems.

Large scale reclamation plants using membranes to treat wastewater are currently in
operation or under construction. One such site, the Ground Water Replenishment System
in Orange County, will use a MIMS consisting of MF followed by RO to treat 57 million
gallons of wastewater per day. The permeate from this MIMS will be returned to the
county’s depleted aquifers.

Despite the increasing water challenges Californians will continue to face in the coming
years and despite the proven success of MIMS pilots and systems to reclaim and treat
wastewater, most wastewater treatment plants throughout the state continue to send their

54
effluent to drain. It is essential for California to go beyond the study of MIMS and to
move quickly to implement the findings of such studies for large scale systems.

The findings of this particular study can be used for the design of a full scale MIMS
using UF or MF as membrane pretreatment to an RO. The choice of UF or MF depends
on the size of the plant as discussed previously above.

Another option to be considered is the piloting of a MIMS consisting of a Membrane

Bioreactor (MBR) as pretreatment to RO. In many cases the MBR is a better alternative
to MF or UF. The piloting of an MBR as pretreatment to the RO was beyond the scope
of this study.

Benefits to California
Within a few miles of the LaSalina Waste Water Treatment Plant, where this study was
conducted, lie several sites that could potentially benefit from reclaimed wastewater
treated by a membrane system. These sites include wetlands, agricultural areas,
industrial parks, highway irrigation, and golf courses. These sites are representative of
the numerous environmental, agricultural, commercial, and industrial water users
throughout the entire state of California that require water for non-potable uses and
would benefit from a dependable source of reclaimed wastewater. This study
demonstrates how an integrated membrane system can reclaim and treat wastewater to a
level that would benefit such users. This study demonstrates how an integrated
membrane system can successfully be used to address California’s water shortage
problems in the coming decades.

55
 References
Abi-Samra, B.C., Cook, P., Lange, P., Zepeda, J., Design and Operation of a
Microfiltration and Reverse Osmosis System Treating Secondary Effluent and Providing
High Quality Reclaimed Water. American Membrane Technology Association Biannual
Conference, Tampa Bay, 2002.

Adham, S., Gabliardo, P., Merlo, R., Trussell, R., Evaluation of Three Pretreatment
Processes on RO Membrane Performance on Reclaimed Water., AWWA Membrane
Conference, San Antonio, TX, March 2001.

Alexander, K., Guendert, D., Pankratz, T., Comparing MF/RO Performance on

Secondary and Tertiary Effluents in Reclamation/Reuse Applications. International
Desalination Association Annual Conference, Bahamas, 2003.

Alexander, K., Alt, S., Owens, E., Patel, M., McGovern, L., Low Fouling Reverse
Osmosis Membranes: Evidence to the Contrary on Microfiltered Secondary Effluent.
American Water Works Association Membrane Technology Conference, 2003.

Alonso E., Santos A., Solis G., Riesco P., On the feasibility of urban wastewater tertiary
treatment by membranes: a comparative assessment, Desalination 141, (2001), 39-51.

Andes, K., Bartels C., Iong, J., Wilf, M., Design Considerations for Wastewater
Treatment by Reverese Osmosis., International Desalination Association Annual
Conference, Bahamas, 2003.

DeCarolis, J., Hong, S., Taylor, J., Naser, S., Effect of Operating Conditions on Flux
Decline and Removal Efficiency of Solid, Organic and Microbial Contaminants During
Ultrafiltration of Tertiary Wastewater for Water Reuse. American Water Works
Association Membrane Conference Proceedings. San Antonio, TX, March 2001.

Schimmoller L., McEwen B., Ferric chloride pretreatment improves membrane

performance: pilot result for Denver’s new 30 MGD water reclamation plant, AWWA
Membrane Technology Conference Proceedings, 2001.

Wilf M., Alt S., Application of Low Fouling RO Membrane Elements for Reclamation of
Municipal Wastewater. Desalination 132 (2000), pp 11-19.

56