Desalination 385 (2016) 83–92
Contents lists available at ScienceDirect
Desalination
journal homepage: www.elsevier.com/locate/desal
Advanced organic and biological analysis of dual media filtration used as
a pretreatment in a full-scale seawater desalination plant
Sanghyun Jeong a,b, Robert Vollprecht c, Kyungjin Cho d, TorOve Leiknes a, Saravanamuthu Vigneswaran b,⁎,
Hyokwan Bae d, Seockheon Lee d
a
Water Desalination and Reuse Center (WDRC), King Abdullah University of Science and Technology (KAUST) Thuwal, 23955-6900, Saudi Arabia
Faculty of Engineering and IT, University of Technology, Sydney (UTS), PO Box 123, Broadway, NSW 2007, Australia
Degremont Pty Ltd, Perth Seawater Desalination Plant/Lot 3003, Barter Road, Naval Base, WA 6165, Australia
d
Center for Water Resource Cycle Research, Korea Institute of Science and Technology, 39-1 Hawolgok-Dong, Seongbuk-Gu, Seoul 136-791, Republic of Korea
b
c
H I G H L I G H T S
•
•
•
•
This paper focuses on the monitoring of organic and biological foulants using advanced techniques.
Optimization of dual media filter (DMF) to maximize the reduction of foulants was made.
Microbial community analysis of DMF filtered seawater and medium, and cartridge filter was conducted.
Semi-pilot scale DMF columns were operated on-site at the biofiltration mode.
a r t i c l e
i n f o
Article history:
Received 10 December 2015
Received in revised form 1 February 2016
Accepted 9 February 2016
Available online 19 February 2016
Keywords:
Biofouling
Dual media filter
Microbial community
Organic fouling
Seawater desalination
a b s t r a c t
Dual media filter (DMF) is being used as a primary pretreatment to remove particulate foulants at seawater desalination plants. However, many plants experience organic and biological fouling. The first part of this paper focuses on the monitoring of organic and biological foulants using advanced analytical techniques to optimize
functioning of DMF at Perth Seawater Desalination Plant (PSDP) in Western Australia. In addition, microbial community analysis in DMF filtered seawater, and on DMF media (DMF-M) and cartridge filter (CF) was conducted
using terminal restriction fragment length polymorphism (T-RFLP) and 454-pyrosequencing. In the full-scale
DMF system, the bacterial community structure was clustered along with the filtration time and sampling positions. For the DMF effluent samples, the bacterial community structure significantly shifted after 4 h of filtration
time, which corresponded with the permeability reduction trend. The dominant bacterial communities in the
DMF effluent were OTU 13 (Phaeobacter) and OTU 19 (Oceaniserpentilla). The different biofilm-forming bacteria
communities were found in the biofilm samples on DMF-M and CF. In the second part of the study, semi-pilot
scale DMF columns were operated on-site under same operating conditions used in PSDP. It demonstrated the
advantage of operating DMF at the biofiltration mode for improving the reduction of biofoulants.
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
Many seawater desalination plants (either thermal or reverse osmosis (RO) based) use granular media (conventional) filtration or dual
Abbreviations: AOC, assimilable organic carbon; ATP, adenosine tri-phosphate; BP, biopolymers; BB, building blocks; CA, correspondence analysis; CF, cartridge filter; CIP, chemical cleaning in place; DMF, dual media filter; DOC, dissolved organic carbon; HS, humic
substances; LC-OCD, liquid chromatography with organic carbon detection; LMW-A, low
molecular weight acids; LMW-N, low molecular weight neutrals; NMDS, non-metric multidimensional scaling; PSDP, Perth Seawater Desalination Plant; RO, reverse osmosis; SDI,
silt density index; TEP, transparent exopolymeric particles; T-RFLP, terminal restriction
fragment length polymorphism.
⁎ Corresponding author.
E-mail address:
[email protected] (S. Vigneswaran).
http://dx.doi.org/10.1016/j.desal.2016.02.017
0011-9164/© 2016 Elsevier B.V. All rights reserved.
media filtration (DMF) as a primary pretreatment strategy prior to the
main desalting processes [1–3]. After screening trash from the intake
seawater, which constitutes the initial pretreatment, coagulation/flocculation, media filtration (typically single-stage DMF; anthracite and
sand), and cartridge filter (CF) steps are usually used as pretreatment
to improve the feed water quality for the RO-based desalination plant
[4]. Ferric salts are often used to coagulate and flocculate colloidal particles and dissolved organics in feed water (seawater). Sulfuric acid
(H2SO4) is dosed to provide optimal pH for improving the coagulation
performance. DMF coupled with coagulation/flocculation works well
reduces the particulate fouling on RO membranes (especially during
summer or when algal bloom occurs). The filter media is backwashed
on a regular basis (once in 24–36 h) with air scour followed by the
DMF to remove captured SS from the filters. CF is usually used to
84
S. Jeong et al. / Desalination 385 (2016) 83–92
prevent the sudden occurrence of particulate matter after DMFs. DMF
and CF consume less energy compared to membrane-based pretreatment [5]. However, conventional DMF is not effective in inhibiting the
organic and biofouling [6]. As a result, a number of desalination plants
face the problem of biofouling both on CF and RO membranes downstream of the DMF system [7]. In addition, DMF filtered seawater still
has high fouling potential that results in fouling of CFs and these have
to be replaced once every 2–8 weeks [8]. Furthermore, non-optimized
DMF may lead to frequent chemical cleaning of RO membranes. Although currently the cost of chemical cleaning is relatively low, the
non-production of water during chemical cleaning in place (CIP) results
in a loss of production and a significant increase in workload. The optimization of DMF, in turn, can reduce the operational and energy costs of
RO plants and increases the lifetime of the CFs (usually used as a security barrier after DMF). It is therefore important to establish the proper
design and operational guidelines for DMF through a detailed study on
organic and biofouling.
At present, six large seawater desalination plants are operating in
Australia [9]. Four desalination plants use the DMF as a primary pretreatment and one of these full-scale plants is Perth Seawater Desalination Plant (PSDP), which is located in Western Australia. This plant was
commissioned in November 2006 with a capacity of 144 MLD that produces up to 17% of Perth's drinking water [10].
This study focuses on the following: i) monitoring of DMF by using
organic and biological parameters, and ii) facilitating DMF to operate
in the biofiltration mode. The performance of DMF at PSDP was monitored in terms of organic foulants using liquid chromatography with organic carbon detector (LC-OCD), and biological foulants (assimilable
organic carbon; AOC, standard blocking index by ultrafiltration; Ks-UF
and adenosine tri-phosphate; ATP).
Molecular level analyses on the DMF and CF systems can help to optimize the DMF operation. For this reason a detailed analysis of the microbial community in DMF filtered seawater and on DMF and CF media
was conducted. The above samples were analyzed using terminal restriction fragment length polymorphism (T-RFLP) and 454-pyrosequencing.
A pilot-scale DMF study was also conducted with same filter media
and coagulant used in PSDP together with the same operating conditions
such as filtration velocity. The effects of filter run-time (or backwashing
frequency) and backwashing duration on biofouling reduction were
assessed in terms of biological activity to operate the DMF in biofiltration
mode.
2. Materials and methods
2.1. Processes and samples description
2.1.1. Dual media filter (DMF) operation
At the PSDP, DMF is operated under pressure as opposed to gravity
to ensure long runs. Each DMF consists of a 300 mm layer of sand
with an effective size of 0.3 mm and an 800 mm layer of anthracite
with an effective size of 1.6 mm. A 100 mm gravel bed is provided so
that the nozzles are protected against clogging by the sand. After operation, flocculated suspended solids, dissolved solids and organics and
biofilms together with ferric hydroxide (oxidized ferric coagulant) are
present on the DMF media.
Ferric sulfate solution (Orica Chemicals, Australia) serves as a primary
coagulant in the pretreatment system of the PSDP. This assisted in the
precipitation of dissolved solids and ions and flocculation of suspended
solids in the feeding seawater. Ferric sulfate was dosed into a dilute sulfuric acid stream flowing into the seawater pipeline to the DMFs. pH was
adjusted using sulfuric acid to 6.8 (between 6.7 and 7.0) for optimum coagulation. Floquat FL 4526 PWG (Polydiallyldimethylammonium Chloride (PolyDADMAC)) was added as a cationic polymeric coagulant aid. It
should be noted that ferric sulfate and Floquat FL 4526 PWG are referred
to as coagulant and coagulant aid (or flocculant and flocculant aid).
The DMFs require regular backwashing to maintain the filter bed in a
good filtration condition and to remove the solids retained in the filter
media. Backwashing interval is determined by the monitoring of
pressure drop using differential pressure transmitters and the hydraulic
permeability measured using a flow meter. The condition of the
backwashing process is summarized in Table 1.
2.1.2. DMF samples
DMF filtered seawater samples collected at different filtration times
were analyzed in this study (DMF-0 h, 1 h, 4 h, 8 h, 12 h, 18 h, 24 h and
30 h). They were also used in a study of the microbial community. DMF
media (DMF-M) samples were taken from DMF number 9 (DMF 9) at
PSDP in March 2013 for the analysis. They were washed twice with
1 × phosphorus buffer solution (PBS) followed by mild sonication for
extraction of biofilm on DMF-M. The mild sonication was carried out
with the DMF-M in a beaker containing 1 × PBS using an ultrasonic
bath (Powersonic 420, Thermoline Scientific, 300 W) for a short time
(10 min).
2.1.3. Cartridge filters (CFs)
DMF filtered seawater was then passed through CFs. In PSDP, 5 μm
CFs (1251-10C, Pall Corporation, USA) are used as security filters. It is
made of high consistency polypropylene melt blown cartridge. In this
study, the fouled CFs were taken from CF number 14 (CF-14) at PSDP
in March 2013 for microbial analyses. The CFs were frozen immediately
after their removal from the housing. Several segments were then cut
from the fouled CF for analysis. Biofilm on CF was extracted using the
same method used for DMF-M extraction (see Section 2.1.1). A detailed
description of samples used in this study is given in Table 2.
2.2. Monitoring methods
2.2.1. Silt density index (SDI)
SDI is being used as a main fouling index or water qualitymonitoring indicator at desalination plants [11]. In the SDI test, the
time required to filter a fixed volume (500 mL) of feed water through
a standard 0.45 μm pore size microfiltration (MF) membrane with a diameter of 47 mm under constant-pressure (207 ± 3 kPa) is measured.
The initial time (ti) and the time of a second measurement taken to filter
500 mL water (tf) (after silt built-up) is noted and SDI value is calculated
using Eq. (1).
SDI ¼
1 t i =t f
100
T
ð1Þ
where, ti and tf are the initial and final time in seconds required to
collect the 500 mL permeate respectively, and T is the total elapsed
flow time (in this study, flow time of 5 min was used, thus is called
SDI5).
2.2.2. Liquid chromatography with organic carbon detector (LC-OCD)
Dissolved organic carbon (DOC) concentrations in DMF effluent and
organic foulant on CF were measured using DOC-LABOR Liquid Chromatography — Organic Carbon Detector (LC-OCD). The LC-OCD system
separates and measures hydrophilic DOC compounds according to
their molecular size using both an OCD detector (after inorganic carbon
Table 1
Main backwashing procedure of DMF in PSDP.
Operation
DMF
Normal flow
Rinse drain
Water scour
Air scour
Backwash
Maturation
800 m3/h
3 min
800 m3/h for 2 min
55 m/h for 7 min
1400 m3/h for 13 min
12 m/h for 18 min
S. Jeong et al. / Desalination 385 (2016) 83–92
Table 2
Description of samples used in this study (DMF and CF).
Sample
no.
Location Description
Analysis
DMF
DMF 9
Organic and biofoulants
and microbial
community
DMF-M DMF 9
CF
CF-14
Seawater SDI = 10.89 ± 0.10,
filtered seawater SDI (2.18 ± 0.24),
ferric dose = 0.68 mg/L and
coagulant aid = 0.28 mg/L
and 30 h operated (n = 14).
DMF media (anthracite and sand
with biofilm)
Old CF after replacement
Microbial community
Organic foulants and
microbial community
2.2.5. Adenosine tri-phosphate (ATP)
ATP is an indicative parameter of active biomass since it plays a critical role in cell energy metabolism and also serves in a number of cell
signaling processes [17]. In this study, BacTiter-Glo Microbial Assay kit
was used to measure ATP. It also served to indicate biological activity
in DMF filtered seawater and on biofilm on DMF-M and CF media. A
96-well luminometer (Wallac 1420 VICTOR2™ multilabel, multitask
plate reader, PerkinElmer Inc., US) was used to measure luminescence
produced from ATP reaction at room temperature. The analytical procedure was followed according to the manufacturer's guidelines.
2.3. Semi-pilot DMF study
purging) and an ultraviolet detector (UVD) (absorbance at 254 nm). The
analysis was carried out using dual columns with retention time of
180 min. Information regarding LC-OCD and relevant procedures are
described in another study in detail [12]. Depending on the LC-OCD
chromatogram, the different organic fractions were calculated using a
software program (ChromCALC DOC-LABOR, Karlsruhe, Germany).
DOC in seawater samples contains mainly biopolymers (BP), humic substances (HS or humics), building blocks (BB) and low molecular weight
neutrals (LMW-N) and acids (LMW-A) [13].
2.2.3. Assimilable organic carbon (AOC)
AOC can be used as an indicator to quantify the relative biofouling
potential or biofilm formation of water samples [14]. This study used a
bioluminescence method using Vibrio fischeri MJ-1 strain. A 24-well
multi-well, tissue culture treated plate (353047 — BD Falcon™) was
used. For quantification of AOC in the sample, the samples were firstly
filtered through a 0.45 μm PES filter to remove large particles and, the
bacteria naturally present in the sample were removed through inactivation at 70 °C for 30 min, and filtering through a 0.22 μm PES filter.
To quench the chlorine residual, 2 drops of 10% sodium thiosulfate solution were added to the water samples. 3.0 × 104 CFU/mL of V. fischeri
MJ-1 from a concentrated (3.0 × 106 CFU/mL) stock solution was inoculated into water samples (20 μL into 2000 μL). To monitor growth of
V. fischeri MJ-1, luminescence was measured using a Wallac 1420
VICTOR2™ plate reader (PerkinElmer Inc., US) at 30 °C (temperature
controlled). Glucose was used with artificial seawater as a general carbon source because a better cell growth was obtained even at low concentration (~0.01 μg-C/L). Thus, AOC concentration is expressed as μg-C
glucose equivalents/L. Detailed procedure can be found in a previous
study [14].
2.2.4. Standard blocking index by ultrafiltration (Ks-UF)
The standard blocking index obtained from molecular weight cut-off
(MWCO) 10 kDa ultrafiltration (UF) membrane (Ks-UF) was found to
be a representative a practical method for determining biofouling potential index [15]. This is because a linear relationship was observed between Ks-UF and AOC. UF membrane (material: Polyethersulfone, PES)
used in this study is made from NADIR®RM and its diameter was
47 mm. The raw seawater and DMF filtered seawater were pressurized
at 2.0 bar (207 ± 3 kPa) using N2 gas at room temperature.
Standard (pore) blocking index (Ks) was calculated with the following equation (Eq. (2)):
t
1 ks
¼ þ t
V=A J 0 2
85
ð2Þ
In this equation, t is the filtration time, V is the cumulative permeate
volume, A is the effective filtration area (0.001735 m2), J0 is the initial
flux, and ks (1/m) is the index (or coefficient) of standard (pore)
blocking. A linear relationship was observed when t/(V/A) was plotted
against t. This is a proof of the standard blocking model [16].
In addition to the above analyses of DMF filtered water and filter
media from PSDP, a semi-pilot filter study was conducted (with similar
conditions used in PSDP) to examine in more detail the benefits of operating DMF in biofiltration mode. Different backwashing intervals (BI)
(24 h and 35 h) and backwashing durations (BD) (2 min and 10 min)
were tested through four semi-pilot scale DMFs. The DMF at PSDP usually operates with 35 h-BI with 10–13 min of BD (as shown in Table 1).
In this semi-pilot study, shorter 24 h-BI and 2 min-BD were tested to see
whether biological activity increased with the mild backwashing. The
DMF columns of effective media height of 900 mm (packed with sand
(250 mm at the bottom) and anthracite (650 mm above the sand
layer)) were operated in parallel with different backwashing conditions: i) DMF-A: 35 h-BI and 2 min-BD; ii) DMF-B: 35 h-BI and
10 min-BD; iii) DMF-C: 24 h-BI and 2 min-BD; and DMF-D: 24 h-BI
and 10 min-BD. They were run at the Sydney Institute of Marine Science
(SIMS) for 14 d with continuous seawater feeding. Filtration velocity in
all four DMF columns was maintained at 10 m/h. The corresponding
empty bed contact time (EBCT) was 5.4 min. In-line coagulation was
provided with the addition of 0.68 mg Fe3+/L of ferric sulfate solution
and 0.28 mg/L of coagulation aid at pH 6.5 (with sulfuric acid). The filter
media and chemicals used were same as those used at PSDP. In addition,
the difference of seawater quality in PSDP and pilot-test site was only
± 10–15% in terms of organic content (DOC) and biological content
(ATP and AOC). Samples were occasionally collected (2 d, 4 d, 8 d and
14 d). DMF medium sample (5 cm3) was taken from three sampling
ports (top, mid and bot −850, 550 and 250 mm from the bottom, respectively) on the 15th day of operation. The biofilm was extracted
from the DMF medium by voltexing and mild sonication. To evaluate
the biofilter function of DMF, bioavailable organic reduction by DMF
and biological growth on DMF were measured in terms of AOC and
ATP concentrations, respectively.
2.4. Microbial community on DMF effluents and medium, and cartridge
filter
DNA samples on the DMF-M, in the DMF effluents (filtered seawater) and fouled CF were extracted using a modified CTAB-PEG protocol
[18]. To extract DNA in the DMF effluent, a 1 mL of DMF filtered seawater was placed in 2 mL tubes. Similarly, DNA was extracted from 1 mL
(cm3) DMF-M samples using centrifugation at the speed of 13,000 g
for 5 min. To extract the surface bound community from fouled CF, a
piece of membrane (1 cm × 1 cm × 1 cm) was cut into small pieces
and these were put into 2 mL tubes with glass beads (Bead-beating).
Details regarding the DNA extraction procedure have been outlined in
another study [18].
2.4.1. Terminal restriction fragment length polymorphism (T-RFLP)
Bacterial DNA encoding of a section of the 16S ribosomal subunit
was amplified using a 6-carboxyfluorescein-labeled 27F (5′-AGAGTT
TGATCMTGGCTCAG-3′) primer and an unlabelled 519R (5′-GWAT
TACCGCGGCKGCTG-3′) primer (IDT DNA, Baulkham Hills, NSW,
Australia). Fragment analyses of (cleaned and digested) DNA samples
were conducted at the Ramaciotti Centre for Genomics (University of
86
S. Jeong et al. / Desalination 385 (2016) 83–92
New South Wales, Australia). The analysis was done on an Applied
Biosystems AB3730 DNA Analyzer (Applied Biosystems, California,
USA) using a GeneScan LIZ-1200 size standard (Applied Biosystems,
California, USA). PeakScanner v1.0 (Applied Biosystems, California,
USA) used to identify fragments and their size was compared to the
LIZ-1200 size standard. The labels and “raw” data were exported to
the correct compatible formatting required by the T-REX T-RFLP Analysis tool [19]. This allowed the dominant T-RFLP at the outflow to determine the relative abundance of the T-RF between samples.
2.4.2. 454-pyrosequencing
454 high-throughput sequencing of the 16S rRNA gene (Roche 454
FLX instrument, Roche, Indianapolis, IN) served to identify microbial
community compositions. In this study, the 454-pyrosequencing was
conducted for the DMF-12 h and DMF-24 h samples in the DMF effluents (liquid phase). These were selected based on the T-RFLP analysis
and DMF performance result. In addition, two biofilm samples (CF and
DMF-M) were analyzed in order to study the dynamics of the bacterial
community from the liquid phase to biofilm samples. One hundred ng
of extracted DNA was sent to the Research and Testing Laboratory (Lubbock, TX, United States of America) for analysis of bacterial rRNA genes
using FLX amplicon pyrosequencing. The forward primer was constructed with the Roche A linker (5′-CCATC TCATC CCTGC GTGTC TCCGA
CTCAG-3′), an 8–10 bp barcode, and the Bac28F (5′-GAGTT TGATC
NTGGC TCAG-3′). The reverse primer was constructed with a biotin
molecule, the Roche B linker (5′-CCTAT CCCCT GTGTG CCTTG GCAGT
CTCAG-3′), and the Bac519R (5′-GTNTT ACNGC GGCKG CTG-3′) [20].
Amplification was performed on 25 μL reactions with Qiagen
HotStarTaq master mix (Qiagen Inc., Valencia, California), 1 μL of a 5μM solution of each primer and 1 μL of template. Reactions were executed on ABI Veriti thermocyclers (Applied Biosytems, Carlsbad, California)
under the following PCR program: 95 °C for 5 min, then 35 cycles of
94 °C for 30 s, 54 °C for 40 s, 72 °C for 1 min, followed by one cycle of
72 °C for 10 min and 4 °C hold. Sequencing analysis was executed according to the manufacturer's protocols (454 Life Sciences).
The sequences that have low quality barcode or short length
(250 bp) were screened. Following this, de-noising was conducted
using the USEARCH-based algorithm [21]. The chimeras were checked
using UCHIME [22]. BLAST was performed using the ribosomal 16S
rRNA gene database of Silva as reference information. Operational taxonomic unit (OTU) analysis for community richness was undertaken
using the CD-HIT-OTU (http://weizhong-lab.ucsd.edu/cd-hit-otu/) and
Mothur software (1.33.0) (http://www.mothur.org/).
2.4.3. Statistical analysis of T-RFLP and 454 pyrosequencing data
Non-metric multidimensional scaling (NMDS) and correspondence
analysis (CA) were undertaken to analyze the similarity of the microbial
community structure among the samples. Both statistical methods were
done based on Sorensen (Bray–Curtis) distance way using PC-ORD v.5.0,
MjM software (Gleneden Beach, OR). Concerning statistical analysis of
T-RFLP, the peak intensities were normalized as described previously
[23]. The relative abundance of the T-RFLP peak intensities was used
to perform the NMDS and CA. For the pyrosequencing data, the relative
abundance of the OTU results was used for CA.
Fig. 1. The relationship between permeability (Lp) in DMF and SDI of filtered seawater by
DMF (March 2013).
time and remained between 2.26 and 2.37. Feed seawater SDI was
10.38 ± 0.17. On the other hand, Lp declined from 471 m3/m2 Pa h at
3 h operation to 251 m3/m2 Pa h after 30 h of operation. This confirms
that SDI cannot explain the DMF fouling. For this reason, this study focused on the monitoring of DMF performance using other organic and
biological parameters (i: DOC and organic fractions; ii: AOC; iii: Ks-UF
and iv: ATP). To do this, DMF filtered seawater was collected at different
filtration times.
3.1.1. DOC and organic fractions
The concentrations of total DOC and different organic fractions
(mainly BP, HS and LMW-N) of DMF filtered seawater are presented
in Fig. 2. Total DOC concentration in DMF filtered seawater (effluent)
samples varied (changed) during filtration (at different filtration
times). Variation of LMW-N completely followed as that of DOC during
filtration, which means that LMW-N led to the variation of total DOC
during filtration run. It is interesting that about 60% of DOC compounds
in DMF filtered seawater was LMW-N. Simon et al. [24] characterized
organic matter in seawater and pretreated seawater in desalination
line by LC-OCD. They observed a similar organic distribution. HS and
LMW-N were the main components. Variation in BP and HS concentration of DMF effluent was not significant with the filtration time. BP and
HS reduction by DMF was relatively stable during filtration time (57.7%
and 14.9%, respectively) while total LMW-N reduction was quite low
(2.7% on average).
DOC concentration in DMF filtered seawater was high at the beginning of filter run (1.71 mg/L after 2 h of filter runtime. This is probably
due to backwashing. After backwashing, DMF underwent maturation,
which is filter-to-waste period of filter operation. However, DOC concentration was lower at 1.37 mg/L from 7 to 8 h to 24 h of filter operation. Further, following 24 h of operation, DOC concentration gradually
3. Results and discussion
3.1. Monitoring of DMF performance
After intake, the coagulated seawater was passed through the DMF.
At present, SDI (see Section 2.2.1) is used as a primary fouling and performance indicator along with hydraulic permeability (Lp (m3/m2 Pa h),
seawater flow (m3/h) / (dP (Pa) × filtration surface (m2)) at PSDP. Fig. 1
shows the relationship between Lp and SDI with filtration time of DMF.
However, as shown in Fig. 1, there was no relationship between Lp and
SDI. SDI of DMF filtered seawater did not change much with filtration
Fig. 2. Variation of DOC and organic fractions in filtered seawater by DMF during filtration
(March 2013).
S. Jeong et al. / Desalination 385 (2016) 83–92
increased up to the end of the filtration run. Thus, DOC variation pattern
during DMF operation is closely linked to LMW-N compounds. Baghoth
et al. [25] also stated that LMW-N has a good correlation with DOC. This
indicates that LMW-N can be used as a DOC indicator and main monitoring organic fraction. In other words, accurate monitoring and effort
to reduce LMW-N compounds are required for better organic reduction.
3.1.2. AOC
AOC in DMF filtered seawater was measured to evaluate the amount
of bioavailable organics removed by DMF (Fig. 3). During the initial operational period of DMF, filtered seawater containing the highest of AOC
(38.46 ± 0.96 μg-C glucose equivalents/L) produced from DMF. A relatively low concentration of AOC (15.32 ± 0.38–22.52 ± 0.56 μg-C glucose equivalents/L) detected in DMF filtered seawater that produced
during 5 h and 24 h of DMF operation. Interestingly, the pattern of
AOC is very similar to that of the LMW-N value obtained from LC-OCD
analysis. AOC increased again after 24 h of operation. From this analysis,
approximately 5% of LMW-N represented the portion of AOC compounds in DMF filtered seawater. Jeong et al. [15] in their earlier study
also found that LMW-N concentration was also roughly 22 times the
AOC concentration. Therefore, AOC measurement can be used for measurements of both biofouling potential and LMW-N concentration.
3.1.3. Ks-UF
DMF effluent was filtered through a 10 kDa UF membrane. In a previous study [15], different MWCO UF membranes were tested for determining biofouling potential index and 10 kDa was found to provide a
linear relationship between Ks-UF and AOC. Thus, in this study, pore
blocking coefficient (Ks) value obtained from UF test with 10 kDa (KsUF) was used as an indicator of biofouling potential. Filtering 1 L of
DMF effluent in this test required approximately 90 to 120 min of filtration time. Ks-UF values of DMF filtered seawater samples are presented
in Fig. 4a. They varied during filter runs. At the start of operation, Ks-UF
of DMF filtered seawater was 1.282 (1/m). This value was a low between 0.461 and 0.725 (1/m) during the filter operational period of 524 h. Ks-UF of DMF effluent samples showed a linear relationship
(R2 = 0.978) with AOC concentration (Fig. 4b). This implies that the
Ks-UF test can be used as an indicator of AOC, which is biofouling potential. A more laborious equipment is required for measurement AOC and
LMW-N. However, Ks-UF can be easily measured on-site (i.e. desalination plant) using a simple fouling index device.
3.1.4. ATP
It should be noted that ATP concentration is used to measure the active biomass or biological activity. In this study, the amount of active
biomass released from DMF during the filtration period was analyzed
(Fig. 5). As expected the un-stabilized DMF (at initial stage of operation;
after backwashing and maturation) released the highest amount of
Fig. 3. Variation of AOC concentration in filtered seawater by DMF during filtration (March
2013) (dotted line indicates logarithmic trendline).
87
biomass (159.8 ± 11.2 nmol/L). However, low ATP concentration was
observed in DMF filtered seawater (2.9 ~ 14.7 nmol/L) during the stable
operational period (5–24 h). This showed similar pattern as those of
LMW-N and AOC. Also, after 24 h of filter operation, the amount of released active biomass increased and it was 60.2 ± 15.4 nmol/L when
the DMF operation ended. This shows the relation between DMF operation and biomass loss. It can thus be used as a useful reference for
selecting the best period for backwashing.
It should be noted that the feed water and filtered water quality obtained from a full-scale desalination plant fluctuated although general
pattern of filtered water quality variation followed a uniform trend
(Figs. 2 to 5).
3.2. Bacterial community structure in DMF effluent, and DMF and CF media
(from T-RFLP analysis)
The bacterial community structures in the samples were examined
using NMDS and CA methods based on T-RFLP analysis. An analysis
was conducted of: firstly, bacterial community structure of the eight liquid samples — DMF effluent collected at different times of filtration (i.e.
0 h, 1 h, 4 h, 8 h, 12 h, 18 h, 24 h, and 30 h); and secondly, two biofilm
samples (CF and DMF-M). The 2D ordination plot and dendrogram illustrate that the bacterial community structure was clustered with the filtration times and sampling points; liquid phase samples, CF, and DMFM (Fig. 6a and 6b). The stress and instability values of the 2D plot
were 0.01 and 1.0 × 10−5, respectively, which indicate that the results
were statistically acceptable (i.e. stress value b20 and instability value
b0.001) [26].
For the DMF effluent samples, the bacterial community structure
could be classified into two groups based on the operation period (i.e.
i) DMF-0 h, -1 h and -4 h, and ii) DMF-8 h, -12 h, -18 h, -24 h, and
-30 h). This indicates that the bacterial community structure significantly shifted between 4 h and 8 h of the operational period. The pores of
DMF bed media become clogged, which led excess head-loss and reduced productivity due to the biofilm growth. However, a fraction of
the bacterial biomass detached from the filter bed can be released during start-up, after backwashing [27]. This occurrence was observed in
ATP analysis (Fig. 5), high numbers of active biomass released from
the DMF were detected in filtered effluent. This can be explained by
the “ripening” phenomenon or the inadequacy in the number of collectors within the pores of the filter grains. In addition, this filtered water
quality degradation may be due to filtration rate changes that occurred
by clogging filter as well as higher filtration rates [28]. Interestingly, this
community change corresponded to the reduction in permeability in
the DMF system (Fig. 1). For example, the permeability (Lp) was more
than 450 m3/m2 Pa h until 4 h of filtration time; this drastically decreased from 420 m3/m2 Pa h at 6 h to 294 m3/m2 Pa h at 8 h; and it
was maintained until the 24 h of filtration time. This comparable outcome implies that the dominant bacterial community that emerged
after 8 h may have resulted in filtration rate change in the DMF system.
This may have significant impact on the pretreatment of DMF since it
would tend to increase biofouling on subsequent processes such as CF
and RO. Microorganisms escaped with the filtered effluent will be
passed to the next step. In fact, regarding the biofilm community on
DMF-M and CF, the bacterial community structure of the latter was
very similar to that of the DMF effluent samples collected after 8 h in
the 2D plot (Fig. 6a). It could be explained by the bacterial community
structure of CF probably originating from the effluent produced by the
DMF process. However, according to the dendrogram, less than 75%
similarity was shared by the bacterial community structure between
the CF and other DMF effluent samples (Fig. 6b). This means that the
bacterial community structure of CF differs somewhat from those of
the DMF effluent samples even though the CF and DMF effluent samples
were very close in the 2D plot derived from the T-RFLP results (Fig. 6a).
In contrast to the CF, the bacterial community structure of the DMF-M
was independent as shown in the 2D plot and dendrogram (Fig. 6a
88
S. Jeong et al. / Desalination 385 (2016) 83–92
Fig. 4. (a) Variation of Ks-UF value during filtration (dotted line indicates logarithmic trendline), and (b) correlation between Ks-UF and AOC values (solid line indicates linear trendline).
and 6b). According to dendrogram, similarity of bacterial communities
between in DMF effluent produced at the initial operation after
backwashing and on DMF-M is around 40%. This indicates that some
of biofilm-forming bacteria developed on DMF-M are present in the filtered effluent. However, more than half of them still remained on filter
media after backwashing. This indicates that the unique biofilm community structure had developed on the DMF-M during its operation
even though this community was exposed to the DMF liquid phase (filtered seawater) and them to the CF media.
In summary, from the T-RFLP-based approach, interesting clustering
patterns of the bacterial community structure along with the operation
period and sampling positions were observed; DMF effluent, DMF-M,
and CF. However, the T-RFLP results provide only a limited understanding of the bacterial community structure because it cannot identify the
microorganisms present in the communities. For this reason, additional
16S-based pyrosequencing analysis (454-pyrosequencing) was conducted to identify the microorganisms present in the samples.
3.3. Microbial community composition in DMF effluent, and DMF and CF
media (from 454-pyrosequencing analysis)
3.3.1. Overall bacterial community structure
In this study, a total of 7299 bacterial sequences were generated
through the 454-pyrosequencing analysis and then assigned to 86
OTUs based on the 3% cutoff of genetic distance. The similarity of bacterial community structure was evaluated using the CA method based on
the relative abundance of OTUs (Fig. 7). The similarity of the two DMF
effluent samples, DMF-12 h and DMF-24 h, were evaluated and they
were nearly 100% identical. This result was consistent with that of the
T-RFLP-based CA result where — after 8 h of operation — the DMF
Fig. 5. Variation of ATP in filtered seawater by DMF during filtration (March 2013).
effluent samples were well clustered (Fig. 6b). These findings show
that the bacterial community structure in the DMF liquid phase did
not significantly change between 12 h and 24 h of the operational
period.
The bacterial community structure of the biofilm samples was significantly different from the DMF effluent samples (DMF-12 h and DMF24 h) because the CF and DMF-M were grouped at sub-level compared
to the effluent samples (Fig. 7). In particular, a unique bacterial community structure was observed in the CF (i.e. the similarity with nearest
samples (DMF-M) was only approximately 50%) as expected based on
the dendrogram derived from the T-RFLP results (Fig. 6b). However,
based on the results of the T-RFLP based-dendrogram shown in
Fig. 6b, the CF's bacterial community structure was more similar to the
DMF effluent samples than the DMF-M, which had a different clustering
pattern with pyrosequencing-based statistical results (Fig. 7). This phenomenon may be caused by the difference of resonance between the TRFLP and pyrosequencing. However, both dendrograms confirm that
the CF has a different bacterial community structure.
The CA results also demonstrate a similar tendency with the relative
abundance at the phylum level. For example, the phylum Proteobacteria
completely dominated the DMF-12 h and DMF-24 h samples and their
abundances were 96.5% and 98.9%, respectively. Although it was observed that the Proteobacteria was the dominant phylum in the CF
and DMF-M samples, its relative abundance was significantly lower
than that of the DMF-12 h and DMF-24 h. The major phylum consisted
of Proteobacteria (53.8%), Bacteroidetes (26.8%), and Firmicutes (18.4%)
for the CF and Proteobacteria (42.1%), Bacteroidetes (24.5%), and
Planctomycetes (8.7%) for the DMF-M. In addition, the diverse phylum
such as Deferribacteres (4.9%) and Gemmatimonadetes (4.6%) were
observed only in the DMF-M. Chun et al. [29] have noted that
Proteobacteria and Firmicutes were the dominant bacterial groups in
biofouled CFs.
3.3.2. Dominant bacterial composition
In order to facilitate a detailed discussion, the dominant bacterial
OTUs in each sample were identified based on the genus level and
then correlated to certain growth characteristics, for example substrate
type, oxygen utilization, salt tolerance and biofilm-forming potential.
Table 3 shows the dominant OTUs with the closest matched genus in
the DMF process. The dominant OTUs in both the DMF-12 h and DMF24 h were similar and as previously stated the dominant OTUs of the
biofilm samples differed significantly from those in the DMF effluent
samples. Furthermore the bacterial similarity between CF and DMF-M
was approximately 50%, which indicates that the CF and DMF-M had
significantly different bacterial community structures.
For the DMF effluent samples, i.e. DMF-12 h and DMF-24 h, the dominant OTUs were very similar. The four bacterial OTUs, OTU 13
S. Jeong et al. / Desalination 385 (2016) 83–92
89
Fig. 6. (a) The 2D ordination plot based on the T-RFLP results and (b) Dendrogram of the CA based on the T-RFLP results.
(Phaeobacter), OTU 19 (Oceaniserpentilla), OTU 21 (Pelagibacter), and
OTU 37 (Citrobacter), were ranked in the Top 5 species in both samples.
Of the OTUs, OTU 13 (Phaeobacter) and OTU 19 (Oceaniserpentilla) dominated the DMF effluent samples as presented in Table 3. The genus
Phaeobacter (which is chemoheterotrophic and obligately aerobic bacteria) grows the best in saline conditions (i.e. the optimum range is
from 0.2 to 0.68 M NaCl) [30]. The genus Oceaniserpentilla is also obligately aerobic bacteria while Oceaniserpentilla haliotis, a type strain of
Oceaniserpentilla, was isolated in the blacklip abalone Haliotis rubra,
which is a commonly found species in southern Australia. The optimal
growth of O. haliotis occurs at 2–8 °C and can be adapted to 22 °C [31].
Table 3
The dominant bacterial OTUs with closest genus in the DMF process.
Sample
CF
DMF-12 h
DMF-24 h
DMF-M
Fig. 7. CA based dendrogram of the relative abundance of OTU results.
Rank
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
OUT
Relative abundance (%)
Closet genus (Similarity %)
OTU 6
OTU 10
OTU 1
OTU 2
OTU 12
OTU 13
OTU 19
OTU 27
OTU 21
OTU 37
OTU 19
OTU 13
OTU 21
OTU 37
OTU 40
OTU 3
OTU 4
OTU 5
OTU 14
OTU 16
24.9
12.1
11.6
10.9
9.1
34.5
30.1
8.8
6.2
4.4
29.5
19.3
13.6
11.4
6.8
13.6
11.7
11.6
6.5
4.6
Pseudoalteromonas (99.3)
Bacillus (99.5)
Mesonia (99.3)
Olleya (97.4)
Cobetia (99.3)
Phaeobacter (99.8)
Oceaniserpentilla (99.8)
Bradyrhizobium (99.5)
Pelagibacter (99.8)
Citrobacter (99.6)
Oceaniserpentilla (99.8)
Phaeobacter (99.8)
Pelagibacter (99.8)
Citrobacter (99.6)
Sphingomonas (99.6)
Flavobacterium (98.5)
Desulfomicrobium (96.3)
Thioclava (92.8)
Ulvibacter (94.0)
Desulfovibrio (85.6)
90
S. Jeong et al. / Desalination 385 (2016) 83–92
The unique habitat characteristics of Oceaniserpentilla spp. depend on experimental location (in Australia) and operational season (in our case,
March 2013). Moreover, both genera, Phaeobacter and Oceaniserpentilla
are known as biofilm-forming bacteria [31–33]. These dominant bacterial
species probably influenced the reduced permeability in the DMF process
(Fig. 1).
Interestingly, the dominant bacterial OTUs were significantly different between the CF and DMF-M even though both biofilms were in contact with the DMF effluent samples. This difference was consistent with
that of the T-RFLP result. In the CF, OTU 6 (Pseudoalteromonas) was
dominant. Four other genera, namely, OTU 10 (Bacillus), OTU 1
(Mesonia), OTU 2 (Olleya), and OTU 12 (Cobetia) were also observed
in similar proportions (Table 3). The genus Pseudoalteromonas, which
is heterotrophic bacteria, is frequently found in the open ocean or coastline seawater [34]. The Pseudoalteromonas spp. is one of the extracellular polysaccharides (EPS) producing bacteria, which facilitates the
initiation of biofouling [35]. Like the genus Pseudoalteromonas, the
other dominant genera, Bacillus [36], Mesonia [35], Olleya [37] and
Cobetia [38], have biofilm-forming potential. This means that these
biofilm-forming bacteria could primarily cause the biofouling problem
in the CF being studied. Completely different genera were found in CF
from DMF effluent sample, which was different from the T-RFLP
analysis.
For the DMF-M, OTU 3 (Flavobacterium) was the most dominant
OUT, comprising 13.6% of the total bacterial community. The genus
Flavobacterium, which is aerobic chemoheterotrophs, is frequently discovered in freshwater condition. However, several Flavobacterium spp.
grown in the polar region are salt tolerant. For example, Flavobacterium
frigidarium, which was isolated from Antarctic seawater, can be
enriched in marine sediment conditions [39,40]. The genus
Flavobacterium is also one of the representative biofilm-forming bacteria in the desalination process [6]. Several sulfur-related bacteria were
observed to be the dominant OTUs in the DMF-M; two sulfatereducing bacteria (SRB), OTU 4 (Desulfomicrobium-like) and OTU 16
(Desulfovibrio-like), and one sulfate-oxidizing bacteria (SOB), OTU 5
(Thioclava-like). The SRB, which are anaerobic microorganisms, utilize
the sulfate as an electron acceptor to degrade organic compounds, so
they can easily grow in sulfur-rich conditions [41]. The DMF media (anthracite) had a sulfur content of ~2 wt.%, which could have encouraged
the growth of Desulfomicrobium-like and Desulfovibrio-like bacteria.
Sulfite produced by SRB can be utilized as the substrate to SOB, which
may promote the active growth of Thioclava-like bacteria on the DMF
media. Additionally, Thioclava can also provide favorable growth conditions to SRB consuming the dissolved oxygen in seawater. In a real
situation, the co-existence of SRB, Desulfovibrio spp., and SOB,
Thiomicrospira spp., was observed in the marine biofilm community in
a seawater flooding system [42]. Therefore, the unique bacterial community structure in DMF-M is probably attributed to the characteristics
of DMF media, such as S content.
3.4. Simulated DMF operation to reduce biofouling potential
3.4.1. Biological activity growth
ATP concentration was used as an indicator of biological growth on
the filter medium. After 14d of operation, samples were taken from
the top, middle (mid) and bottom (bot) layers of DMF column. Bacterial
growth on the DMF medium varied depending on backwashing intensity (A: 2 min and B: 10 min) and frequencies (intervals) (24 h and 35 h)
tested. As can be seen from Fig. 8a and 8b, a 24 h-BI with a shorter BD
(2 min) indicated a higher and more stable biological activity on the medium. In the case of DMF operation, a relatively high concentration of
ATP was detected in all the filter layers (Bot to Top: 0.329 ± 0.010 to
0.537 ± 0.032 nmol/cm3). On the other hand, 10 min-BD resulted in
less biological activity (less than 0.2 nmol/cm3 in all the layers) on the
medium (Fig. 8a). A 35 h-BI led to much biological growth in the top
layers regardless of BD (Fig. 8b). However, biological activity was not
maintained at the mid and bot layers of the DMF with 35 h-BI and
10 min-BD. This indicated that shorter and frequent backwashing
helped the biological growth on the DMF medium.
3.4.2. AOC reduction by DMF
AOC concentration in DMF filtered seawater was measured to evaluate how backwashing conditions impacted in reducing the bioavailable organics. AOC reduction was slightly higher (2 min-BD: 46% and
10 min-BD: 39%) with a 24 h-BI during operation lasting 14 days
(Fig. 8c). In particular, after 14 days of operation, AOC concentration decreased more than 50% with 2 min-BD (AOC in filtered seawater =
40.704 ± 0.102 μg-C glucose equivalents/L). On the other hand, longer
filter backwashing frequency (35 h-BI) and backwashing duration
(10 min-BD) resulted in less AOC reduction (around 20%) and even increased after 8d of operation (Fig. 8d).
Previous studies have proven that biofiltration is a feasible, attractive, environmentally friendly method of protecting against biofouling
through biological degradation (involving biological activity) and coagulation of the nutrients derived from particulate and dissolved organic
matter and minerals, which accumulate in biofouling [18,43,44]. DMF
filters at the desalination plant can be configured as a method of
biofiltration to remove a portion (20–40%) of the soluble bioavailable
organics present in the seawater [2].
The present study, therefore, suggests that backwashing in DMF operation (i.e. frequency and intensity) can improve the removal of bioavailable organic carbon through increased biological activity. DMF
operating in the biofiltration mode constitutes an environmentally
friendly strategy that reduces the biofouling problem.
4. Conclusions
The performance of DMF (used as a primary pretreatment at the seawater desalination plant) was monitored in terms of organic foulants
(DOC and organic fractions by LC-OCD) and biological foulants (AOC,
Ks-UF and ATP). The monitored parameters provided useful information in improving biological activity in DMF. LMW-N was found to
have a linear relationship with AOC. There is a linear relationship between AOC compounds and Ks-UF, indicating that Ks-UF can be used
as an indicative parameter of biofouling. ATP also showed a similar pattern. Laborious equipment is required for the measurement of AOC and
LMW-N (using LC-OCD). However, Ks-UF can be easily used as an onsite measurement method. This uses a simple filtration device similar
to SDI set-up.
Microbial communities in DMF filtered seawater (after 4–5 h —
stabilized operation time) affected the biofilm that formed in CF
(which is located after DMF). The bacterial community structure
was investigated using T-RFLP and pyrosequencing with statistical
methods. This microbial structure was clustered along with the operation period and sampling positions, i.e. DMF effluent, DMF-M,
and CF. In DMF effluent samples, OTU 13 (Phaeobacter) and OTU
19 (Oceaniserpentilla) dominated persistently at DMF-12 h and
DMF-24 h. In biofilm samples (taken both from DMF and CF
media), the bacterial community, which has biofilm-forming potential, was dominant. In particular, for the DMF-M, the sulfur-related
bacterial consortia dominated, this being mainly attributed to the
presence of sulfur in the media composition.
A detailed DMF pilot study conducted showed that optimization of
backwashing (frequency and duration) to operate in the biofiltration
mode led to a reduction in biological fouling. Based on pilot scale results
backwashing frequency of 24 h and duration of 2 min were found to be
the most suitable to reduce biofoulants by DMF. However it needs to be
optimized further. The optimization of DMF operating parameters to
operate in biological mode will lead to a feasible and environmentally
friendly operation of DMF at desalination plants.
S. Jeong et al. / Desalination 385 (2016) 83–92
91
Fig. 8. Biological activity on DMF media after 14d of operation (a: 24 h-BI and b: 35 h-BI) and variation of AOC concentration in filtered seawater by DMF (c: 24 h-BI and d: 35 h-BI) (RSW =
raw seawater, A: 2 min-BD and B: 10 min-BD).
Acknowledgments
This study was supported by the National Centre of Excellence in Desalination Australia (NCEDA), which was funded by the Australian Government through Water for the Future initiative (Project code: 09148).
The authors also acknowledge Mr. Damien Marwick (Degremont) and
Ms. Ushi Jismi (Water Corporation) for their contribution to the sampling. PSDP is owned by Water Corporation while the operation and
maintenance is conducted in alliances with Degremont and Suez
Environment.
References
[1] N. Prihasto, Q.-F. Liu, S.-H. Kim, Pre-treatment strategies for seawater desalination
by reverse osmosis system, Desalination 249 (2009) 308–316.
[2] N. Voutchkov, Considerations for selection of seawater filtration pretreatment system, Desalination 261 (3) (2010) 354–364.
[3] S. Jamaly, N.N. Darwish, I. Ahmed, S.W. Hasan, A short review on reverse osmosis
pretreatment technologies, Desalination 354 (2014) 30–38.
[4] K. Gaid, A Large Review of the Pretreatment, INTECH Open Access Publisher, 2011.
[5] L.O. Villacorte, S.A.A. Tabatabai, D.M. Anderson, G.L. Amy, J.C. Schippers, M.D.
Kennedy, Seawater reverse osmosis desalination and (harmful) algal blooms, Desalination 360 (2015) 61–80.
[6] A. Matin, Z. Khan, S.M.J. Zaidi, M.C. Boyce, Biofouling in reverse osmosis membranes
for seawater desalination: phenomena and prevention, Desalination 281 (2011)
1–16.
[7] L.F. Greenlee, D.F. Lawler, B.D. Freeman, B. Marrot, P. Moulin, Reverse osmosis desalination: water sources, technology, and today's challenges, Water Res. 43 (2009)
2317–2348.
[8] S. Latteman, Development of an Environmental Impact Assessment and Decision
Support System for Seawater Desalination Plants, Technical University Delft, The
Netherlands, 2010.
[9] N. Palmer, A review of desalination in Australia, IWA Desal Newsletter, 2012.
[10] G. Crisp, E.A. Swinton, N. Palmer, A brief review of desalination in Australia in 2010,
Int. J. Nucl. Desalin. 4 (1) (2010) 66–75.
[11] A. Alhadidi, A.J.B. Kemperman, B. Blankert, J.C. Schippers, M. Wessling, W.G.J. Van
der Meer, Silt density index and modified fouling index relation, and effect of pressure, temperature and membrane resistance, Desalination 273 (1) (2011) 48–56.
[12] S. Jeong, S. Vigneswaran, Assessment of biological activity in contact flocculation filtration used as a pretreatment in seawater desalination, Chem. Eng. J. 228 (2013)
976–983.
[13] S.A. Huber, A. Balz, M. Abert, W. Pronk, Characterisation of aquatic humic and nonhumic matter with size-exclusion chromatography–organic carbon detection–
organic nitrogen detection (LC-OCD-OND), Water Res. 45 (2) (2011) 879–885.
[14] S. Jeong, G. Naidu, S. Vigneswaran, C.H. Ma, S.A. Rice, A rapid bioluminescence-based
test of assimilable organic carbon for seawater, Desalination 317 (2013) 160–165.
[15] S. Jeong, S. Vigneswaran, Practical use of standard pore blocking index as an indicator of biofouling potential in seawater desalination, Desalination 365 (2015) 8–14.
[16] C.H. Wei, G. Amy, Membrane fouling potential of secondary effluent organic matter
(EfOM) from conventional activated sludge process, J. Membr. Sep. Technol. 1 (2)
(2012) 129–136.
[17] S. Velten, M. Boller, O. Köster, J. Helbing, H.U. Weilenmann, F. Hammes, Development of biomass in a drinking water granular active carbon (GAC) filter, Water
Res. 45 (19) (2011) 6347–6354.
[18] S. Jeong, S.A. Rice, S. Vigneswaran, Long-term effect on membrane fouling in a new
membrane bioreactor as a pretreatment to seawater desalination, Bioresour.
Technol. 165 (2014) 60–68.
[19] S.W. Culman, R. Bukowski, H.G. Gauch, H. Cadillo-Quiroz, D.H. Buckley, T-REX: software for the processing and analysis of T-RFLP data, BMC Bioinf. 10 (2009) 171.
[20] H.D. Ishak, R. Plowes, R. Sen, K. Kellner, E. Meyer, D.A. Estrada, S.E. Dowd, U.G. Mueller,
Bacterial diversity in Solenopsis invicta and Solenopsis geminata ant colonies characterized by 16S amplicon 454 pyrosequencing, Microb. Ecol. 61 (4) (2011) 821–831.
[21] R.C. Edgar, Search and clustering orders of magnitude faster than BLAST, Bioinformatics 26 (19) (2010) 2460–2461.
[22] R.C. Edgar, B.J. Haas, J.C. Clemente, C. Quince, R. Knight, UCHIME improves sensitivity
and speed of chimera detection, Bioinformatics 27 (16) (2011) 2194–2200.
[23] S. Jeong, H. Bae, G. Naidu, D. Jeong, S. Lee, S. Vigneswaran, Bacterial community
structure in a biofilter used as a pretreatment for seawater desalination, Ecol. Eng.
60 (2013) 370–381.
[24] F.X. Simon, Y. Penru, A.R. Guastalli, S. Esplugas, J. Llorens, S. Baig, NOM characterization by LC-OCD in a SWRO desalination line, Desalin. Water Treat. 51 (7–9) (2013)
1776–1780.
[25] S.A. Baghoth, S.K. Sharma, G.L. Amy, Tracking natural organic matter (NOM) in a
drinking water treatment plant using fluorescence excitation–emission matrices
and PARAFAC, Water Res. 45 (2) (2011) 797–809.
92
S. Jeong et al. / Desalination 385 (2016) 83–92
[26] M.G. Quintana, O.D. Salomón, M.L. De Grosso, Distribution of phlebotomine sand
flies (diptera: psychodidae) in a primary forest-crop interface, Salta, Argentina, J.
Med. Entomol. 47 (6) (2010) 1003–1010.
[27] B. Liu, L. Gu, X. Yu, G.Z. Yu, H.N. Zhang, J.L. Xu, Dissolved organic nitrogen (DON) profile during backwashing cycle of drinking water biofiltration, Sci. Total Environ. 414
(2012) 508–514.
[28] X. Liao, C. Chen, J. Zhang, Y. Dai, X. Zhang, S. Xie, Operational performance, biomass
and microbial community structure: impacts of backwashing on drinking water
biofilter, Environ. Sci. Pollut. Res. 22 (1) (2015) 546–554.
[29] Y. Chun, P.T. Ha, L. Powell, J. Lee, D. Kim, D. Choi, R.W. Lovitt, I.S. Kim, S.S. Mitra, I.S.
Chang, Exploring microbial communities and differences of cartridge filters (CFs)
and reverse osmosis (RO) membranes for seawater desalination processes, Desalination 298 (2012) 85–92.
[30] T. Martens, T. Heidorn, R. Pukall, M. Simon, B.J. Tindall, T. Brinkhoff, Reclassification
of Roseobacter gallaeciensis Ruiz-Ponte et al. 1998 as Phaeobacter gallaeciensis gen.
nov., comb. nov., description of Phaeobacter inhibens sp. nov., reclassification of
Ruegeria algicola (Lafay et al. 1995) Uchino et al. 1999 as Marinovum algicola gen.
nov., comb. nov., and emended descriptions of the genera Roseobacter, Ruegeria
and Leisingera, Int. J. Syst. Evol. Microbiol. 56 (6) (2006) 1293–1304.
[31] A. Schlösser, A. Lipski, J. Schmalfuß, F. Kugler, G. Beckmann, Oceaniserpentilla
haliotis gen. nov., sp. nov., a marine bacterium isolated from haemolymph serum
of blacklip abalone, Int. J. Syst. Evol. Microbiol. 58 (9) (2008) 2122–2125.
[32] I. Vandecandelaere, O. Nercessian, M. Faimali, E. Segaert, A. Mollica, W. Achouak, P.
De Vos, P. Vandamme, Bacterial diversity of the cultivable fraction of a marine
electroactive biofilm, Bioelectrochemistry 78 (1) (2010) 62–66.
[33] H. Dang, R. Chen, L. Wang, S. Shao, L. Dai, Y. Ye, L. Guo, G. Huang, M.G. Klotz, Molecular characterization of putative biocorroding microbiota with a novel niche detection of epsilon- and zetaproteobacteria in Pacific Ocean coastal seawaters, Environ.
Microbiol. 13 (11) (2011) 3059–3074.
[34] C. Holmström, S. Kjelleberg, Marine pseudoalteromonas species are associated with
higher organisms and produce biologically active extracellular agents, FEMS
Microbiol. Ecol. 30 (4) (1999) 285–293.
[35] H. Dang, T. Li, M. Chen, G. Huang, Cross-Ocean distribution of rhodobacterales bacteria as primary surface colonizers in Temperate Coastal Marine waters, Appl. Environ. Microbiol. 74 (1) (2008) 52–60.
[36] H. Sarvamangala, J. Gopal, P. Muraleedharan, R.P. George, R.K. Dayal, K.A. Natarajan,
Biomineralization of manganese by bacillus spp isolated from a marine biofilm,
Miner. Metall. Process. 25 (3) (2008) 149–155.
[37] C.M. Nichols, J.P. Bowman, J. Guezennec, Olleya marilimosa gen. nov., sp. nov., an
exopolysaccharide-producing marine bacterium from the family Flavobacteriaceae,
isolated from the Southern Ocean, Int. J. Syst. Evol. Microbiol. 55 (4) (2005)
1557–1561.
[38] S. Mieszkin, P. Martin-Tanchereau, M.E. Callow, J.A. Callow, Effect of bacterial
biofilms formed on fouling-release coatings from natural seawater and cobetia marina, on the adhesion of two marine algae, Biofouling 28 (9) (2012) 953–968.
[39] J.-F. Bernardet, J. Bowman, in: M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, E.
Stackebrandt (Eds.), The Prokaryotes, Springer New York 2006, pp. 481–531.
[40] Y. Nogi, K. Soda, T. Oikawa, Flavobacterium frigidimaris sp. nov., isolated from Antarctic seawater, Syst. Appl. Microbiol. 28 (4) (2005) 310–315.
[41] G. Muyzer, A.J.M. Stams, The ecology and biotechnology of sulphate-reducing bacteria, Nat. Rev. Microbiol. 6 (6) (2008) 441–454.
[42] M.A. Al-Moniee, I. Chatterjee, G. Voordouw, P.F. Sanders, T.Y. Rizk, Microbial Community structure in a seawater flooding system in Saudi Arabia, Saudi Aramco J.
Technol. Spring (2013) 46–51.
[43] D.B. Mosqueda-Jimenez, P.M. Huck, Effect of biofiltration as pretreatment on the
fouling of nanofiltration membranes, Desalination 245 (1) (2009) 60–72.
[44] G. Naidu, S. Jeong, S. Vigneswaran, S.A. Rice, Microbial activity in biofilter used as a
pretreatment for seawater desalination, Desalination 309 (2013) 254–260.