Separation and Purification Technology 272 (2021) 118921

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur

Ultrafiltration as tertiary treatment for municipal wastewater reuse

J. Yang a, M. Monnot a, T. Eljaddi a, L. Ercolei b, L. Simonian a, P. Moulin a, *
a
Aix-Marseille Univ., Laboratoire de Mécanique, Modélisation et Procédés Propres, Equipe Procédés Membranaire (EPM-M2P2-CNRS-UMR 7340), Europôle de l’Arbois,
BP 80, Bat. Laennec, Hall C, 13545 Aix-en-Provence cedex 04, France
b
Société des Eaux de Marseille Métropole, 25 Rue Edouard Delanglade, B.P. 29 – 13006, Marseille Cedex 06, Marseille, France

A R T I C L E I N F O A B S T R A C T

Keywords: Water reuse is an enduring topic that benefits the society and future generations of mankind. Ultrafiltration (UF)
Ultrafiltration is one of the most cost-effective treatment technologies for improving water quality. In this study, a semi-
Tertiary treatment industrial UF pilot plant with periodical classic backwash (CB) and air backwash (AB) was operated automati­
Wastewater treatment
cally to evaluate its feasibility and sustainability for municipal wastewater reuse and find out the optimized
Water reuse
Fouling management
filtration condition. This study carried out 15 filtration conditions to investigate the impacts of flux (J in
L⋅h− 1⋅m− 2), filtration cycle time (t in min), and air backwash frequency (BW) on membrane hydraulic filtration
performance and membrane fouling management. Through comparative analysis of all conditions in water
quality, permeability variation, irreversible fouling management, and water recovery rates, the sustainable
conditions J80t40BW1/3 (flux of 80 L⋅h− 1⋅m− 2, filtration cycle time of 40 min, 1 AB followed with 3 CBs),
J60t60BW1/4 (flux of 60 L⋅h− 1⋅m− 2, filtration cycle time of 60 min, 1 AB followed with 4 CBs), and J60t60BW1/
3 (flux of 60 L⋅h− 1⋅m− 2, filtration cycle time of 60 min, 1 AB followed with 3 CBs), stood out from the others with
higher overall performances. Additionally, air backwash showed excellent reversibility on membrane fouling
control, which was around 1.25–2 times that of CB in average. After all, long term operation on condition
J60t60BW1/3 in winter and in summer confirmed that the UF system could provide sustainable and adaptable
filtration performance regardless of the temperature and feed water quality. The UF permeate quality is good
enough to be reused in non-potable applications as it met reuse guidelines of the World Health Organization,
reuse standards of France and the most recent EU regulation for agricultural irrigation. This work confirms the
great interest of UF as tertiary treatment for water reuse and gives operational indications for future industrial-
scale production of reclaimed water.

1. Introduction membrane filtration is one of the main unit technologies. Generally,

ultrafiltration (UF) has been widely accepted as one of the most cost-
Water scarcity has long been a worldwide crisis under the increasing effective technologies to deal with secondary treated effluent on non-
pressure of population growth, climate change, and various pollutions. potable reuses, such as agricultural irrigation, landscape irrigation,
Water reuse, defined as using wastewater or reclaimed water from one urban reuse, and process water for industries [4–7]. UF enables high
application to another application, is a promising approach to alleviate removal efficiency on total suspended solids (TSS), turbidity, organic
water crisis sustainably. However, the total volume of reused water all matters, and microorganisms, etc [4,8]. On reuse purpose, the studies by
over the world, which is approximately 14.2 billion m3⋅y− 1, represents Muthukumaran et al. [9], and Pollice et al. [10] or Falsanisi et al. [11] all
less than 4% of the total volume of domestic wastewater [1]. Moreover, confirmed that UF (with prefilter) can support qualified permeate water
the reuse proportion differs a lot among different countries and regions. for reuse, all with respect to World health organization (WHO) reuse
In Europe, approximately 2.4% of the total treated wastewater effluents guidelines. However, regulations are becoming stricter in terms of
was reused in 2015 [2], while France only reused around 1% [3]. Water concentrations of E. coli, biological oxygen demand (BOD5), TSS and
reuse market is still young and limited, which needs to be highly turbidity such as in the higher reclaimed water quality class (A) of the
improved in the future. Nowadays, more and more advanced treatment Regulation (EU) 2020/741 of the European Parliament and of the
technologies for municipal wastewater reuse have emerged, and Council of 25 May 2020 on minimum requirements for water reuse [12].

* Corresponding author.
E-mail address: [email protected] (P. Moulin).

https://doi.org/10.1016/j.seppur.2021.118921
Received 23 February 2021; Received in revised form 26 April 2021; Accepted 5 May 2021
Available online 9 May 2021
1383-5866/© 2021 Elsevier B.V. All rights reserved.
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

Therefore, there is a need to confirm if UF can meet these new standards. Table 1
In addition, fouling formation limits the economical operation of UF Raw wastewater and UF feed quality.
process on wastewater treatment, [13]. Generally, the total fouling Parameters Campaign 1 Campaign 2
resistance in UF filtration is mainly composed of hydraulic reversible (October 2019 – March (July 2020)
resistance and irreversible resistance [14]. The macro-solute or particle 2020)
deposition on top of membrane surface mainly caused by suspended WWTP Outlet WWTP Outlet
solids is usually reversible and non-adhesive [15]. The pore blocking influent effluent (UF influent effluent (UF
and internal pore adsorption on membranes caused by organic matters is feed) feed)

usually irreversible and adhesive [16]. To control fouling, filtration E. coli (CFU 100 mL− 1) 1.6 × 108 (3.4 ± 2.6) > 1.4 × (8.5 ± 9.4)
conditions (flux, filtration time, cleaning) need to be optimized because × 104 106 × 103
7
Enterococci (CFU 100 2.2 × 10 (1.3 ± 1.0) (3.1 ± (8.6 ± 9.5)
they can directly affect either the formation and nature of the fouling
mL− 1) × 104 3.4) × 107 × 102
layer, or the removal mechanism of the foulants. For example, the Anaerobic sulphito- 5.6 × 103 268 ± 253 (1.3 ± (9.6 ± 5.1)
filtration flux and filtration cycle time will affect the density and the reducers (spores) 1.7) × 104 × 102
thickness of the fouling layer deposited on the membranes [17]. Besides, (CFU 100 mL− 1)
frequent cleanings for membranes are required, including physical and Specific F-RNA 4.5 × 103 < 30 (1.7 ± < 30
bacteriophages(PFP 1.8) × 103
chemical washing, to maintain good filtration performance. Chemical 100 mL− 1)
washing is effective on irreversible foulants removal through the reac­ COD (mgO2⋅L− 1) 1124 20 ± 9 552 ± 61 45 ± 21
tion between fouling and chemicals. The choice of chemical agents, BOD5 (mgO2⋅L− 1) n.m. n.m. 205 ± 82 <3
mostly using acids, alkalis, or oxidants, is based on the membrane ma­ TSS (mg⋅L− 1) 77 4±2 367 ± 12.1 ± 8
166
terials and fouling types [18]. Physical cleanings such as classic back­
TOC (mgC⋅L− 1) n. m. 18 ± 9 n.m. n.m.
wash (CB) and air assisted backwash (AB) are commonly used methods Turbidity (NTU) n. m. 2.3 ± 0.9 208 ± 38 2.1 ± 1.6
in membrane fouling control [19]. CB can loosen and detach fouling pH n. m. 7.2 ± 0.4 7.4 ± 0.3 7.5 ± 0.1
cake partially from membrane surface, and AB is a more effective anti- Conductivity (μS⋅cm− 1) n. m. 1 168 ± 128 n.m. n.m.
fouling method compared to CB [15,20,21]. Normally, AB can be per­ CFU: colony-forming unit; PFP: Polyhedral, filamentous, and pleomorphic; n.
formed either by air sparing in backwash water, or by air injection into m.: not measured.
membrane fibers. A previous study by Y. Ye et al [20] on a model COD: Chemical oxygen demand; BOD5: Biochemical oxygen demand TSS: Total
mixture of bentonite and alginate demonstrated that the membrane suspended solids; TOC: Total organic carbon.
system using air bubble backwash in moderate air flow rate (e.g. 154
mL⋅min− 1) could remove more foulants without recompression poten­ The standard deviations (SD) were measured with 2–4 samples.
tial while the loosen and residual foulants after CB would be recom­
pressed on membrane surface when filtration restarts. P.J. Remize et al.
2.2. Ultrafiltration pilot plant description
[22] also stated that AB with air injection enhanced particles removal
during long-term filtration in drinking water production. Additionally,
A semi-industrial UF pilot plant manufactured from Aquasource
C. Cordier et al. [23] confirmed the higher removal ability of AB with air
(France) has been established for this study. The nominal capacity of the
injection, and it also destroyed the cells’ integrity by air injection, such
pilot plant is 20 m3 water per day. The pilot plant can be performed in
as oysters’ oocytes and spermatozoa in seawater. Frequent AB can bring
manual mode, half automatic mode and automatic mode. Operating
some other disadvantages such as pore blocking inside membranes [20];
parameters shown in the flow diagram of the pilot plant in Fig. 1 are
partial drying which can cause membrane integrity problems [19]; more
recorded automatically every minute. The secondary effluent of WWTP
energy and equipment cost requirements. Considering the pros and cons
as feed will be drawn into a feed tank T3 when the feed water drops to
of AB and CB, it is therefore interesting to alternate cleaning by CB and
the lowest settled level in the pilot plant system. Tanks T1 and T2 are
AB during filtration and it is very important to study the optimum
containers for permeate and backwash water. A concentration of 5 ppm
backwash frequency of AB and CB as it has never been done on
of chlorine was added in T1 while in T2, corresponding to the water
municipal secondary wastewater effluent. In addition to optimum AB
production tank, no chlorine was added . A 130 μm disk prefilter is
frequency, filtrations conditions such as flux and filtration time have
designed before UF membrane to avoid UF hollow fiber clogging.
also never been optimized in the scientific literature for this effluent
whereas these are crucial to control fouling and enable the wide
development of water reuse by UF. 2.3. UF membrane and module
In this context, a semi-industrial UF pilot plant was set-up on-site
with real secondary effluent for reuse application and was operated with The membrane module is an ALTEONTM I (Aquasource, France)
periodic CB and AB. The targets of this study are: (i) to check the multichannel UF hollow fiber, with detailed information listed in
permeate quality for reuse according to EU regulation adopted in May Table 2. The module is operated in dead-end filtration mode with an
2020 among others, (ii) to find out the sustainable filtration conditions inside-out configuration. Dead-end ultrafiltration is considered as an
(flux, filtration cycle time, AB frequency) enabling sustainable flux and energetic efficient operation mode in large-scale water treatment sys­
low operating costs, (iii) to check the UF performances as tertiary tems [24]. The hydraulic resistance of new membrane modules was
treatment for municipal wastewater reuse in the long term. measured with pure water to be 4 × 1011 m− 1 at 20 ◦ C (pure water
permeability = 900 L⋅m− 2⋅h− 1⋅bar− 1). The maximum permeability (Lp)
2. Material and methods of UF membrane when filtrated with the feed water is around 600
L⋅m− 2⋅h− 1⋅bar− 1 at 20 ◦ C. Therefore, each filtration condition was
2.1. Feed water quality started from initial Lp at around 600 L⋅m− 2⋅h− 1⋅bar− 1, and the corre­
sponding membrane resistance of 6 × 1011m− 1 at 20 ◦ C was considered
The feed water is the secondary effluent of a municipal wastewater for future calculations.
treatment plant (WWTP) located in Châteauneuf-les-Martigues, France.
The WWTP uses a conventional activated sludge process to treat raw 2.4. Membrane cleaning
wastewater. Two experimental campaigns were conducted in winter and
summer. Table 1 shows the quality of the inlet and outlet effluent of the There are four types of membrane cleanings in the system: (1)
WWTP. The outlet effluent of the WWTP was the UF feed of this work. Classical backwash (CB) using permeate water from T2 without chlorine

2
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

Fig. 1. Diagram of semi-industrial pilot plant.

with 2.0 m3⋅h− 1 – 3.0 m3⋅h− 1 of feed water from bottom to top of the UF
Table 2
module. The forward flushing can assist to decrease aggregation or
UF membrane and module characteristics.
dense attachment of particulates from the membrane surface [26].
Characteristics Data

Material Polyethersulfone
2.5. Filtration conditions
MWCO 200 kDa
Pore size 0.02 µm
Length 1.2 m The experiments were operated under constant flux mode (20 to 100
Internal diameter 0.9 mm L⋅h− 1⋅m− 2). During three months’ operation, the study conducted
Number of channels 7 various filtration conditions to investigate the impact of filtration cycle
Filtration Surface 9 m2
time, AB frequency, and flux on hydraulic filtration performances. The
Volume of fibers 2.0 L
Maximum TMP 2.5 bar AB frequency , represented by “1/n” in the following contents, such as
pH tolerant value 1–13 1/3, 1/4, 1/6, and 1/9, represents (n) times of CBs followed with 1 AB.
15 Different designed conditions were studied in this test and were
separated into 3 main parts to study the impact of filtration cycle time,
with a flow rate of 2.5 m3⋅h− 1. The duration and the volume of water
flux, and AB frequency on filtration performance, respectively (Table 3).
used during one CB is 52 s and 36 L, respectively. (2) Air backwash (AB)
Each condition was conducted continuously for >40 h for stable
includes 3 steps: empty fibers, dry the fouling layer with air injection,
behavior. After each condition, several CEBs were operated manually to
and CB using permeate water from T1. This type of backwash was
clean the membrane until the initial permeability reached to 600
described by Cordier et al. [23,25]. The duration and the volume of
L⋅m− 2⋅h− 1⋅bar− 1, so as to maintain the similarly initial membrane state
water used during an AB is 67 s and 52 L, respectively. (3) Before new
for all filtration conditions. To be noted, the parameters including flux
operating conditions start or if the permeability dropped below 200
and Lp, that could be affected by the temperature have been normalized
L⋅m− 2⋅h− 1⋅bar− 1 at real temperature, the chemical enhanced backwash
to a standard temperature (20 ℃) to account viscosity fluctuations with
(CEB) was conducted. Before CEB, the system carries out a backwash
these parameters.
related to the backwash sequence that is in turn. The CEB starts with
chemical injection into membrane modules and then soaking happens
with chemicals to degrade foulants. Chemicals are either sulfuric acid 2.6. UF performance analysis
([H+] = 1 000 ppm), or sodium hydroxide ([OH–] = 800 ppm) and
chlorine ([active Cl] = 50 ppm). The duration of soaking is 1200 s. After 2.6.1. Permeability
soaking, the system discharges the soaking solutions by gravity and then As the flux J is constant in the semi-industrial system, the relation­
starts with a long-term CB without chlorine until pH of discharge water ship between J (L⋅m− 2⋅h− 1) and Lp (L⋅m− 2⋅h− 1⋅bar− 1) can be expressed
comes back to neutral. The duration and the volume of water used as Eq. (1).
during a CEB is 1560 s and 250 L, respectively. (4) In addition, a forward J
flushing happens in the middle of each filtration cycle for about 10 s Lp = (1)
TMP

3
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

Table 3 (n+1)
Filtration conditions. TMPnend − TMP(n+1) Rn − Rini
Reversibility(n) = ini
= endn (7)
TMPnend − TMPnini Rend − Rnini
Flux (J) Filtration AB frequency
(L⋅h− 1⋅m− 2) time (t) (min) (Air Backwash/
Classic 2.6.4. Water recovery rate
Backwash) The water recovery rate represents the proportion of net produced
Impact of Filtration cycle 60 20 1/3 water quantity when compared to the total inlet feed water quantity.
time on filtration 30 This was calculated with Eq. (8).
performance 40
60 Vnet Vinlet − VCB − VAB − VCEB − VCon
80 30 Recovery rate = × 100% = × 100%
Vinlet Vinlet
40
60
(8)
Impact of Flux on 30 30 1/3
filtration performance 60 where Vinlet refers to the volume of inlet feed water (m3), Vnet refers to
80 the net volume produced (m3). The calculation of Vinlet is related to the
100 filtration flux and filtration time, as in Eq. (9). The volume unit is the
60 60 1/4
liter.
80
Impact of AB frequency on 60 60 1/3 Vinlet = J × tf (9)
filtration performance 1/4
1/6 VCon refers to the concentrate volume, which is known to be 2 L
1/9 (volume of fibers presented in Table 2). VCB, VAB, and VCEB represent the
No AB
volume of permeate consumed for CB, AB and CEB which are 36 L, 52 L
and 250 L, respectively.
The pressure difference over a membrane is called the Trans­
membrane Pressure (TMP), unit bar, which is calculated by Eq. (2) 2.6.5. Impact of air backwash on irreversible resistance increase
To find out the influence of air backwash frequency on UF membrane
Pfeed + Pconcentrate
TMP = − Ppermeate (2) filtration performance, the variation of irreversible fouling during
2
operation can be an interesting parameter to follow. The relationship of
Rirr/Rm variation with feed water turbidity integration (Eq. (10)) will be
2.6.2. Fouling resistances
considered to investigate the influence of AB frequency on fouling for­
According to Darcy’s law [27], hydraulic resistance of the fouled
mation whatever the quality of the feed water:
membrane was measured with Eq. (3):
(∑∫ )
(Rirr /Rm )n
TMP 1 =f Turbidity.dt (10)
Rt = Rirr + Rre + Rm = = (3) (Rirr /Rm )initial
μ⋅J μ⋅Lp
where the turbidity integration with time is shown as follows with n =
where, TMP is the transmembrane pressure (Pa), μ is the viscosity of
number of filtration cycle, and t the duration of a filtration cycle:
permeate (Pa⋅s) and J is the applied flux (m⋅s− 1). The total membrane
resistance (Rt) includes three parts of resistances (m− 1): irreversible
⎡ t ⎤
(∑∫ ) ∑n ∫
fouling resistance (Rirr) which cannot be removed by backwashes, f Tur.dt = ⎣ Turbidity.dt⎦ (11)
reversible fouling resistance (Rre) which can be removed by backwashes, i=1
0 i
and membrane resistance (Rm).
The flux at 20◦ C is calculated through Eq. (4) [28]:
2.7. Water quality assessment
μ(T)
J(20◦ C) = ⋅J(T) = J(T) × e[0.0239×(20− T) ]
(4)
μ(20◦ C) Total suspended solids (TSS) were measured with a 47 mm glass-
In two adjacent filtration cycles, it can be considered that the fiber-filter (Whatman). The filter was firstly rinsed with distilled water
composition of fouling resistance in end of cycle (n-1) is: and dried during minimum 2 h at 105 ◦ C after filtration of the sample.
The TSS concentrations were then obtained by calculating the difference
Rend (n − 1) = Rm + Rirr (n − 1) + Rre (n − 1) (5) of weight before and after filtration divided by the volume of sample.
While after a CB, the total resistance in the beginning of cycle (n) is: The turbidity of feed was measured and recorded every minute by the
pilot plant using a probe TurbiMax W CUS31 (Endress Hauser). Besides,
Rini (n) = Rm + Rirr (n − 1) (6) the turbidity of permeate was also tested punctually in laboratory with a
Through difference in Eq. (7) (8), values of Rirr and Rre in the n-1 turbidity meter (Turb 550 IR, WTW, Germany). The electrical conduc­
filtration cycle can be found. tivity value was determined with a conductometer (Sension + EC7,
Hach, USA). The pH was measured with a pH-meter (Sension + pH31,
2.6.3. Reversibility Hach, USA). During Campaign 1, COD analysis was conducted through
Filtration performance can be evaluated by parameters such as reagent vials (COD cell test C3/25), with test range from 10 to 150
backwash effectiveness, fouling rate, and fouling resistance[29]. Back­ mg⋅L− 1. A volume of 3 mL of a homogeneous water sample was injected
wash effectiveness can be indicated by fouling reversibility which was into a COD digestion reagent vial. Then, the vial contents were stirred
calculated after each filtration cycle (n) according to [19,30]. Revers­ and subsequently placed in the COD reactor at 148 ◦ C for 2 h. Finally,
ibility after each filtration cycle could then be calculated using the initial the samples after cooling were tested by a spectrophotometer (190 –
TMP and final TMP values of the cycle (n) (TMPnend and TMPnini ) as well as 1100 nm, Photolab 6600 UV–Vis, WTW). A TOC-L machine (Shimadzu,
Japan) based on the 680 ◦ C combustion catalytic oxidation method was
the initial TMP of the next filtration cycle (TMPini ). As described
(n+1)
used to measure the concentration of total organic carbon (TOC). The
above, the flux and viscosity at 20 ◦ C are both constant during each
non-purgeable organic carbon (NPOC) method with a detection limit of
filtration condition. Therefore, the reversibility can be calculated as
4 μg⋅L− 1 was used. During Campaign 2, COD, BOD5, microorganisms
follow Eq. (7).
including SARS-CoV-2 were all tested by an external laboratory (Carso –

4
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

Laboratoire Santé Environnement Hygiène de Lyon) certified by the French because it is related to the concentration of suspended solids. Fig. 2
Ministry of Health for water analysis. The methods for COD and BOD5 shows the variations of feed water turbidity and raw wastewater flow­
are respectively ISO 15,705 and NF EN ISO-5815–1. The microbiological rate versus time for a standard day. The feed water quality variation
analyses were also tested by Carso laboratory and consisted in the seems positively correlated to the variations of raw wastewater flow rate
measurements of E. Coli (NF EN ISO 9308–1, NF EN ISO 9308–3), as the two curves follow similar trends. Therefore, the feed water
Enterococci (NF EN ISO 7899–1, NF EN ISO 7899–2), anaerobic sulphito- turbidity varies regularly everyday which has the similar peak and
reducers (spores) (NF EN 26461–2), specific F-RNA bacteriophages (NF trough periods with flow rate variation of raw wastewater: this was
EN ISO 10705–1), and SARS-CoV-2 by real-time quantitative polymer­ undergone during all experiments.
ase chain reaction (qPCR). The detection methods NF EN ISO 9308–1 for The variations of permeability and feed water turbidity obtained in
E. coli, and NF EN ISO 7899–2 for Enterococci were used in samples condition J60t20BW1/3 (J = 60 L⋅h− 1⋅m− 2, t = 20 min, 1 AB with 3 CB)
during Campaign 1 with both detection limits at < 1 CFU⋅100 mL− 1. The are shown in Fig. 3. The variations of Lp showed several unregular over-
detection methods NF EN ISO 9308–3 for E. coli, and NF EN ISO 7899–1 increase periods, such as in circles (a) and (b) which correspond to the
for Enterococci are used in samples of Campaign 2 with both detection moments when the turbidity became lower (trough periods). It can be
limits at < 56 CFU⋅100 mL− 1. For SARS-CoV-2 analysis, the envelope considered that trough periods with lower feed turbidity (Fig. 2) brought
protein gene E is first tested and the nucleoprotein gene N is tested only less fouling load on membrane thus the permeability decreased slower
if the envelope protein gene E is detected. The detection and identifi­ than in peak periods. In trough periods, the feed water contains less
cation of microplastics were analyzed by Spotlight 400 Fourier-trans­ turbidity, suspended solids, and probably also fewer organic matters
form infrared spectroscopy (FT-IR) imaging system from PerkinElmer which results in lower foulant potential. Therefore, with periodical
after filtration of the samples through a 3 μm gold-coated filter. Before backwash (1 AB with 3 CB), the UF membrane would be cleaned more
filtration, the water samples are first mixed with hydrogen peroxide thoroughly in trough periods than that in peak periods, resulting in
solution (30 % w/v, fisher chemical) in a ratio of 1:1 for one day to higher permeability recovery. The above analysis demonstrates that the
remove microorganisms. Data treatment is realized by siMPle (System­ over-increase of permeability during long term filtration is normal and
atic Identification of MicroPlastics in the Environment) software regular, and usually occurs at the trough periods during a day, like the
developed by Aalborg University, Denmark and Alfred Wegener Insti­ periods (a) and (b) in Fig. 3. The variations shown in Fig. 2 and Fig. 3 for
tute, Germany. For all analyses described previously, inlet and outlet two conditions are representatives from the others. All the conditions
(UF feed) of the WWTP and permeate were sampled and analyzed almost are faced with the same or similar impacts from external environment.
at the same time.
4. Permeate quality
3. Results and discussion
Permeate quality analysis is critical to the realization of water reuse.
3.1. Variation of feed water quality To ensure the safety and reliability of reused water in the frame of this
study, the permeate quality was compared to the three kind of guidelines
The flowrate of raw wastewater in the WWTP changes regularly and standards. Table 4 shows the UF permeate quality during both
between day and night. Generally, periods with higher flowrate mean Campaign 1 (October 2019 – March 2020) and Campaign 2 (July 2020)
higher concentration of pollutants for the activated sludge process in the with comparison to WHO guidelines, French reuse standard, and Euro­
WWTP potentially decreasing its treatment efficiency [31]. Besides, pean parliament regulations for water reuse [12,34,35]. The removal
higher flowrates lead to shortened sedimentation time resulting in rates between feed and permeate were calculated to evaluate the benefit
decreased particles’ removal rates [32]. Additionally, the weather can of UF. Moreover, as required by the standards, the removal rates be­
also affect the flowrate and the quality of the wastewater [33]. After all, tween WWTP feed and UF permeate were also calculated when possible.
the increase of raw wastewater flowrate will largely lead to deteriorated As a result, the UF system can greatly retain the microorganisms,
quality of WWTP effluent/UF feed, which is the main cause of UF including E.coli, Enterococci, spores of anaerobic sulphito-reducers, and
membrane fouling. Turbidity can be an indicator of feed water quality bacteriophages, all under the detection limitation of the analysis

Fig. 2. Variation of flow rate and turbidity as the function of time.

5
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

Fig. 3. Permeability and feed water turbidity variations.

Table 4
Mean UF permeate qualities compared to WHO guidelines for water reuse, French reuse standard, and European parliament regulation.
Parameters Campaign 1 Campaign 2 WHO French European
(October 2019 – March 2020) (July 2020) guidelines standard parliament
(A)* regulation (A) *
UF permeate UF permeate

Mean RRs* by RRs* by 6th 16th 28th RRs* by RRs* by

± SD WWTP + UF (%) July July July WWTP + UF (%)
UF (%) 2020 2020 2020 UF (%)

E. coli (CFU⋅100 mL− 1) <1 >6.7 (log*) >4.8 <56 <56 <56 >6.5 ± 1.4 >2.0 ± ≤ 200 ≤ 250 ≤ 10 / ≥ 5 (log*)
(log*) (log*) 0.5
(log*)
Enterococci (CFU⋅100 <1 >6.2 (log*) >4.4 <56 <56 <56 >5.4 ± 1.4 >1.1 ± – ≥ 4 (log*) –
mL− 1) (log*) (log*) 0.3
(log*)
Anaerobic sulphito- <1 >4.1 (log*) – <1 <1 <1 >3.5 ± 0.3 >2.8 ± – ≥ 4 (log*) ≥ 5 (log*)
reducers (spores) (log*) 0.1 /Absence
(CFU⋅100 mL− 1) (log*)
Specific F-RNA <1 >3.6 (log*) – <30 <1 <1 >2.6 ± 0.8 – – ≥ 4 (log*) ≥ 6 (log*)
bacteriophages (log*) /Absence
(PFP⋅100 mL− 1)
COD (mgO2⋅L− 1) <10 >98% >44% <30 <30 <30 >95.4 ± – – <60 –
2.0%
BOD5 (mgO2⋅L− 1) – – – <3 <3 <3 >98.1 ± – – – ≤ 10
0.3%
TSS (mg⋅L− 1) <2 >97% >56% – <2 <2 >98.5 ± >84 ± ≤ 30 <15 ≤ 10
0.8% 6%
− 1
TOC (mgC⋅L ) 4.61 ± 94 ± 3% 70 ± – – – – – – – –
0.61 11%
Turbidity (NTU) 0.42 ± – 85 ± 7% <0.1 <0.1 0.31 >99.9 ± >86 ± ≤2 – ≤5
0.35 0.0% 11%
pH 7.46 ± – – 7.6 7.4 7.5 – – – – –
0.07
− 1
Conductivity(μS⋅cm ) 1056 – – – – – – – – – –
± 70

CFU: colony-forming unit; PFP: Polyhedral, filamentous, and pleomorphic; log*: log removal calculated from the raw wastewater quality; - : not measured.
(A)*: There are ABCD four different levels of water quality in French water reuse standards and EU regulation. Level A being the best.
RRs*: Removal rates.

methods. Besides, the COD, BOD5, TSS concentrations and turbidity of absence. In addition of the parameters in Table 4, the parameters of
permeate all fulfill the three standards. However, because of the Salmonella, Legionella spp., and viable pathogenic helminth eggs, were
detection limit, the log removals of the spores of anaerobic sulphito- tested during Campaign 2 as mentioned in the EU regulation, and they
reducers and virus (Specific F-RNA bacteriophages) cannot fully reach were not found in any of the permeates. Therefore, regardless of the
to 4. In fact, the concentrations of the spores and virus in permeate were filtration conditions and the variation of feed water quality during
very low and under the detection limits which can be considered campaign 1, UF as tertiary treatment can provide high-quality water for

6
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

reuse as it reaches the WHO guidelines for reuse and the French reuse
standards with quality “A”. According to the results of campaign 2, UF
permeate can also reach the EU regulation for reuse with quality “A”.
Because of the international Covid-19 pandemic, coronavirus SARS-
CoV-2 was measured in the raw wastewater, UF feed, and UF permeate
in November 2020. In results, the raw wastewater was tested positive
with the envelope protein gene E above the detection limit (17 000
genome units per liter) but below the quantification limit (170 000
genome units per liter) and the nucleoprotein gene N was not detected.
The UF feed and UF permeate were tested negative to SARS-CoV-2.
While the limit of detection of the PCR method is quite high, it is un­
certain to say that the virus is completely absent in the feed and
permeate, or not. According to recent publications in this field, the
findings indicate that secondary wastewater treatment can largely
remove SARS-CoV-2 from wastewater, and tertiary treatment or disin­
fection can enhance the inactivation of SARS-CoV-2 in WWTPs [36,37]. Fig. 5. The retention rates of microplastics by UF membrane at different size
Bogler et al. [38] stated briefly in their review that SARS-CoV-2 is ex­ ranges of particles based on major dimension/minor dimension. The detection
pected to be removed reliably by UF. Based on the above analysis and limit value of dimensions is 6.25 μm for the permeate and 25 μm for the feed.
detection, the UF permeate after UF treatment in this test would have no
infectivity of SARS-CoV-2 and could be reused safely. Besides, in order to 4.1. Impact of filtration time and flux on hydraulic performances
fully assess UF as a barrier to SARS-CoV-2, lower detection limits in PCR
should be sought or new analytical tools should be developed to quantity Making the assumption that the feed water quality variation is
infectious SARS-CoV-2 for example. cyclical for one operating condition, permeate flux and filtration time
Recently, microplastic contamination have attracted worldwide are the key operating parameters [40]. According to Table 3, different
attention as a new emerging pollutant, especially because of its eco- filtration cycle time (t) and feed water flux (J) were studied to analyze
toxicological effects in marine environments and accumulation in their impact on membrane hydraulic performance. Filtration perfor­
bodies through food chain transmission [39]. Therefore, it is necessary mance was evaluated through backwash effectiveness, CEB frequency
to investigate the microplastics removal by UF in this study. Fig. 4(a) and fouling resistances. To evaluate the influence of flux and filtration
shows the types and quantity of polymers in feed water which includes cycle time, conditions with gradual increase of filtration cycle time (t) at
45.8% Polyethylene (PE), 45.8% Polypropylene (PP), 4.8% Poly­ same constant flux, and with gradual increase of filtration flux (J) at
ethylene terephthalate (PET), 2.4% Polystyrene (PS), and 1.2% Poly­ constant filtration cycle time (t) of 30 min (or 60 min) were designed
amide (PA). Fig. 4(b) shows the distribution of microplastics in feed and and operated.
permeate. Fig. 5 shows the capacity of UF on microplastics removal at The average reversibility of CB and AB obtained under different
different size of the microplastic particles (MPs)’ surface morphologies. conditions is shown in Fig. 6, which is indicative of the overall backwash
The result showed the excellent removal capacity of UF membrane on effectiveness. From the figure, reversibility of AB fluctuates from 110%
MPs with side dimension > 70 μm, which is > 95%. However, the MP to 180% in all conditions which was conspicuously higher than revers­
retention rates by UF will decrease as the particle size becomes smaller. ibility of CB (between 70% and 90%). It is therefore very interesting to
When MPs’ side dimension is ≤ 40 μm, the retention rate is <17%. In use AB in long-term fouling management.
macro vies, there are about 207.5 MPs⋅L− 1 in the feed and 17.5 MPs⋅L− 1 A classification of all conditions is proposed in Table 5 based on the
in the permeate, resulting in 91.6% removal of MPs. Therefore, the fouling reversibility, the total filtrated volume during a cycle (Vf-cyc), the
microplastics in feed water can be largely removed by UF membrane initial TMP of a cycle and the number of CEBs in 40 h operation. From
especially on particles with side dimension > 70 μm. Nevertheless, Fig. 6, J30t30BW1/3 and J60t20BW1/3 were performed in higher AB
despite the apparent larger size of the microplastic particles compared to removal capacity than the others with reversibility values of 150% and
the UF pore size (0.02 μm), the retention is not complete especially for 180%, respectively. These conditions with low flux or low filtration time
the smallest particles and this suggests a need of further research in this had the smallest ultrafiltrated volume and initial TMP as well as no CEB
field to understand why. needed during 40 h operation. Consequently, these conditions are
referred as soft conditions. Some other conditions, such as J80t60BW1/

Fig. 4. (a) The polymer types in feed water and their size distribution. (b) The quantity and size distribution of microplastics in feed and permeate. The size of the
data circles in both (a) and (b) represents the relative surface area of the microplastic particles.

7
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

Fig. 6. Fouling reversibility by CB and AB in different conditions.

constant filtration cycle time can directly increase TMP (0.167 bar) from
Table 5 the beginning of the filtration process. The increase of TMP may
Classification of the different operating conditions considering the total filtrated
compress the fouling layer (i.e. cake density) which will promote the
volume during a cycle, the initial TMP and the number of CEB in 40 h-operation.
smallest particles penetration into membrane pores thus enhancing
Conditions Vf- Initial TMP* CEB times in 40 h Types irreversible fouling [40]. In this case, although backwashes could flow
(bar) operation
cyc
away the cake foulant on membrane surface significantly, they could not
(L)
remove the smallest compounds adsorbed onto the membrane material
J30t30BW1/3 135 0.05 0 Soft conditions
effectively, thus decreasing CB and AB reversibility [20]. As the irre­
J60t20BW1/3 180 0.1 0
J60t30BW1/3 270 0.1 0 Standard versible fouling is mainly caused by increased TMP, the irreversible
J60t40BW1/3 360 0.1 0 conditions fouling was more sensitive to flux increase than filtration cycle time
J60t60BW1/3 540 0.1 0 extension in these tests: this is in agreement with [15].
J60t60BW1/4 540 0.1 0 In order to confirm the above results, the variations and composition
J80t30BW1/3 360 0.133 0
J80t40BW1/3 480 0.133 0
of fouling resistances before backwash versus time are proposed for each
J80t60BW1/3 720 0.133 6 Harsh condition (Fig. 7). J60t20BW1/3, J60t60BW1/3, and J100t30BW1/3
J80t60BW1/4 720 0.133 5 conditions are taken as representatives of the 3 types of conditions.
J100t30BW1/ 450 0.167 10
3
- In soft condition J60t20BW1/3, membrane resistance (Rm) was in
Initial TMP*: calculated when Lp is 600 L⋅m− 2⋅h− 1⋅bar− 1
at 20℃. dominant position during whole filtration and Rirr even not reached
to one-third of Rm.
3, J80t60BW1/4, and J100t30BW1/3, were performed in relatively - In standard condition J60t60BW1/3 with 40 min filtration time
weaker removal ability of CB and AB. These conditions with high flux or longer than J60t20BW1/3 in a cycle, the irreversible fouling Rirr
high filtration times need at least 5 CEBs during 40 h operation. Thus, gradually reached the value of Rm after 35 h’s operation and stayed
these conditions are referred as harsh conditions. Despite the soft and in the same order of magnitude than Rm.
harsh conditions, the rest is considered as standard conditions with CB - In harsh condition J100t30BW1/3 with highest flux among all tested
reversibility ranging between 80% and 90%, and AB reversibility conditions, the irreversible fouling Rirr increased fastest from
ranging between 120% and 145%. The harsh conditions of J80t60BW1/ beginning and exceeded the value of Rm. Higher values of reversible
3, J80t60BW1/4 could filtrate the highest volume (720 L) during one resistances than in soft or standard conditions are also to be noted.
cycle which also means the large quantity of fouling load would be
retained by membranes. The highly increased fouling resistance thereby Through comparison, increase of flux and filtration cycle time both
reduced the fouling reversibility of CB and AB. This phenomenon can be contribute to the increasing rate of Rirr during filtration process. The
explained by the higher filtration volume over a filtration cycle (Vs-cyc) increase of filtration flux (comparison between J60t60 and J100t30)
as a combined result of flux and time, which means more solids, particles seems to have a higher impact on Rirr increasing rate than the increase of
and organic matters retained by membrane thus increasing both irre­ filtration time (comparison between J60t20 and J60t60).
versible and reversible fouling. For J100t30BW1/3 condition, even if the From above, the harsh conditions all resulted in occurrence of CEB
total filtrated volume in a cycle Vf-cyc (450 L) was lower than in standard during 40 h’s operation, especially in condition J100t30BW1/3 with the
conditions (such as J60t60BW1/3 and J80t40BW1/3), the reversibility most frequent CEB needed. In practical uses, the CEB interval usually
of both CB and AB were still lower, and with more frequent CEBs. Ac­ ranges from hours to weeks among different UF processes [42]. Because
cording to Darcy’s law, the increase of flux (100 L⋅m− 2⋅h− 1) under one CEB consumes more chemical agents, permeate water, and energy

8
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

Fig. 8. Water recovery rates: conditions at backwash frequency of 1/3. *The

uncertainty of each column is between 1 and 2%.

frequent AB and CB, and low flux. The water recovery rates of harsh
conditions J100t30BW1/3 and J80t60BW1/3, were at an intermediate
level, ranging from 84.7% to 89.9%. Except the conditions in backwash
frequency of 1/3, the conditions J60t60BW1/4 and J80t60BW1/4,
resulted in water recovery rates of 92.8% and 90.8% respectively. The
condition J60t60BW1/4 showed almost the same water recovery rate
with J60t60BW1/3, with only 0.1% difference because of different
permeate consumption between AB and CB. Therefore, the higher water
recovery rates among these conditions were obtained from the condi­
tions of J80t40BW1/3, J60t60BW1/4, and J60t60BW1/3 with potential
to be the optimized operating conditions in this study.

4.2. Impact of air backwash frequency

To evaluate the impact of AB frequency on filtration performance

and confirm that the frequencies 1/3 and 1/4 previously studied were
the best, 5 different backwash frequencies (1/3, 1/4, 1/6, 1/9, and No
AB) (Table 3) were tested, all under the constant flux and filtration cycle
time (J = 60 L⋅m− 2⋅h− 1 and t = 60 min). The variation of (Rirr/Rm)n/
(Rirr/Rm)ini versus turbidity integration based on Eq. (10) reflecting the
Rirr increasing rate was calculated and is shown in Fig. 9. The turbidity
integration enables to consider feed water quality variation over time
considering turbidity as a main indicator of feed quality. It should be
noted that the filtration process between 2 CEBs was considered as a new
and independent filtration phase to be calculated. Therefore, turbidity
integration will be re-calculated from 0, and corresponding to the re-
calculated (Rirr/Rm)ini. This explains why some conditions present
several phases.
In Fig. 9(a) and (b), two main variation can be described. On one
hand, Fig. 9(a) with AB frequency of 1/9 and no AB presents very high
Fig. 7. Resistances’ variations versus time at conditions: (a) J60t20BW1/3, (b) (Rirr/Rm)n/(Rirr/Rm)initial increase with turbidity integration and 2 CEBs
J60t60BW1/3 and (c) J100t30BW1/3. occurring during 2 days of filtration. On the other hand, Fig. 9(b) with
AB frequencies of 1/3, 1/4 and 1/6 where (Rirr/Rm)n/(Rirr/Rm)initial
compared with an AB or a CB, it is necessary to prolong the CEB interval slowly increased with turbidity integration. AB frequency in conditions
during filtration. It has been previously estimated that sustainable of BW1/3, BW1/4 are the best with excellent Rirr removal abilities,
operating conditions lead to less than 2 CEB occurences per day [41]. which can maintain membrane permeability over 200 L⋅m− 2⋅h− 1⋅bar− 1
Therefore, the harsh conditions cannot be considered as the optimized in 2 days of filtration. No CEB occurred in these conditions. Therefore,
conditions for long term operation. optimum conditions BW1/3 and BW1/4 used in 3.3 are confirmed as
Furthermore, the permeate water was produced for reuse, but it was they are more suitable for long term operation at industrial scale. This
also used for cleanings. The permeate consumption in CB, AB, and CEB result is similar to the result on seawater [17], who concluded that AB
conversely decreased the productivity of the system. Therefore, it is was more effective to control fouling resistance in conditions of BW1/3
necessary to investigate the water recovery rate in different conditions and BW1/5, than that in conditions with lower AB frequency BW1/7 and
to make a comprehensive comparison, shown in Fig. 8 with conditions of BW1/9, all under the same J = 60 L⋅m− 2⋅h− 1 and t = 60 min.
AB frequency of 1/3. Through calculation, the highest water recovery
rates occurred in the standard conditions, especially in conditions of 4.3. Hydraulic filtration performance in long-term operation in optimized
J60t60BW1/3 and J80t40BW1/3 which resulted in 92.7% and 94.5% conditions
water recovery, respectively. In addition, the soft conditions showed the
lowest water recovery rates between 70.7% and 72.2% because of From the above sections, the three best operating conditions on short

9
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

term experiments are J80t40BW1/3, J60t60BW1/4, and J60t60BW1/3.

To verify the feasibility and sustainability of the optimized conditions,
long term tests were conducted on condition J60t60BW1/3 both in
winter (from 26/2/2020 to 18/3/2020) and summer (from 1/7/2020 to
27/7/2020). Fig. 10 and Fig. 11 show the permeability variation for
more than 20 days, respectively, in winter and in summer.
From Fig. 10 and Fig. 11, the UF pilot plant showed stable and
continuous filtration performance, and great permeability recoveries
(by CB, AB, and CEBs) during long-term operation, both in summer and
winter. The occurrence of CEB were all less than once per day which
meets sustainable objectives. However, the frequency of CEB is unstable,
which is found to be related to the feed water quality and temperature
changes. In summer, the permeability drops seem faster than in winter,
but it also seems that permeability recoveries by CEB are higher than in
winter. Faster permeability drops in summer can be related to more
algae observed in the effluent of WWTP resulting in more suspended
solids and organic matters. Therefore, the fouling potential on mem­
brane in summer was relatively higher than in winter but at the same
time the mean water temperature in summer was 27.7 ± 1.3 ℃, while in
winter it was 20.2 ± 3 ℃ which could lead to better reversibility of
cleanings including AB, CB, and CEB [43–45]. Indeed, the mean
reversibility of AB in summer and in winter are respectively 144 ± 46%
and 135 ± 21%. Besides, the mean permeability recoveries by CEB in
summer and in winter are respectively 308 ± 15% and 246 ± 30%.
However, the water temperature increase had little effects on CB per­
formance: the mean reversibility of CB in summer and winter are all
around 80%. Although there was higher fouling potential on membrane
in summer, the corresponding temperature increase improved cleaning
efficiency of AB and CEB, finally resulting in sustainable permeability
recovery and reasonable CEB frequency.

5. Conclusion

In this study, ultrafiltration is considered as tertiary treatment for

Fig. 9. (a) Variation of irreversible resistance as a function of the integral of municipal wastewater reuse in semi-industrial operation and for the first
turbidity vs. time for different BWs of1/9, and No AB; (b) Variation of irre­ time some conclusions can be given:
versible resistance as a function of the integral of turbidity vs. time for different
BWs of 1/3, 1/4, 1/6; all at flux = 60 L⋅m− 2⋅h− 1 and t = 60 min.
(i) The UF permeate quality was detected to be good enough to be
reused in non-potable applications as it met reuse guidelines of

Fig. 10. Permeability variation in winter from 26/2/2020 to 18/3/2020.

10
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

Fig. 11. Permeability variation in Summer from 1/7/2020 to 27/7/2020.

the WHO, reuse standards of France and the most recent EU Declaration of Competing Interest
regulation for agricultural irrigation. In addition, the UF mem­
brane was confirmed to be effective on most microplastic parti­ The authors declare that they have no known competing financial
cles’ removal. interests or personal relationships that could have appeared to influence
(ii) Concerning the most suitable operating conditions for UF: soft the work reported in this paper.
conditions with lower flux or frequent physical cleaning (short
filtration time) resulted in too little productivity and were dis­ Acknowledgement
carded. Harsh conditions with higher flux or longer filtration time
were also discarded because of high occurrence of CEB which led The authors would like to acknowledge Perkin Elmer for microplastic
to difficult permeability stabilization and high consumption of particles analyses and the China Scholarship Council for partial financial
permeate and chemicals. As the permeate quality was not support (201801810046).
affected by filtration conditions, the performance comparison
was based on AB and CB reversibility, irreversible fouling resis­
References
tance increase, occurrence of CEBs and water recovery rates.
Finally, standard conditions J80t40BW1/3, J60t60BW1/4, and [1] I.F.G. Tejero, V.H.D. Zuazo, Water Scarcity and Sustainable Agriculture in Semiarid
J60t60BW1/3 stood out from the others with higher overall Environment: Tools, Strategies, and Challenges for Woody Crops, Academic Press,
performances. Of course, these conditions are function of the 2018.
[2] S. Mudgal, L. Van Long, N. Saïdi, R. Haines, D. McNeil, P. Jeffrey, H. Smith, J.
WWTP and feed water quality, but still, this study defined some Knox, European Commission, Directorate-General for the Environment, BIO by
ranges of operating conditions for future water managers and Deloitte, ICF International, Cranfield University, Optimising water reuse in the EU:
engineers for municipal wastewater tertiary treatment before final report, Publications Office, Luxembourg, 2015. http://bookshop.europa.eu/
uri?target=EUB:NOTICE:KH0115207:EN:HTML (accessed May 22, 2019).
reuse. [3] S. Carraud, France wants to reduce water use as rivers dry up, Reuters., 2019.
(iii) By analyzing the influence of AB frequency on irreversible fouling https://uk.reuters.com/article/uk-france-water-idUKKCN1TW3DZ (accessed July
management, the irreversible fouling resistance increased faster 20, 2020).
[4] D.M. Warsinger, S. Chakraborty, E.W. Tow, M.H. Plumlee, C. Bellona,
with the decrease of AB frequency during operation. Optimum AB
S. Loutatidou, L. Karimi, A.M. Mikelonis, A. Achilli, A. Ghassemi, L.P. Padhye, S.
frequencies of BW1/3 and BW1/4 under constant flux of 60 A. Snyder, S. Curcio, C.D. Vecitis, H.A. Arafat, J.H. Lienhard, A review of polymeric
L⋅m− 2⋅h− 1 and t of 60 min found in (ii) were confirmed. membranes and processes for potable water reuse, Prog. Polym. Sci. 81 (2018)
209–237, https://doi.org/10.1016/j.progpolymsci.2018.01.004.
(iv) Concerning the selected sustainable conditions, more than 20
[5] J. Yang, M. Monnot, L. Ercolei, P. Moulin, Membrane-Based Processes Used in
days of operation in condition J60t60BW1/3 in winter and in Municipal Wastewater Treatment for Water Reuse: State-of-the-Art and
summer confirmed that the UF pilot plant could provide sus­ Performance Analysis, Membranes 10 (2020) 131, https://doi.org/10.3390/
tainable and adaptable filtration performance regardless of the membranes10060131.
[6] L. Shi, J. Huang, L. Zhu, Y. Shi, K. Yi, X. Li, Role of concentration polarization in
temperature and feed water quality. This works confirms the cross flow micellar enhanced ultrafiltration of cadmium with low surfactant
great interest of UF as tertiary treatment of conventional WWTP concentration, Chemosphere 237 (2019), 124859, https://doi.org/10.1016/j.
for water reuse. chemosphere.2019.124859.
[7] L. Shi, Y. Lei, J. Huang, Y. Shi, K. Yi, H. Zhou, Ultrafiltration of oil-in-water
emulsions using ceramic membrane: Roles played by stabilized surfactants,
CRediT authorship contribution statement Colloids Surf. Physicochem. Eng. Asp. 583 (2019), 123948, https://doi.org/
10.1016/j.colsurfa.2019.123948.
[8] F. Qu, H. Wang, J. He, G. Fan, Z. Pan, J. Tian, H. Rong, G. Li, H. Yu, Tertiary
J. Yang: Investigation. M. Monnot: Supervision. T. Eljaddi: Inves­ treatment of secondary effluent using ultrafiltration for wastewater reuse:
tigation. L. Ercolei: Resources. L. Simonian: Investigation. P. Moulin: correlating membrane fouling with rejection of effluent organic matter and
Supervision. hydrophobic pharmaceuticals, Environ. Sci. Water Res. Technol. 5 (2019)
672–683, https://doi.org/10.1039/C9EW00022D.

11
J. Yang et al. Separation and Purification Technology 272 (2021) 118921

[9] S. Muthukumaran, D.A. Nguyen, K. Baskaran, Performance evaluation of different [28] J. Howell, R. Field, D. Wu, Ultrafiltration of high-viscosity solutions: Theoretical
ultrafiltration membranes for the reclamation and reuse of secondary effluent, developments and experimental findings, Chem. Eng. Sci. 51 (1996) 1405–1415,
Desalination 279 (2011) 383–389, https://doi.org/10.1016/j.desal.2011.06.040. https://doi.org/10.1016/0009-2509(95)00315-0.
[10] A. Pollice, A. Lopez, G. Laera, P. Rubino, A. Lonigro, Tertiary filtered municipal [29] A. Resosudarmo, Y. Ye, P. Le-Clech, V. Chen, Analysis of UF membrane fouling
wastewater as alternative water source in agriculture: a field investigation in mechanisms caused by organic interactions in seawater, Water Res. 47 (2013)
Southern Italy, Sci. Total Environ. 324 (2004) 201–210, https://doi.org/10.1016/j. 911–921, https://doi.org/10.1016/j.watres.2012.11.024.
scitotenv.2003.10.018. [30] H. Chang, H. Liang, F. Qu, S. Shao, H. Yu, B. Liu, W. Gao, G. Li, Role of backwash
[11] D. Falsanisi, L. Liberti, M. Notarnicola, Ultrafiltration (UF) Pilot Plant for water composition in alleviating ultrafiltration membrane fouling by sodium
Municipal Wastewater Reuse in Agriculture: Impact of the Operation Mode on alginate and the effectiveness of salt backwashing, J. Membr. Sci. 499 (2016)
Process Performance, Water 2 (2010) 872–885, https://doi.org/10.3390/ 429–441, https://doi.org/10.1016/j.memsci.2015.10.062.
w2040872. [31] A. Aygun, B. Nas, A. Berktay, Influence of High Organic Loading Rates on COD
[12] European Parliament, Regulation (EU) 2020/741 of the European Parliament and Removal and Sludge Production in Moving Bed Biofilm Reactor, Environ. Eng. Sci.
of the Council of 25 May 2020 on minimum requirements for water reuse (Text 25 (2008) 1311–1316, https://doi.org/10.1089/ees.2007.0071.
with EEA relevance), 2020. http://data.europa.eu/eli/reg/2020/741/oj/eng [32] P. Rodríguez López, A.G. Lavín, M.M. Mahamud López, J.L. Bueno de las Heras,
(accessed January 29, 2021). Flow models for rectangular sedimentation tanks, Chem. Eng. Process. Process
[13] X. Zheng, M.T. Khan, J.-P. Croué, Contribution of effluent organic matter (EfOM) Intensif. 47 (2008) 1705–1716. DOI: 10.1016/j.cep.2007.09.020.
to ultrafiltration (UF) membrane fouling: Isolation, characterization, and fouling [33] S.S. Rashid, Y.-Q. Liu, Assessing environmental impacts of large centralized
effect of EfOM fractions, Water Res. 65 (2014) 414–424, https://doi.org/10.1016/ wastewater treatment plants with combined or separate sewer systems in dry/wet
j.watres.2014.07.039. seasons by using LCA, Environ. Sci. Pollut. Res. (2020), https://doi.org/10.1007/
[14] S.F.E. Boerlage, M. Kennedy, Z. Tarawneh, R. De Faber, J.C. Schippers, s11356-020-08038-2.
Development of the MFI-UF in constant flux filtration, Desalination 161 (2004) [34] Ministère de la Santé et des Sports, La Ministre de l’énergie et de la mer, la Ministre
103–113, https://doi.org/10.1016/S0011-9164(04)90046-X. des affaires sociales et de la santé et le Ministre de l’Agriculture, de
[15] E. Akhondi, F. Wicaksana, A.G. Fane, Evaluation of fouling deposition, fouling l’agroalimentaire et de la forêt, Arrêté du 2 août 2010 relatif à l’utilisation d’eaux
reversibility and energy consumption of submerged hollow fiber membrane issues du traitement d’épuration des eaux résiduaires urbaines pour l’irrigation de
systems with periodic backwash, J. Membr. Sci. 452 (2014) 319–331, https://doi. cultures ou d’espaces verts - Légifrance (JORF n◦ 0201 du 31 août 2010), 2010.
org/10.1016/j.memsci.2013.10.031. https://www.legifrance.gouv.fr/loda/id/JORFTEXT000022753522/2020-10-11/
[16] K. Katsoufidou, S.G. Yiantsios, A.J. Karabelas, An experimental study of UF (accessed October 14, 2020).
membrane fouling by humic acid and sodium alginate solutions: the effect of [35] World Health Organization, Guidelines for the safe use of wastewater, excreta and
backwashing on flux recovery, Desalination 220 (2008) 214–227, https://doi.org/ greywater-Volume 4, World Health Organiztion, 2006.
10.1016/j.desal.2007.02.038. [36] S.G. Rimoldi, F. Stefani, A. Gigantiello, S. Polesello, F. Comandatore, D. Mileto,
[17] C. Cordier, T. Eljaddi, N. Ibouroihim, C. Stavrakakis, P. Sauvade, F. Coelho, M. Maresca, C. Longobardi, A. Mancon, F. Romeri, C. Pagani, F. Cappelli,
P. Moulin, Optimization of Air Backwash Frequency during the Ultrafiltration of C. Roscioli, L. Moja, M.R. Gismondo, F. Salerno, Presence and infectivity of SARS-
Seawater, Membranes 10 (2020) 78, https://doi.org/10.3390/ CoV-2 virus in wastewaters and rivers, Sci. Total Environ. 744 (2020), 140911,
membranes10040078. https://doi.org/10.1016/j.scitotenv.2020.140911.
[18] X. Shi, G. Tal, N.P. Hankins, V. Gitis, Fouling and cleaning of ultrafiltration [37] S.P. Sherchan, S. Shahin, L.M. Ward, S. Tandukar, T.G. Aw, B. Schmitz, W. Ahmed,
membranes: A review, J. Water Process Eng. 1 (2014) 121–138, https://doi.org/ M. Kitajima, First detection of SARS-CoV-2 RNA in wastewater in North America: A
10.1016/j.jwpe.2014.04.003. study in Louisiana, USA, Sci. Total Environ. 743 (2020), 140621, https://doi.org/
[19] H. Chang, H. Liang, F. Qu, B. Liu, H. Yu, X. Du, G. Li, S.A. Snyder, Hydraulic 10.1016/j.scitotenv.2020.140621.
backwashing for low-pressure membranes in drinking water treatment: A review, [38] A. Bogler, A. Packman, A. Furman, A. Gross, A. Kushmaro, A. Ronen, C. Dagot,
J. Membr. Sci. 540 (2017) 362–380, https://doi.org/10.1016/j. C. Hill, D. Vaizel-Ohayon, E. Morgenroth, E. Bertuzzo, G. Wells, H.R. Kiperwas,
memsci.2017.06.077. H. Horn, I. Negev, I. Zucker, I. Bar-Or, J. Moran-Gilad, J.L. Balcazar, K. Bibby,
[20] Y. Ye, V. Chen, P. Le-Clech, Evolution of fouling deposition and removal on hollow M. Elimelech, N. Weisbrod, O. Nir, O. Sued, O. Gillor, P.J. Alvarez, S. Crameri,
fibre membrane during filtration with periodical backwash, Desalination 283 S. Arnon, S. Walker, S. Yaron, T.H. Nguyen, Y. Berchenko, Y. Hu, Z. Ronen, E. Bar-
(2011) 198–205, https://doi.org/10.1016/j.desal.2011.03.087. Zeev, Rethinking wastewater risks and monitoring in light of the COVID-19
[21] L. Li, H.E. Wray, R.C. Andrews, P.R. Bérubé, Ultrafiltration Fouling: Impact of pandemic, Nat. Sustain. (2020) 1–10, https://doi.org/10.1038/s41893-020-
Backwash Frequency and Air Sparging, Sep. Sci. Technol. 49 (2014) 2814–2823, 00605-2.
https://doi.org/10.1080/01496395.2014.948964. [39] B. Ma, W. Xue, C. Hu, H. Liu, J. Qu, L. Li, Characteristics of microplastic removal
[22] P.J. Remize, C. Guigui, C. Cabassud, Evaluation of backwash efficiency, definition via coagulation and ultrafiltration during drinking water treatment, Chem. Eng. J.
of remaining fouling and characterisation of its contribution in irreversible fouling: 359 (2019) 159–167, https://doi.org/10.1016/j.cej.2018.11.155.
Case of drinking water production by air-assisted ultra-filtration, J. Membr. Sci. [40] M. Raffin, E. Germain, S.J. Judd, Influence of backwashing, flux and temperature
355 (2010) 104–111, https://doi.org/10.1016/j.memsci.2010.03.005. on microfiltration for wastewater reuse, Sep. Purif. Technol. 96 (2012) 147–153,
[23] C. Cordier, C. Stavrakakis, P. Sauvade, F. Coelho, P. Moulin, Air Backwash https://doi.org/10.1016/j.seppur.2012.05.030.
Efficiency on Organic Fouling of UF Membranes Applied to Shellfish Hatchery [41] R.W. Field, G.K. Pearce, Critical, sustainable and threshold fluxes for membrane
Effluents, Membranes 8 (2018) 48, https://doi.org/10.3390/membranes8030048. filtration with water industry applications, Adv. Colloid Interf. Sci. 164 (2011)
[24] C.M. Chew, M.K. Aroua, M.A. Hussain, Advanced process control for ultrafiltration 38–44, https://doi.org/10.1016/j.cis.2010.12.008.
membrane water treatment system, J. Clean. Prod. 179 (2018) 63–80, https://doi. [42] G. Gilabert-Oriol, Optimization of ultrafiltration membrane cleaning processes.
org/10.1016/j.jclepro.2018.01.075. Pretreatment for reverse osmosis in seawater desalination plants, 2013.
[25] C. Cordier, C. Stavrakakis, B. Dupuy, M. Papin, P. Sauvade, F. Coelho, P. Moulin, [43] M.C. Almecija, A. Martinez-Ferez, A. Guadix, M.P. Paez, E.M. Guadix, Influence of
Ultrafiltration for environment safety in shellfish production: Removal of oyster the cleaning temperature on the permeability of ceramic membranes, Desalination
gametes in hatchery effluents, Aquac. Eng. 84 (2019) 80–90, https://doi.org/ 245 (2009) 708–713, https://doi.org/10.1016/j.desal.2009.02.041.
10.1016/j.aquaeng.2018.12.008. [44] M.T. Alresheedi, O.D. Basu, Effects of feed water temperature on irreversible
[26] W. Gao, H. Liang, J. Ma, M. Han, Z. Chen, Z. Han, G. Li, Membrane fouling control fouling of ceramic ultrafiltration membranes, J. Water Process Eng. 31 (2019),
in ultrafiltration technology for drinking water production: A review, Desalination 100883, https://doi.org/10.1016/j.jwpe.2019.100883.
272 (2011) 1–8, https://doi.org/10.1016/j.desal.2011.01.051. [45] L. Lintzos, K. Chatzikonstantinou, N. Tzamtzis, S. Malamis, Influence of the
[27] M.R. Wiesner, P. Aptel, Mass transport and permeate flux and fouling in pressure- Backwash Cleaning Water Temperature on the Membrane Performance in a Pilot
driven process. Water Treatment Membrane Handbook, McGraw-Hill, New York, SMBR Unit, Water 10 (2018) 238, https://doi.org/10.3390/w10030238.
1996.

12