Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202

Contents lists available at ScienceDirect

Process Safety and Environmental Protection

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

Use of membrane technology for oil field and

refinery produced water treatment—A review

Selvaraj Munirasu a , Mohammad Abu Haija b , Fawzi Banat a,∗

a Department of Chemical Engineering, The Petroleum Institute, PO Box 2533, Abu Dhabi, United Arab Emirates
b Department of Applied Chemistry, The Petroleum Institute, PO Box 2533, Abu Dhabi, United Arab Emirates

a r t i c l e i n f o a b s t r a c t

Article history: With the advent of modern drilling technology namely sand-tar, hydraulic fracturing and
Received 12 July 2015 enhanced oil recovery, the amount of waste water to be treated before reuse and/or dis-
Received in revised form 16 charge to the environment has increased manifold in recent time. The treatment of produced
December 2015 water and refinery waste water from the oil industry has been traditionally done by phys-
Accepted 14 January 2016 ical as well as chemical processes. The use of membrane technology for the produced and
Available online 25 January 2016 refinery waste water treatment has been recent phenomenon and active research has been
focused to enhance the efficiency and life time of the membrane during the operation of
Keywords: the waste water treatment. In this review we briefly focus on the produced and refinery
Produced water waste water treatment by primary and secondary treatment in historical perspective fol-
Oil refinery waste water lowed by focusing on various membrane technologies starting from microfiltration (MF),
Membrane technology ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO). Finally we also focus on the
Water reuse membrane distillation (MD) in combination with forward osmosis (FO) as potential future
Responsive surface technology.
Oil–water separation © 2016 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
1.1. Constituents of PW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
1.2. Challenges and needs for the PW treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .187
2. Primary and secondary treatment of PW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
3. Produced water treatment by membrane technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
3.1. Microfiltration (MF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
3.2. Ultrafiltration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
3.3. NF and RO membranes for PW treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .194
3.4. Thermal desalination and forward osmosis (FO) for the PW treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
4. Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

1. Introduction sources, the fossil fuel is the most convenient source of energy
for more than century and even today it is the primary energy
It is no exaggeration to state that the modern life style depends source for humankind. Among the fossil fuels, except coal,
on the reliable energy supply. Among the various energy all other fuels namely oil, gas, coal bed methane (CBM), and

∗
Corresponding author. Tel.: +971 26075200.
E-mail address: fbanat@pi.ac.ae (F. Banat).
http://dx.doi.org/10.1016/j.psep.2016.01.010
0957-5820/© 2016 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
184 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202

Fig. 1 – (a). General schematic representation of oil & gas well with the formation (produced) water in the reservoir and (b).
The refinery process which produces process waste water (The figure is for representation only and not mentioned in the
true scale bar).

recent advent of hydraulic fracturing of shale oil and gas are
produced by deep drilling of the earth to recover the hydrocar-
bon energy sources. In the drilling process, water has always
been produced as by-product which is generally known as pro-
duced water (PW) or co-produce water (co-PW) or flowback
water. The term co-produce water or flowback water refers the
fresh water comes out with formation water from the well. The
fresh surface water was generally used for the pressurization
and thereby pumping out the crude oil and/or gas from the
well. The volume of PW is particularly vast quantity for the
oil industry whereas the gas drilling produces comparatively
less quantity of water. The amount of PW varies from well to
well as well as the age of reservoir. The volume of PW typi-
cally increases with the age of reservoir and in certain cases
it can reach up to 98% of the total fluid volume (Igunnu and
Chen, 2012; Alzahrani and Mohammad, 2014). However, it is
generally recognized as 1:3 ratio for oil & water for most of the
oil well. The typical source of produced water is depicted in
Fig. 1a (Igunnu and Chen, 2012). The produced water is part of
natural formation water existing beneath or within the oil/gas
reservoir. Thus the constituents of PW are based on natu-
ral geological formation and it is often acidic in nature with
various soluble mineral ions depending on the geology of the
particular the reservoir.
Invariably all the recovered crude oil and gas by drilling the
reservoir are further subjected to “refinery processing” before
the final use. Thus the refinery processes for oil and sweeten-
ing of the gas also produce large amount of waste water which
is generally called as “process water” (Fig. 1b). Particularly the
oil refinery industry uses large volumes of fresh water for the
processing. The major contributing processes in the oil refin-
ery are desalter effluent, sour water, tank bottom draw and
spent caustic (Petroleum, 2010). The major difference between
produced water and process water is that the produced water
contains the majority of the dissolved mineral ions whereas
the refinery process water generally contains comparatively
less dissolved inorganic (mineral) ions. However, the refin-
ery process creates many breakdown chemical compounds
during the chemical transformation which are generally less
or nonexistence in the produced water. The major notable
constituent increment in the process water due to the refin-
ery processing is phenols, ammonia, H2 S and BTEX (benzene,
toluene, ethylbenzene and xylenes). The chemical structure of
these compounds are shown in Fig. 2. Thus, the produced and
process water differs considerably in terms of concentration
of the pollutant present. In general the produced water con-
tains high salt content whereas the refinery process water
contains high organic matters (see Tables 1 and 2) (Igunnu and
Chen, 2012; Fakhru’l-Razi et al., 2009; Olsson et al., 2013). The
pre-treatment for both produced and process water using MF
and/or UF membrane can be considered comparable because
of the removal of macro constituents by these membrane
technologies. Since the quality and quantity of the produced
water (and oil and gas also) varies from well to well therefore
accordingly the refinery processing, we discuss the treatment
of produced and processing water as single progression in this
review. In many literature, the model refinery process water is
prepared using crude oil and therefore it is considerably simi-
lar in nature compared to the produced water (Al-Malack and
Siddiqui, 2013; Chakrabarty et al., 2008). Additionally there are
studies that deals with both produced and process water using
similar membrane treatment protocol for the water purifica-
tion (Ayse, 2009). Therefore hereafter in this review PW refers
both produced and process (refinery) water (Table 3).
In general, the treatment of PW can be simplified in the
following three classes based on the macro and molecular
level separation. (1) Removal of organics which include dis-
persed, dissolved and emulsified oil, grease and gases. (2) The
removal of dissolved inorganic matters which is commonly
referenced as TDS. If naturally occurring radioactive mate-
rials (NORM) is present in PW, then special care should be
Fig. 2 – The chemical structure of few selective compounds
present in the refinery waste water.
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202 185

and secondary treatment and in some cases biological treat-

Table 1 – Commonly existing constituent of pollutants
from the produced water (adapted from (Fakhru’l-Razi ment followed by tertiary treatment. The use of membrane
et al., 2009; Olsson et al., 2013) (* JHM, 2009, 170, 530, # technology as tertiary treatment for the PW is fast evolving
EES, 2013, 70, 3895). application in the oil and gas industry. There are literature
Constituents Oil field PW* Gas field PW* Damme 3 reviews available concerning the produced water or refinery
well# waste water treatment as standalone treatment (Alzahrani
and Mohammad, 2014; Fakhru’l-Razi et al., 2009; Diya’uddeen
TOC 0–1500 67–38,000 NA
et al., 2011; Coday et al., 2014; Padaki et al., 2015; Shaffer
COD 1200 2600–120,000 NA
TSS 1–1000 8–5484 NA
et al., 2013; Szep and Kohlheb, 2010; dos Santos et al., 2014;
pH 4–10 3.1–7 NA Shams Ashaghi and Czermak, 2007; Subramani and Jacangelo,
Total oil 2–565 NA NA 2015), but to the best of our knowledge there is no review arti-
Volatiles 0.4–35 NA NA cle universally deals with both produced and process water
(BTEX) in the oil industry. In this review article, after emphasising
Chloride 80–200,000 1400–190,000 78200
the primary and secondary treatment, we focus on both pro-
Bicarbonate 77–3990 NA NA
duced and process water treatment with main focus on the
Sulfate 2–1650 <0.1–47 8
Ammonia 10–300 NA NA membrane technology as final treatment process. We also
nitrogen discuss the development of FO-MD as potential technology.
Sulfite ∼10 NA NA Additionally, to the best of our knowledge, there are no review
Phenols Up to 23 NA NA articles discussing the futuristic aspect of surface wettability
Volatile 2–4900 NA NA and its potential application and this point also included in
fatty acids
this review article for the first time. Thus this review covers in
Calcium 13–25,800 Up to 51,300 14100
holistic manner, for the first time, all the existing and potential
Sodium 132–97,000 520–120,000 30600
Potassium 24–4300 149–3870 110 future membrane technologies for the treatment of PW.
Magnesium 8–6000 0.9–4300 1800
Iron <0.1–100 Up to 1100 91
Aluminum 310–410 <0.5–83 NA
1.1. Constituents of PW
Boron 5–95 Up to 56 NA
Barium 1.3–650 <1–1740 455 The most commonly existing pollutant constituents from the
Cadmium <0.005–0.2 <0.02–1.21 NA produced and process water are given in Tables 1 and 2 (Igunnu
Chromium 0.02–1.1 Up to 0.03 0.3 and Chen, 2012; Fakhru’l-Razi et al., 2009; Coelho et al., 2006;
Copper <0.002–105 Up to 5 NA
Al Zarooni and Elshorbagy, 2006). It can be largely divided into
Lithium 3–50 18.6–235 5
two categories namely organic and inorganic substances and
Manganese <0.004–175 Up to 63 2.5
Lead 0.002–8.8 <0.2–10.2 1 varying widely from well to well. Therefore it influences the
Strontium 0.02–1000 Up to 6200 1455 treatment process significantly. The inorganic compounds are
Zinc 0.01–35 <0.02–5 0.4 specific in nature and majority of the constituents exist as dis-
Arsenic <0.3 Up to 151 NA solved anions and cations. The common cations in the PW
Mercury <0.002 NA NA are: Na, Ca, K, Mg, Fe, Al, B, Ba, Cu, Li, Zn, Ti, Mn, and also
NA—not available.
toxic Cd, Cr, Pb, Hg, As, Sr, Be. Depending on the geological
condition, some wells contain NORM, like 226 Ra, 228 Ra, 238 U
and 235 U in the PW. The major anions in the PW are: chlo-
needed for the removal and disposal of NORM (Zhang et al., ride, sulphate, sulphite, (bi)carbonate, nitrate, nitrite, etc. The
2014a). (3) Removal of particulates which include suspended organic components are very complex in nature starting from
particle, clay and sand. Depending on the oil and gas pro- oil, grease, and other organic compounds produced by Bac-
duction well, the production chemicals and chemicals used tria. The organics can vary from aliphatic to aromatic and
in the refinery process also should be removed and it can be changing from polar to non-polar characteristic with many
included in the above three categories based on its original possible functional groups like alcoholic OH, amino NH2 ,
constituent. The treatment of PW typically involves primary carboxylic COOH, phenolic OH etc. Additionally the organic

Table 2 – Commonly existing pollutant constituents from refinery waste water. Data adapted from * Petroleum (2010), ∼
Coelho et al. (2006), a Al Zarooni and Elshorbagy (2006).
Contaminant Desalter* Sour water* Tank bottom* Cooling tower* Sour water Brazil∼ Al Ruwaisa

COD 400–1000 600–1200 400–1000 150 850–1020 825

Free hydrocarbons Up to 1000 <10 Up to 1000 <5 NA 750@
Suspended solids Up to 500 <10 Up to 500 Up to 200 ND NA
Phenols 10 to 100 Up to 200 NA NA 98–128 25
Benzene 5 to 15 0 NA NA NA 85#
Sulphides Up to 100 <10 Up to 100 NA 15–23 3900ˆ
Ammonia Up to 100 <100 NA NA 5.1–21.1 75&
Dissolved solid NA NA NA Up to 700 NA NA

NA—not available, ND—non-detectable.

&
As Kjeldahl N.
#
As PAH.
@
As TPH.
ˆ
As sulfate concentration.
186 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202

Table 3 – Selected summary of membrane treatment of produced and refinery waste water from the literature.
S. no Membrane Membrane material and Water source Parameters studied Reference
category properties

1 MF Ceramic and Synthetic PW Oil–water separation Kose et al. (2012)

polyacrylonitrile
2 0.1 ␮m commercial mixed Real Pw COD, TOC, O&G removal Gondal et al. (2014)
cellulose esters
3 Mullite and mullite–alumina Synthetic and real TOC rejection Li et al. (2014)
ceramic PW
4 Hydrophilic and oleophobic Synthetic and real Oil–water separation & Karimnezhad et al.
PVDF hollow fiber PW fouling resistance (2014)
5 Graphene oxide modified Three different Oil–water separation Campos et al. (2002)
Al2 O3 synthetic waters
6 Zirconium oxide ceramic MF Synthetic PW Cost estimation Abbasi et al. (2010)
7 UF Hydrophilic cellulose based Synthetic PW TOC removal & Fouling Abadi et al. (2011),
UF membrane resistance Zhu et al. (2014)
8 PAN-g-PEO Produced and O & G removal & fouling Ayse (2009)
refinery water study
9 PEGlated PSf-UF membrane PW and soyabean Fouling resistance Chaturabul et al.
oil (2015)
10 PVP hydrophilized Real PW TSS, TOC, COD, O&G Pancharoen et al.
polyethersulfone UF removal (2009)
membrane
11 PVDF/MWCNT Refinery Waste TiO2 Photocatalytic organic Hu et al. (2015)
nanocomposite UF water removal by membrane
membrane
12 PAN350 UF membrane Real PW Effect of membrane Weschenfelder et al.
operating conditions—T, (2015)
TMP, CFV
13 Commercial UF membranes, Refinery waste Membrane operational Wandera et al. (2011)
PAN (MWCO–20 kDa) and PSf water parameter optimization—T,
(MWCO–30 kDa) TMP, CFV and pH
14 NF & RO Commercial NF & RO Oil sands Effect of pretreatment on Seyed Shahabadi
membranes (GE-Osmanics) process-affected NF and RO membrane and Reyhani (2014)
water process
15 Dopamine-PEG Flowback Effect of membrane Qiao et al. (2008)
hydrophilized UF & RO water—Barnett modification for the fouling
Shale resistance
16 Commercial ceramic MF, UF Synthetic & real Effect of pretreatment, Jiang et al. (2013),
and NF membrane PW backwashing, chemical Loganathan et al.
cleaning & TOC removal (2015), Vatai et al.
(2012)
17 Commercial MF, UF, NF & RO Oil & gas well PW Effect of various membrane Ebrahimi et al. (2013)
combinations and cost
estimation
18 Commercial NF & RO Petroleum refinery Comparative study of NF Xu and Drewes
PW and RO membranes (2006), Xu et al.
(2008)
19 Commercial NF & RO PW Comparative study of NF Kim et al. (2011),
and RO membranes & Alpatova et al. (2014)
fouling study
20 Commercial RO membrane Real PW Electrocoagulation as Miller et al. (2013)
pretreatment and then
direct RO process
21 Commercial UF & RO Coal-bed methane TDS, COD, NH3 and Cl− Ebrahimi et al. (2009)
PW removal
22 Commercial RO Refinery waste Effect of TMP, CFV, Temp., Ebrahimi et al.
water pH (2010), Ebrahimi
et al. (2012)
23 Commercial UF and RO Refinery waste Large scale demonstration Çakmakce et al.
membrane water (100 m3 /h) for more than (2008)
four years operation
24 MD DCMD Oil field PW Microfiltration and Francis et al. (2014)
activated carbon
pretreatment and cost
calculation for the DCMD
25 FO & MD FO & MD Synthetic PW FO showed >99.9% oil Maab et al. (2012)
rejection & MD showed
99.99% salt rejection
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
matters can also exist in the PW as dissolved, dispersed and
emulsified forms. In addition to the inherent composition of
PW from its source, additional chemical compounds are used
for the oil recovery from the well. They are namely corrosion
and scale inhibitor, demulsifier, polyelectrolyte, methanol and
glycols (Igunnu and Chen, 2012). Apart from dissolved and dis-
persed compounds, the PW also contains solids which ranges
from clay, waxes, bacteria, sand and other suspended solids
based on production chemicals. Additionally dissolved gases
also present in the PW and the three major gases are CO2 , O2
and H2 S.
1.2. Challenges and needs for the PW treatment
The conventional way of dealing with the PW in the past
was mainly physical separation of oil and water using gravity
and subsequent discharge to the environment. However, the
regulation is progressively becoming more stringent as more
evidence suggested that the impact of PW on the environ-
ment (Witze, 2015). The PW water discharged into sea should
contain less than 40 ppm of OIW (oil-in-water) for majority of
the countries, however more countries are attempting for fur-
ther stringent regulatory standards (Igunnu and Chen, 2012).
Consequently most oil companies are attempting to reach
“zero-discharge” of harmful compounds to the environment. It
was reported that about 80% of PW has been re-injected to the
well in order to manage the PW and maintain the pressure of
the oil well. Since many oil producing regions are in the water
stressed zone it makes more economical sense to utilize the
PW for the beneficial use of irrigation or live stocks (Igunnu
and Chen, 2012). Therefore, efficient treatment of PW should
meet the dual target of cleaning the polluted water as well as
recovering the purified water for the beneficial use. Among all
the available water purification technologies, the membranes
can be cost effective water treatment method with required
criterions of the treated water for the final use or discharge
(Mulder, 1996; Baker, 2004). One of the main issues in the PW
treatment using membrane technology is to reduce and solve
the complexity of the organic matter in the pre-treatment
stage in order to reduce the fouling of the membranes. The
high content of inorganic matters in the PW on the other hand
makes it inherently high osmatic and therefore high operating
pressure for the nanofiltration (NF) and reverse osmosis (RO)
membrane process. Additionally, the presence of high TDS is
one of the main reason for the frequent scale and silt forma-
tion on the membrane surface. Thus, it is imperative to reduce
the inorganic as well as organic matters as much as possible
in the pretreatment stages in order to make the membrane
treatment practically feasible as well as economically viable
process. In this review article, we focus on the use of mem-
brane technology for the PW treatment in holistic manner in
historical perspective followed by recent developments in the
membranes for the PW treatment. In the following section,
we first discuss the primary and secondary treatment of PW
followed by membrane technology.
2. Primary and secondary treatment of PW
The earlier use of membrane technology for the PW treat-
ment has not been successful due to the presence of high
fouling constituents in the PW (Igunnu and Chen, 2012).
In order to overcome this shortcoming due to the inherent
nature of PW, the pretreatment of the PW before feeding to
187
the membrane process is most important aspect of any PW
treatment process. Most of the pre-treatment methods for
the PW treatment were similar to the municipality waste
water treatment namely primary and secondary treatment.
Primary treatment includes screening, grit removal, gravity
assisted sedimentation, demulsification (Deng et al., 2005),
coagulation, adsorption, dissolved air precipitation (Thoma
et al., 1999) and induced air flotation (IAF) (Zhang et al.,
2005). The secondary treatment involves the removal of
dissolved components through chemical and biological treat-
ment. The chemical treatment includes chemical oxidation
(Shokrollahzadeh et al., 2012), including advance oxida-
tion process (AOP) (Buthiyappan et al., 2015; Hasan et al.,
2012), biological oxidation (Palmer et al., 1981), electrodialysis
(Dallbauman and Sirivedhin, 2005), electrocoagulation, electro
flocculation (Wang et al., 2014) etc. The biological treatment
includes trickling filter and activated sledge treatment etc.
and, as a recent development, the biological treatment in com-
bination with membrane known as “membrane bioreactor”
(MBR) technology is one of the active research area. The prin-
cipal aim of primary and secondary treatment is to reduce the
pollutant level as low as possible with cost effective manner. In
many cases, more than one primary or secondary treatment
processes are combined in order to have effective pollutant
removal. In the following section we discuss selective primary
and secondary treatments for the PW from the literature fol-
lowed by summary of current status of technology.
The combination of Induced Air Flotation (IAF) as primary
and photo-Fenton process as secondary treatment has been
reported for the treatment of oil in water (da Silva et al., 2015).
The real crude oil was mixed with standard synthetic saline
water to mimic the PW. The resulting synthetic effluent con-
tained 300 ppm of oil. The IAF process in combination with
non-ionic surfactant, ethoxylated fatty alcohol, removed the
oil content in the effluent up to 90%. The remaining ∼35 ppm
oil in the water was removed by photo-Fenton process using
H2 O2 and FeSO4 ·7H2 O. The pH was maintained at 3.0 with
the addition of H2 SO4 . The effect of surfactant concentra-
tion in the IAF and dissolved oxygen in photo-Fenton process
also studied. It was shown that at the optimum condition
of 10 min., IAF in combination with 45 min., photo-Fenton
process, 99% of the oil in the effluent can be removed. In
an earlier work, the experimental conditions were optimized
using the same combination of IAF and photo-Fenton process
for the removal of xylene from the synthetic waste water (da
Silva et al., 2012). Apart from usage of chemicals in the photo
oxidation process, the acidification of pH 3 and possible neu-
tralization later stage will be a major drawback of this process
for the large scale application.
The judicial use of coagulant and flocculent is one of the
robust options for the primary treatment of PW. Based on
the nature of PW, sometime the standalone process of coag-
ulation or flocculation alone was not effective. For example,
the removal of TSS from the PW using FeCl3 or Al2 (SO4 )3
alone as coagulant were showed unsatisfactory result due to
formation of smaller size flocs which was not removed by
sedimentation (Paolo Roccaro et al., 2014). By combining with
ionic polymers as flocculent, much improved TSS removal was
reported and the resulting water contained TSS within accept-
able limit of 80 mg/L. In the study, cationic polyelectrolyte
DF 492 and anionic polyelectrolyte DF 973 were used as floc-
culent. The cationic DF 492 using 5 ppm showed better TSS
removal in combination with 5 ppm Al2 (SO4 )3 . It was claimed
that the cationic polymer formed stronger attachment with
188 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
the floc formed hence better removal. The sequence of the pre-
treatment also has considerable impact on the outcome of the
process. The effect of softening sequencing on electrocoagula-
tion (EC) for the treatment of PW was reported by Esmaeilirad
et al. (2015) The influence of chemical softening on EC was
studied and for the chemical softening the pH was manip-
ulated by addition of NaOH and HCl. About 99% of turbidity
removal for one month old PW was achieved by softening
before EC, compared to 88% for the EC followed by softening. It
was reported that it was more effective to carry out chemical
softening followed by EC for the turbidity removal from the
PW. Neither configuration was effective for the fresh flowback
PW (one and two day samples). It was hypothesized that this
might be due to the decrease in organic matter for the older
sample.
The combination of coagulation and bioaugmentation were
used for treatment of oil field wastewater originating from
the use of alkaline, surfactant and polymer (Yan et al.,
2014). For coagulation, polymeric ferric sulfate (PFS), polymeric
aluminium chloride, polyaluminum ferric chloride (PAFC),
aluminum sulfate (alum) and ferric sulfate were used with
varying dosage. The PFS showed about 80% COD removal
at 600 mg/m3 dosage. The indigenous bacterial communities
were used for the biological hydrolysis/acidification treatment
and followed by bio-contact oxidation treatment. The bacterial
communities include Pseudomonas, Acinetobacter, Bacil-
lus, Flavobacterium, sulfate-reducing prokaryotes, a PAM-
degrading bacterium (Bacillus cereus NY001), a petroleum
sulfonate-degrading strain (Burkholderia sp. NY003), and a
petroleum hydrocarbon-degrading organism (Pseudomonas
spp. NY012). The aerobic hydrolysis/acidification treatment
removed more than 90% BOD5 and 70% COD after 40 h of
hydraulic retention time. As the final stage of bio-contact oxi-
dation at HRT of 20 h, the average values of COD, BOD5 , oil,
PAM, and NH3 N of the effluent decreased to 116, 19, 4, 11, and
20 mg/L, respectively from the PW, which corresponds to the
removal efficiencies of 65.3%, 86.1%, 52.7%, 35.2%, and 78.4%,
respectively.
Though the use of electrolysis for PW treatment in large
scale is uncommon, the COD removal by electrolysis using
Ti/RuO2 electrode was reported for the real PW treatment
(Santos et al., 2013). The two different types of samples, before
flotation (BF) and after flotation (AF), were collected from the
refinery were subjected to the electrolysis. Because of the
partial elimination of COD during the flotation, the current
density or time required varied for these two samples. For
example, with a current density of 10 mA/cm2 with 120 min
electrolysis, 96% COD removal was observed for the AF sam-
ple whereas only 87% observed for the BF sample. Complete
COD removal was achieved by increasing current density to
30 mA/cm2 . At this current, the BF sample needed 60 min elec-
trolysis and for the AF sample only 30 min electrolysis needed.
The calculated cost of complete COD removal based on this
work is US$ 38/kg COD for the BF and US$ 28/kg COD for the AF
sample. Though the cost for AF is considerably less compared
to BF, at this cost, this process might be prohibitively expensive
for PW containing high COD especially from gas field which
has COD of as high as 120,000 mg/L or refinery waste water
(see Table 1).
In an interesting study, the use of plasma gliding arc for
the removal of bicarbonate in PW was reported (Wright et al.,
2014). The aim of the work was to study the effect of plasma on
the removal of bicarbonate and the effect of high concentra-
tion of calcium in the absence of bicarbonate on scaling. The
plasma treatment effectively removed the bicarbonate to neg-
ligible level and no calcium scaling was observed up to 30 h at
high temperature in the absence of bicarbonate. Although this
method can be useful for high scale forming PW, the cost effec-
tiveness and large scale feasibility study were not reported.
The beneficial use of combination of pre-treatment pro-
cesses rather than single process in order to achieve the
effective treatment of PW was reported by El-Naas et al. (2014)
A three-step process combination of electrocoagulation (EC),
spouted bed bioreactor (SBBR), and adsorption on activated
carbon were the three methods used for the study. The authors
used the real PW for the COD and phenol removal by these
three combined methods by varying the process configuration.
As a standalone process, none of three methods provided the
satisfactory treatment of PW for the discharge. At first, the EC
treatment followed by activated carbon and finally SBBR were
attempted for the PW treatment. In this configuration, the EC
as first stage removed 46% COD. The activated carbon as sec-
ond stage removed the cumulative COD of 83%. The final SBBR
reactor did not improve the COD removal at al. Thus the overall
COD removal in this configuration was about 85%. The acti-
vated carbon reduced the COD initially to Zero in the first 3 h
and then gradually saturated. The same configuration showed
the phenol reduction of 30% by EC followed by more than 85%
reduction using activated carbon adsorption as second stage.
The SBBR further reduce the phenol to above 95%. In another
configuration, the authors exchanged the treatment sequence
of SBBR and activate carbon. In this, the SBBR treatment was
done ahead of activated carbon adsorption. In this configura-
tion, the SBBR removed the COD by about 73% cumulatively
with EC and activated carbon finally removed up to 97% COD
cumulatively. It is interesting to note that, SBBR as final treat-
ment for the COD does not show any improvement at all in
the previous configuration. The SBBR as second stage in the
process removed the phenol up to 61% followed by activated
carbon process which essentially removed all the phenol com-
pletely. Thus it was demonstrated that the superiority of the
proper combination of the process configuration and the COD
was removed from initial 4190 ppm to 110 ppm and phenol
and cresols (originally 12.2 ppm & 75 ppm, respectively) were
removed completely.
The use of membrane sequencing batch reactor (MSBR) for
the hypersaline oily waste water treatment has been reported
(Pendashteh et al., 2012). The effect of varying organic and salt
concentrations was reported in the study. The organic loading
was varied from 0.281 to 3.372 kg COD/m3 .day and salt con-
centration from 35,000 to 250,000 mg/L with MSBR cycle time
of 12 to 48 h. Using synthetic PW with an organic loading of
1.124 kg COD/m3 .day with 35,000 mg/L of salt concentration,
removal efficiencies of 97.5%, 97.2%, and 98.9% for the COD,
total organic carbon (TOC), and oil and grease (O&G), respec-
tively, were achieved in 48 h reaction time. However, the real
PW showed less efficiency vis-à-vis 86.2%, 90.8%, and 90% for
the same conditions. At the highest TDS of 250,000 ppm, the
COD removal was only about 90.4% and 17.7% for synthetic
and real PW. It is important to note, in both cases of two dif-
ferent salinities, the efficiency difference between real and
synthetic PW by MSBR process is vast and more so in the case
of high salinity PW. The use of hollow fibre submerged mem-
brane bioreactor for the treatment of brackish oil and natural
gas field PW was reported (Kose et al., 2012). After one year lab
scale operation, stable performance was reported despite vari-
ations in the influent at different sludge retention time (SRT).
The COD removal was about 80–85% and the removal of oil
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
and grease increased from 60% to 85% by increasing the sludge
age. The linear hydrocarbon C9–C13 was effectively removed
up to 99% efficiency. The physical cleaning restored 60% of
initial flux and further 95% of initial flux was re-established
after subsequent chemical cleaning. In most cases the biolog-
ical treatments are more effective in less toxic PW and this
can be facilitated by the effective use primary and secondary
chemical treatment.
In recent times the use of surface wettability by tuning the
surface chemistry of filters gained attention for the separa-
tion of immiscible liquids. By changing the surface energy
of the filter to very low or high, it is possible to change the
wettability of the surface in to superhydrophobic or super-
hydrophilic. With optimum porosity and surface energy, one
can separate the oil-water mixture based on the wettabil-
ity of PW on the filter. Using this principle, simple filtration
method based on the surface energy of the filter and medium
has been studied as bench scale demonstration. Thus super-
hydrophilic and underwater superoleophobic nature of the
surface was used for the oil–water separation (Gondal et al.,
2014). To obtain the surface with specific wetting character,
TiO2 nanoparticles were spray coated on stainless steel mesh
(50 & 100 ␮m). Due to the TiO2 coating, the surface of SS
mesh becomes superhydrophilic as well as underwater super-
oleophobic. Thus the modified SS mesh showed the gravity
driven oil-water separation up to 99% separation efficiency.
The hierarchical superhydrophobic surface made of layer-by-
layer (LBL) self-assembly for the oil-water separation has also
been reported (Li et al., 2014). The SS mesh or SS fibres felt
were used to coat SiO2 nano particles having different size
by serious of LBL method. After obtaining the required rough
surface, the silanol OH group was surface modified using per-
flourotriethoxysilane by chemical vapour deposition (CVD).
The resulting superhydrophobic fibers separated Hexane-in-
water emulsion by non-sieving coalescence mechanism with
separation efficiency up to 99.4%.
Utilising similar concept, alumoxane modified Kevlar fab-
ric has been reported for the separation of oil/water (n-hexane
in water) (Karimnezhad et al., 2014). The surface modified alu-
moxane was attached to the Kevlar fabric using the hydrogen
bond and the resulting enhanced hydrophilic fabric was used
for the dead-end filtration by the gravity force. The neutral
pH proved to be optimum with flux of 7392 LM−2 /h at rejec-
tion of 89%. The fouling gel layer formed on the surface was
removed by distilled water and acidic solution which resulted
a flux recovery of 89%. The unmodified Kevlar fabric showed
oil rejection of less than 20% due to less hydrophilic in nature
compared to the alumoxane modified fabric. The highly
hydrophilic nature of alumoxane modified fabric induced the
selective passage of water by acting as barrier to oil. However
all the reported surface energy based separations are labora-
tory scale studies and their long term stability is an issue. The
separations are not based on molecular level and separation
occurs wetting based sieving on macro scale. In future, wet-
ting based filters may compete and complement the physical
separation methods like skimmer. The main disadvantage of
wetting based filters is when the oil droplet size is less than
20 ␮m size. These are effective when the size of droplets are
more than 20 ␮m and it is unclear the effectiveness for finely
dispersed organic compounds. For obvious reason, the dis-
solved organic compound separation is not possible by this
methodology. Additionally more research studies are needed
at the pilot scale in order to evaluate the suitability for the
industrial scale application.
189
To summarize, for the primary and secondary treatment of
PW many existing technologies can be effectively applied for
the PW treatment. After effective initial screening and sepa-
ration of oil based on gravity, the chemical treatments along
with coagulants could be the choice to reduce the contami-
nant further in order to have reasonably good quality feed for
further membrane filtration. Based on the nature of PW, espe-
cially for the less toxic PW, the biological treatments can be
cost effective method. Recent developments in advanced oxi-
dation process and membrane bioreactors will facilitate the
future applications. After initial screening and gravity sepa-
ration, wetting based oil separation could also have potential
application for the PW treatment in future.
3. Produced water treatment by membrane
technology
The principle of membrane separation of the constituents is
based on exclusion, mostly but not always, based on size using
a selective barrier layer. Most of the modern day membranes
are prepared using synthetic organic polymers; nevertheless
inorganic ceramic membranes are also widely used. For water
purification practice, according to the pore size, the mem-
branes can be classified as microfiltration (MF), ultrafiltration
(UF), nanofiltration (NF) and reverse osmosis (RO). The pore
size of the selective layer decreases from MF to UF and then NF
and finally RO, and therefore hydrodynamic resistance for the
liquid passage increases. Because of the pore size change, the
applied pressures for the membrane operations vary signifi-
cantly and the operating pressure increases dramatically as
the pore size of the membrane decrease. The MF and UF oper-
ation are based on convective pore-flow mechanism (Darcy’s
law) whereas the RO membrane operation is based on solu-
tion diffusion mechanism (Fick’s law). The NF membrane is in
between pore-flow and diffusion mechanism and sometime
it is called loose RO membrane (Baker, 2004). As the name
implies, the microfiltration operates in the pore size range
of 0.1–10 ␮m and accordingly it can reject suspended solids,
bacteria etc., but allows the passage of the virus, divalent ions
etc. The pore size of ultrafiltration is in the range of 2–100 nm
and it rejects most of the macromolecular colloids and virus,
but most of the dissolved ionic species passes through the
membrane. It has been identified that the occurrence of transi-
tion between pore-flow to solution-diffusion is about 0.5–1 nm
diameter and this is the range where NF membrane pore size
should be for the ideal operation (Baker, 2004). Thus, the ideal
NF membrane should reject the divalent and multivalent ions
like Mg2+ , Al3+ , SO4 −2 etc., whereas it should allow the pas-
sage of monovalent ions like Na+ , Cl− etc. Since RO membrane
operates based on solution diffusion mechanism, the ideal RO
membrane should reject all the ionic species and only water
molecule should pass through the membrane and in most
cases the rejection of solutes from water is more than 99%.
Based on these principles, in the following sections, we focus
the use of MF, UF, NF and RO membrane for the PW treatment.
Apart from the principle of separation based on the nature
of pores, the major difference between MF, UF and NF, RO, is
the membrane preparation process. The MF and UF are usu-
ally integral membrane, and most of modern UF membranes
are asymmetric in nature too, where whole membrane is a
single polymeric material with or without mechanical sup-
port. The mechanical support can be non-woven polyester or
other supports. On the other hand the RO and majority of the
190 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
NF membranes are composite membranes where the top thin
separating layer in the range of 100–200 nm are usually cross-
linked polymers on the support of UF membrane. A typical
RO or NF membrane can be prepared on the top of the UF
membrane by interfacial cross-linking polymerization.
There is one important factor to be considered when mem-
brane technologies are used in real application. There is a large
gap between the intrinsic membrane capability and actual
membrane performance. For example, MF & UF can have high
pure water flux whereas in real life application like food and
dairy industries, it can be less than 10% of pure water flux. This
is mainly due to the concentration polarization and fouling of
the membrane. Therefore in reality, rather than membrane’s
intrinsic property, it is the feed factors like concentration
polarization and fouling will decide the performance of the
membrane. In summary, most of the membranes can’t achieve
its theoretical limit in real application. One has to keep this is
mind when membrane technology is applied for the separa-
tion process. For more detailed information about the basics
of membrane technology the readers are directed to follow the
books about membrane science and technology (Mulder, 1996;
Baker, 2004).
3.1. Microfiltration (MF)
MF is a process of sieving particulates based on the pore size
of the membrane. The pore size of MF is about 0.1 ␮m or more,
therefore its application in PW treatment should be consid-
ered as a pre-treatment in order to enhance the effectiveness
of UF, NF or RO processes. Polymeric or ceramic MF (Li and Lee,
2009; Cui et al., 2008) membrane are widely available for the
treatment of PW. One of the earliest report dealt with oil-water
separation is the use of ceramic and polyacrylonitrile (PAN)
(Mulder, 1996) MF membranes (Jeffrey Mueller et al., 1997). The
ceramic MF of 0.8, 0.2 ␮m and 0.1 ␮m PAN membranes were
used for the study and the synthetic PW was made by using
heavy crude oil (API 12 weight, density 0.972 g/cm3 ) in the con-
centration range of 250–1000 ppm. It has been observed that
the increased oil concentration decreased the water flux and
cross flow velocity (CFV) and transmembrane pressure (TMP)
and temperature (T) had little impact on the final flux. The
permeate contained less than 6 ppm of oil and interestingly
addition of suspended solids (250 ppm diatomaceous earth)
increased the final flux of the membrane by one order of mag-
nitude. It was hypothesized that that the suspended solids
adsorb the oil, break up the oil layer on the membrane sur-
face thereby increasing the permeation flux. The bigger pore
size membranes, i.e., 0.8 and 0.2 ␮m ceramic membranes were
found to have both internal and external fouling whereas the
0.1 ␮m PAN had only external fouling; i.e. membrane top sur-
face layer fouling. All the surfaces of the membranes were
found to be hydrophobic due to oil fouling and these oily lay-
ers were resistance to hydrodynamic shear method for the
removal. In another early work, Campos et al. (2002) reported
the use of MF and biological process for the PW treatment.
The coarse filtration was used as pre-treatment and the PW
was subjected to 0.1 ␮m commercial mixed cellulose esters MF.
The MF process removed the COD, TOC, oil & grease (O&G) and
phenol of 35%, 25%, 92%, and 35%, respectively. Subsequently
the polystyrene particle beads air-lift reactor was used as
biodegradation. The lowest studied hydraulic retention time
(HRT) of 12 h showed the COD, TOC, phenol and ammonium
removal of 65%, 80%, 65% and 40%, respectively. The com-
bined MF and biodegradation process showed the COD and
TOC within the discharge limit (230 and 55 mg/L, respectively),
but the phenol and ammonium concentrations were above the
permissible limit. It was clear from these earlier studies that
the use of MF alone will not be sufficient for the treatment
of PW and most importantly the high fouling of membrane
discouraged the wide use in the PW treatment.
The use of mullite and mullite–alumina ceramic MF mem-
branes for the treatment of PW was reported by Abbasi et al.
(2010) The mullite-alumina MF membranes were prepared
using kaolin clay with ␣-alumina by extruding the water mix-
ture followed by room temperature drying. Finally the room
temperature dried membrane was calcinated at 1250 ◦ C for
3 h followed by strong alkali leaching to remove the free sil-
ica. The prepared MF ceramic membrane was subjected to
study the effect of pressure, temperature, CFV, oil and salt
concentration. It has been observed that the permeation flow
increased with increasing flow rate, temperature, pressure and
alumina content of the mullite–alumina membranes whereas
permeation decreased with increasing oil concentration in the
PW. At elevated temperature and high CFV, the fouling resis-
tance decreased. The rejection was slightly higher with the
increased pressure, salt concentration, oil concentration and
alumina content of the mullite–alumina ceramic membranes.
It has been found that the TOC rejection of 94% for the syn-
thetic oily wastewater but only 84% of the TOC rejection for
the real PW using the same membrane. It is important to
note that the reported study showed less permeation flux and
lower rejection for the real PW compared to synthetic PW. A
tubular ceramic ␣-Al2 O3 MF was reported for the treatment
of Tehran refinery (Abadi et al., 2011). The microfiltration pro-
duced the water with oil and grease content of 4 mg/L which
was within allowed limit. The operating parameters like TMP,
CFV and temp. on the effect of permeate flux, TOC removal
and fouling resistance were studied. It was also reported that
the backwashing could prevent the flux decline significantly.
Recently, Zhu et al. (2014) reported the use of hydrophilic
and oleophobic PVDF hollow fiber MF membranes for the
oil-water separation. For this novel membrane, the P(VDF-
co-CTFE)-g-(PMAA-co-fPEG) has been synthesized by grafting
tert-butylmethacrylate (tBMA) on poly(vinylidenefluoride-co-
chlorotrifluoroethylene), P(VDF-co-CTFE), using ATRP. Sub-
sequent hydrolysis of tBMA segment resulted the PMAA
grafted on P(VDF-co-CTFE) backbone. The acid group of
the grafted MAA was further esterified either with perflu-
oroalkylpolyethyleneglycol(fPEG) or polyethyleneglycol (PEG)
to obtain the additive polymer of P(VDF-co-CTFE)-g-(PMAA-
co-fPEG or P(VDF-co-CTFE)-g-(PMAA-co-PEG). The modified
polymers were added as additives to the PVDF dope solution in
order to make hydrophilic and oleophobic PVDF hollow fiber
MF membranes. Three different types of oils namely ‘H-oil’
(Hexane in water), ‘C-oil’ (crude oil in water) and ‘P-oil’ (palm
oil in water) were studied for the oil-water separation using the
new hydrophilic and oleophobic PVDF hollow fiber membrane.
The ‘H-oil’ removal efficiency of >98% was achieved and the ‘C-
oil’ removal was impressive, >99%. However the ‘P-oil’ showed
only about 70% separation efficiency due to the smaller size of
oil droplet in the feed water. The main aim of the work was
fouling resistance and this novel membrane showed excel-
lent flux recovery after back washing and importantly the
rate of flux reduction is considerably lower compared to the
controlled PVDF hollow fiber MF membrane which has been
prepared without the hydrophilic additive polymer (Fig. 3).
From the figure, it is clear that the PVDF membrane with-
out hydrophilic additive loses most it its water permeability
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202 191

Fig. 3 – Fouling resistance hydrophilic PVDF membrane for the oil–water separation. M1 (blue line) represent the control and
red and green lines (M2 & M3) represent the hydrophilic additive PVDF membrane. Experimental condition: First 30 min DI
water filtration (conditioning) followed by 500 ppm crude oil in water. After 2 h filtration, the membrane washing 10 min
followed by 20 min backwashing with DI water and then second cycle of oil-water separation (Reproduced from Zhu et al.,
2014). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

irreversibly within few hours of operation. However, the mem-
branes prepared with hydrophilic additives showed more than
95% flux recovery by simply washing and backwashing with
DI water at the membrane operating pressure, ca. 0.34 MPa.
Because of it excellent flux recovery using well known PVDF
polymer, which has been used as commercial membrane for
long time and the notable exceptional property of new mem-
brane compared to the existing PVDF membrane makes it one
of interesting candidate for further pilot and industrial scale
study.
The integration of coalescer bed and MF membrane pro-
cess was reported by Motta et al. (2014) for the synthetic
oil/water separation. The cation exchange resins were used
for the coalescer bed formation working in up-flow condi-
tion with polyetherimide hollow fiber MF submerged module
with permeation area of 0.5 m2 . Using optimum condition, the
efficiency of 93.0% to 100% oil removal achieved and this corre-
sponded to an O&G concentration in the final effluent between
0.1 and 14.8 mg/L. The coalescer bed first removed about half
of the O&G and the remaining O&G was removed by the
membrane process. The combination of these two processes
was suggested as pretreatment to minimize the operational
problems like fouling for further treatment by UF and RO mem-
brane. Additionally it was suggested that the coalescer bed and
MF processes can be made as a loop, thereby the retentate of
the MF membrane can be sent back to coalescer bed to recover
the oil.
The use of Hollow fiber supported liquid membranes
(HFSLM) for the removal of mercury (II) from the PW water
was reported by Chaturabul et al. (2015) The HFSLM was made
by impregnating the commercial hollow fiber module with
surfactant Aliquat 336 by toluene solution circulation and
thiourea was used as stripping agent. It has been reported
that 99.7% extraction with a recovery of 90% was possible
using the optimized condition. A theoretical model was devel-
oped and the prediction agreed well with the experimental
result. The treated water contained less than 5 ppb of mer-
cury which is within the limit for the discharge. In a similar
previous study, Cyanex® 923 (a mixture of phosphine oxide),
tri-n-butylphosphate (TBP), bis(2,4,4-trimethylpentyl) dithio-
phosphinic acid (Cyanex® 301), tri-n-octylamine (TOA) and
methyltrioctylammonium chloride (Aliquat 336) were used
as the extractants and NaOH solution as stripping agent
(Pancharoen et al., 2009). Since Hg is highly toxic substance
and its removal is high importance in PW. However it might
be interesting to check the longevity and large scale viability
of these types of system in long term operation.
Recently graphene oxide modified Al2 O3 based MF mem-
brane was reported for the synthetic oil/water separation by
Hu et al. (2015) It has been observed that the efficiency of GO
modified membrane was higher for the oil removal with ∼28%
higher permeate flux compared to unmodified membrane.
However, the membrane stability to withstand the backwash-
ing and chemical cleaning has not been reported. In the future,
more research on the applications of these new materials will
assist to understand the effect of these materials in the mem-
brane technology.
The cost for the PW treatment using membrane technol-
ogy is very important for any viable application. Thus the cost
estimation for the synthetic PW treatment using zirconium
oxide ceramic MF was reported by Weschenfelder et al. (2015)
The water flux for the synthetic produced water with differ-
ent oil and salt contents was studied with varying CFV and
TMP. The optimum CFV has been found at 2.0 m/s and based
on the optimization, the capital and operating cost was calcu-
lated for the water recovery rate of 50%, 80% and 95%. The cost
of membrane regeneration by backwashing with 1000 mg/L
of each NaOH followed by sodium hypochlorite and finally
distilled water wash was also included. Based on the exper-
imental results, the capital and operating cost was calculated
for the recovery rate of 95% water for CFV from 0.3 to 4 m/s as
a function of the unit volume of treated water in m3 for the
PW treatment of 1000 m3 /h. A total US$ 3.21/m3 was reported
as the lowest possible cost. The salt rejection has not been
reported in this study and it can be assumed that it should
be negligible due the use of microfiltration process. It should
be also noted that it has been calculated using NaCl alone as
salt constituent with refinery oil. However the real PW sam-
ples are much more complex in terms of oil contents, other
organic and inorganic pollutants. Nevertheless the cost esti-
mation is useful tool for the oil removal and MF treatment can
be considered as pre-treatment and further cost estimation
can be done for NF and/or RO treatment for the salt removal.
To summarize, the MF can be used as cost effective pre-
treatment for the PW treatment. It can be used after removing
the bulk of the oil component by using the primary treatment
of sedimentation, coagulation, flocculation & sedimentation
process. The MF process can effectively remove the dispersed
oil droplets and other particulates with the size of more than
100 nm. However, the smaller droplets and particulates can
192 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
pass though the MF membrane and for more effective removal
of smaller size, the UF membrane can be employed. Another
major limitation of MF is the lack of removal of dissolved com-
ponent.
3.2. Ultrafiltration
UF process is one of the initial methods attempted along with
MF for the oil removal from the petroleum industry waste
water. However, due to high permeation flux, UF membranes
were easily susceptible to the fouling. Consequently one of
the main objectives in UF membrane study is to reduce the
fouling on the surface of UF membrane. Hence most of the
literature dealing with UF membrane reports the effect of foul-
ing and how to manage or reduce it effectively. The most
effective way of reducing fouling is to make the membrane
surface more hydrophilic and reduce the roughness. There-
fore, in the following section, we discuss some of the selective
reported literature dealing with UF membrane for the PW
treatment. First we focus on the new lab made hydrophilized
UF membranes followed by the examples of commercial UF
membranes.
The fouling resistance hydrophilic cellulose based UF
membrane for the synthetic PW treatment was reported
by Husson et al. (Wandera et al., 2011) In this study,
the low molecular weight cutoff regenerated cellulose
UF membranes was modified by surface initiated atom
transfer radical polymerization (ATRP). Firstly, thermos-
sensitive poly(N-isopropylacrylamide), PNIPAAm, was grafted
followed by poly(oligoethyleneglycolmethacrylate), (POEGMA)
grafting. The resulting poly(N-isopropylacrylamide)-block-
oligoethyleneglycolmethacrylate), PNIPAAm-b-POEGMA, and
PNIPAAm homopolymer grafted membranes were compared
with the unmodified cellulose UF membrane. The synthetic
PW was prepared using soybean oil (5 g/L) and NaCl (0.5 g/L).
The modified membranes showed less water flux compared
to unmodified UF membrane due to the surface pore coverage
by the grafted polymers. However, the slop of flux decline is
far less for the modified membrane compared to the unmodi-
fied membrane. Most interestingly, the modified membranes
showed 100% permeation flux recovery by simple cold water
rinse whereas the unmodified membrane showed only 81% of
initial flux recovery. Though it is 81% of initial flux recovery,
the water wash of unmodified membrane showed insignifi-
cant improvement from the point of recovery point, i.e., the
flux value before and after cold water wash is nearly same.
The TOC removal efficiency of >94% was achieved for all three
membranes and the most importantly, 100% flux recovery
from the modified membranes. Based on the complete flux
recovery, it was claimed that the modified membrane will
be useful as fouling resistance membrane. The same group
further studied the effect of grafting density on the fouling
resistance (Wandera et al., 2012). It was found that the increas-
ing grafting density decreased the permeate flux, but the flux
decline over continues operation is significantly less com-
pared to sparsely grafted membrane and this effect is more
pronounced when compared with the unmodified membrane.
It has been determined that the initial lower permeation flux is
better trade-off in long term use when one considers the rate
of flux decline and washing. In an another earlier study, the
commercial NF membrane (NF270) was subjected to the simi-
lar surface initiated polymerization by anchoring the initiator
on the NF membrane (Tomer et al., 2009). However the mod-
ified membranes showed less permeate flux and importantly
does not show any improvement in flux recovery. The authors
suggested that this might be due to increase in hydrophobic-
ity after the modification with dense polymer layer. Unlike the
NF membranes, the UF membranes showed better flux recov-
ery and thus it was claimed to have better fouling resistance
compared to the unmodified membranes.
The effect of hydrophilic segment inclusion in the PAN UF
membrane for the fouling resistance in the PW treatment was
reported (Ayse, 2009). The effect of the addition of 20 wt% PAN-
g-PEO in the PAN UF membrane preparation for the treatment
of two different PWs and one refinery waste water was studied.
For the comparison, the commercial Serpo PAN400 was used
to differentiate the fouling resistance due to the hydrophilic
segment addition in the UF membrane. The dead-end filtra-
tions of commercial and manufactured PAN membranes were
successful in removing free and dispersed oil & grease. The
hydrophilic PAN recovered the flux by backwashing whereas
the PAN membrane lost more than 25% initial flux irreversibly.
Additionally the rate of flux decline for the hydrophilic UF
membrane is significantly less compared to the plain PAN
UF membrane without hydrophilic PAN addition. This report
clearly indicate the superiority of hydrophilized membrane
compared to the conventional commercial PAN membranes
for the PW treatment.
The effect of PEG grafting density on PSf-UF membrane for
the oil separation from water was reported for two different
type oily waste waters, namely soybean oil and industrial PW
(Yuan et al., 2014). The azide-functionalized PSf-UF membrane
was prepared by phase inversion method. The azide group on
the side chain of polysulfone backbone was PEGlated by click
chemistry using alkyne end-functionalized PEG. Two different
degrees of azide functionalities (17% and 45%) and three differ-
ent MWs of PEG (120, 750 & 1300) were used to study the effects
of grafting density as well as molecular weights of PEG. The
high molecular weight PEG with densely grated PEG showed
more hydrophilic than the rest of the combinations. The mem-
branes with high grafting density using low molecular weight
PEG showed better fouling resistance and permeate flux. The
optimized membrane showed complete oil rejection with a
flux of 120 LMH with over 95% flux recovery by simple water
rinsing. Since PSf is widely used for the UF membrane man-
ufacturing due to the exceptional pore formation with good
mechanical stability, the result might be interesting for the
further study.
The lab-made polyethersulfone (PES) asymmetric UF mem-
brane for the PW treatment has been reported (Salahi et al.,
2015). The hollow fiber membrane was prepared using 17 wt%
PES and 2.5 wt% PVP dope solution for the dry-jet-wet spinning
process. The PVP was added in order to make the membrane
more hydrophilic and to change the surface and subsurface
morphology more porous. The real PW was used for the study
and the use of PES UF membrane showed TOC rejection of
96.3%, COD (83.1%), turbidity (99.3%), and oil and grease con-
tent (99.7%), and complete rejection for TSS with a reasonably
high permeation flux of 84.1 L/m2 h and fouling resistance of
63.0%. It was reported that the quality of the UF treated PW is
better than the cooling tower feed water used in the Tehran
oil refinery. The authors also used adaptive neuro-fuzzy infer-
ence system (ANFIS) model to predict permeation flux in UF
process and it was found good agreement of the model with
experimental data.
The hybrid photocatalytic reactor and PVDF/MWCNT
nanocomposite UF membrane process for treatment of
petroleum refinery wastewater has been reported (Moslehyani
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
et al., 2015). The pristine and oxidized MWCNT were used for
the fabrication of PVDF nanocomposite UF membrane. The
waste water was subjected to photocatalytic reaction using
UV light in combination with 200 ppm TiO2 . More than 90%
of organic matters were decomposed by the photocatalytic
reaction in 6 h UV radiation. After the photocatalytic reaction,
the feed was passed through PVDF/MWCNT nanocompos-
ite UF membrane and the permeate showed >99% of organic
pollutant removal. Interestingly the TiO2 photocatalyst also
recovered in single filtration process of UF membrane with an
efficiency of more than 99% from the catalytic reactor. Thus
the UF membrane was used for dual purpose to filter the water
as well as to recover the photocatalyst.
The physiochemical nature of membrane is important for
the effective separation and in addition to that the membrane
operating conditions are also deciding factor for the practi-
cal use of membrane technology. The effect of membrane
operating conditions like temperature (T), transmembrane
pressure (TMP), and cross flow velocity (CFV) on permeate
flux, fouling resistance, and TOC rejection using commercial
PAN350 UF membrane was reported by Seyed Shahabadi and
Reyhani (2014). According to this study, TMP has high impact
on the permeate flux and fouling resistance where as CFV
has more impact on the TOC removal. The experimental and
full factorial design methodology was used to optimize and
narrow down the impact of each operating conditions. For
the PAN350 UF membrane, it was reported that the signifi-
cance of operating parameters on the outcome as follows: (a)
For permeate flux: TMP > T > CFV > T–TMP > T–CFV, (b) For foul-
ing resistance: TMP > T > CFV > T–CFV > T–CFV > TMP–CFV, (c)
For TOC rejection: CFV > T > TMP > TMP–CFV > T–TMP. By utiliz-
ing best operating conditions, 100% oil and grease, 100% TSS,
99% turbidity, and about 50% TOC removal from real PW was
reported. This result clearly indicates that it is important to
optimize the operating parameters of membrane unit based
on the PW quality in order to achieve the best result.
In similar operational parameter optimization, the refinery
waste water from Tehran API unit was subjected to UF pro-
cess under different operating conditions and the permeate
quality was compared with the existing conventional biolog-
ical treatment which has been used in the in refinery for
the PW discharge (Salahi et al., 2012a). Two commercial UF
membranes, PAN (molecular weight cut-off—20 kDa) and PSf
(molecular weight cut-off—30 kDa) were studied for the pilot
scale operational optimization. After studying the effect of
TMP, CFV, pH and temperature, it was found that the optimum
conditions as following: TMP—3 bar, CFV—1 m/s, T—40 ◦ C and
pH—9. The PAN UF membrane showed better water flux and
fouling resistance than PSf UF membrane. The chemical clean-
ing using EDTA as chelating agent and SDS as surfactant were
effective to recover the permeation flux up to 90%. It was
demonstrated that all parameters, except COD removal, were
superior by UF process than the existing biological treatment
in the refinery. A pilot-scale plant involving aeration tank, air
floatation, sand filter and UF has been reported for the Daqing
oilfield waste water treatment (Qiao et al., 2008). It was claimed
that the hybrid process removed the oil, SS contents in the
oily wastewater to less than 0.5 mg/L, 1.0 mg/L, respectively.
It was further claimed that Fe, reductive substances, sulfate
reduction bacteria (SRB), metatrophic bacteria (TGB) and iron
bacteria (IB) were also reduced to the level of discharging limit.
The combination of ceramic MF & UF with ion exchange
resin for the PW treatment was reported (Jiang et al., 2013).
Three different MF membranes with pore size of 1.4, 0.8,
193
0.2 ␮m and two UF membranes with pore size of 20 and 5 nm
were used for the study. The Marcellus shale gas PW water,
with characteristic TOC of 720 mg/L, TSS of 881 mg/L, TDS
of 48,000 mg/L with conductivity of 67,000 ␮s/cm, was passed
through combination of MF-MF or MF-UF membrane process.
All the combination of membranes removed the TSS > 99%
and MF-MF combination (1.4 ␮m & 0.02 ␮m pore size) showed
about 97% turbidity removal. The MF-UF combinations were
effective to remove the turbidity 100%. The TOC removal was
in the range of 10–15% with negligible decrease in conductivity
which suggests that nearly all the TDS remain in the mem-
brane treated PW. As a final stage cleaning, the ion exchange
resins, DOWEX G-26 with a mean diameter of 0.6 mm and
Marathon A resin with a mean diameter of 0.5 mm were used.
After the ion exchange treatment, the quality of the water
met the criteria (Pennsylvania, USA) for the surface water dis-
charge. The cost estimation for the PW treatment using the
membranes and ion exchange resin will be about 18.4 US$/m3 .
It was highlighted that the ion exchange resin treatment alone
has the share of nearly 3/4 of the treatment cost. There-
fore it can be assumed that the reduction of cost should
come from ion exchange resins or using alternative final stage
cleaning.
The effect of pre-treatment on the performance of the
ceramic UF membrane for the treatment of oil sands tail-
ings pond recycle water was reported in a pilot scale study
(Loganathan et al., 2015). Two different methods of pre-
treatment employed and the first pretreatment step was the
use of coagulant. It was reported that the coagulant removed
the solids completely. The second method was use of coag-
ulant with softening agent (MultifloTM system). The use of
softening agent helped to reduce the membrane fouling by
removing the hardness causing ions to negligible level. More
than 99% of turbidity and less than 20% of TOC were removed
by the pretreatment. Thus pre-treated water showed relatively
high flux, ca. 120 L/m2 h, with a recovery of more than 90%
water in the ceramic UF. For improving the water quality fur-
ther, it was suggested that the ceramic UF treated water can
be further subjected to RO process for the beneficial use. In a
direct comparison study, the polymeric and ceramic UF mem-
branes (polyaryletherketone and zirconia respectively) were
used for the oil-in-water emulsion separation (Vatai et al.,
2012). These two membranes were compared under various
operating parameters and ceramic membranes are sensitive
to oil penetration at lower CFV and high TMP. Compared to the
polymeric membrane, optimum performance of the ceramic
membranes were found under low feed flow rate and TMP. It
was reported that more studies needed in order to understand
the systems fully.
The use of rotating ceramic filter disc for the produced
water treatment was reported (Ebrahimi et al., 2013). The
commercially available MF (0.2 ␮m pore size) and UF (7 nm
pore size) ceramic membrane were used for the PW treat-
ment. Two different PWs, namely tank dewatering-produced
water (TDPW) and oil model solution (OMS) prepared using
crude oil, were used for the study. The TDPW had high
pollutant compared to OMS with characteristic of dispersed
oil 200–1000 mg/L, conductivity 20,000–80,000 ␮S/cm, TOC
200–2000 mg/L, iron 66 mg/L and zinc 0.55 mg/L. It has been
shown that more than 99% oil and more than 98% of TOC
can be removed for both TDPW and OMS PW by the ceramic
UF membrane. The rotational speed of the disc filter affected
the performance of the membrane. At high speed rotation,
more permeation flux and less fouling was observed. It was
194 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
also reported that the rejection efficiency was independent of
membrane rotational speed.
In summary, the UF process can be effective membrane
technology for the PW treatment especially for the low saline
and less toxic PW. If the PW is more benign in nature, then
the UF treatment alone can be achieve closer to the discharge
limit, for example in offshore oil field. In more practical terms,
the UF process can be ultimate pretreatment for the NF or
RO membrane where one can get the purified water with rea-
sonably acceptable quality for the beneficial use like irrigation
or live stocks. The main focus of UF membrane development
will be fouling reduction and this alone will be the major
impediment to adapt the UF process in large scale. Therefore
fouling resistance, smart UF membrane development will be
the ongoing and future research development in membrane
technology.
3.3. NF and RO membranes for PW treatment
Both NF and RO membranes are operated at relatively high
pressure compared to MF and UF membranes. In RO mem-
brane, the water permeation flux is directly proportional to
the operating pressure whereas the salt permeation is inde-
pendent of pressure. Therefore, membrane is more selective
when operated in high pressure (Baker, 2004). Both NF and
RO membranes are effective in removing inorganic miner-
als. The major difference between NF and RO membrane is
the selectivity. The RO membrane rejects all the ionic species
including monovalent ions whereas the NF is more selective
for the divalent ions and partially allows the monovalent ions
like Na+ and Cl− . Most of the NF and nearly all of the mod-
ern RO membranes are made as thin film composite (TFC)
membrane on an asymmetric UF membrane support. Since
these membranes are highly selective at the high pressure
operation, these membranes are easily prone to fouling and
therefore require relatively clean feed water in terms of sus-
pended solids, clay, organic foulants etc. In this section we
discuss selective literature report dealt with the PW treatment
using NF and RO membranes.
In one of the earliest study, Xu and Drewes (2006) stud-
ied RO, ultra-low pressure RO (ULPRO) and NF membranes
for the beneficial use of methane produce water. The sam-
ples were collected from sandstone aquifer and considered as
brackish groundwater based on the components in the PW,
namely 5500 of TDS with conductance of 10,000 ␮s/cm with
pH of 8.5. The following commercial membranes were used
for the PW treatment: One RO (TFC-HR, Koch Membrane Sys-
tems), three ULPRO membranes: XLE (Dow/Filmtec), TFCULP
(Koch) and TMG-10 (Toray America), and three NF membranes:
NF-90 (Dow/Filmtec), TFC-S (Koch), and ESNA (Hydranautics).
All the membranes were characterized for contact angle and
mean roughness in order to study the fouling of membranes
by the PW. It was inferred that more hydrophilic and less rough
(smooth) surfaces showed better fouling resistance, and high
permeation flux membranes (for pure water) fouled faster
than low permeation flux membranes. Based on ESEM and
ATR-FTIR study, it was hypothesized that the organic foul-
ing might not be the major mechanism for the flux decline.
Various methods, like hydraulic backwashing by DI water,
acidic (HCl or citric acid) & basic (NaOH) wash, EDTA and SDS
cleaning were studied in order to obtain the original perme-
ation flux. Among all the studied membranes, TMG-10 showed
better recovery by either basic or SDS wash. The remaining
membranes showed below 90% flux recovery. In a continuation
study, 2-stage laboratory scale membranes were employed to
estimate the cost-effectiveness and economic viability of the
membrane treatment (Xu et al., 2008). Based on the previous
optimizations, TFC-HR, TFC-ULP, TMG10 and NF90 selected
for the cost-estimation study. Among chosen membranes,
NF90 required the lowest operating pressure for high per-
meate flux and the rejection was lower than RO and ULPRO
membranes. Though NF membrane met the primary drinking
water standard, it lacked the secondary drinking water stan-
dards with high TDS and chloride. The two ULPRO membranes
showed high flux with comparable rejection of RO membrane.
All the tested membranes showed lower rejection of boron and
it was suggested that further treatment required meeting the
drinking water standard. The PW water of this particular well
contained high iodine concentration (∼50 mg/L) and it was
suggested that the concentrated brine with excess of 100 mg/L
can be used for the iodine recovery. And the final cost analy-
sis showed about one US$/m3 for operating energy cost alone.
The ULPRO membranes provided marginally better operation
and maintenance (O&M) cost than RO membrane and ULPRO
will be more cost effective if the cost of future energy is higher
than the calculated period.
The effect of pretreatment on NF and RO membrane pro-
cess for the desalination of oil sands process-affected water
(OSPW) was studied (Kim et al., 2011). The commercial Dry
flat-sheet NF (DK and DL) and RO (AK and AG) membranes
from GE-Osmonics were used for the PW treatment. A bench
scale coagulation–flocculation–sedimentation (CFS) process
was performed for three different ways of pre-treated PW
samples and these results were compared; The three dif-
ferent methods of pretreatment procedures are as follows:
(1) Solid materials in PW were settled by gravity without
chemical agents, (2) OSPW was treated with CFS plus the
coagulant alum as pretreatment, 3. OSPW was treated with
CFS, alum and coagulant aid–poly(dimethyldiallylammonium
chloride). Without pre-treatment, all four membranes were
fouled within 4–5 h operation with flux reduction of 80% or
more. By utilizing gravity settling as pre-treatment, the flux
reduction about 50% observed for the same operating con-
dition. It was shown that by utilizing CFS with or without
coagulant aid, the flux reduction can be reduced below 40%. All
four membranes showed similar behaviour in the flux reduc-
tion for all three combinations of pre-treatments. The CFS
pre-treatment with or without coagulant aid reduced the tur-
bidity to 0.45 NTU from 71.6 NTU thereby reducing the SDI15 to
1.5 from 6.13. However, the pre-treatments alone do not reduce
any COD or TDS from the feed water which suggests that
the fouling reduction is due the solid content removal by the
CFS pre-treatment. Among the four studied membranes, the
efficient AK-RO membrane with alum pretreatment, achieved
98.5% desalination of OSPW. The chemical cleaning, using
acid and base, was required to remove the fouling layers on
the membranes. In a continuation work, using similar pre-
treatment process, the 1 kDa molecular weight cutoff (MWCO)
ceramic ultrafiltration membrane was subjected for more
detailed study of fouling by the PW on the UF membrane
(Alpatova et al., 2014). Expectedly, the direct filtration of PW
without pre-treatment caused irreversible membrane fouling.
Though permeation flux increased with increasing TMP, the
higher flux decline was observed for the highest TMP. The
lower CFV caused more fouling than higher CFV.
The combination of surface modified UF & RO membranes
for the treatment of flowback water from the Barnett Shale
region of Texas was reported (Miller et al., 2013). The UF &
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
RO membrane surfaces were modified by in situ polymeriza-
tion of dopamine in order to make the membrane surface
more hydrophilic and fouling resistance. Subsequently the
polydopamine modified UF membrane surface was further
modified by reacting with PEG-NH2 . The polydopamine-PEG
modified UF membranes showed relatively higher permeate
flux, lower transmembrane pressure and improved cleaning
efficiency compared to the unmodified UF membrane. The
polydopamine modified RO membrane showed higher salt
rejection than the unmodified RO membrane. Interestingly the
fouling behaviour of the modified RO membrane did not show
any difference compared to the unmodified RO membrane. It
was hypothesized that it might be due to removal of organic
foulants in the feed by the modified UF membrane and the
authors did not report any further study to verify this claim.
Ebrahimi et al. (2009) reported the use of commercially
available ceramic MF, UF and NF membrane combination for
the treatment of model (synthetic) and real PW. Though the
ceramic membranes showed more than 99% oil removal, the
TOC removal was modest. It has been observed that differ-
ent membranes showed different flux reduction and the real
PW showed more drastic flux reduction compared to synthetic
PW. In a follow up study, the authors used DAF as pretreatment
in combination with the ceramic membranes (Ebrahimi et al.,
2010). The DAF process alone removed the oil up to 90% in
90 min with airflow rate of 0.8 Nl/s. The ceramic membrane
treatment process provided up to 99.5% oil removal and TOC
removal was about 50%. It was further reported that the back
flushing was more effective than chemical cleaning using lye
solution for the membrane cleaning (Ebrahimi et al., 2012). The
efficiency of ceramic NF membrane is not comparable with the
polymeric TFC-NF membrane and presumably it might be due
to the high MWCO of the ceramic membrane.
In one of the comparative study, Koyuncu et al. reported the
use of combination of various type membranes for the pro-
duced water treatment (Çakmakce et al., 2008). The primary
treatment namely Dissolved air floatation (DAF), acid crack-
ing, coagulation with lime and precipitation were utilized for
the preliminary treatment. 5 and 1 ␮m cartridge filters were
also employed before subjecting the PW for the membrane
treatment. All four types of membranes namely MF (cellu-
lose acetate), UF (ultrafilic—Osmanics), NF (filmtech—NF200)
in combination with RO membrane (ST10 & AG—Osmanics)
were used for the PW treatment. The MF and UF membranes
were used as pre-treatment to polish the PW and NF and/or
RO membrane as final treatment. It was observed that NF
membrane did not produce the water quality for the direct
beneficial use or discharge due to high TDS, but the COD
removal was within discharge limit. A relatively stable flux was
observed in NF process whereas the RO membranes showed
improved water quality with lesser permeation flux. Based on
the study, it has been proposed that the following configura-
tion for the effective treatment of PW: Primary sedimentation
for the oil separation from water followed by DAF for the
smaller oil particle removal, then 1 ␮m cartridge filter followed
by 0.2 ␮m MF then use of activated carbon and finally RO mem-
brane. It was reported that after studying various combination
of membrane systems, the PW treatment would cost about
5–6 US$/m3 depending on the configuration of the pretreat-
ment and membrane units.
The comparative study of NF and RO membranes for the
PW treatment was also reported (Alzahrani et al., 2013a,b). In
this study, the treated water quality and toxicity removal effi-
ciency of NF and RO membranes were assessed. The authors
195
collected the industrially treated PE before discharge in to
the environment for further NF and RO membrane treatment
for the beneficial reuse. The hydrophilic nanofiltration NF1
and reverse osmosis RO-BW30 from AMFOR INC® , China were
used for the study. Based on contact angle measurements,
it was claimed that the studied membranes were the most
hydrophilic among the commercially available membranes.
Admittedly, the NF membrane removed the divalent cations
more efficiently than monovalent ions with efficiency ran-
ging from 33 to 94% whereas the anion removal was in the
range of 3 to 77%. The NF membrane showed poor removal
of Cl and F anions compared to the RO membrane. A similar
trend was observed for the Na cation removal. The NF mem-
brane showed reasonably improved removal of Mn, Ba, and
surprisingly Fe was more efficiently removed by NF membrane
than RO membrane. Among the 103 parameters studied for
the water quality, 74 international standard parameters were
met by NF membrane alone. The notable exceptions, which NF
membrane could not be effective, were molybdenum, ammo-
nia and boron. The RO membrane removed the boron but
failed to remove molybdenum and ammonia effectively. Thus
treated water further subjected for the toxicity study using
marine luminescent bacterium (Vibrio fischeri). The NF treated
water showed EC50 of 13.65% which is close to the raw PW
which has the EC50 of 14.27. After studying the TOC, COD and
BOD removal by NF and RO membranes, the authors postu-
lated that it could be from the additive chemicals and chemical
product from the oil industry which might have contributed to
the toxicity of the treated PW by NF membranes. The RO mem-
brane treated water showed no toxicity towards V. fischeri. It
was suggested that the NF membrane could be used as pre-
treatment to the RO membrane in order to reduce the fouling.
Mondal et al. (2008) reported the use of nanofiltration and
reverse osmosis for the treatment of co-PW (Mondal and
Wickramasinghe, 2008). In this report, three commercially
available FilmTec membranes namely NF270, NF90 (nanofil-
tration) and BW30 (low pressure reverse osmosis) were studied
for three different PWs with varying concentration of contami-
nants. The high flux NF270 nanofiltration membrane was most
hydrophilic and displayed least reduction in permeation flux.
However it showed less effective in TDS removal whereas the
BW30, low pressure reverse osmosis membrane, provided bet-
ter removal of TDS. Based on the treated water quality, it was
proposed that the membrane treated water can be beneficially
used in lives stock and crop watering. The fouling behaviour
of the three membranes based on its surface roughness and
hydrophilicity was compared. It was inferred that the smooth
and more hydrophilic membranes are less susceptible for the
fouling.
The importance of the pre-treatment for the effective
PW treatment can be understood from the report of Zhao
et al. (2014) The authors studied the real PW for the removal
of hardness, COD and turbidity by electrocoagulation as
pretreatment. The pre-treated produced water was directly
subjected to RO process. The pH, current density and time
were varied and optimized for the effective pretreatment.
At optimum condition of pH 7.36, 85.81% hardness, 66.64%
of COD and 93.80% of turbidity was removed. Thus the pre-
treated water showed steady permeation flux of 87.83%, ca:
22 L/m2 h, whereas the raw produced water without pre-
treatment showed the flux decline of more than 80% in less
than 100 h of RO membrane operation. From the flux decline by
fouling, it is imperative to carry out the effective pretreatment
in order to achieve effective membrane treatment.
196 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
In a pilot scale study, the use of RO membrane in combi-
nation with three different types pretreatment was reported
for the treatment of PW from Coal-bed methane(CBM) (Qian
et al., 2012). A stages of pretreatment namely Manganese sand
filter, sand filter and UF membrane were used before subject-
ing the PW to the RO membrane treatment. The Manganese
sand filter alone removed the Fe and Mn ∼ 100%, which was
originally present at 0.22 and 0.04 ppm in the PW. The Man-
ganese sand filter reduced the turbidity up to 97%. Further use
of sand filter, UF and finally RO filtration combination showed
complete removal of turbidity and F− . Further 98% of TDS,
81.6% of COD, 85.4% of NH3 N and 96.99% of Cl− removal was
reported for the combined treatment. The authors used theo-
retical model to predict the performance membrane and the
prediction agreed well with the experimental result. Interest-
ingly, it was claimed that the treated water is well within the
Chinese drinking water standard.
Salahi et al. (2012b) reported the purification of biologically
treated Tehran refinery waste water using the RO membrane
process. The collected the refinery waste water was sub-
jected directly to the RO membrane as final treatment. The
effect of transmembrane pressure (TMP 8, 15 and 20 bar),
cross flow velocity (CFV 0.5, 1 and 1.5 m/s), pH (4, 7, 10) and
feed temperature (27.5, 37.5 and 50 ◦ C) were studied at three
different levels. Except pH, all three parameters showed pos-
itive effect on the permeation flux with small decrease in
the rejection. The effect of pH on the permeate flux was
reported as complex. The characteristics of Tehran refinery
waste water is given in Table 1. The biologically pre-treated
refinery waste water had the following characteristic (in mg/L):
TSS–4, TDS—1953, O&G—7.2, COD—160, BOD5 —86, TOC—48
and turbidity 1.1 NTU. The RO membrane process had the
permeation flux of 50 LMH with permeate water quality as fol-
lows: TSS—0, TDS—253, O&G—1, COD—8, BOD5 —4, TOC—4.9
and turbidity 0.2 NTU. From the permeate water quality, it is
clear that the treated water quality is reasonable good. How-
ever as suggested in the report, the long term membrane
fouling behaviour needs to be studied further to evaluate the
practical feasibility of direct use of RO membrane from the
biologically treated refinery waste water. In an earlier study,
the research group used the UF as pre-treatment for the same
refinery waste water (Salahi et al., 2012c). In this case also the
effect of transmembrane pressure (TMP—1.5, 3, 4.5 bar), cross
flow velocity (CFV—0.25, 0.75, 1.25 m/s), pH (4, 7, 10) and feed
temperature (25, 37.5, 50 ◦ C) were studied at three different
levels for the optimization of UF process. Using the optimum
TMP (3 bar), CFV (1 m/s), pH (9) at the temperature of 40 ◦ C, the
permeate from the UF membrane was used as feed to the RO
membrane. The RO permeate showed 100%, 98%, 98%, 95%,
and 100% reduction for O&G, TOC, TDS and turbidity respec-
tively. It was reported that the permeate water quality from
UF and RO membrane treatment was better than the cooling
water tower used in the Tehran refinery.
In a large scale demonstration of use of membrane tech-
nology, the use of biological treatment in combination with
UF and RO membrane process was reported for the Sinopec
Yanshan refinery plant waste water (Wang et al., 2012). The
treatment plant was run more than four years with the water
treatment capacity above 100 m3 /h. The biological treatment
comprised of biological aerated filter (BAF), coagulation set-
tlement, fibrous filtration followed by activated carbon filter.
After the biological treatment, the refinery water was sub-
jected to UF membrane which showed 34% TOC removal and
the COD removal was in the range of 0–60% depending on
the molecular weights of the organics. High molecular weight
organics were effectively removed by coagulation and multi-
media filter, but low molecular weight organics easily passed
through UF membrane. The first module of three RO mem-
brane train showed pressure drop more than the second or
third module due to the organic fouling. After more than
30 chemical cleanings, each per month, the RO membrane
showed the rejection of more than 97% with water recovery of
80%. It was reported that the membrane treated refinery waste
water met the entire parameters for the reuse in the plant. This
is one of the few report deals with reasonably large quantity
of real waste water with long period of plant operation. Most
importantly, the treated water was very much within reuse
quality.
In summary, the use of NF can be very selective for the oil &
gas industries. Many oil fields use fresh water for the pressure
maintenance and to pump out the crude. For this purpose,
the water quality may not necessarily high. The fresh water
should be free of corrosion, pore blocking and scale forming
ions. To meet these criteria, the use of sea water with free of
above mentioned ions is the most cost effective way, especially
in off-shore oil field. The NF membrane can effectively remove
all the multivalent ions which cause the problems. Therefore
the use of NF membrane for the sea water softening for poly-
mer flooding for oil recovery was reported (Su et al., 2012). The
UF membrane was used as pretreatment and four commercial
NF membranes, namely NF90-2540, NF200-2540, NF270-2540
and DL-2540, were used for the softening of sea water. All the
four NF membranes showed more than 95% removal of SO4 2− .
However the divalent cations, Ca2+ and Mg2+ , removal varied
for each NF membrane. The NF90-2450 showed the effective
divalent ions, both cation and anion, removal of more than
99%. Using the NF90-2540 treated sea water, the authors stud-
ied the polymer viscosity and stability which was found to
be stable for long time at oil-well formation temperature. It
was further reported that the core sensitivity and compatibil-
ity test were used to prove the NF permeate water quality for
the flooding of oil well. From the result, it is possible to use the
NF membrane for the treatment of sea water for the use in oil
& gas industry. The same methodology can be easily adapted
with appropriate modification for the PW treatment and thus
treated water can be beneficially reused by the oil & gas indus-
try. It is important to note that the operating pressure of NF
membrane is significantly less, therefore less operation cost,
compared to the RO membrane.
3.4. Thermal desalination and forward osmosis (FO)
for the PW treatment
One of the earliest desalination technology adopted in large
scale is thermal desalination. Two well-known desalination
methods which are still in use (mostly in the Middle East
regions) are multistage flash distillation (MSF) and multiple
effect distillation (MED). Though these two distillation meth-
ods are energy intensive, it scores two important points over
membrane technology namely capability of very large scale
installation and long period of plant operation, in some cases
over 20 years. However these thermal technologies are in the
retreat due to high cost and advent of more energy efficient
membrane technologies. Recently a pilot scale study for the
coal seam gas (CSG) produced water treatment by combining
ultrafiltration, reverse osmosis and multi-effect distillation
was reported (Nghiem et al., 2015). The CSG PW water was
considered as brackish with moderate TDS. The UF membrane
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
was operated as dead-end mode with repeated 40 s of air
scouring and 30 s of back-flushing for each 17 min production
cycle. About 76% of water recovered in the UF and permeate
was sent as feed for the RO filtration. The RO system was oper-
ated to recover about 76% of feed water and the RO concentrate
was send to MED for the further water recovery. The MED pro-
cess recovered about 80% of the feed from RO retentate. Overall
an impressive 95.2% of purified water recovered from the
initial PW. Since the RO and MED permeate waters are highly
pure, it was suggested to use as diluent for the UF permeate
to the beneficial use in order to reduce energy consumption.
The membrane distillation (MD) is another important ther-
mal desalination technique which acquired more attention by
the research community in recent times (Wang and Chung,
2015; Prince et al., 2014; Francis et al., 2014; Maab et al., 2012,
2013; Leonard et al., 2015; Alkhudhiri et al., 2012; Al-Asheha
et al., 2006; Banat and Qtaishat, 2005; Banat et al., 2003). The
energy requirement, i.e. the heating requirement, is consid-
erably less for the MD and therefore it can be economically
competitive in some special cases. The membrane technolo-
gies discussed so far for the PW treatment use the pressure
difference as incentive to separate the constituents by using
selective barrier layers. On the other hand, the MD operates
based on the thermal difference between the feed and per-
meate which is separated by hydrophobic membrane layer.
The MD can be a crucial separation component for the PW
treatment because of high water recovery. All the pressure
driven membranes separate the pollutants and at the end of
the process it has highly “concentrated brine” as retentate.
It is assumed volume of waste water is reduced to “man-
ageable volume”. The majority of the literature reports deals
with water recovery from PW and no meaningful sugges-
tions were made about the “concentrated brine” retentate.
Therefore in reality, what all the pressure driven membrane
separation does is the recovery of water as permeate and very
highly concentrated PW as retentate. The disposal or mean-
ingful use of this concentrated retentate of PW from pressure
driven membranes will be the final frontier in the treatment
of PW. By its intrinsic nature, these saturated or “concen-
trated brine” is beyond the practical operating pressure of NF
or RO membranes due to very high osmatic pressure. Since MD
operation is not based on pressure driven but only tempera-
ture difference between feed and permeate side is the driving
force, the final treatment of the “concentrated brine” can be
effectively done using MD process. It has been reported that
the effect of salt concentration in the feed water is negligi-
ble on the MD process compared to pressure driven process
like RO or NF. Only 5% flux decline was observed in the MD
when the TDS of the feed water was increased from 35,000
to 75,000 mg/L (Al-Obaidani et al., 2008). Therefore MD can be
used as last stage recovery process and it has been reported
that near complete removal of salt is possible in the MD
process (Maab et al., 2012, 2013). However the presence of
organic compounds in the feed water is problematic for the
successful MD process. The organic compounds, both polar
and non-polar in nature, first, affect the wetting properties of
the membrane. The non-wetting phenomenon (contact angle)
is one of the crucial requirements for the effective MD pro-
cess and the presence of organic compounds increases the
wetting of the feed water. Additionally, if the organic com-
pounds are highly volatile, then these low boiling compounds
easily passes though the membrane during the distillation
process. Obviously, the gasses also easily pass through the
MD. Therefore it is imperative that the feed water for the MD
197
process should contain as much as less organic compounds,
particularly the low boiling constituents, for the successful MD
process. Otherwise permeate might contain the organic and
dissolved gases which might require further post treatment
before the beneficial use.
There are four different modes of operation in the MD
process which are: (1) Direct contact MD (DCMD), (2) Air gap
MD (AGMD), (3) Sweep gas MD (SGMD) and (4) Vacuum MD
(VMD). The operating principles of these methods are pictori-
ally explained in Fig. 4. As the name implies, in the DCMD
the feed and permeate is in direct contact with the mem-
brane (Fig. 4a). In VMD, instead of circulating cool water in
the permeate side, a vacuum has been applied and the result-
ing vapours are condensed (Fig. 4b). In AGMD, air gap is created
between membrane and condensing surface (Fig. 4c). In SGMD,
cold air is circulated on the permeate side in order to collect
the vapours (Fig. 4d). In all these four variants, only the perme-
ate collection side was varied according to the condensation
methods. The feed side remains similar for all four processes
which is hot circulating water. The permeate side should be
at least 20 ◦ C cooler than feed side for the practically effective
distillation and it can be liquid, air or vacuum to collect the
water vapours from the feed side.
The use of DCMD for treatment of real oilfield produced
water was reported by Macedonio et al. (2014) The PW water
was pre-treated using microfiltration and activated carbon
in order to remove the oil, suspended solids and H2 S. The
pre-treated PW was subjected for the DCMD using four dif-
ferent membranes namely two lab made PVDF membranes
and two commercially available PP membranes. The perme-
ate flux of 4–9 L/m2 h was obtained with feed temperature of
50 ◦ C with permeate circulation at 25 ◦ C. After optimizing the
operating conditions, the TDS removal of >99.5% and >90%
of total carbon removal was obtained. The cost of the DCMD
distillation process alone has been calculated as 0.72 US$/m3
for the PW feed temperature of 50 ◦ C and 1.28 US$/m3 if the
PW feed starting temperature is 20 ◦ C. Singh and Sirkar (2012)
used DCMD for the treatment of simulated produced water
of steam assisted gravity drainage (SAGD). In this study, for
the first time, the authors subjected the feed hot water up
to 130 ◦ C temperature for the circulation. The performance
of commercial PTFE membrane (W. L. Gore and Associates),
was studied in the presence of 45 ppm phenol, 45 ppm cresol,
10 ppm naphthenic acid 3000 ppm NaCl. Even at high tempera-
ture of 130 ◦ C, there was no leakage in the membrane and the
highest permeation flux of 195 kg/m2 h was observed. About
∼5 ppm of phenol and cresol and around 2 ppm of naphthenic
acid passed through the PTFE membrane and it was attributed
due to the fact that the boiling of these compounds are in the
range of 165–212 ◦ C.
Hilal et al. reported the study of three commercially avail-
able membranes for the treatment of ARAMCO gulf PW by
AGMD (Alkhudhiri et al., 2013). It was reported that as the
pore size of the membrane increased, the water flux also
increased. As with typical MD process, the permeate flux
increased directly proportional to the feed temperature and
inversely proportional to the coolant temperature. It has been
observed that exponential increase in the permeate flux by ris-
ing the feed temperature. It also has been reported that three
membranes with different pore size showed no effect on the
energy consumption. However, the water flux is low and it has
be increased at least an order of magnitude for economically
viable. One of the main advantages of MD process is the capa-
bility to treat the high saline PW or concentrated PW. Utilizing
198 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
Fig. 4 – Four different modes of membrane distillations: (a) DCMD, (b) VMD, (c) AGMD, (d) SGMD.
this benefit, Duong et al. (2015) reported a pilot study to treat
the RO brine from coal seam gas PW by spiral-wound air gap
membrane distillation. First UF&RO membranes were used to
recover 75% fresh water from the raw CSG produced water.
The concentrated brine was subjected to AGMD. High packing
density of spiral-wound module was used to overcome the
low permeate nature of membrane. By adding anti-scalant,
a stable 80% recovery was achieved in AGMD. So far, the
reported commercial membranes showed very low permeate
flux, around 20 LMH or less. At this rate of water flux, it might
be tall order to compete with the other purification meth-
ods. However the lab made membranes for distillation showed
much promising results and it needs to be upscale to evaluate
the comparative advantages of MD process.
Forward osmosis (FO) is another membrane technology
which is getting more attention recently for the treatment of
PW (Cath et al., 2006). In the RO process, pressure has to be
applied to overcome the osmotic pressure of the feed to force
the water molecule to pass through the membrane. In the
FO, the permeate side contains very high osmotic solution as
draw solution and the separation occurs spontaneously due to
osmotic pressure difference between feed and draw solution.
In FO, the draw solution should have more osmatic pressure
than the feed and the feed side osmotic pressure (from the PW)
should be less compared to the draw solution for the efficient
separation. The more osmotic pressure difference between
feed and draw solution the more efficient separation occurs.
It should be noted that in some cases the PW itself contains
high to very high TDS and therefore the draw solution should
have even more salt content than the feed from PW. Various
draw solutions starting from conventional salts like NaCl &
MgSO4 to novel chemical compounds including thermolytic
salts and switchable polarity solvents, were studied for the
FO (Shaffer et al., 2013; Cath et al., 2006). The FO operates
at low pressure, normally at ambient condition, therefore
the irreversible fouling should be less compared to the RO
process. Since the fouling formation likely to be less compact,
it was suggested that the removal of fouling layers in the FO
processes should be easier than high pressure membranes
like RO (Shaffer et al., 2013). During the FO process, the draw
solution becomes diluted as the osmosis occurs and it should
be regenerated, i.e. should be concentrated, in order to have
practical viability. Since the draw solution contains very high
TDS and in many cases beyond the operating pressure of RO
membrane, the MD might be the viable option for the draw
solution regeneration. It has been proposed that the combi-
nation of FO and MD might be one of the potential options
for the PW treatment in the future (Shaffer et al., 2013).
Indeed the combination of FO and MD has been reported by
Zhang et al. (2014b) for the water recovery from synthetic PW.
The lab made CTA hollow fiber TFC membrane for the FO and
PVDF for the MD were used for the separation study. 2 M NaCl
was used as draw-solution in the FO process and synthetic
oily water (4000 ppm) with 1000 ppm acetic acid as feed. The
FO membrane showed the fouling initially and after 24 h the
flux has been stabilized. The FO showed >99.9% oil and ∼80%
acetic acid rejection. The MD showed 99.99% salt rejection and
47% acetic acid at operating temperature of 60 ◦ C.
These are the some promising results utilizing the com-
bined advantage of FO and MD. Since PW is considered as
one of the “difficult to treat” waste water, the novel combi-
nation of FO and MD will be further studied in future for the
practical application of PW treatment. In the FO process, high
flux fouling resistance membrane is the need of hour. In the
MD process, the existing commercial membranes are having
uncompetitive low water permeation flux and here also the
permeation flux has to be increased at least an order of mag-
nitude in order to be economically viable process.
In summary, for the successful treatment of PW using
membrane technology, apart from the effectiveness of each
membrane process, the capital and operating cost might be a
deciding factor. For example, a typical operating cost of sea-
water RO membrane replacement is less than 25% whereas
it can be 30–50% for the UF membrane due to high fouling
in the UF process (Baker, 2004). The microfiltration process is
well developed and mature technology with competitive price.
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202 199

Majority of the fluorine based MF membranes like PVDF, PTFE

were tested for MD process also, but the permeation flux is
very low. Therefore novel hydrophobic membranes need to be
developed for the cost effective MD process. For the FO, the
existing thin film composite (TFC) membranes are used and
new membrane development might speedup for the success-
ful upscaling for the PW treatment. Apart from the production
cost of membranes, the operational condition and module
design also play significant role in the final cost. For example
high pressure modules are more expensive than low pres-
sure module and hollow fiber modules are much cheaper than
spiral-wound module. The following cost estimation were
reported for hollow fiber, capillary fiber, spiral-wound, plate-
and-frame and Tubular modules (in US$/m2 ): 5–20, 10–50,
5–100, 50–200 and 50–200 (Baker, 2004). The real cost of the
membrane might be decided by the demand-supply scenario
too. At some point, the spiral-wound RO membranes were
cheaper than the UF module due to large volume production
of RO membrane compared to the few UF membrane manu-
facturer (Baker, 2004). Therefore the true cost of membrane
is vibrant depending on market condition, nature of module
Fig. 5 – The schematic diagram of overall PW treatment
and operating conditions etc. The most important parameter
processes and its final destination of the treated PW.
that decides the effectiveness and longevity of the membrane
process is the feed quality of PW. Therefore, to select the mem-
brane process, it is imperative to keep all these conditions in reports dealing the real and synthetic PW should be carefully
mind to select the membrane and the operating condition. evaluated. As demonstrated in few examples (Abbasi et al.,
2010; Ebrahimi et al., 2009), the effectiveness of membrane
process for the real and synthetic PW is vast and the real
4. Summary and outlook PW is invariably makes membrane less effective due to foul-
ing. Therefore, the choice of membrane technology has to be
The complexity of the produced and process water from oil carefully chosen based on the real feed water quality. Since
& gas industry makes it one of the toughest waste water to produced and process water quality change according to the
manage within acceptable level of contamination at afford- geology of the oil & gas well and also with the age of well,
able cost. The membrane technology is clearly one of the main each industrial need of membrane has to be optimized for the
contenders to solve the intrinsic problem of produced waste effective treatment of PW. The real bottleneck of PW treat-
water treatment. Because of high complexity of PW, the future ment is the lack of large scale plant literature report. There are
research development will have to be concurrently developed very few literature reports that discuss about the large scale
with the advanced pretreatment methods and the success treatment plants (Alzahrani et al., 2013a,b; Wang et al., 2012).
of the membrane process will entirely depend on the initial The three large scale examples using RO as final stage treat-
quality of the PW feed after pretreatment. The overall PW ment are Chennai Petroleum Company Limited (430,000 L/h),
treatment processes and its final destination of the treated San Ardo water reclamation project in California (50,000 bbl/d)
PW is given in Fig. 5. It is clear from the reported literature and oilfield in Wellington, Colorado (93,600 g/d) and even the
that the potable water quality can be achieved only by employ- scarcely available report does not provide complete details to
ing RO membranes. If the water is used for drinking purpose, understand the process and thereby improving the technol-
then special post treatment might be needed to obtain the ogy. The operational data from of existing of large scale plants
acceptable level of water quality. For example, presence of will help to understand the current state of technology bet-
high level NORMs or boron will make it unacceptable for the ter and thereby developing more adaptive smart membranes
drinking purpose. Depending on the pre-treated PW quality, in the future. The results reported by Zhu et al. (2014) and
NF membranes can be useful to achieve water quality for reuse hydrophilized PSf membranes (Yuan et al., 2014) showed that
in the oil industry itself, irrigation and livestock. Though the it is possible to reduce the fouling by designing the membrane
RO membrane provides better water quality than NF mem- appropriately. At the end, the cost of PW treatment will be
branes, the NF membrane can be more cost effect for recycle within acceptable level in order to make the oil industry waste
and reuse in the oil & gas industry. Recently the FO mem- water truly usable resource instead of polluting water.
branes in combination with thermal distillation, particularly
MD, are getting more attention. The combination of FO and Acknowledgements
MD technology needs further study for the practical suitability
for the PW treatment. Finally, all the membrane technologies The authors are grateful to The Petroleum Institute, Abu Dhabi
used for the PW treatment result highly “concentrated brines” for funding the project.
as final effluent at “manageable volume”. Very little study
has been focused on this concentrated brine and the recent References
development of MD will be an ideal candidate for solving this
problem. However the issue revolving around MD process itself Abadi, S.R.H., Sebzari, M.R., Hemati, M., Rekabdar, F.,
should be first resolved i.e. High cost (energy), scale forming, Mohammadi, T., 2011. Ceramic membrane performance in
and heat loss by thermal conduction etc. Finally, the literature microfiltration of oily wastewater. Desalination 265, 222–228.
200 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
Abbasi, M., Mirfendereski, M., Nikbakht, M., Golshenas, M.,
Mohammadi, T., 2010. Performance study of mullite and
mullite–alumina ceramic MF membranes for oily wastewaters
treatment. Desalination 259, 169–178.
Al Zarooni, M., Elshorbagy, W., 2006. Characterization and
assessment of Al Ruwais refinery wastewater. J. Hazard.
Mater. 136, 398–405.
Al-Asheha, F.B.S., Qtaishat, M., Al-Khateeb, M., 2006.
Concentration of sucrose solutions via vacuum membrane
distillation. Desalination 195, 60–68.
Alkhudhiri, A., Darwish, N., Hilal, N., 2012. Membrane distillation:
a comprehensive review. Desalination 287, 2–18.
Alkhudhiri, A., Darwish, N., Hilal, N., 2013. Produced water
treatment: application of air gap membrane distillation.
Desalination 309, 46–51.
Al-Malack, M.H., Siddiqui, M., 2013. Treatment of synthetic
petroleum refinery wastewater in a continuous
electro-oxidation process. Desalin. Water Treat. 51, 6580–6591.
Al-Obaidani, S., Curcio, E., Macedonio, F., Di Profio, G., Al-Hinai,
H., Drioli, E., 2008. Potential of membrane distillation in
seawater desalination: thermal efficiency, sensitivity study
and cost estimation. J. Membr. Sci. 323, 85–98.
Alpatova, A., Kim, E.-S., Dong, S., Sun, N., Chelme-Ayala, P.,
Gamal El-Din, M., 2014. Treatment of oil sands
process-affected water with ceramic ultrafiltration
membrane: effects of operating conditions on membrane
performance. Sep. Purif. Technol. 122, 170–182.
Alzahrani, S., Mohammad, A.W., 2014. Challenges and trends in
membrane technology implementation for produced water
treatment: a review. J. Water Process Eng. 4,
107–133.
Alzahrani, S., Mohammad, A.W., Hilal, N., Abdullah, P., Jaafar, O.,
2013a. Comparative study of NF and RO membranes in the
treatment of produced water—Part I: Assessing water quality.
Desalination 315, 18–26.
Alzahrani, S., Mohammad, A.W., Hilal, N., Abdullah, P., Jaafar, O.,
2013b. Comparative study of NF and RO membranes in the
treatment of produced water II: Toxicity removal efficiency.
Desalination 315, 27–32.
Ayse, A.M.M., 2009. Asatekin oil industry wastewater treatment
with fouling resistant membranes containing amphiphilic
comb copolymers. Environ. Sci. Technol. 43, 4487–4492.
Baker, R.W. (Ed.), 2004. Membrane Technolgoy and Applicatoins. ,
second ed. John Wiley & Sons, West Sussex, England.
Banat, S.A.-A.F., Qtaishat, M., 2005. Treatment of waters colored
with methylene blue dye by vacuum membrane distillation.
Desalination 174, 87–96.
Banat, F., Al-Rub, F.A., Bani-Melhem, K., 2003. Desalination by
vacuum membrane distillation: sensitivity analysis. Sep.
Purif. Technol. 33, 75–87.
Buthiyappan, A., Abdul Aziz Abdul, R., Wan Daud Wan Mohd, A.,
2015. Degradation performance and cost implication of
UV-integrated advanced oxidation processes for wastewater
treatments. Rev. Chem. Eng. 31, 263.
Çakmakce, M., Kayaalp, N., Koyuncu, I., 2008. Desalination of
produced water from oil production fields by membrane
processes. Desalination 222, 176–186.
Campos, R.M.H.B.J.C., Oliveira Filho, A.M., Nobreg, R., Sant’Anna
Jr., G.L., 2002. Oilfield wastewater treatment by combined
microfiltration and biological processes. Water Res. 36, 95–104.
Cath, T., Childress, A., Elimelech, M., 2006. Forward osmosis:
principles, applications, and recent developments. J. Membr.
Sci. 281, 70–87.
Chakrabarty, B., Ghoshal, A.K., Purkait, M.K., 2008. Ultrafiltration
of stable oil-in-water emulsion by polysulfone membrane. J.
Membr. Sci. 325, 427–437.
Chaturabul, S., Srirachat, W., Wannachod, T., Ramakul, P.,
Pancharoen, U., Kheawhom, S., 2015. Separation of mercury(II)
from petroleum produced water via hollow fiber supported
liquid membrane and mass transfer modeling. Chem. Eng. J.
265, 34–46.
Coday, B.D., Xu, P., Beaudry, E.G., Herron, J., Lampi, K., Hancock,
N.T., Cath, T.Y., 2014. The sweet spot of forward osmosis:
treatment of produced water, drilling wastewater, and other
complex and difficult liquid streams. Desalination 333, 23–35.
Coelho, A., Castro, A.V., Dezotti, M., Sant’Anna Jr., G.L., 2006.
Treatment of petroleum refinery sourwater by advanced
oxidation processes. J. Hazard. Mater. 137, 178–184.
Cui, J., Zhang, X., Liu, H., Liu, S., Yeung, K.L., 2008. Preparation
and application of zeolite/ceramic microfiltration membranes
for treatment of oil contaminated water. J. Membr. Sci. 325,
420–426.
da Silva, S.S., Chiavone-Filho, O., de Barros Neto, E.L.,
Nascimento, C.A., 2012. Integration of processes induced air
flotation and photo-Fenton for treatment of residual waters
contaminated with xylene. J. Hazard. Mater. 199–200, 151–157.
da Silva, S.S., Chiavone-Filho, O., de Barros Neto, E.L., Foletto, E.L.,
2015. Oil removal from produced water by conjugation of
flotation and photo-Fenton processes. J. Environ. Manage. 147,
257–263.
Dallbauman, L., Sirivedhin, T., 2005. Reclamation of produced
water for beneficial use. Sep. Sci. Technol. 40, 185–200.
Deng, S., Yu, G., Jiang, Z., Zhang, R., Ting, Y.P., 2005.
Destabilization of oil droplets in produced water from ASP
flooding. Colloids Surf., A: Physicochem. Eng. Aspect 252,
113–119.
Diya’uddeen, B.H., Daud, W.M.A.W., Abdul Aziz, A.R., 2011.
Treatment technologies for petroleum refinery effluents: a
review. Process Saf. Environ. Prot. 89, 95–105.
dos Santos, E.V., Bezerra Rocha, J.H., de Araujo, D.M., de Moura,
D.C., Martinez-Huitle, C.A., 2014. Decontamination of
produced water containing petroleum hydrocarbons by
electrochemical methods: a minireview. Environ. Sci. Pollut.
Res. Int. 21, 8432–8441.
Duong, H.C., Chivas, A.R., Nelemans, B., Duke, M., Gray, S., Cath,
T.Y., Nghiem, L.D., 2015. Treatment of RO brine from CSG
produced water by spiral-wound air gap membrane
distillation—a pilot study. Desalination 366, 121–129.
Ebrahimi, M., Ashaghi, K.S., Engel, L., Willershausen, D., Mund, P.,
Bolduan, P., Czermak, P., 2009. Characterization and
application of different ceramic membranes for the oil-field
produced water treatment. Desalination 245, 533–540.
Ebrahimi, M., Willershausen, D., Ashaghi, K.S., Engel, L., Placido,
L., Mund, P., Bolduan, P., Czermak, P., 2010. Investigations on
the use of different ceramic membranes for efficient oil-field
produced water treatment. Desalination 250, 991–996.
Ebrahimi, M., Kovacs, Z., Schneider, M., Mund, P., Bolduan, P.,
Czermak, P., 2012. Multistage filtration process for efficient
treatment of oil-field produced water using ceramic
membranes. Desalin. Water Treat. 42, 17–23.
Ebrahimi, M., Schmitz, O., Kerker, S., Liebermann, F., Czermak, P.,
2013. Dynamic cross-flow filtration of oilfield produced water
by rotating ceramic filter discs. Desalin. Water Treat. 51,
1762–1768.
El-Naas, M.H., Alhaija, M.A., Al-Zuhair, S., 2014. Evaluation of a
three-step process for the treatment of petroleum refinery
wastewater. J. Environ. Chem. Eng. 2, 56–62.
Esmaeilirad, N., Carlson, K., Omur Ozbek, P., 2015. Influence of
softening sequencing on electrocoagulation treatment of
produced water. J. Hazard. Mater. 283, 721–729.
Fakhru’l-Razi, A., Pendashteh, A., Abdullah, L.C., Biak, D.R.,
Madaeni, S.S., Abidin, Z.Z., 2009. Review of technologies for oil
and gas produced water treatment. J. Hazard. Mater. 170,
530–551.
Francis, L., Ghaffour, N., Alsaadi, A.S., Nunes, S.P., Amy, G.L., 2014.
Performance evaluation of the DCMD desalination process
under bench scale and large scale module operating
conditions. J. Membr. Sci. 455, 103–112.
Gondal, M.A., Sadullah, M.S., Dastageer, M.A., McKinley, G.H.,
Panchanathan, D., Varanasi, K.K., 2014. Study of factors
governing oil-water separation process using TiO(2) films
prepared by spray deposition of nanoparticle dispersions. ACS
Appl. Mater. Interfaces 6, 13422–13429.
Hasan, D.u.B., Abdul Aziz, A.R., Daud, W.M.A.W., 2012. Oxidative
mineralisation of petroleum refinery effluent using
Fenton-like process. Chem. Eng. Res. Des. 90, 298–307.
 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
Hu, X., Yu, Y., Zhou, J., Wang, Y., Liang, J., Zhang, X., Chang, Q.,
Song, L., 2015. The improved oil/water separation
performance of graphene oxide modified Al2 O3 microfiltration
membrane. J. Membr. Sci. 476, 200–204.
Igunnu, E.T., Chen, G.Z., 2012. Produced water treatment
technologies. Int. J. Low-Carbon Technol. 9, 157–177.
Jeffrey Mueller, Y.C., Robert, H., Davis, 1997. Crossflow
microfiltration of oily water. J. Membr. Sci. 129, 221–235.
Jiang, Q., Rentschler, J., Perrone, R., Liu, K., 2013. Application of
ceramic membrane and ion-exchange for the treatment of the
flowback water from Marcellus shale gas production. J.
Membr. Sci. 431, 55–61.
Karimnezhad, H., Rajabi, L., Salehi, E., Derakhshan, A.A., Azimi,
S., 2014. Novel nanocomposite Kevlar fabric membranes:
fabrication characterization, and performance in oil/water
separation. Appl. Surf. Sci. 293, 275–286.
Kim, E.-S., Liu, Y., Gamal El-Din, M., 2011. The effects of
pretreatment on nanofiltration and reverse osmosis
membrane filtration for desalination of oil sands
process-affected water. Sep. Purif. Technol. 81, 418–428.
Kose, B., Ozgun, H., Ersahin, M.E., Dizge, N., Koseoglu-Imer, D.Y.,
Atay, B., Kaya, R., Altınbas, M., Sayılı, S., Hoshan, P., Atay, D.,
Eren, E., Kinaci, C., Koyuncu, I., 2012. Performance evaluation
of a submerged membrane bioreactor for the treatment of
brackish oil and natural gas field produced water.
Desalination 285, 295–300.
Leonard, Y.C.l., Tijing, D., June-SeokChoi, Lee, Sangho, Kim,
Seung-Hyun, Shon, Ho Kyong, 2015. Fouling and its control in
membrane distillation—a review. J. Membr. Sci. 475,
215–244.
Li, L., Lee, R., 2009. Purification of produced water by ceramic
membranes: material screening, process design and
economics. Sep. Sci. Technol. 44, 3455–3484.
Li, X., Hu, D., Huang, K., Yang, C., 2014. Hierarchical rough
surfaces formed by LBL self-assembly for oil–water
separation. J. Mater. Chem. A 2, 11830.
Loganathan, K., Chelme-Ayala, P., El-Din, M.G., 2015. Effects of
different pretreatments on the performance of ceramic
ultrafiltration membrane during the treatment of oil sands
tailings pond recycle water: a pilot-scale study. J. Environ.
Manage. 151, 540–549.
Maab, H., Francis, L., Al-saadi, A., Aubry, C., Ghaffour, N., Amy, G.,
Nunes, S.P., 2012. Synthesis and fabrication of nanostructured
hydrophobic polyazole membranes for low-energy water
recovery. J. Membr. Sci. 423–424, 11–19.
Maab, H., Al Saadi, A., Francis, L., Livazovic, S., Ghafour, N., Amy,
G.L., Nunes, S.P., 2013. Polyazole hollow fiber membranes for
direct contact membrane distillation. Ind. Eng. Chem. Res. 52,
10425–10429.
Macedonio, F., Ali, A., Poerio, T., El-Sayed, E., Drioli, E.,
Abdel-Jawad, M., 2014. Direct contact membrane distillation
for treatment of oilfield produced water. Sep. Purif. Technol.
126, 69–81.
Miller, D.J., Huang, X., Li, H., Kasemset, S., Lee, A., Agnihotri, D.,
Hayes, T., Paul, D.R., Freeman, B.D., 2013. Fouling-resistant
membranes for the treatment of flowback water from
hydraulic shale fracturing: a pilot study. J. Membr. Sci. 437,
265–275.
Mondal, S., Wickramasinghe, S.R., 2008. Produced water
treatment by nanofiltration and reverse osmosis membranes.
J. Membr. Sci. 322, 162–170.
Mondal, S., Hsiao, C.l., Ranil Wickramasinghe, S., 2008.
Nanofiltration/reverse osmosis for treatment of coproduced
waters. Environ. Prog. 27, 173–179.
Moslehyani, A., Ismail, A.F., Othman, M.H.D., Matsuura, T., 2015.
Design and performance study of hybrid photocatalytic
reactor-PVDF/MWCNT nanocomposite membrane system for
treatment of petroleum refinery wastewater. Desalination
363, 99–111.
Motta, A., Borges, C., Esquerre, K., Kiperstok, A., 2014. Oil
Produced Water treatment for oil removal by an integration of
coalescer bed and microfiltration membrane processes. J.
Membr. Sci. 469, 371–378.
201
Mulder, M., 1996. Basic Principles of Membrane Technology.
Kluwer academic Publishers, Dordrecht, Netherlands.
Nghiem, L.D., Elters, C., Simon, A., Tatsuya, T., Price, W., 2015.
Coal seam gas produced water treatment by ultrafiltration,
reverse osmosis and multi-effect distillation: a pilot study.
Sep. Purif. Technol. 146, 94–100.
Olsson, O., Weichgrebe, D., Rosenwinkel, K.-H., 2013. Hydraulic
fracturing wastewater in Germany: composition, treatment,
concerns. Environ. Earth Sci. 70, 3895–3906.
Padaki, M., Surya Murali, R., Abdullah, M.S., Misdan, N.,
Moslehyani, A., Kassim, M.A., Hilal, N., Ismail, A.F., 2015.
Membrane technology enhancement in oil–water separation.
A review. Desalination 357, 197–207.
Palmer, L.L., Beyer, A.H., Stock, J., 1981. Biological oxidation of
dissolved compounds in oilfield produced water by a field
pilot biodisk. J. Pet. Technol. 33, 1136.
Pancharoen, U., Poonkum, W., Lothongkum, A.W., 2009.
Treatment of arsenic ions from produced water through
hollow fiber supported liquid membrane. J. Alloys Compd.
482, 328–334.
Paolo Roccaro, G.L., Federico, G.A., Vagliasindi, 2014. Optimization
of the coagulation process to remove total suspended solids
(TSS) from produced water. Chem. Eng. Trans. 39,
115–120.
Pendashteh, A.R., Abdullah, L.C., Fakhru’l-Razi, A., Madaeni, S.S.,
Zainal Abidin, Z., Awang Biak, D.R., 2012. Evaluation of
membrane bioreactor for hypersaline oily wastewater
treatment. Process Saf. Environ. Prot. 90, 45–55.
Petroleum, 2010. Petroleum Refining Water/Wastewater use and
Management. IPIECA Operation Best Practice Series, IPIECA,
London, UK.
Prince, J.A., Rana, D., Matsuura, T., Ayyanar, N.,
Shanmugasundaram, T.S., Singh, G., 2014. Nanofiber based
triple layer hydro-philic/-phobic membrane—a solution for
pore wetting in membrane distillation. Sci. Rep. 4,
6949.
Qian, Z., Liu, X., Yu, Z., Zhang, H., JÜ, Y., 2012. A pilot-scale
demonstration of reverse osmosis unit for treatment of
coal-bed methane co-produced water and its modeling. Chin.
J. Chem. Eng. 20, 302–311.
Qiao, X., Zhang, Z., Yu, J., Ye, X., 2008. Performance characteristics
of a hybrid membrane pilot-scale plant for oilfield-produced
wastewater. Desalination 225, 113–122.
Salahi, A., Mohammadi, T., Rahmat Pour, A., Rekabdar, F., 2012a.
Oily wastewater treatment using ultrafiltration. Desalin.
Water Treat. 6, 289–298.
Salahi, A., Mohammadi, T., Nikbakht, M., Golshenas, M., Noshadi,
I., 2012b. Purification of biologically treated Tehran refinery
oily wastewater using reverse osmosis. Desalin. Water Treat.
48, 27–37.
Salahi, A., Badrnezhad, R., Abbasi, M., Mohammadi, T., Rekabdar,
F., 2012c. Oily wastewater treatment using a hybrid UF/RO
system. Desalin. Water Treat. 28, 75–82.
Salahi, A., Mohammadi, T., Mosayebi Behbahani, R., Hemmati,
M., 2015. Asymmetric polyethersulfone ultrafiltration
membranes for oily wastewater treatment: synthesis,
characterization, ANFIS modeling, and performance. J.
Environ. Chem. Eng. 3, 170–178.
Santos, I.D., Dezotti, M., Dutra, A.J.B., 2013. Electrochemical
treatment of effluents from petroleum industry using a
Ti/RuO2 anode. Chem. Eng. J. 226, 293–299.
Seyed Shahabadi, S.M., Reyhani, A., 2014. Optimization of
operating conditions in ultrafiltration process for produced
water treatment via the full factorial design methodology.
Sep. Purif. Technol. 132, 50–61.
Shaffer, D.L., Arias Chavez, L.H., Ben-Sasson, M., Romero-Vargas
Castrillon, S., Yip, N.Y., Elimelech, M., 2013. Desalination and
reuse of high-salinity shale gas produced water: drivers,
technologies, and future directions. Environ. Sci. Technol. 47,
9569–9583.
Shams Ashaghi, M.E.K., Czermak, P., 2007. Ceramic ultra- and
nanofiltration membranes for oilfield produced water
treatment—a mini review. Open Environ. Sci. 1, 1–8.
202 Process Safety and Environmental Protection 1 0 0 ( 2 0 1 6 ) 183–202
Shokrollahzadeh, S., Golmohammad, F., Naseri, N., Shokouhi, H.,
Arman-mehr, M., 2012. Chemical oxidation for removal of
hydrocarbons from gas–field produced water. Procedia Eng.
42, 942–947.
Singh, D., Sirkar, K.K., 2012. Desalination of brine and produced
water by direct contact membrane distillation at high
temperatures and pressures. J. Membr. Sci. 389, 380–388.
Su, B., Dou, M., Gao, X., Shang, Y., Gao, C., 2012. Study on
seawater nanofiltration softening technology for offshore
oilfield water and polymer flooding. Desalination 297, 30–37.
Subramani, A., Jacangelo, J.G., 2015. Emerging desalination
technologies for water treatment: a critical review. Water Res.
75, 164–187.
Szep, A., Kohlheb, R., 2010. Water treatment technology for
produced water. Water Sci. Technol. 62, 2372–2380 (a journal
of the International Association on Water Pollution Research).
Thoma, G.J., Bowen, M.L., Hollensworth, D., 1999. Dissolved air
precipitation/solvent sublation for oil-field produced water
treatment. Sep. Purif. Technol. 16, 101–107.
Tomer, N., Mondal, S., Wandera, D., Wickramasinghe, S.R.,
Husson, S.M., 2009. Modification of nanofiltration membranes
by surface-initiated atom transfer radical polymerization for
produced water filtration. Sep. Sci. Technol. 44, 3346–3368.
Vatai, G.N., Krstic, D.M., Koris, A.K., Gáspár, I.L., Tekic, M.N., 2012.
Ultrafiltration of oil-in-water emulsion: comparison of
ceramic and polymeric membranes. Desalin. Water Treat. 3,
162–168.
Wandera, D., Wickramasinghe, S.R., Husson, S.M., 2011.
Modification and characterization of ultrafiltration
membranes for treatment of produced water. J. Membr. Sci.
373, 178–188.
Wandera, D., Himstedt, H.H., Marroquin, M., Wickramasinghe,
S.R., Husson, S.M., 2012. Modification of ultrafiltration
membranes with block copolymer nanolayers for produced
water treatment: the roles of polymer chain density and
polymerization time on performance. J. Membr. Sci. 403–404,
250–260.
Wang, P., Chung, T.-S., 2015. Recent advances in membrane
distillation processes: membrane development, configuration
design and application exploring. J. Membr. Sci. 474, 39–56.
Wang, D., Tong, F., Aerts, P., 2012. Application of the combined
ultrafiltration and reverse osmosis for refinery wastewater
reuse in Sinopec Yanshan Plant. Desalin. Water Treat. 25,
133–142.
Wang, Y., Yan, J.N., Li, Z.Y., Wang, L.G., Wu, J., Tao, Y., You, L.C.,
2014. The mechanism of removing the organic matter in
heavy oil sewage by the electric flocculation method. Pet. Sci.
Technol. 32, 1529–1536.
Weschenfelder, S.E., Mello, A.C.C., Borges, C.P., Campos, J.C., 2015.
Oilfield produced water treatment by ceramic membranes:
preliminary process cost estimation. Desalination 360,
81–86.
Witze, A., 2015. Race to unravel Oklahoma’s artificial quakes.
Nature 520, 418–419.
Wright, K.C., Kim, H.S., Cho, D.J., Rabinovich, A., Fridman, A., Cho,
Y.I., 2014. New fouling prevention method using a plasma
gliding arc for produced water treatment. Desalination 345,
64–71.
Xu, P., Drewes, J.E., 2006. Viability of nanofiltration and ultra-low
pressure reverse osmosis membranes for multi-beneficial use
of methane produced water. Sep. Purif. Technol. 52,
67–76.
Xu, P., Drewes, J.E., Heil, D., 2008. Beneficial use of co-produced
water through membrane treatment: technical-economic
assessment. Desalination 225, 139–155.
Yan, H.Y., Xiao, M., Zhang, Z.Z., Li, J.Q., Shi, B.Q., 2014.
Remediation of oilfield wastewater produced from
alkaline/surfactant/polymer flooding by using a combination
of coagulation and bioaugmentation. Pet. Sci. Technol. 32,
1521–1528.
Yuan, T., Meng, J., Hao, T., Zhang, Y., Xu, M., 2014. Polysulfone
membranes clicked with poly(ethylene glycol) of high density
and uniformity for oil/water emulsion purification: effects of
tethered hydrogel microstructure. J. Membr. Sci. 470,
112–124.
Zhang, J.C., Wang, Y.H., Song, L.F., Hu, J.Y., Ong, S.L., Ng, W.J., Lee,
L.Y., 2005. Feasibility investigation of refinery wastewater
treatment by combination of PACs and coagulant with
ultrafiltration. Desalination 174, 247–256.
Zhang, T., Gregory, K., Hammack, R.W., Vidic, R.D., 2014a.
Co-precipitation of radium with barium and strontium sulfate
and its impact on the fate of radium during treatment of
produced water from unconventional gas extraction. Environ.
Sci. Technol. 48, 4596–4603.
Zhang, S., Wang, P., Fu, X., Chung, T.S., 2014b. Sustainable water
recovery from oily wastewater via forward
osmosis-membrane distillation (FO-MD). Water Res. 52,
112–121.
Zhao, S., Huang, G., Cheng, G., Wang, Y., Fu, H., Hardness, 2014.
COD and turbidity removals from produced water by
electrocoagulation pretreatment prior to Reverse Osmosis
membranes. Desalination 344, 454–462.
Zhu, X., Tu, W., Wee, K.-H., Bai, R., 2014. Effective and low fouling
oil/water separation by a novel hollow fiber membrane with
both hydrophilic and oleophobic surface properties. J. Membr.
Sci. 466, 36–44.