Reverse Osmosis Water Desalination: Quest for Sustainability

Full text

Perspective
A Perspective on Reverse Osmosis Water
Desalination: Quest for Sustainability
Yoram Cohen
Dept. of Chemical and Biomolecular Engineering, Institute of the Environment and Sustainability,
University of California, Los Angeles, CA
Raphael Semiat
Wolfson Faculty of Chemical Engineering, Technion – Israel Institute of Technology, Technion City,
Haifa 32000, Israel
Anditya Rahardianto
Dept. of Chemical and Biomolecular Engineering, Institute of the Environment and Sustainability,
University of California, Los Angeles, CA
DOI 10.1002/aic.15726
Published online April 19, 2017 in Wiley Online Library (wileyonlinelibrary.com)
Keywords: Desalination, Water Treatment and Water Purification, Water Sustainability, Membrane-based Separations,
Desalination Energy, Reverse Osmosis, Environmental Impact of Desalination
Introduction—Water Scarcity and the
Rationale for Water Desalination
Water is needed at different qualities for various types
of municipal/public, agricultural, and industrial
consumption which account for about 23, 67, and
9%, respectively, of total fresh water withdrawals in the
United States, excluding thermoelectric use.
1
Although water
reclamation, recycling, and reuse could be advanced so as to
provide a significant percentage (20–50%) of the water use
portfolio, at present it remains in the single digits in most
regions of the United States.
2
Water and energy are inextrica-
bly linked; nationwide, there is an urgent need to reduce the
energy cost of water delivery, which represents a major por-
tion of the energy usage in various sectors.
3
For example, the
total energy consumption for direct water services in the
United States by the US residential, commercial, and industrial
sectors was estimated to be about 20.2, 12.7, and 4.1%, respec-
tively, in 2010.
3
Population growth, aging water infrastructure
requiring significant upkeep and retrofitting investment,
increased demand for crop production (for both food and bio-
fuels), shrinking freshwater resources, increased soil, ground-
water, and surface water salinity, contamination of
groundwater resources, fossil fuel extraction, and increased
security concerns have also created multiple threats to the
water sustainability in the United States and globally.
4–8
Given
the above realities, along with the impending impact of global
climate change, it is not surprising that the development of
new water resources is high on the priority list of various
water agencies around the United States and the globe.
In addition to the various approaches to increasing the
utilization of water resources (e.g., water use efficiency,
water recycling, and reuse), there has been increased inter-
est in augmenting potable water supplies through desalina-
tion of seawater and inland brackish water (BW).
9–12
Water
salinity is typically categorized based on the concentration
of total dissolved solids (TDS; i.e., mineral salts) as shown
in Table 1. The recommended salinity level for drinking
water is 500 mg/L TDS,
13
while seawater salinity is typi-
cally in the range of 33,000–42,000 mg/L TDS. BW is
classified as having salinity above 1000 mg/L but below
that of seawater, with further classification from mild brack-
ish to heavily BW.
14
In assessing the potential availability
of saline water resources for desalination, it should be rec-
ognized that about 97% of the world water is saline. About
1.74% of the word’s water is trapped in ice caps and glaciers
and the remainder as groundwater (1.69%), surface water
(0.014%), and moisture (0.002%).
15
Clearly, freshwater
supplies (for potable and nonpotable use) are limited and
constitute but a small fraction of the world water resources.
While the majority of the world’s saline water volume is in
seas and oceans, inland water (e.g., groundwater aquifers,
lakes, and wastewater) resources can also be saline, but are
typically at lower salinity levels (i.e., brackish) relative to
seawater. The potential contribution of desalted brackish
groundwater to the overall potable water supply can be sig-
nificant considering that subsurface saline aquifers are
abundant throughout much of the continental United States
and many parts of the world.
10,12,16
Thus, it is not surprising
that to augment dwindling freshwater supplies, various
Correspondence concerning this article should be addressed to Yoram Cohen at
yoram@ucla.edu.
V
C2017 American Institute of Chemical Engineers
AIChE Journal 1771June 2017 Vol. 63, No. 6

inland communities across the United States and globally
have resorted to desalting brackish groundwater reserves
(1000–10,000 mg/L TDS).
9
The process of desalination involves the treatment of a
saline water source to separate dissolved salts from the feed
stream to produce low salinity product water and a residual
stream of high salinity water. Given major advances in both
thermal and nonthermal based membrane desalination technol-
ogies,
11,12,16,17
seawater desalination has developed rapidly in
a number of countries (e.g., Singapore, Australia, Israel, Spain,
Saudi Arabia, and Gulf-States), with others also adding desali-
nation to supplement freshwater supplies.
9
In addition to its
use for desalting brackish surface and groundwater, membrane
based desalination technology is utilized in indirect potable
water reuse applications (e.g., aquifer recharge, industrial use,
and irrigation) to reduce salinity and provide a barrier against
multiple contaminants (e.g., organics, viruses, bacteria, in
addition to toxic inorganics such as boron
18
and selenium), as
well as for reclamation of industrial, mining, and agricultural
wastewaters.
10,12,19,20
Water desalination is practiced via various approaches of
which reverse osmosis (RO) desalination has become the dom-
inant technology.
16,21–24
Accordingly this perspective presents
an overview of water desalination and comparison of energy
requirements by the main desalination technologies practiced
today, followed by a brief account of the basics of RO
desalination process configuration and the challenges imposed
by RO concentrate management, membrane fouling and min-
eral scaling, and feed pretreatment. The various factors that
contribute to the overall cost of RO desalination are then intro-
duced and the energy cost of RO desalination is assessed con-
sidering the thermodynamic limit and the implications for
feasible future reductions in energy utilization and other
aspects of RO process operation that could provide further
reduction in the cost of RO desalination.
Water Desalination
Overview
The total number of desalination (seawater and BW) plants
in the world (in 2015) is reported to be 18,426 with a total
water desalination capacity of about 86.5 million m
3
/day.
9
The majority of the world desalination installation capacity
(Figure 1a) is for desalting of seawater (59%), followed by
BW (22%), with the remainder (19%) being for river water,
wastewater, and BW.
9
In terms of worldwide desalination
capacity by technology (Figure 1b), RO has emerged as the
leading desalination technology practiced today (65%), fol-
lowed by multi-stage flash evaporation (21%). The remaining
14% of the capacity include multiple effect distillation, vapor
compression, and electrodialysis (ED)/ED reversal (EDR).
Other desalination technologies have also been proposed in
recent years including, at a smaller scale, membrane distilla-
tion
25
and forward osmosis.
26–28
The total water desalination capacity in the United States is
about 4.5 million m
3
/day
29
with most being inland BW desali-
nation with the majority via RO technology. More than 300
municipal BW desalination plants operating in the United
States are primarily located in Florida (45%), California
(14%), and Texas (9%).
30
The largest online brackish RO
water desalination plants in the United States is the Orange
County Groundwater Replenishment System having a capacity
of 378,000 m
3
/day with planned expansion to 492,000 m
3
/day.
The largest operating seawater RO (SWRO) plants in the
Table 1. Characterization of Water Salinity
Typical TDS Levels of Different Water Sources
Type TDS (mg/L)
Drinking water 500
Fresh water <1000
Mildly brackish water 1000–5000
Moderately brackish water 5000–15,000
Heavily brackish water 15,000–35,000
Seawater 35,000
Source: Watson et al.
14
Figure 1. (a) Desalination installation capacity by water source. (b) Desalination capacity by technology (Total: 86.5 million
m
3
/day).
9
1772 DOI 10.1002/aic Published on behalf of the AIChE June 2017 Vol. 63, No. 6 AIChE Journal

United States are: (a) the Tampa Bay plant in Florida (capacity
of 94,500 m
3
/day), and (b) the new plant in Carlsbad, Califor-
nia (capacity of 189,000 m
3
/day) began its operation in late
2015.
Comparison of the energy cost of different energy desalina-
tion methods based on the published literature is not trivial
given the variability of feed salinity and quality encountered
in the various operating plants. Therefore, it is not surprising
that the reported specific energy consumption (SEC, i.e.,
energy per volume of desalted water product) can vary by up
to an order of magnitude in some cases (Table 2). In general,
the SEC of desalination via RO is lower for seawater than by
thermal methods.
11,31,32
BW desalination via RO and EDR
can be competitive depending on the salinity level.
10,11
Both
RO and EDR require only electrical energy, while processes
based on evaporation/distillation also require thermal energy.
It is noted that heat- and osmotically driven (with thermolytic
draw solutions, e.g., FO) desalination have been reported to be
cost-effective only if low-cost heat source (or “waste heat”) is
readily available locally to drive the process or to provide for
regeneration of a process stream (e.g., in FO).
27,33–35
RO desalination
RO Emergence as the Standard Technology for Water
Desalination. The popularity of RO desalination technology
and its growing market share
9
have been, in part, due to its
simplicity and reliable operations and maintenance support by
diverse and well-established supply chain of off-the-shelf
components and consumables (e.g., membrane elements, pre-
filters, compatible water treatment additives and membrane
cleaning chemicals, etc.).
10,16,17
Moreover, the process is eas-
ily scalable whereby both small- and large scale RO plants can
use the similar types of membrane elements and pressure ves-
sels (Figure 2). RO desalination relies on a semi-permeable
membrane (housed in a membrane element inside a pressure
vessel) that allows water permeation but rejects dissolved sol-
ids. The feed side of the membrane is pressurized to a level
above the osmotic pressure so as to provide the desired perme-
ate water flux. The RO process is driven by a pump for fluid
conveyance and for generating the required raw feed pressure.
The overall desalination process consists of process trains that
includes feed intake system, feed water pretreatment, the desa-
lination separation system, product water post-treatment, and
Table 2. Summary of Reported Specific Energy Consumption (SEC) for Seawater and Brackish Water
a
Desalination Technology
SEC
b
(KWh/m
3
)
Total SEC
c
(KWh/m
3
)
Total Water
Production Cost
d
($/m
3
)Electrical Thermal
Reverse osmosis (Seawater) 3
e
–7.5 NA 3
e
–7.5 0.53–1.72
Reverse osmosis (brackish water) 0.3–3 NA 0.3–3 0.2–1.33
Electrodialysis (brackish water) 0.5–1.8 NA 0.5–1.8 0.6–1.05
Multi-effect distillation 4–20.2 1.5–2.5 5.5–22.7 0.52–1.5
Multi-stage flash (brackish water) 7.5–30.3 2.5–5 10–35.3 0.56–1.75
Thermal vapor compression – 16.3 16.3 0.27–1.6
Mechanical vapor compression 7–12 – 7–12 –
a
Source: Refs. 11, 16, 17, 31, 36 and 83. Note: reported specific energy consumption (SEC) is for 1 m
3
of desalted product water. The wide range of costs is
due to differences in source water salinity, plant size, pump, and other system components efficiencies, product water recovery (i.e., product volume/feed
volume), and heat quality (for thermal desalination processes), and in some cases the inclusion of intake and discharge pumping energy.
b
Thermal desalination processes require thermal energy and electrical energy. In the absence of available thermal energy source electrical energy is used (con-
verted) to provide the needed thermal energy.
c
Reported range of total SEC for the desalination process.
d
Total water cost per 1 m
3
of desalted product water includes energy, capital and O&M costs.
e
Note: The theoretical minimum energy consumption at 50% recovery for seawater desalination using a single stage RO with 100% efficient pump and energy
recovery is estimated to be 1.6 KWh/m
3
based on the theoretical analysis in Ref. 31.
Figure 2. Left: Photo of the desalination plant at Orange County (Capacity of 378,000 m
3
/day).
47
Right: a small mobile RO
desalination plant (capacity of 45.4 m
3
/day).
48,49
The same size membrane elements (800) are used in both plant
demonstrating the scalability of RO desalination.
AIChE Journal June 2017 Vol. 63, No. 6 Published on behalf of the AIChE DOI 10.1002/aic 1773

concentrate management (Figure 3). An intake system is
required to convey the raw feed water from the source to the
plant, and concentrate management is required to handle the
high salinity concentrate (or brine). Pretreatment of the raw
RO feed water is a crucial component of RO desalination tech-
nology (Figure 3). The levels of suspended solids, organics,
and microorganisms in the RO feed must be reduced to a suffi-
cient level to avoid fouling of the RO membranes.
12,16
The
extent of fouling has been described to depend on a variety of
factors including hydrodynamic forces, membrane properties
(e.g. surface charge, permeability, surface roughness), and sol-
ute on ionic matrix.
24,37,38
RO operation can also be negatively
impacted by mineral scaling which can occur when the con-
centrations of sparingly soluble minerals (e.g., SiO
2
) and min-
eral salts (e.g., gypsum (CaSO
4
2H
2
O), BaSO
4
, SrSO
4
,
CaCO
3
, etc.) in the RO feed rise, as product water is extracted,
above their solubility limit.
39–45
The consequence is mineral
precipitation and scale formation on the RO membrane. Mem-
brane fouling and mineral scaling lead to membrane surface
blockage and, as a result, degradation of membrane perfor-
mance (i.e., permeate flux decline and deteriorating salt rejec-
tion) which can shorten membrane longevity and increase
operational costs.
43,46
Mitigation of Fouling and Mineral Scaling. Current prac-
tice of membrane fouling prevention emphasizes the removal
of potential foulants, prior to any RO desalting unit operations,
using conventional flocculation/coagulation, disinfection (e.g.,
via chlorine dosing or UV pretreatment), media filters or mem-
brane filtration (microfiltration or ultrafiltration) processes for
RO feed water pretreatment.
16,32,43,48,50,51
In some cases,
sodium metabisulfite, sodium bisulfite, or activated carbon
may also be used for dechlorination of the RO feed to avoid
RO membranes degradation due to chlorine;
52,53
it is noted,
however, that the use of chlorine in RO feed is now typically
avoided in modern plants.
11
These feed pretreatment pro-
cesses, while effective for minimizing membrane fouling due
to particulate deposition, organic adsorption, and biological
growth, do not remove mineral scale precursors. Mineral scal-
ing is typically mitigated via: (a) the addition of polymeric
antiscalants additives to the feed to suppress the nucleation
and/or growth of mineral crystals and promote dispersion of
crystals,
37,41,43,46,54
(b) feed pH adjustment if calcium carbon-
ate is an issue of concern,
16,54
(c) regulation of water recovery
to keep the level of retentate concentration within acceptable
limits (as governed by mineral scaling kinetics),
40,43
(d) peri-
odic membrane cleaning (e.g., via fresh water flush, osmotic
backwash, feed flow reversal, or chemical cleaning),
12,16,55,56
and (e) removal of scale precursors via crystallization, nanofil-
tration, or ion exchange.
12,20,57–59
In certain cases where boron
and carbon dioxide removal are necessary and where silica
scaling is to be avoided (at high recovery), RO operation at
high pH may be necessary.
42,44,60–63
Chemical post-treatment
of the product water may also be necessary and, depending on
the end use, can include disinfection, remineralization, and pH
adjustment.
64
Concentrate Management. RO desalination is typically
carried out at a product water recovery of 40–50% for seawa-
ter desalting and in the range of 50–90% for BW desalt-
ing.
10,12,16,32
As a result a concentrate retentate (or brine)
stream is generated which must be managed in an environ-
mentally compatible manner.
20,65–69
In seawater RO desalina-
tion, the brine stream is discharged to the sea through an
elaborate outfall system designed to disperse the concentrate
in a manner that reduces local elevated salinity impacts and
precipitation along the discharge pipes.
16,32,70
Environmental
considerations with respect to the implementation of ocean
discharge of RO concentrate include, for example, the salinity
tolerance of marine organisms, discharge toxicity (due to pos-
sible chemical water treatment/cleaning additives), long-term
local salinity buildup, and the need for meeting effluent water
quality standards.
10,11,30
At inland locations concentrate man-
agement options are more limited.
10,12,20,71
In some cases
sewer discharge may be feasible (for sufficiently low salinity
concentrate), but may be restricted by regulatory restrictions.
Disposal of RO concentrate in evaporation ponds may be pos-
sible in some locations, but may be limited due to large area
requirement and constrained holdup and possible concerns
with concentrate toxicity. Deep well injection is another
option which can be costly (in some cases, due to the high
injection pressure requirements) and in some areas deep geo-
logical injection may be restricted. In some cases, ocean dis-
posal via a long waste discharge pipeline may be feasible but
can be burdened by mineral scaling.
70
In principle, thermally
driven evaporative and crystallization systems, the so-called
zero-liquid-discharge technologies, are capable of eliminating
liquid RO concentrate discharge, but are often prohibitively
energy intensive for processing large volumes.
20,57–59,63,72
Consequently, high recovery inland desalination (if not limited
by mineral scaling) is desirable to reduce the volume of gener-
ated brine.
57–59
Configurations of RO Desalination Systems
RO membrane desalination relies on a semipermeable mem-
brane that allows water permeation but rejects dissolved solids
Figure 3. Process elements in water desalination.
1774 DOI 10.1002/aic Published on behalf of the AIChE June 2017 Vol. 63, No. 6 AIChE Journal

(Figure 4). The product water flux (J
w
) is dictated by the dif-
ference between the applied pressure (DP) and the osmotic
pressure (Dp) differences across the membrane as given by the
classical expression, Jw5LpðDP2rDpÞ, in which L
p
is the
membrane permeability and ris the so-called reflection coeffi-
cient,
73
which is in the range of [0,1] and nearly unity for high
salt rejecting membranes. The solute flux through
the membrane, J
c
, is typically described as Jc5JwCp5
BDC1ð12rÞJw

C,inwhichC
p
isthepermeatesolute
concentration, Bis the solute (membrane) transport coef-
ficient and 
Cis the average solute concentration across
the membrane. As water permeates across the membrane,
the rejected solute accumulates near the membrane sur-
face resulting in a concentration polarization layer; thus,
the osmotic pressure at the membrane surface increases
and hence the requirement for higher applied pressure
for a given target water permeate flux.
Most RO systems employ membrane channels operated in a
cross flow operation which serves to reduce salt concentration
buildup (i.e., concentration polarization) at and in the vicinity
of the membrane surface (Figure 4a).
74–76
The flow and con-
centration fields within membrane channels are complex, but
the solute concentration at the membrane surface, C
m
, can be
estimated, to within a reasonable level of approximation, as
Cm=Cb5CP5ð12RoÞ1Roexp ðJw=kcÞ, where C
b
is the solute
concentration in the bulk of the feed channel, CP is the so-
called concentration polarization modulus, R
O
is the observed
solute rejection (i.e., R
o
212C
p
/C
b
) and k
c
is the feed-side sol-
ute mass transfer coefficient. Salt concentration in the retentate
stream (“brine” or “concentrate”) increases as it flows down-
stream along the membrane channel. For example, for seawa-
ter RO operation at 40% recovery, 60% of the processed feed
volume will be a brine stream with a salt concentration a fac-
tor of 1.67 above that of the raw feed. At a higher recovery of
80%, which would more typical for low salinity BW, the brine
salt concentration would be a factor of five above that of the
raw feed.
Membrane elements are general constructed as multiple
membrane sheets arranged (using spacers) to form flow chan-
nels through which pressurized saline feed water flows
(Figure 4b). Standard commercial spiral-wound RO mem-
brane elements which are housed in pressure vessels allow
operationuptoamaximumpressureof1200 psi (8.3 MPa)
and 600 psi (4.1 MPa) for seawater and BW membranes,
respectively. Standard RO process configurations (Figure 5)
typically comprise of one or more pumps and one or more
membrane arrays, coupled with a throttling valve for concen-
trate depressurization.
10,77,78
To avoid wasting the pressure
energy of the discharged brine stream, energy recovery devi-
ces (ERDs) can be integrated with the RO system.
10,79
Typi-
cally, single stage RO processes are integrated with ERDs
(Figure 6) such as a pressure intensifier/booster or parallel
feed pumping,
10,78,79
which presently are mostly limited to
seawater desalination applications. Reduction in energy con-
sumption through the use of multi-stage RO with inter-stage
(booster) pumps is also possible and is most suitable for BW
desalination in which a high product level of product recov-
ery is needed.
10,78
Conventional RO systems typically operate over a narrow
range of water recovery. The above limitation is because the
system feed flow rate is directly coupled with the main feed
pump inlet, the membrane array feed and, for the case in
which ERD is utilized, the ERD device feed- and concentrate-
side flow rates. Also, flow-range restrictions, imposed by pro-
cess components for effective and energy-optimal operation
often restrict the range of water recovery in conventional sys-
tems. The alternative RO configuration of concentrate recy-
cling/recirculation, at partial or total recycle (i.e., batch
Figure 4. (a) Cross-flow reverse osmosis in a membrane channel. (Q and C and P denote flow rate, salt concentration, and
hydraulic pressure, respectively, and where subscripts f, b, c, p, and m denote the feed, bulk feed-side, concentrate,
permeate, and membrane surface, respectively; J
w
is the water flux through the membrane. (b) Schematic of a
spiral-wound RO membrane element.
AIChE Journal June 2017 Vol. 63, No. 6 Published on behalf of the AIChE DOI 10.1002/aic 1775

operation with permeate withdrawal and periodic system feed
flushing) can improve operational flexibility and increase RO
recovery for a given footprint.
80,81
The Cost of RO Desalination
Over the years of RO development, there have been signifi-
cant technology improvements with respect to membranes and
membrane modules, plant standardization, operational effi-
ciency, and energy recovery. This has led to a steady decline
in the cost of SWRO desalination (Figure 7). It is projected
that the cost of SWRO desalination is approaching or nearly
comparable in certain cases with the cost of traditional water
sources.
82
Water production cost in a typical RO desalination plant
generally consists of the cost of energy consumption, equip-
ment, membrane replacements, residual concentrate (i.e.,
brine) management, labor, maintenance, and financial
charges. The cost breakdown for desalination plants can vary
greatly depending on plant size and location, quality and
salinity of the water source, and local electrical energy cost.
For a typical large seawater RO desalination plant, energy,
and capital costs (Table 3) constitute the major portion of the
overall water production cost (e.g., in terms of $/m
3
product)
with the remainder attributed to operation and maintenance
costs.
11,16,32
The SEC is strongly affected by water product recovery,
energy recovery (from the high pressure RO concentrate),
pumping energy (including feed intake), operating conditions
(e.g., RO feed channel velocity and feed-side pressure), and
plant configuration in terms of membrane modules arrange-
ments.
10,11,16,31,78,83–85
The various contributions to the total
energy consumption for a seawater RO desalination plant
typically includes the desalination stage and discharge
(67%), RO feed pre-treatment (13%), intake (7%) and
Figure 5. Typical arrangements of pressure vessels for RO/NF membrane elements.
16
(F, C, and P designate the feed, concen-
trate and permeate streams, respectively).
Figure 6. Conventional RO membrane system configuration
integrating energy recovery (ER) devices: (a) feed
pressure booster/intensifier, (b) parallel feed
pumping, and (c) inter-stage pressure booster.
(MA – membrane array; P1 and P2: feed and
booster pumps, respectively; o, f, c, p, and w des-
ignate the raw water source, and RO unit feed,
concentrate, permeate and discharge streams,
respectively).
Figure 7. The cost of water seawater RO desalination rela-
tive to traditional water sources.
82
1776 DOI 10.1002/aic Published on behalf of the AIChE June 2017 Vol. 63, No. 6 AIChE Journal

post-treatment (13%).
36
Because seawater RO is energy
intensive, the costs of seawater RO is often considered high.
However, this can be misleading as often desalination costs
are not compared to the costs of locally available (and feasi-
ble) alternatives (Figure 7, Ref. 31) It has also been argued
that in certain cases, the energy costs of pumping/conveying
water from large distances can be higher than the energy
required for water production by a large desalination plant
located in proximity to water consumers and when such a
plant utilizes off-peak electrical energy.
11,27,31
Another com-
mon concern regarding seawater desalination is the discharge
of high salinity concentrate (up to twice the salinity of sea-
water) to the sea. The concentrate stream may contain chemi-
cals (e.g., surfactants, antiscalants, and acid/base for pH
control) used for membrane cleaning and scale prevention in
addition to rejected backwash from feed water pretreatment.
Today there are continuing efforts to introduce “green” water
treatment chemicals so as to avoid potential environmental
impacts. At the same time, to reduce potential environmental
impacts, modern technologies of concentrate disposal to the
sea provide for sufficient concentrate dispersion through
multiple nozzles, upward (subsurface) concentrate discharge,
and where feasible mixing the concentrate with power station
cooling water.
11
The energy cost of desalting BW is lower than for seawater
(given the lower salinity of BW, Table 1). The energy cost is
in the range of 20–30% of the total water production cost, but
can be higher for heavily BW.
10,12,16
However, unlike seawa-
ter RO plants, the management of the RO concentrate from
inland brackish desalination can represent a major challenge
given the often limited options for concentrate dis-
posal.
10,12,20,30
Inland concentrate management costs can be as
high as 0.40–1.78 $/m
3
and in certain cases represent more
than 50% of the cost of water production.
86
Hence, there is a
significant drive to increase RO recovery so as to minimize
the volume of generated RO residual stream.
10,12,30,43,57–59,86
The Energy Cost of Ro Desalination
Minimum energy consumption for a reversible
desalination process
The minimum isothermal reversible work, W
rev
, for separat-
ing salt from water, irrespective of the separation mechanism,
is given as follows:
31,87
Wrev5ð
n2
n1
DFdn5ð
n2
n1
RTln awdn5ð
n2
n1
Ps

Vwdn (1)
in which DFis the change in free energy associated with the
transition from the initial (1) to final (2) states of a solution as
its salt concentration is altered, Ris the ideal gas constant, Tis
the absolute temperature, a
w
is the water activity, nis the num-
ber of moles of water, Psis the solution osmotic pressure, and

Vwis the water molar volume. As the feed water salinity
increases so does the osmotic pressure (reasonably approxi-
mated for a wide salinity range by Ps5uCRT, where uis the
van’t Hoff factor and Cis the salt molar concentration
83
). Fol-
lowing Eq. 1, it can be shown that the minimum energy of
desalination, at the limit of zero product water recovery, for
desalting seawater of salinity of 35,000 mg/L TDS (osmotic
pressure is 27 bar) is about 0.79 kWh/m
3
. Increasing the
water recovery would elevate the osmotic pressure, thereby
increasing the required reversible desalination work. The
dependence of the minimum required isothermal reversible
work to achieve a given water recovery can be deduced from
Eq. 1
88,89
and expressed in terms of the SEC normalized with
respect to the feed water osmotic pressure (p
o
):
NSECrev5SECrev
po
51
po
Wrev
Vf2Vc

52 ln 12YS
ðÞ
YS
(2)
where V
f
and V
c
are the feed and concentrate volumes, respec-
tively, and Y
S
is the fraction of product water recovered from
the feed. Based on Eq. 2, for example, desalination of seawater
at a water recovery of say 50% would require 1.0 kWh/m
3
.
However, desalting at the above level of energy consumption
for a reversible process (at infinitesimal permeate flux) is
clearly impractical. It is noted that the most energy efficient
SWRO desalination, under non-reversible thermodynamic
operation with a 100% efficient pump and energy recovery, is
estimated to be about 1.6 kWh/m
3
(Table 2). Given the current
state of RO efficiency, further RO SEC reduction is likely to
require a higher capital cost for the same level of target perme-
ate productivity.
RO energy consumption and the thermodynamic limit
The minimum transmembrane pressure that must be applied
to ensure permeate production along the entire membrane area
(in the axial fluid flow direction), for crossflow membrane
channels, must be such that DPDpexit, where Dpexit is the
osmotic pressure difference (across the membrane) at the RO
channel exit.
83
Using the approximate linear relation for the
Table 3. Elements of Capital Expenses (CAPEX) and
Operational and Maintenance (O&M) Costs for Seawater
RO Desalination
Cost of RO Desalination
a
Percent of Total Water
Production Cost
Capital 34
Energy 38
Chemicals 5
Labor 3
Replacement parts 9
Membranes replacement 5
Overhead 5
Insurance 1
Capital Expenses (CAPEX)
a
Percent of Total CAPEX
Desalination system 31
Power system 26
Pre-treatment 12
Intake and outfall (discharge) 11
Design and permitting 7
Other 13
O&M
b
Percent of Total O&M
Fixed cost 28–50
Energy 32–44
Operation and maintenance 18–28
Source: Adapted from Ref. 16, 36.
a
Cost for a typical large plant. Actual costs will vary depending on plant
design, product water recovery, level of automation, outfall requirements.
b
O&M costs can vary considerable depending on plant design, level of auto-
mation, intake and outfall, and feed pretreatment requirements.
AIChE Journal June 2017 Vol. 63, No. 6 Published on behalf of the AIChE DOI 10.1002/aic 1777

osmotic pressure dependence on salt concentration, it has been
shown that the minimum required applied pressure difference
is DPðÞ
min poRt=ð12YSÞ(in which R
t
is the membrane salt
rejection), which clearly increases with both rising feed
osmotic pressure and water recovery.
83
With the old genera-
tion of relatively low permeability (Lp) cellulose acetate mem-
branes, it was necessary to operate at a feed pressure that was
much higher than the brine osmotic pressure to produce a rea-
sonable level of permeate flux for a viable desalting operation.
However, with the development of highly permeable polyam-
ide based thin-film composite membranes and their improve-
ments over the last two decades, higher permeate fluxes can be
attained at significantly lower feed pressure enabling cross-
flow RO operation near the limit imposed by the thermody-
namic restriction.
78,90
The energy footprint of RO desalination can be expressed in
terms of the SEC normalized with respect to the feed osmotic
pressure (po):
NSEC SEC
po
51
poqp
X
N
i51
_
Wi
gp;i
! (3)
where _
Wiand gp;iare the rate of work and efficiency of pump i
in a multi-stage system with Ninter-stage pumps, q
p
is the
total permeate flow rate and pois the RO feed osmotic pres-
sure. In principle, NSEC of conventional single-stage RO with
an ERD, at the limit of cross-flow thermodynamic restriction
(tr), can be described by the following relationship:
85
NSECtr 512gERD 12YS
ðÞðÞRt
gpYS12YS
ðÞ (4)
where gpand gERD are the pump and ERD efficiencies, respec-
tively, and YS5qp=qo(where q
o
and q
p
are the raw feed and
product water flow rates, respectively) is the product water
recovery and R
t
is the salt rejection. In the absence of an ERD
(i.e., gERD50) NSECtr 5Rt=gpYS12YS
ðÞ

:Thus, the energy
savings of utilizing and ERD is given by gERD Rt=gpYS

,
which is more pronounced at lower water recovery. A plot of
Eq. 4 (Figure 8) illustrates that the energy consumption, for
RO process operation (for the case of a membrane of 100%
salt rejection) up to the thermodynamic limit, is highly depen-
dent on the product water recovery, as well as pump and ERD
efficiencies, but it is independent of membrane permeability
(Figure 3).
The global minimum (w.r.t energy consumption) for RO
operation, without an ERD and assuming constant pump effi-
ciency, is at recovery of 50%. For seawater desalination (salin-
ity of 35,000 mg/L TDS) and ideal pump (i.e., gp51) this
implies an energy consumption of 3.2 kWh/m
3
(NSEC 54).
This energy consumption level may be reduced to 1.6 kWh/m
3
h (NSEC 52) with an ideal ERD (i.e., gERD 51) and to 1.68
kWh/m
3
(NSEC 52.1) for ERD of 95% efficiency. In princi-
ple, RO plants that are designed to operate close to the thermo-
dynamic limit, which can be feasible with currently available
commercial RO membranes, should be able to operate opti-
mally so as to reduce energy consumption. When the RO plant
is incapable of operating close to the thermodynamic limit
(due to design and equipment restrictions), one can drive the
RO process toward the optimal operating condition via suit-
able model-based control.
49,91
In such an approach one must
also consider constraints imposed by the target water produc-
tion capacity and limitations placed on the operability of
plants components (e.g., maximum allowable operating pres-
sure for pumps and pressure vessels).
It is stressed that for RO process operation up to the thermo-
dynamic limit, the SEC is not impacted by membrane perme-
ability (Eq. 4). Therefore, utilization of more permeable
membranes will not provide additional reduction in energy con-
sumption, unless the plant is operating away from the thermo-
dynamic limit.
78,90
For a plant operating at the thermodynamic
limit, the use of higher permeability membranes will, however,
allow reduction in plant footprint owing to the higher achiev-
able water flux which will enable the thermodynamic restriction
to be approached with a smaller membrane surface area. Mem-
brane fouling, however, is more severe at high flux operation
and thus high flux membranes would have to be more fouling
resistant and/or feed pretreatment will need to be more
effective.
Multi-stage RO desalination
The energy foot-print of RO desalting can be reduced with
increasing number of pumping stages. Unlike a single stage
RO configuration that requires pressurizing the entire feed, at
a pressure equal or above the concentrate pressure, in a multi-
stage system membrane units are staged serially with inter-
stage booster pumps (Figure 77). This enables incremental
increase in the feed pressure to each stage and correspondingly
an incremental reduction in the required feed pumping capac-
ity for each successive membrane stage. The NSEC at the
limit of thermodynamic restriction (for each membrane stage,
assuming complete salt rejection and ideal pump efficiency) is
given by:
83,92
Figure 8. Variation of the normalized specific energy con-
sumption with product water recovery for a single
stage RO operation at the thermodynamic limit
(excluding frictional pressure losses which ae typ-
ically <5–10%; note: gpand gERD are the RO
pump and ERD efficiencies, respectively; frictional
pressure losses).
1778 DOI 10.1002/aic Published on behalf of the AIChE June 2017 Vol. 63, No. 6 AIChE Journal

NSECtr 5Nð12YSÞ21=N2N112gERD

=YS(5)
where Nis the number of pumping stages and Y
S
is the overall
system water recovery, which is related to the individual stage
water recovery (Yn) by the relationship 12YS
ðÞ512Yn
ðÞ
N.
Note that Eq. 5 reflects the energy-optimal condition at which
that the individual stage water recovery is identical for all
stages.
83,92
In the limit of infinite number of pumping stages
and ideal energy recovery (gERD 51), Eq. 5 for the multi-stage
RO process reduces to Eq. 2 for the reversible desalination
process. Analysis of Eq. 5 reveals that at recovery above about
68%, infinite-stage RO without an ERD would have lower
energy consumption relative to single-stage RO with an ideal
ERD.
92
With a finite number of stages with inter-stage pumps
(without ERD), the advantage over a single-stage RO with an
ERD will be shifted to even higher recovery.
Concentrate recirculation
Membrane process systems are composed of highly inte-
grated and interdependent process components. Optimal oper-
ation of each system component is often limited to a narrow
flow range, which in turn restricts the integrated process sys-
tem operational range. Pumps (and associated motors) and
energy recovery devices, for example, typically have narrow
flow ranges at which energy efficiencies are optimal. Further-
more, membrane modules are inherently limited with respect
to the flow ranges for optimal operation, while maintaining
conditions that protect physical integrity, maximize membrane
area utilization, minimize concentration polarization, and
avoid excessive membrane fouling and/or mineral scaling.
Such module-level operational limits can significantly restrict
water recovery and permeate productivity ranges of the inte-
grated system, which is highly dependent on the specific mem-
brane array design.
Existing methods for improving operational flexibility of
RO membrane based processes rely on partial recirculation of
RO concentrate (to RO feed); this provides an additional
degree of freedom for steady-state process regulation. The
simplest approach is to recirculate a portion of depressurized
RO concentrate to the RO feed pump (Figure 9a). Although
simple, RO with low pressure concentrate recirculation
(LPCR) is highly energy intensive with energy consumption
(for single stage RO) given by:
NSECLPCR1;tr 51
gpYMA 12YS
ðÞ (6)
Compared to the NSEC for conventional RO without concen-
trate recirculation (Eq. 6), LPCR has a higher energy footprint
due to the required higher flow capacity of the main feed
pump (by a factor of Y
MA
/Y
S
) for achieving the same produc-
tivity. As a consequence, conventional LPCR is often deemed
undesirable and only utilized in limited applications when
operational flexibility is significantly more important than the
energy footprint (e.g., as proposed certain small-scale indus-
trial applications
80
)
RO with high pressure concentrate recirculation (HPCR)
(Figure 9b) is a more energy efficient process when compared
to LPCR. In HPCR, only a portion of the concentrate stream
(w) is depressurized (and discharged) using a throttling valve,
while the remaining concentrate (rc) is recycled to the main
pump (P1) outlet without depressurization.
80
A recirculation
pump (P2) is needed to match the main pump outlet pressure
and make up for the relatively small axial pressure loss in the
membrane array. However, the concentrate pressure energy
remains unrecovered in the concentrate discharge stream (w),
making HPCR suboptimal with respect to energy utilization.
For example, assuming that axial pressure drop in the mem-
brane array (MA) is small such that P2 energy requirement is
minimal and that complete salt rejection can be achieved, the
minimum HPCR1 energy footprint would be no better than
that of standard single stage RO (Eq. 4). Another critical draw-
back of conventional HPCR (Figure 9b), is that, because the
feed pump (P1) is placed in the raw feed line (o), any changes
in Y
S
(while keeping productivity target constant) will change
the pump flow rate, which can shift the pump operation away
from its best efficiency point and thus increase system energy
footprint. Thus, increased operational flexibility in conven-
tional RO with HPCR may be at the cost of reduced energy
efficiency, depending on variation of the pump efficiency with
the feed water flow rate.
HPCR operation of RO in unsteady-state cyclic operational
modes have recently been introduced commercially (i.e.,
closed circuit desalination).
81,93
In this approach, the RO sys-
tem is operated in alternating periods of complete and no con-
centrate recirculation in a cyclical manner. In each cycle, salt
buildup in the RO system during the period of complete con-
centrate recirculation is purged during subsequent period of
concentrate discharge. In addition to enhanced operational
flexibility, recent theoretical analysis under idealized plug-
flow conditions suggests that for a single stage RO, the
approach can potentially be utilized as an alternative to using
an energy recovery device.
94,95
Nevertheless, existing analysis
have yet to consider the impact of non-ideal flow conditions in
real RO systems, especially during the concentrate withdrawal
period. Hence, the practicality of the approach with respect to
energy consumption still needs to be explored.
Figure 9. Conventional single-stage reverse osmosis (a)
with low (LPCR1) and (b) high pressure (HPCR1)
concentrate recirculation. MA-membrane array,
FD- flow diverter, P1-main feed pump, P2: concen-
trate recirculation pump, TV-throttling valve (for
back pressure regulation); o, f, p, c, rc, o, and
w—source water, and RO unit feed, permeate,
concentrate, recycled concentrate, and discharge
streams, respectively.
AIChE Journal June 2017 Vol. 63, No. 6 Published on behalf of the AIChE DOI 10.1002/aic 1779

The Path to Improving RO Desalination
In the forthcoming years improvements in the design and
operation of RO plants will lead to further reduction in the
cost of RO desalination. For example, more effective process
configurations could allow operational flexibility (i.e., wide
recovery range) with efficient energy recovery.
83–85
Improve-
ments in membrane element design
96,97
could also be benefi-
cial to reducing pressure losses while possibly enabling
reduction in concentration polarization, thereby allowing
higher recovery operation per element and thus decreasing the
plant footprint and possibly plant components. Operational
considerations include optimal plant control that will ensure
energy optimal operation while meeting the target permeate
productivity.
49,50
In recent years there have also been mounting efforts to
develop effective methods of plant fault detection and isola-
tion
49,91,98–100
so as to avoid plant failures and reduce mainte-
nance costs. Effective self-adaptive RO feed pretreatment
48,51
that can handle temporal variability of raw feed water quality
and production demand is also likely to lead to significant
reduction in the use of feed treatment chemicals while reduc-
ing the frequency of membrane cleaning and replacement. All
of the above would benefit from real time monitoring of mem-
brane fouling and mineral scaling.
44,51,77,101–103
It is also
stressed that real-time monitoring of feed and product water
quality (e.g., w.r.t emerging contaminants) and of membrane
integrity
104,105
would contribute to establishing effective oper-
ational strategies to mitigate fouling/scaling and toward gain-
ing acceptability of RO for direct potable reuse applications.
While efforts to increase membrane permeability are con-
tinuing with claims of this approach as being a viable path to
achieving significant reduction in RO energy consumption,
studies about a decade ago have shown that this is unlikely to
be realized (section “RO energy consumption and the thermo-
dynamic limit”). The classical polyamide membranes appear
to be well entrenched in the RO desalination industry.
106
How-
ever, the development of promising membranes,
107,108
such as
those based on graphene, carbon nanotubes, aquaporins, self-
organizing block copolymers, crystalline polymers, as well as
molecularly printed and nanostructured,
109,110
and surface
modified membranes,
103
could pave the way for commercial
RO elements capable of high flux (and thus high recovery per
element) operation with high membrane selectivity. These
future membranes, if successful, could allow significant reduc-
tion in plant footprint. However, such membrane will need to
match or be better than current membranes with respect to
their fouling propensity. Moreover, the development of such
membranes will also have to ensure their mechanical and
chemical (e.g., due to exposure to oxidants, extreme pH condi-
tions) stability.
Summary
Water scarcity and groundwater contamination around the
globe have sparked major efforts to preserve and diversify
regional water portfolios through water desalination and water
reuse. While various desalination technologies exist, RO
membrane desalination has emerged as the dominant technol-
ogy for desalting of seawater and BW in applications that
range from small- to municipal-scale applications. The market
growth of RO desalination is attributed, in part, to its simplic-
ity and technological maturity. While it has been often
reported that RO desalination is energy intensive such state-
ments are often misleading as examples already exist where
RO desalination of both brackish and seawater are competitive
with traditional water sources. Admittedly, the energy cost of
water desalination is typically the major cost contributor to
water production via seawater desalination, but this is not
always the case (e.g., capital cost can be higher in some cases).
In the case of BW desalination, energy (while being a signifi-
cant contributor) is often not the major cost component. While
the overall cost for seawater RO desalination has been
reported to be as low as about $0.53/m
3
, there is a wider range
of reported costs for both seawater and BW desalination. Cost
variability is due to numerous factors including, but not lim-
ited to, differences in plant design and location, equipment,
operation (e.g., feed salinity and quality, recovery level, auto-
mation, and real time process optimization), and environmen-
tal compliance costs (e.g., regulatory requirements for inflow
and outflow safeguards).
It is questionable if the energy cost of desalination can be sig-
nificantly reduced in the coming years given the current avail-
ability of high permeability RO membranes which should
enable operation that approaches the cross-flow thermodynamic
restriction. Higher permeability membranes of low fouling pro-
pensity and high resistance to disinfectants should, however,
allow reduction in plant footprint and increased in membrane
lifetime. Moreover, cost reduction can also be achieved through
advances in RO operational flexibility, energy optimal plant
control and self-adaptive operation, real-time monitoring of
membrane fouling and integrity, smaller plant footprint, mem-
brane element design improvements, and higher efficiency
pumps and energy recovery devices. It is stressed that all of the
above are inextricably linked whereby an improvement in one
area is also likely to impact the others. Although there are vari-
ous proposed alternatives to RO desalination, to date, none have
been shown in field studies to be superior to RO desalination
for large-scale applications of potable water production. How-
ever, it is conceivable that future developments will drive the
competition which will ultimately lead to expansion and
broader acceptability of desalination technologies as an impor-
tant step toward water sustainability.
References
1. Maupin MA, Kenny JF, Hutson SS, Lovelace JK, Barber
NL, Linsey KS. Estimated Use of Water in the United
States in 2010. Reston, VA: U.S. Geological Survey Cir-
cular, 1405, 2014.
2. Global Water Intelligence. Municipal Water Reuse Mar-
kets 2010. Oxford, England: Global Water Intelligence,
2010.
3. Kelly TS, Michael EW. Evaluating the energy consumed
for water use in the United States. Environ Res Lett.
2012;7:034034.
4. American Society of Civil Engineers. 2009 Report Card
for America’s Infrastructure. Washington, DC: ASCE,
2009. Available from www.asce.org/reportcard.
5. Cook BI, Ault TR, Smerdon JE. Unprecedented 21st
century drought risk in the American Southwest and
Central Plains. Sci Adv. 2015;1:e1400082.
1780 DOI 10.1002/aic Published on behalf of the AIChE June 2017 Vol. 63, No. 6 AIChE Journal

6. Committee on Future Options for Management in the
Nation’s Subsurface Remediation Effort, National
Research Council. Alternatives for Managing the
Nation’s Complex Contaminated Groundwater Sites.
Washington, DC: National Academies Press, 2013.
7. Biggs EM, Bruce E, Boruff B, Duncan JMA, Horsley J,
Pauli N, McNeill K, Neef A, Van Ogtrop F, Curnow J,
Haworth B, Duce S, Imanari Y. Sustainable development
and the water–energy–food nexus: a perspective on live-
lihoods. Environ Sci Policy. 2015;54:389–397.
8. DiGiulio DC, Jackson RB. Impact to underground sour-
ces of drinking water and domestic wells from produc-
tion well stimulation and completion practices in the
Pavillion, Wyoming, Field. Environ Sci Technol. 2016;
50:4524–4536.
9. VirgiliPankratz F, Gasson TJ. IDA desalination yearbook
2015-2016. Oxford, UK: Global Water Intell. 2016.
10. Semiat R, Hasson D. Seawater and brackish-water desali-
nation with membrane operations. In: Peinemann K-V,
and Nunez SP, editors. Membrane Technology. Wein-
heim, Germany: Wiley-VCH Verlag GmbH & Co.
KGaA, 2010:147–168.
11. Miller S, Shemer H, Semiat R. Energy and environmen-
tal issues in desalination. Desalination. 2015;366:2–8.
12. Y, Cohen B, McCool A, Rahardianto M-M, Kim J. Faria
Membrane desalination of agricultural drainage water.
In: Chang AC, Brawer Silva D, editors. Salinity and
Drainage in San Joaquin Valley, California. Nether-
lands: Springer, 2014:303–341.
13. U.S. National Archives and Records Administration. Sec-
ondary Maximum Contaminant Levels, Code of Federal
Regulations, Title 40, Section § 143.3, 2016.
14. Watson IC, Morin OJ Jr, Henthorne L. Desalting Hand-
book for Planners, 3rd ed. Desalination and Water Puri-
fication Research and Development Program Report No.
72. Denver, CO: U.S. Bureau of Reclamation, 2003.
15. Shiklomanov I. World fresh water resources. In: Gleick
PH, editor. Water in Crisis: A Guide to the World’s
Fresh Water Resources. New York: Oxford University
Press, 1993.
16. Gray S, Semiat R, Duke M, Rahardianto A, Cohen Y.
Seawater use and desalination technology, treatise on
water science. Water Qual Eng. 2011;4:73–109.
17. Semiat R. Desalination - Present and future. Water Int.
2000;25:54–65.
18. G€
uler E, Kaya C, Kabay N, Arda M. Boron removal
from seawater: state-of-the-art review. Desalination.
2015;356:85–93.
19. Venkatesan AK, Ahmad S, Johnson W, Batista JR.
Salinity reduction and energy conservation in direct and
indirect potable water reuse. Desalination. 2011;272:
120–127.
20. Gabelich CJ, Xu P, Cohen Y. Concentrate treatment for
inland desalting. In: CE Isabel, IS Andrea, editors. Sus-
tainability Science and Engineering, Chapter 10. Amster-
dam, The Netherlands: Elsevier, 2010, 295–326.
21. Shenvi SS, Isloor AM, Ismail AF. A review on RO
membrane technology: developments and challenges.
Desalination. 2015;368:10–26.
22. Amy G, Ghaffour N, Li Z, Francis L, Linares RV,
Missimer T, Lattemann S. Membrane-based seawater
desalination: present and future prospects. Desalination.
2017;401:16–21.
23. Ghaffour N, Missimer TM, Amy GL. Technical review
and evaluation of the economics of water desalination:
current and future challenges for better water supply sus-
tainability. Desalination. 2013;309:197–207.
24. Henthorne L, Boysen B. State-of-the-art of reverse osmo-
sis desalination pretreatment. Desalination. 2015;356:
129–139.
25. Schwantes R, Cipollina A, Gross F, Koschikowski J,
Pfeifle D, Rolletschek M, Subiela V. Membrane distilla-
tion: solar and waste heat driven demonstration plants
for desalination. Desalination. 2013;323:93–106.
26. Coday BD, Xu P, Beaudry EG, Herron J, Lampi K,
Hancock NT, Cath TY. The sweet spot of forward osmo-
sis: treatment of produced water, drilling wastewater,
and other complex and difficult liquid streams. Desalina-
tion. 2014;333:23–35.
27. Semiat R, Sapoznik J, Hasson D. Energy aspects in
osmotic processes. Desalin Water Treat. 2010;15:228–
235.
28. Sagiv A, Zhu A, Christofides PD, Cohen Y, Semiat R.
Analysis of forward osmosis desalination via two-
dimensional FEM model. J Memb Sci. 2014;464:
161–172.
29. Ziolkowska JR, Reyes R. Impact of socio-economic
growth on desalination in the US. J Environ Manag.
2016;167:15–22.
30. Mickley MJ, Jordahl J, Arakel A. Development of a
knowledge base for desalination concentrate and salt
management. Alexandria, VA: Water Reuse Research
Foundation. 2011.
31. Semiat R. Energy issues in desalination processes. Envi-
ron Sci Technol. 2008;42:8193–8201.
32. Voutchkov N, Semiat R. Seawater desalination. In:
Advanced Membrane Technology and Applications.
Hoboken, NJ: Wiley, 2008:47–86.
33. Camacho LM, Dum
ee L, Zhang J, Li J, Duke M, Gomez
J, Gray SR. Advances in membrane distillation for water
desalination and purification applications. Water. 2013;5:
94–196.
34. Tavakkoli S, Lokare OR, Vidic RD, Khanna V. Systems-
level analysis of waste heat recovery opportunities from
natural gas compressor stations in the United States. ACS
Sustain Chem Eng. 2016;4:3618–3626.
35. Gingerich DB, Mauter MS. Quantity, quality, and avail-
ability of waste heat from United States thermal power
generation. Environ Sci Technol. 2015;49:8297–8306.
36. P, Rao A, Aghajanzadeh P, Sheaffer I, William R,
Morrow S, Brueske C, Dollinger K, Price P, Sarker N,
Ward J. Cresko. Volume 1: Survey of Available Informa-
tion in Support of the Energy-Water Bandwidth Study of
Desalination Systems, Energy Technology Area, LBNL-
1006424, 2016.
37. Rahardianto A, Shih W-Y, Lee R-W, Cohen Y. Diagnos-
tic characterization of gypsum scale formation and con-
trol in RO membrane desalination of brackish water.
J Membr Sci. 2006;279:655–668.
AIChE Journal June 2017 Vol. 63, No. 6 Published on behalf of the AIChE DOI 10.1002/aic 1781

38. Shih W-Y, Rahardianto A, Lee R-W, Cohen Y. Morpho-
metric characterization of calcium sulfate dihydrate (gyp-
sum) scale on reverse osmosis membranes. J Membr Sci.
2005;252:253–263.
39. Greenberg G, Hasson D, Semiat R. Limits of RO recov-
ery imposed by calcium phosphate precipitation. Desali-
nation. 2005;183:273–288.
40. Hasson D, Drak A, Semiat R. Inception of CaSO4 scal-
ing on RO membranes at various water recovery levels.
Desalination. 2001;139:73–81.
41. Hasson D, Semiat R, Bramson D, Busch M, Limoni-
Relis B. Suppression of CaCO3 scale deposition by anti-
scalants. Desalination. 1998;118:285–296.
42. Semiat R, Sutzkover I, Hasson D. Technique for evaluat-
ing silica scaling and its inhibition in RO desalting.
Desalination. 2001;140:181–193.
43. Thompson J, Rahardianto A, Gu H, Uchymiak M,
Bartman A, Hedrick M, Lara D, Cooper J, Faria J,
Christofides PD. Rapid field assessment of RO desalina-
tion of brackish agricultural drainage water. Water Res.
2013;47:2649–2660.
44. Thompson J, Rahardianto A, Kim S, Bilal M,
Breckenridge R, Cohen Y. Real-time direct detection of
silica scaling on RO membranes. J Membr Sci. 2017;
528:346–358.
45. Thompson J, Lin N, Lyster E, Arbel R, Knoell T, Gilron
J, Cohen Y. RO membrane mineral scaling in the pres-
ence of a biofilm. J Membr. Sci. 2012;415–416:181–191.
46. Shemer H, Hasson D, Semiat R. Review of the state of
the art of antiscalant selection. In: Zahid Amjad, editor.
Mineral Scales in Biological and Industrial Systems.Boca
Raton, FL: CRC Press, 2013:227–256.
47. OCWD, OCSD, Groundwater Replenishment System
Technical Brochure. Orange County Water District and
Orange County Sanitation District. Fountain Valley, Cal-
ifornia, 2016.
48. Gao LX, Rahardianto A, Gu H, Christofides PD, Cohen
Y. Novel design and operational control of integrated
ultrafiltration—reverse osmosis system with RO concen-
trate backwash. Desalination. 2016;382:43–52.
49. Gao L, Rahardianto A, Gu H, Christofides PD, Cohen Y.
Energy-optimal control of RO desalination. Ind Eng
Chem Res. 2014;53:7409–7420.
50. Gao LX, Gu H, Rahardianto A, Christofides PD, Cohen
Y. Self-adaptive cycle-to-cycle control of in-line coagu-
lant dosing in ultrafiltration for pre-treatment of reverse
osmosis feed water. Desalination. 2017;401:22–31.
51. Gu H, Rahardianto A, Gao LX, Christofides PD, Cohen
Y. Ultrafiltration with self-generated RO concentrate
pulse backwash in a novel integrated seawater desalina-
tion UF-RO system. J Membr Sci. 2016;520:111–119.
52. Surawanvijit S, Rahardianto A, Cohen Y. An Integrated
approach for characterization of polyamide reverse
osmosis membrane degradation due to exposure to free
chlorine. J Membrane Sci. 2016;510:164–173.
53. Gu J-E, Jun B-M, Kwon Y-N. Effect of chlorination con-
dition and permeability of chlorine species on the chlori-
nation of a polyamide membrane. Water Res. 2012;46:
5389–5400.
54. Rahardianto A, McCool BC, Cohen Y. Reverse osmosis
desalting of inland brackish water of high gypsum scal-
ing propensity: kinetics and mitigation of membrane
mineral scaling. Environ Sci Technol. 2008;42:4292–
4297.
55. Gu H, Bartman AR, Uchymiak M, Christofides PD, Cohen
Y. Self-adaptive feed flow reversal operation of reverse
osmosis desalination. Desalination. 2013;308:63–72.
56. Sagiv A, Semiat R. Backwash of RO spiral wound mem-
branes. Desalination. 2005;179:1–9.
57. Segev R, Hasson D, Semiat R. Improved high recovery
brackish water desalination process based on fluidized
bed air stripping. Desalination. 2011;281:75–79.
58. Gabelich CJ, Rahardianto A, Northrup CR, Yun TI,
Cohen Y. Process evaluation of intermediate chemical
demineralization for water recovery enhancement in
production-scale brackish water desalting. Desalination.
2011;272:36–45.
59. Rahardianto A, McCool BC, Cohen Y. Accelerated desu-
persaturation of reverse osmosis concentrate by
chemically-enhanced seeded precipitation. Desalination.
2010;264:256–267.
60. Freger V, Shemer H, Sagiv A, Semiat R. Chapter 8 -
Boron removal using membranes A2 - Kabay, Nalan. In:
Bryjak M, Hilal N, editors. Boron Separation Processes.
Amsterdam, Netherlands: Elsevier, 2015:199–217.
61. Semiat R, Sutzkover I, Hasson D. Scaling of RO mem-
branes from silica supersaturated solutions. Desalination.
2003;157:169–191.
62. Sagiv A, Semiat R. Analysis of parameters affecting
boron permeation through reverse osmosis membranes.
J Membrane Sci. 2004;243:79–87.
63. Lisitsin D, Hasson D, Semiat R. The potential of CO
2
stripping for pretreating brackish and wastewater desali-
nation feeds. Desalination. 2008;222:50–58.
64. Shemer H, Hasson D, Semiat R. State-of-the-art review
on post-treatment technologies. Desalination. 2015;356:
285–293.
65. Xu P, Cath TY, Robertson AP, Reinhard M, Leckie JO,
Drewes JE. Critical review of desalination concentrate
management, treatment and beneficial use. Environ Eng
Sci. 2013;30:502–514.
66. P
erez-Gonz
alez A, Urtiaga AM, Ib
a~
nez R, Ortiz I. State
of the art and review on the treatment technologies of
water reverse osmosis concentrates. Water Res. 2012;46:
267–283.
67. Giwa A, Dufour V, Al Marzooqi F, Al Kaabi M, Hasan
SW. Brine management methods: recent innovations and
current status. Desalination. 2017;407:1–23.
68. Morillo J, Usero J, Rosado D, El Bakouri H, Riaza A,
Bernaola F-J. Comparative study of brine management
technologies for desalination plants. Desalination. 2014;
336:32–49.
69. Leong J, Tan J, Charrois J, Ladewig BP. Review of high
recovery concentrate management options. Desalin
Water Treat. 2014;52:7609–7627.
70. Semiat R, Hasson D, Zelmanov G, Hemom I. Threshold
scaling limits of RO concentrates flowing in a long waste
disposal pipeline. Water Sci Technol. 2004;49:211–219.
1782 DOI 10.1002/aic Published on behalf of the AIChE June 2017 Vol. 63, No. 6 AIChE Journal

71. Mickley MC. Membrane Concentrate Disposal: Practi-
ces and Regulation. Desalination and Water Purification
Research and Development Program Report No. 123.
Denver, CO: U.S. Department of the Interior, Bureau of
Reclamation, 2006.
72. Segev R, Hasson D, Semiat R. Modeling CaCO
3
precipi-
tation in fluidized bed CO
2
stripping desalination pro-
cess. Desalination. 2013;311:192–197.
73. Spiegler KS, Kedem O. Thermodynamics of hyperfiltra-
tion (reverse osmosis): criteria for efficient membranes.
Desalination. 1966;1:311–326.
74. Lyster E, Au J, Rallo R, Giralt F, Cohen Y. Coupled 3-
D hydrodynamics and mass transfer analysis of mineral
scaling-induced flux decline in a laboratory plate-and-
frame reverse osmosis membrane module. J Membrane
Sci. 2009;339:39–48.
75. Lyster E, Cohen Y. Numerical study of concentration
polarization in a rectangular reverse osmosis membrane
channel: permeate flux variation and hydrodynamic end
effects. J Membrane Sci. 2007;303:140–153.
76. Sutzkover I, Hasson D, Semiat R. Simple technique for
measuring the concentration polarization level in a reverse
osmosis system. Desalination. 2000;131:117–127.
77. Antony A, Low JH, Gray S, Childress AE, Le-Clech P,
Leslie G. Scale formation and control in high pressure
membrane water treatment systems: a review.
J Membrane Sci. 2011;383:1–16.
78. Zhu A, Rahardianto A, Christofides PD, Cohen Y.
Reverse osmosis desalination with high permeability
membranes—cost optimization and research needs.
Desalin Water Treat. 2010;15:256–266.
79. Huehmer R, Alt S, Lozier J, Kupp A, Nading T, Egrican
K, Emerson B. Evaluation and optimization of emerging
and existing energy recovery devices for desalination
and wastewater membrane treatment plants. Project No.
WRF-08-14. Alexandria, VA: Water Reuse Research
Foundation, 2014.
80. Dow Water & Process Solutions. FILMTEC reverse
osmosis membranes technical manual, Form No. 609-
00071-1009. Available from http://dowac.custhelp.com/
app/answers/detail/a_id/3428, 2014.
81. Efraty A. Closed circuit desalination series no-3: high
recovery low energy desalination of brackish water by a
new two-mode consecutive sequential method. Desalin
Water Treat. 2012;42:256–261.
82. Stikker A. Market Profile, Desalination Mark. Oxford,
UK: Global Water Intelligence, 2006.
83. Zhu A, Christofides PD, Cohen Y. Effect of thermody-
namic restriction on energy cost optimization of RO
membrane water desalination. Ind Eng Chem Res.in
Press, 2009.
84. Zhu A, Christofides PD, Cohen Y. Minimization of
energy consumption for a two-pass membrane desalina-
tion: effect of energy recovery, membrane rejection and
retentate recycling. J Membrane Sci. 2009;339:126–137.
85. Zhu A, Christofides PD, Cohen Y. Effect of stream mix-
ing on RO energy cost minimization. Desalination.
2010;261:232–239.
86. McCool BC, Rahardianto A, Faria JI, Cohen Y. Evalua-
tion of chemically-enhanced seeded precipitation of RO
concentrate for high recovery desalting of high salinity
brackish water. Desalination. 2013;317:116–126.
87. Spiegler KS, El-Sayed YM. The energetics of desalina-
tion processes. Desalination. 2001;134:109–128.
88. Spiegler KS, El-Sayed YM. A Desalination Primer:
Introductory Book for Students and Newcomers to Desa-
lination. Santa Maria Imbaro, Italy: Balaban Desalination
Publications, 1994.
89. Water Science and Technology Board. Review of the
Desalination and Water Purification Technology Road-
map. Washington, DC: National Academies Press, 2004.
90. Zhu A, Christofides PD, Cohen Y. On RO membrane
and energy costs and associated incentives for future
enhancements of membrane permeability. J Membr Sci.
2009;344:1–5.
91. Bartman AR, Zhu A, Christofides PD, Cohen Y. Mini-
mizing energy consumption in reverse osmosis mem-
brane desalination using optimization-based control.
J Process Control. 2010;20:1261–1269.
92. Li M. Energy consumption in spiral-wound seawater
reverse osmosis at the thermodynamic limit. Ind Eng
Chem Res. 2014;53:3293–3299.
93. Efraty A, Barak RN, Gal Z. Closed circuit desalina-
tion—a new low energy high recovery technology with-
out energy recovery. Desalin Water Treat. 2011;31:95–
101.
94. Qiu T, Davies PA. Comparison of configurations for
high-recovery inland desalination systems. Water. 2012;
4:690.
95. Warsinger DM, Tow EW, Nayar KG, Maswadeh LA,
Lienhard V JH. Energy efficiency of batch and semi-
batch (CCRO) reverse osmosis desalination. Water Res.
2016;106:272–282.
96. Karabelas AJ, Kostoglou M, Koutsou CP. Modeling of
spiral wound membrane desalination modules and plants
– review and research priorities. Desalination. 2015;356:
165–186.
97. Karabelas AJ, Koutsou CP, Kostoglou M. The effect of
spiral wound membrane element design characteristics
on its performance in steady state desalination—a para-
metric study. Desalination. 2014;332:76–90.
98. Pascual X, Gu H, Bartman A, Zhu A, Rahardianto A,
Giralt J, Rallo R, Christofides PD, Cohen Y. Fault detec-
tion and isolation in a spiral-wound reverse osmosis
(RO) desalination plant. Ind Eng Chem Res. 2014;53:
3257–3271.
99. Garcia-Alvarez D, Fuente MJ. A UPCA-based monitor-
ing and fault detection approach for reverse osmosis
desalination plants. Desalin Water Treat. 2014;52:1272–
1286.
100. McFall CW, Bartman A, Christofides PD, Cohen Y.
Control and monitoring of a high recovery reverse
osmosis desalination process. Ind Eng Chem Res. 2008;
47:6698–6710.
101. Ahmed F, Lalia BS, Kochkodan V, Hilal N, Hashaikeh
R. Electrically conductive polymeric membranes for
fouling prevention and detection: a review. Desalina-
tion. 2016;391:1–15.
102. Uchymiak M, Rahardianto A, Lyster E, Glater J, Cohen
Y. A novel RO ex situ scale observation detector
AIChE Journal June 2017 Vol. 63, No. 6 Published on behalf of the AIChE DOI 10.1002/aic 1783

(EXSOD) for mineral scale characterization and early
detection. J Membr Sci. 2007;291:86–95.
103. Miller DJ, Dreyer DR, Bielawski CW, Paul DR,
Freeman BD. Surface modification of water purification
membranes. Angew.Chem. Int.Ed. 2017;56:4662–4711.
104. Pype M-L, Lawrence MG, Keller J, Gernjak W.
Reverse osmosis integrity monitoring in water reuse:
the challenge to verify virus removal – a review. Water
Res. 2016;98:384–395.
105. Surawanvijit S, Thompson J, Rahardianto A, Frenkel V,
Cohen Y. Pulsed marker method for real-time detection
of reverse osmosis membrane integrity loss. Desalina-
tion. 2015;370:25–32.
106. Gohil JM, Ray P. A review on semi-aromatic polyam-
ide TFC membranes prepared by interfacial polymeriza-
tion: potential for water treatment and desalination.
Separation and Purification Technol. 2017;181:159–
182.
107. Werber JR, Osuji CO, Elimelech M. Materials for next-
generation desalination and water purification mem-
branes. Nat Rev Mater. 2016;1:16018.
108. Jiang Y, Biswas P, Fortner JD. A review of recent
developments in graphene-enabled membranes for water
treatment. Environ Sci Water Res Technol. 2016;2:915–
922.
109. Cohen Y, Lin N, Varin KJ, Chien D, Hicks RF. Mem-
brane surface nanostructuring with terminally anchored
polymer chains. In: Duke M., Zhou D., Semiat R., and
Lu M. Functional Nanostructured Materials and Mem-
branes for Water Treatment. Weinheim, Germany:
Wiley-VCH Verlag GmbH & Co. KGaA, 2013:85–124.
110. Jhaveri JH, Murthy ZVP. A comprehensive review on
anti-fouling nanocomposite membranes for pressure
driven membrane separation processes. Desalination.
2016;379:137–154.
1784 DOI 10.1002/aic Published on behalf of the AIChE June 2017 Vol. 63, No. 6 AIChE Journal