Journal of Membrane and Separation Technology, 2012, 1, 79-93 79

E-ISSN: 1929-6037/12 2012 Lifescience Global

Forward Osmosis: Potential use in Desalination and Water Reuse
Ali Altaee
*

School of Engineering University of the West of Scotland Paisley PA1 2BE, the UK
Abstract: There has been a recurring interest in using Forward Osmosis (FO) process in water treatment and
desalination. Despite the promising results from pilot and bench scale experiments the technology is still not
commercialized yet. This is due to the complicated nature of the process which usually involves multiple stages of
treatment in addition to the FO membrane process. Unfortunately, most of the recent studies were focused on studying
the FO process alone and didnt provide enough data about the actual cost of the process as whole which includes the
osmotic agent regeneration stage/s. This issue resulted in some uncertainties about the total cost of the water treatment
by the process. Furthermore, more data are required to evaluate the impact of the osmotic agent losses on the overall
cost and efficiency. In case if the draw solution is regenerated by membrane treatment, a suitable membrane should be
selected to ensure an optimal salt rejection. For power generation by Pressure Retarded Osmosis (PRO) process, there
was an evident progress. However, the process is site specific; i.e. it is dependent of the availability of the draw and
donor solution. This suggested that the process is applicable to certain areas but cant be generalized.
Keywords: Please provide keywords.
1. INTRODUCTION
Due to the increasing demands on fresh water
supply and contamination of ground waters,
desalination becomes the choice option for water
supply in water shortage areas. Thermal and
membrane technologies are the forefront processes for
seawater desalination. Multi Stage Flashing (MSF) and
Multi Effect Distillation (MED) are the leading
processes in thermal technologies while Reverse
Osmosis (RO) dominates the membrane desalination
technologies [1-3]. The thermal processes are widely
used in the Gulf region of the Middle East due the
difficult nature of gulf water such as high salinity and
concentrations of impurities. In the last decade, this
trend has been changed in favor of the RO process
due to the development of high performance RO
membranes for seawater desalination. Worldwide, RO
process is the most common technology for seawater
desalination [3, 4]. Mainly, this is because of its
reliability and lower power consumption compared to
the thermal technologies [4, 5]. The cost of water
desalination by the RO process is less than USD 1
$/m
3
[6, 7]. Unfortunately, the cost of RO desalination
is still unaffordable to many countries.
As a result, scientists and researchers are
investigating cheaper processes for seawater
desalination such as FO [6-8]. The attractiveness of the
FO process consists in its low power consumption,
easy to scale up, and potential high recovery rate [6, 8,
9]. The operating cost of FO process is much lower

*Address correspondence to this author at the School of Engineering University
of the West of Scotland Paisley PA1 2BE, the UK, Tel +971507356657;
Email: alialtaee@hotmail.com
than RO and thermal processes. In principle, FO relies
on the for water transport across semi permeable
membrane. Freshwater transports across the semi-
permeable membrane from the low to the high
concentration solution due to the osmotic pressure
gradient. A number of chemical compounds have been
used as a draw solution in FO process such as table
salt, magnesium sulfate, glucose, ammonia carbon
dioxide, and magnesium chloride [9, 10]. In practice,
draw solution is recycled and reused to reduce the cost
of desalination. This is typically achieved either by
thermal or membrane filtration processes.
Previous studies have shown there is a number of
physical and chemical factors which affect the
efficiency and cost of the FO process. Amongst these
factors are; salt diffusion from the seawater to the draw
solution, osmotic agent diffusion across the membrane,
and concentration polarization. Because of the adverse
impact of these factors, the actual performance of FO
is significantly lower than the theoretical performance
[11]. Yet, there isnt enough information about the
overall effect of salt diffusion on the product water
quality and the desalination cost. Although osmotic
agent can be regenerated and reused, none of the
aforementioned regeneration processes are capable of
completely recycling the osmotic agent without losses.
Depending on the regeneration process used for
osmotic agent recycling, salt diffusion will affect the
quality and cost of the product water. For example, the
membrane processes is affected by the type of the
osmotic agent and membrane used in the regeneration
process. Different types of membranes were used such
as RO, Nanofiltration (NF), and membrane distillation
(MD).
80 Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 Ali Altaee
In addition to seawater desalination, FO was
proposed for power generation; the technique is known
as Pressure Retarded Osmosis (PRO) [9, 10, 12]. It
was first suggested by Loeb [13] and there is a wealth
of literature about describing its potential and
application in power generation. The mechanism of
PRO operation is similar to that explained above but
after leaving the FO membrane, preferably pressurized,
the draw solution is fed to a turbine system to convert
the hydraulic energy to an electrical power. The
application of PRO in power generation was limited by
the membranes characteristics and efficiency until the
breakthrough made by StatKraft which built the world
first pilot plant for power generation in Norway [14]. The
pilot plant used freshwater as donor solution and
seawater as a draw solution. The optimal performance
of PRO can be achieved when fresh water is used as a
donor solution; this will eliminate the problem of internal
concentrative concentration polarization. StatKraft plant
is the largest in the world which demonstrated the
feasible scalability of PRO from bench to pilot scale. To
date, there is no data about the economical feasibility
of the PRO in power generation. It should be stressed
here that the type and concentration of the osmotic
agent have a significant impact on the performance of
FO process. This paper pinpointed the major
drawbacks in FO applications and commercialization.
The paper also submitted a critical review from
engineering perspective with regard to FO process
optimization.
2. FO FOR SEAWATER DESALINATION
Desalination is an essential process for fresh water
supply in water stressed area. The cost of desalination
is still high and need to be reduced especially for high
salinity feed waters. FO, therefore, received a lot of
attention in the past decade of being a competitive
process for seawater desalination compared to the
conventional processes. There are a number of
technical and operating factors which affect the
efficiency of the process. Technical factors such as
internal and external concentration polarizations were
found to have detrimental impact on the membrane flux
and the recovery rate. The phenomenon of concen-
tration polarization is associated with the membrane
characteristics and it is mainly due to the salt
concentration at the membrane surface relative to bulk
solution [9, 11]. Depending on the flow mode of the
draw and feed solutions, concentration polarization is
classified into internal and external effects [11]. The
internal effect occurs at the lumen or at the support
layer side of the membrane while the external occurs
on the membrane active layer [11, 15]. Concentrative
concentration polarization usually refers to the change
in the concentration of feed solution and it is due to the
increase in the concentration of feed solution at the
membrane surface leading to a decrease in the
osmotic pressure gradients across the membrane [11,
15]. As a result, fresh water flow across the membrane
declines with time. On contrast, the dilutive concen-
tration polarization is associated with the draw solution
which results in a reduction in the concentration of
draw solution. In response to this effect the osmotic
pressure gradient is decreased as well as water flux
across the membrane. The recent advance in the
membrane manufacturing technology successfully
minimized the internal concentration polarization
problem. This was achieved by reducing the thickness
of the membrane support layer [16]. However, the
effect of external concentration polarization is an
inherent nature of the FO membrane process and cant
be avoided.
The effect of internal concentration polarization was
found to be more serious than the external
concentration polarization (Figure 1) [7]. This finding is
especially important when the FO membrane is
operated in the RO mode or the draw solution in the
lumen side of the membrane. Pilot and bench scale
experiments showed a sharp drop in the membrane
flux under the effect of internal concentration
polarization. Although the new FO membranes have
successfully reduced the effect of internal dilutive and
concentrative concentration polarization, it cant be
completely eliminated. Therefore, the efficiency of FO
process will still be affected by the internal concen-
tration polarization phenomenon. The new experi-
mental work showed that FO process operates better in
inside to outside flow mode or when the draw solution
is in the shell side and the feed solution in the lumen
side. Basically, this is due to the higher membrane flux
in the latter operation mode. But, it should be noted
here this fact holds true when the FO process is
applied for seawater desalination in the conventional
way. In some cases, as will be discussed later, it is
preferred to have the draw solution in the lumen not to
improve the membrane flux but to reduce the
membrane fouling propensity [15]. This is especially
true if the donor solution has a high concentration of
fouling agents such as wastewater effluent.
It is highly desirable to achieve a high recovery rate
in the desalination process. At the present, the
recovery rate in RO seawater desalination is less than
50% while it is about 30% in the thermal desalination
Forward Osmosis Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 81
processes (MSF and MED) [4, 8, 17, 18]. Osmotic
pressure gradient is the main mechanism by which
fresh water transports across the membrane in the FO
process. For a given feed water/seawater TDS, the
recovery rate can be increased by increasing the
concentration of draw solution. However, the
experimental work demonstrated that the relationship
between the concentration of draw solution and the
recovery rate is not linear (Figure 2). Water flow across
the membrane dilutes the draw solution at the
membrane surface forming a micro boundary layer
adjacent to the membrane surface. The concentration
at the boundary layer is lower than that in the bulk
solution which results in reducing the driving force of
water transport from the feed to the draw solution. As
shown in Figure 2, the increase in the osmotic pressure
gradient resulted in a proportional increase in the
recovery rate. Although the relationship was linear at
low osmotic pressure gradient, it is changed at high
osmotic pressure gradients. Primarily, this is due to the
concentration polarization effect at the membrane-
solution interface. It should be noted here that the
energy requirement for regeneration and fresh water
extraction is increased with increasing the concen-
tration of draw solution. This is because of the higher
concentration of the diluted draw solution has to be
regenerated (Figure 2). Accordingly, a subsequent
increase in the concentration of draw solution wouldnt
necessarily lead to the desirable improvement in the
performance of FO process.
The regeneration process of draw solution is the
most expensive stage in the FO desalination process.
Conversely to the FO process, which is driven by the
natural osmosis phenomenon, the regeneration
process is relatively more power intensive. Typical
osmotic pressure of the diluted draw solution is equal
or higher than the osmotic pressure of the feed
solution. Both membrane and thermal processes were
proposed for freshwater extraction and draw solution
regeneration [9, 10]. NF, RO, and MD membranes
were proposed for the regeneration of draw solution
and fresh water extraction [9, 12]. Each type of mem-
branes has its own advantages and disadvantages
which need to be considered upon choosing a suitable
membrane for the regeneration process. Thermal
processes, otherwise, are used for the regeneration of
draw solution [20].
2.1. Membrane Processes for Draw Sol uti on
Regenerati on
The experimental work demonstrated the
applicability of membrane processes such as NF, RO
and MD for the draw solution regeneration and

Figure 1: Concentration polarization in FO membrane (a) A symmetric membrane (b) An asymmetric membrane draw solution
facing the active layer (c) An asymmetric membrane feed solution facing the active layer (J effery et al., 2006).
82 Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 Ali Altaee
concentration. The application of membrane process is
affected by:
1. Type of the osmotic agent used in the draw
solution
2. The concentration of draw solution
3. The concentration of the feed water
4. Type of the membrane
For instance, the permeability of NF membranes is
higher than RO membranes but they exhibit a lower
rejection rate compared to the RO membranes. NF
membranes, therefore, are more suitable in the
regeneration of low concentration and multivalent
osmotic agent. Commercial NF membranes are
manufactured to tolerate a maximum feed pressure not
exceeding a forty bar. Practically, this pressure is out of
the range required for the treatment a diluted draw
solution for seawater desalination i.e. feed water
osmotic pressure is around 27 bar. Otherwise the
recovery rate of NF membrane, if used in seawater
desalination, will not be economical. Therefore, NF
membranes are suitable for the regeneration of a draw
solution having an osmotic pressure less than seawater
such as brackish water. For simplicity, it is assumed
here that the osmotic pressure of the brackish water to
be treated by FO process is 1.6 bar (feed TDS 2000
ppm NaCl) and the recovery rate is 75%. Ideally, in FO
process the osmotic pressure of the diluted draw
solution will be at least equal to that of the feed solution
concentrate; i.e. 6.5 bar (based on 75% recovery rate).
ROSA software by Filmtec was used in this paper to
estimate the membrane feed pressure for the
regeneration process. Two different draw solutions
were tested here (NaCl and MgSO
4
) to investigate the
effect of draw solution on the membrane performance
and product water quality. The cost of the regeneration
process and product water quality is also affected by
the type of membrane used in the regeneration
process. For comparison purpose, NF and BWRO
membranes were applied for the draw solution
regeneration and concentration. The simulation results
showed that the permeate TDS was lower when
BWRO membrane was used for draw solution
regeneration (Figure 3). The permeate TDS was also
affected by the type of draw solution (Figure 3). For a
given BWRO membrane, a lower permeate
concentration was achieved when NaCl was used as a
draw solution. This is because of the lower
concentration of NaCl was required for the generation
of same osmotic pressure compared to MgSO
4
. The
required concentration of NaCl and MgSO
4
in the draw
solution to generate an osmotic pressure equivalent to
6.5 is 8250 ppm and 29972 ppm respectively. As such,
NaCl is likely to be more efficient osmotic agent than
MgSO
4
if BWRO membrane is used for the draw
solution regeneration. It is clearly shown in Figure 3
that there was a proportional increase in the permeate
TDS with increasing the recovery rate [8, 19]. On the
other hand NF membranes are not suitable for the
regeneration of monovalent osmotic agent due to the
low rejection rate to monovalent ions (Figure 3).
Eventhough BWRO membrane showed a higher
rejection rate to mono and multivalent ions osmotic
agent but this was at the cost of higher energy
consumption (Figure 4). As expected, the specific
energy consumption was higher in case of BWRO
membrane especially when the draw solution was
made of MgSO
4
. This is because:

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates are 21%, 29%, 37% and 39%, Cf 34760
Figure 2: Effect of osmotic pressure on recovery rate and the concentration of diluted draw solution.
Forward Osmosis Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 83
1. MgSO
4
has more affinity to the membrane
surface because of the higher ionic charge
compared to NaCl
2. The higher concentration of MgSO
4
required
for the generation of osmotic pressure
equivalent to that of NaCl Salt. Higher
concentration causes higher concentration
polarization.
3. Higher membrane rejection to MgSO
4

C
w
= C
b
*CP [1]
Where C
w
is the concentration at the membrane
wall, C
b
is the bulk concentration and CP is the
concentration polarization factor.
Regardless the type of osmotic agent, the specific
energy consumption decreased when NF was used
instead of BWRO membrane for the draw solution
regeneration. The specific energy consumption in-
creased with the recovery rate increase which is typi-
cally observed in the membrane filtration processes.
BWRO membranes have a tighter structure than NF
membranes which render them an expensive option for
0
200
400
600
800
1000
1200
1400
1600
1800
60.5 61 61.5 62 62.5 63 63.5 64 64.5 65 65.5
%R
C
p

(
p
p
m
)
MgSO4 DS: NF90-400
MgSO4 DS:BW30-440i
NaCl DS: NF90-400
NaCl DS:BW30-440i

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates are 61%-65%, Cf 34760, Qf 7 m
3
/h, Cf MgSO429971 ppm, CfNaCl 8250 ppm
Figure 3: Effect of recovery rate on the permeate concentration.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
60.5 61 61.5 62 62.5 63 63.5 64 64.5 65 65.5
%R
E
s

(
k
W
h
/
m
3
)
MgSO4 DS: NF90-400
MgSO4 DS: BW30-440i
NaCl DS: NF90-400
NaCl DS: BW30-440i

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates are 61%-65%, Cf 34760, Qf 7 m
3
/h, Cf MgSO429971 ppm, CfNaCl 8250 ppm
Figure 4: Power consumption as a function of recovery rate.
84 Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 Ali Altaee
treating the draw solution because of the high feed
pressure requirement (Figure 5). The simulation results
showed that the feed pressure was higher when
BWRO membrane was used. In case of NF membrane
the power consumption decreased when NaCl was
used as a draw solution because of the lower NF
rejection rate to monovalent ions. As such NaCl is not
recommended as an osmotic agent if NF membranes
are used for the regeneration of the draw solution. The
BWRO membrane required a higher feed pressure for
the filtration of MgSO4 than NaCl (Figure 5). This was
probably attributed to the high rejection rate of MgSO
4

by BWRO membrane and hence the concentrate
pressure was higher than that for NaCl (Figure 6).
Indeed, the concentrate osmotic pressure increased
with increasing the recovery rate in both draw
solutions; i.e. NaCl and MgSO
4
. Therefore, it is
preferable to use NaCl or any monovalent ions of high
osmotic pressure in FO process if tight membranes
such as BWRO/RO are used in the regeneration
process.
For seawater desalination, RO membranes should
be used in conjunction with FO membranes. High
pressure seawater RO membranes are preferable for
water extraction and draw solution concentration.
ROSA was used to demonstrate the applicability of RO
membranes for the draw solution concentration and
regeneration. Three different types of osmotic agents
were investigated; MgSO
4
, MgCl
2
and NaCl and a
number of recovery rates ranged between 42 and 54 at
3% intervals were examined. The results showed that
0
5
10
15
20
25
60.5 61 61.5 62 62.5 63 63.5 64 64.5 65 65.5
%R
P
f

(
b
a
r
)
MgSO4 DS:NF90-400
MgSO4 DS:BW30-440i
NaCl DS:NF90-400
NaCl DS:BW30-440i

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates are 61%-65%, Cf 34760, Qf 7 m
3
/h, Cf MgSO429971 ppm, CfNaCl 8250 ppm
Figure 5: Effect of the recovery rate on the feed pressure for NF and RO membranes.

12
13
14
15
16
17
18
19
20
60.5 61 61.5 62 62.5 63 63.5 64 64.5 65 65.5
%R
_
c

(
b
a
r
)
MgSO4 DS: NF90-400
MgSO4 DS:BW30-440i
NaCl DS: NF90-400
NaCl DS: BW30-440i

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates are 61%-65%, Cf 34760, Qf 7 m
3
/h, Cf MgSO429971 ppm, CfNaCl 8250 ppm
Figure 6: Concentrate osmotic pressure at different recovery rates.
Forward Osmosis Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 85
the feed pressure requirements were the highest in
case of MgSO
4
followed by MgCl
2
and NaCl
respectively (Figure 7). As mentioned earlier here, the
reason for that was due to the high rejection rate and
concentration of chemicals used to generate a
desirable osmotic pressure for FO seawater
desalination when MgSO
4
was used as a draw
solution. Figure 8 shows that the concentrate osmotic
pressure was the highest in case of MgSO
4
because of
the high membrane rejection to divalent ions and
complemented by the high concentration polarization
problem. This substantiated the fact why MgSO
4
draw
solution required a higher feed pressure than MgCl
2

and NaCl. As a result, the specific power consumption
was the highest when MgSO
4
was used as a draw
solution (Figure 9). Furthermore, the diffusion of SO
4

ions is lower than Cl which may aggravate the intensity
of concentration polarization of MgSO
4
draw solution
[29, 30].

It is evident from Figure 9 that the specific power
consumption was affected by the recovery rate and it
depends on the type of the osmotic agent in use. In
case of MgSO4 draw solution, the simulation results
show the optimal specific power consumption could be
achieved at around 42% recovery rate. Then the power
consumption increased with the recovery rates
increase above 42% (Figure 9). Therefore, the optimal
recovery rate for MgSO
4
in the regeneration stage
30
40
50
60
70
80
90
100
40 42 44 46 48 50 52 54 56
%R
P
f

(
b
a
r
)
MgSO4
MgCl2
NaCl

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates 42%-54%, Qf 5 m
3
/h, Cf MgSO4120386 ppm, CfNaCl 36600 ppm, CfmgCl 36600 ppm
Figure 7: Effect of recovery rate on the feed pressure.

40
45
50
55
60
65
70
75
80
85
90
40 42 44 46 48 50 52 54 56
%R
_
c
(
b
a
r
)
MgSO4
MgCl2
NaCl

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates 42%-54%, Qf 5 m
3
/h, Cf MgSO4120386 ppm, CfNaCl 36600 ppm, CfmgCl 36600 ppm
Figure 8: Effect of recovery rate on the concentrate concentration.
86 Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 Ali Altaee
should be around 42%. When MgSO4 was replaced by
MgCl
2
, the optimal power consumption was achieved at
a recovery rate about 48% (Figure 9). Accordingly, a
higher recovery rate can be achieved if MgCl
2
was
used instead of MgSO
4
as a draw solution. Finally,
when NaCl was used as a draw solution the optimal
power consumption was achieved at a recovery rate
around 51% (Figure 9). Based on these results, a
higher recovery rate can be achieved when NaCl was
used in the draw solution. This is probably one of the
advantages of using low molecular weight osmotic
agent in the draw solution.
The advantage of using MgSO
4
was the lower
permeate TDS compared to MgCl
2
and NaCl draw
solutions (Figure 10). Primarily, this was due to the
large molecular size and higher ionic charge of MgSO
4

which was highly rejected by the RO membrane. The
high permeate concentration may require an additional
membrane filtration to reduce the concentration to a
desirable level which leads to a higher treatment cost.
The lowest permeate concentration was observed
when MgSO
4
was used as a draw solution. Noting that
the difference between the permeate TDS in case of
MgSO
4
and MgCl
2
was insignificant. The permeate
TDS, however, decreased with increasing the recovery
rate (Figure 10). This was due to increasing the per-
meate dilution factor with the recovery rate increase.
Although membrane hyperfiltration processes were
suggested, so far, for the draw solution separation and
regeneration, MD process was also investigated for the
3
3.5
4
4.5
5
5.5
6
38 40 42 44 46 48 50 52 54 56
%R
E
s

(
k
W
h
/
m
3
)
MgSO4
MgCl2
NaCl

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates 42%-54%, Qf 5 m
3
/h, Cf MgSO4120386 ppm, CfNaCl 36600 ppm, CfmgCl 36600 ppm
Figure 9: Effect of recovery rate on the Specific power consumption.

0
100
200
300
400
500
600
40 42 44 46 48 50 52 54 56
%R
C
p

(
p
p
m
)
MgSO4
MgCl2
NaCl

Testing condition: feed temperature 25
o
C, pH 7.6, recovery rates 42%-54%, Qf 5 m
3
/h, Cf MgSO4120386 ppm, CfNaCl 36600 ppm, CfmgCl 36600 ppm
Figure 10: permeate concentration at different recovery rates.
Forward Osmosis Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 87
regeneration of draw solution [20]. MD process was
demonstrated to be feasible in the regeneration of
some draw solution such as ammonia carbon dioxide
[7, 9]. In the latter process, the diluted draw solution
was heated up to 60
o
C before it was fed into the MD
unit. Inside the membrane, ammonia carbon dioxide
evaporated and condensed in the permeate side of the
membrane leaving behind a fresh water in the
concentrate side of the membrane. The concentration
of ammonia carbon dioxide in the draw solution varies
depending on the feed water solution. For instance, the
FO membrane flux for seawater desalination using a
draw solution contains 12% ammonia carbon dioxide
was 10 L/h [7]. After leaving the MD unit, the
concentrated ammonia carbon dioxide is mixed with
distilled water to prepare a draw solution of desirable
concentration. Despite the low cost of draw solution
regeneration by MD, the process suffers from several
draw backs [21]:
1. Membrane wetting which results in a reduction
in the permeate flux and increasing the
permeate TDS.
2. Low recovery rate
3. Small membrane area of the MD unit
In addition, the residues of osmotic agent in the
concentrate side of the MD membrane (the product
water) will affect the product water quality. According to
the WHO, the concentration of ammonia in the drinking
water should be less than 1 ppm. Upon chlorine
reaction with ammonia to produce chloramine, the
concentration of chloramine shouldnt exceed 1 ppm as
recommended by the environmental and health
agencies. A high concentration of disinfectant in the
drinking water triggers the formation of disinfection by-
products such as Trihalomethane (THM) which is a
carcinogenic compound. The level of disinfectant
should be adjusted to the desirable level; this in turn
will increase the cost of treatment. It is also expected
that the product water quality is affected by the salt
diffusion (mostly NaCl) from the seawater to the draw
solution side of the FO membrane. Salt concentration
in the draw solution side is not affected by the MD
process and it will remain in the product water stream.
The concentration of salt in the product water will
increase due to the draw solution recycling till it
reaches the actual concentration level of salt diffusion
across the FO membrane. The lower the membrane
salt rejection rate the higher the salt diffusion.
Practically, the salt diffusion in FO process using high
salt rejection membrane (Rejection>99%) was 2051
ppm; feed water concentration 34590 ppm (results as
not shown here). The results in Figure 11 show the
increase of salt concentration in the product water as a
function of draw solution recycling. It is evident from
Figure 11 that the concentration of NaCl in the product
water increased with the number of draw solution
recycling. After recycling 15 times the concentration of
NaCl in the draw solution reached 2000 ppm. The
concentration of NaCl in the product water reached the
actual salt diffusion concentration, 2051.093 ppm, after
recycle number 60 and remained unchanged.
The accumulation of NaCl in the product water
renders it brackish. As a result it should be purged to
reduce the salt concentration to the desirable level.
Low pressures BWRO membrane can be employed to
adjust the concentration of product water TDS.
0
500
1000
1500
2000
2500
1 5 9
1
3
1
7
2
1
2
5
2
9
3
3
3
7
4
1
4
5
4
9
5
3
5
7
6
1
6
5
6
9
Recycl e No.
S
a
l
t

c
o
n
c
e
n
t
r
a
t
i
o
n

(
p
p
m
)
Breakthrough point of
maximum salt
diffusion


Figure 11: Concentration of NaCl salt in product water as function of draw solution recycle.
88 Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 Ali Altaee
Consequently, the total cost of the seawater treatment
will be higher. There are a number of parameters
affecting salt diffusion across the FO membrane such
as feed concentration, type of the membrane, feed and
draw solutions flow rate and temperature which need a
further investigation.
2.2. Thermal Processes for Draw Solut ion Regene-
ration
Thermal processes such as MSF and MED were
proposed for the regeneration of draw solution and
recycling [20]. MSF is very popular in the Gulf region of
the Middle East due to the high salinity of the Gulf
water. At the beginning of the 21
st
century the new
generation of MED was introduced and proven to be
very competitive to the traditional MSF design [4, 22,
23]. MSF can be operated either by once through or
brine recycle modes. The recovery rate in the once
through mode is 10% while in the brine recycle is 33%
[24]. Most of the current MSF plants are operated in the
brine recycle mode to achieve a high recovery rate [4,
24]. In contrast, MED doesnt have the operation
flexibility in MSF. A recovery rate of 33% can be
reached in the MED plant. The high performance of
MED and low energy requirements compared to MSF
attracted a lot of attention especially when the fuel
price is high. The top brine temperature in MSF is 110
o
C while in MED is 65
o
C [25].
The cost of draw solution regeneration by thermal
processes is expected to be higher than membrane
filtration. In the latter case the feed water to the thermal
processes is the diluted draw solution from the FO
process. Thermal processes are more suitable for the
regeneration of ammonia carbon dioxide than organic
and inorganic salts such as sucrose, glucose, MgSO4,
MgCl2, etc. because of the low temperature required
for the evaporation of ammonia carbon dioxide. The
following point should be observed upon using thermal
processes for the regeneration of draw solution:
1. The potential of scale problems caused by the
draw solution such as MgSO4 treatment by
MSF
2. The diluted draw solution needs to be recycled
more than often to achieve the desirable water
recovery and draw solution concentration. As
such the cost of regeneration will increase
As mentioned before, the expected concentration of
ammonia dioxide in the draw solution is about 10% for
seawater desalination, TDS 35000 ppm. In case of
MSF process, once thorough operating mode could be
adequate for ammonia carbon dioxide regeneration
(concentration 10% in draw solution). However, for
other draw solution a recovery rate over than the
conventional 30% is required to achieve the target
osmotic agent concentration in the draw solution.
Keeping in mind that if thermal processes are used in
conjunction with FO for ammonia carbon dioxide
regeneration, there will be salt residues in the fresh
water as explained above in the FO-MD process.
Therefore, an additional membrane/chemical treatment
is required for the removal of salt from the product
water.
3. FORWARD OSMOSIS FOR POWER GENERATION
Power generation from renewable sources such as
solar, wind, geothermal has received a lot of attention
due to the continuous increase in the fossil fuel prices
and environmental awareness about green house
gases emission. The idea of using FO in power
generation was dated back to the 70s of last century
[26]. Sidney Loeb was first who suggested using
osmotic energy in what so called Pressure Retarded
Osmosis (PRO) in power generation [26, 27]. There are
analogy between the application of FO in seawater
desalination and power generation. In the latter
process, two solutions of different concentrations are
fed into the FO membrane. The high concentration
solution is known as the draw solution while the low
concentration solution is the donor solution. Fresh
water transports across the membrane barrier from the
low to the high concentration solution due to the
osmotic pressure gradient. After leaving the FO
membrane, the diluted draw solution is fed into turbine
for power generation (Figure 12).
PRO can be used alone or in combination with RO
process for power generation and seawater
desalination [27]. As such, seawater is applied into the
RO membranes for desalination. Permeate from the
RO system is the product water while the concentrate
is the donor solution in the FO process. This process
has the advantage of reducing the brine concentrate
discharged to sea. Although PRO was suggested long
time ago, its commercial application wasnt achieved till
2009 when StatKraft company, Norway, built the world
first power plant operates by the osmotic energy
(Figure 12). The process was slightly different from that
suggested by Sidney Loeb by in which an Energy
Recovery Instrument (ERI) was incorporating to
enhance the overall performance of the process
according to the following equation:
Forward Osmosis Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 89

PV
Es = [2]
Where Es is the power generation from the PRO, P
is the pressure of feed solution to the turbine, V is the
volume of feed solution to the turbine, and is the
pump efficiency. The generated power from PRO
process increases with increasing the volume and
pressure of the feed solution to the turbine system.
Earlier FO membrane exhibited a low flow due to the
adverse impact of concentration polarization (Figure 1).
External and/or internal salt accumulation at the feed
side of the membrane surface reduces the osmotic
pressure gradient and hence the driving force for fresh
water extraction. Luckily, new FO membranes dealt
with this problem through reducing the thickness of the
support layer which resulted in reducing salt
accumulation at the membrane surface (Figure 14)
[16]. This development encouraged scientists and
pushed the PRO process a step ahead towards
commer-cialization. The concerted efforts were
culminated by building the world first power generation
plant by using the PRO process [14]. Fresh water was
used as feed solution while seawater was the draw
solution. Using fresh water as a feed solution will
eliminate the effect of concentrative concentration
polarization at the membrane surface. However, the
process is site specific; i.e. it is dependent on the
availability of draw and feed solutions. In many
countries affected by water shortage it is rather
impossible using fresh water as a feed solution for
power generation by the PRO process. Since water
shortage problem has affected many areas around the
world, seawater was proposed to be the feed solution
provided that the draw solution must be a solution of
higher concentration. In the latter design, the
concentrative concentration polarization plays an
important role in determining the net water flux across
the membrane [15]. Indeed, PRO process for power
generation using fresh water as a feed solution and
seawater as a draw solution is a site specific process
and cant be generalized worldwide. It depends on the
abundancy of the feed and draw solutions in a
particular area.
Alternatively, wastewater effluent was proposed to
replace the fresh water as a feed solution in the PRO
process to overcome the fresh water shortage problem
[10, 12]. Any impaired solution with low salinity can be
used as a feed solution. The PRO process diagram
using wastewater effluent and seawater as feed and
draw solution respectively is shown in Figure 14.
Wastewater effluent contains a number of impurities
such as organic matters, total nitrogen (T-N), total
phosphorus (T-P) and suspended solids (TSS).
Organic matters presence, in particular, in feed
wastewater effluent increases the propensity of
membrane fouling propensity and affecting the overall
water transport across the membrane [15, 28]. The
coupled effect of organic matter and the concentration


Figure 12: Schematic diagram of PRO process.
Donor solution
Power
90 Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 Ali Altaee
polarization effect were well investigated in the
literature [15]. The experimental work showed that
PRO operates better when the wastewater effluent
facing the membrane surface while the draw solution
facing the support layer [15]. Although such design
reduces the osmotic driving force across the
membrane but it is more efficient in reducing the
coupled effect of organic matter fouling and
concentration polarization.
Shung et al. suggested a conceptual PRO design
using wastewater effluent as a feed solution while
seawater was the draw solution (Figure 15). Part of the
diluted draw solution is passed through a pressure
exchanger for energy reuse then it is mixed with the
rest of the diluted draw solution and sent to a second
FO membrane. In the latter membrane a custom
design draw solution is used for water extraction from
the diluted seawater. Although using wastewater
effluent as a feed solution will overcome the problem of


Figure 13: StatKraft PRO power generation plant (from StatKraft website).


Figure 14: Concentration polarization in RO and FO membranes.
RO membrane
FO membrane
Forward Osmosis Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 91
fresh water shortage, the process performance will be
lower than the fresh water feed. Organic matter fouling
is the main drawbacks of using wastewater effluent as
a feed in PRO. To alleviate the effect of organic fouling,
wastewater effluent should face the selective layer of
FO membrane while the draw solution faces the
support layer. This operating mode is renown of
yielding a lower membrane flux but more effective in
reducing the FO membrane fouling [15]. Practically, the
concentration of organic impurities in wastewater
effluent varies depending on the type and level of
wastewater treatment.
The PRO design shown in Figure 15 has two of FO
membrane systems which makes it rather complicated.
The diluted seawater from the first FO process can be
treated directly by thermal or membrane processes for
fresh water extraction and draw solution recycling
(Figure 16). This will reduce the FO membrane area
and cost. Additionally, the plant foot plant will be less.


Figure 15: An integrated osmotic MBR, osmotic power generation and seawater desalination system (Tai-Shung Chung et al.,
2010).














Figure 16: PRO process using wastewater effluent as a feed and seawater as the draw solution.
92 Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 Ali Altaee
The wastewater concentrate leaving the FO membrane
can be used for irrigation or discharge to a proper
water system. Either design in Figure 15 or 16 are an
alternative to the use of freshwater and they need to be
confirmed experimentally.
4. Conclusion
Despite the wealth of literature and experimental
work conducted in FO membrane process, its
application is still limited to bench and some pilot plant
studies. In seawater desalination, the process is still
under investigation. Its wide application in seawater
desalination was hampered, at the beginning, due to
the lack of appropriate membrane. Understanding the
phenomenon of concentration polarization in the FO
process has resulted in the development of a suitable
FO membrane for seawater desalination. The real
challenge encountered the commercial application of
FO process was the economic feasibility of the FO and
if it can be competitive to the existing membrane
desalination technologies such as RO. Any successful
application of FO requires a cost-effective regeneration
process. This is because most of the energy required in
FO desalination is spent in the regeneration process.
Results from previous research studies suggested
using NF membrane in the regeneration of tailored
design draw solution constituted of large divalent ions
such as MgSO4. Such design is more suitable for
brackish water desalination as most of the available NF
membrane cant tolerate feed pressure more than 40
bar. Different organic and inorganic salts were
suggested to be used as draw solution. The simulation
results in this study showed that NaCl is more efficient
than MgSO
4
and MgCl
2
due to the higher recovery rate
that can be achieved at lower power consumption.
Osmotic agent of small molecular weight, probably, is
more efficient draw solution than large molecular
weight osmotic agent due to the higher osmotic
pressure possessed by the former osmotic agent.
One of the inherent problems in FO is the salt
diffusion from seawater to the draw solution side of the
membrane. In particular, this is important when
MD/thermal processes are used for draw solution
evaporation and concentration such as in ammonia
carbon dioxide. Low pressure BWRO membrane
process can be used for salt removal from permeate to
the desirable level. But the cost of the process be
higher than the basic conventional design.
Additionally, FO process has the potential of
application in power generation by what so called PRO
process. The only commercial application of such
process in power generation is the pilot plant built by
StatKraft in Norway. Such process is a site specific as
it uses fresh water and seawater as feed and draw
solution respectively. Wastewater effluent was
proposed as a feed solution and hence the geographic
application of PRO is extended to include water
shortage areas. However, membrane fouling by the
organic matters in the wastewater effluent should be
further investigated to reduce the treatment cost.
REFERENCES
[1] Mohammad Al-Sahali, Hisham Ettouney. Developments in
thermal desalination processes: Design, energy, and costing
aspects, Desalination 2007; 214: 227-240.
http://dx.doi.org/10.1016/j.desal.2006.08.020
[2] Ioannis C. Karagiannis, Petros G. Soldatos, Water
desalination cost literature: review and assessment,
Desalination 2008; 223: 448-456.
http://dx.doi.org/10.1016/j.desal.2007.02.071
[3] Kah Peng Lee, Tom C. Arnot, Davide Mattia, A review of
reverse osmosis membrane materials for desalination
Development to date and future potential, J ournal of
Membrane Science 2011; 370: 1-22.
http://dx.doi.org/10.1016/j.memsci.2010.12.036
[4] Toufic Mezher, Hassan Fath, Zeina Abbas, Arslan Khaled,
Techno-economic assessment and environmental impacts of
desalination technologies, Desalination 2011; 266: 263-273.
http://dx.doi.org/10.1016/j.desal.2010.08.035
[5] Lauren F. Greenlee, Desmond F. Lawler, Benny D. Freeman,
Benoit Marrot, Philippe Moulin, Reverse osmosis
desalination: Water sources, technology, and today's
challenges, Water Research 2009; 43: 2317-2348.
http://dx.doi.org/10.1016/j.watres.2009.03.010
[6] Robert L. McGinnis, Menachem Elimelech, Energy
requirements of ammoniacarbon dioxide forward osmosis
desalination, Desalination 2007; 207: 370-382.
http://dx.doi.org/10.1016/j.desal.2006.08.012
[7] J effrey R. McCutcheon, Robert L. McGinnis, Menachem
Elimelech, Desalination by ammoniacarbon dioxide forward
osmosis: Influence of draw and feed solution concentrations
on process performance, J ournal of Membrane Science
2006; 278: 114-123.
http://dx.doi.org/10.1016/j.memsci.2005.10.048
[8] Kah Peng Lee, Tom C. Arnot, Davide Mattia, A review of
reverse osmosis membrane materials for desalination
Development to date and future potential, J ournal of
Membrane Science 2011; 370: 1-22.
http://dx.doi.org/10.1016/j.memsci.2010.12.036
[9] Tzahi Y. Cath, Amy E. Childress, Menachem Elimelech,
Forward osmosis: Principles, applications, and recent
developments, J ournal of Membrane Science 2006; 281: 70-
87.
http://dx.doi.org/10.1016/j.memsci.2006.05.048
[10] Tai-Shung Chung, Sui Zhang, Kai Yu Wang, J incai Su, Ming
Ming Ling, Forward osmosis processes: Yesterday, today and
tomorrow, Desalination 2012; 287: 78-81.
http://dx.doi.org/10.1016/j.desal.2010.12.019
[11] J effrey R. McCutcheon, Menachem Elimelech, Influence of
concentrative and dilutive internal concentration polarization
on flux behavior in forward osmosis, J ournal of Membrane
Science 2006; 284: 237-247.
http://dx.doi.org/10.1016/j.memsci.2006.07.049
[12] Andrea Achilli, Tzahi Y. Cath, Amy E. Childress, Power
generation with pressure retarded osmosis: An experimental

Forward Osmosis Journal of Membrane and Separation Technology, 2012 Vol. 1, No. 2 93
and theoretical investigation, J ournal of Membrane
Science2009;343: 42-52.
http://dx.doi.org/10.1016/j.memsci.2009.07.006
[13] Sidney Loeb, Energy production at the Dead Sea by
pressure-retarded osmosis: challenge or chimera?,
Desalination 1998; 120: 247-262.
http://dx.doi.org/10.1016/S0011-9164(98)00222-7
[14] World Desalination Report, Vol 45, Issue 39, 26 October
2009.
[15] Chuyang Y. Tang, Qianhong She, Winson C.L. Lay, Rong
Wang, Anthony G. Fane, Coupled effects of internal
concentration polarization and fouling on flux behavior of
forward osmosis membranes during humic acid filtration,
J ournal of Membrane Science2010; 354: 123-133.
http://dx.doi.org/10.1016/j.memsci.2010.02.059
[16] Sidney Loeb, Leonid Titelman, Emmanuel Korngold, J oseph
Freiman, Effect of porous support fabric on osmosis through
a Loeb-Sourirajan type asymmetric membrane, J ournal of
Membrane Science 1997: 129: 243-249.
http://dx.doi.org/10.1016/S0376-7388(96)00354-7
[17] M. A. Darwish, Hisham El-Dessouky, The heat recovery
thermal vapour-compression desalting system: A comparison
with other thermal desalination processes, Applied Thermal
Engineering 1996; 16: 523-537.
http://dx.doi.org/10.1016/1359-4311(95)00034-8
[18] Ian Lomax, Experiences of Dow in the field of seawater
reverse osmosis, Desalination 2008; 224: 111-118.
http://dx.doi.org/10.1016/j.desal.2007.02.085
[19] Mark Wilf, Kenneth Klinko, Optimization of seawater RO
systems design, Desalination 2001; 138: 299-306.
http://dx.doi.org/10.1016/S0011-9164(01)00278-8
[20] Abdulsalam Al-Mayahi (Worcestor Park, GB); Adel Sharif
(Guildford GB, GB). Solvent Removal Process. United States
patent US 20110114558. 2011 May.
[21] Mohamed Khayet, Membranes and theoretical modeling of
membrane distillation: A review, Advances in Colloid and
Interface Science 2011; 164: 56-88.
http://dx.doi.org/10.1016/j.cis.2010.09.005
[22] Narmine H. Aly, Adel K. El-Figi, Thermal performance of
seawater desalination systems, Desalination 2003; 158: 127-
142.
http://dx.doi.org/10.1016/S0011-9164(03)00443-0
[23] J acques Andrianne, Flix Alardin, Thermal and membrane
processe economics: Optimized selection for seawater
desalination
,
Desalination 2003; 153: 305-311.
http://dx.doi.org/10.1016/S0011-9164(02)01128-1
[24] Ahmed M. Helal, Once-through and brine recirculation MSF
designs a comparative study, Desalination 2005; 171: 33-
60.
http://dx.doi.org/10.1016/j.desa1.2004.02.104
[25] Suresh Patel, Michael A. Finan, New antifoulants for deposit
control in MSF and MED plants, Desalination 1999; 124: 63-
74.
http://dx.doi.org/10.1016/S0011-9164(99)00089-2
[26] Loeb S., Norman R. S. (1975). "Osmotic Power Plants".
Science 2003; 189: 654655. .
http://dx.doi.org/10.1126/science.189.4203.654
[27] Loeb S, "Energy Production at the Dead Sea by Pressure-
Retarded Osmosis: Challenge or Chimera?", Desalination
1998; 120: 247262.
http://dx.doi.org/10.1016/S0011-9164(98)00222-7
[28] Sangyoup Lee, Menachem Elimelech, Salt cleaning of
organic-fouled reverse osmosis membranes, Water Research
2007; 415: 1134-1142.
http://dx.doi.org/10.1016/j.watres.2006.11.043
[29] Bernard Vigneault, Peter G. C. Campbell, Andr Tessier and
Richard De Vitre, Geochemical changes in sulfidic mine
tailings stored under a shallow water cover, Water Research
2001; 35: 1066-1076.
http://dx.doi.org/10.1016/S0043-1354(00)00331-6
[30] Mustafa, Syed, Shah, Khizar, Naeem, Abdul, Waseem,
Muhammad, Ahmad, Tauqeer, Khan, Sadullah, Water, Air,
and Soil Pollution 2011; 217: 57-69.



Received on 15-08-2012 Accepted on 08-12-2012 Published on 31-12-2012

DOI: http://dx.doi.org/10.6000/1929-6037.2012.01.02.2