The exploration of coupling surfactants with highly charged phosphate as novel draw solutions for forward osmosis (FO) to minimize salt leakage and enhance draw solute recovery at low energy is presented for the first time. In this work, anionic (sodium dodecyl sulfate (SDS)), cationic (cetyltrimethylammonium bromide) and non-ionic (polyethylene glycol tert-octylphenyl ether, TX-100) surfactants were used to couple with Na3PO4 for enlarging draw solution particle size that could improve the solute retention. Three kinds of nanofiltration (NF) membranes (HL, DL and TS80 with molecular weight cutoff (MWCO) of 100–500) and three kinds of ultrafiltration (UF) membranes (GE, GH and GK with MWCO of 1,000–3,000) were used to evaluate the draw solution recovery. The formed second layer, micelles, MWCO and charged valent were determined to be the dominant mechanisms for diluted draw solution recovery. The results show that coupling non-ionic surfactant with 0.02 M Na3PO4 can be regenerated with recoveries as high as 60% and 96% when using UF-GE and NF-TS80, respectively. The overall performance demonstrates that coupling surfactant with highly charged phosphate is promising as draw solutes for application to FO.

INTRODUCTION

Forward osmosis (FO) is a process of water driven by the osmotic pressure difference across the semi-permeable membrane without the aid of external energy. Unlike other membrane process such as reverse osmosis (RO), FO can be operated at a low cost because of the absence of hydraulic pressure which makes it possible to be a promising technology. Recently, FO process has been applied in many places, such as power generation (McGinnis & Elimelech 2008; Jia et al. 2014), food processing (Petrotos et al. 1998, 1999), desalination (McCutcheon et al. 2005, 2006; Cath et al. 2006), osmotic membrane bioreactor (Achilli et al. 2009; Kim 2014) and medical and pharmaceutical applications (Yang et al. 2009). However, its application is still hindered due to the lack of effective diluted draw solution recovery. Draw solution regeneration using RO or thermal decomposition proved highly efficient but costly (McCutcheon et al. 2006; Achilli et al. 2010; Yen et al. 2010; Wang et al. 2011), meanwhile magnetic separators is troublesome with particle aggregation during recycling (Ling et al. 2010; Ge et al. 2011). Some draw solutions (CuSO4 and MgSO4) were recovered by using metathesis precipitation reaction (Alnaizy et al. 2013a, b), but adding chemicals will increase the cost of upgrading and secondary pollution to the environment.

Basically, the small molecular weight electrolytes and solutes used as a draw solution may not be economical because of the challenge of achieving low-energy recovery and the problem of salt leakage due to reverse diffusion of the draw solution across the membrane. Hence, in this study, three surfactants were proposed to couple with highly charged species of phosphate as the draw solutions to minimize the reverse flux of ions in FO and increase efficiency recovery of diluted draw solution using ultrafiltration (UF) or nanofiltration (NF) systems at low energy. Surfactants present dual structural units: a hydrophilic group (polar group, the head) and a hydrophobic group (long hydrocarbon chain, the tail), which are classified according to the charged groups in the hydrophilic head including (i) anionic, (ii) cationic, and (iii) non-ionic. Micelles will be formed spontaneously in solution at concentrations above the critical micelle concentration (CMC), and micellar systems have been applied in food processing, chemicals and biotechnology (Schramm et al. 2003). In principle, at the CMC, surfactant solution aggregate to form micelles that could couple with highly charged phosphate to enlarge molecular size of draw solution, resulting in enhancing draw solute recovery with the aid of membrane process. The purpose of this research is to investigate the recovery of highly charged species of phosphate coupled with surfactants as the draw solutions for the following: (1) effect of different surfactants on water flux and reverse salt flux; (2) comparison of three kinds of NF membranes and three kinds of UF membranes; and (3) investigate the effects of coupling anionic (SDS), cationic (cetyltrimethylammonium bromide (CTAB)) and non-ionic (TX-100) surfactants for draw solution recovery.

MATERIALS AND METHODS

FO membrane

The flat-sheet cellulose triacetate (CTA) FO membranes used in this study were supplied by Hydration Technology Innovations (HTIs OsMem™ CTA Membrane 130806, Albany, OR, USA) with a size of 15 × 22 cm for each piece. The overall thickness of the membrane was approximately 50 μm, and the FO membrane was negatively charged at pH >4.5.

NF/UF membranes

In the last few years, the application of NF membranes have become popular, because of offering a variety of advantages such as low operating pressure, relatively low capital and maintenance cost and high rejection rates for multivalent salts (Liang et al. 2014). UF has attracted considerable attention because of its relatively low energy cost compared to RO and NF, and the decreasing cost of membranes (Gao et al. 2011). Table 1 shows the characteristics of the UF and NF membranes provided by Osmonics and TriSep for the draw solution recovery.

Table 1

Characteristics of ultrafiltration and nanofiltration membranes

 UF-GEUF-GHUF-GKNF-HLNF-DLNF-TS80
Manufacturer GE osmonics GE osmonics GE osmonics GE osmonics GE osmonics TriSep 
Material Composite polyamide Thin film Thin film Thin film Thin film Polyamide 
MWCO (Da) 1,000 2,000 3,000 150–300 150–300 150 
25°C
Range pH 
1–11 1–11 1–11 3–9 2–10 2–11 
Rej-size 1K-PEG 2K-PEG 3K-PEG 98% – MgSO4 96% – MgSO4 99% – MgSO4 
Typical flux/psi
LMH@PSI 
31/400 34/150 29/75 66/100 53/100 34/110 
Contact angle (°C) 66.2 70.96 58.66 41 ± 2 (Verliefde 200854.4 (Cornelis et al. 200548 ± 2 (Verliefde 2008
 UF-GEUF-GHUF-GKNF-HLNF-DLNF-TS80
Manufacturer GE osmonics GE osmonics GE osmonics GE osmonics GE osmonics TriSep 
Material Composite polyamide Thin film Thin film Thin film Thin film Polyamide 
MWCO (Da) 1,000 2,000 3,000 150–300 150–300 150 
25°C
Range pH 
1–11 1–11 1–11 3–9 2–10 2–11 
Rej-size 1K-PEG 2K-PEG 3K-PEG 98% – MgSO4 96% – MgSO4 99% – MgSO4 
Typical flux/psi
LMH@PSI 
31/400 34/150 29/75 66/100 53/100 34/110 
Contact angle (°C) 66.2 70.96 58.66 41 ± 2 (Verliefde 200854.4 (Cornelis et al. 200548 ± 2 (Verliefde 2008

Chemicals

All the chemicals used in this study were of laboratory grade, and purchased from Sigma-Aldrich, UK and using deionized (DI) water as feed solution. The three surfactants used in this study, including their names, chemical formulae and properties, are listed in Table 2. The structural formulae of surfactants are shown in Figure 1.

Table 2

Summary of the surfactant properties

SurfactantTypeMolecular weight (g/mol)AbbreviationCritical micelle concentration (mM)
Cetyltrimethylammonium bromide Cationic 364.45 CTAB 0.9 
SDS Anionic 288.38 SDS 
Polyethyleneglycoltert-octylphenyl ether Non-ionic 647 TX-100 0.4 
SurfactantTypeMolecular weight (g/mol)AbbreviationCritical micelle concentration (mM)
Cetyltrimethylammonium bromide Cationic 364.45 CTAB 0.9 
SDS Anionic 288.38 SDS 
Polyethyleneglycoltert-octylphenyl ether Non-ionic 647 TX-100 0.4 
Figure 1

The structural formulae of SDS, CTAB and TX-100.

Figure 1

The structural formulae of SDS, CTAB and TX-100.

Experiment setup

The experiment setup is shown in Figure 2 and the total effective FO membrane area was 41.40 cm2. The feed and draw solutions on both sides of the FO membrane were circulated by two peristaltic pumps (Master Flux L/S Drive, Model 7518-00) with a cross flow rate of 6.4 cm/s. Two water baths were used to maintain the temperature at 25 ± 0.5 °C during the experiment. Conductivity and pH sensors were installed in the containers to monitor any changes in the feed and draw solutions. A 1 L of feed solution tank was placed on a weighting scale (BW12KH, Shimadzu, Japan) which was connected to a computer data logging system to monitor weight and volume changes at regular time intervals. After FO, the dilute draw solution was recovered through the different UF and NF membranes for reuse by using a cross flow module (CF042 DelrinAcetal Cross Flow Cell, USA) under hydraulic pressures of 60 and 6.9 bar. The efficiency of the UF and NF is defined by the rejection factor, R: 
formula
1
Figure 2

Experimental setup of laboratory-scale FO/UF/NF hybrid system using highly charged phosphate coupled with surfactants as the draw solutions.

Figure 2

Experimental setup of laboratory-scale FO/UF/NF hybrid system using highly charged phosphate coupled with surfactants as the draw solutions.

where CP and CR are the concentrations of the solutes in permeate and retentate, respectively.

Analytical methods

The concentration of phosphate was measured by colorimetrically according to Method 8178-Amino Acid Method using a UV–Vis spectrophotometer (Hach Model DR-4000) at a wavelength of 530 nm. The concentration of sodium was measured by direct potentiometric analyses using an ion meter (Metrohm Model 781). Surfactant concentrations were prepared using 0.45 μm filter paper, and then measured using an Aurora1010C TOC Analyzer (O.I. Analytical Corporation, USA). The osmolality of solutions was measured using an osmometer (Model 3320, Advanced Instruments, Inc., USA). The conductivity was determined by a conductivity meter (Sension156, Hach, China).

RESULTS AND DISCUSSION

Effect of different surfactants on water flux and reverse salt flux

Figure 3 shows the reverse salt fluxes (Js) and water fluxes (Jw) for four draw solutions with various surfactants coupled with a Na3PO4 concentration of 0.1 M. The concentrations of all surfactants were used at the CMC. It is observed that reverse flux reduced significantly when coupling SDS and TX-100 to Na3PO4 draw solution. Compared with pure Na3PO4 (Js = 0.617 g/m2 h), 0.1 M Na3PO4 draw solution coupled with 10 mM SDS and 0.5 mM TX-100 exhibited less reverse salt flux (Js = 0.184 g/m2 h for 10 mM SDS and Js = 0.0786 mM TX-100). Strong hydrophobic interactions between tail groups of SDS and TX-100 with the membrane are believed to be the main mechanism for minimizing salt leakage. In contrast, the reverse salt flux of 1 mM CTAB coupled with 0.1 M Na3PO4 draw solution was not reduced compared with SDS and TX-100. Moreover, the experiment results showed that water fluxes decreased slightly when using CTAB (Jw of 2.814/m2 h), SDS (Jw of 2.648 L/m2 h) and TX-100 (Jw of 2.508 L/m2 h) coupled with Na3PO4 draw solution whereas their osmolality increased as shown in Table 3. This may be due to the increase in viscosity of draw solution when adding surfactants to Na3PO4.

Table 3

Comparison of osmolality of different draw solutions

Draw solutionNa3PO4Na3PO4 +10 mM SDSNa3PO4 + 1 mM CTABNa3PO4 + 0.5 mM Triton X-100
Osmolality (mOsm/kg•H2O) 300 317 354 368 
Draw solutionNa3PO4Na3PO4 +10 mM SDSNa3PO4 + 1 mM CTABNa3PO4 + 0.5 mM Triton X-100
Osmolality (mOsm/kg•H2O) 300 317 354 368 
Figure 3

Comparison of water flux and reverse salt flux with coupling different surfactants to 0.1 M Na3PO4 draw solution (pH of 8, using DI water as feed solution, CTA membrane).

Figure 3

Comparison of water flux and reverse salt flux with coupling different surfactants to 0.1 M Na3PO4 draw solution (pH of 8, using DI water as feed solution, CTA membrane).

Effect of different membranes on recovery of diluted draw solution

Recovery of diluted draw solutions of 0.04 M Na3PO4 coupled with 0.5 mM TX-100 in FO process was performed with the selected three UF and three NF membranes. Table 4 presents the experiments performed with the specific values of operating conditions. The results record that UF-GE achieved the highest recovery of 55.8%, followed by UF-GH (47.17%), and then UF-GK (45.56%) from the conducted experiments, indicating that UF-GE had the smallest pore size (molecular weight cutoff (MWCO) of 1000 Da), leading to retaining the diluted draw solution instead of passing through the membrane. By contrast, UF-GK generated the highest permeate flux of 24.7 L/m2 h under hydraulic pressure of 5.5 bar, indicating that UF-GK was the loosest membrane and the production of permeate across the UF membrane was easier. Similarly, diluted draw solution was recovered using three NF membranes and the results demonstrated that NF-TS80 could achieve a retention of 97.05%, followed by NF-DL (91.40%), and NF-HL (90.24%).

Table 4

Permeate flux (Jp) and retentions of different kinds of membranes using 0.04 M Na3PO4 coupled with 0.5 mM TX-100 as feed stream

MembranesTMP (bar)Jp (L m−2 h−1)TDS of permeate (g/L)Rej (%)
UF-GE 5.5 4.8 2.77 55.80 
UF-GH 5.5 10.3 3.27 47.17 
UF-GK 5.5 24.7 3.37 45.56 
NF-HL 6.9 8.2 0.543 91.40 
NF-DL 6.9 6.3 0.613 90.24 
NF-TS80 6.9 3.7 0.185 97.05 
MembranesTMP (bar)Jp (L m−2 h−1)TDS of permeate (g/L)Rej (%)
UF-GE 5.5 4.8 2.77 55.80 
UF-GH 5.5 10.3 3.27 47.17 
UF-GK 5.5 24.7 3.37 45.56 
NF-HL 6.9 8.2 0.543 91.40 
NF-DL 6.9 6.3 0.613 90.24 
NF-TS80 6.9 3.7 0.185 97.05 

Concentrate of different diluted draw solutions

The effect of different kinds of surfactants on recovering diluted draw solution is shown in Figure 4. The diluted draw solutions were prepared from 0.02 M Na3PO4 coupling with 0.5 mM TX-100, 1 mM CTAB and 10 mM SDS. The results showed that the TX-100 achieved the highest retention efficiency of phosphate. Non-ionic surfactants are known to adsorb strongly to pores of UF membranes, therefore causing flux decline and a retention increase (Jijnssonl & Jijnsson 1991). In Figure 5, the retention decreased from 76 to 54% with increase in the phosphate concentrations from 0.02 to 0.1 M in the UF process. This behavior could be explained by the shielding phenomenon of the charged membranes (Baticle et al. 1997; Paugam et al. 2004; Jermann et al. 2007). However, the retention of phosphate was high and stable (approximately 96.52%) when the concentrations of Na3PO4 were increased from 0.02 to 0.08 M using NF-TS80 as shown in Figure 6. An explanation for this phenomenon would be electrostatic repulsion and steric-hindrance effect. At pH 8, Na3PO4 was converted to divalent of and formed complex of (simulated from Mineql+ based on chemical equilibrium model from the thermodynamic database) according to Figure 7. Since NF-TS80 membrane was negatively charged, negative charged membrane repulsed the negative charged and leading to increased efficiency of solute rejection. In addition, the formed second layer of TX-100 on membrane surface enhanced solute retention (Ku et al. 2005).

Figure 4

Variation of PO4-P retention and permeate flux in the presence of 0.02 M Na3PO4 coupled with 0.5 mM TX-100, 1 mM CTAB, and 10 mM SDS using UF-GE membrane at a pressure of 5.5 bar.

Figure 4

Variation of PO4-P retention and permeate flux in the presence of 0.02 M Na3PO4 coupled with 0.5 mM TX-100, 1 mM CTAB, and 10 mM SDS using UF-GE membrane at a pressure of 5.5 bar.

Figure 5

Variations of PO4-P retention and permeate flux at various Na3PO4 concentrations coupled with 0.5 mM TX-100 using UF-GE membrane at a pressure of 5.5 bar.

Figure 5

Variations of PO4-P retention and permeate flux at various Na3PO4 concentrations coupled with 0.5 mM TX-100 using UF-GE membrane at a pressure of 5.5 bar.

Figure 6

Variations of PO4-P retention and permeate flux at various Na3PO4 concentrations coupled with 0.5 mM TX-100 using NT-TS80 membrane at a pressure of 6.9 bar.

Figure 6

Variations of PO4-P retention and permeate flux at various Na3PO4 concentrations coupled with 0.5 mM TX-100 using NT-TS80 membrane at a pressure of 6.9 bar.

Figure 7

Complex formation between and Na+ from Mineql+ for 0.02 M Na3PO4 (reproduced from the data of Mineql+).

Figure 7

Complex formation between and Na+ from Mineql+ for 0.02 M Na3PO4 (reproduced from the data of Mineql+).

CONCLUSIONS

This study presents the retention comparison of coupling CTAB, TX-100 and SDS with Na3PO4 by FO and UF/NF membranes in a cross flow module, and the results showed that the rejection order is TX-100 > SDS > CTAB. It was found that the formed second layer, micelles, MWCO, charged valent and hydrophobicity interactions were the dominant mechanism for the retention of solutes. For FO tests, the reverse salt fluxes of 0.1 M Na3PO4 coupled with 10 mM SDS and 0.5 mM TX-100 were reduced three and eight times more than pure Na3PO4, resulting in a significant decrease in the cost for replenishing lost draw solute. Moreover, the result of using different UF and NF membranes demonstrated that the UF-GE and NF-TS80 achieved the highest solute retention.

REFERENCES

REFERENCES
Achilli
A.
Cath
T. Y.
Marchand
E. A.
Childressa
A. E.
2009
The forward osmosis membrane bioreactor: a low fouling alternative to MBR processes
.
Desalination
239
,
10
21
.
Achilli
A.
Cath
T. Y.
Childress
A. E.
2010
Selection of inorganic-based draw solutions for forward osmosis applications
.
Journal of Membrane Science
364
(
1–2
),
233
241
.
Alnaizy
R.
Aidan
A.
Qasim
M.
2013a
Copper sulfate as draw solute in forward osmosis desalination
.
Journal of Environmental Chemical Engineering
1
(
3
),
424
430
.
Alnaizy
R.
Aidan
A.
Qasim
M.
2013b
Draw solute recovery by metathesis precipitation in forward osmosis desalination
.
Desalination and Water Treatment
51
(
28–30
),
5516
5525
.
Baticle
P.
Kiefer
C.
Lakhchaf
N.
Larbot
A.
Leclerc
O.
Persin
M.
Sarrazin
J.
1997
Salt filtration on gamma alumina nanofiltration membranes fired at two different temperatures
.
Journal of Membrane Science
135
,
1
8
.
Cath
T.
Childress
A.
Elimelech
M.
2006
Forward osmosis: principles, applications, and recent developments
.
Journal of Membrane Science
281
(
1–2
),
70
87
.
Cornelis
G.
Boussu
K.
Bruggen
B. V. d.
Devreese
I.
Vandecasteele
C.
2005
Nanofiltration of nonionic surfactants: effect of the molecular weight cutoff and contact angle on flux behavior
.
Industrial and Engineering Chemistry Research
44
,
7652
7658
.
Gao
W.
Liang
H.
Ma
J.
Han
M.
Chen
Z.-l.
Han
Z.-S.
Li
G.-B.
2011
Membrane fouling control in ultrafiltration technology for drinking water production: a review
.
Desalination
272
(
1
),
1
8
.
Jia
Z.
Wang
B.
Song
S.
Fan
Y.
2014
Blue energy: current technologies for sustainable power generation from water salinity gradient
.
Renewable and Sustainable Energy Reviews
31
,
91
100
.
Ku
Y.
Chen
S.
Wang
W.
2005
Effect of solution composition on the removal of copper ions by nanofiltration
.
Separation and Purification Technology
43
(
2
),
135
142
.
Ling
M. M.
Wang
K. Y.
Chung
T.-S.
2010
Highly water-soluble magnetic nanoparticles as novel draw solutes in forward osmosis for water reuse
.
Industrial and Engineering Chemistry Research
49
,
5869
5876
.
McCutcheon
J. R.
McGinnis
R. L.
Elimelech
M.
2005
A novel ammonia
carbon dioxide forward (direct) osmosis desalination process
.
Desalination
174
,
1
11
.
McGinnis
R. L.
Elimelech
M.
2008
Global challenges in energy and water supply: the promise of engineered osmosis
.
Environmental Science & Technology
42
(
23
),
8625
8629
.
Paugam
L.
Taha
S.
Dorange
G.
Jaouen
P.
Quéméneur
F.
2004
Mechanism of nitrate ions transfer in nanofiltration depending on pressure, pH, concentration and medium composition
.
Journal of Membrane Science
231
(
1–2
),
37
46
.
Schramm
L. L.
Stasiuk
E. N.
Marangoni
D. G.
2003
Surfactants and their applications
.
Annual Reports Section ‘C’ (Physical Chemistry)
99
,
3
48
.
Verliefde
A. R. D.
2008
Rejection of Organic Micropollutants by High Pressure Membranes (NF/RO)
.
Water Management Academic Press Delft, The Netherlands
.
Wang
K. Y.
Teoh
M. M.
Nugroho
A.
Chung
T.-S.
2011
Integrated forward osmosis–membrane distillation (FO–MD) hybrid system for the concentration of protein solutions
.
Chemical Engineering Science
66
(
11
),
2421
2430
.
Yang
Q.
Wang
K. Y.
Chung
T.-S.
2009
A novel dual-layer forward osmosis membrane for protein enrichment and concentration
.
Separation and Purification Technology
69
(
3
),
269
274
.
Yen
S. K.
Mehnas Haja
N. F.
Su
M.
Wang
K. Y.
Chung
T.-S.
2010
Study of draw solutes using 2-methylimidazole-based compounds in forward osmosis
.
Journal of Membrane Science
364
(
1–2
),
242
252
.