The polyamide reverse osmosis (RO) membranes were prepared through interfacial polymerization of m-phenylenediamine (MPD) and trimesoyl chloride (TMC). The use of dimethyl sulfoxide (DMSO) and glycerol as additives for the formation of thin-film composite (TFC) was investigated. We studied the effect of DMSO and glycerol addition on membrane property and RO performance. Microscopic morphology was examined by atomic force microscopy and scanning electron microscopy. The surface hydrophilicity was characterized on the basis of water contact angle and surface solid–liquid interfacial free energy (−ΔGSL). Water flux and salt rejection ability of the membranes prepared with or without the additives were evaluated by cross-flow RO tests. The results reveal that the addition of DMSO and glycerol strongly influences the property of the TFC RO membrane. Compared to the MPD/TMC membrane fabricated without DMSO and glycerol, the MPD/TMC/DMSO/glycerol membrane has a rougher surface and is more hydrophilic, showing smaller water contact angle and larger −ΔGSL value. Without decrease in salt rejection ability, the MPD/TMC/DMSO/glycerol membrane shows water flux significantly larger than that of the MPD/TMC membrane. The unique property of the MPD/TMC/DMSO/glycerol membrane is attributed to the cooperative effect of DMSO and glycerol on membrane structure during the interfacial polymerization process.

INTRODUCTION

Reverse osmosis (RO) membranes are widely used in desalination and water treatment (Kheriji et al. 2015; Ricci et al. 2016). In addition, RO membranes are applied in a variety of separation operations in environmental, chemical, and food manufacture sectors (Akbari et al. 2015). Being high in thermal, mechanical, and hydrolytic stability, polyamide membranes of advanced thin-film composites (TFC) dominate the RO membrane market (Zuo et al. 2014). The TFC membranes are composed of three polymers: an outer ultra-thin polyamide layer, a microporous middle polysulfone support, and a polyester web that acts as a substrate (Safarpour et al. 2015). Each layer can be independently modified to optimize selectivity and/or mechanical strength.

The ultra-thin polyamide layer is generally formed by interfacial polymerization of m-phenylenediamine (MPD) in aqueous phase and trimesoyl chloride (TMC) in organic phase (Ng et al. 2013; Kim et al. 2014; Jahangiri et al. 2015). Further development of TFC membranes demands excellent water permeability without sacrificing salt rejection ability (Shao & Kurth 2013; Tu et al. 2014). The aromatic polyamide TFC membranes with high water flux and satisfactory salt rejection ability were synthesized by MPD/TMC interfacial reaction in the presence of different types of additives, such as alcohols, ethers, sulfur-containing compounds, and monohydric aromatic compounds. Among the additives, dimethyl sulfoxide (DMSO) was reported to enhance water flux. Being able to dissolve both polar and nonpolar compounds, DMSO is used as a solvent for extractive distillation (Shen et al. 2015). Also, since it is miscible in water as well as in a wide range of organic solvents, it is utilized in chemical reactions as well as in biochemistry and cell biology processes (Dominik et al. 2014; Kurt et al. 2014).

According to Kim et al. (2005), the presence of DMSO in the aqueous phase during MPD/TMC interfacial reaction enhances the miscibility between the aqueous and the organic phases (by reducing the solubility difference of the two immiscible solutions), thereby facilitating the diffusion of MPD into the organic TMC phase. The consequence is modification of surface morphology, variation in polymer chain organization, and change of molecular nature during the formation of TFC membranes. On the other hand, glycerol is a passive flux-preserving agent commonly employed to prevent loss of porosity during oven drying (Kuehne et al. 2001). To the best of our knowledge, few studies were conducted to understand the reason of permeability enhancement without having salt separability seriously affected when DMSO and glycerol are used as additives for the generation of TFC RO membranes.

In this study, we prepared samples of MPD/TMC and MPD/TMC/DMSO/glycerol membranes. During interfacial polymerization, DMSO and glycerol were added for the generation of the latter. As for the former, there was no deployment of any additives. We compared the two in terms of RO membrane separation performance (i.e. water flux and salt rejection) and surface nature. Atomic force microscopy (AFM) and scanning electron microscope (SEM) were used to characterize surface morphology while contact angle to measure the hydrophilicity of the membranes.

EXPERIMENTAL

Materials and reagents

Polysulfone transparent beads with average molecular weight of 35,000 Da was purchased from Solvay Advanced Polymers, LLC. The chemicals used for the fabrication of the RO membranes, including TMC (purity >99.5%), MPD (purity >99.5%), DMSO, (+)-10-champhor sulfonic acid, sodium dodecyl sulfate, sodium hydroxide, and glycerin were purchased from Sigma–Aldrich, Co., Ltd. The organic solvent Isopar G selected for preparing TMC solution was from Gallade Chemical, Inc. The polyvinyl alcohol with a typical molecular weight of approximately 125 kg/mol was purchased from Sekisui Specialty Chemicals.

Other reagents such as N,N-dimethylformamide (DMF), 2-methoxyethanol and sodium chloride were AR grade and used as received without further purification.

Synthesis of the RO membranes

Microporous polysulfone membrane

Using a film applicator and after degassing, a homogeneous solution contained 74 wt% of DMF, 16 wt% of polysulfone, and 10 wt% of 2-methoxyethanol was casted onto a glass plate coated with a layer of polyester non-woven fabric. The plate was immediately immersed into a coagulation bath (deionized water, 15 °C). After 30 min, the polysulfone membrane supported on non-woven fabric was removed from the coagulation bath and separated from the glass plate. The membrane was washed thoroughly with deionized water and stored at 5 °C in a refrigerator.

TFC membranes

The ultra-thin polyamide layer of TFC membrane was prepared by interfacial polymerization of MPD in aqueous phase and TMC in organic phase on the polysulfone support. The aqueous solution adopted in this study contained 2.0 wt% of MPD, 0.05 wt% of sodium dodecyl sulfate, 5.0 wt% of DMSO and 2.0 wt% of glycerol. The pH of the aqueous solution was adjusted to 6–8 by using camphor sulfonic acid solution of 1.0 wt% and sodium hydroxide solution of 0.6 wt%. As for the organic phase containing 0.23 wt% of TMC, it was prepared by dissolving 1.38 g of TMC in 600 g of Isopa G.

First, the microporous polysulphone support was clamped between two Teflon frames (thickness: 0.8 cm; inner length: 18 cm; and inner width: 15 cm), then put in contact with the aqueous amine solution for 10 s to allow penetration of the solution into the pores. The surface was then rolled with a smooth steel roller to remove the excess solution. Afterwards, the top surface of the polysulphone membrane was exposed to the organic solution for 10 s, and left in air for 30 s at room temperature. The curing of the as-prepared membrane was performed at 90 °C for 6 min. Finally, the resulted membrane was rinsed with deionized water for 10 min, dipped in an aqueous solution containing 9 wt% of glycerol for 3 min for the coating of poly(vinyl alcohol) on the polyamide skin layer, and then dried at 80 °C in an oven for 10 min.

Characterization methods

The TFC membranes were washed several times with deionized water, and vacuum dried before characterization. The SEM analysis of samples was performed with FEI Quanta 200 equipment. The AFM analysis for surface morphology was conducted with Agilent 5500 AFM/SPM using silicon probes (curvature radius 10 nm) in the tapping mode.

Hydrophilicity of the polyamide skin layer was estimated by water contact angle measurement. The angle between the membrane surface and air-water interface was measured at 25 °C using the sessile-drop method over a DSA100 contact angle analyzer (KRUSS GmbH Co, Germany). Ten random locations of the membrane surface were tested for contact angle measurement, and the average was adopted to minimize experimental errors.

Evaluation of RO performance

The RO performance tests of the membranes were performed using a cross-flow membrane filtration equipment. There were three parallel permeation cells having an active surface area of 64 cm2 (8 cm × 8 cm) and a channel height of 2 mm. The membrane coupons were loaded with the active skin layer facing the incoming feed. All the permeation tests were conducted at 25 ± 1 °C and pH = 7.0–8.0 with the feed stream rate set at 0.25 m3/h.

At the beginning of each experiment, the membrane coupon was compacted using deionized water at 100 psi for at least 1 h. Then an appropriate volume of NaCl solution was added to reach the desired salt concentration, and the membrane was equilibrated within 60 min under a selected working pressure. When a stable flux was obtained, the water flux was determined by direct measurement of the permeate flow as follows: 
formula
1
The salt rejection level was calculated using the following equation: 
formula
2
The salt concentration of the feed and that of the permeated solution were measured using a conductivity meter (Oakton CON 11). To minimize experimental errors, each measurement of membrane separation property was duplicated with a total of three samples, and the averaged results were adopted.

Performance degradation tests

The membrane coupons (effective area: 64 cm2) were compacted using deionized water at 100 psi for at least 1 h on the cross-flow membrane filtration equipment before the start of the degradation test. Then NaCl solution (concentration: 1,500 ppm) was added to the feed tank. In the first 1 h of the experiment, the water flux and salt rejection level of the membranes were determined after a stable flux was reached (150 psi, 25 ± 1 °C, pH = 7.0–8.0). After 24 h, the water flux and salt rejection level of the membranes were again measured. To minimize experimental errors, each measurement was duplicated with a total of three samples, and the averaged values were calculated.

RESULTS AND DISCUSSION

Morphological study

Shown in Figures 1 and 2 are the surface SEM images of the MPD/TMC and MPD/TMC/DMSO/glycerol membranes, respectively. It is observed that both membranes show small-scale surface roughness of a ‘ridge-and-valley’ structure. In the SEM images, the white parts represent ridges and the black parts correspond to valleys, similar to those of the previous report (Ma et al. 2016). The surface morphology of the MPD/TMC/DMSO/glycerol membrane is obviously different from that of the MPD/TMC membrane. The latter shows a ‘ridge-and-valley’ structure that is more compact, implying a skin layer rougher than that of the former.
Figure 1

SEM image of the MPD/TMC membrane.

Figure 1

SEM image of the MPD/TMC membrane.

Figure 2

SEM image of the MPD/TMC/DMSO/glycerol membrane.

Figure 2

SEM image of the MPD/TMC/DMSO/glycerol membrane.

In AFM analysis, the parameters of surface roughness are: maximum peak-to-valley distance , average height , average roughness , root means square roughness , and relative surface area (Δ) (Al-Jeshi & Neville 2006).

The relative surface area Δ is defined as the actual surface area divided by the planar area (Tiraferri et al. 2011). denotes the height difference between the highest peak and the lowest valley in the selected areas: 
formula
3
is the mean roughness defined as the average deviation of peaks and valleys from the center plane. is expressed by the following: 
formula
4
In Equation (4), S is the specific surface area, is the height in the specified area, a and b are the two side lengths of the area. And is the mean height which is calculated by the equation: 
formula
5
is given by the following equation: 
formula
6

All the parameters of surface roughness, such as and are obtained from the AFM images by an AFM software program.

The high values of and low value of Δ are in accord with their significant effect on surface roughness. Figures 3 and 4 show the AFM surface images with a projection area of 10 μm × 10 μm for MPD/TMC and MPD/TMC/DMSO/glycerol. The quantitative analyses of surface roughness were performed with three to five replication images for the individual membrane, and the arithmetic means are depicted in Table 1. According to the AFM results, the MPD/TMC membrane has surface roughness lower than that of the MPD/TMC/DMSO/glycerol membrane. The AFM results are consistent with those of SEM study, both showing surface features of high similarity. It is noted that compared to the SEM images, the AFM ones are brighter and sharper.
Table 1

Parameters of surface roughness for the TFC membranes

Membrane z0(μm) Rp−v(μm) Ravg(μm) Rrms(μm) Δ 
MPD/TMC 0.165 0.493 0.053 0.066 1.26 
MPD/TMC/DMSO/glycerol 0.180 0.496 0.057 0.070 1.22 
Membrane z0(μm) Rp−v(μm) Ravg(μm) Rrms(μm) Δ 
MPD/TMC 0.165 0.493 0.053 0.066 1.26 
MPD/TMC/DMSO/glycerol 0.180 0.496 0.057 0.070 1.22 
Figure 3

AFM image of the MPD/TMC membrane.

Figure 3

AFM image of the MPD/TMC membrane.

Figure 4

AFM image of the MPD/TMC/DMSO/glycerol membrane.

Figure 4

AFM image of the MPD/TMC/DMSO/glycerol membrane.

Hydrophilicity of TFC membranes

The surface hydrophilicity of the composite membranes was assessed by water contact angle and surface solid–liquid interfacial free energy. Generally speaking, a contact angle of less than 90 ° is hydrophilic, while larger than 90 ° is hydrophobic. The lower the contact angle is, the greater the tendency for water to wet a surface (Yu et al. 2009).

The relative hydrophilicity of TFC membranes was also investigated by surface solid–liquid interfacial free energy (−ΔGSL), which is a modified version of the Young–Dupre equation and widely used in this kind of study (Wang et al. 2013). The solid–liquid interfacial free energy is expressed as: 
formula
7
where θ is the average contact angle, is the pure water surface tension, and Δ is the relative surface area. For pure water at 25 °C, the value is 72.8 mJ/m2. The solid–liquid interfacial free energy of a smooth surface ranges from to , and the measured contact angle ranges from 90° to 0°, correspondingly. For a rough surface, the contact angle is smaller than what it would be on a smooth surface of the same material (Ghosh et al. 2008). In comparison to the observed contact angles, the values of the solid–liquid interfacial free energy are better hydrophilicity representation of a surface. The results of the hydrophilicity values for the present study are given in Table 2. It can be seen that the contact angle of the MPD/TMC/DMSO/glycerol membrane is lower while the value larger than that of the MPD/TMC membrane. In other words, the former is more hydrophilic than the latter. Since the wettability of a solid surface can be related to surface geometrical structure, and is attributable to the surface solid–liquid interfacial free energy, it is deduced that the roughness of the skin layer of the MPD/TMC/DMSO/glycerol membrane is the major factor that contributes to the higher hydrophilicity in comparison to that of the MPD/TMC membrane.
Table 2

Variation of hydrophilicity and flux decline level of the two TFC membranes

Membrane Contact angle (deg) −ΔGSL(mJ/m2Flux decline level (%) 
MPD/TMC 58 ± 4.1 96.8 10.2 
MPD/TMC/DMSO/glycerol 49 ± 4.2 102.5 12.3 
Membrane Contact angle (deg) −ΔGSL(mJ/m2Flux decline level (%) 
MPD/TMC 58 ± 4.1 96.8 10.2 
MPD/TMC/DMSO/glycerol 49 ± 4.2 102.5 12.3 

The hydrophilic property of a TFC membrane is closely related with its degradation. The experiment to show the dynamic performance of the TFC membranes was conducted using a NaCl solution that was 1,500 ppm in concentration. From Table 2, it is clear that the MPD/TMC/DMSO/glycerol membrane suffered higher loss of water flux than the MPD/TMC membrane. In other words, the value is negatively related to the degradation property of the TFC membrane. It can be seen that the MPD/TMC membrane has better resistance to degradation than the MPD/TMC/DMSO/glycerol membrane because the MPD/TMC membrane is lower in hydrophilicity.

Permeation evaluation of the TFC membranes

To evaluate the effects of DMSO and glycerol addition on separation performance, we tested the polyamide RO membranes in terms of water flux and salt rejection ability using NaCl solutions of different concentrations (150 psi, 25 ± 1 °C, pH = 7–8) as well as under different working pressures (1,500 ppm NaCl concentration, 25 ± 1 °C, pH = 7–8). Presented in Figure 5 are the separation performances of the two TFC membranes. The water flux of the MPD/TMC/DMSO/glycerol membrane is larger than that of the MPD/TMC membrane, while there is only slight difference in salt rejection level. From Figure 6, one can see that under different working pressures the rejection abilities of the RO membranes are similar while the MPD/TMC/DMSO/glycerol membrane shows higher flux performance.
Figure 5

Separation performance of the TFC membranes versus the variation of NaCl concentration. Test conditions: 150 psi, 25 ± 1 °C, pH = 7.0–8.0.

Figure 5

Separation performance of the TFC membranes versus the variation of NaCl concentration. Test conditions: 150 psi, 25 ± 1 °C, pH = 7.0–8.0.

Figure 6

Separation performance of the TFC membranes versus the variation of working pressures. Test conditions: NaCl concentration 1,500 ppm, 25 ± 1 °C, pH = 7.0–8.0.

Figure 6

Separation performance of the TFC membranes versus the variation of working pressures. Test conditions: NaCl concentration 1,500 ppm, 25 ± 1 °C, pH = 7.0–8.0.

The enhanced permeability can be related to the higher hydrophilicity of the MPD/TMC/DMSO/glycerol membrane as a result of DMSO and glycerol addition during interfacial polymerization. The parameters of surface roughness also suggest enhancement of water flux because the rough surface enlarges the effective contact area (Ma et al. 2016). The surface of the MPD/TMC/DMSO/glycerol membrane is rougher than that of MPD/TMC membrane. The rougher surface is a result of fluctuating interface through reducing the immiscibility between aqueous and organic phases by DMSO. And the surface roughness of the MPD/TMC/DMSO/glycerol membrane enables better contact with water molecules, and hence the higher permeability. In this work, the water flux is positively related to surface roughness, which coincides well with literature records. Besides, glycerol is a flux-preserving agent during interfacial reaction, preventing the loss of porosity during oven drying (Kuehne et al. 2001). Hence, we deduce that the higher permeate flux of the MPD/TMC/DMSO/glycerol membrane in comparison to that of the MPD/TMC membrane is a cooperative effect of DMSO and glycerol on the membrane structure during interfacial polymerization.

CONCLUSION

The aim of the study is to investigate the effects of DMSO and glycerol additives on the performance of polyamide MPD/TMC/DMSO/glycerol membrane. The results reveal that the use of the additives strongly influences the properties of the membrane. According to the SEM and AFM results, there is increase of surface roughness when DMSO works to increase the miscibility of the aqueous and organic phases. The studies on hydrophilicity indicate that the MPD/TMC/DMSO/glycerol membrane is more hydrophilic than the MPD/TMC membrane. With similar salt rejection ability, the former shows water flux significantly larger than that of the latter. It is deduced that the cooperative effect of DMSO and glycerol during interfacial polymerization results in a membrane surface higher in roughness and hydrophilicity.

REFERENCES

REFERENCES
Akbari
A.
Fakharshakeri
Z.
Mojallali Rostami
S. M.
2015
A novel positively charged membrane based on polyamide thin-film composite made by cross-linking for nanofiltration
.
Water Science & Technology
73
(
4
),
776
789
.
Dominik
S.
Angela
N.
Jeet
S.
Renato
F.
Daniele
M.
Fabien
B.
Sumudhu
S. P.
Kristina
E.
Beat
L.
Konstantin
R.
Jacqueline
T.
Mittmann
T.
2014
Oligodendrocyte precursor cells modulate the neuronal network by activity-dependent ectodomain cleavage of glial NG2
.
PLoS Biology
12
(
11
),
e1001993
e1001993
.
Ghosh
A. K.
Jeong
B. H.
Huang
X. F.
Hoek
E. M. V.
2008
Impacts of reaction and curing conditions on polyamide composite reverse osmosis membrane properties
.
J. Membr. Sci.
311
(
1–2
),
34
45
.
Jahangiri
F.
Mousavi
S. A.
Farhadi
F.
Vatanpour
V.
Sabzi
B.
Chenari
Z.
2015
Effect of CO2-laser irradiation on properties and performance of thin-film composite polyamide reverse osmosis membrane
.
Korean J. Chem. Eng.
33
(
3
),
1028
1036
.
Kim
H. J.
Choi
K.
Baek
Y.
Kim
D.
Shim
J.
Yoon
J.
Lee
J.
2014
High-performance reverse osmosis CNT/polyamide nanocomposite membrane by controlled interfacial interactions
.
Appl. Mater. Interfaces
6
(
4
),
2819
2829
.
Kuehne
M. A.
Song
R. Q.
Li
N. N.
Petersen
R. J.
2001
Flux enhancement in TFC RO membranes
.
Environmental Progress
20
(
1
),
23
26
.
Kurt
V.
Bell
T. W.
Victor
V. P.
Anne
G.
Sam
N.
Sandra
L.
Dominique
S.
Enno
H.
Kalies
K. U.
Marsh
M.
2014
Signal peptide-binding drug as a selective inhibitor of co-translational protein translocation
.
PLoS Biology
12
(
12
),
e1002011
e1002011
.
Ricci
B. C.
Ferreira
C. D.
Marques
L. S.
Martins
S. S.
Amaral
M. C. S.
2016
Assessment of nanofiltration and reverse osmosis potentialities to recover metals, sulfuric acid, and recycled water from acid gold mining effluent
.
Water Science & Technology
74
(
2
),
367
374
.
Shen
W. F.
Dong
L. C.
Wei
S. A.
Li
J.
Benyounes
H.
You
X. Q.
Gerbaud
V.
2015
Systematic design of an extractive distillation for maximum-boiling azeotropes with heavy entrainers
.
AIChE J.
61
(
11
),
3898
3910
.
Tiraferri
A.
Vecitis
C. D.
Elimelech
M.
2011
Covalent binding of single-walled carbon nanotubes to polyamide membranes for antimicrobial surface properties
.
ACS Applied Materials & Interfaces
3
(
8
),
2869
2877
.