Selecting the best neat membrane. The effect of polymer concentration (wt%) on the permeation flux and rejection are provided in Table 1.
Table 1
The flux of permeation and rejection at different PVC concentrations.
Membrane
|
PVC
(%)
|
Water permeate flux
( kg.m− 2h− 1)
|
Rejection
(%)
|
M1
|
13
|
68.90
|
51.5
|
M2
|
14
|
57.36
|
69.0
|
M3
|
15
|
46.53
|
71.0
|
M4
|
16
|
36.97
|
73.0
|
M5
|
17
|
23.95
|
86.0
|
M6
|
18
|
7.24
|
83.0
|
As shown in Table 1, the highest water flux is related to M1 which is synthesized with a PVC concentration of 13 wt%. On the other hand, this membrane exhibits the lowest rejection among the synthesized membranes. The highest rejection (86%) is obtained at a PVC concentration of 17 wt%. However, this membrane (M5) doesn’t show the highest flux of permeation. The best performance among the six lab-made membranes belongs to M5, which shows the highest rejection and a relatively high permeation (Table 1).
The effects of CNs on membrane structure and performance. The results of water permeation through mixed matrix membranes (M5-1 to M5-3) are provided in Table 2. The highest permeation and rejection is obtained by membrane M5-2.
Table 2
Performance results of mixed matrix membranes
Membrane
|
Nanoparticles
(%)
|
Water permeate flux
( kg.m− 2h− 1)
|
Rejection
(%)
|
M5
|
0
|
23.95
|
86.00
|
M5-1
|
1
|
35.62
|
89.25
|
M5-2
|
2
|
40.72
|
94.78
|
M5-3
|
3
|
38.19
|
91.19
|
Figure 1 shows the flux of water permeation for mixed matrix membranes containing different concentrations of CNs. Enhancement in the flux of permeation and rejection of mixed matrix membranes can be explained by FESEM images and contact angle results.
As can be seen, the flux of membrane M5-3 has decreased compared to membrane M5-2, due to the accumulation of CNs on the surface of the M5-3 membrane (Fig. 4c). The accumulation of nanoparticles on the membrane surface can be clearly apperceived in Fig. 4c. As a result, according to Table 2, it was observed that the highest water flux and also the highest rejection related to the M5-2 mixed matrix membrane. Figure 2 shows the results of contact angle of neat as well as the mixed matrix membranes. As we know, the contact angle indicates the degree of hydrophilicity of the membrane surface, and factors such as porosity, functional groups, surface charge, and membrane roughness affect the contact angle18,23. The lower the contact angle, the greater the hydrophilicity. Reduction of the contact angles of the mixed matrix nanofiltration membranes are observed due to the presence of hydroxyl groups, extremely hydrophilic, and very high porosity of CNs24,26. CNs showed good dispersion at low concentrations in the polymeric matrix. However, by increasing the clay concentration beyond 3 wt%, part of nanoparticles accumulated on the membrane surface. Increasing the viscosity of the polymeric solution at high concentrations of nanoparticles affects the mechanism of membrane formation. The latter increase leads to an increase in the thickness of the top layer and a denser structure with lower porosity is formed. Lower porosity causes higher roughness and also lower surface hydrophilicity25,10. Higher roughness can be explained by accumulation of nanoparticles on the membrane surface (Fig. 4c). As can be seen in Figs. 4e and 4f, upon increasing the nanoparticles' concentration from 2 to 3 wt%, the thickness of the top layer increased. Furthermore, the porosity decreases and pore size increases (Table 3). As such, the flux of permeation and rejection of M5-3 decreases comparted with M5-2. Generally, the membrane porosity increases and the pore size decreases in the presence of CNs. Embedding hydrophilic CNs in the PVC membrane increases the DMF/THF-water exchange rate and creates a porous structure. As the concentration of nanoparticles increases, the porosity and pore size of the membranes increases. However, by further increasing the amount of nanoparticles, these nanoparticles accumulate on the membrane surface and reduce the porosity and pore size. The addition of nanoparticles created more gaps between the hydrophilic groups on the CNs and the hydrophobic chains of PVC, which led to the creation of a more porous structure26–28.
Table 3
Porosity and average pore size
Membrane
|
Porosity (%)
|
Average pore radius (nm)
|
M5
|
74.33
|
8.4
|
M5-1
|
80.11
|
4.3
|
M5-2
|
89.67
|
1.8
|
M5-3
|
85.29
|
2.7
|
The FTIR spectra of the mixed matrix membrane with nanoclay concentration of 2 wt% is shown in Fig. 3. The C-Cl peak at 600 cm− 1, CCl-H peak at 950 cm− 1, and CH2 peak at 2900 cm− 1 are strong identifications of PVC, which are present in both PVC and PVC/nanoclay membranes29.
For CNs, the peaks in the range of 1100, 1250, 3240, and 3450 are related to the Si-O stretching bond (silicon group), the -COOC- stretching bond, and the -OH stretching bond, and the -OH stretching bond (hydroxyl group), respectively30,31. According to the above interpretations, it can be concluded that CNs have been well dispersed on the membrane structure.
Incorporation of nanoparticles as a hydrophilic additive affects the kinetics and thermodynamics of the phase inversion process, which leads to different morphologies of the synthesized membrane32. SEM show that the number of pores increases with the composition of CNs (Fig. 4e). Compared to neat PVC nanofiltration membranes, the mixed matrix membrane provides a surface with a porous structure and no roughness (AFM analysis, Fig. 5).
Also decrease in roughness can be observed in the mixed matrix membrane containing 2 wt% of CNs (Fig. 4b) compared to the neat membrane (Fig. 4a).
One of the important morphological features that can affect the performance of the membrane is the structure of the inner pores of the membrane33. The effect of additives on the internal morphology of the membrane was investigated by cross-sectional FESEM analysis of the membrane. According to the cross-sectional images, it was observed that the membranes have a normal asymmetric and porous structure with a thin top layer (less than 2.5 µm) followed by a layer with a finger structure and macro-sized cavities (Figs. 4d, 4e, 4f). The morphology of the mixed matrix membranes containing CNs is characterized by a coarse surface layer of pores, a bottom layer having a spongy pore-like structure with fine pores, and an intermediate layer with finger-like pores and fine pores (Figs. 4e, 4f)14,16,34. The structure of the inner cavities of the membrane is significantly affected after adding CNs. CNs has a layered structure (Fig. 8) and in interaction with the polymer, its layers are separated and distributed throughout the polymeric matrix15,35. The presence of CNs in the polymer solution makes the solution more thermodynamically unstable, which leads to rapid phase separation during immersion in the coagulation bath. Increasing the affinity of insoluble polymer solution due to the presence of hydrophilic groups in the membrane structure leads to faster polymer deposition. As a result, the growth of the top layer is reduced and the formation of finger pores under the layer is improved, and eventually, a more porous structure is formed (Figs. 4d and 4e). FESEM images of the cross-sectional area of the mixed matrix confirms the presence of CNs in the membrane structure17,36.
The cross-sectional size of the finger cavities of the mixed matrix membranes with the nanoparticles is larger than that of the neat membrane (Figs. 4e, 4f compare with Fig. 4d). Since CNs have a high tendency to absorb water, the exchange rate of solvent-nonsolvent as a key factor of membrane formation mechanism can be reduced37. Fast phase inversion and the formation of the membrane surface layer prevent the penetration of solvents to coagulation bath and as a result increase the time of solvent-nonsolvent exchange. This leads to more growth of the polymer-free phase, and eventually larger cavities are formed in the sub-layer of the membrane, which leads to increased membrane porosity16,38,39.
Using three-dimensional surface images, the surface roughness of the membranes was also investigated (Fig. 5). The dark areas and the light areas are related to the valleys and peaks of the membrane surface, respectively. According to the results, it was observed that the incorporation of nanoparticles leads to a significant reduction in surface roughness. The rather uniform dispersal of nanoparticles on the membrane surface and the probabilistic scattering of the nanoparticles to the membrane surface as result of their low-densities formed a smoother surface for the mixed matrix membrane. The surface morphology in the images also shows that the CNs are relatively dispersed uniformly in the polymeric solution, which reduces the roughness of the membrane surface. Surface roughness is considered a criterion for the membrane surface deposition. The results of roughness analysis certified the efficiency of the mixed matrix membranes in eliminating scale and adsorbent from the membrane surface (increasing membrane permeation flux and decreasing membrane fouling). One of the reasons for rougher surfaces could be the effects of increased colloidal interactions and adsorption on the membrane surface39,40.
Membrane surface roughness parameters such as mean surface roughness (Ra), mean second root roughness (Rq), maximum profile valley depth (Rv), and maximum peak profile height (Rp) are given in Table 4.
Table 4
Surface roughness parameters
Membrane
|
Rp (nm)
|
Rv (nm)
|
Rq (nm)
|
Ra (nm)
|
M5
|
35.2841
|
-43.2945
|
10.8915
|
8.6482
|
M5-2
|
24.0874
|
-21.2481
|
6.6390
|
5.1642
|
XRD test is a non-destructive method with multiple applications that provide comprehensive information on the chemical composition as well as the crystalline structure of both natural and industrial materials41. XRD analysis was used to characterize and confirm the crystal structure of the mixed matrix membrane in the presence of CNs. As shown in Fig. 6, the XRD pattern contains a peak at 2θ = 25º for pure PVC, 2θ = 24.9º for M5-1, and 2θ = 22º for M5-2. In other words, the above results show that the silicates of the CN layer are located in the PVC matrix and shows that the CNs do not change the crystal structure of PVC in the membrane. It was also observed that the neat membrane curve refers to a non-crystalline or amorphous polymer and remains unchanged compared to the XRD pattern of the origin polymer. It is also observed that for the membranes containing nanoparticles, the characteristic peaks of montmorillonite in the presence of membrane curves were not detected, indicating a peeled and/or partially peeled structure. The reason why the characteristic peak of montmorillonite does not appear can be attributed to the efficient fusion of polymer chains between the clay layers13,15,42.
EDAX spectroscopy was performed to obtain the elemental composition of the membrane. X-ray energy dispersive spectroscopy was performed to confirm the outer surface of the membranes to confirm FESEM micrographs. Figure 7 shows the EDAX spectrum and the composition of the specified elements in the mixed matrix membrane. The corresponding atomic fractions are shown in Table 5. As the results in Table 6 reveal, silicon (Si) scattered on the outer surface of the membrane indicates the presence of CNs in the polymer matrix. The presence of Si can increase the hydrophilicity and porosity of the membrane. The micrographs also show the uniform dispersion of the nanoparticles in the PVC matrix (membrane surface without cracks).)16,44.
Table 5
Elemental analysis of the M5-2 membrane.
Element
|
Atom (%)
|
Mass Norm (%)
|
Mass (%)
|
Cl
|
94.82
|
95.69
|
31.90
|
Si
|
4.65
|
3.72
|
1.24
|
K
|
0.53
|
0.59
|
0.20
|
Ca
|
0.00
|
0.00
|
0.00
|
Sum
|
100.00
|
100.00
|
33.34
|
A comparison between the lab-made mixed matrix membrane with the previous reported studies reveals an enhancement in the performance of the synthesized membrane (Table 6). As can be seen, appropriate fluxes of rejection and permeation were simultaneously obtained with PVC/ nanoclay mixed matrix nanofiltration membrane. In other researches, rejection and permeation flux have not increased simultaneously. Reference 46, which has a higher rejection, show much less permeation flux than the current lab-made membrane. Only ref. 48 reported simultaneous enhancement in both permeation flux and rejection by incorporation of ZnO, which is much more expensive nanoparticles than nanoclay. In other words, in the current study, permeation flux and rejection are improved simultaneously by using a low cost and unlimited natural resources nanoparticle as an additive.
Table 6: Comparison between the present study and some recent studies
Reference
|
Contact angle
|
Rejection
(%)
|
Permeation flux
( kg.m-2h-1)
|
NP
|
Membrane type
|
12
|
66.0
|
88.0
|
48.50
|
Clay
|
PVDF
|
16
|
65.0
|
94.7
|
23.00
|
Clay
|
PS
|
45
|
59.5
|
91.2
|
19.40
|
SiO2
|
PES
|
45
|
64.3
|
89.5
|
16.20
|
ZnO
|
PES
|
45
|
55.3
|
88.2
|
23.32
|
GO
|
PES
|
45
|
60.5
|
96.9
|
18.20
|
Al2O3
|
PES
|
46
|
32.0
|
92.6
|
64.23
|
Ge-TA
|
PES-PVP
|
47
|
28.0
|
98.9
|
166.73
|
ZnO
|
PEES
|
Lab made membrane
|
63.3
|
94.8
|
40.71
|
Clay
|
PVC
|