Mechanism of grafted/crosslinked CA-RO membrane
As shown in Fig. 1, grafting of CA-RO membrane 1 is conducted by reaction with different concentration of the hydrophilic monomer N-IPAAm 3 at in the presence of KPS to obtain the grafted CA/RO membranes 2. Firstly, the surface of the CA-RO membrane was treated with 0.014 wt % NaOH for partially deacetylation to facilitate the polymerization on the surface of the membrane 1. Secondly, the surface was initiated using KPS initiator to generate free radicals on the surface of the CA-RO membrane 2 could be attributed to the abstraction of hydrogen atom of the free C6-OH group due to its reactivity as a primary alcohol5,33. The propagation step started when the in situ radical terminals are added to vinyl terminal of N-IPAAm monomer to furnish the targeted intermediate 4. Portion of the latter intermediate 4 is quenched with water to terminate the reaction and produces the grafted CA/RO membrane 5, while the rest of 4 is further treated with different concentrations of MBAAm 6 to produce the targeted grafted/crosslinked CA-RO membrane 7 by creating a new link between the two monomers. The addition of the hydrophilic monomer N-IPAAm introduces polar functional groups to the surface of the membrane and improves the hydrophilicity. The grafting process could be influence the performance of the CA-RO membrane by enhancing the salt rejection and water flux because it forms a new hydrophilic layer with good coverage to the surface to enhance the dense layer properties and treat any defects on the membrane surface. To avoid the probability of the monomer penetration or blocking of the pores which decreases the flux of the membrane by grafting, the crosslinker was inserted to increase the interchain between polymer chain and formed a net structure28.
Structural property of the grafted CA-RO membrane
The FTIR spectra of CA-RO, grafted CA-RO and grafted/crosslinked CA-RO membranes are presented in Fig. 2. The FT-IR spectrum of the CA-RO membrane exhibits absorption bands at 3475 cm−1, corresponds to OHstr, at 2941 cm−1 attributes to the H-C groups, at 1739 cm−1 belongs to CO ester group, and at 1257 cm−1 due to the C-O-C symmetric ether bond as illustrated in Fig. 2. The grafted CA-RO membrane absorption band at υ 1649 cm−1 is attributed to the CONH amide, at 1550 cm−1 corresponds to N-H bending, a broad band at 3500 cm−1 is belonged to OH and N-H str and a band at 2953 cm−1 is due to the H-C groups12,34. On the other hand, the grafted/crosslinked membrane displays two successive sharp absorption bands at 1743 cm−1 and 1747 cm−1, correspond to the CO group as shown in Fig. 2. Noteworthy, the intensity of these bands is increased with increasing grafting agent concentration. The absorption bands at 1649 cm−1, 1537 cm−1 and 1404 cm−1 correspond to the CO amide, NH bending and CN amide bands, respectively11,12.
The morphologies of the surface, bottom and the cross section of the CA-RO membrane and grafted 0.05%-, 0.1%-, 0.2%- and 0.3 wt %-N-IPAAm membranes are dispalyed in Fig. 3. The SEM image of CA-RO membrane shows a dense top layer with ridge, valley shape and porous sublayer as presented in Fig. 3(a). However, the cross-section photograph has small voids and large portion sponge-like structure, indicating the highly viscous casting solution35. The SEM images of the grafted CA-RO membranes using the above-mentioned grafting concentrations are shown in Fig. 3(b–e). The images depict denser skin top layers which increase from 0.92 µm of the pristine CA membrane to 1, 1.2, 1.3 and 1.33 µm, respectively for the different N-IPAAm concentrations of 0.05, 0.1, 0.2 and 0.3 wt % with smoother top surfaces than the pristine. Noteworthy, the ridge and valley shape are decreased and a high porosity with better pore shape and density is observed with increasing the grafting concentration.
The cross-section images of the grafted samples exhibit interesting morphologies. The cross sections have also asymmetric structures with dense top layers and channel or macrovoids like structures. For instance, the SEM image of a 0.05 wt % sample shown in Fig. 3(b) displays a channel-like structure, while the 0.1 wt % membrane (Fig. 3c) displays two layers of macrovoids and sponge like structure. Further increasing of the grafting agent concentrations to 0.2 wt %, and 0.3 wt% affords macrovoids layer morphologies as presented in Fig. (3d,e).
The bottom surfaces illustrate porous structures and the number of pores is enhanced and increased with the excess of the grafting concentration. These pores work as a water pump, while the fingers and macrovoids work as cannels. Consequently, if these pumps are blocked the water will not pass through the cannels and the pores of the bottom layer are the dominant parameter which will influence on the water flux.
The 0.1 wt % N-IPAAm grafted CA-RO membrane was selected for further crosslinking and morphological studies. The crosslinking process is conducted using 0.006 wt%, 0.01 wt%, 0.013 wt% and 0.02 wt% of the crosslinker MBAAm. The SEM images exhibits higher thickness of denser top layer than the grafted membranes in the range from 1.7 to 2 µm from 0.006 to 0.02 wt % of MBAAm. The cross-section images shown in Fig. 4(a–d) clearly indicate the formation of finger-like macrovoids structure. Noticeably, larger numbers of the macrovoids are observed for the 0.013 wt % MBAAm sample (Fig. 4c). Further rising of the crosslinking concentration up to 0.02 wt % furnishes the tight and lower voids structure as demonstrated in Fig. 4(d).
For structure-morphology correlation purpose, selected samples of CA-RO, 0.1 wt % N-IPAAm-grafted CA-RO and 0.1 wt % N-IPAAm grafted-/0.013 wt% MBAAm crosslinked CA-RO membranes are subjected to AFM scan area studies. The bright and dark regions are the peaks and valleys, respectively. Ra represents the average roughness and Rz is the difference in height between the average of the five highest peaks and the five lowest valleys along the assessment length of the profile36. Ra of the CA-RO membrane is 42.99 nm as shown in Fig. 5(a) which decreases to 11.6 nm for the grafted one (Fig. 5b) and to 27.1 nm in case of the grafted/crosslinked membrane. These results can be explained based on the grafted and grafted/crosslinked hydrophilic monomers fill the pores onto the surface and covered any defects and become more homogenous. Consequently, it is expected that the fluxes of the grafted and grafted/crosslinked RO membrane are declined. Whereas the salt rejection will be improved. Generally, the CA-RO surface roughness is susceptible to fouling compared to smooth surface37. Compared to the surface roughness of grafted CA-RO samples, the grafted/crosslinked CA/RO samples exhibit higher roughness than the grafted. This is attributed to the lower solubility of the crosslinker on the CA surface. Moreover, these results are attributed to the formation of additional network structure linked between the polymers chains which could be affected the smooth of the surface and formation of peaks and valleys more than the grafted monomer. Table (1) shows Ra and Rz values of the pristine CA, grafted CA and grafted/crosslinked CA-RO membranes. The grafted membrane presents a lower Ra and RZ compared to the pure CA-RO membrane. On the other hand, the crosslinking increases the Ra and Rz values compared to the grafted CA-RO membranes.
Effect of grafting on the hydrophilicity of grafted CA-RO membranes
The contact angle is a measurement of the hydrophilicity of the membrane surface. Higher hydrophilicity reduces the fouling and vice versa. The contact angle is measured by placing a microliter droplet of distilled water on the membrane surface. The angle is determined for both sides of the droplet and the mean value is then calculated. The contact angles versus concentration change of the grafting and/or crosslinking agents are presented in Fig. 6. From Fig. 6(a), the contact angle of pristine CA-RO membrane is 66.28°. This value is decreased to 58°, 49.7°, 53.8° and 49.6° by increasing the grafting concentrations to 0.05 wt%, 0.1 wt%, 0.2 wt % and 0.3 wt %, respectively. Interestingly, the contact angles of samples measured upon addition of different crosslinker concentrations of 0.006 wt%, 0.01 wt%, 0.013 wt%, and 0.02 wt% to the 0.1 wt % N-IPAAm grafted membrane are 59.02°, 57.85°, 52.6° and 52.6°, respectively. It is observed that the contact angle of the crosslinked CA-RO membrane is increased with respect to the grafted CA-RO membrane as a reference from 49.7° to 59. 02° for 0.006 wt % of the crosslinker. This is attributed to the raising of the roughness of the crosslinked RO membrane. Furthermore, increasing the crosslinker concentrating to 0.013 and 0.02 wt % increases the polar functionality of the surface and, thus improves the hydrophilicity to 52°. The hydrophilicity and hydrophobicity of the RO membrane are controlled by the electrostatic and/or hydrogen-bond interactions between the water molecules and functional groups. Because of the strong hydrogen-bond interactions between the water molecules and the surface functional groups, the RO-membrane affinity to water molecules is strong and the water droplet is spread on the membrane surface. The improvement of the hydrophilic properties for the grafted and crosslinked RO membrane is due to the presence of hydrophilic atoms (O and N) increases the wettability of the membrane. These groups form hydrogen bonds with the water molecules.
Salt Rejection and water flux of untreated CA-RO membranes
A testing RO cell was used to measure the flux and salt rejection values of the untreated CA-RO membrane at different operating pressures as represented in Fig. 7. It is found that the salt rejection of CA-RO membrane is 93.7% at 12 bar and decreases to 74.5 at 22 bar. The rejection decline is referred to the concentration polarization which formed boundary NaCl layer close to the surface and thus, the observed rejection of the feed bulk concentration is lower than the real rejection of the solute concentration. The permeate water flux of this membrane is 1.6 L/m2h at 12 bar and increases to 5.17 L/m2h at 22 bar. Hence, it is concluded that the water flux is increased proportionally with increasing the operating pressure and is matched with the solution diffusion model38.
Salt rejection and water flux of the grafted CA-RO membranes
The salt rejection and permeate water flux of the treated CA-RO membranes with different grafting concentrations are evaluated and the results are shown in Fig. 8(a,b). The 0.05% grafted, 0.1% grafted, 0.2% grafted, and 0.3% grafted membranes have salt rejection values of 97.5%, 98.9%, 97.9% and 96.7%, respectively at the 12 bar (Fig. 8a). Interestingly, conducting the measurements at 18 bar, the salt rejection values are 78.6%, 89%, 88%, 81.84%, respectively. The salt rejection results inversely proportional to increasing pressure. The small pores in the grafted membrane exhibit better salt rejection than the dense skin untreated analogue39. On the other hand, the permeate water flux of the above-mentioned grafted samples conducted at 12 bar are, respectively, 1.18 L/m2h, 1.31 L/m2h, 1.03 L/m2h and 1.19 L/m2h as presented in Fig. 8(b). However, conducting the tests of the previously mentioned grafted membranes at 18 bar, the water fluxes are improved to 6.8 L/m2h, 2.81% L/m2h, 2.9 L/m2h and 2.59 L/m2h, respectively. This result could be explained in terms of either the formation of compact layer or the chain pores blocking by the grafting monomer molecules28,40.
Salt rejection and water flux performance of the grafted/crosslinked CA-RO membranes
Based on the above results of different grafting concentrations, the relatively efficient 0.1 wt % N-IPAAm-grafted membrane is selected for further MBAAm crosslinking-performance studies as displayed in Fig. 9. Salt rejection results conducted at 12 bar of (0.1% grafted-/0.006 wt % crosslinked-), (0.1 wt % grafted-/0.01 wt % crosslinked-) and (0.1 wt % grafted-/0.013 wt % crosslinked) CA-RO membranes are 97.53, 97.0, and 94%, respectively as depicted in Fig. 9(a). Noticeably, the experiments at 26 bar for the same mentioned grafted/crosslinked CA-RO membranes, the salt rejection values are found to be 76.3, 67 and 78.2%, respectively. Likewise, the salt rejection of 0.1 wt % grafted/0.013% wt crosslinked-CA-RO membranes at 10 bar is 97.77% as shown in Fig. 9(a). It is noted that grafting/crosslinking process improved the salt rejection and this result could be explained in terms of increasing the crosslinker-CA matrix interaction, which subsequently increase the free volume throughout the surface, and improve water transport40. In addition, crosslinker may reduce the degree of crystallinity of CA and thus, facilitate water diffusion41,42.
Otherwise, conducting the salt rejection for 0.1 wt % grafted/0.02 wt % crosslinked CA-RO membrane leads to surface blocking and no water passage is noticed at 12 bar. However, water transport is only observed starting at 24 bar. For this specific membrane, salt rejection value is 99% and decreases to 77% on further increase of the operating pressure up to 34 bar. On the other hands, the above grafted/crosslinked membranes appear water flux at 12 bars of 2.23 L/m2h, 1.7 L/m2h and 3.25 L/m2h, respectively as illustrated in Fig. 9(b) and the water flux results at 26 bar are 11.67 L/m2h, 22.9 L/m2h and 16.12 L/m2h, respectively.
Figure 10 summarizes the results of the water flux and salt rejection versus different concentrations of grafting N-IPAAm onto the CA-RO membrane surface at 12 bar. It can be concluded that the 0.1 wt % N-IPAAm/CA-RO membrane produced the highest salt rejection of 98.9% with water flux 1.3 L/m2h. On the other hand, the salt rejection and water flux at 26 bar of 0.1 wt % grafted/0.013 wt % crosslinked CA-RO membrane are 78.2% and 16.12 L/m2h, respectively.