Systematic investigation of MAX phase (Ti3AlC2) modified polyethersulfone membrane performance for forward osmosis applications in desalination

2.1 Materials

Polyether sulfone (PES) (MW = 58000 g/mol) was obtained from Solvay Advanced Polymer, (Belgium). Hexane (n-Hexane, 95 %) was obtained from Gain land Chemical Co. Ltd (UK). N,N-Di methyl acet amide (DMAc) (MW = 87.12 g/mol), MAX phase (Ti3AlC2) (Particle size 0.7 µm), Sodium Chloride, M-phenylene diamine monomer (MPD, >98 %), trimesoyl chloride monomer (TMC, >99.85 %) were all purchased from Sigma‐Aldrich (Shanghai, China). All chemicals were analytical grades and no further purification was required.

2.2 Membrane characterization

The morphological changes on the surface and cross-section of all membranes have been visualized by scanning electron microscopy (SEM) (Mira-3 FEG SEM, Tescan Orsay Italy, ITM CNR). Initially, the specimens were fractured in liquid nitrogen and sputter coated with a thin layer of gold.

An atomic force microscope (AFM, Bruker multimode 8 with Nanoscope V controller, Model TAP150, Burker) was harnessed to scan the surface topography of the samples.

The composition and functional groups of the membrane samples were identified by Fourier-transform infrared spectroscopy (FTIR) (PerkinElmer, universal ATR sampling). Wide range (4000–400 cm⁻¹) spectra were recorded.

The hydrophilicity of the membranes was conducted using the sessile drop method. 3 µl of DI water drops were placed on a flat membrane surface and the contact angle between the membrane surface and the drop was captured. For each sample, at least 10 drops at different locations were measured and the average value was recorded.

The gravimetric technique was utilized to determine the membrane's porosity. 2 cm * 2 cm samples of the membranes were cut and their dry weight was measured. The membranes were then soaked in DI water for 1 hr and excessive water drops were removed by a tissue and their weights were recorded. The porosity (ε) is the volume of the pores divided by the sum of the membrane's volumes. The porosity of the sheets was calculated using the following Eq. (1) (Ursino et al., 2020; Al-Ani et al., 2020):

The mean pore radius (rp, nm) of membrane substrates was calculated using the Guerout-Elford-Ferry Equation (Kadhim et al., 2020).

2.2.1

2.2.1 FO testing rig

A lab-scale filtration setup was used to test the intrinsic permeability parameters of the membranes, which identify how water and salt move across this layer. These parameters are water permeability, salt permeability, and salt rejection. The water flux refers to the water molecules that pass through a membrane from the feed side to the draw side. A larger water flux denotes a membrane's better permeability. Similarly, the reverse salt flux (RSF) is the amount of draw solute that will pass through a membrane backwards from the draw side to the feed side during a FO operation. A low RSF implies a high membrane interception capability. The schematic diagram of the FO system is illustrated in Fig. 1. The system comprises a FO membrane cell with external dimensions of 110 × 55 × 100 mm, internal dimensions of 46 × 92 mm, and 2.3 mm for the slot depth. The feed and draw solutions are separated by the cell's internal membrane. Two 1 L stainless steel tanks were used for the feed solution (FS) and draw solutions (DS). The feed and draw solutions were circulated throughout the system using two gear pumps (DP-75). The flow rates of the FS and DS were determined using two flow meters while the water flux in the FO system was measured using a digital balance. The same tanks were used to recycle the permeate from the FO cell. Each experiment lasted for about 120 min. Flowmeters were used to regulate the feed and draw solution flow rates. When the feed solution's volume changes (v) are divided by the module's effective membrane area (A) and a certain experimentation time (t), the result is the water flux J_w, which is determined by the following equation (Wang, 2022):

(4)

$$J_{W=\frac{\mathrm{\Delta }V}{A.t}}$$ where J_w is water flux (LMH), ΔV is volume change of FS (litre), A is the membrane area (m²), and t is the time of testing (h).

Close

Fig. 1

Schematic diagram of the FO test unit.

Export to PPT

Reverse salt flux (J_s) was estimated according to the following equation:

(5)

$$J_{S=\frac{\mathrm{\Delta }CV}{A.t}}$$ where Js is reverse salt flux (gMH), ΔC is the change in feed solution concentration (g/l), V is feed solution volume at time t (litre), A is the effective area of membrane (m²), and t is the experiment time (hr). The change in salt concentration at the feed side was monitored using the calibrated conductivity meter.

3.1 Scanning electron microscope (SEM) of the substrate membrane

The morphological changes in the surface and cross-section of the substrate, TFC membranes have been observed by SEM and depicted in Figs. 2 and 3. Fig. 2A and B displayed the cross-sectional and surface images of the control substrate membrane (M1), respectively. As could be seen, a typical structure of the PES membrane was observed in the cross-section image where a thin active layer and small thin micropores are supported on huge interconnected macrovoids. Meanwhile, a smooth top surface with some small pores could be detected. Adding the MAX phase to the PES polymeric matrix has induced several morphological changes in the cross-section. The number of micropores has significantly increased while smaller macrovoids were formed at the bottom surrounded by a spongey structure (Fig. 2D). Likewise, the number of pores has significantly increased as could be seen in Figs. 2E and S4 explain the surface section of B, E and H. This could be induced by the enhanced hydrophilicity associated with MAX phase incorporation in the PES polymeric matrix. Triplicating the content of the MAX phase (0.075 g) in the PES membrane substrate produced a denser membrane structure where much thinner micropores and sponge-like structures were formed (Fig. 2F). This agreed with the preceding literature and attributed that to the enhanced casting solution viscosity due to the presence of 2D MAX phase (Salehi et al., 2020). Increasing the casting solution induces a delayed mixing-demixing process during the NIPS and produces a dense membrane structure with a more sponge-like structure at the bottom of the membrane (Amini et al., 2019). Meanwhile, constructing the polyamide layer on the surface of the membranes was confirmed by the typical leaf-like structure as observed in Fig. 2C, F and I. The PA was uniform and homogenously rendered on the PES substrate with a defect-free structure for all samples. Fig. S3 explain the cross section of C, F and I.

Close

Fig. 2

SEM surface and cross-sectional images of (A) M₁ cross-section, (B) M₁ surface, (C) M₄ surface, (D) M₇ cross-section, (E) M₇ surface, (F) M₁₁ surface, (G) M₉ cross-section, (h) (M₉) surface, (I) M₁₃ surface.

Export to PPT

Close

Fig. 3

SEM surface images showing the impact of MAX phase addition in the aqueous and organic phase during the PA formation: (a) M₁₅, (b) M₁₇, (c) M₁₉ and (d) M₂₁.

Export to PPT

In the meantime, incorporating of MAX phase within the PA layer during the IP process manifested a distinguished variation in their appearance depending on whether the material were added to the aqueous or organic phase. Fig. 3A and B depicted the SEM surface images after adding the MAX phase to the aqueous phase at 0.025 and 0.075 g, respectively. As could be seen, the structure of PA was still displaying the leaf-like structure but slightly was denser and more compacted at higher MAX phase content. However, adding the MAX phase into the organic phase manifested an apparent discrepancy in the PA structure (Fig. 3). The PA structure had a sponge-like appearance with a non-uniform porous structure at a low (0.025 g) MAX phase (Fig. 3C). This observed structure turned to a more dense spongy structure upon increasing the content to 0.075 g in the aqueous phase (Fig. 3D). The reason for this is attributed to the interaction that took place during the IP process. The addition of the MAX phase to the aqueous phase (MPD) stimulated the reaction between MPD and TMC, and ultimately, increased the strength and thickness of the PA layer. In contrast, when they were added to the organic phase, it could be concluded that these interactions were inhibited by the addition of the MAX phase. The quantity of fillers in the aqueous phase and membranes with a high loading % of the MAX phase, which had a thinner PA layer, both influenced how quickly MPD monomers diffused into the organic phase (Bagherzadeh et al., 2020).

3.2 Chemical functional groups of membranes (Fourier-transform infrared (FT-IR) spectroscopy)

Fourier-transform infrared (FT-IR) spectroscopy has been devoted to examining membrane composition following the 2D MAX phase incorporation. The amount of infrared light radiation (IR) that will be absorbed by a sample of materials is calculated using this technique. The materials’ wavelengths and the chemical bonds and functional groups of their molecular components have all been recognized as each functional group is represented by a wavelength. Herein, the FTIR was used to compare the chemical characteristics of the PES membrane before and after MAX phase were added, and before and after IP reaction occurred. The FT-IR spectra of the neat PES, polyamide layer (TFC) and PES/MAX phase membranes, have been detected within the spectrum 4000–650 cm⁻¹ (Fig. S1). For the 16 % PES flat sheet membranes, it is noticeable that the PES polymer characteristics produced peaks which starts with the peak at 701.751 cm⁻¹ attributed to phenyl groups and ends with the peak (3365.97 cm⁻¹) that represents the root (O—H) stretching alcohol intermolecular bonded (Sadidi et al., 2023). Additionally, it contains many roots between them such as (1578.69 cm⁻¹) which represents (C⚌O) stretching peak. The peak (1407.94 cm⁻¹) was attributed to C⚌C aromatic ring stretching (Abbas et al., 2022). The peaks at1322.81 cm⁻¹,1298.78 cm⁻¹, 1241.54 cm⁻¹, 1150.43 cm⁻¹ were ascribed to O—H bending phenol, nitro compound, C—O stretching alkyl aryl ether and S⚌O stretching sulfone, respectively (Abbas et al., 2022). The peaks (1105.57 cm⁻¹) and (1072.84 cm⁻¹) are characterized by the symmetric and asymmetric stretching vibrations unique to the sulfone group (Sadidi et al., 2023). Also, the spectra at (872.77 cm⁻¹) are C—H bending 1,2 disubstituted, while (836.80 cm⁻¹), (797.93 cm⁻¹) and (718.08 cm⁻¹) are C⚌C bending alkene trisubstituted. In comparison between the PES substrate spectrum and the TFC membrane spectrum, two roots absorption peaks at 1611.85 cm⁻¹ and 1543.83 cm⁻¹ associated with the polyamide skin layer are observed in the spectra of all TFC membranes as shown in Fig. S1(b), which can correspond to the amide I band (C⚌O stretching), aromatic ring breathing, and the amide II band (C—H stretching), respectively (Darabi et al., 2017). So, this establishes the NH2 amide layer's existence. Also, it can be noted when adding the MAX phase to the polymer, in the case of concentration (0.025), the radical appears (3430.96 cm⁻¹) instead of the peak (3365.97 cm⁻¹) that represents the apparent presence of —OH with solid hydrogen bonding on the surface of (Ti3AlC2) membrane (Sadidi et al., 2023). And when the concentration of the MAX phase is increased, the peak (3430.96 cm⁻¹) disappears, as the range is limited between (1578.87–701.72) cm⁻¹. In the case of adding MAX phase (Ti3AlC2) (0.025) to the MPD concentration, the peak (1611.86 cm⁻¹) appears and it is representing the amide I band (C⚌O stretching) (Darabi et al., 2017). Finally, the peak (627 cm⁻¹) appears and it representing the bond (Ti—C) (Abood et al., 2023).

3.3 Atomic force microscopy AFM

Surface roughness parameters have a critical impact on the entire performance of any membrane. The rougher surface is expected to endow a higher fouling tendency on the membrane surface and vice versa. The vertical distance between valleys and nodules is what referred to as roughness. The rough surface can stop some impurities from getting through the membrane, but it can also cause more fouling (Li et al., 2017; Alsalhy et al., 2013). The surface topography parameters of the PES support layer, and polyamide layer (PES-TFC), were determined by AFM. The 0.5 µm × 0.5 µm 3D images were captured by the AFM where the dark region referred to the lowest points (valleys) and light regions referred to the highest points (peaks) (Fig. 4). As shown in the figure and Table 1, the PES-TF membrane manifested the lowest average roughness (Ra = 6 nm) amongst all other samples whereas Ra was around 9 nm for pristine PES membrane and at other higher MAX phase content. Apparently, the surface was not regular as there were high roughness peaks, related to surface defects and dirt. The tip of the microscope risks sticking to the sample. There are filaments which, from a qualitative analysis, do not appear to be made of a material of a different nature. TFC membranes revealed slightly higher roughness parameters, 9.6 nm for M7 while recorded 14.6 nm upon increasing the MAX phase content in the PES polymeric matrix (M9). Incorporating the MAX phase within the PA layer to prepare TFC membranes disclosed much higher roughness parameters compared to the aforementioned membranes. Upon adding only a few content of MAX phase (0.025 g) within the aqueous phase, the Ra value reached up to 60 nm for M11, while this value noticeably dropped to 42 nm at higher content of MAX phase (M13). In the meantime, incorporating 0.025 g MAX phase in the organic phase resulted in the highest average roughness value (71.4 nm) compared to 41 nm when adding higher MAX phase content which showed only 41 nm average roughness. The results then revealed that the membranes modified with 0.025 wt% MAX phase introduced to the organic solution (TMC) were rougher than those added to the aqueous solution (MPD). The improvement in roughness occurred as a result of the deposition of the material on the surface of the membrane and because the MAX phase is hydrophilic material so the contact angle decreases (Abood et al., 2023).

Close

Fig. 4

AFM analysis in 3D was used to investigate the surface properties of the support layer and polyamide layer (PES-TFC), for (a) M₁, (b) M₄, (c) M₇, (d) M₉, (e) M₁₁, (f) M₁₃, (g) M₁₅, (h) M₁₇, (i) M₁₉, (j) M₂₁.

Export to PPT

3.4 Membranes hydrophilicity

Membrane hydrophilicity is a critical characteristic that gives a deep insight into membrane ageing and performance. Contact angle measurements can be used to examine the membrane's hydrophilicity and link the results to how well the membrane performs. Lower contact angles denote higher hydrophilicity and vice versa. Fig. S2 illustrates the analysis of the water contact angle of all membranes, including various ratios of pristine PES, and TFC membranes. As shown, the pristine PES membranes manifested higher CA values compared to other samples. The average contact angle of M₁ was 64.914° and this value increased slightly upon increasing the PES wt.%. This could be interpreted due to the denser membrane formed at 17 and 18 wt% where CA relies on many other characteristics of the membrane such as roughness, pore size and porosity and not only on the membrane’s material nature. On the other hand, TFC exhibited lower CA values (48.2°) as shown for the M₄ membrane whereas the CA was 58.4° for the M₉ membrane. Besides that, the CA measurements of the M₁₃ membrane manifested 44.8° which confirmed the potential of the MAX phase to enhance the surface hydrophilicity by decreasing the water CA at high loading content (0.075 g). Particularly, incorporating a hydrophilic group is a well-established technique to impart that hydrophilicity to the designated membrane (Sadidi et al., 2023). The addition of the TFC and MAX phase as an embedded material had a favorable effect on the hydrophilic nature of the PES/ MAX/ TFC membrane surface, as confirmed by the witnessed decline in the value of the CA of almost 20°.

3.5 Effect of MAX phase wt.% content on the porosity, pore size and thickness

The permeation flux of the membrane is known to be influenced by the interplay of many variables, including the pore size and porosity of the membrane's surface, roughness, the active layer thickness and hydrophilicity. Therefore, the impregnation of the MAX phase is expected to bestow desirable traits on the permeation characteristics of the membranes. Table 2 listed the values of the porosity, pore size and active layer thickness of the produced membranes. Unsurprisingly, as the PES wt.% increased from 16 to 17 and then 18 %, the porosity of the membranes decreased from 79.9 to 71 and 65.7 for M₁, M₂ and M₃, respectively. Similarly, the mean pore radius revealed a gradual decline in size as the dope casting solution increased and showcased 10.3, 9.2 and 8.7 nm for M₁, M₂ and M₃, respectively. The membrane thickness has shown a significant increase upon increasing the PES wt.% and was two-fifths (125 µm) higher for M₃ compared to M₁. This could be attributed to the higher casting solution viscosity induced by high polymer wt.%. Higher casting solution viscosity diminishes the exchange rate between solvent and non-solvent during the NIPS process. Ultimately, produces lower porosity, smaller pore size and higher thickness (Abood et al., 2023). On the other hand, increasing the MAX content from 0.025 to 0.075 wt% has only imparted a trivial influence on the porosity values as could be seen in the table. However, the pore size displayed a significant increase in both mean pore size and thickness as the MAX phase increased and reached 17.6 nm and 123.1 µm, respectively. The porosity effect on the membrane hydrophilicity, when the concentration of polymer PES increased, the porosity decreased and the hydrophilicity decreased because the contact angle increased (Ibraheem et al., 2023).

The pore size effects the flow, water flux and reverse salt flux increased when the pore size increase according to the attached results.

3.6 FO performance of TFC membranes

3.6.1

3.6.1 Effect of substrate PES membrane concentration on the TFC membrane

The concentration of PES and PA has a substantial impact on the water flux (Jw) and the reverse salt flux (Js) of the FO system. Performance tests of the FO process were carried out at a flow rate of 300 mL/min with DI water and 1 M NaCl as the feed and draw solutions, respectively. As shown in Fig. 5, both water flux and reverse salt flux diminish as the concentration of the TFC membrane substrate increases from 16 to 18 wt%. With a water flux of 6.03 L/m²h and a reverse salt flux of 2.85 g/m²h, the M4 membrane seems to disclose the best-performing membrane Fig. 5a. Compared to the M4 membrane (6.03 L/m²h), M5 and M6 membranes revealed water flux magnitudes of 5.74 L/m²h and 4.71 L/m²h.

Close

Fig. 5

(a) water flux and (b) reverse salt flux of TFC membranes prepared with different PES concentrations (1 M NaCl as DS, DI water as FS, 300 mL/min).

Export to PPT

In Fig. 5(b), the reverse salt flux (RSF) decreased when the PES concentration increased from 16 % to 18 %. This could be due to two factors: (i) Water flux normally declines when RSF declines because RSF was moving in the opposite direction to water flux, and (ii) an increase in the salt rejection value of the TFC membranes caused RSF to decline. Therefore, a lower PES concentration result was found to display a better permeability in the structure of the TFC membrane, which led to a favorable performance in the FO process. This also agrees with other studies (Saeedi-Jurkuyeh and Jafari, 2019; Al-Musawy et al., 2021; Tiraferri et al., 2011)). They believed that FO performance would be enhanced by the substrate membrane having a low polymer concentration (porous structure).

3.6.2

3.6.2 Effect of MAX phase concentration on water flux and reverse salt flux of membrane

Fig. 6 displayed the impact of MAX phase content in the membrane on the water flux and the reverse salt flux. As shown in Fig. 6a, the M14 membrane, which comprises a 0.1 wt% MAX phase, manifested the highest water flux compared to the M11 membrane (0.025 wt% MAX phase). The water flux duplicated from 7.57 to 14.17 L/m²h when the concentration of the MAX phase increased from 0.025 to 0.1 wt%. In the meantime, the M14 membrane revealed the highest reverse salt flux compared to the M11 membrane. The reverse salt flux showcased an increase from 0.88 to 3.60 g/m²h, as shown in Fig. 6b. This resulted from the fact that at the higher concentration of MAX phase, more salt ions and water molecules were rejected at once due to an increase in cross-linking and layer thickness of polyamide on the top side of the M7 (16 wt% PES + 0.025 wt% MAX phase) membrane, which is consistent with the findings of preceding studies (Rastgar et al., 2019). Yet, M11 has the best performance as a FO membrane due to the best-compromised value between water permeability and reverse salt rejection, resulting in the highest performance of the FO membrane.

Close

Fig. 6

(a) water flux and (b) reverse salt flux of TFC membranes prepared with different MAX phase concentrations (1 M NaCl as DS, DI water as FS, 300 mL/min).

Export to PPT

3.6.3

3.6.3 Effect of MAX phase concentration change on MPD and TMC performance

Fig. 7 illustrates how the MAX phase concentration affects the water flux and the reverse salt flux through FO membranes. According to the experimental findings, the water flux substantially enhanced from 6.83 to 12.55 L/m²h when the MAX phase content increased from 0.025 to 0.1 wt%, as shown in Fig. 7a, while the reverse salt flux increased from 2.49 to 5.77 g/m²h, as shown in Fig. 7b. This resulted from the fact that a higher amount of MAX phase causes the M18 (16 wt% PES) membrane top side to reject more salt ions and water molecules at the same time due to an increase in cross-linking and layer thickness of polyamide, which is in keeping with the results of other (Rastgar et al., 2019).

Close

Fig. 7

The water flux and reverse salt flux when adding the MAX phase to aqueous solution MPD and to the organic phase (TMC).

Export to PPT

Relevantly, compared to the M15 membrane, which has 0.025 wt% MAX phase, and the M16 membrane, which has 0.05 wt% MAX phase, the M18 membrane, which has 0.1 wt% MAX phase, has a high water flux and a high back-salt flow, respectively. Because M15 has the appropriate balance of polyamide film thickness, it provides an optimal balance of reverse salt rejection and water permeability, which ultimately leads to the high performance of the FO membrane when compared to M16, M17, and M18.

The water flux and the reverse salt flux as measured by the TFC membrane are shown in Fig. 7. By varying the amount of MAX phase added to the organic phase TMC during the IP process, the PA layer was produced on the surface of the M1 (16 wt% PES) substrate. The MAX phase was raised from 0.025 wt% to 0.1 wt%, which resulted in an increase in the water flux and reverse salt flux of TFC membranes from 6.63 L/m²h and 1.41 g/m²h to 10.09 L/m²h and 5.97 g/m²h, respectively. These findings demonstrate that the functionality of the FO membrane is more significantly influenced by the thickness of the PA layer. Additionally, with the MAX phase, cross-linking intensity also increases (Kadhom and Denga, 2019; Zhou et al., 2014). The substrate M19, in particular, produced the best outcomes in terms of water penetration and reduced draw solution losses. The TMC molecules must therefore be evenly distributed and absorbed over the support layer in order to achieve the best cross-linking when interacting with the MPD. The FO membrane with the best performance will ultimately result from this.