3.1. Preparation of [email protected] membrane
The preparation mechanism of [email protected], which combined interfacial ion migration technology and unilateral spraying treatment was illustrated in Scheme 1a. At -12 oC, the fabric (Scheme 1c-ⅰ) was treated with 7 wt % NaOH/12 wt % urea solution for two hours. NaOH destroyed the intramolecular and intermolecular hydrogen bonds of cellulose, causing cellulose to swell and even partially dissolve (Scheme 1c-ⅱ) (Cai and Zhang 2005; Fan et al. 2018). At the same time, hydroxide ions, sodium ions, and alkali cellulose macromolecules constituted the internal ionic system of cotton fibers, also called an intra-membrane system (Scheme 1c-ⅲ). Alkali cellulose macromolecules were immobilized within the membrane, while hydroxide ions and sodium ions were free to pass through the membrane. Treated cotton fabrics carrying OH- were immersed in MgCl2∙6H2O aqueous solution. The solution outside the fibers contained magnesium ions and chlorine ions were considered as an extra-membrane system. There was a concentration gradient of mobile ions between inside and outside the membrane. Therefore, Mg2+ tended to migrate to the intra-membrane system and OH- was inclined to move to the extra-membrane system. Mg2+ reacted with the OH- immediately to form magnesium hydroxide at the interface of the surface of cotton fibers. As shown in SEM images, the raw CF surface was smooth, whereas the fibers of [email protected] were fully clad by Mg(OH)2 with a micro-nano structure. Next, stearic acid was used as a finishing agent to create a protective layer on the unilateral surface of the fabric. Stearic acid imparted a low unilateral surface free energy value to the substrate and conferred a protective layer acting like a quilt.
Fig. 1a shows the FTIR spectra obtained for raw CF and [email protected] The spectra of raw cotton and [email protected] were confirmed to be generally consistent, verifying that the primary structure of the treated cotton fabric remained unchanged. Characteristic peaks at 2968 cm-1 and 2918 cm-1 were belonging to the asymmetrical stretching vibration of the C-H bond in the -CH3 and -CH2 groups, respectively. The characteristic peak at 2848cm-1 was due to the symmetrical stretching vibration of the C-H bond in the -CH2 group. The weak peaks at 1540cm-1 and 1457cm-1 were caused by the functional group of the carbonyl group (C=O) (Wen et al. 2018; Khattab et al. 2020; sharif et al. 2020), demonstrating that the stearic acid covering the fabric was in minor amounts. In the XRD pattern (Fig. 1b), both samples show three peaks centered around 14.94 o, 16.69 o, and 22.79 o, which were in accordance with the typical (1 -1 0), (1 1 0), and (2 0 0) peaks of natural cellulose fibers (French 2014). It was clear that the characteristic peaks indicated Mg(OH)2 phase (2θ=19.22 o, 38.30 o, 50.78 o, and 58.96 o), which were attributed to their corresponding indices (0 0 1), (1 0 1), (1 0 2), and (1 1 0), respectively. From the wide scan spectrum (Fig. 1c), both CF and [email protected] had two main peaks of C1s and O1s. [email protected] has core level peaks of Mg 2p, Mg 2s, and Mg 1s at 49.6 eV, 88.1 eV, and 1303.2 eV, while Mg (KLL) peak appeared at about 306.1 eV (Keikhaei and Ichimura 2019). As in Fig. 1d, the C 1s spectrum was divided into four parts. The peaks were concentrated at 284.7 eV, 286.2 eV, corresponding to C-C/C-H, C-O of cellulose, respectively. In addition, the peaks are concentrated at 287.8 eV and 289.8 eV, which correspond to C=O and COO- respectively (Fan et al. 2018), demonstrating that stearic acid was covered onto the surface of the fabric. As shown in Fig. 1e and 1f, Mg 2p peak at 49.6 eV and the O 1s peak at 530.9 eV in XPS spectra was related to Mg(OH)2 (Ardizzone et al. 1997; Zhu et al. 2011). The deposition of Mg(OH)2 on the surface of cotton fabrics was further confirmed by XPS.
3.2. Water permeability
After dripping water onto the obverse side of the [email protected] membrane, the change of contact angle was recorded (Fig. 2a). The obverse side had hydrophobic abilities with a high WCA of about 141.1 o. When the water was dropped onto the obverse side, the water droplet could spontaneously penetrate [email protected] from the hydrophobic side to the hydrophilic side. It took about 15.4 s for the water droplet to penetrate the membrane completely. Fig. 2c provides a set of images that also confirmed the transfer behavior of water in the air/air system.
3.3. Evaluation of the emulsion separation performance
A series of experiments were carried out to evaluate the separation performance of the [email protected] membrane. Emulsions were stabilized by three different ionic types of surfactants (SDS, CTAB, or Tween-60) respectively. The anionic surfactant stabilized emulsion (1,2-dichloroethane/SDS/H2O) separation was evaluated at first.
The emulsion was directly injected into a sand core funnel device, and the volume of the emulsion was fixed as 100 mL. The emulsion was separated rapidly with water almost entirely penetrating the membrane while the oil droplets surrounded by SDS were blocked at the upper beaker. Optical microscopic images (Fig. 3e) of the emulsion and filtrate indicated that both emulsions contained numerous oil droplets, while the filtrate did not. Furthermore, as shown in the digital photographs, all the filtrates were clear and transparent. To demonstrate the absence of SDS in the filtrate, the samples were characterized by FTIR (Fig. 3b). To further evaluate the separation performance of the membrane, the change in the composition of the liquid before and after separation was tested. Characteristic peaks at 2971.73 cm-1 and 2868.71 cm-1 were mainly dominated by the symmetric and asymmetric stretching bands of C-H at the hydrophilic end of the surfactant. Peaks at 1200.15 cm-1 resulted from asymmetric S-O stretching, whereas the peaks at 1004.29 cm-1 corresponded to symmetric S-O stretching (Gao and Chorover 2010; Ramimoghadam et al. 2012; Warsi et al. 2020). Compared with the emulsion, the peaks of SDS (2971.73 cm-1, 2868.71 cm-1) were not highlighted in the curve of the filtrate. FTIR pattern of the filtrate was similar to that of distilled water, which showed that the separation effect was satisfactory.
For emulsions derived from different kinds of oil (Fig. 3a), 1,2-dichloroethane/SDS/H2O, isooctane/SDS/H2O, xylene/SDS/H2O, the separation flux was 532 L m-2 h-1, 609 L m-2 h-1 and 426 L m-2 h-1, respectively. After calculation, separation efficiencies were more than 99.3 % each time. For emulsions by surfactants with different ionic types, as shown in Fig. 3c and 3d, CTAB/Oil/H2O emulsions and the Tween-60/Oil/H2O emulsions could also be successfully separated with high separation efficiency, indicating that the [email protected] membrane could be used to separate oily wastewater containing unspecified surfactants.
3.4. Separation mechanism
The separation process based on the simulation was demonstrated in Fig. 4a to further showing the mechanism of the [email protected] separation performance. In the separation process, water droplets dropped onto the hydrophobic side should form a hump (i.e., PL) on the opposite side of the membrane pores (Fig. 4a left). Because of the presence of hydrophilic capillary force, the hydrophilic layer underneath would eliminate PL once the “hump” contacted the hydrophilic layer (Hou et al. 2019; Chen et al. 2019). Therefore, water permeated the membrane as a continuous phase. After surfactants were added to the oil-water mixture, the surfactant coating structure of the oil droplet was gradually assembled. Oil droplets surrounded by hydrophobic tails of surfactant tended to form the core, while the hydrophilic heads were exposed outside in contact with the aqueous environment. Oil droplets surrounded by surfactants were difficult to aggregate due to their thermodynamic stability. The Mg(OH)2 decorated CF with an increase in micro/nanostructure and a reduction in the contact area between oil droplets and membrane (Zhang et al. 2013). Unilateral modification with stearic acid further exposed a large number of carbon chains on the surface of micro/nanostructures, resulting in a reduction in surface energy. Therefore, the stearic acid layer acted as a protective barrier, like a quilt, covering the magnesium hydroxide (Fig. 4a right). The polarity of the hydration layer was high. It was difficult for the oil droplets stabilized by surfactants to demulsify. Oil droplets surrounded by surfactants as a dispersed phase were rejected by the membrane. As time progressed, oil droplets physically accumulated on top of the membrane due to gravity, and the “cream layer” was gradually formed. It could be seen from the optical microscopic image that the “cream layer” was composed of undemulsified emulsion droplets (Fig. 4b right), while there were no oil droplets in the filtrate (Fig. 4b left).
3.5. Protection performance
The “cream layer” gradually formed on the surface of the membrane as the oil droplets physically accumulated. To evaluate the performance of the protective layer of the membrane, the “cream layer” obtained after the separation of 50 mL, 100 mL, and 150 mL emulsion was observed by optical microscope respectively. From the optical microscope pictures (Fig. 5b), it could be seen that the accumulation of emulsion droplets became more and more obvious with the increase of emulsion content, which showed that the “cream layer” was composed of undemulsified emulsion droplets. Oil droplets accumulated gradually driven by gravity on the membrane surface to form a “cream layer”, which facilitated to improve the separation efficiency to a certain extent. However, the “cream layer” also gradually covered the surface pores and reduced the effective separation area of the membranes, resulting in a gradual reduction of separation flux. It was also demonstrated by the flux calculated after each 5 mL of filtrate collected (Fig. 5a). As the separation process went on, flux gradually decreased from 1403 m-2 h-1 to about 412 L m-2 h-1. Impressively, after rinsing with distilled water, the “cream layer” was easily washed off and the membrane was not contaminated by oil droplets (Fig. 5c). The separation flux of multiple cycles also proved that the “cream layer” was formed only through physical accumulation. There was no obvious decrease in liquid flux and separation efficiency after 10 cycles, which indicated robust reusability (Fig. 5d).