Introduction

During the past several decades, oily pollution, originated from oil spillage, domestic lives, and industrial activities, have caused huge threats to current ecological systems [1, 2]. Thus many methods for oily wastewater purification have been developed such as skims, air flotations, adsorbents, chemical or biological treatments, and filtration membrane (FM) [3, 4]. Among these methods, FM-based purification technique has gained special attention for their low cost, high separation efficiency, and less energy consumption [5,6,7,8,9,10,11,12].

For the lower viscosity of water than these of oils, FMs with superhydrophilicity/underwater superoleophobicity have been widely investigated in oil-in-water (O/W) emulsion purification. The absorbed water layer on the surface of FMs could effectively reduce the direct contact between oil droplets and the surface of FMs, lowering the risk of oil fouling. [13,14,15,16,17]. Li fabricated superhydrophilic FM by decorating MXene on electrospun polylactide NM. The rejection efficiency of the as-prepared FM against different O/W emulsions was up to 99% [18]. Guzman and co-workers reported polydopamine-sulfobetaine methacrylate nanoparticles/cellulose acetate composite membrane for oil–water separation. The separation efficiency against different oil-in-water emulsions ranged from 95 to 99% [19]. Although many successes have been realized, the oil-fouling issues for long-term O/W emulsions treatment are still inevitable, in which some small-sized oil droplets could penetrate into the inner voids of NMs or adhere to the surface of NM, reducing the filtration efficiency [20,21,22,23,24]. To solve this bottleneck problem, many routes have been explored. Jiang and co-workers reported a double-defensed membrane by grafting poly(sulfobetaine) brushes onto poly(hydroxyethyl methylacrylate) hydrogel modified poly(vinylidene fluoride) porous membrane [25]. The membrane could continuously separate different surfactant-stabilized O/W emulsions for 6 h with the unchanged flux. Jin and co-workers fabricated zwitterionic nanosized hydrogels grafted poly(vinylidene fluoride) porous membrane [26]. The membrane could separate Tween80-stabilized isooctane-in-water emulsion for 10 h with 99% recovery of flux. Although many successes have been obtained, most literatures focus on the surfactant-stabilized non-viscous O/W emulsions separation.

Another factor, hindering the practical application of FMs in oily wastewater emulsion separation, originates from the threats of corrosive media (e.g. HCl, NaOH, and NaCl solution) [27,28,29,30,31]. To solve this problem, many efforts have been carried on to enhance the harsh conditions tolerance of FMs [32,33,34,35,36,37]. For example, Zhan et al. used engineering poly(phenyl-propionic nitrile) (PEN) to construct ternary PEN/halloysite nanotubes/graphene oxide FM. Even under harsh conditions including high-temperature resistance (up to 90 °C) and corrosive media (pH = 1 and pH = 14), the separation efficiency of as-prepared FM for O/W emulsions was still higher than 99% [33]. Wu and Chen constructed hydrophilic nanofibrous membrane (NM) with poly(N-isopropylacrylamide-co–N-methylolacrylamide) nanofiber as matrix and chitin nano-whisker as reinforcement via co-crosslinking route. The NM possessed high separation efficiency of > 99.5% with good flux of 1, 100–1, 300 L m–2 h–1 against Tween 80-stablized O/W emulsions under broaden pH values (pH = 1–13) [38]. Although many successes have been realized, the FMs, with both high anti-corrosive performance and ultra-long-term oil-fouling recovery for large-scale surfactant-stabilized viscous O/W emulsion separation are still highly required.

Herein, we proposed an alternative route to solve these two challenges above by constructing pre-oxidized polyacrylonitrile (p-PAN) NM. Given the cyclization, dehydrogenation, and oxidation within pre-oxidation, both the conjugated ladder molecular structures and the hydrophilic groups formed throughout the p-PAN [39,40,41,42,43,44,45]. The former provided p-PAN NM with high robust anti-corrosive capacity and high mechanical performance, the latter enabled the NM with superhydrophilic/underwater superoleophobic wettability. Consequently, our NM could continuously separate surfactant-free/surfactant-stabilized viscous (soybean oil or crude oil) O/W emulsions under harsh conditions (e.g., 1 M HCl, 10 wt% NaCl, 1 M NaOH) for 265 h with both high separation efficiency and robust oil-fouling recovery. What is more, robust stability against some organic corrosive media (e.g. N, N-dimethyl formamide (DMF), dimethylacetamide (DMAC), formic acid (FA), or dimethyl sulfoxide (DMSO)) has been also realized.

Experimental Section

Materials

Polyacrylonitrile (PAN) with an average molecular weight of about 150, 000, ethanol, N,N-dimethylformamide (DMF: ≥ 99.8%), dimethyl sulfoxide (DMSO: ≥ 99.9%), formic acid (FA: ≥ 96%), and dimethylacetamide (DMAC: ≥ 99.8%) were purchased from Sigma-Aldrich (Shanghai, China) Trading Co., Ltd. Diesel, soybean oil, HCl, NaOH, NaCl, and sodium dodecyl sulfonate (SDS) were provided by Kelong Chemical Co., Ltd. (Chengdu, China). Ampliflu Red and Ocean Blue were purchased from Aladdin Chemistry Co., Ltd. (Shanghai, China). The crude oil was provided by Daqing Oilfield and diluted with diesel fuel.

Preparation of Electrospun PAN Nanofibrous Membrane

1 g of PAN was added to 10 mL DMF, and the solution was heated under magnetic stirring at 70 °C for 12 h. PAN NM was fabricated through electrospinning, with the temperature, feed rate, applied voltage, and distance between the spinneret and collector fixed at 25 ± 2 ℃, 1 mL/h, 18 kV, and 20 cm, respectively.

Pre-oxidation Process

The PAN NM was placed in an oven and heated from 40 to 238 ℃, then to 248 ℃, and finally to 260 ℃ in air. Each stage was dwelled for 30 min with the heating rate of 5 ℃/min.

Preparation of Oil-in-Water Emulsion

Surfactant-free oil-in-water emulsions were achieved by mixing oil into deionized water (DI water) (Voil:Vwater = 1:99) under magnetic stirring (1, 200 rpm) for 30 min. Then, the mixture was sonicated (output power: 550 W) for 3 h to form a milk-like emulsion. As for surfactant-stabilized oil-in-water emulsions, SDS was chosen as the emulsifier and dissolved into the aforementioned oil/water system. The concentration of SDS was fixed at 0.1 g/L within the final emulsions.

Separation of Oil-in-Water Emulsion

Two routes have been used for emulsions separation.

  1. (i)

    As for small amount oil-in-water emulsions separation: The separation test was performed using a dead-end equipment setup, where the pre-wetted p-PAN nanofiber membrane, with the filtration area of 2.01 cm2, was secured between the glass tube and the filter sand core. 20 mL of emulsions were added to the glass tube for filtration under a pressure of 0.09 MPa.

  2. (ii)

    As for large-scale viscous oil-in water emulsions separation: The separation test was performed using a homemade setup, in which the p-PAN nanofiber membrane, with a stainless-steel mesh as a substrate, was fixed to a glass tube (with an area of 2.06 cm2) using hydrophilic tape as a binder. The entire glass tube was immersed in a 2 L beaker containing 1.1 L of surfactant-free/surfactant-stabilized viscous oil-in-water emulsion, which was stirred magnetically at 400 rpm to simulate a flow oily wastewater system. Two filter flasks (2.5 L each) were connected between the glass tube and the water pump. One flask was used to collect the filtrate, while the other flask, acting as a buffer flask, was used to prevent backflow from the water pump. The entire filtration process was conducted under a pressure of 0.09 MPa.

Filtration function: The water flux J (L m−2 h−1 bar−1) could be calculated by using Eq. (1) [46]:

$${\text{J}} = \frac{V}{A \times t},$$
(1)

where V (L) is the permeation volume of filtrate permeated through the membrane, t (h) is the filtration time, A (m2) is the useable filtration area.

The filter efficiency R (%) is defined as Eq. (2)

$$R\left( \% \right) = \left( {1 - \frac{{C_{f} }}{{C_{p} }}} \right),$$
(2)

where Cp is the concentration of prepared solution and Cf is the concentration of the filtrate solution. As for large-scale viscous O/W emulsion separation, we recorded the time per 50 mL O/W emulsion separation. Thus, the flux versus time (volume) can be calculated.

Anti-fouling Assessment

To clearly demonstrate the anti-fouling performance, the viscous oil droplets were dyed with Ampliflu Red. After filtration, the fouled nanofiber membrane (NM) was placed in an 80 ℃ water bath and subjected to ultrasonic treatment for 5 min to remove the fouled oil droplets. Ultraviolet (UV) lamp was used to illuminate both the fouled membrane and the recovered membrane.

DFT Calculation Details

All simulation and calculation experiments were carried out under the Materials Studio. The generalized gradient approximation of Perdew–Burke–Ernzerh (PBE) was used in Dmol3 modulus with OTFG ultra-soft pseudopotential. The kinetic energy cutoffs were all set to 480 eV, and the Brillouin zone was sampled with a special Monkorst-park K-point grid of 4 × 4 × 1 for geometry optimization and energy task. The SCF tolerance, maximum force, maximum stress and maximum displacement were all set as 2.0 × 10–5 eV/atom, 0.05 eV/ Å, 0.1 GPa and 0.001 Å, respectively.

Characterizations

A field-emission scanning electron microscope (Thermo Scientific Apreo 2 C, USA) was utilized for SEM analysis. Atomic force microscope (AFM) was used for roughness analysis using a BRUKER Dimension ICON, (Germany). The chemical composition of the nanofibrous membrane surface was fully characterized using Fourier transform infrared (FTIR) spectra obtained from a Bruker Lumos FT-IR Microscope (Billerica, MA, USA) in ATR mode. X-ray photoelectron spectroscopy (XPS) was conducted using a Thermo Fisher Escalab 250 Xi instrument (Waltham, MA, USA), with Al Kα as the excitation source. Contact angles (CAs) were measured using an OCA 25 data physics contact angle meter (Germany). Mechanical properties test was performed using a universal testing machine (UTM, CS 225, Lloyd Instruments, Ltd., UK). The average total organic carbon (TOC) content in the feed solutions and corresponding filtrates was determined using a Shimadzu TOC-L total organic carbon analyzer (Japan). Optical images were captured using a Canon EOS 50 D camera (Tokyo, Japan). Dynamic light scattering (DLS) characterization was performed using a Zetasizer instrument (Nano-ZS, Malvern, UK).

Results and Discussion

Synthesis and Characterization of p-PAN NM

Figure 1 illustrates the whole synthetic procedures for p-PAN NM. Briefly, PAN NM was obtained by electrospinning PAN/DMF solution. Pre-oxidation was then employed to treat the PAN NM, in which cyclization, dehydrogenation, and oxidation occurred simultaneously [47, 48], resulting in the formation of conjugated ladder molecular structures and hydrophilic groups throughout the p-PAN NM. Besides, the roughness of p-PAN NM was smaller than that of PAN NM (ESI; Fig. S1), attributing to the stretching of the membrane during pre-oxidation process. Thus, p-PAN NM with strong anti-corrosion properties, good mechanical performance, and superhydrophilicity/underwater superoleophobicity was fabricated for highly efficient oil/water emulsion separation.

Fig. 1
figure 1

Schematic diagrams for the preparation of p-PAN NM and its application in oil/water emulsion separation

Full size image

To confirm the reactions during the pre-oxidation process, Fourier transform infrared spectra (FTIR) with ATR model and X-ray photoelectron spectra (XPS) were combined to characterize the PAN and p-PAN NM. From the FTIR spectra (Fig. 2a), PAN NM exhibited two peaks (2, 941 cm−1 and 1, 452 cm−1) corresponding to C-H stretching bonds, as well as a peak (2, 243 cm−1) for the C≡N vibrating band [49]. After pre-oxidation, the relative intensities of both the C≡N vibrating bands (2,243 cm−1) and the C-H vibrating bonds (2,941 cm−1) decreased. Simultaneously, a new and strong band emerged at 1,596 cm−1, indicating the formation of C=N vibrating bands. Additionally, a new peak at 807 cm−1, associated with (C = C in the aromatic rings), was also observed [50] confirming the cyclization and dehydrogenation reactions. Furthermore, two weak peaks at 1, 725 cm−1 and 1, 252 cm−1, corresponding to the C=O and C–O groups, respectively, were detected. The broadened peak around 3, 235 cm−1, indexing to the OH group, confirmed the oxidation reaction. Figure 2b displays the XPS survey spectra of PAN NM and p-PAN NM, revealing the appearance of strong O 1 s peaks after pre-oxidation. Figure 2c presents the high-resolution spectrum of C 1 s from PAN NM, showing two peaks at 284.8 and 283.5 eV for C≡N and C–C/C–H, respectively [51]. In the case of p-PAN NM, the C 1 s peak can be deconvoluted into four peaks when considering the elements and groups directly bonded to the carbon atoms (Fig. 2d) [52]. The N 1 s spectrum of p-PAN NM exhibited a new peak at around 399.9 eV, indicating aromatic C=N was formed during peroxidation (Fig. 2e) [53]. The O 1 s peak of p-PAN NMs was deconvoluted into two peaks at 531.4 and 533.1 eV, corresponding to C–O and C=O, respectively (Fig. 2f) [54].

Fig. 2
figure 2

a FITR-ATR spectra and b XPS survey spectra of PAN and p-PAN NM. c C 1 s high-resolution of PAN NM. d-f C 1 s, N 1 s, and O 1 s high-resolution of p-PAN NM, respectively

Full size image

Variation of Surficial Wettability and Theoretical Elucidation

The variation of PAN NM surficial wettability before and after pre-oxidation has been systematically investigated. To vividly demonstrate the wettabilityNM with the aid of a UV lamp, indicating the hydrophobicity/lipophilicity of PAN NM and the amphiphilicity of p-PAN NM in air. Figure 3c illustr, Fig. 3a, b presents the photos of water and oil droplets on PAN and p-PAN NM. For PAN NM, water droplets remained on the surface while oil droplets (1, 2-dichloroethane was used) penetrated into the NM. In contrast, both water (dyed with Ocean Blue) and oil droplets (dyed with Ampliflu Red) quickly penetrated the p-PAN ates the water and oil contact angles (WCA and OCA) of PAN NM were 135° and 0°, respectively. The underwater oil contact angle (UWOCA) and underoil water contact angle (UOWCA) were 126° and 145°, respectively. For p-PAN NM (Fig. 3d), the WCA, OCA, UWOCA, and UOWCA were 0°, 0°, 156°, and 150°, respectively, confirming the superamphiphilicity in air, underwater superoleophobicity, and underoil superhydrophobicity of p-PAN NM. UWOCAs of different oil droplets (1, 2-dichloroethane, chloroform, diesel, crude oil, N-hexane, and soybean) have been measured (Fig. 3e), proving the generality of underwater superoleophobicity of p-PAN NM. The underwater oil adhesion on PAN NM and p-PAN NM was also investigated by measuring the underwater oil sliding angle (UWOSA). Even PAN NM was tilted to 10°, the oil droplet still tightly adhered to the NM (Top of Fig. 3f). In contrast, p-PAN NM (Bottom of Fig. 3f) exhibited a lower UWOSA (~ 5°), indicating the reduced oil adhesion of p-PAN NM.

Fig. 3
figure 3

a and b Photos of water and oil droplets on PAN and p-PAN NM. c and d WCA, OCA, UWOCA, and UOWCA of PAN and p-PAN NM. e UWOCAs of p-PAN NM against different oil droplets. f UWOSA of PAN and PAN NM

Full size image

To clearly understand the variation of surficial wettability, all spin-unrestricted density functional theory (DFT) computations were implemented using the Dmol3 code. Electron exchange and correlation effects were evaluated using the Perdew-Burke-Ernzerhof (PBE) generalizing gradient approximation (GGA). The optimized adsorption structures of H2O and oils (It is important to note that three representative oils (cyclohexane, 1, 2-dichloroethane, and chloroform) with well-defined structures were chosen as models to replace the traditional viscous oils (soybean oil or crude oil) owning to their not defined structures) on the surface of PAN and p-PAN were studied. The adsorption energy (Ea) of PAN and p-PAN for both water and oils was calculated at 25 ℃. From Fig. 4a, it could be observed that the Ea of PAN for water was positive, indicating the hydrophobicity (as proven in Fig. 3c). In contrast, the Ea values of PAN for oils were all negative, confirming the lipophilicity. For p-PAN NM (Fig. 4b), all Ea values were negative, demonstrating the super-amphiphilicity of p-PAN NM in air environment (as observed in Fig. 3d).

Fig. 4
figure 4

Electrostatic potential diagrams on Ea of the water, cyclohexane, 1, 2-dichloroethane, and chloroform on a PAN NM and b p-PAN NM at 25 ℃, respectively

Full size image

Measurement of Anti-harsh Conditions Tolerance

Strong tolerance to harsh conditions is of great importance for effectively separating oily wastewater systems coupled with corrosive media [55,56,57,58,59]. To evaluate the tolerance of our p-PAN NM against various normal inorganic solutions (e.g., 100 ℃ boiling water, 1 M HCl, 1 M NaOH, 10 wt% NaCl, and DI Water) and strong corrosive organic solvents (DMF, DMAC, DMSO, and FA) at room temperature (25 ± 2 ℃) was measured by immersing the NM in these media for 7 days (Fig. 5a and b). Additionally, the stability of p-PAN NM under ultrasonic treatment in an 80 ℃ hot water bath for 1 h was also tested (ESI, Fig. S2). From these data, we could observe that our NM maintained its integrity and the media remained transparent, implying no reactions occurred between our NM and the media. To further validate our hypothesis, the surficial wettability of p-PAN NM before and after these treatments was evaluated by measuring the UWOCA with 1, 2-dichloromethane as model (Fig. 5c). These unchanged UWOCA indicated that the surficial wettability of p-PAN NM was not affected. In contrast, PAN NM exhibited instability when exposed to these media (ESI, Fig. S3). PAN NM became curled when contacting with the inorganic media and organic FA solvent. As for organic solvents (DMF, DMSO, and DMAC), PAN NM dissolved completely within 1 s. The mechanical properties of PAN NM and p-PAN NM have been measured (ESI, Fig. S4). Given the cyclization reaction within pre-oxidation process, the tensile stress of p-PAN was higher than that of PAN NM.

Fig. 5
figure 5

Photographs of p-PAN NM stored within a inorganic solvents, b organic solvents, respectively. c Underwater oil contact angles of p-PAN NM after being treated under harsh conditions

Full size image

O/W Emulsion Separation Performance

To evaluate the efficiency of separating oil-in-water (O/W) emulsions, a series of O/W emulsions, including both SDS-free and SDS-stabilized emulsions, were prepared and filtered via p-PAN NM. Optical microscopy and digital imaging were employed to visually track the changes in the feeds and filtrates with SDS-free/SDS-stabilized cyclohexane-in-water emulsions as model system (Fig. 6a, b). It was observed that all the creamy feeds turned transparent and the oil droplets completely disappeared after filtration. The size distribution variations of the O/W emulsions in the feeds and filtrates were measured using dynamic light scattering (DLS) (ESI; Fig. S5). The effectiveness of this high separation efficiency against different types of O/W emulsions was evaluated by filtering eight different SDS-free/SDS-stabilized O/W emulsions (containing dichloroethane, cyclohexane, soybean oil, and crude oil) using p-PAN NM (Fig. 6c). All the separation efficiencies were found to exceed 99.7%. Figure 6d illustrates the cycling stability of SDS-free cyclohexane-in-water emulsions for ten cycles. It can be observed that both the flux and separation efficiency remained constant over the course of ten cycles, demonstrating the excellent cycling stability of our p-PAN NM.

Fig. 6
figure 6

The optical microscope and digital images for feeds and filtrates of a surfactant free and b surfactant stabilized O/W emulsions. c Separation efficiencies and fluxes for different emulsions. d Cycling stability for cyclohexane O/W emulsion for ten times

Full size image

Long-Term Separation of SDS-Free/SDS-Stabilized Liter-Scale Soybean/Crude O/W Emulsions Under Harsh Conditions

To assess the separation efficiency of our p-PAN NM for large-scale viscous O/W emulsion separation, we employed a home-made filtration device (ESI; Fig. S6). A beaker containing 1.1 L of viscous O/W emulsion was subjected to moderate magnetic stirring (300 rpm) to simulate the existing O/W emulsions in a flowing aquatic system. The left flask collected the filtrate while the right one prevented sucking-back. The entire setup was driven by a water pump. Data in Fig. 7a, b indicated that both liter-scale creamy SDS-free soybean and crude O/W emulsions (feed) turned transparent after filtration (filtrate). The size distribution variations of the viscous O/W emulsions in the feeds and filtrates were measured using dynamic light scattering (DLS) (ESI; Fig. S7). To evaluate the long-term separation efficiency and cycling stability, Fig. 7c, d depicts the variation in permeating flux versus time during five cycles of separation. From the curves, it can be observed that the permeating flux of both SDS-free O/W emulsions could be completely recovered, achieving 100% flux recovery ratio (FRR) (Fig. 7e and f) for 170 h of continuous separation.

Fig. 7
figure 7

a Digital photos of the process of liter-scale viscous O/W emulsions separation. b Photographs of the feed and filtrate of SDS-free soybean/crude O/W emulsions before and after separation. c and d are the variation of flux versus time of SDS-free soybean/crude O/W emulsions separation for 5 cycles, respectively. e and f the variation of separation efficiency for SDS-free soybean/crude O/W emulsions separation for 5 cycles, respectively. g and h are SDS-stabilized soybean and crude O/W emulsions separations coupled with different media for 5 cycles. i is the separation efficiency of these two SDS-stabilized O/W emulsions for 5 cycles. j and k are separation efficiency of these two SDS-stabilized O/W emulsions coupled with different media

Full size image

The components present in real oil-contaminated effluents, generated from industrial processes and daily activities, are rather complex. Various pollutants, including acids, alkalis, salts, and surfactants, often coexist with oil droplets. To evaluate the purification capacity of p-PAN NM under harsh conditions for real oil-contaminated effluents, we conducted long-term continuous separation experiments using liter-scale SDS-stabilized soybean O/W and crude O/W emulsions in the presence of corrosive media. Figure 7g illustrates the long-term cycling performance of p-PAN NM for SDS-stabilized soybean O/W emulsion coupled with H2O, 1 M HCl, 1 M NaOH, or 10 wt% NaCl over five cycles. From these curves, it can be observed that our p-PAN NM exhibits excellent recycling stability. Even when the separation time reached 265 h (Fig. 7h) for SDS-stabilized crude O/W emulsion coupled with H2O, 1 M HCl, 1 M NaOH, and 10 wt% NaCl over five cycles, a 100% FRR was achieved, demonstrating the robustness and ultra-long-term cycling stability of p-PAN NM under harsh conditions. Figure 7i, j, and k clearly demonstrates the maintained high separation efficiency (≥ 99.0% for crude O/W emulsions and 99.2% for soybean O/W emulsion), regardless of the corrosive media. This exceptional ultra-long-term continuous separation capacity for surfactant-stabilized viscous O/W emulsions under harsh conditions surpasses most of the reported literature. To highlight the advantages of this work, Table 1 provides a comparison of separation efficiency, continuous separation time, targets, FRR, and harsh conditions with similar literature.

Table 1 Comparison of separation efficiency (SE), continuous separation time (CST), separated targets (ST), FRR, and harsh conditions Testing (HCT) in similar literatures and our work
Full size table

Fouling-Recovery/Theoretical Clarification and Mechanism of Separating O/W Emulsions

To vividly demonstrate the fouling recovery through ultrasound treatment in an 80 ℃ hot water bath, we utilized Ampliflu Red dyed SDS-stabilized soybean and crude O/W emulsions. Upon separation, it was observed that the surface of fouled p-PAN NM was completely covered by viscous oil cakes, which were clearly visible under a UV lamp. However, after subjecting the p-PAN NM to sonication treatment, these cakes disappeared entirely (Fig. 8a). Figure 8b displays the activation energy (Ea) values of p-PAN against water, cyclohexane/dichloroethane/chloroform vapors at 80 ℃. The data reveal an increased Ea value for water on p-PAN, indicating an enhanced affinity between water and p-PAN at 80 ℃. Conversely, the Ea values between p-PAN and these oil vapors decreased. Furthermore, the viscosity of viscous oil decreases at high temperature [69]. Considering the heightened affinity between water and p-PAN, reduced affinity between oils and p-PAN, decreased viscosity of viscous oil, as well as the vibrational fluid motion during ultrasound treatment [70], it can be inferred that these oil cakes on the surface or within the voids of p-PAN NM will be completely removed.

Fig. 8
figure 8

a Photos on the surficial variations of viscous oil fouled NM before and after ultrasound. b Electrostatic potential diagram on Ea of the water, hexane, dichloroethane and chloroform absorbed on p-PAN NM at 80 °C. c Schematic diagram of the mechanism on separating O/W emulsions

Full size image

Figure 8c illustrates the emulsion separation mechanism of p-PAN NM. When water comes into contact with the surface of p-PAN NM in air, it diffuses and penetrates the membrane directly (left side of Fig. 8c), forming a stable water layer. According to Eq. (3), the porous surface of the water-covered p-PAN NM can withstand a certain intrusion pressure (ΔP > 0) when submerged in water (right side of Fig. 8c), as the contact angle of oil droplets on the surface is greater than 90°.

$${\Delta }P = \frac{2\gamma }{R} = - \frac{{l\gamma \left( {\cos \theta } \right)}}{A},$$
(3)

where γ is the oil–water interfacial tension; R is the meniscus’s radius; l is the circumference of the pore; A is the cross-sectional area of the pore; θ is the advancing contact angle of oil on the nanofiber surface. The water film on the surface of p-PAN NM creates underwater superoleophobicity and oil-repellent properties. Throughout the separation process, the demulsified viscous oil droplets approach the NM surface and steadily accumulate due to transmembrane pressure, resulting in a reduction in permeating flux.

Conclusion

Electrospun pre-oxidized polyacrylonitrile (p-PAN) NM, with robust anti-corrosion, firm fouling recovery, and superhydrophilicity/underwater superoleophobicity has been prepared via one-pot Pre-oxidation process. Our p-PAN NM displayed high separation efficiency (~ 99%) and ~ 100% oil-fouling recovery for continuously (265 h) separating large-scale SDS-stabilized soybean/crude oil-in-water emulsions coupled with corrosive media. DFT computations have been adopted to clarify the robust fouling recovery and the variation of surficial wettability after Pre-oxidation. These outstanding filtration functions open a route for membrane filtration for durable and massive viscous oils/water emulsions separation under harsh conditions in practice.