Introduction

Oil bearing wastewater derived from oil spillage and industrial organic solvent discharge usually contains toxic substances, which have brought severe threats to aquatic ecosystems and even the health of human beings (Oliveira et al. 2021; Chen et al. 2019; Grabowski et al. 2017). Traditional technologies of isolating these pollutants usually suffer some limitations, such as high costs, and time consumption, resulting in secondary contamination, and low efficiency. Therefore, developing functional materials for high-efficiency separation of oil/water mixtures and emulsions has become an urgent demand.

Various functional materials, including meshes, sponges, aerogels, and fabrics, with superhydrophobicity/superoleophilicity (Shang et al. 2019, 2020; Zhang et al. 2020a, b; Renjith et al. 2021) have been designed and prepared for removing oils from wastewater owing to their high selectivity, high separation performance, and low energy consumption. Among these, cotton fabric has received broad attention because of its renewability, biodegradability, porosity, and good flexibility. Nevertheless, the intrinsic hydrophilicity of cotton fabric limits its application in oily wastewater treatment and oil purification. Hence, surface modification is highly demanded to transform cotton fabrics into superhydrophobic materials. A variety of approaches, such as dip-coating (Bai et al. 2020; Song et al. 2020), solvothermal reaction (Zhu et al. 2017; Ning et al. 2017), spray method (Park et al. 2020; Cao et al. 2020), electrophoretic deposition (Saji 2021; Ogihara et al. 2011), and chemical vapor deposition (Mosayebi et al. 2020; Zheng et al. 2009), have been developed for manufacturing superhydrophobic surfaces. All researches have proved that the combination of rough surface structure and low-surface-energy components is the most representative strategy in the design of superhydrophobic materials (Wang et al. 2020; Hu et al. 2020; Zhang et al. 2021; Si et al. 2018). Many studies have utilized nanoparticles to build micro/nano-roughness in superhydrophobic coatings (Shang et al. 2018; Bao et al. 2020). However, the roughness of the superhydrophobic surfaces based on nanoparticles is easily destroyed by mechanical abrasion, leading to the reduction of the superhydrophobicity. Furthermore, expensive fluorine-containing reagents are often utilized as low-surface-energy materials for superhydrophobic surfaces (Klimov et al. 2021; Zhang et al. 2018), which may increase costs and cause harm to human health and the environment. Therefore, the exploration of durable, nanoparticles-free, and fluorine-free superhydrophobic surfaces has become a research hotspot in the field oil–water separation.

As a new type of phenolic resin, benzoxazine is an important inexpensive alternative to prepare superhydrophobic surface because of its low cost, low water retention, mechanical robustness, and thermal stability (Wang et al. 2017; Malakooti et al. 2019; Zhang et al. 2020a, b; Qin et al. 2020). Benzoxazines with flexible molecular structure can be synthesized using phenols, primary amines, and formaldehyde through Mannich condensation reaction. Due to the shortage of petrochemical resources, the exploration of biomass with sustainable and renewable characteristics has received wide attention (Amarnath et al. 2019; Lyu and Ishida 2019). As a renewable phenol source, cardanol is a by-product of cashew nut shell processing. It has a C15 long aliphatic side chain with different degrees of unsaturation. Due to its renewability and commercially scalability, cardanol has been utilized as a substitute for petrochemical phenols to synthesize bio-based benzoxazine resin for oil–water separation (Prabunathan et al. 2020; Arumugam et al. 2021). For example, Arumugam et al. (2021) reported the oil–water separation efficiency of the cardanol-based polybenzoxazines coated cotton fabric was 95%. Yao et al. (2019) also reported the cardanol-based polybenzoxazine/SiO2 modified fabric was used for separating of several heavy oils/water mixtures with oil fluxes in the range of 6500–9500 L·m−2·h−1 and separation efficiencies of all over 90%. Although excellent separation performance of cardanol-based superhydrophobic surfaces has been achieved for oil/water mixtures, the investigation for separating emulsified mixtures is still lacking. Recently, our group developed superhydrophobic cotton fabric with silylated cardanol-based polybenzoxazine for efficient separating of oil/water mixtures and emulsions through the sequential enzyme-etching, dip-coating, and thermal curing process (Shang et al. 2021). However, the enzyme-etching process for creating rough structure may limit the large-scale preparation of the superhydrophobic cotton fabric. Therefore, it is urgent to simplify the preparation process of cardanol benzoxazine-based superhydrophobic materials for oil/water emulsion separation.

In this work, a novel cardanol/POSS-based benzoxazine (C-aPOSS) was synthesized by using cardanol, paraformaldehyde and aminopropyllsobutyl-POSS through Mannich condensation. Then, a durable superhydrophobic cardanol/POSS-based polybenzoxazine (PC-aPOSS) coated cotton fabric was fabricated through dip-coating technique and thermal curing process. Thanks to the rigid cubic cage structure of POSS, a hierarchical rough structure was formed on the surface of the modified cotton fabric without additional treatment. Excellent mechanical and chemical durability of the modified cotton fabric were observed because of the remarkable chemical and physical stability of the PC-aPOSS. Due to its superhydrophobic and superoleophilic properties, the modified cotton fabric exhibited high performance in the separation of a variety of oil/water mixtures and emulsions. This simple and facile synthesis approach provides an avenue for large-area creation of superhydrophobic materials for oily wastewater treatment.

Experimental section

Materials

Cardanol was supplied by Jining Huakai Resin Co., Ltd. Aminopropyllsobutyl-POSS (NH2-POSS) was purchased from Hybrid Plastics (USA). Oil red, Span80, cyclohexane, paraformaldehyde and petroleum ether were obtained from Energy Chemical (Shanghai). Absolute ethanol, toluene, 1, 4-dioxane, trichloromethane, dichloromethane and dichloroethane were purchased from Nanjing Chemical Reagent Co., Ltd. Cotton fabric (Material: 100% cotton; Style: plain woven; Fabric Squared Weight: 150 g·m−2; Yarn Count: 20S*20S; Ends Picks: 60*60) was purchased from local mall.

Synthesis of C-aPOSS

Paraformaldehyde (0.4 mol), cardanol (0.2 mol) and NH2-POSS (0.2 mol) were added into a round-bottom flask and then dissolved in 200 mL of 1, 4-dioxane. The solution was heated to 90 °C and refluxed for 8 h. After the reaction, reddish-brown C-aPOSS was obtained by removing the solvent using rotary evaporation.

Fabrication of PC-aPOSS/CF

CF was washed using an ethanol/water mixture to remove the surface impurities before use. C-aPOSS was dissolved in absolute ethanol at concentrations of 1.0 wt.%, 1.5 wt.%, 2.0 wt.%, 2.5 wt.% and 3.0 wt.%. Subsequently, CF (3 cm × 3 cm) was immersed in 50 mL of different C-aPOSS concentration ethanol solution for 30 min, then taken out and dried at 70 °C for 30 min. After further heating at 190 °C for 1 h, the polybenzoxazine coated CF (PC-aPOSS/CF) was formed. In addition, PC-aPOSS/CF was prepared by immersing CF in the 2.0 wt% C-aPOSS solution with different time (10 ~ 60 min) to study the effect of dipping time on the hydrophobicity.

Characterizations

The surface morphologies and elements of the samples were analyzed by Field Emission Scanning Electron Microscopy (FE-SEM, Zeiss supra 55) equipped with an Energy Dispersive Spectrometer (EDS). The 3D topography and surface roughness of the samples were detected by using a Confocal Laser Scanning Microscopy (CLSM, KEYENCE VK-150 K). The chemical structure of the samples was analyzed by Fourier Transform infrared spectroscopy (FT-IR, Nicolet iS 10) and X-ray photoelectron spectroscopy (XPS, ESCALAB250Xi). The water contact angle (WCA) and sliding angle (SA) of the samples were recorded by a contact angle meter (CV-705B, CVOK, Dongguan) with 5 μL of water. At least five positions were measured for each sample. The emulsion droplet images were characterized using an Olympus Optical Microscope (BH53). The size distributions of the emulsion droplets were tested using a Mastersizer 2000. The air permeability of the samples was characterized by using a fabric permeability instrument (YG461E) with an operating pressure of 100 Pa. The test of mechanical property was performed on a UTM 4304 universal tester with a speed of 2 mm·min−1.

Mechanical durability and environmental stability

The mechanical durability of the PC-aPOSS/CF was investigated by abrasion test and ultrasound treatment. In the abrasion test, the PC-aPOSS/CF (3.0 cm × 3.0 cm) placed on 1000 grit sandpaper, loaded with a weight of 200 g, and then moved linearly by 10 cm for an abrasion cycle. The WCA and SA were recorded every 10 cycles. In the ultrasound treatment, the PC-aPOSS/CF was dipped into ultrapure water and put into the ultrasonic cleaning machine (KH7200DB, Hechuang Ultrasonic). The WCA and SA were evaluated every 10 min.

The corrosive resistance was tested to evaluate the durability of the PC-aPOSS/CF in harsh operating environment. For the corrosive resistance, corrosive liquids with pH value between 1 and 13 were prepared by diluting HCl and NaOH solutions. The corrosive liquids were dropped on the PC-aPOSS/CF surface and then the WCA was recorded. Furthermore, PC-aPOSS/CF was immersed in HCl solution (1 M) or NaOH solution (1 M) for 24 h to investigate its corrosive resistance.

Oil/water separation

Toluene, cyclohexane, petroleum ether, trichloromethane, dichloromethane and dichloroethane were chosen for oil/water separation experiment. Oil/water mixtures were prepared by mixing oils with ultrapure water in a volume ratio of 1:1. Surfactant-free water-in-oil emulsions were prepared by mixing ultrapure water and oils (v: v, 1:99) using an emulsion machine (T10, IKA) at 10,000 rpm for 10 min. Six kind of surfactant-stabilized water-in-oil emulsions were prepared by adding Span80 (1 g) and ultrapure water (10 ml) into oils (90 ml) respectively. The mixtures were emulsified at 10,000 rpm for 10 min to obtain the stable emulsions.

The experiment for separating oil/water mixtures was performed on the homemade separation device by gravity using a piece of PC-aPOSS/CF as the filter membrane. For separating surfactant-free and surfactant-stabilized water-in-oil emulsions, a polyurethane (PU) sponge was wrapped by two pieces of PC-aPOSS/CF to form sandwich-like composite filter membrane. The separation experiment was performed under external pressure (0.2 bar). The water content in the filtrate was tested by using a Karl Fischer Moisture Titrator (Mettler Toledo, ET-08). The oil purity was calculated according to the water content in the filtrate.

The fluxes for oil/water mixture and emulsion were measured according to Eqs. (1) and (2) respectively.

$${\text{Flux}} = \frac{V}{A \times t}\quad \left( {{\text{L}}\;{\text{m}}^{ - 2} \;{\text{h}}^{ - 1} } \right)$$
(1)
$${\text{Flux}} = \frac{V}{A \times t \times P}\quad \left( {{\text{L}}\;{\text{m}}^{ - 2} \;{\text{h}}^{ - 1} \;{\text{bar}}^{ - 1} } \right)$$
(2)

where V is the oil volume (L), A is the separation area (m2), t is the separation time (h) and P is the operating pressure (bar).

The repeated flux and oil purity for the oil/water mixture and emulsion were recorded to evaluate the reusability of the PC-aPOSS/CF. The contaminated PC-aPOSS/CF was cleaned by using ethanol and dried after each cycle for further use.

Results and discussion

Synthesis and characterization of C-aPOSS

A novel C-aPOSS monomer was synthesized through Mannich condensation reaction between cardanol, paraformaldehyde and NH2-POSS. The synthesis route of C-aPOSS is revealed in Fig. 1a. The chemical structure of C-aPOSS was characterized by FT-IR and 1H NMR. Figure 1b displays the FT-IR spectra of cardanol, NH2-POSS and C-aPOSS. In the C-aPOSS spectrum, the typical peak at 1253 cm−1 for the C–O–C asymmetric vibration, the peak at 1365 cm−1 for the C–N–C bending vibration and the peak at 887 cm−1 for the oxazine ring attached in the benzene ring were detected (Devaraju et al. 2019), which indicated the successful formation of the benzoxazine structure. The peaks at 2953, 2924 and 2853 cm–1 were assigned to the stretching vibrations of methyl and methylene groups on the long alkyl carbon chain of cardanol and the isobutyl of NH2-POSS. The peak at 1462 cm–1 confirmed the C–H asymmetrical deformation vibration, whereas the peaks at 3008 and 1587 cm–1 revealed the unsaturated double bond and the benzene ring structure of cardanol. Also, the peak at 1094 cm–1 denoted the asymmetric stretching vibration of Si–O–Si group. Figure 1c reveals the 1H NMR spectrum of C-aPOSS. From the spectrum, the peaks of oxazine ring were appeared at 3.75–3.95 (a) and 4.70–4.90 (b) ppm, which were ascribed to Ar–CH2–N– and –O–CH2–N– groups respectively. The peaks detected at 2.27 (c) and 1.58 (d) ppm correlated with the propyl chain of –N–CH2–C– and –C–CH2–C groups respectively. The methylene peak of –C–CH2–Si– group was detected at 0.57–0.62 (e) ppm. The methylene peak of NH2-POSS was observed at 1.85 (f) ppm. The peaks for the methyl protons of cardanol and NH2-POSS appeared at 0.85–0.91 (4) and 0.92–1.00 (g) ppm respectively. The peaks at 5.00 (6), 5.35 (5) and 5.80 (10) ppm corresponded to the unsaturated protons of cardanol. The peaks at 6.48–7.15 (11 ~ 13) ppm were assigned to the aromatic groups of cardanol. The peaks at 1.34 (3), 1.65 (2), 2.05 (8, 9) and 2.85 (7) ppm were attributed to the methylene protons of the long alkyl chain of cardanol. These results demonstrate that the C-aPOSS was successfully synthesized from cardanol and NH2-POSS.

Fig. 1
figure 1

a Synthesis route of C-aPOSS. b FT-IR spectra of cardanol, NH2-POSS and C-aPOSS. c 1H NMR spectrum of C-aPOSS

Full size image

Fabrication of superhydrophobic PC-aPOSS/CF

The schematic preparation strategy of superhydrophobic PC-aPOSS/CF is presented in Fig. 2. Initially, C-aPOSS was dispersed in ethanol to prepare a uniform dispersion. The C-aPOSS coated CF (C-aPOSS/CF) was prepared by immersing a piece of pristine CF into the C-aPOSS ethanol solution with ultrasonic treatment for 30 min. After drying at 70 °C for 30 min, the C-aPOSS/CF was cured at 190 °C for 1 h to obtain the superhydrophobic PC-aPOSS/CF. The C-aPOSS monomers was polymerized on the fiber surface through thermal treatment ring-opening reaction to form polybenzoxazine. During the polymerization process, the C–O bond of the oxazine ring was cleaved and reacted with the aromatic ring of another C-aPOSS molecule, which resulted in a cross-linking network structure of polybenzoxazine on the fiber.

Fig. 2
figure 2

Schematic diagram of fabrication of superhydrophobic PC-aPOSS/CF by in-situ polymerization

Full size image

It is well known that low-surface-energy chemistry and hierarchical rough structure are essential for the fabrication of superhydrophobic surfaces. Herein, the polybenzoxazine PC-aPOSS was used as the low-surface-energy material due to the oxazine rings and long alkyl chains in its molecule, while the hierarchical rough morphology on the CF surface originated from the agglomeration of the PC-aPOSS polymers.

The effect of C-aPOSS concentration on the surface morphology and hydrophobicity of the PC-aPOSS/CF was investigated, as shown in Fig. 3a–g. The changes of 3D topography and roughness of the PC-aPOSS/CF with different C-aPOSS concentrations were characterized by CLSM (Figure S1). Figure 3a shows that the fiber of the pristine CF had a flat surface with intrinsic winkles. The surface profile of the pristine CF showed the interweaving structure of fibers with an arithmetic average of the roughness profile (Sa) of 24.2 μm (Figure S1(a)). The pristine CF had a WCA of 0° and an oil contact angle of 0° (Fig. 5a), indicating its intrinsic superhydrophilicity/superoleophilicity. After coating PC-POSS, the surface profiles of the PC-aPOSS/CF were similar with that of the pristine CF, but the values of Sa changed significantly (Figure S1(b-f)). When the concentration of C-aPOSS was 1.0 wt.%, the intrinsic winkles of the fiber were filled by the PC-aPOSS polymers and the rough structure was formed on the fiber surface due to the agglomeration of POSS in the polybenzoxazine (Fig. 3b). The Sa of the PC-aPOSS/CF for 1.0 wt.% C-aPOSS increased to 25.4 μm (Figure S1(b)). The PC-aPOSS/CF for 1.0 wt.% C-aPOSS had a WCA of 149.5° ± 1.2° and a SA of 7.2° ± 0.6° (Fig. 3g), indicating its high hydrophobicity. Increasing the C-aPOSS concentration to 1.5 wt.%, more POSS aggregates were formed on the fiber surface with increasing roughness (Sa = 27.1 μm), as shown in Fig. 3c and Figure S1(c). The WCA of PC-aPOSS/CF for 1.5 wt.% C-aPOSS increased to 152.2° ± 1.1° (Fig. 3g), demonstrating the superhydrophobicity of the PC-aPOSS/CF. When the C-aPOSS concentration was 2.0 wt.% (Fig. 3d), the fiber surface was covered by many protuberances with a roughness of Sa = 31.2 μm (Figure S1(d)). The PC-aPOSS/CF had a maximum WCA of 155.0° ± 0.3° and a minimum SA of 5.5° ± 0.5° (Fig. 3g). These results indicated that the superhydrophobic surface of the PC-aPOSS/CF was in the Cassie-Baxter state. According to the Cassie-Baxter theory (Cassie and Baxter, 1944), the pores and intervals between the fibers and rough structures could trap masses of air to provide a composite surface. The trapped air formed a cushion layer, which could effectively prevent water droplets from penetrating into the composite surface interior, resulting in superhydrophobicity. With the further increasing of C-aPOSS concentration to 2.5 wt.% and 3.0 wt.%, the continuous polymer coatings with massive POSS aggregates were formed on the fiber surfaces (Fig. 3e, f). However, the roughness of the PC-aPOSS/CF was reduced to Sa = 29.6 μm for 2.5 wt.% C-aPOSS (Figure S1(e)) and Sa = 29.2 μm for 3.0 wt.% C-aPOSS (Figure S1(f)), which might be caused by filling excess PC-aPOSS in the gaps between the fibers. Accordingly, the WCA of the PC-aPOSS/CF decreased to 151.9° ± 0.5° for 2.5 wt.% C-aPOSS and 151.4° ± 1.1° for 3.0 wt.% C-aPOSS (Fig. 3g). Thus, the 2.0 wt.% C-aPOSS solution was used in the follow-up study.

Fig. 3
figure 3

SEM images of the CF (a) and PC-aPOSS/CF prepared with different C-aPOSS concentrations (1.0 wt.% (b), 1.5 wt.% (c), 2.0 wt.% (d), 2.5 wt.% (e) and 3.0 wt.% (f)). Effects of (g) C-aPOSS concentration and (h) dipping time on WCA and SA of the PC-aPOSS/CF

Full size image

Furthermore, the effect of dipping time on the hydrophobicity and surface morphology of the PC-aPOSS/CF with 2.0 wt.% C-aPOSS was studied (Fig. 3h and Figure S2). The dipping time of the CF in C-aPOSS solution could also affect the loading content of PC-aPOSS on the CF (Figure S2). Figure 3h shows that the WCA and SA of the PC-aPOSS/CF altered with dipping time and the most appropriate one was 30 min. Further increasing the dipping time could cause an aggregation of PC-aPOSS and decrease the surface roughness (Figure S2), which resulted in the reduction of the hydrophobicity. Therefore, the superhydrophobic PC-aPOSS/CF prepared with 2.0 wt.% C-aPOSS and 30 min dipping time was chosen for the following studies.

Chemical structure and composition of the PC-aPOSS/CF

The chemical structure of the C-aPOSS, pristine CF and PC-aPOSS/CF were characterized by FT-IR (Fig. 4a). In the spectrum of PC-aPOSS/CF, the peaks of the oxazine ring at 1365, 1253 and 887 cm−1 disappeared, indicating that the C-aPOSS monomer was completely cross-linked and polymerized to form polybenzoxazine. Meanwhile, some new peaks were detected in the spectrum of PC-aPOSS/CF compared with that the pristine CF. The peaks at 2953, 2924 and 2853 cm–1 were assigned to the methyl and methylene stretching vibrations of the long alkyl carbon chain of cardanol moiety and the isobutyl groups of NH2-POSS. And the Si–O–Si peak was detected at around 1094 cm–1 in the PC-aPOSS/CF. XPS was used to further characterize the chemical compositions of the pristine CF and PC-aPOSS/CF. As revealed in Fig. 4b, the surface of the pristine CF was mainly composed of C and O elements. However, C, N, O and Si elements were detected on the surface of the PC-aPOSS/CF. The N element showed a peak at 399.9 eV (N 1 s), and the Si element displayed two peaks at 102.3 eV (Si 2p) and 153.1 eV (Si 2 s), indicating the presence of PC-aPOSS on the surface of the cotton fabric. EDS results further proved that the C, O, N and Si elements were the main components of the superhydrophobic PC-aPOSS/CF, as shown in Fig. 4c. The element mappings in Fig. 4c reveal that the PC-aPOSS coating successfully covered the surface of cotton fabric.

Fig. 4
figure 4

a FT-IR spectra of the C-aPOSS, pristine CF and PC-aPOSS/CF. b XPS spectra of the pristine CF and PC-aPOSS/CF. c EDS spectrum and element mappings of PC-aPOSS/CF

Full size image

Superhydrophobicity/superoleophilicity of PC-aPOSS/CF

The super-wettability of pristine CF and PC-aPOSS/CF were investigated. As shown in Fig. 5a, the water-based liquids and cyclohexane could spread on the surface of the pristine CF, which confirmed the superhydrophilicity/superoleophilicity of the pristine CF. Figure 5b shows that the water-based liquids even including corrosive liquids were repelled by the PC-aPOSS/CF and formed ball-like shapes due to the superhydrophobicity of the PC-aPOSS/CF. However, cyclohexane was adsorbed upon contact with PC-aPOSS/CF, indicating the superoleophilicity of the PC-aPOSS/CF. The superoleophilicity was attributed to the oleophilicity of the alkyl chains in the PC-aPOSS, which was enhanced by the rough structure of the PC-aPOSS/CF. In addition, after modifying PC-aPOSS coating, the color of the cotton fabric changed from white to yellowish-brown (Fig. 5a, b), which was caused by the reddish-brown color of cardanol-based benzoxazine monomer and the yellowing of the benzoxazine monomer after polymerization (Li et al. 2018). However, the PC-aPOSS/CF exhibited uneven color due to the aggregation of the benzoxazine monomers caused by the rapid evaporation of ethanol during drying.

Fig. 5
figure 5

Photographs of water-based droplets and cyclohexane on (a) pristine CF and (b) PC-aPOSS/CF. c PC-aPOSS/CF on the oil–water interface. d Mirror-like phenomenon of PC-aPOSS/CF underwater. e Low adhesion between PC-aPOSS/CF and water droplet

Full size image

Figure 5c exhibits that the pristine CF sank to the bottom due to its superhydrophilicity, while PC-aPOSS/CF could float on the interface of oil and water phases because of its superhydrophobicity/superoleophilicity. Figure 5d shows that the PC-aPOSS/CF exhibited a mirror-like phenomenon after immersing in water because of the trapped air in the rough surface, confirming its excellent superhydrophobicity. As shown in Fig. 5e, the low adhesion of the PC-aPOSS/CF to water droplet was investigated. After the process of approaching, deforming and leaving, the water droplet could be completely separated from the PC-aPOSS/CF surface, which demonstrated the ultralow adhesion between the PC-aPOSS/CF and water droplet.

Mechanical durability and environmental stability of PC-aPOSS/CF

In real-world applications, superhydrophobic surfaces usually lose their superhydrophobicity due to the damage of rough structure and depletion of low-surface-energy component after exposure to abrasion, ultrasound, or acid/alkali environment. In this work, the abrasion test and ultrasound treatment were used to investigate the durability of the PC-aPOSS/CF. Figure 6a exhibits the change of WCA and SA with abrasion cycles. The PC-aPOSS/CF had a WCA of 147.0° ± 0.5° and a SA of 14.5° ± 1.0° even after 200 sandpaper abrasion tests, indicating its great mechanical durability. In addition, Fig. 6b displays the wettability of the PC-aPOSS/CF after ultrasonic shaking in water for different time. After ultrasonic shaking for 60 min, the WCA was maintained at 147.7° ± 0.6° and the SA was lower than 9°, demonstrating that the PC-aPOSS/CF had outstanding ultrasonic resistance. The excellent mechanical durability of the superhydrophobic PC-aPOSS/CF prepared by this method might be due to the intrinsic stability of POSS and the cross-linking structure of polybenzoxazine.

Fig. 6
figure 6

The changes of the WCA and SA of the PC-aPOSS/CF with (a) different abrasion cycles and (b) different ultrasonic time. c WCA and SA of corrosive liquids (pH = 1 ~ 13) on the PC-aPOSS/CF. d Photographs of the PC-aPOSS/CF after soaking in acid and alkaline solutions for 24 h

Full size image

To explore the stability of the PC-aPOSS/CF under aqueous acid and alkaline solutions, the WCA and SA of corrosive liquids (pH = 1 ~ 13) on the PC-aPOSS/CF surface were measured (Fig. 6c). Compared with pure water, the CAs of corrosive liquids was reduced because the corrosive liquids destroyed the Si–O–Si bonds of POSS in the polybenzoxazine coating. However, the PC-aPOSS/CF could maintain high hydrophobicity after soaking in acid and alkaline solutions for 24 h (Fig. 6d), demonstrating that it had great chemical corrosion resistance.

Subsequently, the mechanical properties and air permeability of the pristine CF and PC-aPOSS/CF were investigated, as shown in Table 1. It can be seen from the table, the tensile strength and elongation at break of the PC-aPOSS/CF decreased compared with that of the pristine CF, which was caused by the high temperature treatment during the thermal curing process. In addition, the air permeability of the fabric was also reduced after the modification of PC-aPOSS coating, which might be due to the reduction of the gaps between the fibers by the loaded polybenzoxazine resin.

Table 1 Physical properties of the pristine CF and PC-aPOSS/CF
Full size table

Separation of oil/water mixtures

Taking the advantages of the superhydrophobility/superoleophilicity, the as-prepared PC-aPOSS/CF is very suitable for oil/water separation. As revealed in Fig. 7a, b, a homemade filter equipment with the PC-aPOSS/CF as filter element was used to separate various oil/water mixtures. A variety of light oils and weight oils were employed as models for the oil/water separation. The fluxes of pure oils were measured (Fig. 7c). The fluxes of toluene, petroleum ether, cyclohexane, dichloromethane, trichloromethane and dichloroethane were 5587 ± 34, 5954 ± 45, 4075 ± 26, 10,431 ± 416, 8810 ± 437 and 6050 ± 25 L·m−2·h−1 respectively. The flux difference was mostly related to the densities of the used oils/organic solvents. Due to the impediment of water, the fluxes of the light oil/water mixtures were dramatically lower than that of the pure oils, but they were still higher than 3600 L·m−2·h−1 (Fig. 7d). The oil purities towards these oil/water mixtures were all higher than 99.86% (Fig. 7d). Particularly, the oil purities of petroleum ether and cyclohexane reached 99.99%, which was greater than other superhydrophobic fabrics (Pal et al. 2021; Yao et al. 2020; Lu et al. 2020; Xu et al. 2020; Sam et al. 2021). Furthermore, the recyclability of the PC-aPOSS/CF for separating trichloromethane/water mixture was investigated, as shown in Figure S3. After 50 separation cycles, the oil purity of trichloromethane was higher than 99.90% and the oil flux was remained at 8000 L·m−2·h−1, demonstrating the long-term recyclability of the PC-aPOSS/CF. Figure S4 shows the WCA and SA of the PC-aPOSS/CF after different separation cycles. The WCA decreased and the SA slightly increased with the increase of separation cycles. However, after 50 separation cycles, the WCA remained above 150° and the SA was lower than 7°, which indicated the robust superhydrophobicity of the PC-aPOSS/CF.

Fig. 7
figure 7

The process of oil/water separation: a cyclohexane/water mixture, b dichloromethane/water mixture. c Fluxes of various pure oils. d Oil purities and fluxes of various oil/water mixtures

Full size image

Separation of water-in-oil emulsion

In addition to the immiscible oil/water mixture, water-in-oil emulsions also have adverse effects on oil quality because of their stability and small size. Here, a composite filter membrane with sandwich structure was designed by wrapping a polyurethane (PU) sponge with two pieces of PC-aPOSS/CF (PC-aPOSS/CF@PU sponge@ PC-aPOSS/CF) (Fig. 8a). Due to the large interspace between interwoven fibers of the cotton fabric, low oil purity (~ 85%) was obtained by directly using PC-aPOSS/CF as a filter membrane. Consequently, the designed composite filter membrane was employed to improve the oil purity. Six kinds of surfactant-free water-in-oil emulsions, including water-in-toluene, water-in-petroleum ether, water-in-cyclohexane, water-in-dichloromethane, water-in-trichloromethane and water-in-dichloroethane, were prepared as working solutions for separation experiments. As depicted in Fig. 8b, the oil fluxes for the tested surfactant-free emulsions were 170,089 ± 8021, 245,930 ± 6403, 158,731 ± 4705, 248,954 ± 10,437, 166,627 ± 3874 and 127,289 ± 9977 L·m−2·h−1·bar−1. The oil purities of these surfactant-free emulsions were all higher than 99.93% (Fig. 8b). The water-in-cyclohexane emulsion was milky white liquid and numerous water droplets were detected in the optical microscope image (Fig. 8c). After filtering, the filtrate became clear and transparent, and almost no water droplets were observed in the filtrate (Fig. 8c). The other five emulsions presented similar results after separation with the designed composite filter membrane (Figure S5). Figure 8d, e reveals that the size distribution of water droplets in emulsion and filtrate. After filtration with the PC-aPOSS/CF@PU sponge@PC-aPOSS/CF composite filter membrane, the size distribution of water droplets in the filtrate was 0.6 ~ 1.5 nm, demonstrating its outstanding separation performance.

Fig. 8
figure 8

a Separation setup for water-in-oil emulsion. b Oil purities and fluxes for six kinds of surfactant-free water-in-oil emulsions. c Photograph and optical microscope images of the surfactant-free water-in-cyclohexane emulsion before and after the filtration. Size distributions of surfactant-free water-in-cyclohexane emulsion (d) and filtrate (e)

Full size image

It is noteworthy that the improved oil purity of the PC-aPOSS/CF@PU sponge@PC-aPOSS/CF composite filter membrane compared with that of the single PC-aPOSS/CF was due to the great adsorption ability of the PU sponge for the tiny water drops passing through the top layer of PC-aPOSS/CF. The comparison of the separation performance with the previously reported superhydrophobic separation materials is shown in Table 2. Compared with other separation materials, the composite filter membrane exhibited comparable or higher separation performance and flux for the tested water-in-oil emulsions. The recyclability of the composite filter membrane was assessed by using the cyclic separation tests of surfactant-free water-in-trichloromethane emulsion, as illustrated in Figure S6. The oil purity of the composite filter membrane was remained above 99.99% even after 50 separation cycles. The flux presented a slight decrease at the initial 5 separation cycles and had no significant fluctuation in the following separation process. After 50 separation cycles, the flux was maintained at 128,691 L·m−2·h−1·bar−1. These results proved that the composite membrane had the outstanding separation performance and long-term durability.

Table 2 Comparison of the separation properties of various superhydrophobic materials for separating of water-in-oil emulsions
Full size table

To further study the separation ability of the composite filter membrane, six kinds of surfactant-stabilized water-in-oil emulsions with water volume fraction of 10% were prepared. As shown in Fig. 9, the oil purities of all the surfactant-stabilized emulsions were higher than 99.87%, with a few up to 99.99%, indicating the excellent separation efficiencies. The fluxes of the six surfactant-stabilized water-in-oil emulsions based on an external pressure were 26,786 ± 1120, 30,157 ± 2170, 27,087 ± 1274, 25,956 ± 1805, 19,390 ± 3251 and 17,895 ± 2527 L m−2 h−1 bar−1 for toluene, petroleum ether, cyclohexane, dichloromethane, trichloromethane and dichloroethane respectively (Fig. 9). The oil purities of the surfactant-stabilized emulsions were close to those of surfactant-free emulsions. Although the fluxes were lower than those of the surfactant-free emulsions, they were still extremely high compared to those of other separation materials, such as a ZnO nanoflower@SiC composite ceramic membrane with a flux of ~ 1300 L m−2 h−1 bar−1 (Wei et al. 2022) and MnO2 & Carbon sphere composite membrane with fluxes in the range of 3032–15,199 L m−2 h−1 bar−1 (Chen et al. 2022). To clearly observe the effectiveness of this separation process, the optical microscope images of both emulsion and filtrate using surfactant-stabilized water-in-toluene emulsion as an example are shown in Figure S7. The emulsion was obviously converted into transparent liquid after being filtered. No droplets were observed in the filtrate, indicating the outstanding performance of the composite filter membrane for separating surfactant-stabilized water-in-oil emulsions.

Fig. 9
figure 9

Oil purities and fluxes for six kinds of surfactant-stabilized water-in-oil emulsions

Full size image

Figure 10 presents a possible emulsion separation mechanism to give a more comprehensive understanding of the separation process of the composite filter membrane for water-in-oil emulsion. When water-in-oil emulsion approached the composite filter membrane, the emulsified water droplets were captured and coalesced, then were repelled due to the superhydrophobicity of the upper PC-aPOSS/CF. Simultaneously, oil was allowed to pass through the upper PC-aPOSS/CF because of its superoleophilicity. The separation process was referred to the size-sieving filtration and coalescence. The PU sponge in the interlayer has three-dimensional porous structure and amphiphilicity for oil and water. Tiny water droplets could be adsorbed and coalesced with other water droplets to generate larger droplets in the long and interconnected microchannels of the PU sponge. When contacting the bottom superhydrophobic PC-aPOSS/CF, the large water droplets were repelled to achieve complete separation from the emulsion.

Fig. 10
figure 10

Separation process of the composite filter membrane for water-in-oil emulsion

Full size image

Conclusions

This work presents a facile method to fabricate an eco-friendly superhydrophobic/superoleophilic fabric from cardanol/POSS-based polybenzoxazine. The as-prepared superhydrophobic PC-aPOSS/CF had excellent separation performance for various oil/water mixtures, with high oil purity (> 99.86%) and high flux (> 3600 L m−2 h−1). Furthermore, the designed composite filter membrane with sandwich-like structure could be used to separate surfactant-free and surfactant-stabilized water-in-oil emulsions upon the application of external pressure. High oil fluxes in the range of 127,289–248,954 and 17,895–30,157 L m−2 h−1 bar−1 were obtained for surfactant-free emulsions and surfactant-stabilized emulsions respectively, with outstanding separation performance (oil purity > 99.86%). More importantly, the superhydrophobic PC-aPOSS/CF exhibited the excellent durability and long-term reusability, which fully meets the requirements for the recovery of oil pollutants. Moreover, the facile and environmentally friendly method will effectively motivate the practical applications of superhydrophobic fabric and expand its application prospects in real life.