Cellulosic paper-based membrane for oil-water separation enabled by papermaking and in-situ gelation

Cellulosic paper-based membranes reveal incalculable application values for oil–water separation owing to their renewability and biodegradability. In this work, a cellulosic paper-based membrane was prepared via papermaking and in-situ gelation. The porosity of the membrane can be tailored by pulp refining in papermaking to further adjust the flux of membranes. The as-prepared membrane can separate oil–water emulsion efficiently with separation efficiency > 98.5%. The membranes also showed good underwater oil repellency due to the hydration layer. Besides, the wet strength of the paper-based membranes was enhanced by micro-dissolved and in-situ gelation to suit the underwater separation process. The membrane is expected to be a low-cost, highly-efficient material for oily wastewater purification. This work demonstrates a new idea for the development of oil–water separation and papermaking, which provides a feasible strategy for large scale production of fully biodegradable oil–water separation membranes.


Introduction
The rapid development of industry has led to an increase in the use of petroleum products, which produces tremendous oily wastewater pollution during the transportation and usage. The industrial effluent caused severe environmental problems as well as harm to human health (Cai et al. 2019;Chen et al. 2016;Dalton and Jin 2010). Thus, constructing oil-water separation materials for the effective treatment of oily wastewater is critical. (Li et al. 2017). The oil in oily wastewater exists primarily in three states, including floating, dispersed and emulsified (Chakrabarty et al. 2008). The separation of floating oil and dispersed oil from wastewater is easy, whereas the separation of emulsified oil is relatively difficult. Among the existing methods of oil-water separation, the membrane separation is a more convenient and feasible way (Zhu et al. 2014). Membranes with special wettability are ideal for oil-water separation (Ahmad et al. 2011;Feng and Jiang 2006;Wang et al. 2015). Among them, the membranes with superhydrophilicity and underwater superoleophobicity properties have wide potential application in oil-water separation due to their excellent underwater oil repellence and great oil-water separation performance (Ao et al. 2018; Abstract Cellulosic paper-based membranes reveal incalculable application values for oil-water separation owing to their renewability and biodegradability. In this work, a cellulosic paper-based membrane was prepared via papermaking and in-situ gelation. The porosity of the membrane can be tailored by pulp refining in papermaking to further adjust the flux of membranes. The as-prepared membrane can separate oil-water emulsion efficiently with separation efficiency > 98.5%. The membranes also showed good underwater oil repellency due to the hydration layer. Besides, the wet strength of the paper-based membranes was enhanced by micro-dissolved and in-situ gelation to suit the underwater separation process. The membrane is expected to be a low-cost, highlyefficient material for oily wastewater purification. This work demonstrates a new idea for the development of oil-water separation and papermaking, which provides a feasible strategy for large scale production of fully biodegradable oil-water separation membranes. Hu et al. 2016;Zhang et al. 2018). However, the majority of materials used in oil-water separation membranes are non-renewable or non-degradable materials, posing a risk of secondary pollution to the environment (Anis et al. 2021;Ao et al. 2020;Gu et al. 2017). Therefore, developing environmentally friendly oil-water separation materials is critical.
Cellulose is considered as a suitable material for oil-water separation due to its renewability and degradability (Huang et al. 2021;Xie et al. 2019;Xu et al. 2021). However, most cellulose-based membranes are superoleophobic (Li et al. 2018b) and can be easily polluted or even blocked by oily wastewater (Chu et al. 2015;Zhao et al. 2021). The oil fouling with high viscosity is difficult to wash off, shortening the service life of these materials (Lin et al. 2010). The application of the oil-water separation membrane with superhydrophilicity and underwater superoleophobicity properties has attracted more and more attention in recent years. These membranes can absorb water to form a hydration layer to separate oily wastewater (Li et al. 2018a;Yang et al. 2019). Meanwhile, the hydration layer on the membranes can prevent oil pollution to realize self-cleaning function.
As the first industry to systematically utilize cellulose, papermaking has the advantages of mature technology, stable equipment and strong production capacity (Gharehkhani et al. 2015;Hou et al. 2011). Thus, it is feasible to produce the membrane substrates on a large scale through papermaking (Li et al. 2021a(Li et al. , 2021bSehaqui et al. 2011). Pulp refining is a crucial step in the papermaking process that can result in a variety of changes in pulps, including internal fibrillation, external fibrillation, fiber shortening or cutting, and fines generation by peeling off cellulose fibers (Banavath et al. 2011). After pulp refining, the increased specific area of fibers and the strong bonding between fibers will considerably improve the performance of paper (Afra et al. 2013). Furthermore, the amount of hydroxyl groups on fibrillated fibers increases, causing fiber swelling to increase. At the same time, the pore in the paper structure is reduced, so as to have better barrier property (Xi et al. 2021b). Hence, the paper with refined pulps can absorb water immediately and form a hydration layer on its surface to repel oil, so as to endow the paper with the ability to separate the oily wastewater. However, it is necessary to find a way to stabilize the structure of paper in water, or the paper would be easily disintegrated because of the low wet strength.
Cellulosic self-reinforced material is a type of novel environmentally material (Nishino et al. 2004) that can be prepared using one of two methods: mixing cellulose fibers with cellulose solution or immersing cellulose fibers into the cellulose solvent to dissolve the cellulose fibers partially. The strength of cellulosic materials prepared by using the methods described above will be greatly enhanced (Piltonen et al. 2016;Shibata et al. 2013). Therefore, the strength of paper-based materials can be enhanced through the dissolution-regeneration.
Herein, cellulosic paper-based membranes for continuous oil-water separation were prepared by papermaking and in-situ gelation (Cellulosic materials were partially dissolved and regenerated in its original position). It is noteworthy that pulp refining held a remarkable impact on oil-water separation processes of the membranes. Because of the reduced pore size and enhanced water retention, a hydration layer was formed on the membranes to protect it from oil droplets. The gel structure formed by in-situ gelation and the multilayer cellulose framework formed by papermaking were interpenetrated to form a network, which enhanced the strength of membranes. In addition, the effects of different regeneration solutions on the properties of the membranes were investigated, including water, ethanol, sulphate solution and acetic acid.

Materials
Bleached softwood pulp was provided by Shun Pu Paper Co. Ltd (Zhejiang, China). Its polymerization (DP) was measured to be 846 by viscosity measurements. NaOH, toluene and xylene were supplied by Guangzhou Chemical Reagent Company (Guangzhou, China). Urea was supplied by Guangdong Guanghua Sci-Tech Co. LTD (Guangzhou, China). Acetic acid glacial, sodium sulfate, petroleum ether and liquid paraffin were supplied by Qiangsheng Functional Chemical Co. LTD (Jiangsu, China). N-hexane and Tween-80 were purchased from Kermio Chemical Reagent Co. LTD (Guangzhou, China). Pump oil is a commercial product.

Preparation of membranes
Softwood pulp was refined to 20°SR, 25°SR, 30°SR, 35°SR, 40°SR, 45°SR, 50°SR (°SR: The beating degree of pulps) using PFI refiner (Mark V1, HAMJERN MASKIN 621, Norway) according to ISO 5264-2:2002, and the unrefined pulp was 10°SR. The paper substrates were prepared by Paper sheet former (MESSMER 255, USA) with an average grammage of 80 g/m 2 according to TAPPI T 205 sp-12. The substrates were dried at 105 °C (10 min), and then stored at 25 °C and 50% RH for 24 h or more. These substrates were named Px (where x is the beating degree). The paper-based membrane were prepared by partial dissolving the paper substrates with NaOH / urea solution (w/w, 8: 12: 80) (Liquid absorption, 400 ± 10 g/m 2 ) at − 11 °C for 15 min and soaking them in regeneration solutions for 15 min. The excess reagent was removed by washing with water, and drying at 90 °C (0.1 MPa).
There were four different regeneration solutions adopted in this study, including water (Li et al. 2012), ethanol , 5 wt% Na 2 SO 4 solution (Mao et al. 2006) and 2/4/6 wt% acetic acid solution (Yang et al. 2019), and the membrane regenerated using these solutions were named as Px-y (where x is the beating degree of pulps, y is the order of the above solutions), respectively.

Oil-water separation process
The emulsion was prepared by homogenizing toluene and water at 1/99 (v/v) with 1 wt% tween-80 as surfactant-stabilized at 7000 rpm for 10 min (Ao et al. 2020). The separation process was driven by the gravity. The separation efficiency (R) was determined by measuring the volume concentration of toluene in the solution before and after separation by UV-vis spectra, and calculating according to the equation: where C p is the volume concentration of emulsion and C o is the volume concentration of filtrate. The flux (J) (L·m −2 ·h -1 ) was expressed by the equation below: where v (L) is the volume of filtrate, t (h) is the separating time and S (m 2 ) is the effective area of membranes.

Characterization
The scanning electron microscope (SEM; S-3700 N, Hitachi, L td. Tokyo, Japan) was used to carry out the SEM observations of membranes. . The crystalline structure of the membrane was investigated by wide angle X-ray diffraction (Bruker D8 Advance, Germany).
The pore diameter calculation. The pore diameter was tested by mercury porosimeter (AutoPore IV 9520). In our work, the highest intrusion pressure was about 207 MPa (30,000 psia), corresponding to pores of about 0.007 μm in diameter. The pore size was calculated according to the equation (Moura et al. 2005): where γ is the surface tension, γ = 0.52 N/m; is the contact angel, = 130°; d (nm) is the pore diameter; p (psia) is the applied pressure.

Morphology and properties of membranes
The membranes were prepared by papermaking and in-situ gelation. Figure 1 showed the route for the preparation of the membranes. Firstly, pulp refining was used to adjust the pore structure of paper substrates and improve the swelling of fibers, so that the paper could separate oily wastewater. The wet strength of membranes was then improved by in-situ gelation, making it suitable for the underwater separation procedure. The fibrillation of fibers during pulp refining lowered the porosity and pore diameters of paper substrates. Therefore, pulp refining could improve the barrier performance of paper. With the occurrence of gelatinization, the fibers in paper were covered by gel. And the gelation structure could endow the membranes with the underwater strength stability (Fig. S3). The SEM images (Fig. 2) showed the microstructural changes of paper substrates and membranes. With the gradual increase of beating degree, the structure of paper and membranes became compact (Fig. S1). The boundary between fibers in P35 (Fig. 2d) was clear in the surface and cross section, whereas the boundary between fibers in P35-2 was obscured (Fig. 2e). However, the fiber skeleton of P35-2 was still distinct. It indicated that the gelation occurred on the surface of fibers and did not disrupt the membrane's fiber skeleton. The membrane exhibited a multilayer structure in the cross-sectional direction, which was caused by papermaking. This structure was critical for the oil-water separation process.

The effect of pulp refining
Pulp refining could improve the connection between fibers in paper and membranes while also lowering porosity and pore diameters. As shown in Table 1, the average pore diameters and porosity of the paper and membranes decreased with the increase of beating degree. After gelation, the average pore diameters and porosity of the membranes composed of pulps with beating degree less than 35°SR showed a downward trend compared with the paper substrates. In contrast, the membranes composed of pulps with beating degree more than 35°SR showed a opposite trend due to the different fibrillating levels of pulps. The difference was caused by the different absorption degree of fibers to NaOH/urea solution at different beating degree and the changed pore structure at different beating degrees. Under the condition of low beating degrees, the bonding between fibers was weak, and the gel generated by gelation would partly fill the pores. Therefore, the porosity and average pore diameters of the paper-based membranes (P20-1 ~ P30-1) were reduced. On the contrary, the fibers in the paper could pack tightly at high beating degree, while the gelation would dissolve the fibrillated fibers partially and formed a gelled layer attached to the fibers, so that the porosity and average pore diameters of the paper-based membranes (P35-1 ~ P50-1) were increased after the gelation process ( Fig. 3a-b).
The wet strength of P10 was too low to support the oil-water separation and gelatinization. The oil-water separation performance of the paper substrates and membranes (Px-1) was shown in Fig. 3c-d. In oil-water separation, the reduced porosity and average pore diameters could improve the separation efficiency of the paper and membranes, but decrease the flux. When the beating degree of the paper and membranes was ≥ 35°SR, the increase of beating degree would only affect the flux, and the separation efficiency would remain > 98.5% (Fig. 3c-d). Therefore, the subsequent investigations of the paper-based membranes were based on 35°SR. The water flux of Px, Px-1, Px-2 was showed in Table S4.

Properties of different membranes
The pore structure of membranes The performance of the membrane was different due to the different regeneration rates of the gels in different solutions ). The ability of regeneration solutions to dissolve the urea complex  was crucial in regulating the pore structure of membranes during regeneration. The worse the solubility of urea, the faster the regeneration of gels, and the more tightly the bonding between gels and fibers. The tight bonding of the gels to the fiber skeleton helped to increase the porosity and the strength of the membrane. SEM was used to characterize the microstructure changes of different membranes. As illustrated in Fig. 4, the porosity of membranes from large to small was P35-2, P35-1, P35-3, P35-4, P35-5, P35-6. Among them, the porosity of P35-1, P35-2, P35-3 was higher than that of the paper substrates (P35) while the porosity of P35-4, P35-5, P35-6 was less than that of the paper substrates. The porosity of P35-y was consistent with the SEM characterization. The average pore diameters and porosity of P35-y were shown in Table 2.

Wetting behavior of membranes
The wettability of the membrane (P35-y) was shown in Fig. 4. The water wetting behavior of the membrane (Fig. 5b) was characterized in air, while the oil wetting behavior (Fig. 5a) was characterized under water. The wettability of the membrane was mainly determined by the hydrophilicity of fibers and gels. And the difference in the wettability of membranes was caused by the different pore structure. The dense pore structure of the membrane would prevent the intrusion of water droplets . Therefore, the denser the structure of the membrane, the worse its hydrophilicity. Nonetheless, all membranes still showed the superhydrophilicity (The water contact angels were nearly 0°) and underwater  The water contact angel of the membranes; c High-speed camera photo of water droplets on the membrane; d High-speed camera photo of oil droplets (1,2-dichloroethane) on the membrane superoleophobicity (The oil contact angels were all greater than 150°) (Ge et al. 2018). The contact processes between oil/water droplets and the surface of the membranes were photographed using a high-speed camera to further confirm the wettability of the membranes. As shown in Fig. 5c, the membrane could absorb water in 1.53 s after placing a drop of water on its surface in air and the water contact angel was nearly 0°, indicating the superhydrophilicity of the membrane. Meanwhile, when the membrane was under water (Fig. 5d), the membrane was hardly penetrated by oil droplets even if the oil droplets were driven down by the needle. And after the needle was lifted up, the oil droplets could be easily lifted without deformation, indicating the low oil adhesion of the membrane (He et al. 2015). Besides, the membranes would not be invaded by oil even in air after absorbing water (Fig. S4). This result suggested that the hydration layer on the membranes could effectively block the invasion of oil droplets.

Oil-in-water emulsion separation
The separation efficiency of the membranes (P35-y) was shown in Fig. 6a-c, and the paper substrate (P35) was the control group. At the first separation process (Fig. 6a), the separation efficiency of the membrane was > 98.5%. In addition, the separation efficiency of the membrane was > 98% after three separation cycles. This proved that the separation efficiency of the membranes was hardly affected by the kind of regenerated solution, and the kind of regeneration solution could only affect the flux of the membranes.
The stability was essential for the membrane to separate oil-in-water emulsion successfully. As shown in Fig. 6d-e, the stability of the membrane Fig. 6 a The first separation efficiency of the membranes; b The second separation efficiency of the membranes; c The third separation efficiency of the membranes; d The separation cycles of P35; e The separation cycles of P35-2; f The size of oil droplets in feed emulsion; g The photograph and optical microscope images of the feed emulsion and the filtrate was better than the paper. The separation efficiency of P35-2 could maintain > 95% after 20 separation cycles while the paper could only separate the emulsion 5 ~ 6 times. Meanwhile, the flux of P35-2 could remain within a certain range, while the flux of P35 changed greatly.
By comparing the emulsion and the filtrate under the polarizing microscope (Fig. 6b), it could be found that the feed emulsion was full of toluene droplets which disappeared in the filtrate. The purity of the filtrate also could be confirmed by UV-vis spectrum (Fig. S9). The average size of oil droplets in feed emulsion was 2.4 μm (Fig. 6f).

The strength of membranes
The strength of paper mainly comes from fiber strength, hydrogen bonding and entanglement between fibers. The increased fiber-to-fiber contact caused by fiber fibrillation has a positive effect on paper strength (Afra et al. 2013). The formation of gels increased entanglement and hydrogen bonding between fibers. However, the dry strength (Fig. 7a, c) of the membrane had no significant change owing to the decrease of fibrillated fibers and fine fibers in gelatinization. On the contrary, the wet strength of the membranes (P35-y) was remarkable improved owing to the decreased fibers swelling (Fig. 7b,  d). The decreased fiber swelling in the membranes was caused by the coverage of gels, which could be reflected by the water retention value (WRV) of the pulps (Table S2). Compared with paper substrates, the wet tensile index of membranes was improved about 15 to 18 times, and 4 times for the wet burst index. In order to compare the wet strength of the membranes and the paper intuitively, a 1 kg weight was placed on the wet paper and the wet membrane. As shown in Video S1, S2, the wet paper was crushed instantly, while the wet membrane was intact.
The FT-IR spectra of the membranes (P35-y) were shown in Fig. 7e, P35 was the control group. The peak around 1060 cm −1 was attributed to the skeleton vibration of C-O-C pyranose (Kafy et al. 2016). And the peak around 2900 cm −1 was the asymmetric stretching vibration of CH 2 (Xi et al. 2021a). The peak around 1645 cm −1 was caused by water absorption of the cellulose in the membranes . The board peak around 3388 cm −1 was due to the stretching vibration of hydroxyl groups on cellulose fibers (Fan et al. 2015). After gelatinization, the peak around 3388 cm −1 was higher, as a result of the cellulose macromolecule chain broken by NaOH / Fig. 7 a The tensile index of the membranes; b The wet tensile index of the membranes; c The burst index of the membranes; d The wet burst index of the membranes; e The FT-IR of the membranes; f The WAXD of the membranes urea solution. In addition, there were no other absorption peaks in the FT-IR spectra of the membranes. It suggested that there were no other chemical reactions occurring.
The X-ray diffraction profiles of the membranes at 15 min pretreatment were shown in Fig. 7f, and P35 was the control group. The absorbances of CNC-I occurred at 2 θ = 15°,16°and 22.5° (Duchemin et al. 2015). The peak around 16° was the peak in the amorphous regions, and the peak around 22.5° was the peak in the crystalline regions (Fan et al. 2017). The X-ray diffraction profiles showed that there were no obvious changes in the crystalline region of fibers in the membrane, and the gelation of fibers only occurred in the amorphous region due to the high dissolution concentration. All these indicated that there were no other chemical reactions except gelation took place in the membranes during the process, and the fiber skeleton structure of the membrane displayed no significant changes.

Separation mechanism
The hypothetical principles schematic was represented in Fig. 8. The paper substrates and membranes with refined pulps could quickly absorb and retain water to form a hydration layer (Fig. 8b-c). The hydration layer was a barrier of the membrane to against oil droplets. The polarizing microscope was used to observe the formation of the hydration layer on the surface of paper substrates and membranes. As Fig. 8a-c showed that it was impossible to form a hydration layer on the surface of P10, while the hydration layer could be observed on the surface of Px and Px-y composed of refined pulps (Fig. S5-7). The hydration layer was formed by fiber swelling. And fiber swelling could be reflected by the WRV of pulps. Therefore, the WRV of pulps was an important evidence of hydration layer on the paper substrates and membranes (Table S2).
The role of the hydration layer is speculated as follows: the oil droplets will collide with the hydration layer under gravity, causing the emulsion to partially break. The exposed oil droplets would be intercepted by the hydration layer (Koh et al. 2019). The water in the emulsion would be absorbed continuously by the membranes and separated under gravity, and a water replacement was formed in the membrane. With the accumulation and aggregation of oil droplets, the contact between oil droplets would be rise, and the emulsion would be break. The demulsified oil droplets would gather with the oil droplets in contact with them to form large oil droplets, and float in or on the water. Therefore, the paper substrates and membranes could block oil droplets although the average pore diameters of them was larger than that of the oil droplets.

Conclusions
In summary, we have fabricated a type of cellulosic paper-based membranes by papermaking and in-situ gelation. The role of hydration layer produced by fiber swelling in oil-water separation was explored. The membranes showed the superhydrophilicity and underwater superoleophobicity. Meanwhile, the remarkable increase in wet strength by in-situ gelation is in line with the expected application of the membranes. The flux of the membranes could be adjusted by pulp refining and regeneration process. The reduction of pore size and the improvement of water retention performance are considered to be the key. The separation efficiency for oil-in-water emulsion is > 98.5%. The prepared membranes are composed entirely of cellulose, making them convenient in usage and waste disposal. Besides, the preparation of the membrane adopts the papermaking process to provide a possibility for large-scale preparation in the future.