Qian Zhang,Lei Li,*,Lixia Cao,Yanxiang Li,Wangliang Li,*

1 CAS Key Laboratory of Green Process and Engineering,Institute of Process Engineering,Chinese Academy of Sciences,Beijing 100190,China

2 University of Chinese Academy of Sciences,Beijing 100049,China

3 Innovation Academy for Green Manufacture,Chinese Academy of Sciences,Beijing 100190,China

Keywords:Oil–water separation Fluorocarbon polymer Amphiphobic Coalescence

ABSTRACT Discharging untreated oily wastewater into the environment disrupts the ecological balance,which is a global problem that requires urgent solutions.Superhydrophilic and superoleophilic fibrous medium(FM)effectively separated oil–water emulsion as it was hydrophobic underwater.But its separation efficiencies(SEs)first increased to 98.9%,then dropped to 97.6%in 10 min because of oil-fouling.To tackle this problem,FM deposited with 0%–10%silica nanoparticle (NPsFMs),then coated by fluorocarbon polymer (X-[CH2CH2O]nCH2CH2O-Y-NH-COOCH2C4F9)(FCNPsFMs),was used to enhance its roughness and regulate its initial wettability to improve the anti-fouling property.FCFM and FCNPsFMs were hydrophobic and oleophobic in air and oleophobic underwater.Their water contact angles,oil contact angles and oil contact angles were 115.3°–121.1°,128.8°–136.5°,and 131.6°–136.7°,respectively,meeting the requirement of 90°–140°for coalescence separation.FCNPsFM-5 had the best separation performance with a constant value of 99.8%in 10 min,while that of FCNPsFM-10 slightly decreased to 99.5%.Theoretical released droplet(TRD)diameter,calculated by the square root of the product of pore radius and fiber diameter,was used for the evaluation of coalescence performance.Analyzed by two ideal models,TRD diameter and fiber diameter showed a parabola type relationship,proving that the separation efficiency was a collaborative work of wettability,pore size and fiber diameter.Also,it explained the SEs reduction from FCNPsFM-5 to FCNPsFM-10 was revelent to the three parameters.Moreover,FCNPsFMs effectively separated emulsions stabilized by cationic surfactant CTAB(SEs:97.3%–98.4%)and anionic surfactant SDBS(SEs:91.3%–93.4%).But they had an adverse effect on nonionic surfactant Tween-80 emulsion separation(SEs:94.0%–71.76%).Emulsions made by diverse oils can be effectively separated:octane(SEs:99.4%–100%),rapeseed oil(SEs:97.3%–98.8%),and diesel(SEs:95.2%–97.0%).These findings provide new insights for designing novel materials for oil–water separation by coalescence mechanism.

1.Introduction

Many industrial processes,such as textile,metal,petroleum chemicals and food,produce a large amount of oily wastewater and cause a great many harms to the ecosystem.Oily wastewater penetrates into the soil and forms an oil film,preventing the transportation of water and air and retarding the growth of plants.Untreated oily wastewater discharged into the ocean destroys the living environment of marine creatures and disrupts the ecological balance.Hence,the treatment of oily wastewater before releasing to the environment is a necessary process.Novel separation technologies for oil–water emulsion have been reported and developed by a great many researchers[1–6].

Sponges,foams and powders are used to absorbed oil as their huge specific surfaces and volume capacities,especially for the oil leakage in the ocean [7–9].Gravity separation and flotation can separate oil/water mixture,but not effective for oil–water emulsions because of its small size of a few to several tens of microns.The gravity separation can only affect droplets larger than 100 μm.Membranes,filters and meshes can effectively retain or coalesce emulsified droplets to achieve separation [7–10].Superhydrophobic and superoleophilic porous materials such as modified polytetrafluoroethylene(PTFE)and polydimethylsiloxane(PDMS)membranes were used as“oil-removing materials”for their different selectivity towards water and oil[11].The pore size of these materials was smaller than the emulsified droplet size to strengthen the retention effect[12–14].During this process,unavoidable oil adhesion and high viscosity may lead to the membrane fouling and destroy the system stability.Moreover,this fouling changed the intrinsic surface composition of the membrane and increased the hydraulic resistance,resulting in low permeability and separation efficiency[15].In our system,the emulsion was prepared by a homogenizer at a stirring speed of 25,000 r?min?1for 10 min and the diameters of emulsified droplets ranged from 2 μm to 3 μm[16].And,if the pore size of the membrane was less than 2～ 3 μm,the system is easy to reach an unfavorable high-pressure drop.To tackle this problem,a membrane with a pore size slightly larger than the emulsified droplets was fabricated,and the separation experiment was carried out by the coalescence mechanism[17,18].In the coalescence process,small droplets are captured and coalesced in the medium,then enlarged droplets are released with the flow and finally separated by gravity in the filtration.Therefore,membrane with anti-fouling property and slightly larger pore size may meet the requirement for high effective oil–water emulsion separation.

The surfaces of lotus leaf and fish scales are hydrophobic and oleophobic surfaces,respectively,which inspires people to mimic and fabricate the artificial surface to achieve self-cleaning and anti-fouling.Wenzel and Cassie-Baxter models proved that rough surface with a nano/microstructure can enhance its intrinsic wettability [19].Feng et al.proved that the PFDAE-g-ZnO-PTFE membrane showed antifouling performance as its hierarchical structural and amphiphobic surface[20].Wang et al.fabricated a robust water/oil-proof superamphiphobic nanofibrous membranes by the assembling of the fluorinate polyurethane(FPU)and incorporated SiO2nanoparticles[19].Also,Liang et al.grifted superhydrophilic silica nanoparticles on the low fouling resistance polyvinylidene fluoride(PVDF)ultrafiltration membranes to obtain hierarchical micro/nanostructure and to improve its hydrophobicity to resist oil-pollution [21].Inspired by these studies,an amphiphobic materials with roughness surface was designed for oil–water emulsion separation.

Due to the convenient preparation and low cost,the fiber medium was prepared as a substrate for further modification[22].The high surface energy of FM made it susceptible to contamination.In our previous study,the surface property of the glass fiber coated with fluorocarbon polymer(FC:X-[CH2CH2O]nCH2CH2O-Y-NH-COOCH2C4F9)was moderate hydrophobic with a contact angle of 114.1°.Moreover,FC showed a not easily wettable property towards different solvents.The contact angles of hexadecane,polyethylene glycol,diiodomethane and diesel on it were 106.2°,101.7°,95.9°,80.2°,respectively[23,24].Therefore,the wettability of FM can be regulated by FC coating to meet the contact angles requirement of 90°-140°for coalescence separation[25].Before coating,silica nanoparticles were deposited on the FM to enhance the roughness for better performance.The theoretical released coalesced droplet size,fiber diameter and pore size were applied for two ideal mathematical models to give a deep insight into the coalescence process[26].

Furthermore,in the oil-surfactant-water system,surfactants adsorbed on both oil–water interface and solid–liquid surface to improve the emulsion stability and change the initial surface properties,resulting in low selectivity,membrane fouling and separation efficiency reduction.Rosen explained the complicated behaviors of surfactants adsorbed on the solid–liquid surface and believed that the adsorption was decided by solid substrate,polar groups of surfactant active and pH and so on[27].In our research,CTAB,SDBS,Tween-80 as cationic,anionic and nonionic surfactants were added into the emulsion to evaluate the separation efficiencies of FCNPsFMs in the surfactant system.Hexadecane,Octane,rapeseed oil and diesel were used to evaluate the separation efficiencies for emulsion prepared by diverse oils.Deep insights for designing novel materials were given and oil–water separation by coalescence mechanism was discussed.

2.Materials and Method

2.1.Materials

Glass wool(diameter 1 μm,Shenyang Dongxiang Glass Fiber Co.,Ltd),glass fiber(diameter 7 μm,Taishan Fiberglass Inc.),and cellulose fiber(diameter 11 μm,Lenzing Fiber(Hong Kong)Ltd.)were used to fabricate fibrous fiber medium.Fluorochemical emulsion was designed and provided by a local company (Mw:9100 g?mol?1,Ms.:X-[CH2CH2O]nCH2CH2O-Y-NH-COOCH2CF2CF2CF2CF3,X,Y:undefined segment).Silica nanoparticle(diameter 10–20 nm)was purchased from Sigma.Isopropyl alcohol(IPA)was purchased from Aldrich.Hexadecane was purchased from Haltermann GmbH.CTAB(C19H42NBr)and SDBS(C18H29SO3Na)were purchased from Sinopharm Chemical Reagent Co.,Ltd.,Tween-80(C32H60O10) was purchased from Guangdong Guanghua Sci-Tech Co.,Ltd.All the chemicals were used without further treatment.

2.2.Preparation of fibrous filter medium

The preparation of the fibrous filter medium is described as our previous work[16].Glass fiber wool,glass fiber and cellulose fiber mixed with a mass ratio of 2∶1∶1 and a total mass of 3 g.All fibers dispersed in 400 ml pure water and a high-speed shearing machine was used to make the mixture a homogenous suspension.Fibrous dispersion was transferred into a handsheet model.Water was drained by a copper screen and the dewatered fibrous medium was obtained.Fibrous filter medium was transferred to a plate heater and dried at 60°C for 4 h.Then the fibrous filter medium(diameter 300 mm)was evenly cut into small round pieces with a diameter of 19 mm and a thickness of 0.5 mm.Then,small round pieces FM with the same pore size of (7.8 ± 0.2) μm were used as substrates for further modification.

2.3.Fluorochemical and silica nanoparticles coated fibrous medium

Fluorochemical emulsion,isopropyl alcohol and deionized water mixed with a weight ratio of 10∶5∶85.Per 10,50,100 μl 0.23 g 1 %(mass)SiO2dispersion was diluted to 100 ml by deionized water,respectively.As shown in Fig.1,the diluted solution went through the FM by a filtration device to obtain SiO2nanoparticles deposition fibrous mediums (recorded as NPsFM-1,NPsFM-5 and NPsFM-10).NPsFMs were dried at a 45 °C constant thermostatic oven for 3 h and dipped coating into the diluted FC emulsion for 5 min.Then,the FC coated NPsFMs were dried at a 45°C constant thermostatic oven for 3 h and solidified at a 135°C vacuum oven for 10 h to get mediums FCNPsFM-1,FCNPsFM-5 and FCNPsFM-10.FCFM was prepared by dipping coating FC emulsion,dried and solidified with the same method.Thus,FM,FCFM,FCNPsFM-1,FCNPsFM-5 and FCNPsFM-10 were prepared and stored for further research.

2.4.Characterization

The static water or oil contact angles of the mediums were measured by a contact angle goniometer at ambient temperature(DM-701 Kyowa Interface Science Co.,Ltd.).A 2 μl water or oil droplet was dispensed from the needle and settled on the fibrous surface.The photographs were recorded after 65 ms of droplet contacted with the medium surface and analyzed by the software.Deionized water (10 ml) was poured into a transparent quartz,fiber medium fixed on the upside of the stage,immersed into the water with slight pressure to drive out the air bubbles.A 2 μl oil droplet was dispensed towards the medium underwater and the photographs were recorded and analyzed.The pore size distributions of the fibrous medium were measured by a porometer through-pore size analyzer(Porometer3Gzh,Quantachrome Instruments,USA).The surface morphologies of the fibrous mediums were characterized by field emission scanning electron microscope(JSM-6700F)operated at 5 KV.

2.5.Oil–water emulsion separation experiments

As Fig.1 illustrated,1.3 ml hexadecane was added into 1.00 L deionized water and the mixture was emulsified by a 25,000 r?min?1homogenizer(Shinetek Instruments Co.,Ltd)for 5 min.The concentration of the prepared hexadecane-water emulsion was 4.4 × 10?3mol?L?1.The emulsion was continuously stirred at 500 r?min?1with a magnetic stirrer and pumped by a peristaltic pump with a flow velocity of 10 r?min?1(Volume:9.5×10?4m3?m?2?s?1,2.7×10?7m3?s?1).Fibrous medium was placed in a filter holder(effective area:283.5 mm2).Emulsion went through the filter and the filtration was collected by a tube every 2 min on a time scale of 10 min.2.5 ml sample was collected from the middle of the filtration to measure the chemical oxygen demand by a COD detector(Beijing Lianhua Science and Technology Development Co.,Ltd).The coalescence separation efficiency of the fibrous medium was calculated as:

Fig.1.Preparation of FCNPsFMs for oil–water separation.

where C0is the initial chemical oxygen demand (COD) of the hexadecane-water emulsion as a constant 3.1 g?L?1,and Cfis the COD of the filtration sample.0.01 g CTAB,SDBS,Tween-80 were added into 1 L hexadecane oil–water mixture(hexadecane concentration:4.4×10?3mol?L?1),respectively,and then the mixture was emulsified by the homogenizer under the same condition.The separation experiment was proceeded with the same method mentioned above.Octane,rapeseed oil and diesel were used to explore the separation effects of the modified mediums on diverse oils.Each experiment was repeated three times to obtain the average data.

3.Results and Discussion

3.1.Contact angles measurements

The wettability of the solid surface is the contribution of its chemical composition and surface topography.In our system,as Fig.2 described,the apparent contact angles of water and oil on FM were hardly observed at the initial 65 ms and recorded as a result of 0°.Therefore,FM composed of -OH group.SiO2and Al2O3had super-hydrophilicity and superoleophilicity.However,after modified with silica nanoparticles and coated with FC polymer,FCFM,FCNPsFM-1,FCNPsFM-5 and FCNPsFM-10 were hydrophobic and oleophobic as water contact angles were 128.8°,129.4°,133.7°,136.5°,respectively.Oil contact angles were 115.3°,119.3°,120.6°,121.1°,respectively.Contact angles increased with the content of silica nanoparticles was due to the improvement of the roughness.

Fig.2.The water contact angles,oil contact angles in air and oil contact angles underwater on FM,FCFM,FCNPsFM-1,FCNPsFM-5 and FCNPsFM-10.

As shown in Wenzel model,surface performs a“sticky”state when water directly contacts with the solid surface.While,as shown in Cassie-Baxter model,the surface is in a“slippy”state when the stage tilts as the air trapped into the grooves[28,29].However,for the porous material,water can penetrate into the pore with a convex interfacial curvature when the material was hydrophobic or with a concave interfacial curvature when the materials was hydrophilic [30,31].This phenomenon is driven by the capillary force which is hard to be eliminated.For the apparent contact angles,it is hard to conclude the specific contribution from the capillary depletion process or the surface composition.As our former work has mentioned,FC was coated on the glass slide and the contact angles of water,hexadecane,polyethylene glycol,diiodomethane and diesel were 114.1°,106.2°,101.7°,95.9°,80.2°,respectively[23,24].It is proved that the wettability of the modified film was due to the contribution of FC and SiO2nanoparticles.

Furthermore,Koji et al.had proved that,when the—CF2—unit on a side chain of FC polymer was 2–6,the fluoroalkyl chain was mobile as its no crystallization state.Under certain conditions such as the stage tilting,the reorientation of the mobile fluoroalkyl chain expose the carboxyl groups and change surface wettability[32].The contact angles and solid free energies of FC and PVDF on a glass slide were 114.1°,95°and 10.9,31.4 mN?m?1,which implied that PVDF was more hydrophilic than FC[23].But in the sliding angle test,the adhesive energies of FC and PVDF were 0.5 mJ?m?2and 0.4 mJ?m?2,which meant that removing water from the FC surface consumed more energy[23].Hence,as the—CF2— unit in FC was 4,we supposed that the no crystallization state fluoroalkyl group reoriented and exposed the carboxyl group to some extent under some special conditions.Then the hydrophobicity of FC was the contribution of —CF2CF2CF2CF3group and the oleophobicity came from—COO group.The resistance of the FC surface towards diverse solvents was favorable for anti-fouling.

In the aqueous phase,fibrous mediums were fixed on a stage upside and wetted by the water,the air bubbles were drained from the mediums by gently pressing[33].An oil droplet was dispensed by an upward bend needle and the contact angle underwater was measured by the same method.As shown in Fig.2,the OCA under water of FM,FCFM,FCNPsFM-1.FCNPsFM-5 and FCNPsFM-10 were 140.9°,136.7°,133.5°,131.6° and 135.2°,respectively,meeting the requirement of 90°-140°for coalescence separation.

3.2.Pore size

As shown in Fig.3(a)-(e),the SEM photographs illustrated the surface topography of the five mediums.Compared with FM,adjacent fibers in FCFM were connected by the solidified FC casting emulsion,reducing the pore size to some extent[Fig.3(a)-(b)].Silica nanoparticles were not homogeneous in the images as they were easily aggregated by electrostatic interaction in the casting solution.Aggregated silica nanoparticles attached on the fiber surface or filled the vacant space in the medium,enhanced the surface roughness and reduced the pore sizes,as Fig.3(c)-(e)described.By FC coating,SiO2nanoparticles were tightly sandwiched between fibers and FC surface,and not easily detached.The average pore sizes of FM,FCFM,FCNPsFM-1,FCNPsFM-5 and FCNPsFM-10 were 7.8 μm,6.4 μm,6.7 μm,4.7 μm and 3.5 μm,respectively.Correspondingly,their porosities were 90.4%,82.2%,74.6%,60.3%,44.9%,as illustrated inside Fig.3(f).The SEM photographs descriptions were consistent with the pore size distribution results in Table 1.In the meantime,the reduction of pore size caused a rise in pressure drop and induced an adverse effect on separation.The permission flux of the medium was an important parameter to evaluate the membrane performance[7,8].In our system,the flux of the medium was a constant 9.5×10?4m3?m?2?s?1and the face velocity was 9.42×10?4m?s?1.The pressure drops between the two sides of medium can be calculated by the classic fluid theory Hagen-Poiseuille equation[34,35]:

Table1 The pore sizes of the five fibrous mediums

where J was the flux of the medium,ε was the porosity,rpwas pore diameter,μ was the viscosity of the liquid and l was the thickness of the medium.In our system,μ was water viscosity 890 Pa?s,the thickness of the medium l was 0.5 mm and the effective area was 283.5 mm2.Compared with FM (55.6 Pa)in Table 2,the pressure drops of FCFM(90.9 Pa),FCNPsFM-1(90.9 Pa),FCNPsFM-5(229.7 Pa),and FCNPsFM-10(556.3 Pa)were 1.6,1.6,4.1 and 10.0 times.The rise in FCNPsFM-10 pressure drops was due to the increased SiO2concentration.

Table2 The pressure drops(Δp,Pa)of the five fibrous mediums

In conclusion,silica nanoparticles deposited on FM can enhance the surface roughness,providing more opportunities for droplets collision.But it induced adverse effects that porosities reduction and pressure drop rise.

3.3.Separation efficiency measurement

3.3.1.Hexadecane-water emulsion coalescence separation

Fig.3.The SEM photographs of FM(a),FCFM(b),FCNPsFM-1(c),FCNPsFM-5(d)and FCNPsFM-10(e);The pore size distribution of the five fiber mediums and porosities(f).

As shown in Fig.4a,FM or FCNPsFM was placed in a plastic filter,and hexadecane-water emulsion in a bottle was pumped by a peristaltic pump to go through the filter with a flux of 9.5 × 10?4m3?m?2?s?1(face velocity(u):9.42×10?4m?s?1).Sample from the middle of the filtration was collected for COD measurement.The separation efficiency can be calculated by the Eq.(1).The initial COD value of 4.4×10?3mol?L?1hexadecane-water emulsion C0was a constant 3.1 g?L?1and Cfwas the COD value of the filtration.The results were shown in the Fig.4(c),the separation efficiencies of FM first increased from 98.1% to 98.9% and then decreased to 97.6%overtime.This reduction may be due to the inherent oleophilicity of the FM leading to oil droplet adhesion.FCNPsFM-5 had the highest separation efficiency of 99.8%and maintained in 10 min.The separation efficiencies of FCFM and FCNPsFM-1 were almost identical:first increased from 99.2%to 99.6%and then without any variation.And the separation efficiencies of FCNPsFM-10 increased from 98.8% to 99.5%,which were lower than other FCNPsFMs.It indicated that FCFM and FCNPsFMs had excellent anti-fouling performance,and the separation efficiency was sensitive to the pore size variation.

3.3.2.Coalescence mechanisms for two models

Coalescence process in the fiber bed has been analyzed by a great many researchers [36–38].When emulsified droplets go through the fiber bed,coalescence occurs in three steps:approach,attachment and detachment process.According to Fig.4b,this process can be described as:Oil droplets anchored on the fiber surface and moved at a specific speed.Adjacent droplets collided and squeezed the oil–water interfaces to the ultimate limit state,causing the interfaces ruptured and droplets merged into a larger one.Finally,enlarged droplets (diameter >100 μm)was released by the fiber bed with the flow.Then they can be affected by the gravity in the filtration and floated on the top of the water phase,achieving the oil–water separation.The discussion was as follows:

Fig.4.The scheme of the coalescence separation process(a);droplets approach towards each other with certain velocities on a single fiber,coalesce or bounce off after they collide(b);The separation efficiencies of the five fibrous mediums for hexadecane-water emulsion(c).

In the approach process,capture efficiency was an important parameter to evaluate fiber bed properties.The capture efficiency was the sum of interception,diffusion,London von der waal's force,inertial impaction and sedimentation efficiencies.Details and calculation equations were given and discussed in our former paper [25].As the fibrous medium was made by a mixing of glass wool(df1=1 μm),glass fiber(df2=7 μm)and cellulose fiber(df3=11 μm),the calculation results for each fiber were given in Table 3.Inertial impaction and Sedimentation can be ignored as the Re was 2.503×10?3(<10)and the Sedimentation factor NGwas 8.387×10?6(<10?3).Therefore,the interception mechanism played a major role in the capture process,and proved that fiber with a smaller diameter had higher capture efficiency at a certain face velocity.

Table3 The contributions of interception,diffusion,London von der Waal's,inertial impaction and sedimentation efficiency to capture efficiency

In the attachment process,the wettability of the fiber and the existence of surfactants decided whether colliding droplets coalesce or bounce off.In the Fig.4(b),two droplets were captured by the fiber,approached towards each other with certain velocities,collided and deformed,then they merged into a larger drop or bounce off towards different orientations.Basu had proved that the contact angle (θ) of droplets on the fiber for effective coalescence was 90°–140°[17].Surface wettability played a vital role in this process as emulsified droplets intended to form an oil film changing surface properties when θ ≈0°and fail to attach on the fiber when θ ≈180°.In this work,oil contact angles underwater on FCFM and FCNPsFMs were in the range of 131.6°–136.7°,which were favorable for coalescence.

In the release process,the enlarged droplets detached from the bottom surface of the fiber bed when the hydrodynamic force overcame the adhesive force [38,39].Then the drag force on the droplet was equal to the restraining force of the interfacial tension[25]:

Fig.5.Two ideal models for Case I and Case II,FM(I1,II1),FCFM(I2,II2),FCNPsFM-1(I3,II3),FCNPsFM-5(I4,II4),FCNPsFM-10(I5,II5);The fitting curves of modified fiber diameter(df′)and theoretical released droplet diameter(dp)for Case I(a)and Case II(b).

Where,C was the drag coefficient;ɑ was orifice radius at the downstream side of the fibrous bed;γ was interfacial tension;Δu is the relative face velocity.And if the radius of the pore size and droplet was less than 0.1 mm,the equation can be written as:

According to the Stokes law and Newton's law,the drag force can be written as:

According to the Stokes law:

In our system,Re was 2.50×10?3,lower than 0.3.The formula was fit for the Stokes law and can be simplified as[40,41]:

In this article,1/2 average pore size was used as orifice radius ɑ to calculate;γ was the hexadecane-water interfacial tension 39 mN?m?1;dfwas fiber diameter(1 μm,7 μm,11 μm);η was the water viscosity 890 Pa?s;u was face velocity 9.42×10?4m?s?1.Calculated by Eq.(7),for the FM case,the theoretical released droplet diameter of 1 μm,7 μm,11 μm fibers were 213.3 μm,564.3 μm and 707.4 μm,respectively.All of these enlarged droplets can be separated by gravity as they were larger than 100 μm.Also,it indicated that fiber with larger diameter had a better coalescence effect.

For the modified mediums,the situation became more complex as the surface area for coalescence increased with SiO2deposition and FC coating.Since the release process was insensitive to the materials and several parameters in our system were constants,the revised theoretical released droplet diameter can be obtained by the square root of the product of the pore radius and the revised fiber diameter.Then two ideal models were proposed to give a deep insight into the relationship between the three parameters(dp,df,ɑ).In the models,we assumed that the medium was extremely homogenous and the coalescence area at every cross-section was the same.FM with an effective area of 283.5 mm2and porosity of 90.4%,then its coalescence area was 27.3 mm2.And we supposed that glass fiber wool,glass fiber and cellulose fiber with a ratio of 2∶1∶1 occupied the smallest area unit as Fig.5I-II depicted:Case I:all of the fibers stand up on this square;Case II:all of the fibers lie on the square.Circles in Case I represented the total bottom area of the fibers and the rectangles in Case II represented its total side-face areas of the fibers.In both models,the areas of circle and rectangle were identical as they simulated the same medium.For the FM in Case I,the sum area of the fiber bottom(Scoalesce)calculated by their radius was 135.1 μm2(ε:90.4%)and the smallest area unit was defined as a 1407.1 μm2square.The results of Scoalescefor each modified medium on the smallest area unit were calculated according to their porosities as given in Table 4.The Scoalesceincreased with the addition of FC and the concentration of SiO2.As depicted in Fig.5I1-5 and II1-5,revised fiber diameters(df′)were calculated as we imagined the four fiber as one.For FM in case II,we assumed that the four fibers lied on the square have the same initial length of 6.8 μm.As the concentration of FC and SiO2increased,the length and the diameter of the revised fiber increased with the same level.Revised L and df' in each Case II were given in Table 5.Calculated by Eq.(7),the revised TRD diameter in each case were achieved.As illustrated in Fig.5(b)and(c),the fitted curves showed the same phenomenon in both Case I and II,that is,a parabola type relationship between TRD diameter and fiber diameter.The fitted equation in Case I showed the most effective medium had the maximum TRD diameter,porosity,revised fiber diameter were 880.9 μm,70.2%and 23.1 μm,while the results in Case II were respectively 982.6 μm,64.1%and 25.4 μm.It proved that the separation efficiency in a fiber bed was a collaborative effect of the wettability,pore size and fiber diameter.Also,the SEs reduction from FCNPsFM-5 to FCNPsFM-10 was due to the overloaded SiO2nanoparticles.The low correlation coefficients were due to the real situation was more complex than the ideal models.In the real cases,FCNPsFM-5 with a porosity of 60.1%had the highest separation efficiency.This real situation was closer to the Case II model but with disordered stacked fibers.In summary,by rationally applying the three parameters(dp,df,ɑ),these models can help us design a more effective coalescer.

Table4 The porosities,Ssolid on medium area and orifice radius at the bottom of the fiber bed of the fiver fibrous mediums

Table5 The revised fiber diameter,theoretical released droplet size,curve fitting equation of model Case I and Case II

3.3.3.Coalescence experiments for oil-surfactant-water system

CTAB,SDBS and Tween-80 were added into the 4.4×10?3mol?L?1hexadecane-water emulsion,respectively,to perform the coalescence experiment.The results were shown in Fig.6(a)?(c).In Fig.6(a),FM had the worst performance with separation efficiencies were around 90.2%～92.3%.Cationic surfactant CTAB preferred to adsorb on the negatively charged fibrous medium surface,altering the surface wetting properties.The separation efficiencies of FCFM and FCNPsFM-1,FCNPsFM-5,FCNPsFM-10 were respectively 96.9%?97.5%,97.3%?97.8%,97.7%?98.2% and 94.5% ?95.4%,respectively.Anionic surfactant SDBS was prone to be rejected by the negatively charged FM surface.And oil droplet emulsified by SDBS may be separated by a retention mechanism.The separation efficiencies of FM,FCFM,FCNPsFM-1,FCNPsFM-5,FCNPsFM-10 were respectively 82.5%?93.8%,86.3%?89.6%,89.4%?91.7%,91.5%?94.6%,91.5% ?93.3%,respectively.Obviously,FCNPsFM-5 had the best performance for hexadecane-water emulsion with CTAB and SDBS.Their separation performance in the SDBS emulsion was worse than that in the CTAB emulsion.Nonionic surfactant Tween-80 can beintensively absorbed on the fibrous medium as we have observed by the QCM test[22].The steric hinder offered by its long alkyl chain was hard to overcome.Hence,the separation efficiencies of all the fibrous mediums showed a dramatic reduction from 95.3%to 74.1%?65.3%.

Fig.6.The separation efficiencies of the five fibrous mediums for hexadecane-water emulsion with 0.01 g?L?1 CTAB(a),SDBS(b),Tween-80(c);The separation efficiencies of fibrous mediums for octane(d),rapeseed oil(e),diesel(f)emulsions.

In summary,FCNPsFM has a great separation effect on hexadecane-CTAB-water emulsion and a slight weakening effect on hexadecane-SDBS-water emulsion.All of five fibrous mediums had a bad performance on hexadecane-Tween-80-water emulsion.

3.3.4.Coalescence experiment for diverse oils

Octane,rapeseed oil,and diesel were used to prepare emulsion,and the calescence separation experiments were performed by same method.In Fig.6(d),the separation efficiencies of the five fibrous mediums for octane-water emulsion were 99.2%?100%as octane had a shorter alkyl chain than hexadecane.The separation efficiency of rapeseed oil emulsion was smaller than that of octane but still above 96.1%.The separation efficiencies of FM and FCNPsFMs for diesel had an increasing trend from 91.4%,94.2%,94.5%,95.4%,96.4% to 95.1%,96.5%,96.3%,97.3%,96.5%,respectively.In the three cases,FM had the worst performance while FCNPsFM-5 had the best.Rapeseed oil and diesel were bio-oils containing surface actives that strengthened the emulsion stability and reduced the separation efficiencies.

4.Conclusions

FC coated fibrous medium with SiO2deposition had excellent performances for oil–water emulsion separation by coalescence.After modification,superhydrophilic and superoleophilic FM converted to hydrophobic and oleophobic in air,and oleophobic underwater.Also,oil contact angles(131.6°–136.7°)underwater met the requirement of 90°–140°for coalescence separation.Amphiphobicity of FCFM and FCNPsFMs enhanced the anti-fouling property and the separation performances.FCNPsFM-5 had the best separation efficiency 99.8% for hexadecane-water emulsion while that of FCNPsFM-10 had a slight reduction.Two ideal models described a parabola type relationship between TRD diameter and fiber diameter,and revealed that the separation performance was a collaborative work of wettability,pore size and fiber diameter.Also,FCNPsFMs had a good performance for emulsion with cationic surfactant(CTAB)and a slight weakening effect on emulsion with anionic surfactant(SDBS).For the five mediums,the separation efficiencies were relatively low for the emulsion with nonionic surfactant Tween-80 was due to its intensive adsorption and the strong steric hinder offered by its long alkyl chain.The modified mediums also showed excellent separation performance towards diverse oils.In conclusion,we had a depth insight into the coalescence process,and we hope our work will be helpful to design a more effective coalescer for oil–water emulsion.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work is supported by the National Key Research and Development Program of China under the contract number of 2017YFB0308000,Program of Innovation Academy for Green Manufacture,CAS(IAGM2020C04),the State Key Laboratory of Heavy Oil Processing(SKLOP201903001) and Key Research and Development Program of Hebei Province,China(20374001D).