Introduction

The dye wastewater produced by the textile industry and dye houses is commonly toxic, poor biochemical, and harmful substances which seriously threaten human health and contaminate the aquatic system of the Earth [1,2,3]. Various techniques, such as precipitation, reverse osmosis, photo-degradation, coagulation, and adsorption have been devoted to remove the dyes from solutions system [4,5,6]. Among these approaches, adsorption has been regarded as a very promising technique to uptake dye molecules from wastewater owing to its economic, easy operability, high efficiency and environmental-friendly features [7]. A wide variety of adsorbents have been extensively applied and explored their potential in dye wastewater treatment [8,9,10,11]. Conventional powder adsorbents are difficult to separate and recover from aqueous solutions because of their small particle size. Once they are dispersed into the aquatic system, they may cause potential risk of health and environment. The nanofibers with high porosity and interconnected pore structures can be conveniently separated and collected from the solution system, and regarded as a new generation of adsorption materials for wastewater treatment [12,13,14,15].

A variety of nanofibers have already been investigated as the promising adsorbents and showed excellent adsorption performances because of their high specific surface area, excellent porosity, fine flexibility and good interconnectivity [16]. Compared to traditional nanofibers with smooth surface, constructing nanofibers with various hierarchical nanostructures, such as the porous or coarse structure will increase the specific surface area and activated sites of the nanofibers, provide more adsorption space and show great potential for the wastewater treatment [17,18,19].

Electrospinning is a reliable and universal way to prepare consecutive fibers containing organic and inorganic materials with ranging from micro to nanoscale. Naturally, most published studies about the preparation of porous nanofibers have been reported in this way [20]. Polyvinylidene fluoride nanofiber mat with porous structure was prepared via electrospinning, and titanium dioxide (TiO2) nanoparticles were embedded on the nanofibers following electrospraying by Gao et al. [21]. The decorating of the TiO2 nanoparticles improved the MB adsorption capacity from 46.87 to 85.70 mg/g in a short time, while the photocatalytic efficiency was enhanced and the energy requirement of photocatalysis was reduced. The composite nanofibers owned excellent reusability (> 89.5%) after the fourth recycling adsorption and photocatalytic experiments. Liao et al. [22] prepared poly(vinylidene fluoride) (PVDF) porous fibers via electrospinning method with poly(vinyl pyrrolidone) (PVP) as pore-forming agent which was then dealt with distilled water. The results showed that the fibrous membrane exhibited high porosity and super-wettability, and its super-lipophilicity and super hydrophilicity could be switched freely. The oil adsorption capacities of the fibers were varied in the range of 22.7 ~ 76.0 g/g, and oil and water fluxes were 54737.3 and 56869.9 L/(m2·h), respectively, in oil–water separation experiment. High surface area poly(lactic acid) (PLA)/tea polyphenols (TPs) porous composite nanofiber membranes (CNFMs) were prepared successfully by Wang et al. [23] via electrospinning. The CNFMs were utilized in the adsorption of silver ions. The results illustrated the nanofibers average diameter increased with the increase of TPs content and the porous structure could enhance the adsorption capacity of silver ions.

In recent years, solution-blowing as an innovative and simple technique to prepare nanofibers has attracted considerable attention due to its wide applicability in processing most polymers, simple operation, high productivity, and industrialized prospect [24,25,26,27,28]. Compared with the electrospinning technology, conductivity of spinning solutions and high-voltage electrostatic field support are not required in solution -blowing process, which represents the wider application range, the relatively higher safety and the lower device requirement. Generally speaking, to fabricate porous nanofibers via solution-blowing in one step may rely on rapid volatilization of low-boiling-point solvent and a certain relative humidity and temperature of the environment. The stretching of high-speed air flow will promote the volatilization of the solvent and the formation of hierarchical structure which means a great potential in preparing porous nano-mats. But as we know, few have reported the fabrication of porous nanofibers with ultrahigh specific surface area in one step by this way with no post-processing.

In the paper, PSF porous nanofibers were simply and effectively made in one step via solution-blowing a ternary system of PSF/DMAc/THF at ambient environment. The formation of porous structure is mainly owing to the solvent rapid evaporation pore-forming mechanism during the spinning process. The different ratios of the THF and DMAc in mixed spinning solutions were the key factor to affect the specific surface area (SSA) and pore sizes on the nanofibers. Then the PSF porous nanofibers were modified in two ways: (i) dopamine (DA) was self-polymerized in tris-buffer solution and immobilized on the PSF porous nanofibers surface at mild condition to form an ultrathin layer of polydopamine (PDA); (ii) polyaniline (PANI) was precipitated on the porous nanofibers using in situ oxidative polymerization of anilinium monomer. The PSF porous nanofibers and the modified nanofibers were determined by attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy, field emission scanning electron microscopy (FESEM), static water contact angle (WCA) measurement, and surface area and porosity analyzer, etc. Then the nanofibers were used as adsorbent to remove the methylene blue (MB) and methyl orange (MO) from wastewater. Adsorption mechanisms such as kinetics, isotherms, and different pH condition research for the removal of dye molecules were investigated in detail to explore their adsorption capacity and mechanism. Finally, the reusability of dye-adsorbed nanofibers was also evaluated by performing for four consecutive cycle’s adsorption/desorption experiments.

Experimental

Materials

Polysulfone (PSF, MW: ~ 35000 g/mol) was purchased from Sigma Aldrich. Aniline (99.5%), sodium hydroxide (NaOH, 98%), concentrated hydrochloric acid (HCl, 36.5%), methyl orange (MO), methylene blue (MB), N,N-dimethylacetamide (DMAc, 99%), and tetrahydrofuran (THF, 99%) were purchased from Shanghai Titan Scientific Co., Ltd. (China). Tris (hydroxymethyl) aminomethane (Tris), ammonium peroxydisulfate (APS, 98%), and dopamine hydrochloride (99%) were bought from Sinopharm Chemical Reagent Co., Ltd. (China).

Preparation of PSF porous nanofibers

The PSF power (25 wt%) was dissolved in THF and DMAc mixed solutions at different mass ratios of 10:90, 20:80, 30:70, 35:65, and 40:60, respectively. The PSF power was fully dissolved to form uniform mixed solution under mechanical stirring at ambient temperature for 24 h, then it was left in a vacuum oven at 30 °C for 24 h to eliminate the air bubbles.

The solution-blown spinning process was similar to the previous work and briefed in the following [29]: the extrusion speed of spinning solution was 4 ml/h, the inner diameter needle and outer diameter needle were 0.5 mm and 1.0 mm, respectively, the nitrogen pressure was about 0.1 MPa. The nanofibers were collected on the rotating cylindrical collector with the collecting distance of about 25 cm under irradiation of an incandescent lamp (100 W). The nanofibrous membranes were oven-dried at 80 °C for 12 h. The obtained PSF porous nanofibers were labelled as PSF10, PSF20, PSF30, PSF35, and PSF40, respectively, according to the mount of THF in mixed solvent. The formation process of the porous nanofibers in one step is shown in Fig. 1.

Fig. 1
figure 1

Fabrication route of the PSF porous nanofibers

Full size image

Preparation of PSF/PANI composite nanofibrous membranes

When the optimal addition content of THF was identified, the obtained PSF porous nanofibers were trimmed into specimen size of 3 cm × 3 cm and dipped into the mixed solution with 5 ml of aniline (An) monomer and 20 ml of 1 M hydrochloric acid and kept in an ice-water bath (0 ~ 5 ℃) for 2 h. Then 20 ml of 1 M hydrochloric acid with ammonium persulfate (APS, n(APS): n(An) = 1:1) was dropwise added into the aniline solution at 0 ~ 5 °C. The PANI was synthesized by typical oxidative polymerization and decorated on the surface of PSF porous nanofibers. The coating time was 15 min, 30 min, 60 min, and 120 min, respectively, and the PANI/PSF composite membranes were named as PANI15/PSFX, PANI30/PSFX, PANI60/PSFX, and PANI120/PSFX in turn (X represented the THF content). The obtained PANI/PSF composite membranes were doped with 1 M HCl for 2 h and oven-dried at 80 °C for 12 h.

Preparation of PDA/PSF composite nanofibrous membranes

A certain amount of dopamine powder was added into Tris buffer (pH 8.5, 10 mM) solution under mechanical stirring to dissolve completely with the concentration of 1 g/L. Then the PSF porous membranes (3 cm × 3 cm) were immersed into the mixed solution for 8 h at 25 ℃. The PDA/PSFX (X represented the THF content) composite membrane was cleaned alternately with ethanol and deionized water for 48 h to eliminate the residual impurities [30]. The composite nanofibers were vacuum oven-dried at 60 °C for 12 h.

Dye adsorption, desorption experiments

The effect of pH on dye adsorption

The PSF, PANI/PSFX, and PDA/PSFX nanofibrous membranes were used to explore the adsorption and desorption abilities of MB and MO which had been determined as the model dyes. 50 mg of the nanofibers were immersed in the 100-mL glass stopper flask with 50 ml of MB or MO solutions respectively (200 mg/l) under constant oscillation at 150 rpm and 30 °C. To identify the impact of pH value on the adsorption, the MO solution as well as MB solution of pH, i.e., 1, 4, 7, and 10 were prepared respectively to execute the experiments. After contacting for 24 h to obtain full adsorption equilibrium, the nanofibers were collected and the residual dye solutions absorbance was measured by UV spectrophotometer at maximum absorbing wavelengths of 664 nm (MB) and 464 nm (MO), respectively. The adsorption equilibrium quantity (qe, mg/g) and elimination efficiency (R, %) of dyes were obtained by Eqs. (1) and (2), respectively:

$${q}_{e}=\frac{{C}_{O}-{C}_{e}}{m}\times V$$
(1)
$$R=\frac{{C}_{o}-{C}_{\mathrm{e}}}{{C}_{o}}\times 100\%$$
(2)

where C0 and Ce (mg/l) represented the beginning and equilibrium time concentration of dye in the solutions respectively. V and m represented the solution volume (l) and the nanofibers weight (g), respectively.

Thermodynamic equilibrium sorption

In the isotherm experiments, 50 mg nanofibers and 50 ml different initial concentrations of MB or MO solutions were placed in a series glass-stopper flask of 100 ml by shaking at 100 rpm, and the adsorption tests were executed at the appropriate pH value and 30 °C for 24 h. The concentration of each dye was from 30.0 to 300.0 mg/l. After adsorption finished, the nanofibers were taken out from the dye solutions and the absorbance of MB and MO was tested by using UV spectrophotometer. The amount of MB and MO at adsorption equilibrium (qe) was calculated using the Eq. (1). The equilibrium adsorption results were analyzed via the linear forms of Langmuir and Frendlich equilibrium isotherm models, respectively.

Adsorption kinetics

The adsorption kinetic tests were carried out via adding 200 mg nanofibers to 200 ml MB or MO solutions (300 mg/l), respectively, at 30 ℃ under shaking of 150 rmp. After a fixed time interval, 1 ml dye solution samples were taken out and the amount of MB or MO adsorption at time t, qt (mg/g) was tested and computed by the Eq. (3):

$${q}_{t}=\frac{{C}_{0}-{C}_{t}}{\mathrm{W}}\times V$$
(3)

where C0 and Ct (mg/l) represented the beginning and anytime t concentrations of dye in the solutions respectively. V and m represented the solution volume (l) and the membranes weight (g), respectively. The adsorption results were analyzed using pseudo-first-order and pseudo-second-order model to explain the dye adsorption kinetic process.

Desorption and reusability experiments

The recyclability and stability of nanofibers were evaluated, as well. After the adsorption in dye solutions, the adsorbed PSF and PDA/PSF nanofibers were washed with 0.1 mol/L HCl solution for 24 h and then washed with deionized water for 24 h to rinse the residual dye molecules at room temperature. The PANI/PSF nanofibers were first immersed in 0.1 mol/L NaOH and washed for 24 h to remove the residual impurities and then washed with deionized water, oven-dried, and doped again by HCl (0.1 mol/L). The removal efficiency (R, %) was calculated by Eq. (2) as above. The adsorption-desorption process was repeated for 4 cycles through the similar procedure to determine the reusability potential of the nanofibers.

Characterization

The attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR, VERTEX 70, USA) was used to characterized the nanofibers structure with the range of 4000–600 cm−1.

The surface morphology of nanofibers was observed by field emission scanning electron microscopy (FESEM, ZEISS GeminiSEM 300 from German) at 5 kV acceleration voltage. Prior to imaging, the samples were dried by vacuum at 80 ℃ for 12 h and sprayed gold for 1 min.

The non-contact optical profilometry (Bruker Countor GT-K 3D, German) was applied to determine the roughness parameters (Ra) of samples, and at least 3 spots of each membrane were observed to obtain the average value.

The surface wettability of the nanofibers was executed via using a contact angle goniometer Kino SL200B (USA) by the sessile drop method at ambient temperature. The images were recorded within about 3 s and the mean static water contact angle (WCA) results were calculated by testing at least 10 spots to acquire.

The surface area and porosity analyzer (TriStar II 3020, USA) was used to characterize the surface area and pore width of the nanofibrous mats under nitrogen adsorption at 77 K. Before the BET test, at least 0.2 g of samples were cut into small pieces and degassed for 24 h at 80 ℃.

The UV absorbance of MB or MO solutions before and after adsorption were tested by a UV-9100 spectrophotometer (LabTech, Beijing) to calculate the dyes concentrations.

Results and discussion

Morphology of modified membranes

The PSF porous nanofibers with different additions of THF and their modified nanofibers are shown in Fig. 2. When the THF content was 0 wt%, the morphology of PSF nanofibers with average diameter of about 440 nm were uniform and smooth without pores existing (Fig. S1). The dense porous structure of the nanofibers was due to the rapid evaporation of THF during the spinning and drying process. As the evaporation rate of THF was more quickly than the rate air penetrated into the fibers, the fibers tended to collapse under atmospheric pressure leading to form the porous shape. As we knew, fibers with nanopores on the fibers surface and wrinkled surface could be produced via adding highly volatile solvents (THF, acetone, etc.) into spinning solvent in electrospinning owing to the thermodynamic instability in this process [31, 32]. Compared with electrospinning, stretching effect of the high speed nitrogen flow on spinning jet was also more beneficial to the rapid volatilization of the THF and the porous structure formation. The mechanism of pores formation in nanofibers was comparatively complex. In a word, the solvent evaporation outward the jet and the diffusion process of air penetration inward the jet simultaneously occurred along with the spinning process. This was accompanied by the flash volatilization of solvent on the surface of nanofibers and diffusion from core to surface, bringing the temperatures of polymer fluid jet quite different from the original solution, resulting in thermodynamically unstable jet, and the phase is separated into the polymer and solvent enrichment regions by biaxial and/or spinodal decomposition. Combined with the solvent evaporation and high-speed nitrogen stretching, the concentrated polymer-rich phase solidified into the fibers, while the solvent-rich phase eventually transformed into the meso- and nanopores. The vapor pressure of DMAc was much lower than that of THF at the same temperature. With the increase content of THF, faster volatilization of high vapor pressure solvent could preserve the initial polymer-rich phase and solvent-rich phase, resulting in the nanopores structure. On the contrary, lowering the vapor pressure of the mixed solvent allowed further phase separation to produce a larger pore structure with roughness surfaces. The vapor-induced phase separation was proved to be the primary reasons for the formation of pores in- and outside of the fibers in the paper [33,34,35,36].

Fig. 2
figure 2

SEM images of surface and cross-sections the nanofibers: a and a' PSF10, b and b' PSF20, c and c' PSF30, d and d' PSF35, e PDA/PSF30, f PANI30/PSF30

Full size image

The pores sizes increased with the raise of THF content in the mixed solvent and the morphology of the nanofibers became coarser which meant higher specific surface area and porosity. The specific surface area and pore volume of nanofibers determined by BET measurements would explain obvious effects of the different solvent component in detail later. The PSF nanofibers displayed obvious porous geometry with interconnected three dimensional pores structure with each other. Numerous nano-, meso- and micropores were shown in the surface and cross-sections images (Fig. 2a–d'). It was observed that a mixed system of a solvent could not completely dissolve the PSF powder when the content of THF reached to 40 wt% in the laboratory, so the maximum addition amount of THF was finally controlled at about 35 wt%. The average nanofiber diameter gradually increased with the DMAc content decreased which mainly due to the actual solute contents rising in the spinning solutions. The PSF porous nanofibers including beads or other defects could be observed with reducing the DMAc content which leading to inducing the instable fluid jets formation. Therefore taken together, the optimal THF content was final identified to be 30 wt%. The PSF30 porous nanofibers were selected as the samples for the subsequent PDA and PANI modification experiments. The surface morphology of PDA/PSF30 membrane is depicted in Fig. 2e. The membrane appearance gradually changed from white into dark brown after the PDA hydrophilic layer depositing. The morphology of modified nanofibers was observed to be nearly any apparent difference, while maintaining the nanofibers porous geometry in SEM image. This phenomenon could be mainly ascribed to the stable and ultrathin PAD layer on the nanofibers. The PDA chains could be immobilized onto the PSF molecules by thiols or amines via Michael addition or Schiff-based reactions to obviously enhance the hydrophilicity of the nanofibers. The image of PSF30 porous nanofibers with PANI coating for 30 min is shown in Fig. 2f. The porous structure had basically disappeared which mainly due to the immobilization of the PANI nanoparticles. The porous structure with higher specific surface area would be more favorable for the adhesion of PANI connected small particles during in situ polymerization. Noticeably, except for the disappearance of porous structure, there were conspicuous PANI nanoparticles which attached to the PSF30 nanofibers as shown in magnified figures. In other words, PANI chains had been fixed on the nanofibers with bead shape rather than layer by layer deposition. It was hypothesized that the binding force between PANI and PSF was mainly owing to the hydrogen bonding and conjugated effect interactions which illustrated the PSF nanofibers were a high-affinity substrate to the PANI chains [37]. Even in a short time of deposition for PANI, there would be excess aggregation nanoparticles coating on the surface of the nanofibers, the images of PANI/PSF30 composite nanofibers with different depositing times are shown in Fig. S1c ~ e. It could be observed that the aggregation phenomenon of PANI nanoparticles was more serious with the increasing coating time. While the self-polymerization time was 2 h, PANI nanoparticles even formed a nano-network structure on the fibers surface. This was an interesting phenomenon of its shaping mechanism and controlling the formation of net structure which was worth our attention in future. Naturally, the porous nanofibers could not retain the original morphology after PANI coating, but the composite nanofibers surfaces were still roughness due to the deposition of PANI nanorods.

Composition analysis of the modified membranes

The FTIR spectra of the PSF, PDA/PSF30, and PANI60/PSF30 nanofibrous mats have been described in Fig. 3. As it showed, the characteristic peaks of PSF30 porous nanofibers in the FTIR spectrum had been displayed at wavenumbers of 1150 cm−1 (S = O stretching vibration), 1246 cm−1 (C–O–C stretching vibration), 1585 cm−1 and 1487 cm−1 (benzene-ring (C = C) stretching vibration), and 2960 cm−1 (C-H stretching vibration). After the PDA was coated on the surface of the PSF porous nanofibers, the appearance of the membranes changed from white to black brown, but there were no new outstanding absorption peaks that could be discovered mainly owing to the ultrathin PDA layer was too weak to be characterized, such as 1620 cm−1 (C–C resonance vibration) and 1510 cm−1 (N–H bending vibration), only the intensity of the range of 3500–3300 cm−1 increased due to the -OH and N–H groups on PDA chains. When the PANI was deposited on the PSF30 nanofibers (PANI30/PSF30), the membranes appearance became dark and black. The intensity of peaks at about 1585 and 1487 cm−1 which corresponded to the quinone -(N = Q = N) and benzene-ring (N-B-N) stretching vibration of PANI had obviously enhanced. The peak at 1310 cm−1 was identified as the C-N stretching vibration in the benzenoid ring. The peak centered at 1148 cm−1 was associated with the aromatic C-H bending of the benzenoid ring and vibrations of the C-N–C group. The broad peak ranging from 3500 to 3200 cm−1 related to the asymmetrical and symmetrical N–H stretching vibrations. Combined with appearance variation of membranes and FTIR spectra results, the PDA and PANI were determined to be fixed on the surface of PSF30 porous nanofibers.

Fig. 3
figure 3

FTIR spectra of a PSF membrane, b PDA/PSF30 membrane, c PANI60/PSF30 membrane

Full size image

Analysis of membranes surface roughness

The optical profilometry images and average surface roughness (Ra) of the nanofibrous mats are exhibited in Fig. 4. When the THF addition was 0 wt%, the Ra of the PSF nanofibrous mat was about 4.55 μm due to the fluffy and roughness structure. The Ra of PSF30 mat slightly increased to about 6.02 μm owing to the three-dimensional interconnected and nanoscale hierarchical structure. Then after the PDA ultrathin layer was immobilized on the surface of PSF30 mat, the Ra was about 5.66 μm which demonstrated the modification impact on the variation of surface roughness could be neglected. Through in situ polymerization of PANI on the surface of PSF30 porous nanofibers, the Ra of PANI30/PSF30 membrane was about 5.74 μm. The pores and voids in- and outside of the nanofibers were completely filled with PANI nanocluster along with the increase of deposition time. The Ra value would decrease initially and then increase again with the continuous depositing of PANI nanocomposites on the nanofibers surface.

Fig. 4
figure 4

The optical profilometry images of membranes, a PSF, b PSF30, c PDA/PSF30, d PANI30/PSF30, and e the corresponding Ra of the membranes

Full size image

Surface hydrophilicity of the membranes

The static water contact angles (WCA) of the membranes are shown in Fig. 5. The WCA of PSF membrane was about 127.6° due to its hydrophobic nature and high surface energy. According to Wenzel’s theory, the hydrophobicity would be enhanced with the increase of roughness on the material’s surface while it was larger than 90°. So the WCA of PSF30 membrane reached to about 137.1° mainly due to the porous hierarchical structures which could increase the surface coarseness. When the PSF30 nanofibers were modified by PDA, the water droplet quickly permeated into the membrane surface within 3 s and the WCA sharply decreased to 0°. The PDA/PSF30 nanofibrous membrane was shifted to superhydrophilicity after modification. The superwettability of PDA/PSF30 membrane could be attributed to the existence of extremely hydrophilic constituents such as hydroxyl and amine groups on the PDA chains. After the PSF30 porous nanofibers were decorated by PANI of in suit polymerization doping at pH = 1 for 30 min, the WCA value was significantly decreased from 137.1° to about 58.5° which demonstrated the well hydrophilicity of the composite membrane. The nanoscale hierarchical structures and amino groups of PANI nanocomposites were crucial for the decreasing of WCA value and enhancement of wetting property. Improving the hydrophilicity could effectively reduce the organic pollutants adhesion force, which would be beneficial for their recyclability and stability in wastewater treatment.

Fig. 5
figure 5

The static water contact angle of the membranes

Full size image

Quantitative analysis of porous structure

The N2 adsorption/desorption isotherms of PSF, PSF30, PDA/PSF30, and PANI30/PSF30 membranes are shown in Fig. 6. As we known, the porous structure can be classified as microporous, mesoporous, and macroporous materials due to the different pore sizes. The results of the four membranes corresponded to a typical IV pore model by the BJH method. The hysteresis loops at relative pressure (P/Po) between 0.6 and 1.0 were exhibited in the isotherm of membranes mainly due to their mesopores (2–50 nm pore diameter) feature. The specific surface area (SSA) of PSF nanofibers was determined to be 11.14 m2/g with average pore width of 14.89 nm. Then the SSA substantially increased to about 30.91 m2/g of PSF30 porous nanofibers owing to the interconnected hierarchical porous structure. After modification, the SSA of PDA/PSF30 and PANI30/PSF30 membranes slightly decreased to about 22.94 m2/g and 23.41 m2/g, respectively. We considered the SSA values would further reduce due to the disappearance of porous structure and aggregation of the larger particles on nanofibers surface. The SSA and the average pore diameter of the membranes were depicted in Table 1 which demonstrated that they were mesoporous materials. These results showed that the porous nanofibers would own an excellent adsorption efficiency potential in dye wastewater treatment.

Fig. 6
figure 6

N2 adsorption–desorption isotherm curves of the membranes: a PSF, b PSF30, c PDA/PSF30, d PANI30/PSF30

Full size image
Table 1 The specific surface areas (SSA) and average pore diameter of the membranes
Full size table

Dye adsorption

Effect of pH

The pH of the dye solution was a significant factor during the adsorption process, because it could determine the surface charges of the membranes and the variation of the functional groups. The adsorption capacity with different values of pH, i.e., 1, 4, 7, and 10 were tested and the results are illustrated in Fig. 7. The adsorption property of the MB dyes was displayed in Fig. 7a. The adsorption amount of MB onto PSF nanofibers was about 55.23 mg/g at pH 1 and it slightly increased with increasing pH values, such as 71.92 mg/g at pH 10 which mainly due to the negatively charges of PSF nanofibers while at pH above its isoelectric point. The adsorption amount was no obvious significant increased when the pH value was greater than or equal to 7. This could be assigned to the high SSA of nanofibers, the electrostatic interaction at high pH value and π-π conjugated synergistic effect between the PSF chains and MB molecules. The adsorption performances of PSF30 porous nanofibers was noticeably higher than that of PSF nanofibers primarily owing to the larger surface area of 30.91 m2/g and interconnected nano- and mesopores structure. With increasing pH values, the removal amount of MB gradually enhanced which was similar with adsorbed variation of PSF membrane. The PSF30 membrane at pH 10 showed a dye adsorption capacity of about 124.31 mg/g which was obviously higher than about 72.24 mg/g for PSF smooth nanofibers at the same pH condition. After the depositing of the PDA ultrathin hydrophilic layer, the porous composite membrane exhibited the highest adsorption efficiencies at pH 10 of about 172.87 mg/g which was the optimal abilities among the membranes. The isoelectric point of PDA was approximately 4.0 which indicated that at pH above it the PDA/PSF30 membrane was carrying positive charges [38]. Except for the electrostatic attraction, hierarchical structure and strong π-π stacking effects, high number of hydrogen bonding interactions between PDA and MB molecules were another crucial factor in dye removal efficiency. The results showed the higher surface area and larger negative charges sites on nanofibers could prove a comparatively better adsorbed tendency towards the cationic MB dye. While the PANI nanoparticles were immobilized on the porous substrates through in situ polymerization, the PANI30/PSF30 membrane was selected as a model sample and applied to the MB adsorption at varying values of pH. The high positively charged sites on the PANI surface doped at hydrochloric acid solution was not suitable to remove cationic dyes due to electrostatic repulsion. These results illustrated the importance of pH values to influence the efficiency of dye adsorption. The MO adsorption curves are shown in Fig. 7b. Except for the hydrogen bonding interaction and high SSA characteristic, MO molecules was an anionic dye with negative charged on the surface owing to the anionic sulfonate groups. The positive and negative surface charges variation which was remarkably associated with the pH values of dye solutions would be another important factor to influence the adsorption performance. It was worth noting that the adsorption amounts of MO decreased with the rising of pH values. This could be mainly due to the electrostatic repulsive forces between the negatively charged of MO molecules and the same charged adsorbent at pH above the isoelectric point. The PANI30/PSF30 membrane doped at pH 1 exhibited the best adsorption performances of 160.40 mg/g onto MO dye. To our knowledge, PANI existed in the form of emeraldine salt (ES) with a highly positive charged surface under acidic conditions. Therefore, the anionic MO dye could be more easily adsorbed on the PANI30/PSF30 nanofibrous membranes at low pH values through the electrostatic attraction effect. The active sites with positively charged on PANI chains could interact with the sulfonate groups of MO and thus enhanced the dye adsorption ability of the mats. With the increase of pH value, the ES form would transform into the emeraldine base (EB) form with no positive charge and result in the decreasing of the MO adsorption capacity, such as 105.08 mg/g at pH 10. The proposed adsorption mechanism of MO onto PANI30/PSF30 nanofibers is illustrated in Fig. S2 which revealed that there existed sorts of interaction between them in solutions. The other three membranes also showed the similar attenuation adsorption tendency onto MO with the increasing of pH values particularly in alkaline solutions. On the whole, the MB absorbed by PDA/PSF30 nanofibers at pH 10 and the MO absorbed by PANI30/PSF30 nanofibers at pH 1 exhibited the highest adsorption efficiencies respectively and were chosen to investigate the adsorption mechanism in the following.

Fig. 7
figure 7

The influence of pH value for the adsorption capacity of dyes a MB, b MO

Full size image

Adsorption kinetics

Generally, the research of adsorption kinetics study represented the rate of adsorption and nature of interaction for adsorption process. In this paper, pseudo-first-order and pseudo-second-order models were selected to investigate the adsorption kinetics of porous nanofibers. Pseudo-first-order model demonstrated that the rate of adsorption process was in direct ratio with the number of vacant sites, while pseudo-second-order model illustrated that the adsorption rate was in direct ratio with the square of vacant sites [39].

Adsorption kinetics of MB onto PDA/PSF30 nanofibers and MO onto PANI30/PSF30 nanofibers was investigated at 200 mg/l initial concentrations under 30 ℃ with increasing adsorption time until reaching to equilibrium state and shown in Fig. 8. The experimental kinetic results for dyes adsorption onto porous nanofibers were summarized in Table 2.

Fig. 8
figure 8

Adsorption kinetics onto the membranes of a PDA/PSF30 (MB, 200 mg/l, pH = 1) and b PANI30/PSF30 (MO, 200 mg/l, pH = 10) at 30 ℃

Full size image
Table 2 Parameters obtained from the pseudo-first order and pseudo-second-order kinetics for the adsorption of MB and MO at 30 ℃
Full size table

The kinetics adsorption of dyes on nanofibers was fitted by using Lagergren’s pseudo-first-order and Ho’s pseudo-second-order model. Nonlinear equation of the pseudo-first order model was exhibited as Eq. (4) [40]:

$${\mathrm{q}}_{t}={q}_{e}\left(1-{\mathrm{exp}}^{-{k}_{1}t}\right)$$
(4)

where k1 (1/h) represented the rate constant, the adsorption capacity of nanofibers at time of t and equilibrium time (h) were qt and qe (mg/g), respectively.

Nonlinear equation of the pseudo-second-order model can be expressed as Eq. (5):

$${\mathrm{q}}_{t}=\frac{{q}_{e}^{2}{k}_{2}t}{1+{q}_{e}{k}_{2}t}$$
(5)

where k2 demonstrated the rate constant of Ho’s pseudo-second-order equation (g/mg·h).

The MB adsorption increased rapidly during the initial 4 h, which was owing to the abundant number of the meso- and macropores on the porous nanofibers. Then it gradually levelled off over 10 h due to the limited effective adsorption sites and achieved complete equilibrium. As shown in Table 2, the pseudo-second-order kinetics model exhibited the higher correlation coefficient (R2) of 0.9905 as compared to pseudo-first kinetics models at MB adsorption. Good fitting relationship and very close calculated and experimental values of qe indicated that the MB adsorption onto PDA/PSF30 porous nanofibers was mainly dominated by pseudo-second-order model and chemisorption type in nature. The adsorption features of MO onto PANI30/PSF30 nanofibers displayed the similar curves to the adsorption of MB and could also be well fitted by the pseudo-second-order model in Fig. 8b and Table 2. The MB and MO amounts adsorbed on PSF30 porous nanofibers were 124.90 and 128.22 mg/g respectively and listed in Table S1. The kinetics adsorption of MO and MB onto PSF30 porous nanofibers could also be governed very well by the pseudo-second-order model in Figure S3 and Table S1 which suggest that the adsorption process abided by a chemical process in solid-liquid interface.

Thermodynamic investigations

The adsorption isotherm studies were executed at 30 ℃ to estimate the thermodynamic characteristics of the MB removal via the PDA/PSF30 nanofibers and MO removal via the PANI30/PSF30 nanofibers, the experimental data were evaluated via the Langmuir and Freundlich isotherm models. The result of fitting plots is shown in Fig. 9. The calculated parameters, such as qm and KL from Langmuir isotherm, 1/n and KF from Freundlich isotherm and their corresponding regression coefficients (R2) were shown in Table 3. Nonlinear isotherm equations of the Langmuir (6) and Freundlich (7) could be expressed as follows:

$${\mathrm{q}}_{e}=\frac{{q}_{m}{K}_{L}{C}_{e}}{1+{K}_{L}{C}_{e}}$$
(6)
$${\mathrm{q}}_{e}={K}_{F}{C}_{e}^{1/n}$$
(7)

where Ce represented the equilibrium concentration of dyes in aqueous solution (mg/L), qe and qm (mg/g) were designated as the equilibrium adsorption quantity and theoretical maximum adsorption capacity (mg/g), respectively, and KL (l/mg) was the constant of Langmuir equation, corresponded to the heterogeneity factor, and KF (mg/g) was adsorption capacity when the equilibrium concentration of adsorbate equaled to 1. The removal process of MB or MO onto PSF30 porous nanofibers was well coincidence with the Langmuir model rather than Freundlich model in Fig. S4 and Table S2. R2 values were of 0.9942 and 0.9947 for MB and MO at C0 of 200 mg/L, respectively. Comparing with the correlation coefficients (R2), the Langmuir model with R2 of 0.9912 fitted experimental equilibrium adsorption data better than the Freundlich model with R2 of 0.9157 revealed that MB adsorption by PDA/PSF30 porous nanofibers confirmed monolayer type adsorption process and the uniform binding energy present (uniform distribution adsorption sites) on nanofibers surface. The 1/n value of Freundlich model was 0.2842 < 1, which also confirmed the favorability of normal Langmuir isotherm. Moreover, the KL value of 0.1257 represented the strong affinity between the adsorbent and MB molecules via strong electrostatic interaction. The anionic MO adsorption onto the PANI30/PSF30 nanofibers could also be better described by Langmuir isotherm model (R2 0.9864) comparing with Freundlich isotherm model (R2 0.94519). Except for the electrostatic attraction effect, the functional groups of PANI nanoparticles doped at pH 1 could provide more active sites to adsorb the MO molecules. Moreover, based on the Langmuir isotherm, the PDA/PSF30 and PANI30/PSF30 nanofibers had maximum adsorption capacities (qmax) of 190.51 and 196.48 mg/g, respectively (Table 2) which suggesting the porous nanofibers with PDA or PANI modified were good candidates for dye removal from aqueous solutions.

Fig. 9
figure 9

Equilibrium state adsorption of a MB onto PDA/PSF30 at pH 10 and b MO onto PANI30/PSF30 membranes at pH 1

Full size image
Table 3 Isotherm parameters for MB adsorption onto PDA/PSF30 and MO adsorption onto PANI30/PSF30 membranes
Full size table

Reusability

The structural stability and reusability of adsorbents are of crucial importance for practical applications [41]. The recycling ability of adsorption–desorption process was carried out for four consecutive cycles and the results are shown in Fig. 10. The dye adsorbed PDA/PSF30 nanofibers were washed in the desorbent of 0.1 M HCl solution, the adsorbed cationic MB molecules would be replaced with H+ moieties. While the anionic MO species adsorbed on the PANI30/PSF30 nanofibers were also be removed with OH− moieties in 0.1 M NaOH solution. Then it was doped again in 0.1 M HCl solution to recover the positive active sites for recycling adsorption. It was worth noting that the efficiencies of adsorption and the actively dye adsorbing cites of nanofibers reduced with multiple cycles, nevertheless, the dye equilibrium adsorption quantity (mg/g) could still maintained above 70% in the 4th cycle, respectively. Comparing with the poor reusability of hydrophobic PSF30 porous nanofibers in Fig. S5, maybe the improving hydrophilicity and hierarchical structure of composite nanofibers were the mainly factor to enhance its antifouling and regeneration ability. The above experiments indicated that the adsorbents of porous nanofibers with PDA or PANI layers coating could be considered as a effective and economic technique in water purification industry.

Fig. 10
figure 10

Four cycles of a PDA/PSF30 nanofibers in the removal of MB at pH 10 and b PANI30/PSF30 nanofibers in the removal of MO at pH 1

Full size image

Conclusion

In this paper, PSF porous nanofibers were fabricated in one step via solution-blowing technique mainly due to the rapid volatilization of low boiling point solvent (THF). The optimal THF addition was identified to be 30 wt% in mixed solvent which the obtained porous nanofiber was named as PSF30 nanofibers. Then the PANI and PDA layers were immobilized on PSF30 porous nanofibers surface respectively by in situ polymerization to prepare composite nanofibers. The PDA/PSF30 nanofibers were still porous morphology attributed to the PDA ultrathin layer while the porous structure of PANI/PSF30 nanofibers disappeared due to the serious aggregation phenomenon of PANI nanoparticles. The Ra of PSF30 membrane increased to about 6.02 μm and the wettability improved obviously after decoration. During the MB and MO adsorption process, in addition to high specific surface area, hydrogen bonding and π-π stacking effect, electrostatic interaction was also an important factor for dye adsorption. The adsorption mechanism showed the removal process of MB and MO was well described with pseudo-second-order kinetic model. The results of equilibrium adsorption for MB and MO onto nanofibers were both well coincide with the Langmuir isotherm model than the Freundlich model. The composite nanofibers had a good reusability and stability after 4 cycles of adsorption-desorption process. It was indicated that the composite nanofibrous membranes with rough hierarchical structure had high potential to be applied in practical effluent treatment.