Analysis of surface tension and conductivity
The surface tension and electrical conductivity of the spinning solution are the significant parameters that affect the micro-morphology of nanofibers. Fig. 1 displays the variation trend of the surface tension and electrical conductivity of the spinning solution at various CA and CS ratios. The conductivity of the electrospinning solution shows an ascending trend with the CS content increase. Specifically, the conductivity increased from 302 to 1370 µs•cm−1 with a CA to CS ratio varying from 10:0 to 0:10. This may be due to the chitosan dissolution in the acetic acid/water solvents forming ammonium salts, which act as polyelectrolytes, increasing the electrical conductivity of electrospinning solution (Qasim et al. 2018). The surface tension value just slightly dropped from 35.5 to 33.3 mN•m−1 as the CS content increased from 0–30% and then climbed back to 34.5 mN•m−1 with the further CS content increase to 90%. This demonstrated that the mixing of CA and CS with low concentration hardly affects the surface tension. Among the biocomposite membranes, the 50% CS group showed the optimum electrospinning performance. This is because, in addition to the interaction between CS and CA macromolecules, CA, CS, and PVA can form hydrogen bonds (Zhang et al. 2007) despite the solutions containing insufficient entanglement of molecular chains to achieve a critical level for producing fibers at low concentration (Du et al. 2007).
The morphology of CA/CS biocomposite nanofibers
The macro- and micro-morphologies and diameter distribution of CA/CS biocomposite membranes are shown in Fig. 2. The common composite nanofibrous membranes with CA and CS ratios appeared as porous web-like structures with average nanofibers diameters from 56.22 nm to 94.39 nm, which is smaller than the bio-based nanofibers in previous reports (Wang et al. 2017b; Yang et al. 2017; Wang et al. 2018a; Xu et al. 2020). The average diameters initially decreased and then increased when the CS content varied from 10–90%. The lowest average diameter and narrowest distribution for the conventional composite structure were 56.22 nm at 50% CS loading, attributed to enhanced electrostatic forces and reduced surface tension of the resultant solution (Stefanescu et al. 2012). The surface of 30% CS and 50% CS membranes is smooth and bead-free. The excess CS conversely increased the surface tension of the spinning solution, leading to increased nanofiber diameters. Some mixed thinner nanofibers in Fig. 2(D and E) may be caused by partial deformation and splitting of the charged droplets at the electrospinning process. The physical adsorption capacity was mainly determined by porosity and specific surface area. Compared to our previous CE/CS biocomposite nanofibers, the current CA/CS nanofibers have smaller diameters, resulting in a better physical adsorption for copper ions. Additionally, the core-shell biocomposite nanofibrous membrane with CS as the “shell” and CA as the “core” at 30% CS loading can be clearly observed in Fig. 2G, which showed the lowest average diameter (37.28 nm) but a broader diameter distribution (Fig. 2F) compared to the 50% CS membrane (Fig. 2C).
The three-dimensional (3D) micro-morphology of the CA/CS biocomposite membranes was further examined by AFM. The nanofiber structure of the membrane shows different diameter distribution and smoothness, consistent with SEM images. The surface roughness of the nanofibrous membranes were evaluated by analyzing the AFM images with the WSXM software. The detailed pore depth and surface are shown in Fig. 3. A 3D web structure with various interstices was formed by the interconnected ultrafine nanofibers. The 30% CS (306 nm) has a relatively higher Ra than the 90% CS group (195 nm) and the 30% core-shell group (212 nm) because of its larger diameter, more uniform distribution, and smoother nanofibers. A beads and nanofibers mixture was also observed in 90% CS group. Some inhomogeneous regions of the core-shell membrane and the incorporation of interpenetrating nanofibers may lead to extremely small pore sizes.
BET analysis
The adsorption-desorption curves of N2 as well as the distribution of pore width for CA/CS biocomposite membranes are displayed in Fig. 4. Fig. 4a shows that the curves display typical type-IV profiles with narrow hysteresis loops in the range of 0.8 to 0.99, especially for the core-shell and 50%CS groups, implying mesopores in the specimens in agreement with the pore width distribution in Fig. 4b. Table 1 lists the porosity parameters of the samples. The sample at 30% CS loading exhibits the lowest BET surface area (6.65 m2/g) but the largest average pore width (19.42 nm). Conversely, the samples with core-shell composite structure showed the highest surface area but smallest average pore width (11.91 nm). This analysis correlated well with the micro-morphology observation.
Table 1
Porous parameters of CA/CS biocomposite nanofibrous membranes
Samples
|
BET
surface area (m2/g)
|
Average pore
width (nm)
|
DFT cumulative
pore volume (cm3/g)
|
30%CS
|
6.65
|
19.42
|
0.021
|
50%CS
|
8.66
|
19.03
|
0.038
|
90%CS
|
6.66
|
16.99
|
0.023
|
Core-shell
|
12.33
|
11.91
|
0.036
|
Analysis of FTIR
Figure 5 shows the FTIR spectra of the nanofiber membranes with ordinary composite structure at different CA and CS ratios and the core-shell composite nanofibrous membranes before and after adsorption. The adsorption bands at 2936 cm−1 and 2896 cm−1 are associated with the stretching vibration of CH3 and CH2 in both CA and CS, respectively. The 1740 cm−1 band is due to the stretching vibration of C=O derived from the residual vinyl acetate unit in PVA, the acetyl group in CS, and the carbonyl group in CA (Phan et al. 2019). The adsorption bands at 1370 cm−1 and 1230 cm−1 are ascribed to the CH3 acetyl group (Wang et al. 2020). A shoulder peak at 1090 cm−1 can be explained as the stretching vibration of C-O in PVA (Koosha and Mirzadeh 2015). The bands at 1040 cm−1 and 1720 cm−1 are associated with the C-O-C and C=O stretching vibrations in CS. The band at 1620 cm−1 is related to the C-O stretching vibration due to the hydrogen bond between PVA and CS. There is a weaker N-H vibration at 1590 cm−1, assigned to the mixed vibration of CS amide II and NH2 (Li et al. 2012). The broad and strong peaks at 3360 cm−1 are -OH and -NH stretching vibrations. With the increase of chitosan content, the absorption band shifts from 3360 cm−1 to around 3650 cm−1, indicating that many hydrogen bonds were formed in the nanocomposite (Koosha and Mirzadeh 2015). The band at 1260 cm−1 is related to the C-N bond (Yezer and Demirkol 2020). The amino group characteristic band for the 30% CS and core-shell biocomposite membrane shifted from 1590 cm−1 to 1620 cm−1 and an additional peak occurrence at 1640 cm−1 after adsorbing copper ion mainly due to the amino group chelation in CS with copper ions (Chen et al. 2019). Thus, both 30% CS and core-shell membranes have superior chemical adsorption for copper ions than other groups.
Micro-morphology before and after adsorption
The SEM-EDS images of CA/CS biocomposite nanofibrous membranes after copper ions adsorption are presented in Fig. 6. As observed in Fig. 6(A1-A4), many copper particles are evenly adsorbed onto the pure CS, 30% CS, and core-shell nanofibrous membranes, which was evidenced in the EDS images (Fig. 6(C1-C4)). The pores and gaps with various sizes and irregular shapes contribute to the physical adsorption of copper ions. Fig. 6B1 shows that some “cloud-like” thin layers were formed on the 30% CS membrane, and the corresponding enlarged partial image shows that the surface is dense with a rough morphology. The chitosan nanofibers' excessive aggregation on the surface of the composite membrane after copper ions adsorption prevented the further entry of copper ions into the internal porous network of the membrane, leading to the lowest copper atomic content (0.49%). The many irregular pores on the surface of the 30% CS membrane in Fig. 6(A2 and B2) are attributed to hydrogen bonding and Van der Waals forces between CS and CA. It can be noticed from the magnified image that the nanofibers are interwoven and superimposed on each other. The stable porous nanofiber structure keeps the more chemically active sites for the removal of copper ions. The microscopic surface of the core-shell composite membrane is shown in Fig. 6(A3 and B3). Some lumps pieces are observed due to the existence of thin nanofibers and many cross-linking sites. After carbonization of 30% CS composite membrane in Fig. 6(A4 and B4), the surface fibers were twisted and irregular although the porous structure was not destroyed. The adsorption active sites with specific chemical groups were significantly reduced or disappeared during the carbonization process. Combined with the data in Table 2, the 30% CS composite membrane adsorbed the most copper content after adsorption corresponding to a relative atomic content of 1.85%. The relative atomic content of copper for core-shell composite membranes slightly decreased to 1.73%. The carbonized 30% CS composite film adsorption was only 0.11%, which can be explained by the NH2 or NH3+ groups' destruction and transformation into C-N groups during the carbonization process (Wang et al. 2021).
Table 2
Elemental content of copper after adsorption in each different proportion of composite membranes
Samples
|
Element
|
Weight(%)
|
Atomic(%)
|
CS
|
Cu k
|
2.20
|
0.49
|
30%CS
|
Cu k
|
7.93
|
1.85
|
Core-shell
|
Cu k
|
7.45
|
1.73
|
Carbonized 30%CS
|
Cu k
|
0.56
|
0.11
|
Adsorption performance and influencing factors
The adsorption performance of ultrafine CA/CS biocomposite nanofibrous membranes for copper ions (Cu2+) is shown in Fig. 7. The absorption rate of nanofibrous membranes continuously increased with time for all groups (Fig. 7A). The absorption rate shows a sharp increase in the initial stage (t<30min), which mainly attributed to the large specific surface area with many free adsorption sites on the composite nanofibrous membrane (Bock et al. 2012). The absorption rate slowed down and gradually reached a plateau, attributing to the Cu2+ concentration decrease in the solution and the consumption of adsorption sites. After equilibrium, the maximum and minimum adsorption rates corresponding to the 30% CS ordinary composite and core-shell composite membranes were 34.53 mg/g and 27.67 mg/g, respectively. Moreover, 52.8% of the total adsorption capacity was reached around 30 min for 30% CS nanofibrous membranes. The total adsorption equilibrium time is longer, and the Cu2+ diffusion into the composite nanofiber membrane may take more time.
The variation of adsorption capacity of 30% CS ordinary composite membrane at different initial concentrations is displayed in Fig. 7B. The adsorption performance decreases with increasing initial concentration. Especially, there is a significant decline in the concentration range of 20 to 30 mg/L, corresponding to the reduction of adsorption capacity by 47.8%. This is because the total amount of Cu2+ in the solution increases with the initial Cu2+ concentration, but the number of available active sites on the composite membrane is constant. In addition, the high initial concentration promotes the efficient adsorption of the composite membrane with more Cu2+ congregating, hindering the diffusion of Cu2+ into the composite membrane.
The effect of pH values on the 30% CS ordinary composite membrane adsorption performance is illustrated in Fig. 7C. The adsorption performance of the composite membrane increased from 1.16 mg/g to 86.41 mg/g corresponding to the pH varying from 1 to 5. The adsorption performance can be achieved to the optimum level when the pH is equal to 5. This can be explained as the amine groups in chitosan can protonate either NH3+ or (NH2–H3O)+ groups under various pH of the solution. The copper species primarily were unhydrolyzed Cu2+ with partial Cu(OH)+, and Cu(OH)20− in the pH range of 4–6. The pronated amine groups appear to be the predominant active sites for the copper ion adsorption. The coexistence of copper ions and hydrogen ions at low pH of 2-4 forms competitive adsorption, limiting the reaction of chitosan amino groups with Cu2+ (Yang et al. 2019). The amino group of chitosan formed an ammonium salt that is positively charged and repelled each other with the positively charged Cu2+ under acidic conditions, reducing its adsorption capacity (Bates et al. 2021). Conversely, the extent of protonation remarkably decreases as the pH is increased to 6 or higher, which leads to the decrease of adsorption capacity (Hasan et al. 2008).
XPS analysis
To reveal the mechanism of chemical adsorption on the CA/CS biocomposite nanofibrous membranes with conventional 30% CS and core-shell groups, XPS analysis was performed before and after adsorption, and its spectra are shown in Fig. 8. The surface atomic composition and states are listed in Table 3. However, there were no changes for the spectra of C1s and O1s in the binding energy position, indicating that the functional groups that contain carbon and oxygen hardly participate in copper ion binding reactions. The N1s spectra are seen in Fig. 8(A1 and A2, B1 and B2). Two peaks corresponding to the binding energy of 399.1 and 400.70 eV before copper ion adsorption was related with the –NH2 and the –NH– groups and their protonated –NH3+ groups in CS (Liu and Bai 2006; Yang et al. 2021). The two peaks show a significant shift to 399.7 and 401.7 eV after copper ion adsorption, respectively, agreeing with the previous report that chemical shifts are significant (Hasan et al. 2006). Therefore, it is verified that copper ions formed effective chemical coordination complexes with –NH2, –NH– and protonated –NH3+ groups in CS during adsorption probably due to surface bonding formation by sharing lone pair electrons in N atoms with the copper ions. This promoted the states of oxidation and resultant elevated binding energy of the N atoms (Liu and Bai 2006). The area and width of the peak for -NH or -NH2 after adsorption are larger and narrower than that before the adsorption, which is due to the formation of Cu[(-NH2)]2 through covalent bonds with [Cu-NH3+]2+ complex on the surface (Hasan et al. 2008), moreover, the more for core-shell samples compared to conventional samples. Nevertheless, the -NH3+ lower binding energy intensity for the core-shell membrane may be attributed to the shrinkage of the composite membrane surface with an -NH3+ enrichment, as seen in Fig. 6(B3), which leads to a space hindrance to react with the inner active sites, decreasing the adsorption capacity.
Table 3
The compositions of surface atomic elements and chemical groups of CA/CS biocomposite nanofibrous membranes
Composition
|
30%CS
|
30%CS
after Cu (II)-loading
|
Core-shell
|
Core-shell
after Cu (II)-loading
|
Binding energy(eV)
|
Binding energy(eV)
|
-NH- or -NH2
|
339.09
|
399.74
|
339.07
|
399.68
|
Atomic concentration (%)
|
0.67%
|
1.36%
|
0.56%
|
2.05%
|
-NH3+
|
400.70
|
401.75
|
400.70
|
401.68
|
Atomic concentration (%)
|
0.20%
|
0.13%
|
0.23%
|
0.24%
|
Total N1S
|
0.87%
|
1.49%
|
0.79%
|
2.29%
|
Cu 2p3/2
|
—
|
933.47
|
—
|
933.42
|
Atomic concentration (%)
|
0%
|
0.35%
|
0%
|
0.37%
|
Cu 2p1/2
|
—
|
953.27
|
—
|
953.27
|
Atomic concentration (%)
|
0%
|
0.1%
|
0%
|
0.1%
|
Total Cu 2p
|
0%
|
0.45%
|
0%
|
0.47%
|
The Cu2p spectrum is displayed in Fig. 6(C1 and C2). Two characteristic peaks corresponding to 932 eV and 952 eV are associated with Cu2p3/2 and Cu2p1/2, respectively, which proves the effective adsorption of copper ions (Hasan et al. 2006). Although the atomic concentration of Cu2p3/2 for core-shell membranes was higher than that of conventional composite membranes, the corresponding adsorption performance was limited by the agglomeration of chitosan nanofibers on the surface with the blockage of porous network structures as seen in Fig. 6(B3). The two additional fluctuations in the 940–946 eV and 961–965 eV regions can be explained as the unsaturated state of electronic configuration (Hasan et al. 2008). The copper ions may have a coordinated bonding with the amine and hydroxyl groups in chitosan, and the bonding ability varies with the pH of the solution, the concentration of metal ions, and the chitosan contents.
Table 4 compares the average diameter and adsorption capacity of copper ions in this study (ultrafine CA/CS nanofibers) with the previous reports using natural cellulose or chitosan to fabricate micro- or nano-fibrous membranes including cellulose acetate/chitosan (CA/CS) hollow microfibers, cellulose acetate/silicon dioxide (CA/SiO2) nanofibers, TEMPO-oxidized cellulose/chitosan/polyethylene oxide (TOC/CS/PEO) sandwich-like nanofibers, cellulose nanocrystals/chitosan (CNC/CS) nanofibers, chitosan/polyethylene oxide (CS/PEO) nanofibers, phosphorylated cellulose/chitosan/polyethylene oxide (PCF/CS/PEO) bi-layer nanofibers, cellulose acetate/chitosan (CA/CS) nanofibers, cellulose/chitosan (CE/CS) nanofibers as well as the porous CE/CS hydrogel beads. The results obtained from our current study show the significantly thinner nanofibers (37-94 nm) as well as the superior adsorption capacity of copper ions (86.4 mg/g) than those from any other groups. This can be attributed to the ultrafine nanofiber porous and stable structure as well as the ideal distribution of chemical adsorptive sites. Additionally, the toxic chemical reagents or solvents used in previous cellulose- or chitosan-based nanofibers listed in Table 4 like acetone, dimethylacetamide, 2,2,6,6-tetramethylpiperidinooxy, trifluoroacetic acid, dichloromethane, sulfuric acid and sodium hydroxide would lead to the secondary pollution of ecosystem. The current ultrafine CA/CS nanofibrous membrane also has a greener solvent (polyvinyl alcohol) and a more facile preparation way without post-treatment, showing a promising prospect for water treatment.
Table 4
The average diameter and adsorption capacity of copper ions for various bio-based materials
Cellulose or chitosan-based adsorbents
|
Chemical reagents
or solvents
|
Average diameter range (nm)
|
Adsorption capacity of copper ion (mg/g)
|
Reference
|
CA/CS hollow microfibers
|
Formic acid
|
600000-700000
|
4.1
|
(Liu et al. 2005)
|
CA/SiO2 nanofibers
|
Acetone, dimethylacetamide
|
> 500
|
23.0
|
(Gebru and Das 2017)
|
TOC/CS/PEO sandwich-like nanofibers
|
2,2,6,6-Tetramethylpiperidinooxy, acetic acid
|
159
|
36.8
|
(Bates et al. 2021)
|
CNC/CS nanofibers
|
Sulfuric acid, acetic acid, poly(vinyl alcohol)
|
262-498
|
45.0
|
(Wang et al. 2018a)
|
CE/CS hydrogel beads
|
Acetic acid, sodium hydroxide
|
—
|
53.2
|
(Li and Bai 2005)
|
PCF/CS/PEO bi-layer nanofibers
|
Urea, phosphate ester, acetic acid
|
372
|
74.5
|
(Brandes et al. 2020)
|
CA/CS nanofibers
|
Trifluoroacetic acid, dichloromethane
|
450-650
|
—
|
(Salihu et al. 2012)
|
CE/CS nanofibers
|
Trifluoroacetic acid, acetic acid
|
122-349
|
80.7
|
(Phan et al. 2018)
|
CA/CS ultrafine nanofibers
|
Acetic acid, poly(vinyl alcohol)
|
37-94
|
86.4
|
This study*
|