3.1. Characterization of porous CD and derivatives
Figure 3 shows the SEM micrographs of the surface of composites under the magnification. It was observed that the composites exhibit rough surface and porosity. The surfaces of the four adsorbents were dense and porous with a large number of channels, which is conducive to the interaction between the adsorbent and organic pollutants to improve the adsorption efficiency. Figure 3(d) further reveals that, through the two-step cross-linking reaction, X-CDP obtains a larger pore size and a better pore structure than the other three adsorbents, which makes the material have better adsorption performance.
The TGA curves for the pyrolysis of β-CD polymer were shown in Fig. 4. The mass loss of 3% − 5% at 150°C was assigned to the evaporation of the water. At 150–260°C, the TGA curve kept steady, indicating that the β-CD polymer was thermostable at this temperature range. At the temperature above 260°C, there was a mass loss of 30%-40%, ascribed to the decomposition of β-CD monomer and TFTPN. Compared with the other three adsorbents, X-CDP lost the least quality at this stage. The re-crosslinking of TFTPN enhanced the thermal stability of the material. At temperature higher than 370°C, there was still a mass loss of 15–20%, which was caused by the decomposition of the carbon chain. These indicated that the as-synthesized β-CD polymer was thermostable enough for organic pollutants adsorption.
The specific surface area of several different composite materials was measured. SBET of Cl-CDP, NO2-CDP, F-CDP and X-CDP were 16.863 m2·g− 1, 2.125 m2·g− 1, 16.532 m2·g− 1 and 200.974 m2·g− 1, respectively. The obvious increase in the specific surface area of X-CDP indicates that X-CDP is not simply obtained by physical mixing of the first three polymers. During the preparation of X-CDP, the large pores of the polymer were broken, and more small pores appeared. The higher specific surface area makes X-CDP have better adsorption capacity for micro-pollutants.
The FT-IR spectra of TFTPN, β-CD, Cl-CDP, NO2-CDP, F-CDP and X-CDP are shown in Fig. 4b. For there CDPs, the broad bands and peaks show that the composites contain functional groups of isocyanate and β-CD, which is similar to the previous studies (Li et al. 2018). Particularly, the absorption peak at 3423 cm− 1 was ascribed to the –OH groups in β-CD. The absorption at about 1157 cm− 1 was from C-O bond stretching in the C-O-H group, and the peaks at 1105 and 1035 cm− 1 were ascribed to C-O bond stretching in the C-O-C group of the anhydroglucose ring. The peak at 2924 cm− 1 was associated to the stretching vibrations of aliphatic C–H. The peak at 1730 cm− 1 is attributed to the vibrational absorption of C = O, which is a characteristic peak of isocyanate modification.
The peaks at 2242 cm− 1 and 1626 cm− 1 respectively corresponds to the stretching vibration of C ≡ N and the C-C aromatic extension. Both the TFTPN and the final product composites spectra contained C-F stretching vibration peak at 1477 cm− 1. After the cyclodextrins were modified with isocyanate and crosslinked with tetrafluoroteronitrile, the peak intensity was weaker than the reactant material. However, the C-F bond peak can still be clearly observed in the spectrum.
The 13C solid-state NMR spectra of X-CDP are shown in Fig. 4c. The resonance associated with β-CD is exhibit at δ = 74 ppm. Resonances at δ = 98 and 142 ppm correspond to the newly formed alkoxy groups and aromatic carbons, respectively. Resonances at δ = 111 and 124 ppm are related to the two unsubstituted carbons on the benzene ring introduced by isocyanate substitution. This reveals that the cyclodextrin was successfully phenylcarbamoylated and crosslinked by tetrafluoroterephthalonitrile.
3.2. The adsorption of organic pollutants
Cl-CDP, NO2-CDP, F-CDP and X-CDP are used to adsorb the common and commercialized organic pollutants in the field of water purification (Fig. 5). The negative inductive effect of-Cl, -NO2, and -CF3 reduces the electron cloud density on the benzene ring, making it easier for the adsorbent to form hydrogen bonds with pollutants. When adsorbing pollutants containing benzene ring can be adsorbed by π-π interaction with the benzene ring on the adsorbent. Combined with the inherent inclusion effect of the cyclodextrin cavity, especially non-planar compounds are more easily captured by the cavities of cyclodextrin, the adsorbent exhibits good performance in adsorbing pollutants.
Furthermore, X-CDP has a high adsorption capacity for almost all pollutants due to its porous structure and effect synergy of various functional groups which can form various intermolecular forces and hydrogen bonds. In Fig. 6a, X-CDP removed various small organic molecule contaminants faster than other materials. X-CDP reached almost ~ 95% of its equilibrium in 30 s. In contrast, it took 10 min for other adsorbents to reach equilibrium and adsorbed only 46% of its equilibrium value in 30 s.
3.3. Batch adsorption kinetic modeling
As shown in Fig. 7, different adsorbents have different adsorption effects for the same compound. The characteristic peaks of ibuprofen, dichlorophenol, naphthol, norfloxacin, bisphenol A and tetracycline are located at 225 nm, 287 nm, 287 nm, 275 nm, 275 nm, and 355 nm, respectively. Obviously, after adsorption, the characteristic peak intensity decreased with different degrees. Those adsorbents have obvious absorption for the above organic substances. X-CDP has the most obvious adsorption effect on various pollutions. The removal rate of norfloxacin and tetracycline by X-CDP in 30 s is over 95%, and the adsorption equilibrium is basically reached within 2 min. The rapid removal rate of pollutants has never been reported before.
The effect of contact time on adsorption of organic contaminants by X-CDP and other adsorbents are shown in Fig. 8a, 8b. The adsorption of contaminants by X-CDP was high and quick in the initial 2 min due to its porosity. Especially for polycyclic and high molecular weight pollutants, such as tetracycline and nopifloxacin, these pollutants are more likely to be quickly included due to the three-dimensional cavity structure of cyclodextrin.
The pseudo-second-order models were used to investigate adsorption kinetics. Using the pseudo-second-order kinetic model (Eq. (1)), the adsorption process was expressed as:
Where qt (mg·g− 1) is the amount of adsorbed organic contaminants at any time t (min); k (mg·g− 1·min− 1) is the second-order rate constant; and qe (mg·g− 1) represents the amount of adsorbed organic contaminants at equilibrium. Kinetic constants of the pseudo-second-order kinetic model are estimated by the experimental data in Fig. 9. The process of adsorbing organic contaminants for X-CDP fit the pseudo-second-order model well with a high linear relationship between t/qt and t (R2 > 0.999). The apparent pseudo-second-order rate constant (kobs) of the adsorption of tetracycline by X-CDP is 2.49 g·mg− 1·min− 1, which is higher than the other studied adsorbents for tetracycline or any other pollutant removed by mesoporous silicas or carbohydrate-based adsorbents at the same experimental conditions. X-CDP’s superior k for adsorbing organic contaminants indicates that nearly all of its β-CD binding sites are readily accessible, and the mount of binding sites is higher than many adsorbents.
3.4. Batch adsorption isotherm modeling
Adsorption isotherm is important for determining the adsorption behavior of an adsorbent. In order to better investigate the adsorption mechanisms, the Langmuir model (Eq. (2)) was applied to the experimental data using the following equations:
Where qe (mg·g− 1) is the adsorption capacity at equilibrium; qmax (mg·g− 1) is the maximum adsorption capacity; ce (mg·L− 1) is the equilibrium concentration in the solution; The plot of ce/qe versus ce was employed to generate qm and b. The linear Langmuir model data of the four adsorbents for tetracycline were shown in Table. 1.
Table. 1 The Langmuir model data of tetracycline adsorption by four adsorbents.
Adsorbents
|
qmax
|
b
|
R2
|
NO2-CDP
|
47.62
|
0.0017
|
0.9991
|
Cl-CDP
|
1.20
|
0.0223
|
0.9926
|
F-CDP
|
3.13
|
0.0123
|
0.9912
|
X-CDP
|
230.15
|
0.0094
|
0.9982
|
According to the equilibrium adsorption value and the concentration of tetracycline after adsorption, the Langmuir model was fitted. As shown in Table 1, R2 values are quite close to 1, which means the adsorption process accorded with Langmuir model. The adsorption process is monolayer adsorption and there is no interaction between adsorbates. Furthermore, the maximum adsorption capacity of X-CDP at equilibrium was found to be 230.15 mg·g− 1, the qmax value for X-CDP was competitive with the adsorption capacities of tetracycline of other adsorbents, such as chitosan powder (23.92 mg/g) (Huang et al. 2011), honeycomb tubular biochar (123.6 mg/g) (Ma et al. 2018), and NaY zeolite (201.8 mg/g) (Ali et al. 2018).
3.5. The adsorption-desorption cycles of X-CDP
An ideal adsorbent should have high adsorption ability as well as excellent desorption performance, which reduce the operating cost for the adsorbent application. As is shown in Fig. 10, the efficiency (%) was the ratio of the weight of absorbed (or desorbed) tetracycline at any time. Thus, the adsorption capacity of the desorbed X-CDP was examined by using ethanol as eluent and the regenerated X-CDP was reused in tetracycline solution with known concentration. From Fig. 10, the removal rate of tetracycline is still 86.7%, which is only 8.3% lower than the first adsorption after five cycles. This reduction might be ascribed to the loss of binding sites after each desorption procedure. These results demonstrated that the recovery efficiency of X-CDP was relatively high with slightly affected by the five consecutive regeneration cycles and X-CDP showed excellent re-adsorption effect.