Preparing for reactive SiO2 nanospheres (R-SiO2) and R-SiO2@DCX
SiO2 nanospheres with an average diameter of 150 nm were synthesized through the Stöber method. SiO2 nanospheres were preprocessed to obtain reactive SiO2 nanospheres (R-SiO2). SiO2 nanospheres (0.5 g) were added to 20 ml of anhydrous tetrahydrofuran and 1.5 g of trichloro[4-(chloromethyl)phenyl]silane to a mixture with drastic sitrring. Then, 5.0 ml anhydrous tetrahydrofuran and 1.2 ml of triethylamine mixture were added under nitrogen and reacted in a nitrogen/air atmosphere for 8 and 18 hours, respectively, and then reacted in air for 18 hours. The resulting solid precipitate (named R-SiO2) was filtered and sequentially washed with ethanol and tetrahydrofuran and vacuum-dried at 60 ℃ overnight.
Generally, 0.5 g of R-SiO2 nanospheres and 1.6 g of anhydrous ferric chloride were dispersed uniformly in 10 ml dichloroethane. Then, the system was added dropwise into a solution of p-dichlorobenzene in dichloroethane at a constant speed under the protection of nitrogen after reaching 80 ℃.
Synthesis of hierarchical porous polymer (HPP) and hierarchical porous carbon material (HPC)
R-SiO2@DCX was etched with 10% HF for 24 h. The precipitate was filtered and washed with ammonia and pure water to remove the residual HF. Then, the resulting product was vacuum-dried at 60 ℃ overnight to obtain a hierarchical porous polymer (HPP). Subsequently, the HPP was carbonized at 900 ℃ in a tube furnace under the protection of a N2 atmosphere with a flow rate of 80 ml min-1 to obtain a hierarchical porous carbon material (HPC).
Characterization
Scanning electron microscopy (SEM) images were obtained on an S8010 instrument (Hitachi, Japan) at an acceleration voltage and current of 10 kV and 10 μA, respectively. The N2 adsorption-desorption isotherm of the sample was measured at 77 K by an ASAP 2460 adsorption instrument (Micromeritics, USA). The BET method was used to calculate the specific surface area, and the density functional theory (DFT) equilibrium model was used to calculate the pore size distribution. The t-plot method was used to calculate the micropore surface area and external pore surface area. An inVia laser Raman spectrometer from Renishaw Company was used to measure the microcrystalline structure of the sample. The scanning range was from 800-2000 cm-1. The excitation wavelength was 514.5 nm with an exposure time of 30 s.
The adsorption performance test
A certain amount of methylene blue (MB) solution was added to an Erlenmeyer flask, and then HPC was added. The system was kept in a water bath with magnetic field at room temperature. It was placed in a constant temperature water bath with magnetic stirring for adsorption experiments. The solution was filtered with a microporous membrane, and the concentration of the filtrate was further measured by a 722 visible spectrophotometer (Shanghai Precision Scientific Instrument Co., Ltd.). The MB adsorption rate (R) was calculated according to formula (1):
In formula (1): C0 is the concentration of MB solution before adsorption (mg·L-1); Ce is the concentration of MB solution at adsorption equilibrium (mg·L-1);
The adsorption capacity (q) was calculated according to formula (2):
In formula (2), C0 is the concentration of MB solution before adsorption (mg·L-1), Ct is the concentration of MB solution at time t (mg·L-1), V is the volume of MB solution (L), and m is the mass of HPC (g).
Characterization of the HPP
The successful preparation of R-SiO2 nanospheres was proven by a series of characterizations (Fig. S1-S3). The R-SiO2 nanospheres were further supercrosslinked in situ by means of the Friedel-Crafts reaction. As shown in Fig. 1a, the surface of the R-SiO2@DCX nanospheres presented a rough and uneven coating layer, proving the supercrosslinked reaction on the R-SiO2 surface. The unmodified SiO2 was also used to react with DCX, while polymerization was carried out only between the DCX monomer. As shown in Fig 1c, SiO2@DCX shows a flat and agglomerated surface, indicating that the cross-linking reaction did not occur. After removing the silica template (Fig. 1b), the internal cavity of the R-SiO2@DCX nanosphere maintains a complete spherical structure. The Fig. 1d showed that the average size of R-SiO2@DCX nanoarticles are at 150 nm. In addition, a series of characterizations illustrate the successful preparation of R-SiO2@DCX (Fig S4-S7).
For carbon materials, the pore structure and element formation are mainly inherient from precursors12-14. Thus, the hypercrosslinking time was changed to explore the variation in the morphology and structure of HPP to precisely control the nanostructure of HPP. Fig. 2 (SEM characterization) shows that the samples all contained structures with hollow and crossing spheres when the supercrosslinking time was continuously extended from 3 h to 24 h. As the supercrosslinking time was just 3 h (Fig. 2a), the sample showed that a small part of the spherical shells were not connected with only a small part of the crosslinking hollow spheres, and the graft layer originally covered the spherical surface was almost peeled off, leaving a few residues. As the cross-linking time was extended to 6 h (Fig. 2b), the particle size of the sample tended to be uniform, while agglomeration still occurred, and hollow spheres were almost formed with a complete appearance. When the time reached 24 hours (Fig. 2e), the hollow nanosphere shells were well stacked and regularly arranged together, forming a hollow hole-like structure with good morphology. The oligomer on the surface of the nanospheres gradually decreased with the extension of the supercrosslinking time, which was caused by the gradual consumption of the polymerized part, leading to a more complete reaction 15.
As shown in Fig. 3, all the samples presented a typical hierarchical porous structure that included micropores, mesopores and macropores16,17. The pore size distribution of HPP-Xs based on DFT calculations (Fig. 3b) showed peaks at 0.67 nm and 1.2 nm, at 27 nm and at 50-100 nm, corresponding to micropores, mesopores and macropores, respectively. The micropores were derived from the methylene bridge voids of the cross-linking between the nanospheres. As the Friedel-Crafts reaction was rapid, once the reaction occurred, it quickly polymerized to form a methylene cross-linked bridge to form a microporous structure. The changes in the supercrosslinking reaction times deeply affected the crosslinking degree of the samples18.
Table 1 summarizes the DTF calculation results based on the nitrogen absorption/desorption tests. As the reaction time was extended, the SSA and Vmicro of the samples increased 19. HPP-3 showed an SSA of 386 m2·g-1 and a Vmicro of 0.04 cm3 g-1. As the crosslinking time of HPP-3 was shortened, the reaction could not proceed completely, leading to meso/macropores occluding the majority of the pore structure. With the extension of the cross-linking time, the SSA and the Vmicro of HPP-6, HPP-12, HPP-18 and HPP-24 showed an increasing trend. The maximal SSA and Vmicro were reached when the crosslinking time increased to 24 h, where Vtotal increased from 0.04 to 0.88 cm3 g-1 and the SSA increased from 386 m2 g-1 to more than double (889 m2 g-1). The results indicated that prolonging the super-crosslinking time was helpful to the Friedel-Crafts reaction, and the degree of cross-linking rose rapidly when the time reached 18 h. Comparing the value of Vmicro, it can be found that almost all the increase in the specific surface area came from the increase in the number of micropores. Since the microporous structure in the sample was mainly derived from the self-crosslinking reaction of the benzyl group between the benzene rings, the greater the degree of crosslinking between the spheres, the denser the cross-linked bridge structure, thus forming more micropores20,21. In conclusion, HPP-24 contained the highest SSAs and the largest number of micropores, which would be the best choice to be the precursor of HPC.
Table 1
Pore structure parameters of HPP at various hypercrosslinking times
Sample
(h)
|
SBET
(m2g-1)
|
Smic
(m2g-1)
|
Sext
(m2g-1)
|
Vmic
(cm3g-1)
|
Vext
(cm3g-1)
|
Vtotal
(cm3g-1)
|
3
|
386
|
135
|
251
|
0.04
|
0.26
|
0.30
|
6
|
397
|
143
|
201
|
0.04
|
0.27
|
0.31
|
12
|
434
|
186
|
218
|
0.06
|
0.31
|
0.37
|
18
|
450
|
229
|
221
|
0.09
|
0.33
|
0.42
|
24
|
889
|
567
|
332
|
0.23
|
0.65
|
0.88
|
The structure characterization of HPC
Fig. 4a shows an SEM image of a typical HPC, displaying a three-dimensional nanonetwork structure with mesopores and macropores, which was formed by the stack of carbon nanospheres. The broken nanospheres possessed a hollow structure, indicating successful etching of the SiO2 template. Fig. 4b show the Raman spectrum of HPC. The characteristic peak at 1345 cm-1 was the D peak, and the G peak was at 1590 cm-1, where the D peak corresponded to the vibration of carbon atoms and the G peak was the characterization of the graphitized structure, indicating that the carbon skeleton of HPC contained a graphite-like microcrystalline structure22. Fig. 4c illustrates the N2 adsorption-desorption isotherm of HPC. In the low-pressure region (P0/P < 0.1), the adsorption isotherm increased, indicating the microporous structure of HPC, while the isotherm showed a hysteresis loop area in the medium-pressure region, indicating the existence of mesopores; when the relative pressure was close to 1.0, the adsorption capacity increased significantly without the adsorption platform, showing the presence of macropores23. These conclusions were further proven by the pore size distribution curve, which was calculated from the DFT method. As shown in Fig. 4d, the pore size was distributed from the micropores (pore size below 2 nm) and the mesopores (pores from 2-50 nm) to the macropores (pores above 50 nm), indicating a hierarchical structure.
The influence of carbonization temperature on the HPC nanostructure and SSA was further studied. Fig. 5 shows the SEM image of HPC prepared at different carbonization temperatures. This result showed that the special three-dimensional structure, which was formed by stacking hollow nanospheres, can be prepared at all carbonization temperature ranges, indicating that HPC presents good thermal stability and framework strength. Table S1 summarizes the pore structure parameters of HPC prepared under different carbonization conditions. The optimization conditions were a carbonization temperature of 1000 ℃ with a 3 h carbonization time and a carbonization heating rate of 5 ℃ min-1. The as-prepared HPC presented a high SSA of 2388 m2·g-1 and an outstanding hierarchically porous structure.
Adsorption performance of HPC
The results above showed that HPCs were microporous materials with abundant adsorption active sites. In addition, HPCs are hierarchically porous materials with connected pores, which could shorten the transmission distance between the pores, making HPCs a promising adsorbent with high-efficiency adsorption performance24. HPC-1000 with an SSA of 2388 m2 g-1 was selected as the adsorbent, and methylene blue (MB) was selected as the adsorbate to investigate the adsorption performance of HPCs. The mechanism of adsorption was discussed based on the model of adsorption kinetics, adsorption isotherm and adsorption thermodynamics.
Fig. 6 shows the adsorption kinetic curve of HPC-1000 in aqueous solution toward MB. When the adsorption time was 5 minutes, the adsorption rate reached 59%, increased to 64% and 99% at 10 minutes and 60 minutes, respectively, and then did not change substantially as the adsorption time increased, indicating that the adsorption toward MB had reached equilibrium at this time. In the early stage of adsorption, the microporous structure of HPC-1000 has a large number of active sites, leading to a high initial rate of adsorption. As the adsorption time increased, the active sites inside HPC-1000 were filled with MB molecules, resulting in adsorption equilibrium after 60 minutes. Fig. 6 illustrates a digital photo of HPC-1000 adsorbing toward MB in aqueous solution. As the time increased, the color of the MB solution became lighter and was completely colorless after 2 h, indicating the effective adsorption performance of HPC-1000 toward the MB solution.
Fig. 6 also shows the fitting results of the adsorption process using the pseudo-first-order kinetic model, pseudo-second-order kinetic model, and Weber-Morris intraparticle diffusion model25. Table 2 summarized the adsorption kinetic parameters. The correlation coefficient value (R2) of pseudo-second-order kinetics was higher with R2 values greater than 0.99. The equilibrium adsorption capacity (qe-exp) obtained in the experiment was very close to the theoretical adsorption capacity (qe-cal) calculated by the pseudo-second-order kinetic model, indicating that the pseudo-second-order kinetic model was more suitable for describing the adsorption process of HPC toward MB solution. The intraparticle diffusion model was used by the segmentation method to fit the adsorption kinetics data (Fig. 6f). As shown in Table 2, the linear correlation of the segmentation fitting was better (the R2 value can reach 0.9774), indicating that MB was inside the HPC. The diffusion could be divided into two stages, while the first stage was the rapid diffusion of MB in macropores and mesopores, and the second stage was the slow diffusion in micropores. Fig. 6 b and c show the fitting results of the Langmuir model and Freundlich model26 of the thermodynamics of HPC adsorption toward MB solution. The fitting results are summarized in Table 3, where the linear correlation coefficients of the Langmuir model are all higher than those of the Freundlich model, indicating that the adsorption process is more according to the Langmuir isotherm model, which means the main monolayer adsorption of the adsorption process27. The study of the HPC adsorption thermodynamics toward MB solution was further explored. According to the influence of different temperatures on the adsorption in the range of 298 to 318 K, the adsorption thermodynamic constants can be calculated. As shown in Table 4, the Gibbs free energy ∆G value was less than zero at the experimental temperature. As the temperature rises, the absolute value of ∆G increases, indicating that HPC adsorption toward MB solution was a spontaneous behavior. The enthalpy change value (∆H) was more than 0, indicating that the adsorption of MB by HPC is an endothermic process. The entropy change value (ΔS) reflected the degree of disorder of the solid-liquid interface, and the value ΔS>0 indicated that the adsorption process increased the degree of disorder of the molecular motion between the solid-liquid interface, indicating that HPC had a strong affinity toward MB solution28.
Table 2
Kinetic parameters at various MB concentrations
Kinetic equation
|
C0
(mg·L-1)
|
qe-exp
(mg·g-1)
|
Kinetic parameters
|
Pseudo-first-order
ln(qe-qt) = lnqe-K1t
|
10
20
30
|
31.5
32.3
29.2
|
K1 = 0.07266, qe-cal =35.5, R2 = 0.9353
K1 = 0.07159, qe-cal =35.1, R2 = 0.9285
K1 = 0.06710, qe-cal =31.0, R2 = 0.9750
|
Pseudo-second-order
t/qt = 1/(K2qe2)+t/qe
|
10
20
30
|
48.1
57.9
77.8
|
K2 = 0.00279, qe-cal =51.8, R2 = 0.9956
K2 = 0.00337, qe-cal =59.0, R2 = 0.9966
K2 = 0.00439, qe-cal =78.2, R2 = 0.9991
|
Intraparticle diffusion
qt = Kid t1/2 + C
|
10
20
30
|
/
|
Kid1 = 3.71, C1= 19.1, R21 = 0.9774
Kid2 = 0.11, C2= 46.8, R22 = 0.9983
Kid1 = 3.53, C1= 28.4, R21 = 0.9619
Kid2 = 0.17, C2= 54.1, R22 = 0.9054
Kid1 = 2.89, C1= 53.7, R21 = 0.9505
Kid2 = 0.17, C2= 73.8, R22 = 0.9486
|
Table. 3
Langmuir and Freundlich isothermal equation parameters
Temperature
(℃)
|
Langmuir
|
Frenndlich
|
qm(mg·g-1)
|
KL(L·mg-1)
|
R2
|
|
KF(mg·g-1)
|
R2
|
25
|
125.1
|
0.0522
|
0.94996
|
|
9.0346
|
0.92111
|
35
|
121.1
|
0.058
|
0.99316
|
|
10.8137
|
0.97651
|
45
|
123.2
|
0.0588
|
0.99726
|
|
10.305
|
0.96459
|
Table. 4
Adsorption thermodynamic parameters
Temperature(K)
|
thermodynamic parameters
|
∆G(kJ·mol-1)
|
∆H(kJ·mol-1)
|
∆S(J·mol-1·K-1)
|
298
|
-4.86
|
|
|
308
|
-5.34
|
14.910
|
41.947
|
318
|
-5.70
|
|
|