Facile Synthesis of Microporous Ferrocenyl Polymers Photocatalyst for Degradation of Cationic Dye

Microporous organic polymers (MOPs) were prepared by condensation reactions from substituent-group-free carbazole and pyrrole with 1,1′-ferrocenedicarboxaldehyde without adding any catalysts. The resultant MOPs were insoluble in common solvent and characterized by FTIR, XPS, TGA and SEM. An N2 adsorption test showed that the obtained polymers PFcMOP and CFcMOP exhibited Brunauer–Emmett–Teller (BET) surface areas of 48 and 105 m2 g−1, respectively, and both polymers possessed abundant micropores. The MOPs with a nitrogen and ferrocene unit could be potentially applied in degrading dye with high efficiency.

Dye molecules, which have been extensively utilized in various industries, have caused tremendous environmental pollution due to the poor biodegradation of organic dyes [29]. Therefore, the treatment of dye wastewater is of great importance for the environmental protection. To this end, a large number of catalysts, such as hydrogenated graphitic carbon nitride [30], TiO 2 [31], AC-TiO 2 [32], CaBi 2 O 4 [33], CaCO 3 /Ag 2 CO 3 /AgI/Ag composites [34], Ag@AgCl [35], Cu@MnO 2 nanowires [36], etc., have been reported to decompose organic dyes. In recent years, MOPs with light-responsive catalytic properties have attracted numerous interest [37,38]. Yu et al. have reported serials of microporous polymers photocatalyst (CMP-CSUs) via the oxidative coupling polymerization approach. They reported that CMP-CSUs can be used to catalyze C-3 formylation and the thiocyanation of indoles reactions and Ugi-type reaction [39,40]. Although significant progress toward photoactive MOPs with light-harvesting and catalytic properties has already been achieved, few research studies were available on the direct use of MOPs for photocatalytically degrading organic dye contaminant.
In this work, two hypercrosslinked polymer networks were synthesized by facile catalyst-free polycondensation from substituent-group-free carbazole and pyrrole with 1,1 -ferrocenedicarboxaldehyde, using 1,4-dioxane as a reaction medium. The abundance of carbazole, pyrrole and iron atoms in the polymeric skeleton enables us to examine the photoactive catalysis property of the polymers toward dye molecules, and the cationic methylene blue (MB) was used as a model dye. Notably, the as-prepared MOPs show excellent chemical stability, and they could be used to degrade dye with high efficiency.

Results and Discussion
As shown in Scheme 1, the nitrogen-rich and ferrocene-based microporous polymers networks were synthesized in 1,4-dioxane via one-step polycondensation at 220 • C from substituent-group-free carbazole and pyrrole with 1,1 -ferrocenedi-carboxaldehyde without adding any catalyst. The resultant solid polymers showed complete insolubility in any organic solvent, for example, tetrahydrofuran (THF), dimethyl sulfoxide (DMSO), and N,N-dimethylformamide (DMF), suggesting the hypercrosslinked nature of polymers' structures. The thermal stability was tested through the thermogravimetric analysis (TGA) under inert atmosphere at temperature from 25 to 800 • C using a heating rate of 10 • C/min ( Figure S1, ESI †). A small loss of polymers weight was detected in the preliminary stage ascribed to the residual solvents and the adsorbed moisture inside the pores. PFcMOP began to be degenerated at 200 • C under nitrogen atmosphere, and the degeneration temperature of carbazole-based CFcMOP reached 400 • C, suggesting that the formation of a C-C covalent bond between the aldehyde group and benzene ring in carbazole was more stable than in pyrrole. At temperature from 200 to 800 • C, the continuous weight loss of the polymers was derived from the carbonization of MOPs attributed to the fact that TGA was carried out under the atmosphere of nitrogen. The X-ray diffraction test exhibited that they had amorphous morphology ( Figure S2, ESI †). Like other microporous polymers, PFcMOP consisted of loose agglomerates of tiny particles with a rough surface and irregular shape, as investigated by SEM ( Figure 1A,B). Interestingly, CFcMOP using carbazole as a connect unit can form sphere-like particles ( Figure 1C,D). reaction [39,40]. Although significant progress toward photoactive MOPs wit light-harvesting and catalytic properties has already been achieved, few research studie were available on the direct use of MOPs for photocatalytically degrading organic dy contaminant.
In this work, two hypercrosslinked polymer networks were synthesized by facil catalyst-free polycondensation from substituent-group-free carbazole and pyrrole wit 1,1′-ferrocenedicarboxaldehyde, using 1,4-dioxane as a reaction medium. The abundanc of carbazole, pyrrole and iron atoms in the polymeric skeleton enables us to examine th photoactive catalysis property of the polymers toward dye molecules, and the cationi methylene blue (MB) was used as a model dye. Notably, the as-prepared MOPs show excellent chemical stability, and they could be used to degrade dye with high efficiency.

Results and Discussion
As shown in Scheme 1, the nitrogen-rich and ferrocene-based microporou polymers networks were synthesized in 1,4-dioxane via one-step polycondensation a 220 °C from substituent-group-free carbazole and pyrrole wit 1,1′-ferrocenedi-carboxaldehyde without adding any catalyst. The resultant soli polymers showed complete insolubility in any organic solvent, for example tetrahydrofuran (THF), dimethyl sulfoxide (DMSO), and N,N-dimethylformamid (DMF), suggesting the hypercrosslinked nature of polymers' structures. The therma stability was tested through the thermogravimetric analysis (TGA) under iner atmosphere at temperature from 25 to 800 °C using a heating rate of 10 °C/min ( Figure S1 ESI †). A small loss of polymers weight was detected in the preliminary stage ascribed t the residual solvents and the adsorbed moisture inside the pores. PFcMOP began to b degenerated at 200 °C under nitrogen atmosphere, and the degeneration temperature o carbazole-based CFcMOP reached 400 °C, suggesting that the formation of a C-C covalent bond between the aldehyde group and benzene ring in carbazole was mor stable than in pyrrole. At temperature from 200 to 800 °C, the continuous weight loss o the polymers was derived from the carbonization of MOPs attributed to the fact that TGA was carried out under the atmosphere of nitrogen. The X-ray diffraction test exhibite that they had amorphous morphology ( Figure S2, ESI †). Like other microporou polymers, PFcMOP consisted of loose agglomerates of tiny particles with a rough surfac and irregular shape, as investigated by SEM ( Figure 1A,B). Interestingly, CFcMOP usin carbazole as a connect unit can form sphere-like particles ( Figure 1C,D). The chemical connected structure of MOPs was determined by Fourier transform infrared spectroscopy (FTIR). As exhibited in the FTIR spectra ( Figure 2), following th deformation (1680 cm −1 ) and strong reduction of the aldehyde i 1,1′-ferrocenedicarboxaldehyde, the adsorption bands at 2970 cm −1 were assigned t Csp3-H stretching vibrations indicative of the successful polymerizations reaction o substituent-group-free carbazole and pyrrole with 1,1′-ferrocenedicarboxaldehyde. I addition, the bands at around 1640 cm −1 are assigned to the stretching vibrations of N−H derived from carbazole and pyrrole rings. Furthermore, the elemental content information in the as-prepared polymers was investigated by XPS spectra. As shown i Figures S3 and S4, the iron element signals were observed in the spectra, and the iro content values listed in Table S1 were 4.91% and 4.61% for PFcMOP and CFcMOP respectively. In addition, the measured carbon content was 92.12% in CFcMOP, whic agreed with the theoretical values (92.30%) and the higher carbon contents (91.30%) i PFcMOP than the calculated values (86.95%), which may be due to the absorption of CO in the porous polymers. Furthermore, the nitrogen content values were 3.17% and 3.24% for PFcMOP and CFcMOP, respectively. The difference between the measured value an theoretical value could be ascribed to the XPS test on the surface content of th polymers.  The chemical connected structure of MOPs was determined by Fourier transform infrared spectroscopy (FTIR). As exhibited in the FTIR spectra (Figure 2), following the deformation (1680 cm −1 ) and strong reduction of the aldehyde in 1,1 -ferrocenedicarboxaldehyde, the adsorption bands at 2970 cm −1 were assigned to C sp3 -H stretching vibrations indicative of the successful polymerizations reaction of substituent-group-free carbazole and pyrrole with 1,1 -ferrocenedicarboxaldehyde. In addition, the bands at around 1640 cm −1 are assigned to the stretching vibrations of N−H derived from carbazole and pyrrole rings. Furthermore, the elemental contents information in the as-prepared polymers was investigated by XPS spectra. As shown in Figures S3 and S4, the iron element signals were observed in the spectra, and the iron content values listed in Table S1 were 4.91% and 4.61% for PFcMOP and CFcMOP, respectively. In addition, the measured carbon content was 92.12% in CFcMOP, which agreed with the theoretical values (92.30%) and the higher carbon contents (91.30%) in PFcMOP than the calculated values (86.95%), which may be due to the absorption of CO 2 in the porous polymers. Furthermore, the nitrogen content values were 3.17% and 3.24% for PFcMOP and CFcMOP, respectively. The difference between the measured value and theoretical value could be ascribed to the XPS test on the surface content of the polymers. The chemical connected structure of MOPs was determined by Fourier transform infrared spectroscopy (FTIR). As exhibited in the FTIR spectra ( Figure 2), following the deformation (1680 cm −1 ) and strong reduction of the aldehyde in 1,1′-ferrocenedicarboxaldehyde, the adsorption bands at 2970 cm −1 were assigned to Csp3-H stretching vibrations indicative of the successful polymerizations reaction of substituent-group-free carbazole and pyrrole with 1,1′-ferrocenedicarboxaldehyde. In addition, the bands at around 1640 cm −1 are assigned to the stretching vibrations of N−H derived from carbazole and pyrrole rings. Furthermore, the elemental contents information in the as-prepared polymers was investigated by XPS spectra. As shown in Figures S3 and S4, the iron element signals were observed in the spectra, and the iron content values listed in Table S1 were 4.91% and 4.61% for PFcMOP and CFcMOP, respectively. In addition, the measured carbon content was 92.12% in CFcMOP, which agreed with the theoretical values (92.30%) and the higher carbon contents (91.30%) in PFcMOP than the calculated values (86.95%), which may be due to the absorption of CO2 in the porous polymers. Furthermore, the nitrogen content values were 3.17% and 3.24% for PFcMOP and CFcMOP, respectively. The difference between the measured value and theoretical value could be ascribed to the XPS test on the surface content of the polymers. The porosity of the resultant ferrocene-based polymer networks was evaluated by the N2 adsorption and desorption isotherms at 77 K, as illustrated in Figure 3a. Both the The porosity of the resultant ferrocene-based polymer networks was evaluated by the N 2 adsorption and desorption isotherms at 77 K, as illustrated in Figure 3a. Both the two polymer networks exhibited a continual increase of N 2 uptake at the relative pressure (P/P 0 ) up to 1.0, suggesting the characteristics of mesopores. In addition, CFcMOP showed a steep rise at very low relative pressure (P/P 0 < 0.01), and the N 2 adsorption of CFcMOP increased slowly at relative pressure (P/P 0 ) from 0.1 to 0.9, which can be roughly classified as type I sorption [41]. When the relative pressure exceeds 0.9, nitrogen uptake rapidly increases, suggesting the presence of meso-and macroporous structures probably derived from the interparticulate voids because of the loose packing of tiny particles, as detected by the field-emission SEM micrographs. Utilizing nitrogen as a probe, the pore sizes and distributions for the two microporous polymers were calculated through the nonlocal functional theory (NLDFT) from the adsorption isotherms of nitrogen at 77 K. As shown in Figure 3b, the PFcMOP exhibits the pore size of about 1.86 nm, and the CFcMOP shows a slightly smaller pore size located at 1.36 nm, which belongs to micropore region. Considering the higher molecular size of pyrrole than carbazole, the smaller pore size of carbazole-based CFcMOP than PFcMOP can be attributed to interpenetration effect, which is detected in many porous organic polymers [42,43].
Polymers 2022, 14, x FOR PEER REVIEW 4 of 8 two polymer networks exhibited a continual increase of N2 uptake at the relative pressure (P/P0) up to 1.0, suggesting the characteristics of mesopores. In addition, CFcMOP showed a steep rise at very low relative pressure (P/P0 < 0.01), and the N2 adsorption of CFcMOP increased slowly at relative pressure (P/P0) from 0.1 to 0.9, which can be roughly classified as type I sorption [41]. When the relative pressure exceeds 0.9, nitrogen uptake rapidly increases, suggesting the presence of meso-and macroporous structures probably derived from the interparticulate voids because of the loose packing of tiny particles, as detected by the field-emission SEM micrographs. Utilizing nitrogen as a probe, the pore sizes and distributions for the two microporous polymers were calculated through the non-local functional theory (NLDFT) from the adsorption isotherms of nitrogen at 77 K. As shown in Figure 3b, the PFcMOP exhibits the pore size of about 1.86 nm, and the CFcMOP shows a slightly smaller pore size located at 1.36 nm, which belongs to micropore region. Considering the higher molecular size of pyrrole than carbazole, the smaller pore size of carbazole-based CFcMOP than PFcMOP can be attributed to interpenetration effect, which is detected in many porous organic polymers [42,43]. The specific surface areas and porosity parameters of PFcMOP and CFcMOP were calculated from the adsorption isotherms. Using the Brunauer−Emmett−Teller (BET) model, the CFcMOP possesses specific surface areas up to 104.5 m 2 g −1 . However, the BET surface areas of PFcMOP using pyrrole as a building unit descended to 48.8 m 2 g −1 . That the BET surface areas of CFcMOP are a little higher than PFcMOP probably can be ascribed to the condensation reaction of the aldehyde group with two carbon atoms at the same five-membered ring of pyrrole, while there are different benzene units in carbazole, which affect the reacting activity of monomers. In addition, the pore volume of PFcMOP and CFcMOP reached 0.07 and 0.1 cm 3 g −1 , respectively.
Photocatalytic degradation of dyes was usually seen as an efficient and simple approach for the removal of dye pollutants due to its high efficiency, low cost and simplicity. Considering the existence of the metal ion in the microporous polymer networks, the photocatalytic activity of MOPs in methylene blue degradation under LED light irradiation was explored. Typically, the time-dependent adsorption of MB on 3 mg PFcMOP and CFcMOP was investigated at an initial concentration of 10 mg/L in the test tube with a magnetic stir bar. Firstly, the system was stirred without light for 60 min to keep the adsorption/desorption balance. After stirring at certain time intervals during the reaction processes, the absorbance variation of MB was detected at 660 nm by means of UV-vis spectroscopy in Figure 4. As seen in Figure 5, the efficiency of degrading MB could achieve 54% and 65% within 20 min for PFcMOP and CFcMOP, respectively. After stirring for 100 min, the degradation finally reached 65% and 78% for PFcMOP and CFcMOP, respectively. The higher photocatalytic activity of CFcMOP than PFcMOP can be ascribed to the larger special surface area of CFcMOP, in which MB molecules can easy diffuse into the pore channel of CFcMOP. The specific surface areas and porosity parameters of PFcMOP and CFcMOP were calculated from the adsorption isotherms. Using the Brunauer-Emmett-Teller (BET) model, the CFcMOP possesses specific surface areas up to 104.5 m 2 g −1 . However, the BET surface areas of PFcMOP using pyrrole as a building unit descended to 48.8 m 2 g −1 . That the BET surface areas of CFcMOP are a little higher than PFcMOP probably can be ascribed to the condensation reaction of the aldehyde group with two carbon atoms at the same five-membered ring of pyrrole, while there are different benzene units in carbazole, which affect the reacting activity of monomers. In addition, the pore volume of PFcMOP and CFcMOP reached 0.07 and 0.1 cm 3 g −1 , respectively.
Photocatalytic degradation of dyes was usually seen as an efficient and simple approach for the removal of dye pollutants due to its high efficiency, low cost and simplicity. Considering the existence of the metal ion in the microporous polymer networks, the photocatalytic activity of MOPs in methylene blue degradation under LED light irradiation was explored. Typically, the time-dependent adsorption of MB on 3 mg PFcMOP and CFcMOP was investigated at an initial concentration of 10 mg/L in the test tube with a magnetic stir bar. Firstly, the system was stirred without light for 60 min to keep the adsorption/desorption balance. After stirring at certain time intervals during the reaction processes, the absorbance variation of MB was detected at 660 nm by means of UV-vis spectroscopy in Figure 4. As seen in Figure 5, the efficiency of degrading MB could achieve 54% and 65% within 20 min for PFcMOP and CFcMOP, respectively. After stirring for 100 min, the degradation finally reached 65% and 78% for PFcMOP and CFcMOP, respectively. The higher photocatalytic activity of CFcMOP than PFcMOP can be ascribed to the  The photocatalytic active mechanism of MOPs was studied below. As is well known four species including electron (e − ), active hole (h + ), hydroxyl radical (·OH) and single oxygen ( 1 O2) are presumed to react as active sites for photocatalytically degrading th dye [44]. Different scavengers, namely CuSO4, EDTA-2Na, tert-butanol (TBA) and L-Hi were selectively added to the photocatalytic reaction for quenching electron (e − ), activ hole (h + ), ·OH and 1 O2, respectively. As exhibited in Figure 6a, the presence of CuSO EDTA-2Na, L-His and tert-butanol led to a prominent restraint in the degradation of MB It reflected that e − , h + , 1 O2 and ·OH were reactive species in this degradation reaction. A shown in Figure 6a, the efficiency of PFcMOP degeneration reaction descended to 25% 25%, 20% and 17%, respectively. In addition, the efficiency of CFcMOP degeneratio reaction descended to 38%, 22%, 34% and 20% after adding corresponding scavengers t the degradation system (Figure 6b). Hence, in the MB photodecomposition reactio catalyzed by the MOPs, singlet oxygen and hydroxyl radical could be generated simultaneously as reactive species in the degradation of MB dye.    The photocatalytic active mechanism of MOPs was studied below. As is well four species including electron (e − ), active hole (h + ), hydroxyl radical (·OH) and oxygen ( 1 O2) are presumed to react as active sites for photocatalytically degra dye [44]. Different scavengers, namely CuSO4, EDTA-2Na, tert-butanol (TBA) an were selectively added to the photocatalytic reaction for quenching electron (e hole (h + ), ·OH and 1 O2, respectively. As exhibited in Figure 6a, the presence of EDTA-2Na, L-His and tert-butanol led to a prominent restraint in the degradatio It reflected that e − , h + , 1 O2 and ·OH were reactive species in this degradation reac shown in Figure 6a, the efficiency of PFcMOP degeneration reaction descended 25%, 20% and 17%, respectively. In addition, the efficiency of CFcMOP dege reaction descended to 38%, 22%, 34% and 20% after adding corresponding scave the degradation system (Figure 6b). Hence, in the MB photodecomposition catalyzed by the MOPs, singlet oxygen and hydroxyl radical could be ge simultaneously as reactive species in the degradation of MB dye. The photocatalytic active mechanism of MOPs was studied below. As is well known, four species including electron (e − ), active hole (h + ), hydroxyl radical (OH) and singlet oxygen ( 1 O 2 ) are presumed to react as active sites for photocatalytically degrading the dye [44]. Different scavengers, namely CuSO 4 , EDTA-2Na, tert-butanol (TBA) and L-His were selectively added to the photocatalytic reaction for quenching electron (e − ), active hole (h + ), ·OH and 1 O 2 , respectively. As exhibited in Figure 6a, the presence of CuSO 4 , EDTA-2Na, L-His and tert-butanol led to a prominent restraint in the degradation of MB. It reflected that e − , h + , 1 O 2 and ·OH were reactive species in this degradation reaction. As shown in Figure 6a, the efficiency of PFcMOP degeneration reaction descended to 25%, 25%, 20% and 17%, respectively. In addition, the efficiency of CFcMOP degeneration reaction descended to 38%, 22%, 34% and 20% after adding corresponding scavengers to the degradation system (Figure 6b). Hence, in the MB photodecomposition reaction catalyzed by the MOPs, singlet oxygen and hydroxyl radical could be generated simultaneously as reactive species in the degradation of MB dye.
It reflected that e − , h + , 1 O2 and ·OH were reactive species in this degradation reaction. As shown in Figure 6a, the efficiency of PFcMOP degeneration reaction descended to 25%, 25%, 20% and 17%, respectively. In addition, the efficiency of CFcMOP degeneration reaction descended to 38%, 22%, 34% and 20% after adding corresponding scavengers to the degradation system (Figure 6b). Hence, in the MB photodecomposition reaction catalyzed by the MOPs, singlet oxygen and hydroxyl radical could be generated simultaneously as reactive species in the degradation of MB dye.

Conclusions
In summary, two ferrocenyl microporous organic polymers were successfully synthesized utilizing 1,1 -ferrocenedicarboxaldehyde as building monomers. The resultant polymers networks were examined by FTIR, XPS and SEM in detail. In addition, MOPs showed good thermal stability and possessed high porosity with a BET surface area from 48 to 105 m 2 g −1 , and MOPs could be used in degrading dyes with high efficiency.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/polym14091900/s1, Figure S1: TGA curves of the two ferrocenyl microporous polymer networks, Figure S2: PXRD curves of the two ferrocenyl microporous polymer networks, Figure S3: XPS curves of the two ferrocenyl microporous polymer networks, Figure S4: XPS curves of Fe element in the two ferrocenyl microporous polymer networks, Table S1: C, Fe, and N element percent of PFcMOP and CFcMOP.

Conflicts of Interest:
The authors declare no conflict of interest.