3.1 Characterization analysis
To provide information for crystalline structure and phase composition of the as-synthesized catalysts in the presence of different morphologies, X-ray diffraction (XRD) was carried out. As illustrated in Fig. 1a, a series of peaks indexed as (003), (006), (012), and (110) located at 11.02°, 22.98°, 34.39°, and 59.26° were observed in all samples, which can be well-indexed to NiFe-LDH structure (JCPS card no. 40–0215) (Miranda et al. 2014). Compared with the peaks of P-NiFe-LDH, the (003) reflection of FH-NiFe-LDH was shifted, while the (006) reflection was significantly shifted to lower angles, but the (110) reflection remained the same. This result may be attributed to lattice distortion reaction caused by residual stress, thus leading to more surface defects. Meanwhile, the peak width of FS-NiFe-LDH and FH-NiFe-LDH is significantly larger than P-NiFe-LDH, indicating that the crystal size in clusters is smaller. Through Bragg’s equation, the intralayer spacing value of P-NiFe-LDH was calculated to be ~ 7.72Å from the (003) peak. However, the intralayer spacing value of FS-NiFe-LDH and FH-NiFe-LDH were calculated to be ~ 8.16 and ~ 8.31 Å, respectively. This data demonstrated that the interlayer spacing of the catalysts obviously increased, which was beneficial to pollutants’ entering.
Figure 1b displays the FT-IR spectra of the as-prepared catalysts. The broad bands around the 3400–3600 cm− 1 region are assigned to the -OH stretching vibrations of the surface-bound -OH group. The obvious -OH stretching vibration of P-NiFe-LDH may possess a superior affinity with water molecules to reveal strong hydrophilicity. A sharp absorption at ~ 1375 cm− 1 is observed in all samples, which ascribed to the asymmetric stretching type of CO32−. Meanwhile, the peaks at ~ 795, ~626, and ~ 432 cm− 1 are assigned to M − O vibrations and M − O−H bending (M represent metals) (Li et al. 2008). Synchronously, two bands at ~ 2935 and ~ 1051 cm− 1 were watched in both FS-NiFe-LDH and FH-NiFe-LDH which were belonged to the bending vibration absorption of C − H and the C − O stretching of ethanol (Gao et al. 2014). The stretching vibration of the interlayer water is located at ~ 1634 cm− 1, while the carbonate in the layer of NiFe-LDH is observed by the strong band at 1385 cm− 1 (Wang et al. 2015). The results further manifested the existence of intercalated -OH and CO32− groups, as well as H2O.
To further investigate the morphology and microstructure of the as-synthesized LDH, SEM was conducted (Fig. 2). A uniform morphology of plate-like hexagonal can be observed with an average lateral size 500 − 600 nm, while FH-NiFe-LDH and FS-NiFe-LDH composed of hierarchical flower-like spheres assembled by nanosheets with an average lateral size 800 − 900 nm. Simultaneously, the morphology of FS-NiFe-LDH and FH-NiFe-LDH is similar, except the latter is a spherical structure with a hollow inside. Moreover, both solid and hollow structures are obviously constructed by the unit of the hexagonal sheet, indicating that a typical 2D/3D homojunction structure is successfully constructed. It has been reported that 2D/3D homojunction with natural matching chemical and electronic structures between structural units was more beneficial to transfer electrons at the interface (Liu et al. 2019b).
HR-TEM images of the as-prepared three catalysts with different morphologies were further analysed (Fig. 3). The images revealed the lattice fringe to be 0.334, 0.195, and 0.219 nm, which were consistent with (113), (112) and (115) planes of NiFe-LDH, respectively. Remarkably, the preferentially exposed {110} planes of FH-NiFe-LDH possess higher surface energy, which was verified by the reported theoretical calculation. Likewise, the HRTEM images also exhibited that FS-NiFe-LDH and FH-NiFe-LDH had mainly exposure of {110} facets. However, the {100} facets of P-NiFe-LDH are the exposed facets, which is well match with previous study (Zhang et al. 2021). Moreover, it was worth noticing that there were some voids in the overall morphology for FS-NiFe-LDH and FH-NiFe-LDH, indicating different amounts of defects on the surface of samples which more intuitively verifies the judgment on the location of different OVs defects.
To comprehend the origin of the catalytic activity, the specific surface and pore size distribution of the structures were explored by a N2 adsorption/desorption analyzer. Figure 4a and b present the Brunauer-Emmett-Teller (BET) sorption isotherms and the corresponding distribution of Barrette-Joynere-Halenda (BJH) pore size. The pore size of NiFe-LDH was distributed in the mesoporous region, and specific surface areas (SSAs) decreased in the order FH-NiFe-LDH > FS-NiFe-LDH > P-NiFe-LDH, which corresponded to catalytic performance. This is because the increase of SSAs can accelerate mass transfer and maintain superior utilization efficiency of exposed active sites, thus enhancing catalytic performance (Liu et al. 2019b). Meanwhile, the isotherms exhibited hysteresis loops of varying intensities, which are consistent with the capillary condensation under high pressure. According to IUPAC classification, they could be categorized to the type IV with H3 type hysteresis loops (Song et al. 2015). The calculated BET specific surface area of the FH-NiFe-LDH was ~ 125.3487 m2·g− 1, higher than that of FS-NiFe-LDH (~ 74.7166 m2·g− 1) and P-NiFe-LDH (~ 60.1322 m2·g− 1). In addition, the total pore volumes of FS-NiFe-LDH and P-NiFe-LDH were ~ 0.2158 cm3·g− 1 and 0.0832cm3·g− 1, respectively, while the total pore volume of FH-NiFe-LDH increased to ~ 0.3636 cm3·g− 1. The increase of pore volume might create more efficient transport routes to catalysts’ internal voids and was conducive to the pollutant adsorption.
Considering the reactive facets can enrich abundant active sites, more exposed active facets are considered to be a decisive reason for improving catalytic performance. Generally, by virtue of the positive correlation between surface energy and chemical reactivity, the relatively less stable {110} facets theoretically exhibited higher reactivities than the stable {100} facet (Zhang et al. 2021). The conclusion proved that the catalytic activities of NiFe-LDH were primarily associated with the appearance of exposed facet which possessed high surface energies.
The surface chemical states and the elemental compositions of as-prepared catalysts were conducted by XPS (Fig. 5). The wide-scan high-resolution XPS spectra of the samples (Fig. 10a) demonstrate characteristic peaks of Ni 2p, Fe 2p, O 1s, and C 1s, indicating that there was no new element introduced to the catalysts. As clearly revealed in Fig. 5a, the XPS spectrum of Ni 2p3/2 exhibited no apparent difference in three samples, indicating that the morphology did not change the chemical environment of Ni 2p. In comparation of the P-NiFe-LDH, the binding energy of Fe 2p 3/2 in FS-NiFe-LDH and FH-NiFe-LDH showed weak shift to lower region about 1.1 and 1.3 eV, respectively, evidencing that the Fe element in the latter accepted the electron. Correspondingly, a red-shifted was watched in both FS-NiFe-LDH and FH-NiFe-LDH by the spectra of O 1s, revealing that more electrons were transmitted (Wang et al. 2022). Hence, we speculated that crystal morphology-engineering could induce lattice distortion and thus resulting to the appearance of oxygen vacancies.
The XPS spectra of Ni 2p, Fe 2p, and O 1s were detailedly analyzed to further verify the above conclusion. Figure 6a displayed the XPS spectra of Ni 2p, the peaks centered at 853.1 and 855.2 eV were considered as Ni(Ⅱ) and Ni(Ⅲ), respectively (Liu et al. 2017, Wang et al. 2022). Another two peaks around 858.7 and 861.1 eV could be assigned to the satellite of Ni(Ⅱ) and Ni(Ⅲ), respectively. After calculation, the element proportions of the three catalysts were obtained. The proportion of Ni(Ⅱ) in FS-NiFe-LDH and FH-NiFe-LDH accounted for 40.12% and 41.07%, respectively, higher than P-NiFe-LDH (32.77%). As for the Fe element (Fig. 6b), the peaks centered at 708.5, 721.8, and 727.3 eV were corresponded to Fe(Ⅲ), another peak at 713.2 eV was attributed to Fe(Ⅱ) which coordinated to -OH (Wang et al. 2016). The proportion of Fe(Ⅱ) in P-NiFe-LDH accounted for 21.35%, while for the FS-NiFe-LDH and FH-NiFe-LDH samples, the proportion of Fe(Ⅱ) further increased to 38.84% and 38.16%, respectively, elucidating that the morphological change alters the surface electronic structure and leads to the conversion of valence state.
Due to a large number of localized electrons and surface suspended bonds, the OVs enable modulating the electronic states as well as the coordination structures of catalysts, thereby influencing its kinetics and catalytic performance (Sun et al. 2015). According to reported experimental and theoretical studies, the implantation of OVs may be beneficial to the adsorption and activation of PS molecules (Wu et al. 2020). Since morphology-engineering can regulate OVs concentration, we thoroughly believe that different morphology of NiFe-LDH crystals is expected to produce different concentrations of OVs, thus tuning their catalytic performance. Considering that the existence of OVs changes the coordination of the O atoms, the XPS of O 1s may be altered (Xu et al. 2021). Figure 6c exhibited the analysis of OVs species and concentrations by using O 1s spectrum. Three identified peaks at 529.5, 531.3, and 532.1 eV were adopted in the core-level spectra, respectively. In detail, two reflections at 530.8 and 532.1 eV were considered as surface lattice oxygen (OL) and hydroxyl species (OOH) (Zhuang et al. 2017), while another reflection at 531.3 eV could be corresponded to the formation of low oxygen coordination (Liu et al. 2016). The broad area of diverse integrated OV reflection at 531.3 eV was found in FH-NiFe-LDH, demonstrating the maximum concentration of OVs. Remarkably, the integral-area ratios of OV to (OL + OOH + OV) were calculated to 0.3188, 0.3645 and 0.3878 for P-NiFe-LDH, FS-NiFe-LDH, and FH-NiFe-LDH, respectively. The OV/ (OL + OOH + OV) ratios decreased in the order FH-NiFe-LDH > FS-NiFe-LDH > P-NiFe-LDH, which was thoroughly in accordance with the tendency in catalytic performance. Obviously, the catalytic performance can be improved with the increase of OVs concentration. The catalytic activity was positively correlated with OVs concentration, indicating that OVs was the active sites for the promotion of PS activation.
Figure 7a shows the Raman spectra of the three catalysts. In general, three peaks around positions 390, 420, and 460 cm− 1 were assigned to the symmetry of oxygen contact with metal ions in the octahedron (Devi &Srinivas 2017). The observed band around 685 cm− 1 was attributed to the symmetric stretching of Fe/Ni-O bond in tetrahedral coordination (Nandan et al. 2019). The intensity of the peaks may seem to be positively correlated with the amounts of cations in a vibrating substance that produce a signal peak. Moreover, a new band at around 528 cm− 1 appeared from OVs was found in the FH-NiFe-LDH, suggesting that the FH-NiFe-LDH possessed the superior concentration of OVs (Jia et al. 2017, Taniguchi et al. 2009).
To further elucidate that the differences in catalytic activity could be ascribed to the influences of OVs concentrations, EPR spectroscopy was used as a powerful tool to quantificationally analyse the existence and the concentration of OVs in different NiFe-LDH (Li et al. 2018a). As illustrated in Fig. 7b, the composites exhibited a strong symmetrical EPR signal at g = 2.002, which could be attributed to the paramagnetic OVs with 1 + charge state, agreeing with the previous study (Zhang et al. 2016). In addition, the signal intensity revealed that FH-NiFe-LDH possesses more OVs than the other two catalysts, also elucidated by the spectra of XPS. The EPR analysis would confirm that OVs were the active sites that primarily affected its electronic properties and facilitated bonding with PS molecules. It also contributed to the outstanding performance for FH-NiFe-LDH.
Electrochemical tests were conducted to verify the evaluate the electrical properties via morphology-engineering. The CV curves were illustrated in Fig. 8a, compared with other two catalysts, the FH-NiFe-LDH exhibited a larger area of window, evidencing larger capacitance, higher electric current density, and superior reducibility to coordinate a redox reaction. Meanwhile, two symmetrical reflections at -0.5 V and − 0.42 V can be assigned to the redox couples of the valence conversion between Ni(Ⅱ)/Ni(Ⅲ) and Fe(Ⅱ)/Fe(Ⅲ). It was clear that the FH-NiFe-LDH exhibited the most promising peak current, evidencing that the highest catalytic activity. Furthermore, the EIS was also carried out, and Fig. 8b displayed the Nyquist plots of the catalysts. The FH-NiFe-LDH exhibited a relatively smaller electron transfer resistance, demonstrating superior conductivity and electron transfer capacity comparing with P-NiFe-LDH and FS-NiFe-LDH. Therefore, it was confirmed that the introduction of OVs by morphology-engineering could beneficial to enhance electrical conductivity, that is, accelerate the electron transfer between the catalyst and PS, hence dramatically facilitating the catalytic activity.
3.4. Identification of reactive species and possible activation mechanisms
The conversion of Ni(II)/Ni(III) and Fe(II)/Fe(III) occurs during the active catalytic process on the LDH surface. To further explore the mechanism of catalytic active sites involved in the degradation differences among the catalysts, the XPS analysis was applied on the reacted NiFe-LDH. As illustrated the Ni 2p spectra in Fig. 11a and 11b, after the catalytic reaction, the Ni(II) and Fe(II) content of NiFe-LDH accounted for 41.07% and 38.16%, respectively, before the catalytic reaction, while the Ni(III) and Fe(III) content were converted to 35.86% and 26.95%, respectively, after the catalytic reaction, signifying Ni(II)/Ni(III) and Fe(II)/Fe(III) redox reaction process was contained for PS activation. Simultaneously, as displayed in O 1s spectra (Fig. 12c), the relative concentration of OVs decreased after the reaction (20.97% vs 38.78% of the fresh one), evidencing that the OVs in the NiFe-LDH involved in the activation of PS. Of note, the concentration of surface OVs species on the reacted NiFe-LDH can be still maintain a better level.
To uncover the reaction mechanism in the LDH/PS system with different morphologies, radical quenching experiment was performed to investigate the involved reactive species. It is widely recognized that active radicals including •OH, SO4•−, and O2•− are produced in the process of PS activation. Herein, the frequently-used radical scavengers (EtOH and TBA) were chosen to differentiate the contribution of SO4•− and •OH, while p-BQ was used as the scavengers for O2•−. Phenol was used to validate surface-bonded radicals due to its weak polarity and the ability to consume SO4•− and O2•− when approach the solid–liquid interface (Wu et al. 2020). As shown in Fig. 12, the removal efficiency of AO7 was 36.2%, 50.1% and 79.8% without adding any scavenger in the presence of P-LDH, FS-LDH, and FH-LDH, respectively. For the P-LDH/PS system (Fig. 12a), the degradation rate of AO7 decrease to 22.5% and 10.5% in the presence of TBA and EtOH, respectively, manifesting that the existence of both •OH and SO4•−, and notably, SO4•− was more efficient for AO7 removal than •OH. Moreover, in the presence of p-BQ, the decay of AO7 declined to ~ 26.8%, demonstrating that O2•− was also involved in AO7 degradation. Similar phenomena were also observed for FS-LDH/PS and FH-LDH/PS system (Fig. 12b and 12c) in the presence of TBA and EtOH. However, as clearly observed, the inhibitory effect of p-BQ was significantly enhanced in FS-LDH/PS and FH-LDH/PS system, indicating that O2•− play a dominant role. The experimental results further confirmed that tailoring the morphology of LDH could result in different OVs concentrations. Moreover, the removal efficiency of AO7 decreased to 9.5%, 18.2 and 20.5% in the presence of phenol, demonstrating that the reactive species could be originally generated on the surface of LDH and then some of it would diffuse into the liquid phase in LDH/PS system. In summary, three kinds of ROS including •OH, SO4•−, and O2•− corporately contribute to the AO7 elimination in LDH/PS system, and the degradation reaction belongs to surface reaction.
Based on the above detailed analysis, we systematical proposed the overall mechanism of PS activation in LDH/PS system (Fig. 13). In brief, PS molecules diffuse to the surface of NiFe-LDH and adsorb on the surface-active site of catalyst. Then, PS molecules are activated via electron transfer and produce a series of ROS to attack the contaminant. It is worth noting that because of the abundant diffusion pathway combined with nano-confinement and nano-enrichment, the unique hollow structure with 2D/3D homojunction in FH-NiFe-LDH can deeply boost surface adsorption and the mass diffusion. Meanwhile, with the existence of OVs in FH-NiFe-LDH, the efficiency of electron transfer to PS can be facilitated, thus enhancing PS activation.

According to aforementioned analysis in association with the reported accounts (Bu et al. 2021, Guo et al. 2020, Wu et al. 2020), the PS activation and ROS generation mechanism can be detailedly displayed in Eqs. (1)-(12). Expressly, the PS molecule firstly accepts an electron from Ni(II) or Fe(II) on the surface of NiFe-LDH to produce SO4•−. The redox cycle could accomplish via the reaction between Ni(III)/Fe(III) and S2O82− for consecutive production of SO4•−. Then, SO4•− could further convert H2O to •OH. (Eqs. (6)– (8)). Notably, the OVs also play a vital role in facilitating the production of more active sites for PS activation. Impressively, NiFe-LDH featuring abundant OVs could react with O2 to generate O2•− (Eq. (1)). During this process, OVs could accelerate the electron transfer to promote the generation of ROS, meanwhile, facilitate the redox cycle of Fe(III)/Fe(II) and Ni(III)/Ni(II) (Wang et al. 2022). Finally, the active radicals attacked the AO7 and quickly eliminated AO7.