Enhanced photocatalysis activity of ferroelectric KNbO₃ nanofibers compared with antiferroelectric NaNbO₃ nanofibers synthesized by electrospinning

Abstract

Perovskite-type alkaline niobate nanofibers were prepared by electrospinning. KNbO₃ nanofibers with lower BET exhibit higher photocatalytic efficiency than NaNbO₃ with similar phase structure and morphology. The internal electric field, which does not exist in antiferroelectric NaNbO₃, induced by spontaneous polarization in ferroelectric KNbO₃ can promote the separation and accelerate the movement of photogenerated charge carriers.

---

Introduction

Since TiO₂ was first reported by Honda et al.¹ to display photocatalytic activity in hydrogen evolution, semiconductor photocatalysts have been found to have important applications in clean hydrogen energy generation and the treatment of organic pollutants. The study of novel semiconductor photocatalysts, such as Ti-, Nb-, and Ta-based photocatalysts, has been an important issue over the past few decades.^2–4 Among these materials, perovskite-type oxides are attractive because of their high stability, good availability and low toxicity under light illumination. It is well known that the efficiency of photocatalysts is determined not only by their band gap but also by their ability to separate photogenerated charge carriers. Ferroelectric materials exhibit considerable photocatalytic performance because of the existence of an internal electric field created by spontaneous polarization. The internal electric field can effectively promote the separation of electrons and holes and accelerate the movement of electrons and holes in opposite directions.⁵ Thus, ferroelectric materials, such as PSZT⁶ and BiFeO₃,⁷ have attracted a lot of attention in recent decades. However, these materials unfortunately contain large quantities of the toxic elements such as Pb and Bi. Environmentally friendly ANbO₃ (A = K, Na) has attracted substantial interest among researchers and equipment designers because of its unique physical properties.^8,9 Although a few studies have focused on the photocatalytic activity of ANbO₃ powders,^10–13 they do not shed any light on the correlation between the enhanced photocatalytic activity and the internal electric field in the ferroelectric material.

In the light of scientific interest and potential commercial applications, the photocatalytic activities of previously reported ANbO₃ micropowders are still not high enough because of their low Brunauer–Emmett–Teller (BET) surface areas.^10–12 Recently, one-dimensional (1D) nanostructures are expected to result in considerably higher photocatalytic activities due to their large surface-volume ratio and high surface activity.^14,15 Compared to hydrothermal methods¹⁶ and molten salt methods,¹⁷ electrospinning is the simplest and most versatile technique capable of generating nanofibers. Moreover, the electrospun nanofibers have a high aspect ratio, controllable fiber diameters and a precise stoichiometric chemical composition.¹⁸

In this study, ANbO₃ nanofibers were prepared by a sol–gel-based electrospinning technique. We applied ANbO₃ nanofibers in the degradation of rhodamine B (RhB), which is a preventative dye pollutant used in the determination of photocatalytic activity. To the best of our knowledge, this study is the first report on ANbO₃ electrospun nanofibers in photocatalysis. Our experimental results indicate that higher photocatalytic activity could be obtained in ferroelectric KNbO₃ nanofibers compared with antiferroelectric NaNbO₃ nanofibers because of the internal polar field induced by spontaneous polarization. The synthesis and characterization of KNbO₃ and NaNbO₃ nanofibers are shown in the ESI.†

Results and discussion

KNbO₃ and NaNbO₃ binder-containing fibers after drying at 80 °C for 5 h in a vacuum oven are abbreviated as KN-NF and NN-NF, respectively. Fig. 1b displays the Fourier-Transform Infrared (FTIR) spectra of KNbO₃ samples and raw materials. The characteristic band at ca. 1620 cm⁻¹ corresponds to the C=O stretching vibration of acetylacetone. 2-Methoxyethanol exhibits two C–O stretching modes, one at ca. 1124 cm⁻¹, which is characteristic of an ether with a –CH₂–O–CH₂– group, and another at ca. 1066 cm⁻¹, which is characteristic of a primary alcohol.¹⁹ It is difficult to detect these three peaks in the FTIR spectra of KN-NF and NN-NF, which illustrates that the solvent volatilizes completely. The double bands at ca. 1105 and 1066 cm⁻¹ of niobium ethoxide do not appear in the spectra of KN-NF and NN-NF, which demonstrates that niobium ethoxide is decomposed to niobium hydroxide. The bands at ca. 1574 and 1425 cm⁻¹ are assignable to stretching vibrations and antisymmetric stretching vibrations of carboxyl groups, respectively, which appear in the FTIR spectra of KN-NF and NN-NF and demonstrate that alkaline acetates exist in KN-NF and NN-NF.

Fig. 1 (a) TG-DSC curves of KN-NF, (b) FTIR spectra of raw materials and KN-NF calcined at different temperatures, (c) XRD patterns of KN-NF calcined at different temperatures, (d) TG-DSC curves of NN-NF, (e) FTIR spectra of raw materials and NN-NF calcined at different temperatures, and (f) XRD patterns of NN-NF calcined at different temperatures.

KN-NF was calcined at 500 °C, 600 °C, 700 °C, and 800 °C in air, which are abbreviated as KN-NF-500, KN-NF-600, KN-NF-700, and KN-NF-800, respectively. In the KN-NF thermogravimetric-differential scanning calorimetry (TG-DSC) curves, as shown in Fig. 1a, the endothermic peak at 130 °C in the DSC curve should be attributed to the decomposition of CH₃COOK. Subsequently, about 30% weight loss is observed at 280 °C in the TG curve accompanied by a corresponding endothermic peak in the DSC curve, which results from the combustion of PVP. The weak endothermic peak at ca. 400 °C should be assigned to the elimination of water of crystallization in niobium hydroxide, because a part of the water of crystallization is difficult to be eliminated until 400 °C. The endothermic peak at ca. 580 °C should be ascribed to the crystallization of KNbO₃. The bands at about 750 cm⁻¹ and 850 cm⁻¹ in FTIR spectrum of the KN-NF-500 are characteristic bands of the resident organism. All the characteristic peaks that correspond to organic functional groups in the FTIR spectrum have disappeared for KN-NF-600. The band around 500 cm⁻¹ can be assigned to the edge-shared NbO₆ octahedron and the broad, strong band centered at ca. 623 cm⁻¹ represents O–Nb–O stretching vibrations in the corner-shared NbO₆,¹² which demonstrates that a perovskite structure forms when KN-NF is calcined above 600 °C, in agreement with the X-ray diffraction (XRD) patterns shown in Fig. 1c. KNbO₃ could be well indexed to the orthorhombic phase with JCPDS no. 71-0946 without a second phase. The endothermic peak at ca. 670 °C should be ascribed to the grain growth of KNbO₃, because no significant change appears in the FTIR spectra and XRD patterns. Because the crystallinity of KN-NF-600 is low, as seen from the Raman spectra (Fig. S1a†), the calcination temperature is designated to be 700 °C. The weak endothermic peak at ca. 860 °C should be attributed to the fracture of the nanofibers.

The TG-DSC curve of NN-NF shows similar features to that of KN-NF except for a few differences. First, the combustion of PVP occurs from ca. 200 to 380 °C accompanied by an endothermic peak in the DSC curve. Second, the decomposition of CH₃COONa is assigned to the endothermic peak at ca. 390 °C in the DSC curve because CH₃COONa decomposes at a higher temperature than CH₃COOK. Third, the crystallization of NaNbO₃ should occur at ca. 580 °C accompanied by a corresponding endothermic peak. To determine the calcination temperature, NN-NF was calcined at 400 °C, 500 °C, 600 °C, and 700 °C in air, which are abbreviated as NN-NF-400, NN-NF-500, NN-NF-600, and NN-NF-700, respectively. From the FTIR spectra shown in Fig. 1e, the bands of organic functional groups do not disappear until calcination at 500 °C. There are two different phases of NaNbO₃, antiferroelectric (Pbma, JCPDS no. 89-8957) and ferroelectric (P2₁ma, JCPDS no. 82-0606).²⁰ The XRD patterns of NN-NF calcined above 500 °C could be well indexed to the space group Pbma, as displayed in Fig. 1f, which demonstrates that NaNbO₃ nanofibers exhibit antiferroelectric properties. The characteristic peaks of NbO₆ octahedra distinctly appear when NN-NF is calcined at 600 °C, as seen from the Raman spectra (Fig. S1b†). Therefore, the best calcination temperature for the treatment of NN-NF is 600 °C.

From the scanning electron microscopy (SEM) images, shown in Fig. 2a1 and b1, it can be inferred that KN-NF and NN-NF have relatively uniform diameters of ca. 500 nm. After calcining, the surface of the nanofibers is not quite smooth because of grain growth at high temperature, as shown in Fig. 2a2 and b2. These nanofibers compactly stack together with small nanoparticles, as shown in the transmission electron microscopy (TEM) bright-field images presented in Fig. 2a3 and b3. The nanofibers are polycrystalline as inferred from the selected-area electron diffraction (SAED) patterns with several sets of spot patterns obtained from partial nanofibers, as shown in the insets of Fig. 2a3 and b3. Moreover, the regular atomic arrangement, as depicted in high-resolution transmission electron microscopy (HRTEM) images in Fig. 2a4 and b4, indicates that the synthesized nanofibers have good crystallinity. The crystal lattice spacings of 0.406 nm and 0.285 nm correspond to the (011) and (200) crystal planes of orthorhombic KNbO₃, respectively, which agree well with the results obtained from the SAED pattern shown in the inset. The angle between adjacent spots labeled in the SAED pattern is 90°, which is identical to the theoretical value of the angle between the (011) and (200) planes of orthorhombic KNbO₃. Analogously, the crystal lattice spacings of 0.391 nm and 0.389 nm in NN-NF-600 are well ascribed to the (101) and (040) crystal planes of orthorhombic NaNbO₃, respectively. To visualize the grain size of nanofibers, dark-field TEM (DFTEM) images are illustrated in Fig. 2. The DFTEM images in Fig. 2a5 and a6 were obtained using the A and B reflections shown in the inset of Fig. 2a3, respectively. Typically, the nanofibers are stacked one by one with small nanoparticles together with a grain size of about 200 nm. Similar conclusions apply to NN-NF-600. Furthermore, energy-dispersive X-ray spectroscopy (EDS) mapping indicates that all the chemical elements are distributed homogeneously, as shown in Fig. S2.†

Fig. 2 SEM images of (a1) KN-NF; (a2) KN-NF-700; (a3) TEM image of KN-NF-700, inset: the corresponding SAED pattern; (a4) HRTEM image of KN-NF-700, inset: the corresponding SAED pattern; (a5 and a6) DFTEM images obtained using the A and B reflections in Fig. a3; SEM images of (b1) NN-NF; (b2) NN-NF-600; (b3) TEM image of NN-NF-600, inset: the corresponding SAED pattern; (b4) HRTEM image of NN-NF-600, inset: the corresponding SAED pattern; (b5 and b6) DFTEM images obtained using the C and D reflections in Fig. b3.

The photocatalytic activity of the as-prepared samples was evaluated using the degradation of an RhB aqueous solution under UV-vis light illumination. The change in the relative concentration of the RhB aqueous solution as a function of the irradiation time (t) is shown in Fig. 3a. For the blank experiment, the degradation of the RhB aqueous solution under UV-vis light illumination is negligible. The concentration of the RhB aqueous solution decreases with an increase in reaction time, which demonstrates that the RhB molecule is decomposed. The decomposition rate declines because the concentration of the RhB aqueous solution decreases. To evaluate the photoreactivity quantitatively, the reaction rate constants were calculated and are depicted in Fig. 3b. The photodegradation of RhB in the presence of KNbO₃ and NaNbO₃ can be considered as a pseudo-first-order reaction¹³ and its kinetics can be expressed as follows:

C = C₀e^−kt

where k is the degradation rate constant and C₀ and C are the initial concentration of RhB and that at time t, respectively. Fig. 3b shows the relationship between ln(C₀/C) and t with the corresponding fitting curves. It was found that the degradation rate constants with KN-NF-700 and NN-NF-600 were 0.01281 min⁻¹ and 0.00982 min⁻¹, respectively, although NN-NF-600 has a larger BET specific surface area, as shown in Table 1. When the reaction rate constant was divided by the BET specific surface area, it was inferred that KN-NF-700 exhibits better photocatalytic activity than NN-NF-600, as shown in the right-hand graph of Fig. 3c. It is well known that the efficiency of a photocatalyst is determined mainly by its band gap energy (E_g) and ability to separate photogenerated charge carriers (electrons and holes). E_g is estimated from the UV-vis diffuse reflectance spectra in Fig. 3d as follows. The linear part of the spectral line was extended as the black dotted line in Fig. 3d and the intersection of the black dotted line and x-axis was determined. Then, E_g was calculated by dividing 1240 by the wavelength of the intersection. It is obvious that the E_g of NN-NF-600 is slightly larger than that of KN-NF-700, as shown in Table 1, resulting in the low effective utilization of irradiation of NN-NF-600. Moreover, KNbO₃ with an orthorhombic phase is ferroelectric, whereas NaNbO₃ with an orthorhombic phase is antiferroelectric. To determine the ferroelectric properties of the nanoscale KN-NF-700 prepared in this study, its effective piezoelectric coefficient d₃₃* was measured by a modified atomic force microscopy (AFM) system. The Si cantilever tip was Pt-coated and conductive. KN-NF-700 was dropped on Pt-coated silicon (Pt/Ti/SiO₂/Si) substrates after being ultrasonically dispersed in ethanol. An individual KN-NF-700 nanofiber was investigated by AFM in the experiment, as shown in Fig. 4a. Then, the cantilever tip was kept fixed above the point of interest of the nanofiber and an alternating current voltage was applied from −9 to 9 V while recording the piezoelectric displacement by laser. The result of the measurement of d₃₃* is shown in Fig. 4b. The individual nanofiber exhibited piezoelectric properties, as shown by the typical well-shaped displacement-voltage “butterfly” loop, which indicates that KNbO₃ nanofibers are ferroelectric, although the grain size is very small. As shown in the schematic diagram in Fig. 5, a domain exists in the KNbO₃ grains and an internal electric field is induced by spontaneous polarization. When the ferroelectric photocatalyst is illuminated by UV-vis light with photon energy higher than E_g, electrons in the valence band can be excited to the conduction band with the simultaneous generation of the same amount of holes that are left behind. The internal electric field in KNbO₃ nanofibers promotes the separation of electrons and holes and accelerates electrons and holes to move in opposite spatial directions. These phenomena are analogous to the behavior of charge carriers under an external electric field. Bastard et al.²¹ reported that the conduction band and valence band electrons in GaAs quantum wells can be spatially separated by applying an external electric field. However, there is no internal electric field because NaNbO₃ is antiferroelectric. Photogenerated electrons and holes will more easily recombine before they migrate to the surface, leading to depressed photocatalytic activity. Thus, it is reasonable to suppose that photogenerated electron–hole pairs in ferroelectric materials can be effectively separated by the internal polar field that is caused by spontaneous polarization, and hence the photocatalytic activity can be enhanced. In addition, the XRD and XPS patterns of the samples after photocatalytic reactions are nearly the same as those of the fresh photocatalyst (Fig. S4 and S5†), which indicates high stability against chemical reaction during the photocatalytic reaction. KN-NF-700 and NN-NF-600 can be recycled by filtration. Therefore, KNbO₃ and NaNbO₃ nanofibers may be attractive candidates as photocatalysts.

Fig. 3 (a) Photodegradation curves of RhB aqueous solution with photocatalyst, (b) linear least-squares fitting of ln(C/C₀) versus irradiation time plots, (c) reaction rate constant and reaction rate constant per unit area of the photocatalyst, (d) UV-vis diffuse reflectance spectra of KN-NF-700 and NN-NF-600.

Table 1 BET surface areas and band gaps of the as-prepared samples

Sample	BET surface area (m^2 g^−1)	Band gap (eV)
KN-NF-700	10.098	3.626
NN-NF-600	15.668	3.804

---

Fig. 4 (a) Top view of an individual KNbO₃ nanofiber, (b) displacement–voltage curve (black line) and d₃₃-voltage curve (red line).

Fig. 5 Schematic diagram of the domain structure and spontaneous polarization direction in a unit cell in (a) KN-NN-700 and (b) NN-NF-600.

Conclusion

KNbO₃ and NaNbO₃ nanofibers with uniform diameters of ca. 500 nm were prepared by sol–gel-based electrospinning. Perovskite-type KNbO₃ and NaNbO₃ nanofibers were well crystallized after being calcined at 700 and 600 °C, respectively. The nanofibers are composed of small nanoparticles that are compactly stacked one by one, as revealed by DFTEM analysis. Ferroelectric KNbO₃ nanofibers, whose ferroelectric properties were confirmed by modified AFM, with lower BET surface areas exhibit higher photocatalytic activity than antiferroelectric NaNbO₃ nanofibers with a similar phase structure and micromorphology. It was found that spontaneous polarization in the ferroelectric domain can induce an internal polar electric field, which can promote the separation of photogenerated electrons and holes and accelerate the movement of electrons and holes in opposite directions. In addition, KNbO₃ and NaNbO₃ nanofibers with good availability and low toxicity display high stability under irradiation. Therefore, there is a potential to develop nanoscale multifunctional devices because KNbO₃ has applications in actuators, sensors, transducers, and capacitors.

Acknowledgements

The study was supported by the Ministry of Sciences and Technology of China through the National Basic Research Program of China (973 Program 2015CB654604), the National Natural Science Foundation of China for Creative Research Groups (Grant No. 51221291), and the National Natural Science Foundation of China (Grant No. 51272123, 51332002), and also supported by CBMI Construction Co., Ltd W. Hao thanks the National Natural Science Foundation of China for partial support to this study (Grant Nos. 51272015, 51472016).

References

1. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
2. M. R. Hoffmann, S. T. Martin, W. Y. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69–96.
3. Q. J. Xiang, J. G. Yu and M. Jaroniec, Chem. Soc. Rev., 2012, 41, 782–796.
4. Y. Inoue, Energy Environ. Sci., 2009, 2, 364–386.
5. J. Jiang, K. Zhao, X. Xiao and L. Zhang, J. Am. Chem. Soc., 2012, 134, 4473–4476.
6. Y. Inoue, K. Sato and K. Sato, J. Chem. Soc., Faraday Trans. 1, 1989, 85, 1765–1774.
7. F. Gao, Y. Yuan, K. Wang, X. Chen, F. Chen, J.-M. Liu and Z. Ren, Appl. Phys. Lett., 2006, 89, 102506.
8. P. Günter, Phys. Rep., 1982, 93, 199–299.
9. J. Richter, A. Steinbrück, M. Zilk, A. Sergeyev, T. Pertsch, A. Tünnermann and R. Grange, Nanoscale, 2014, 6, 5200–5207.
10. G. Q. Li, T. Kako, D. F. Wang, Z. G. Zou and J. H. Ye, J. Phys. Chem. Solids, 2008, 69, 2487–2491.
11. L. S. Yan, J. Zhang, X. M. Zhou, X. X. Wu, J. Y. Lan, Y. S. Wang, G. Liu, J. G. Yu and L. J. Zhi, Int. J. Hydrogen Energy, 2013, 38, 3554–3561.
12. T. T. Zhang, K. Zhao, J. G. Yu, J. Jin, Y. Qi, H. Q. Li, X. J. Hou and G. Liu, Nanoscale, 2013, 5, 8375–8383.
13. J. Y. Lan, X. Zhou, G. Liu, J. G. Yu, J. C. Zhang, L. J. Zhi and G. J. Nie, Nanoscale, 2011, 3, 5161–5167.
14. Y. Wang, X. Zhan, F. Wang, Q. Wang, M. Safdar and J. He, J. Mater. Chem. A, 2014, 2, 18413–18419.
15. N. C. Hildebrandt, J. Soldat and R. Marschall, Small, 2015, 11, 2051–2057.
16. Y. Xu, Q. Yu and J. F. Li, J. Mater. Chem., 2012, 22, 23221–23226.
17. L. Li, J. Deng, J. Chen, X. Sun, R. Yu, G. Liu and X. Xing, Chem. Mater., 2009, 21, 1207–1213.
18. X. F. Lu, C. Wang and Y. Wei, Small, 2009, 5, 2349–2370.
19. C. Guha, J. M. Chakraborty, S. Karanjai and B. Das, J. Phys. Chem. B, 2003, 107, 12814–12819.
20. T. Y. Ke, H. A. Chen, H. S. Sheu, J. W. Yeh, H. N. Lin, C. Y. Lee and H. T. Chiu, J. Phys. Chem. C, 2008, 112, 8827–8831.
21. G. Bastard, E. Mendez, L. Chang and L. Esaki, Phys. Rev. B: Condens. Matter Mater. Phys., 1983, 28, 3241.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra13680f

---