A simple UV-ozone surface treatment to enhance photocatalytic performance of TiO₂ loaded polymer nanofiber membranes

Abstract

Homogeneously dispersed titanium dioxide loaded polyacrylonitrile nanofiber membranes with increased active mass loading, Ti³⁺ surface defects and hydrophilicity were fabricated by combining electrospinning and UV-ozone surface treatment. The photocatalytic activity improved by a factor of ∼2 and the kinetics of photodegradation switched from pseudo-first order to pseudo-second order with increasing TiO₂ content with a maximum rate constant of 20.7 h⁻¹.

---

Introduction

TiO₂ and ZnO are very promising photocatalysts for utilizing solar energy for degradation of organic dyes and bacteria from wastewater.^1–3 These photocatalysts absorb the right energy photons, generate excitons and cause the oxidation of the biomaterial that accumulates on its surface.⁴ The charge carrier generation, transfer and recombination affect the overall photocatalytic activity and need to be understood individually for optimum performance.⁵

Nanoparticles of these photocatalysts can provide a large number of active sites per unit area but their natural tendency to agglomerate restricts their efficiency. Immobilization of these nanoparticles on a support with homogeneous dispersion is an effective method to increase their photocatalytic performance.⁶ One such support is nanofiber membranes by electrospinning which have several attributes suitable for water filtration and purification such as interconnected open pore sizes in nm to μm range, high permeability, and large surface to volume ratio allowing greater adsorption of the pollutants.^7,8 However, bio-fouling of these membranes by bacteria and viruses is a major limiting factor which determines the efficiency and life of the filter.⁹ Recently, the focus has been on synthesizing self-cleaning filters by immobilizing photocatalyst nanoparticles on selective polymer^10–13 and carbon^14–16 nanofibers. The electrospun nanofiber mesh will carry out the filtering part, isolating bacteria and most viruses, whereas the photocatalyst nanoparticles will oxidize them to CO₂ and H₂O. The need for a safe, reliable and low cost method for efficient removal of organic pollutants can potentially be addressed by such multifunctional nanofiber membranes.

Ding et al.¹⁰ made polyimide/ZnO fibers by electrospinning using a direct ion-exchange process which exhibited pseudo-first order photodegradation of methylene blue (MB) with rate constants in 0.6–1.7 h⁻¹ range. Ramasundaram et al.¹³ used a combination of electrospinning of polyvinylidene fluoride (PVDF) and electrospraying of TiO₂/DMF solution to fabricate PVDF/TiO₂ hybrid nanofibers in which TiO₂ nanoparticles were non-uniformly agglomerated on the surface. Other approaches include composite TiO₂ nanofibers¹⁷ but their fragile nature is not suitable for pressure driven applications. Recently, a promising air jet spinning technique has been used to fabricate TiO₂/polymer composite membranes¹⁸ with mean fiber diameters and pore sizes of 505–901 nm and 1.58–5.12 μm, respectively, but the process needs optimization to achieve uniform TiO₂ dispersion. Moreover, surface defects such as Ti³⁺ on TiO₂ and oxygen vacancies play a vital role.¹⁹ A surface plasma treatment on TiO₂ resulted in enhanced efficiency in dye sensitized solar cells^20,21 due to efficient charge transport. Besides these TiO₂ based composites, various 1D spatially ordered architectures²² and graphene based composite photocatalysts²³ are being intensively investigated for solar every conversion applications.

Enhanced photocatalytic performance of these membranes require optimum morphology, distribution, surface chemical states of TiO₂ and improved hydrophilicity. Here, we report a simple and low cost UV-ozone surface treatment to enhance these factors and obtain improved photocatalytic performance of TiO₂/polymer composite nanofiber membranes. These offer great promise as anti-fouling filters for water remediation.

Experimental

Polyacrylonitrile (PAN, Sigma-Aldrich, M_w = 150 000) was stirred for 12 h in N,N-dimethyl formamide (DMF) at room temperature to form a 9 wt% solution. Separately, TiO₂ agglomerates (P25, Sigma-Aldrich, particle size 21 nm) were ultrasonicated for 4 hours in DMF to homogeneously disperse them. The two solutions were then blended using ultrasonication for additional 4 hours. Electrospinning at 15 kV with 15 cm needle-to-collector distance yielded nanofiber membranes. The composition of the blend solutions were adjusted to obtain 9–47 wt% TiO₂/PAN nanofibers as confirmed by thermal gravimetric analysis (TGA). In this paper, the samples will be identified by their wt% TiO₂. The nanofiber membranes were then exposed to UV-ozone for 15 minutes. Nanofiber sizes, surface morphology and TiO₂ dispersion were characterized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM) equipped with energy-dispersive X-ray spectroscopy (EDS). The oxidation states of Ti, before and after UV-ozone treatment, were characterized by studying the L_2,3 energy-loss near edge structure (ELNES) by electron energy loss spectroscopy (EELS). The membrane hydrophilicity was characterized by water contact angle measurement. Photocatalytic performance of the membranes were evaluated by photodegradation of 10 ppm MB solution by 10 mg of composite nanofiber membranes under UV irradiation and monitored by UV-Vis spectrophotometry. A calibration curve of 0–20 ppm MB solution by serial dilutions was used as a reference. Adsorption of MB on nanofiber surfaces were taken into account prior to all photodegradation experiments by establishing an adsorption–desorption equilibrium. 17 and 47 wt% samples, without UV-ozone treatment, were also tested for comparison.

Results and discussion

PAN was selected because it can be easily electrospun with very smooth surfaces and minimum bead formation. Moreover, it's a widely used precursor for carbon nanofibers (CNFs) thus expanding the scope of this study to making TiO₂/CNFs composites with enhanced photocatalytic performance.²³ In fact, TiO₂–carbon composites have shown similar results to TiO₂–graphene composites on adsorption of pollutants, light absorption intensity and charge carrier lifetimes.²³ Fig. 1a is a representative SEM image showing the interpenetrating random network of 25 wt% sample with interstices on the order of few μm which can trap most of the microorganisms. Fig. 1b shows the diameter distribution of all the samples which is characteristically log-normal. The mean diameter increases from 225 to 335 nm with increasing TiO₂ due to increased blend solution viscosity and hindrance in nanofiber thinning during later stages of electrospinning. Fig. 1c and d shows SEM images of the 47 wt% nanofibers, before and after a 15 min UV-ozone treatment, respectively, which clearly indicates the enhanced exposure of TiO₂ in the UV-ozone treated sample. UV-ozone combines decomposition of organic materials by UV radiations and strong oxidation by atomic oxygen during formation/decomposition of ozone. Here, we used it to selectively etch PAN and expose TiO₂ on the nanofiber surface. The TiO₂ active mass loading can be adjusted by varying the TiO₂ loading in the blend solution as shown in Fig. 1e and f for 40 and 25 wt% samples, respectively. A PAN/TiO₂ blend solution by stirring only, without ultrasonication, results in agglomerated TiO₂ non-uniformly dispersed in PAN nanofibers as shown in Fig. S1 of the ESI.† Thus, ultrasonication of the blend solution gives excellent dispersion of TiO₂ in PAN and a post UV-ozone treatment enhances the exposure of TiO₂ as depicted in a high resolution SEM image in the graphical abstract.

Fig. 1 (a) A representative SEM image of the 25 wt% TiO₂/PAN nanofiber membrane (inset: larger magnification), (b) nanofiber diameter distribution of all the samples which is characteristically log-normal, (c) a 47 wt% nanofiber before UV-ozone treatment and (d) 47 wt% after UV-ozone treatment, (e) 40 wt%, (f) 25 wt% nanofibers after UV-ozone treatment.

The amount of TiO₂ loading in PAN nanofibers was estimated by TGA as the residual weight after complete burning of PAN. Fig. 2 shows the simultaneous DSC and TGA scans of all the samples. The first weight loss occurs at ∼300 °C which coincides with the exothermal DSC peak of PAN stabilization reaction.²⁴ The samples remain relatively stable between 300 and 450 °C. After that, the thermal degradation of PAN occurs as shown by a second exothermic peak. This exothermic event shifts to lower temperatures (by as much as ∼70 °C for 47 wt% sample) as the amount of TiO₂ increases due to the effect of large thermal mass of TiO₂ which aids in the early decomposition.

Fig. 2 DSC (top) and TGA (bottom) scans of TiO₂/PAN nanofibers. The weight percent of TiO₂ in the samples is calculated from TGA as the residual weight at 800 °C after complete thermal degradation of PAN in air.

To verify the morphology and distribution of nanosized TiO₂ on PAN nanofibers, TEM analysis was carried out. Fig. 3a and b shows the bright field TEM images and EDS elemental maps of Ti (green) and O (red) for a 40 wt% nanofiber before and after UV-ozone treatment, respectively. The TiO₂ nanoparticles are finely distributed in PAN nanofibers and are ∼20 nm in size. Moreover, the active particles of TiO₂ are not only located at the surface but are very well embedded in nanofiber bulk as well.

Fig. 3 Bright field TEM image, Ti and O elemental EDS map and water contact angle of 40 wt% nanofiber sample (a) before and (b) after UV-ozone treatment. (c) EELS edge spectra showing Ti L-edges (L₃ and L₂) and O K-edge before and after UV-ozone treatment (inset: zoomed-in of L₃ and L₂ edges).

Several factors affect the photocatalytic activity such as surface chemical states, electron–hole recombination, charge separation and transfer kinetics. Moreover, the dye adsorption is affected by the hydrophilicity of the nanofiber membranes. These factors are now discussed in context of the UV-ozone treatment. It's reported that radio frequency (RF)²⁰ and oxygen²¹ plasma treatment on TiO₂ increases hydrophilicity and Ti³⁺ surface sites resulting in reduced electron–hole recombination rate. Here, the effect of UV-ozone treatment on the chemical states of TiO₂ nanoparticles is characterized by EELS by studying the L_2,3 energy-loss near edge structure (ELNES) in TiO₂ and comparing the L₃/L₂ intensity ratio with the expected value of 0.8.²⁵ L_2,3 near edge fine structure correspond to transitions to unoccupied 3d states and is an indicator of the oxidation state (higher L₃/L₂ ratio corresponds to higher concentration of Ti³⁺).²⁶ The ratio was estimated by using the integrated area above the step level and under each intensity peak (Fig. 3c). The step level is extrapolated from the tail ends of each peak. This ratio increased from 0.76 (untreated) to 0.82 (UV-ozone treated) indicating an increase in the Ti³⁺ oxidation state.²⁷ Thus, UV-ozone treatment results in a surface reaction of TiO₂ (Ti⁴⁺) to Ti₂O₃ (Ti³⁺). This generates oxygen vacancies and excess electrons in Ti at surface which aids the charge transport between dye and photocatalyst and reduces electron–hole recombination rate.²⁷ Kuznetsov et al. reported a ∼14% conversion of Ti⁴⁺ to Ti³⁺ under UV irradiation.²⁸ Similar results have been reported for dye sensitized solar cells.^20,21 Moreover, electrochemical impedance spectroscopy showed that charge transfer resistance is reduced in UV-ozone treated TiO₂.²¹ To determine the change in the hydrophilic nature of we have measured the water contact angle of the nanofiber membrane. The water contact angle of the nanofiber membrane (insets in Fig. 3a and b) decreased from 29.6° to 18.6° after UV-ozone treatment resulting in increased hydrophilicity which aids in the adsorption of dye.

Finally, the photocatalytic performance of the TiO₂/PAN composite nanofiber membranes were evaluated by photodegradation of MB under UV irradiation and monitored by determining the decrease in absorbance by UV-Vis spectrophotometry. However, the decrease in absorbance is also due to adsorption of MB molecules on PAN and TiO₂.²⁹ In this study, the mixture of TiO₂/PAN nanofibers in MB solution were shaken for 2 hours in dark until an adsorption–desorption equilibrium was attained as confirmed by a steady state absorbance value of the mixture. Xu et al.²⁹ reported 30 min to achieve adsorption–desorption equilibrium for TiO₂ nanoparticles in MB solution. Fig. S2 of the ESI† shows the UV-Vis spectrums of MB solutions taken at different times of photodegradation under UV irradiation. There is negligible change in the absorbance of the control sample. The photocatalytic efficiency of nanofiber membranes is quantitatively shown in Fig. 4a as C/C_o versus time plot where C_o is the MB solution concentration after the adsorption–desorption equilibrium has been achieved. The initial rate of photodegradation increases with TiO₂ loading and UV-ozone treatment further enhances it due to the increased exposure of TiO₂, increase Ti³⁺ surface states, and increase hydrophilicity. The total degradation after 60 min of exposure is 92, 93, 95, 97 and 98% for 9, 17, 25, 40, and 47 wt% TiO₂/PAN samples, respectively. Fig. 4a also shows the performance of as-prepared 17 and 47 wt% samples (without UV-ozone treatment) which takes about twice as much time to photodegrade.

Fig. 4 (a) Photocatalytic performance of TiO₂/PAN nanofiber membranes plotted as C/C_o versus time, (b) ln(C) versus t plot for 9, 17, and 25 wt% and (inset) 1/C versus t plot for 40 and 47 wt% membranes, and (c) photocatalytic performance of 45 wt% membrane over 6 cycles.

It is relevant to remark that the UV-ozone treatment didn't cause any significant change in the mechanical strength of the nanofiber membranes as they all were handled easily without any damage. In fact, TiO₂ will act as reinforcement in the flexible PAN matrix to make a strong and tough multifunctional composite membrane which can withstand the swelling pressures generated during water uptake.³⁰ Fig. 4b, which plots ln(C) versus t, shows that the photodegradation follows a pseudo-first order kinetics for 9, 17, and 25 wt% samples with rate constants of 3.9, 4.2, 4.7 h⁻¹, respectively. Interestingly, the kinetics switches to pseudo-second order for 40 and 47 wt% samples with rate constants of 18.8 and 20.7 h⁻¹, respectively, as shown in the plot of 1/C versus t in the inset of Fig. 4b. Xu et al.²⁹ reported a detailed study on photodegradation of MB by TiO₂ powder and observed that the concentration of TiO₂ and MB has a complex impact on the overall kinetics with both desorption and degradation exhibiting pseudo-first order kinetics. He observed a desorption dominated slight increase (less than 10%) in the absorbance of 20 ppm MB solution within first 5 min of UV irradiation attributed to a shift in adsorption equilibrium due to altered electric charges on MB functional group and TiO₂ surface. Thus, our results from a much longer time scale are expected to be negligibly affected by this initial desorption phenomenon. Strong adhesion of TiO₂ to nanofiber membrane is critical for long term and multiple cycles use. The 40 wt% membrane showed a consistent performance over 6 cycles of photodegradation as shown in Fig. 4c. This suggests a strong physical interlocking of TiO₂ on the nanofiber surface by the encapsulating PAN without using any interfacial chemistry based methods to enhance the adhesion. No leaching of TiO₂ in the MB solution was observed. This also implies Ti³⁺ concentration doesn't change over multiple cycles of use.

Conclusions

The simple process of electrospinning from an ultrasonicated PAN/TiO₂ blend solution followed by controlled UV-ozone treatment results in flexible and strong composite nanofiber membranes with excellent photocatalytic activity and reliability over multiple cycles. This enhanced photocatalytic activity is attributed to the enhanced active mass exposure, increased Ti³⁺ surface defects and increased hydrophilicity. At higher TiO₂ concentration the photodegradation kinetics switches to pseudo-second order.

References

1. X. Wang, J. C. Yu, C. Ho, Y. Hou and X. Fu, Langmuir, 2005, 21, 2552–2559.
2. H. Lachheb, E. Puzenat, A. Houas, M. Ksibi, E. Elaloui, C. Guillard and J. Herrmann, Appl. Catal., B, 2002, 39, 75–90.
3. J. Yu and X. Yu, Environ. Sci. Technol., 2008, 42, 4902–4907.
4. K. Nakata, T. Ochiai, T. Murakami and A. Fujishima, Electrochim. Acta, 2012, 84, 103–111.
5. J. Schneider, M. Matsuoka, M. Takeuchi, J. Zhang, Y. Horiuchi, M. Anpo and D. W. Bahnemann, Chem. Rev., 2014, 114, 9919–9986.
6. S. Leong, A. Razmjou, K. Wang, K. Hapgood, X. Zhang and H. Wang, J. Membr. Sci., 2014, 472, 167–184.
7. V. Thavasi, G. Singh and S. Ramakrishna, Energy Environ. Sci., 2008, 1, 205–221.
8. A. Sato, R. Wang, H. Ma, B. S. Hsiao and B. Chu, J. Electron Microsc., 2011, 60, 201–209.
9. J. Prince, S. Bhuvana, V. Anbharasi, N. Ayyanar, K. Boodhoo and G. Singh, Sci. Rep., 2014, 4, 6555.
10. Q. Ding, Y.-E. Miao and T. Liu, ACS Appl. Mater. Interfaces, 2013, 5, 5617–5622.
11. T. He, Z. Zhou, W. Xu, F. Ren, H. Ma and J. Wang, Polymer, 2009, 50, 3031–3036.
12. N. Daels, M. Radoicic, M. Radetic, S. W. H. Van Hulle and K. De Clerck, Sep. Purif. Technol., 2014, 133, 282–290.
13. S. Ramasundaram, A. Son, M. G. Seid, S. Shim, S. H. Lee, Y. C. Chung, C. Lee, J. Lee and S. W. Hong, J. Hazard. Mater., 2015, 285, 267–276.
14. S.-I. Lee, S.-M. Jo, H.-I. Joh, M.-H. Lee and S. Lee, Chem. Commun., 2015, 51, 2718–2720.
15. C. H. Kim, B. H. Kim and K. S. Yang, Carbon, 2012, 50, 2472–2481.
16. B. Pant, N. A. M. Barakat, H. R. Pant, M. Park, P. S. Saud, J.-W. Kim and H.-Y. Kim, J. Colloid Interface Sci., 2014, 434, 159–166.
17. M. Zhang, L. Li and X. Zhang, RSC Adv., 2015, 5, 29693–29697.
18. A. Abdal-hay, A. S. Hamdy Makhlouf and K. A. Khalil, ACS Appl. Mater. Interfaces, 2015, 7, 13329–13341.
19. B. Ohtani, Catalysts, 2013, 3, 942–953.
20. K.-H. Park and M. Dhayal, Electrochem. Commun., 2009, 11, 75–79.
21. R. Liu, W.-D. Yang and L.-S. Qiang, J. Power Sources, 2012, 199, 418–425.
22. S. Liu, Z.-R. Tang, Y. Sun, J. C. Colmenares and Y.-J. Xu, Chem. Soc. Rev., 2015, 44, 5053–5075.
23. Y. Zhang, Z.-R. Tang, X. Fu and Y.-J. Xu, ACS Nano, 2010, 4, 7303–7314.
24. Z. Wangxi, L. Jie and W. Gang, Carbon, 2003, 41, 2805–2812.
25. R. D. Leapman, L. A. Grunes and P. L. Fejes, Phys. Rev. B: Condens. Matter Mater. Phys., 1982, 26, 614–635.
26. E. Stoyanov, F. Langenhorst and G. Steinle-Neumann, Am. Mineral., 2007, 92, 577–586.
27. J. Jun, M. Dhayal, J. H. Shin, J. C. Kim and N. Getoff, Radiat. Phys. Chem., 2006, 75, 583–589.
28. A. I. Kuznetsov, O. Kameneva, A. Alexandrov, N. Bityurin, P. Marteau, K. Chhor, C. Sanchez and A. Kanaev, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2005, 71, 1–7.
29. C. Xu, G. P. Rangaiah and X. S. Zhao, Ind. Eng. Chem. Res., 2014, 53, 14641–14649.
30. M. Satterfield, P. Majsztrik, H. Ota, J. Benziger and A. Bocarsly, J. Polym. Sci., Part B: Polym. Phys., 2006, 44, 2327–2345.

---

Footnote

† Electronic supplementary information (ESI) available: SEM images of the nanofibers without ultrasonication of the TiO₂/PAN blend solution. UV-Vis spectrums of the MB solution after various times of photodegradation. See DOI: 10.1039/c5ra22903k

---