A forest of SiO₂ nanowires covered by a TiO₂ thin film for an efficient photocatalytic water treatment

Abstract

We propose a highly disordered and randomly oriented array (forest) of SiO₂ nanowires (NWs) coated by a thin film of TiO₂ as a performing photocatalytic material for water treatment applications. The SiO₂ NWs with a length of tens of microns were obtained via thermal oxidation of Si NWs, grown by plasma enhanced chemical vapour deposition (PECVD), and covered by a nanometric TiO₂ film deposited by atomic layer deposition (ALD). The remarkable photocatalytic performance of the TiO₂/SiO₂ NWs was demonstrated by the degradation of methylene blue and phenols in water. The enhanced photocatalytic activity in the overlaying TiO₂ film is attributed to the synergetic effect of the morphology and optical properties of the SiO₂ NWs. The fabrication methods based on the well-established Si technology make the proposed materials a promising strategy for water treatment.

---

Introduction

Today, much research is devoted to addressing the world's demand for clean and drinkable water via sustainable technology. Semiconductor heterogeneous photocatalysis has a great potential to treat a wide range of organic pollutants (pesticides, herbicides, detergents, pathogens, viruses, coliforms and spores) in water by degrading them into readily biodegradable compounds, and eventually into carbon dioxide and water.^1–3 To date the use of nanoscale TiO₂ photocatalysts in water suspension is the most efficient method for water treatment because of a large surface-area-to-volume ratio, an efficient charge separation and surface trapping.^4–8 However, this approach faces some critical issues which prevent its real employment in a large-scale water treatment process. Expensive and complicated additional steps are indeed required for post-separation of the nano-catalysts from the treated water to avoid TiO₂ contamination.^1,9,10 On the contrary, the immobilization of the nanoparticles onto a large inert substrate, such as membrane or glass, reduces the amount of active catalyst sites and absorbed light since photon penetration might not reach every surface site due to shadow effects.¹¹

In order to overcome these problems, recent investigations are oriented toward the immobilization of the TiO₂ in the form of thin films with a thickness of a few nanometers useful to provide surface charge carriers for photoreactions.^12–15 However the limiting factor in this approach is the low amount of light absorbed in the TiO₂ films.

In this work we propose a highly disordered and randomly oriented array (i.e. forest) of SiO₂ nanowires (NWs) as three-dimensional (3D) support for a TiO₂ thin film. The idea is to exploit the morphological and optical properties of a SiO₂ NW forest to enhance the photocatalytic efficiency of the TiO₂ film. The nano-structuration imparted by the NWs to the overlaying TiO₂ film provides a large surface area, thus enhancing the amount of photo-generated charges responsible for water purification reactions. In addition, the SiO₂ NW forest provides efficient light absorption in the semiconductor catalyst because of strong light scattering.^16–19 As sketched in Fig. 1, the light passing through a NW forest undergoes multiple scattering events. The scattering events fold the light path many times in a random walk inside the NW forest increasing the effective absorption, with a consequent reduction of the reflectivity at the frequencies absorbed by the composite NWs, i.e. light trapping. The TiO₂ layer benefits from the light trapping since the SiO₂ NWs are transparent in the UV/Vis range. As such, the overlaying TiO₂ film captures and absorbs significantly more photons than an equivalent planar thin film and all the TiO₂ surface can contribute to light absorption without shadowing problems. Similar strategies have found successful applications in photovoltaic energy conversion^20–22 and more recently in refractive index sensors.^23,24 Here we propose to extend the potential benefits of the light trapping at the TiO₂ absorption spectral range to water applications. Furthermore, this approach provides a scalable technology for water treatment. Indeed, (i) the NW forest fabrication procedure is based on well-established Si technology, fully integrated in industrial processes; (ii) there are effective techniques to remove the NWs from the substrate and transfer them onto non-conventional supports,^25,26 such as polymeric films, useful to coat the walls of water containers; (iii) the material architecture prevents nanomaterial being released into the environment.

Fig. 1 Multiple light scattering events in a NW forest.

To show the potential of this approach we fabricated a disordered array of SiO₂ NWs via thermal oxidation of Si NWs grown by plasma enhanced chemical vapour deposition (PECVD) on a Si wafer. The SiO₂ NWs, tens of microns long, where then covered by about 10 nm of TiO₂ deposited by atomic layer deposition (ALD). Since ALD is intrinsically atomic, it allows the deposition of conformal films of high aspect ratio structures and a controlled thickness.^27,28 A complete structural and morphological characterization was performed with scanning electron microscopy (SEM), transmission electron microscopy (TEM) and Raman spectroscopy. The optical properties of the TiO₂/SiO₂ NWs were investigated by UV/Vis spectrophotometer measurements. The photocatalytic performance of the TiO₂/SiO₂ NWs was demonstrated by the degradation of methylene blue (MB) dye and phenols in water.

Experimental

A SiO₂ NW forest was produced by thermal annealing of Si NWs grown by PECVD. To induce the NW growth, a 2 nm thick Au film used as catalyst was evaporated onto Si substrates prior to the growth. The growth was performed using pure SiH₄ as a precursor at a pressure of 1 Torr and substrate temperature of 350 °C. A 13.56 MHz radiofrequency with power density of 50 mW cm⁻² was used to create the plasma. After the growth, the Si NWs were oxidized to form SiO₂ (fused silica) NWs via a thermal treatment in a quartz oven with a controlled O₂ atmosphere at 980 °C for 8 h. The SiO₂ NWs were then covered with a thin layer of TiO₂ (about 10 nm in thickness) by ALD. The TiO₂ thickness value was selected after preliminary photocatalytic tests performed on several nanometric TiO₂ planar films. As a reference, TiO₂ was simultaneously deposited also on a Si wafer and a quartz substrate. The ALD was performed with a Picosun R-200 advanced system. During the deposition, the temperature was kept at 300 °C. The two precursors were titanium tetrachloride (TiCl₄, with a purity >99.9995%) and de-ionized water, while N₂ was used as a carrier and purge gas (purity ≥99.999%). The film thickness (∼10 nm) was evaluated by ellipsometry on a TiO₂ planar film deposited on a Si substrate using a M-2000 spectroscopic ellipsometer by Woollam.

The composition analysis of the oxidized NWs was performed by Raman spectroscopy. Raman spectra were acquired with a DXR Thermo Fisher Scientific Raman Microscope, by exciting the samples at 532 nm with a 10 mW power and a 100× objective. The measured spectral range was 50–3000 cm⁻¹, and each spectrum resulted from 2 s acquisition and 200 accumulations over an area of about 800 nm (laser spot).

The structural characterization was performed by SEM and TEM. SEM analyses were performed by using a field emission Zeiss Supra 25 microscope. TEM analyses were performed by using a JEOL JEM ARM200CF at a primary beam energy of 200 keV, operated both in conventional TEM mode (C-TEM) and in scanning mode (S-TEM) and equipped with a large angle silicon drift detector (SDD) energy dispersive X-ray (EDS) detector. TiO₂/SiO₂ NWs were transferred onto a copper–carbon TEM grid by mechanical rubbing.

The optical properties were studied by measuring the angle integrated total reflectivity in the spectral range between 200 and 1100 nm with a Perkin-Elmer Lambda 35 UV/Vis spectrophotometer equipped with an integrating sphere.

The photocatalytic activity was evaluated by the degradation of the MB organic compound. The samples, 1 cm × 1 cm in size, were immersed in a solution (2 ml) containing MB and de-ionized water with an initial concentration of 1.5 × 10⁻⁵ M. The mixture was irradiated by using a UV lamp (Philips F8T5BLB-8W), peaked at 365 nm with a full width at half maximum of 20 nm and irradiance of 2 mW cm⁻², for a total time of 3.5 hours. Every 30 minutes of irradiation the solution was measured using a UV/VIS spectrophotometer (PerkinElmer Lambda 45) in a wavelength range between 500 and 800 nm. The temperature of the solution was tracked with a thermocouple. During the tests, the measured values are comparable with the room temperature.

The photocatalytic activity of TiO₂/SiO₂ NWs was also tested for the degradation of phenols (purchased by Sigma Aldrich) with an initial concentration of 1.38 mg l⁻¹ (in de-ionized water) using the same procedure with the degradation of MB. The variation of the concentration of the phenol solution was measured spectrophotometrically using a Hatch DR 3900 spectrophotometer.

Results and discussion

The photograph of the oxidized NWs on a 3 inch Si wafer in Fig. 2a shows that these materials form a natural and uniform bright white mat.

Fig. 2 (a) Photograph of SiO₂ NW forest on a 3 inch Si wafer. (b) Raman spectra of the oxidized NWs (red line) and Si substrate (black line). (c) and (d) Plan-view SEM images.

In Fig. 2b the Raman spectra of the oxidized NWs and Si substrate are plotted. The oxidized NWs are characterized by the typical fused silica features such as the strong and diffused band at around 440 cm⁻¹ with a weaker band near 880 cm⁻¹ and the two sharp peaks at around 490 cm⁻¹ and 606 cm⁻¹.²⁹ The sharp and strong peak at around 520 cm⁻¹ is related to the contribution of the underlying crystalline Si³⁰ substrate as clear by comparing the NW and substrate spectra. The SEM images in Fig. 2c and d reveal tens of micrometers long NWs, characterized by a tapered shape with bottom diameter in the range of 150–200 nm and tip diameter around 50–80 nm as determined by combined measurements from plan and cross view SEM. The SiO₂ NWs are interwoven in a dense forest with a macroporous structure easily accessible to liquids such as water.

Fig. 3 reports TEM analyses on the SiO₂ NWs after the deposition of TiO₂. The bright field C-TEM image reported in Fig. 3a shows the excellent conformability of the TiO₂ deposition thanks to the diffusion of the precursors inside the nanostructured SiO₂ template. The thickness of the TiO₂ coating was estimated to be 14 nm. In the inset, the elemental qualitative S-TEM EDS maps of a NW section show Ti, O and Si signals. While O and Si signals are localized within the nanowire and are proportional to thickness, Ti is clearly localized at the NW surface. The dark field C-TEM image reported in Fig. 3b highlights the TiO₂ crystalline domains. The selected area electron diffraction (SAED) pattern, reported in the inset, shows discrete rings attributed to the polycrystalline anatase phase of the TiO₂, since SiO₂ is amorphous and produces only the diffuse halo around the central transmitted spot.

Fig. 3 (a) Bright field C-TEM image showing the continuous TiO₂ deposited layer. In the inset, the elemental EDS maps for Ti (red), O (green) and Si (blue). (b) Bright field C-TEM image showing TiO₂ domains and the relative SAED pattern with crystal plane indexing showing anatase phase of the TiO₂.

This result is in perfect agreement with what is reported in the literature for ALD deposition.³¹ The polycrystalline structure of the TiO₂ coating is the same for both TiO₂/SiO₂ NWs and flat TiO₂ reference film, as expected.

The optical behaviour of TiO₂/SiO₂ NWs was studied by measuring the total integrated reflectivity, R, between 200–900 nm and comparing it with that of two planar reference films consisting of TiO₂ deposited on Si, which is the same bare substrate of the NWs, and quartz substrate. Both the films were characterized by the same thickness of the NW coating.

In Fig. 4 the spectra obtained from the TiO₂/SiO₂ NWs (continuous red line), the planar TiO₂ film on Si (continuous blue line) and quartz (continuous green line) are plotted.

Fig. 4 Total integrated reflectivity spectra of TiO₂/SiO₂ NWs (red lines) and TiO₂ reference film on Si (blue lines) and quartz (green lines). The continuous lines are the experimental data, while the dashed lines are the calculated curves. In the inset the spectrum of the pristine SiO₂ NWs.

The planar TiO₂ on Si shows R values around 35% in the Vis-IR and a continuous increase of R up to 60% for λ < 400 nm, where TiO₂ is absorbing and the photocatalytic processes occur. The high reflectivity of the planar TiO₂ for λ < 400 nm is not reduced if a low refractive index substrate is used as shown by the spectrum obtained from the same planar TiO₂ film deposited on quartz.

The experimental spectra are well reproduced by the calculated ones (dashed red and green lines for TiO₂ on Si and on quartz, respectively) obtained by using a program on the http://filmetrics.com web site.³² On this site, information about layered structures such as composition, refractive index, extinction coefficient and thickness can be used in the complex-matrix form of Fresnel equations to produce the simulated reflectance spectra. The calculations were obtained by considering a film thickness of 10 nm for both the planar TiO₂ samples. The weak discrepancies observed between the experimental and calculated spectra are attributed to the fact that the optical constants used in the calculations are for the rutile TiO₂ (the only option of Filmetrics) whereas our TiO₂ is anatase.

The spectrum of the TiO₂/SiO₂ NWs is characterized by a strong diffusive reflectivity over Vis/IR range and a very weak reflectivity for λ < 400 nm where R strongly decreases down to 10%, a value much smaller than that observed for the planar films. It is worth highlighting that the drastic reduction of R occurs at a wavelength of 384 nm corresponding to the bulk TiO₂ energy band gap in the anatase phase (3.2 eV).⁹ For comparison, the inset of Fig. 4 shows the R spectrum of the bare SiO₂ NWs which is characterized by a strong reflectivity over the whole investigated range with values up to 80% in the UV range. This suggests that the strong reduction of R observed for the TiO₂/SiO₂ NWs in the UV is due to the TiO₂ absorption of the incident light.

The optical behaviour of TiO₂/SiO₂ NWs can be described within the model of diffuse optical reflectors.¹⁶ In that model, the NWs are regarded as a discrete medium consisting of an ensemble of diffuse optical reflectors in air placed onto a partial reflecting substrate. The NWs can absorb, transmit, or reflect the incident light, and their reflectivity, R, is calculated by taking into account light scattering by individual NWs described in the Rayleigh approximation, absorption by the NWs, and substrate reflectivity. Under these conditions, R can be written as:^16,18

where R_N and T are the reflectivity and the transmittance of the NW forest, respectively, and R_s represents the substrate reflectivity at the substrate/air interface. R_N and T are given by the equationswhere N is the number of the scattering events, α and d the absorption coefficient and length, σ the scattering cross section and w the total width of the mat.

In our case, since the SiO₂ NWs are transparent in the whole investigated spectral range, the calculations were performed by considering α and d of the TiO₂ component. In Fig. 4 the calculated spectrum is plotted (dashed red line) and show a very good agreement by using the values of the crystalline anatase TiO₂ absorption coefficient³³ and the measured R_s as fixed parameters, whereas Nd = 7 × 10⁻³ cm and σw = 6 × 10⁻¹⁷ cm³ as fitting parameters. Hence, from our optical study a key feature can be highlighted: the presence of the SiO₂ NWs allows the overlaying TiO₂ film to absorb significantly more photons than an equivalent planar film which is affected by a high reflectivity in the spectral range where the photocatalytic processes occurs. In particular, by comparing the R spectra below λ = 384 nm, the nanostructured TiO₂ absorbs about 40% of the light, that in the planar TiO₂ film is completely lost due to reflection in that range.

In order to check the photocatalytic activity of the TiO₂/SiO₂ NWs in the degradation of organic compounds under UV light irradiation, we performed MB degradation measurements in water.¹³ The results were compared with that of the reference TiO₂ planar film and that of TiO₂/SiO₂ NWs.

Fig. 5a reports the residual concentration of MB, C/C₀, where C is the concentration of MB after the irradiation and C₀ the starting concentration of MB, versus the irradiation time. We tested four samples: MB in absence of any catalyst materials (squares), with SiO₂ NWs (circles), with TiO₂ planar film (triangles), with TiO₂/SiO₂ NWs (diamonds). A preconditioning process was applied to remove possible adsorbed organic pollutants on the surfaces. In particular, samples were irradiated by the UV lamp for 50 min in order to remove the hydrocarbons localized on the sample surface. After the preconditioning process, samples were immersed in the MB solutions and kept in the dark for one hour and half. This step allowed disentangling the photocatalytic effect from the adsorption of MB on the sample surfaces and beaker walls.

Fig. 5 (a) MB degradation under UV irradiation and (b) rate kinetics for four samples: MB (squares), MB with the TiO₂ planar film (triangles), MB with the SiO₂ NWs (circles) and MB with the TiO₂/SiO₂ NWs (diamonds).

After the MB adsorption step, the samples were exposed to UV light to induce the photocatalysis. The observed reaction rate, k, follows a Langmuir–Hinshelwood kinetics model, which can be expressed as follows:

where C is the concentration of organic species, C₀ is the initial concentration of organic species and t is the irradiation time.¹ The reaction rates are reported in Fig. 5b for all the analysed samples. The k value for SiO₂ NWs is positive due to both the MB desorption from the sample surface and the incapability of the SiO₂ NWs to degrade the dyes. Consequently, there is an increase of the MB concentration in the solutions with increasing exposition time. Nevertheless, the variation of the MB concentration with the time is within the experimental error of the measurements (2%). On the other hand, planar TiO₂ film and TiO₂/SiO₂ NWs show negative values for k due to the ability of titanium dioxide to degrade MB. In particular, TiO₂/SiO₂ NWs show the best photocatalytic activity, with a reaction rate more than 10 times the rate of TiO₂ flat film.

In order to demonstrate that the synthesized materials are active in the degradation of dangerous organic pollutants, the photocatalytic properties of TiO₂/SiO₂ NWs were tested for the degradation of phenols. The phenols are considered as priority pollutants since they are harmful for organisms at low concentrations and many of them have been classified as hazardous pollutants because of their potential to harm human health. The major sources of phenol pollution in the aquatic environment are wastewaters from paint, pesticides, coal conversion, polymeric resin, petroleum and petrochemical industries.³⁴ Because of their toxicity, phenols have been included in the list of priority pollutants of US environmental protection agency (EPA) (Federal Register of US EPA, 1987). Also the EU has classified several phenols as priority contaminants and the 80/778/EC directive sets a maximum concentration of 0.5 mg l⁻¹ for total phenols in drinking water.³⁵

Fig. 6 reports the percentage of the phenol degradation after 150 min of irradiation under UV light. The photocatalytic response of the TiO₂/SiO₂ NWs samples were compared to the response of the phenol solution without any catalyst materials, with SiO₂ NWs and with the reference TiO₂ planar film. On the abscissa axis, ‘blank’ indicates a phenol solution in the absence of any catalysts. ‘Blank’ demonstrates a degraded phenol solution of 0.7%, the same percentage for the SiO₂ NWs. The planar TiO₂ shows a phenol degradation of 9%, whereas, the TiO₂/SiO₂ NWs were able to remove 45% of phenols present in the solution.

Fig. 6 Phenol degradation percentage after 150 min of UV-irradiation, for the phenol solution in the absence of the catalyst (blank), for the different investigated samples: SiO₂ NWs, TiO₂ planar film, and TiO₂/SiO₂ NWs.

In both types of photocatalytic tests performed we observe improved performances for TiO₂/SiO₂ NWs.

The marked difference in the photocatalytic response between the TiO₂/SiO₂ NWs and the TiO₂ planar film can be explained taking into account the significant enhancement of the TiO₂ exposed surface and the more efficient collection of photons in the nanostructured sample. Eventually, we expect an improvement of performance by optimizing some key material features, including the NW size and density.

Conclusions

Our results show that the exploitation of a SiO₂ NW forest as 3D-support for TiO₂ thin film is an effective strategy for the creation of performing photocatalytic materials. These materials show morphological and optical properties very suitable for water treatment applications. First, the NWs offer to the photocatalytic semiconductor an increased surface area, easily accessible to water, and hence they allow an increase in the number of photo-generated charges responsible for water purification reactions. Second, the NW forest enhances the absorption in the TiO₂ overlayer because of light scattering in the NW forest. Third the fabrication methodology is based on well-established Si technology and the material architecture does not allow nanomaterial to be released into the environment.

Acknowledgements

This research has been partially supported by the FP7 European Project WATER (Grant Agreement 316082). Part of the work was performed at Beyond-Nano CNR-IMM, Catania, Italy, which is supported by the MIUR under project Beyond-Nano (PON a3_00363). The authors wish to thank Dr Valentina Mussi (CNR-ISC) for Raman characterization, Mr Marco Maiani (CNR-IMM) and Mr Giuseppe Pantè (CNR-IMM) for technical assistance, Dr Eric G. Barbagiovanni (CNR-IMM) for his assistance editing this manuscript and Thermo Fisher Scientific for technological support.

Notes and references

1. M. N. Chong, B. Jin, C. W. K. Chow and C. Saint, Water Res., 2010, 44, 2997.
2. M. A. Barakat, Arabian J. Chem., 2011, 4, 361.
3. S. Malato, P. Fernandez-Ibanez, M. I. Maldonado, J. Blanco and W. Gernjak, Catal. Today, 2009, 147, 1.
4. D. F. Ollis, Environ. Sci. Technol., 1985, 19, 480.
5. A. Fujishima and K. Honda, Nature, 1972, 238, 37.
6. W. Y. Teoh, J. A. Scott and R. Amal, J. Phys. Chem. Lett., 2012, 3, 629.
7. M. A. Lazar, S. Varghese and S. S. Nair, Catalysts, 2012, 2, 572.
8. K. Nagaveni, G. Sivalingam, M. S. Hegde and G. Madras, Environ. Sci. Technol., 2004, 38, 1600.
9. S. A. Lee, K.-H. Choo, C.-H. Lee, H.-I. Lee, T. Hyeon, W. Choi and H.-H. Kwon, Ind. Eng. Chem. Res., 2001, 40, 1712.
10. G. C. C. Yang and C. J. Li, Sep. Purif. Technol., 2007, 58, 159e165.
11. R. L. Pozzo, J. L. Giombi, M. A. Baltanas and A. E. Cassano, Catal. Today, 2000, 62, 175.
12. Y. Zhang, J. Li and J. Wang, Chem. Mater., 2006, 18, 2917.
13. V. Scuderi, G. Impellizzeri, L. Romano, M. Scuderi, M. Brundo, K. Bergun, M. Zimbone, R. Sanz, M. A. Buccheri, F. Simone, G. Nicotra, B. G. Svensson, M. G. Grimaldi and V. Privitera, Nanoscale, 2014, 6, 11189.
14. V. Scuderi, G. Impellizzeri, L. Romano, M. Scuderi, G. Nicotra, K. Bergun, A. Irrera, B. G. Svensson and V. Privitera, Nanoscale Res. Lett., 2014, 9, 458.
15. T. Luttrell, S. Halpegamage, J. Tao, A. Kramer, E. Sutter and M. Batzill, Sci. Rep., 2014, 4, 4043.
16. R. A. Street, W. S. Wong and C. Paulson, Nano Lett., 2009, 9, 3494.
17. A. Convertino, M. Cuscunà and F. Martelli, Nanotechnology, 2010, 21, 355701.
18. A. Convertino, M. Cuscunà, S. Rubini and F. Martelli, J. Appl. Phys., 2012, 111, 114302.
19. O. L. Muskens, S. L. Diedenhofen, B. C. Kaas, R. E. Algra, E. P. A. M. Bakkers, J. Gòmez Rivas and A. Lagendijk, Nano Lett., 2009, 9, 930.
20. L. Hu and G. Chen, Nano Lett., 2007, 7, 3249.
21. O. L. Muskens, J. Gómez Rivas, R. E. Algra, E. P. A. M. Bakkers and A. Lagendijk, Nano Lett., 2008, 8, 2638.
22. S. K. Jana, A. Le Donne and S. Binetti, J. Phys. Chem. Solids, 2012, 73, 143.
23. A. Convertino, M. Cuscunà, F. Martelli, M. G. Manera and R. Rella, J. Phys. Chem. C, 2013, 118, 685.
24. A. Colombelli, M. G. Manera, A. Taurino, M. Catalano, A. Convertino and R. Rella, Sens. Actuators, B, 2016, 226, 589.
25. C. H. Lee, D. R. Kim, I. S. Cho, N. William, Q. Wang and X. Zheng, Sci. Rep., 2012, 2, 1000,  DOI:10.1038/srep01000.
26. K. E. Plass, et al., Adv. Mater., 2009, 21, 325.
27. N. Pinna and M. Knez, in Atomic Layer Deposition of Nanostructured Materials, Wiley-VCH Verlag GmbH & Co. kGaA, Weinheim, Germany, 2012.
28. S. M. George, Chem. Rev., 2010, 110, 111.
29. R. J. Hemley, H. K. Mao, P. M. Bell and B. O. Mysen, Phys. Rev. Lett., 1986, 57, 747.
30. J. H. Parker, D. W. Feldman and M. Ashkin, Phys. Rev., 1967, 155, 712.
31. J. Aarika, A. Aidla and H. Mandar, J. Cryst. Growth, 2000, 220, 531.
32. Filmetrics, reflectance calculator, http://filmetrics.com/reflectance-calculator, accessed June 2016.
33. S. Tanemura, L. Miao, P. Jin, K. Kaneko, A. Terai and N. Nabatova-Gabain, Appl. Surf. Sci., 2003, 212–213, 654.
34. E. Grabowska, J. Reszcznska and A. Zaleska, Water Res., 2012, 46, 5453.
35. I. Rodrıguez, M. P. Lompart and R. Cela, J. Chromatogr. A, 2000, 885, 291.

---