Virus and bacterial removal ability of TiO₂ nanowire-based self-supported hybrid membranes

2.1 Materials

Titanium dioxide (TiO₂) (P25) with an average diameter of 21 nm, ammonium hydroxide (NH₄OH, 25 %) and sodium hydroxide (NaOH) was purchased from Sigma Aldrich (Hungary). The potassium hydroxide (KOH) was purchased from Thomasker Co., (Hungary). Iron chloride hexahydrate (FeCl₃·6H₂O) was purchased from Scharlab, (Hungary). Hydrochloric acid (HCl, 37 %), copper (II) acetate monohydrate (Cu(OOCCH₃)₂·H₂O) was ordered from VWR Chemicals (Hungary). Cellulose micro wires originating from the DIPA ltd (Hungary). Polyvinylidene (PVDF) filter membrane with a diameter of 47 mm and the pore size of 0.1 µm (Durapore-VVLP04700) was used to produce the hybrid membranes.

Bacteriological agar, d-glucose, and sodium dihydrogen phosphate dihydrate (NaH₂PO₄·2H₂O) were purchased from Simga-Aldrich (Switzerland). Calcium chloride dihydrate (CaCl₂·2H₂O), microbiology yeast extract and glycerol were provided by Merck Eurolab (Switzerland). Streptomycin was purchased from AppliChem PanReac (Germany). Tryptone (Difco 0123) and sodium chloride (NaCl) were purchased from Becton Dickinson and VWR International (Switzerland), respectively. Escherichia coli DH5α (SZMC 21399) Gram-negative strain was used for filtration test as a model organism (this bacterial strain was gifted from the Department of Microbiology, Faculty of Science and Informatics, University of Szeged). Escherichia coli (Migula 1895) Castellani and Chalmers 1919 (DSM No.: 5695) colonies were used as host cells for MS2 bacteriophage multiplication (DSM No.: 13767). Dry E. coli pellets and the MS2 phage suspension were purchased from DSMZ (Braunschweig, Germany).

2.2 Preparation of TiO₂ NW-based hybrid membranes

In our earlier manuscript, the preparation and investigation of TiO₂ NW, the TiO₂ NW-Fe₂O₃ and TiO₂ NW-CuO nanocomposite additives (M. A. Shehab et al., 2022) were shown. Briefly, solvothermal process was used for the preparation of TiO₂ NW, and impregnation technique was applied to produce the inorganic composite additives such as TiO₂ NW-Fe₂O₃ and TiO₂ NW-CuO in both cases.

For the synthesis of hybrid membranes, firstly, the TiO₂ NW-Fe₂O₃ and TiO₂ NW-CuO nanocomposites were prepared. The calculated amount of FeCl₃·6H₂O and of (Cu(CH3COO)₂·H₂O) precursors was dissolved in 100 mL of distilled water and 100 mL of EtOH respectively, to obtain a homogeneous solution. In the next step, TiO₂ NWs was added to the solutions and stirred for 1 h, then transferred to autoclave. The products obtained was washed to adjust the pH 7, and then calcinated for 2 h at 500 °C using a static furnace. The load of the Fe₂O₃ and CuO nanoparticles in the final nanocomposites was 5 w/w % each case.

To prepare the TiO₂ NW-Fe₂O₃-cellulose and TiO₂ NW-CuO-cellulose membranes, 0.2 g of the above-prepared composites was dipped into 100 mL of distilled water and EtOH, respectively for 1 h. Finally, 5 g of cellulose (1 w/w%) was added to the solution for another 1 h. Hybrid membranes were produces by vacuum filtration using PVDF membrane (total mass of 250 mg/membrane), then it was dried in a furnace for 30 min at 40 °C (M. A. Shehab et al., 2022).

2.3 Bacterial filtration efficiency (BFE) tests

For bacterial filtration efficiency tests, tenfold serial dilution was made from the initial starter culture in 30 mL Lysogeny Broth (LB) medium. The colony forming unit (cfu) concentration of the sixth tenfold dilution was 1.21 ± 0.24 × 10³ cfu/mL determined by the colony counting method. This suspension was used for the filtration experiment providing a countable number of colonies on the agar plate. Before each filtration, 3 × 100 µL sample was evenly spread on the top of LB agar plate to check the initial colony forming unit concentration (CFU0). The filtration was performed using glass vacuum filtration device (Sartorius Stedim Biotech 16306–25 mm) without vacuum pump, applying gravity driven filtration. After filtration 3 × 100 µL sample was also taken from the filtrated suspension to determine the number of the colony forming units of the filtrate (CFUf) and then to derive BFE% values using the following formula: $$\mathit{BFE}\%=\frac{{\mathit{CFU}}_0-{\mathit{CFU}}_f}{{\mathit{CFU}}_0}\bullet 100\%$$

2.4 Virus filtration efficiency tests

The required media for bacteria and MS2 growth and filtration (such as antibiotic solution, broth, virus dilution buffer, hard and soft agar) were produced as offered by B.M. Pecson et al. (B. M. Pecson et al., 2009). MS2 was replicated using its host Escherichia coli strain and subsequently purified and concentrated in different steps, according to the protocol provided by DSMZ. Because of the sensitive nature of MS2 bacteriophages, the amount of plaque forming units (PFUs) in a suspension had to be determined from time to time. The initial stock of purified MS2 had the concentration of 7.6 × 10⁷ PFU/mL. For further usage, the initial stock was diluted in virus dilution buffer (VDB). Hence, the detection limit of ∼ 4 LRV (Log Reduction Value) or 99.99 % was determined by the phage enumeration that used a maximum sample volume of 2 mL. VDB was prepared using NaH₂PO₄·2H₂O, NaCl and water. The pH (pH = 7.0) was adjusted by adding drops of 0.1 M NaOH solution. Room temperature (23 °C) was maintained throughout the virus filtration experiments. The concentration of the MS2 bacteriophages was determined after the incubation. Each enumeration of MS2 samples was performed twice, and each condition was tested three times.

Filtration experiments, the as-prepared hybrid membranes were placed into a 200 mL glass funnel with a porous glass insert at the bottom. During the experiments so-called gravity driven filtration was applied, consequently vacuum suction was not used to enhance the flux of the filtered liquid. A control experiment was also performed without a filter to investigate the influence of the equipment on the retention of the bacteriophages, given that the system is porous. 30 mL of the virus dilution buffer was infected with the calculated volume of bacteriophage stock to reach the active phage concentration of 10 PFU/mL. This allowed the investigation of retention of up to 2 magnitudes (2 Log). Separate experiments were performed both with membranes containing TiO₂ NW-Fe₂O₃ and TiO₂ NW-CuO nanocomposites. One millilitre of the filtered liquid was collected after each filtration and used to infect E. coli for virus enumeration. Filtration experiments were performed at pH values 7.0 to simulate real-life situations.

2.5 Toxicological tests

Toxicology test was done by the Kisanalitika Kft. (Sajóbábony - Hungary). Fish tests can be used to assess the environmental risk of surface water or to determine the toxic effects of pollution. Acute toxicity is a detectable adverse effect in an organism within a short period of time (within days) after exposure to a substance. In this study, acute toxicity is expressed as the lethal concentration 50 (LC50), i.e. the concentration of the toxic material in water that causes death in 50 % of the test fish group within given period of time. A static test was used, so the test solution does not flow and the solutions remain unchanged throughout the experiments. In our case zebrafish (Danio rerio) were used to perform the tests. The fish must be in good health and free from any visible physical defects. Water toxicity testing was performed using a dilution series, using the water from the fish aquarium as dilution water, which also serves as the water for the control test group. The dilution water in the aquarium was aerated and its oxygen and pH were checked regularly. The test groups were placed in 500 mL of 24-hour soaking water of hybrid membranes or in water filtered also using the tested hybrid membranes. Each test group contained 5 fishes. Fishes were exposed to the test substance added to the water, over the concentration range tested, for a period of 48 h. Fish mortality was recorded at specified intervals, in this case after 2 h, 24 h and 48 h.

2.6 Characterization

The surface morphology and the three-dimensional architecture of hybrid membranes was characterized using YXLON FF35 dual-beam X-ray micro computed tomography equipment (CT) (microfocus X-ray tube, transmission beam, acceleration voltage: 50 kV).

Raman spectroscopy measurements were carried out using a high-resolution Raman spectrometer (Nicolet Almega XR, Thermo Electron Corporation, Waltham, MA, USA) equipped with a 532 nm Nd:YAG laser (50 mW).

Furthermore, our recently published paper presented a wide range of characterization of the as-prepared membranes, such as transmission electron microscopy (TEM), scanning electron microscopy (SEM), X-ray powder diffraction (XRD), specific surface area measurement (BET), etc., and their photocatalytic properties in the decomposition of against methylene blue organic dye (M. A. Shehab et al., 2022).

Electrophoretic measurement was performed by dynamic light scattering (DLS) (ZetaSizer NS-Malvern, UK) device. The measurement is based on the combination of laser Doppler velocimetry and phase analysis of light scattering (PALS) in Malvern’s M3-PALS technique.

Inductively coupled optical emission spectrometry (ICP-OES) measurements were performed in a Varian 720 ES ICP-OES device to determine the amount of Cu and Fe dissolved during the filtration experiments.

The FTIR spectra were recorded on a Bio-Rad Digilab Division FTS65A/896 FT-IR spectrometer equipped with a DTGS detector and a Ge/KBr beamsplitter, between 4000 and 400 cm⁻¹, at 4 cm⁻¹ optical resolution. A Harrick’s Meridian® SplitPea Single Reflection Diamond ATR accessory was used, so no sample preparation was required. The spectrometer was controlled by Win IR Pro v. 3.3 (Bio-Rad Digilab Division) software, and the spectra were analyzed using GRAMS/AI v. 7.0 (Thermo Galactic) software.

3.1 CT, Raman and FT-IR spectroscopy analysis

In our recently published study (M. A. Shehab et al., 2022) TEM and SEM analysis was performed in case of both TiO₂ NW-based composite additives and the hybrid membranes. It was found that the electron microscopy results showed properly the structure of as-prepared nanocomposite additives and the surface properties of hybrid membranes, although these measurements were did not provided information about the 3D structure of the hybrid membranes. Furthermore, there are some other limitations of the usage of the electron microscopy techniques in the field of membrane technology, such as the small size of the investigated area. In order to gain information about the homogeneity and the 3D morphology of the whole hybrid membrane structure, micro CT analysis was performed. Fig. 1 shows the 3D cross section analysis of TiO₂ NW-Fe₂O₃-cellulose (Fig. 1 a, b) and TiO₂ NW-CuO-cellulose (Fig. 1 c, d) membranes.

Close

Fig. 1

3D CT cross section analysis of TiO₂ NW-Fe₂O₃-cellulose (a, b) and TiO₂ NW-CuO-cellulose (c, d) hybrid membranes.

Export to PPT

As can be seen in Fig. 1, no significant different was observed in the membrane structures. In both types of hybrid membrane, the brighter shapes represent the presence of cellulose, while the grey fibrous material was originating from the TiO₂ NW-Fe₂O₃ (Fig. 1 a, b) and TiO₂ NW-CuO (Fig. 1 c, d) nanocomposite additives. It was found, that in good approximation homogeneous distribution of the nanocomposite additives and cellulose in the hybrid membrane structure was formed.

The material volume values of the hybrid membranes were also determined in the full 3D extension as can be seen in Table 1. There was no difference between the measured volume of materials, it was 0.33 mm³ in both cases. Furthermore, the obtained percentage by volume of materials values were also rather similar, the difference between the values was only 0.5 %. Based on the results obtained, it can be concluded that the 3D structure of the membranes produced is practically identical, so that the differences observed in the E. coli and MS2 filtration experiments are certainly due to differences in the material quality of the membranes and not to morphological differences.

In order to prove the chemical interaction between cellulose and TiO₂ NW-based nanocomposite additives, Raman spectroscopy analyses were recorded. Fig. 2a and b illustrates the Raman spectra of TiO₂ NW-CuO and TiO₂ NW-Fe₂O₃ nanocomposites (black curves), the pristine cellulose (red curves) and the as-prepared TiO₂ NW-CuO-cellulose and TiO₂ NW-Fe₂O₃-cellulose hybrid membranes (blue curves), respectively.

Close

Fig. 2

Raman analysis of TiO₂-CuO-cellulose (a) and TiO₂-Fe₂O₃-cellulose (b) hybrid membranes.

Export to PPT

Raman spectra in Fig. 2a show three peaks at 143, 412, and 638 cm⁻¹ corresponding to the E_g (144 cm⁻¹), B_1g (399 cm⁻¹), and E_g (639 cm⁻¹) modes of anatase TiO₂ crystals, respectively (W. T. Pawlewicz et al., 1983), and peaks at 284 and 636 cm⁻¹ belonging to the A_g (296 cm⁻¹) and B_2g (636 cm⁻¹) modes of CuO crystals, respectively (T. Yu et al., 2004).

The characteristic peaks of cellulose can be identified at 138, 1097, 1365 and 1650 cm ^-1, as can be seen in Fig. 2 a red curve (K. Schenzel and S. Fischer, 2001). In Fig. 2 a blue curve shift in both the characteristic peaks of TiO₂-CuO nanocomposite and cellulose were observed in the spectrum of hybrid membrane. Moreover, as can be seen on the spectrum of the hybrid membrane new peaks appear at 1028, 1439 and 1549 cm⁻¹ (marked by green colour in Fig. 2 a). It can be assumed that these bands originated from the interaction of the TiO₂-CuO nanocomposite and the oxygen containing surface functional groups of cellulose while there are no characteristic peaks of the nanocomposite and the cellulose in this region.

In the case of TiO₂-Fe₂O₃-cellulose hybrid membrane the Raman spectra in Fig. 2 b show three peaks at 143, 450 and 659 cm⁻¹ corresponding to the E_g (144 cm⁻¹), B_1g (399 cm⁻¹), and E_g (639 cm⁻¹) modes of anatase TiO₂ crystals, respectively (W. T. Pawlewicz et al., 1983), and peaks at 286, 658 and 880 cm⁻¹ belonging to the E_g (293, 613 and 814 cm⁻¹) modes of α-Fe₂O₃ crystals, respectively (H. Mansour et al.; 2020). Similarly, to the TiO₂-CuO-cellulose membrane (Fig. 2 a) the characteristic peaks of TiO₂-Fe₂O₃ nanocomposite and cellulose were shifted slightly in the spectrum of TiO₂-Fe₂O₃-cellulose hybrid membrane (Fig. 2 b blue curve). Furthermore, new peaks appear at 993 and 1516 cm⁻¹ (marked by green colour in Fig. 2 b). These peaks can also be attributed to the interaction of the TiO₂-Fe₂O₃ nanocomposite and the oxygen containing surface functional groups of cellulose. The above-mentioned phenomena could be explained by the formation of non-covalent interactions, presumably hydrogen bonds, between the TiO₂-CuO, TiO₂-Fe₂O₃ nanocomposites and cellulose fibres, resulting in band shifting and new characteristic peaks (B. El Mrabate et al., 2020).

For further characterization and to answer the question without doubt, that chemical bond formed between the membrane components, FT-IR analysis was performed. Beside the reference cellulose (black curve) and TiO₂ NW/cellulose (red curve), the IR spectra of the hybrid membranes (blue and brown curves) are shown in Fig. 3 in the range of 400 to 800 cm⁻¹. This low frequency range features mainly the metal-oxigen stretching vibrational bands. As can be seen the black curve the cellulose has five bands with an intensity at 434, 518, 557, 609 and 663 cm⁻¹, while the TiO₂ NW/cellulose composite has two bands with considerable intensity at 439 and 584 cm⁻¹. In the case of hybrid membranes, the above mentioned peaks slightly shifted - from 439 cm⁻¹ to 440 and 442 cm⁻¹ and from 584 cm⁻¹ to 586 cm⁻¹ -, presumably due to the chemical interaction between the CuO, Fe₂O₃ nanoparticles and TiO₂ NW. Furthermore, a new band can be observed in both the hybrid membranes spectra at 669 cm⁻¹ (marked by green), which is assigned to the metal–oxygen stretching band due to the presence of interaction between the TiO₂ NW and CuO and Fe₂O₃ nanoparticles, since there are no characteristic bands of the pristine components in this region (O. P. Keabadile et al., 2020; A. Lassoued et al., 2017).

Close

Fig. 3

FT-IR analysis of cellulose (black curve), TiO₂-cellulose (red curve), TiO₂-CuO-cellulose (blue curve) and TiO₂-Fe₂O₃-cellulose (brown curve).

Export to PPT

3.2 Zeta potential and ICP-OES measurements

Zeta potential (ζ) of the hybrid membranes and raw materials was measured at pH 7.0 and the results presented in Table 2. Prior to the experiments, the samples were dispersed in deionized water to reach the final concentration of 0.1 wt%. Each measurement was repeated three times.

By analyzing the ζ potential of cellulose, TiO₂ NWs-cellulose, and the hybrid membranes, we can conclude whether the efficiency of virus retention can be improved by the CuO and Fe₂O₃ nanoparticles adhered on the surface of TiO₂ NWs. The zeta potential of the neat cellulose is negative (–32.2 mV) whereas the hybrid membranes became more positively charged due to the functionalization (-18.2 and −10.4 mV). The surface treatment shifts the isoelectric point (IEP) towards higher values (the surface becomes more positively charged), consequently the filtration efficiency of MS2 could be also increasing, due to the electrostatic interaction between the surface of hybrid membranes and the bacteriophages. Since, the ζ potential of MS2 is approximately −30 mV (V.I. Syngouna and C.V. Chrysikopoulos, 2010), it is not expected a high virus retention value from the pure cellulose with a ζ potential of –32.2 mV.

In contrast to this, Fe₂O₃ and CuO nanoparticles have been shown to have positive ζ potential at pH 7.0 (M. Kosmulski, 2009). It can also be ascertained that the surface functionalization of TiO₂ NWs by applying CuO and Fe₂O₃ nanoparticles shifted the zeta potential values to less negative region in the case of hybrid membranes, as can be seen in Table 2. These findings mean that the electric properties of the surface of TiO₂ NWs-based membranes are more favourable for the retention of MS2 bacteriophages.

Samples for inductively coupled optical emission spectrometry (ICP-OES) measurements have been prepared to determine the total dissolved Cu and Fe content by filtering deionized water of pH 7.0 through the membrane under the previously detailed experimental conditions. The results show in Table 3 that in each case the dissolved amount of metal ions less than the WHO limitations (World Health Organization, 2017) (Cu < 2 mg/L; Fe - no health-based guideline value is proposed for iron); thus, the dissolution of Cu and Fe during filtration does not overstep in extent the international guidelines. Furthermore, based on the values obtained, it is assumed that leaching not affects the results of E. Coli and MS2 filtration experiments.

3.3 E. coli filtration experiments results

Before the filtration tests began, control experiments were performed without applying the as-prepared hybrid membranes to investigate the possible adsorption on the membrane holder of the funnel. The E. coli removal efficiency values are shown in Fig. 4. In both cases, the membranes containing TiO₂ NW-based inorganic nanocomposite additives showed significantly higher E. coli removal (Fig. 4 red and orange columns) than membranes were build using only TiO₂ NW and cellulose without the usage of Fe₂O₃ and CuO nanoparticles (Fig. 4 green column). As can be seen the highest E. coli removal was achieved by the TiO₂ NW-Fe₂O₃-cellulose membrane with an efficiency of 61 % (Fig. 4 red column). Furthermore, the effectiveness of TiO₂ NW-CuO-cellulose membrane was increased (Fig. 4 orange column − 51 %) comparing with the pure TiO₂ NW-cellulose membrane (Fig. 4 green column − 39 %). Although, the removal efficiency of the pure cellulose membrane was also measured it was not presented in the Fig. 3, because of it was not feasible the gravity driven filtration with these membranes.

Close

Fig. 4

E. coli removal efficiency (BFE%) of TiO₂ NW-based hybrid membranes.

Export to PPT

Based on the above-mentioned results, it was found that in terms of E. coli removal, the incorporation of TiO₂ NW-Fe₂O₃ and TiO₂ NW-CuO nanocomposites into the hybrid membrane structures provided increased bacteria removal efficiency. As can be seen in Fig. 4 the application of TiO₂ NW-based additives caused a significant improvement in the E. coli filtration efficiency in both cases. Consequently, the outstanding filtration performance of hybrid membranes indicates that the removal properties of the membranes are mainly attributed to the TiO₂ NW-Fe₂O₃ and TiO₂ NW-CuO nanocomposite additives.

3.4 MS2 filtration experiments results

The results of MS2 filtration experiments are presented in Figs. 5 and 6. It was found that both membranes were capable of retaining MS2 bacteriophages as model contaminants, but the efficiency of the membranes was significantly different.

Close

Fig. 5

MS2 enumeration plates of TiO₂-cellulose (a); TiO₂-Fe₂O₃-cellulose (b) and TiO₂-CuO-cellulose (c) membranes after filtration experiments, investigating 2 LRV value.

Export to PPT

Close

Fig. 6

MS2 retention of the TiO₂ NW-based hybrid membranes in gravity driven filtration.

Export to PPT

As can be seen in Fig. 5 c and Fig. 6 orange column the CuO-coated TiO₂ NW-based membrane provided the most promising adsorption properties. While the TiO₂-CuO-cellulose membrane showed LRV of up to 2 at pH 7 (Fig. 5 c and Fig. 6 orange column), the LRVs of TiO₂-Fe₂O₃-cellulose and TiO₂-cellulose membranes showed significantly lower virus retention capacity. In the case of TiO₂-Fe₂O₃-cellulose membrane (Fig. 5 b and Fig. 6 red column) the virus retention value was 1.2 LRV, while this value was only 0.6 LRV using TiO₂-cellulose membrane (Fig. 5 a and Fig. 6 green column). As it was previously published, during virus filtration not only the specific surface area but also the electrostatic force between the bacteriophages and the adsorbents are very significant parameters for virus rejection (Z. Németh et al., 2019). The enhanced virus retention value in the case of TiO₂-CuO-cellulose membrane can be explained by two main phenomena: the inactivation of virions by CuO nanoparticles and their surface adsorption by electrostatic force (L. Sellaoui et al., 2021).

3.5 Toxicology experiments results

The results of the toxicological tests are given in Table 4. After performing the tests, it was found that none of the membranes showed any toxic properties under the experimental set-up we used (1 membrane soaked in 500 mL dilution water or 500 mL of filtered water, respectively). None of the fish groups showed any mortality during the test period.