Introduction

With the prosperity of shipping industry, the discharge of ship ballast water has become a serious threat to marine ecological environment. It is estimated that about 10 billion tons of ballast water are transferred worldwide each year, resulting in unexpected introduction of non-indigenous organisms to their new habitat. Once the alien organisms (such as microalgae and bacteria) adapt to the new environmental conditions, their proliferation might lead to severe ecological disaster, such as algal blooms. In addition, many researches indicated that heterotrophic microorganisms (bacteria, viruses, etc.) could be introduced into uncontaminated areas by ship ballast water, posing a potential hazard to human health (Leichsenring and Lawrence 2011; Seiden et al. 2010). The International Maritime Organization (IMO) has set stringent standards on living bacteria in discharged ballast water (Escherichia coli: < 250 cfu/100 mL; Enterococci: < 100 cfu/100 mL), and the development of efficient bacterial inactivation techniques has become a hot spot in the field (Tao et al. 2017).

Generally, ship ballast water treatment system is composed of a filtration unit and an inactivation unit. Since the main effort of the filtration unit is to keep out large planktons, bacterial inactivation is mainly achieved in the inactivation unit (Tang et al. 2009; Tsolaki and Diamadopoulos 2010). Among various disinfection methods, ultraviolet (UV) radiation is considered a favorable method with its significant advantages of simple unit configuration, high adaptability under different salinities, none of corrosion to pipelines, and no harmful oxidant substances production (Mamlook et al. 2008). However, the inactivation effect of the UV radiation can be greatly limited at high seawater turbidities, and bacteria can survive the UV treatment due to the effect of photoreactivation and dark repair. Recently, combined UV and nano-TiO2 photocatalysis technology has caught much sight to inhibit the photoreactivation and dark repair of bacteria (Shang et al. 2009). Under UV irradiation, nano-TiO2 catalyst can generate a large number of reactive oxygenated species, which are highly effective to deconstruct cellular structure, accelerate the death of microorganisms, and prevent the regeneration of harmful microorganisms in the ballast water (Martínez et al. 2013).

The combined UV and nano-TiO2 technology has proven an efficient disinfection alternative for ballast water treatment (Roy et al. 2018), and its restrictions mainly lie in the inherent characteristics of nano-TiO2. For instance, TiO2 has a large band gap energy (3.2 eV for anatase phase and 3.0 eV for rutile phase), and the spectral response range is narrow. In order to improve material properties, the modification of nano-TiO2 by doping other elements has attracted much attention (Banerjee et al. 2016). By doing this, the separation probability of electrons and holes can be increased, and the recombination of electron-hole pairs can be inhibited, resulting in improved photocatalytic performance of nano-TiO2. Some kinds of elements have been investigated, such as Au (Radzig et al. 2019; Veziroglu et al. 2018), Ag (Rtimi et al. 2019), Cu (Wu et al. 2018), Co (Smirnova et al. 2017), Fe (Mangayayam et al. 2017; Yu et al. 2019), and nonmetal elements such as F and N (Makropoulou et al. 2018; Milosevic et al. 2017). However, while using the combined UV and metal-doped nano-TiO2 technology, the effect of doping elements (type and ratio) on bacterial inactivation was less investigated, and there is still a lack of knowledge on bacterial inactivation mechanisms.

In our previous researches, the inactivation effect of UV-TiO2 on microalgae and Escherichia coli has been investigated, and the results proved the effectiveness of the UV-TiO2 technique on treating ballast water containing high contents of suspended matter (Lu et al. 2018, Lu et al. 2019). In this study, three representative metal elements (Ag, Fe, and Gd) were doped to nano-TiO2 at different ratios (0.4%, 0.7%, and 1.0%) to generate metal-doped nano-TiO2 materials. The inactivation effect of sole UV, UV and nano-TiO2, and UV and metal-doped nano-TiO2 was compared with Escherichia coli and Enterococci as the inactivation objects. The prepared metal-doped nano-TiO2 materials were characterized by scanning electron microscope (SEM), X-ray diffraction (XRD), and X-ray energy dispersive spectroscopy (EDS). Malondialdehyde (MDA) concentration and superoxide dismutase (SOD) activity were analyzed to reveal bacterial inactivation mechanisms of the combined UV and metal-doped nano-TiO2 technology.

Materials and methods

Artificial ballast water and bacterial cultivation

Artificial ballast water was prepared according to Mocledon’s artificial seawater formula. The 200-mesh calcium-based organic bentonite was supplemented as an additive to make the ballast water having a turbidity of 10 NTU. The pH of the artificial ballast water was adjusted to 8.0. As the inactivation objects, Escherichia coli and Enterococcus were purchased from China General Microbiological Culture Collection Center (CGMCC). Prior to the inactivation process, the two kinds of bacteria were inoculated in autoclaved Luria broth liquid medium. The bacterial culture was carried out in an incubator at 37 °C for 18–24 h and then ready for the following inactivation experiments.

Preparation of original and metal-doped nano-TiO2

The original and metal-doped nano-TiO2 were prepared by the sol-gel method (Yang et al. 2018). The procedures were as follows: 10 mL tetrabutyl titanate solution was slowly added into 35 mL anhydrous ethanol. The mixture was thoroughly agitated for 10 min by using a magnetic stirrer, generating solution A. In addition, 35 mL anhydrous ethanol, 4.08 mL deionized water, and 4.08 mL glacial acetic were mixed and supplemented with nothing (for preparing original nano-TiO2), silver nitrate (for doping Ag), ferric nitrate nonahydrate (for doping Fe), and gadolinium nitrate hexahydrate (for doping Gd), respectively. For each doped metal element, the doping ratio was 0.4%, 0.7%, and 1.0%, respectively. The mixture was named solution B, and its pH was adjusted to less than 3.0 by using nitric acid. Afterwards, while the solution B was heated and stirred in a water bath at 30 °C, the solution A was slowly added into the solution B at a speed of one drop per second. After the stirring of 1 h, the mixture was called solution C. The catalyst in the solution C was loaded on the pretreated nickel foam by using a dip coater, and the loading times were fifteen. After that, the loaded nickel foam was dried at 80 °C for 10 min, put into a muffle furnace at 445 °C for 2 h, and then insulated for 12 h. The original and metal-doped nano-TiO2 were then prepared for the further use.

Inactivation device and operation process

An inactivation device was designed to explore inactivation effect and mechanisms of Escherichia coli and Enterococcus using sole UV, UV and nano-TiO2, and UV and metal-doping nano-TiO2. Figure 1 displays the schematic diagram of the inactivation device, which was mainly composed of a UV lamp, a light shielding plate, a glass reaction vessel (containing bacterial suspension diluted from the cultural solution), and a magnetic agitator. Before the inactivation process, the UV lamp was preheated for 15 min to obtain a stable UV light intensity. Escherichia coli and Enterococcus were separately used as the inactivated object with an initial concentration of about 104 cfu/mL. The device was operated for inactivating the two bacterial suspensions using the sole UV treatment, the combined UV and original nano-TiO2 treatment, and the combined UV and metal-doped nano-TiO2 treatment, respectively. During the treatment process (0 s, 5 s, 10 s, 15 s, and 20 s), bacterial suspension was obtained to analyze the number of living bacteria. In addition, a continuous flow device was used to analyze MDA content and SOD activity as previously described (Lu et al. 2018). For each trial, three parallel samples were analyzed to make sure the data reproducibility. All experiments were carried out at a room temperature of 22 °C.

Fig. 1
figure 1

Schematic diagram of the inactivation device

Full size image

Inactivation rate of Escherichia coli and Enterococcus

Escherichia coli and Enterococcus were numbered according to the following procedures. The bacterial suspension was proportionately diluted and then filtered by a vacuum pump. Bacteria were then entrapped by the membrane, and the membrane was attached to the surface of the Luria broth solid medium for further cultivation (37 ± 0.5 °C for 24 ± 1 h). Afterwards, the number of colonies was counted to calculate bacterial concentrations in the suspension. The logarithmic sterilization rate (η) was used to reflect the inactivation rate of bacteria in the device, and the calculation formula is η = − lg(Nt/N0), where N0 is the number of living bacteria before the inactivation treatment and Nt is the number of bacteria in the sample at different treatment times.

Materials characterization

The microstructure and morphology of the prepared materials (original nano-TiO2 and metal-doped nano-TiO2) were analyzed using SEM (Quanta 200, FEI, Netherlands), and the EDS analysis of the materials was performed by using an energy dispersive spectrometer (TEAM™, EDAX, USA). The crystal structure of the materials was analyzed by an XRD analyzer (D/max 2550 VB+, Rigaku, Japan). The radiation source was Cu Kα (λ = 0.15406 nm), the scanning range was 10–80°, the tube voltage was 40 kV, the current used was 30 mA, and the scanning speed was 5°/min. The average crystallite size D was calculated by the Scherrer formula D = Kγ/(Bcosθ), where D is the average thickness of the grain perpendicular to the crystal plane direction, K is the Scherrer constant, γ is the X-ray wavelength, which is 0.154056 nm, B is the half-height width of the diffraction peak of the measured sample, and θ is the diffraction angle.

Measurement of MDA content and SOD activity

As an important indicator reflecting cell membrane damage, MDA content was measured using the thiobarbituric acid method (Esterbauer 1996). With the capability of eliminating harmful substances produced by cell metabolism, the activity of SOD was analyzed using the pyrogallol method (Marklund et al. 1974). The detailed analytical procedures of MDA content and SOD activity were in accordance with the previous studies.

Results and discussion

Inactivation effect of the combined UV and Ag, Fe, and Gd-doped nano-TiO2

The inactivation effect of the combined UV and Ag, Fe, and Gd-doped nano-TiO2 was investigated with Escherichia coli and Enterococcus as the inactivation objects. Figure 2 shows the logarithmic sterilization rates of the UV and Ag-doped nano-TiO2 treatment. Compared with the sole UV and the UV and nano-TiO2 treatment, the doping of Ag significantly improved the inactivation effect on the two kinds of typical bacteria. Within the Ag content range in this study, the higher the Ag content in the nano-TiO2 material was, the higher the inactivation rates of the two bacteria were. With the doping of 1% Ag, the logarithmic sterilization rates of Escherichia coli and Enterococcus reached 0.915 and 0.805, respectively, at the end, which were 0.266 and 0.179 higher than those of the UV and nano-TiO2 treatment. Figure 3 displays the logarithmic sterilization rates while using the UV and Fe-doped nano-TiO2. It indicates that the doping of Fe slightly increased the inactivation effect of Escherichia coli and Enterococcus. The doping of 1% Fe reached the optimal logarithmic sterilization rates (0.772 for Escherichia coli and 0.746 for Enterococcus), which were 0.123 and 0.120 higher than those of the UV and nano-TiO2 treatment, respectively. The logarithmic sterilization rates using the UV and Gd-doped nano-TiO2 are shown in Fig. 4. The introduction of Gd could also increase the inactivation rate, but the improvement was less obvious. The logarithmic sterilization rates of Escherichia coli and Enterococci using the UV and Gd-doped nano-TiO2 treatment (1% Gd content) were 0.735 and 0.720, respectively, which were only 0.086 and 0.094 higher than the UV and nano-TiO2 treatment.

Fig. 2
figure 2

Logarithmic sterilization rate for the sole UV, the UV and original nano-TiO2, and the UV and Ag-doped nano-TiO2 treatment. a Escherichia coli and b Enterococci

Full size image
Fig. 3
figure 3

Logarithmic sterilization rate for the sole UV, the UV and original nano-TiO2, and the UV and Fe-doped nano-TiO2 treatment. a Escherichia coli and b Enterococci

Full size image
Fig. 4
figure 4

Logarithmic sterilization rate for the sole UV, the UV and original nano-TiO2, and the UV and Gd-doped nano-TiO2 treatment. a Escherichia coli and b Enterococci

Full size image

Under UV radiation, electrons of TiO2 photocatalyst could be excited from valence band to conduction band, and electron/hole pairs then move to material surface and initiate the generation of reactive oxygenated species, resulting in the inactivation of bacterial cells. The high degree of recombination between photo-induced electrons and holes is a key restricting factor on the photocatalytic efficiency, and metal doping proved to be an effective way to solve the problem. In this study, the perfect inactivation effect of the UV and Ag-doped nano-TiO2 might be mainly due to the improvement effort of Ag in nano-TiO2. Many studies have proved that Ag could capture photo-generated electrons, inhibit the recombination of electrons and holes, and generate more hydroxyl radicals for inactivation (Bahadur et al. 2016; Rtimi et al. 2019). With high concentrations of reactive oxygenated species, phospholipid peroxidation led to significant breakdown of bacterial membrane structure and inhibition of bacterial DNA replication, resulting in the death of bacteria. Moreover, it has been reported that the Ag-TiO2 nanoparticles had better antibacterial activity through theoretical quantum chemical calculation and electronic analysis (André et al. 2015). Based on the results, the doping of Ag was an effective strategy to improve the inactivation effect of the UV and nano-TiO2 treatment.

Experimental results indicate that the doping of Fe could also improve the sterilization effect of nano-TiO2. Previous study has reported that the Fe-TiO2 film has better photocatalytic effect than the sole TiO2 (Sun et al. 2009). Since Fe and Ti are of the very similar ionic radius, Fe in the TiO2 film could replace Ti in the lattice, resulting in efficient separation and recombination of photo-generated electrons and holes (Sun et al. 2009, Wang et al. 2003). For the doping of Gd, the insignificant improvement might be because the doping of Gd could simultaneously cause positive and negative efforts on TiO2 photocatalysis. On the positive aspects, the incorporation of Gd into TiO2 could shorten the charge transfer distance (Reddy et al. 2001). The doping of Gd also helps to transfer light-induced charge carriers to surface-adsorbed materials, improving photocatalytic activity by enhancing the electronic band structure and light response. On the opposite, the ionic radius of Gd is larger than Ti, so that Gd locates in the interstitial position and induces a stress in the matrix (Baiju et al. 2010). Hence, Gd might not be suitable as the additive of the nano-TiO2 for bacterial inactivation.

Microstructure and composition of the Ag, Fe, and Gd-doped nano-TiO2

Microstructure of the original nano-TiO2, 1% Ag-doped nano-TiO2, 1% Fe-doped nano-TiO2, and 1% Gd-doped nano-TiO2 was analyzed by SEM analysis (Fig. 5). It indicates that the doping of Ag, Fe, and Gd prevented the aggregation of nanoparticles on the surface of the material. Among the four materials, the microstructure of 1% Ag-doped nano-TiO2 showed the smallest particle diameter and the evenest distribution of the nanoparticles, which was in consistent with the preferable inactivation effort in this case. For 1% Ag-doped nano-TiO2, the smallest grain size after doping meant the largest specific surface area for photocatalysis. Under the UV radiation, there were more photocatalysis sites on the material surface. More reactive oxygenated species could be induced for 1% Ag-doped nano-TiO2, causing the deconstruction of the bacterial cell wall and plasma membrane, the leakage of cellular contents, and then cell death (Allahverdiyev et al. 2011; Li et al. 2012; Yang et al. 2013).

Fig. 5
figure 5

SEM images of the four kinds of materials. a Nano-TiO2, b 1% Ag-doped nano-TiO2, c 1% Fe-doped nano-TiO2, and d 1% Gd-doped nano-TiO2

Full size image

The composition of the original nano-TiO2, 1% Ag-doped nano-TiO2, 1% Fe-doped nano-TiO2, and 1% Gd-doped nano-TiO2 was determined by the EDS analysis (Fig. 6). It indicates that small amounts of Ag, Fe, and Gd were separately doped into the original nano-TiO2. For the four materials, the crystal morphology and phase were analyzed by the XRD analysis (Fig. 7). These samples all corresponded to the crystal structure of anatase TiO2 (JCPDS 01-084-1285) (Durupthy et al. 2007). The average particle diameters of the original nano-TiO2, 1% Ag-doped nano-TiO2, 1% Fe-doped nano-TiO2, and 1% Gd-doped nano-TiO2 were calculated by Scherrer’s formula to be 18 nm, 11 nm, 13 nm, and 15 nm, respectively. The antibacterial activity of nano-TiO2 is closely related to its crystal state, and the anatase phase TiO2 can exert better catalytic effect than rutile and amorphous TiO2 (Xue et al. 2013). The doping of Ag, Fe, and Gd did not change the crystal form of the anatase phase but decreased the grain size to varying degrees. The increased specific surface area might be a key reason for the improvement of photocatalytic inactivation effect.

Fig. 6
figure 6

EDS images of the four kinds of materials. a Nano-TiO2, b 1% Ag-doped nano-TiO2, c 1% Fe-doped nano-TiO2, and d 1% Gd-doped nano-TiO2

Full size image
Fig. 7
figure 7

XRD images of the four kinds of materials. a Nano-TiO2, b 1% Ag-doped nano-TiO2, c 1% Fe-doped nano-TiO2, and d 1% Gd-doped nano-TiO2

Full size image

MDA content and SOD activity during the inactivation process

Under harsh environments, a large amount of reactive oxygenated species is accumulated in the cells to generate more active free radicals. Cell membrane undergoes lipid peroxide reaction to form MDA, which is considered a good indicator to reflect the degree of damage to the microbial cells (Mateos et al. 2005). For Escherichia coli and Enterococci, Fig. 8 shows MDA concentrations under the treatment of the sole UV and the UV and Ag-doped nano-TiO2. Compared with the approximate linear increase of MDA content by the sole UV treatment, the MDA concentration increased rapidly under the UV and Ag-doped nano-TiO2 treatment. At the end of the inactivation process, the MDA concentration under the UV and Ag-doped nano-TiO2 treatment reached 0.0646 μmol/L for Escherichia coli and 0.0529 μmol/L for Enterococci, which were about 1.61 times and 1.71 times the values of the sole UV treatment (0.0402 μmol/L for Escherichia coli and 0.0309 μmol/L for Enterococci), respectively. Since the concentration of MDA can reflect the degree of damage of reactive oxygenated species in cells (Halliwell and Chirico 1993; Onorato et al. 1998), the high MDA contents under the UV and Ag-doped nano-TiO2 treatment indicate that much more reactive oxygenated species were generated in these cases, leading to high inactivation rates of Escherichia coli and Enterococci.

Fig. 8
figure 8

MDA concentration under the treatment of the sole UV and the UV and Ag-doped nano-TiO2. a Escherichia coli and b Enterococci

Full size image

As an important antioxidase, SOD scavenges superoxide radicals (O2−) in organisms to protect cells from the damage of free radicals (Huang et al. 2000). SOD activity is a key factor influencing self-repair process within the cells. The SOD activities in Escherichia coli and Enterococcus under the sole UV treatment and the UV and Ag-doped nano-TiO2 treatment are shown in Fig. 9. Compared with the sole UV treatment, the SOD activity under the UV and Ag-doped nano-TiO2 performed a faster decrease trend during the process. At the end of the inactivation process, the SOD activity under the UV and Ag-doped nano-TiO2 was 0.672 U/mL for Escherichia coli and 0.792 U/mL for Enterococci, which were significantly lower than those of the sole UV treatment (1.207 U/mL for Escherichia coli and 1.187 U/mL for Enterococci). With low SOD activities, normal cell metabolism of Escherichia coli and Enterococci was greatly inhibited by a large number of reactive oxygenated species generated during the UV and Ag-doped nano-TiO2 treatment. The accumulation of reactive oxygenated species accelerated the deconstruction of bacterial cell structure and then led to the death of bacteria. The results indicate that the UV and Ag-doped nano-TiO2 treatment surely has more remarkable inactivation effect than the sole UV treatment.

Fig. 9
figure 9

SOD activity under the treatment of the sole UV and the UV and Ag-doped nano-TiO2. a Escherichia coli and b Enterococci

Full size image

Conclusions

To improve bacterial inactivation effect for ship ballast water treatment, Ag, Fe, and Gd were doped to nano-TiO2 at different ratios (0.4%, 0.7%, and 1.0%) to prepare metal-doped nano-TiO2. Compared with the sole UV and the UV and nano-TiO2, the UV and metal-doped nano-TiO2 increased the inactivation rate of Escherichia coli and Enterococci to different extents. For each metal, high external metal content (1.0%) corresponded to high inactivation effort. Among the three metal elements, the doping of Ag resulted in the optimal inactivation effect, and the addition of Fe and Gd caused unobvious efforts. At the treatment time of 20 s, the UV and 1% Ag-doped nano-TiO2 reached the highest logarithmic sterilization rates of Escherichia coli and Enterococcus with the values of 0.915 and 0.805, respectively. SEM and XRD analysis indicate that the doping of Ag, Fe, and Gd did not change the anatase phase TiO2 crystal form and the microstructure of 1% Ag-doped nano-TiO2 performed the smallest particle diameter and the evenest distribution of nanoparticles. Compared with the sole UV treatment, the UV and Ag-doped nano-TiO2 treatment resulted in higher MDA contents (0.0646 μmol/L for Escherichia coli and 0.0529 μmol/L for Enterococci) and lower SOD activities (0.672 U/mL for Escherichia coli and 0.792 U/mL for Enterococci), which provides theoretical explanation for high inactivation rates in these cases.