Disinfection of Escherichia coli by Mixing with Bulk Ultrafine Bubble Solutions

Abstract

For potential use in wastewater management and health control, this study investigates the disinfection effectiveness of bulk ultrafine bubbles (UFBs) with different bubble number densities and solution pH. Initially, neutral UFB solutions with different bubble concentrations were mixed with E. coli suspension for 120 min, but these solutions did not achieve sterilization. The bubble number density did not affect the disinfection ability of the neutral solution. Next, the pH of the UFB solutions was fixed at 5, 7, and 9. When mixed with E. coli suspension, the acidic UFB solutions reduced the colony counts by 12% after 30 min of cultivation and by 66% after 60 min of cultivation. The colony counts increased slightly in neutral and significantly in alkaline UFB solutions. The acidic UFB solutions had lower zeta potentials and smaller number densities after cultivation, implying that the number density reduced through bubble coalescence rather than increased by bubble collapse. Additionally, the UFBs exhibited insignificant fluorescence intensity, suggesting that the colony counts increased by generated ∙OH radicals. This study revealed that the effect of UFB on E. coli significantly depends on the solution pH. Further, an acidified solvent achieves a bactericidal effect, whereas a neutral or alkaline solvent enhances the growth effect. This result is important when using actual wastewater.

1. Introduction

A fine bubble is generally defined as a tiny bubble less than 100 µm in size. Fine bubbles are further classified into microbubbles (MBs) with sizes of 1–100 µm and ultrafine bubbles (UFBs) with sizes smaller than 1 µm (ISO 20480-1 2017). For fine bubbles involved in flotation, such bubbles can be classified into UFBs, MBs, and macrobubbles (CBs) [1,2]. The UFBs contain bulk nanobubbles and surface nanobubbles. Owing to their different size ranges, MBs and UFBs exhibit different characteristics. MBs appear milky white in water and rise much more slowly than ordinary bubbles. They also shrink during flotation in water and eventually disappear [3,4]. In contrast, UFBs appear clear in water because their average size (~100 nm) is smaller than the wavelength of visible light. In addition, bulk UFBs remain in water and move randomly because their buoyancy force is insignificant. The movement of UFBs is dominated by Brownian motion. Bulk MBs are unstable. Bulk UFBs exhibit long-term stability [1,5]. Despite their different features, both MBs and UFBs have important characteristics, such as negatively charged surfaces and high specific areas, that are lacking in ordinary bubbles [6,7].

Fine bubbles in water treatment can potentially reduce the facility size, cost, and operation time from those of conventional methods. Both MBs and UFBs have already been studied in water-treatment technologies such as bacterial disinfection and the decomposition of organic matter [8]. Several investigations have reported that bacteria are inactivated after interacting with fine bubbles [9,10,11]. Fine bubbles are useful in water-treatment technologies because they generate reactive oxygen species (ROS) under specific conditions. For example, ultrasonic irradiation causes MBs and UFBs to collapse with the continuous generation of ROS [9,12,13]. UFBs mainly release hydroxyl radicals (∙OH) [14], a strongly oxidative substance that can disinfect Escherichia coli and other bacteria [9]. MBs can also produce ∙OH in the absence of a dynamic stimulus. In acidic solution, an MB spontaneously collapses and produces ∙OH [3,15]; this process can be accelerated by adding a copper catalyst [16]. The possible mechanism of ∙OH generation, mediated by the collapse of MBs in acidic solutions, can be inferred from the properties of MBs. When MBs shrink in water, the increasing specific surface area accelerates the increase in negative zeta potential and ion accumulation near the bubble surface [3]. Eventually, the ion accumulation becomes extreme and sufficiently energetic to break oxygen molecules. When reacted with H⁺, this oxygen breakage creates ∙OH radicals [17]. Some studies have reported a poor efficiency of ∙OH production by MBs in alkaline solutions [15], whereas others have argued that alkaline conditions facilitate ∙OH production by MBs because OH− increases the negative zeta potential [17]. Without pH control, MBs exhibit sterilizing ability against several types of bacteria depending on the bubbling time [10].

Because UFBs and MBs have different properties, whether UFBs also exert a disinfection effect under controlled pH or bubbling time is unclear. In addition, the methods that facilitate the disinfection ability of UFBs mainly rely on dynamic stimuli such as ultrasonic waves, which obscure the true sterilization effect of UFBs. The present study investigates the inherent bactericidal ability of UFBs without an external stimulus. The experimental subject was E. coli, one of the commonest biotic indicators in water treatment. To the authors’ best knowledge, the natural disinfection ability of UFBs against E. coli has not been previously examined. As disinfectant agents without a dynamic stimulus, UFBs are expected to be superior to MBs because they are longer living in solution. In particular, UFBs persist in still water [18], so they can be potentially stored and used when required. In the present study, the bactericidal ability of UFBs against E. coli is explored in the absence of external stimuli while varying the bubbling time or pH of the UFB solution.

2. Materials and Methods

2.1. UFB Generation

The setup and specifications of the UFB generation system are presented in Figure 1 and

Table 1. Specifications of the system.

Volume of tank 1	15 L
Volume of water stored in tank 2 during running of the system	1.5 L
Height from the ground to the highest point of the system	1.5 m
Height from the ground to the bottom of tank 2	0.8 m
Inside diameter of the pipe	25 mm
Total length of the pipe	2.9 m
Pump pressure	1 MPa
Pump flow rate	41.7 L/min

, respectively. The UFBs are sustained by a circulatory system comprising a pump, two tanks, a UFB generator, and an exhaust route. Purified water (industrial purified water, MonotaRO Co., Ltd., Hyogo, Japan) was initially stored in tank 1. The pump (MTH2-5/3A–W–AQQV, Grundfos Holding A/S, Brøndby, Denmark) sent the water to tank 2 through the UFB generator (SIO, Sio Co., Ltd., Tokyo, Japan), in which UFBs were produced by a static mixer combined with a high-speed rotational device inside a channel with a Venturi effect. This UFB generator generates no microbubbles. The UFB solution in tank 2 was returned to tank 1 through the exhaust route. The UFB solutions were collected after a set bubbling time because the system was running continuously.

2.2. Preparation of Bacteria

Suspensions of E. coli NBRC 3301 (the experimental subject) were prepared by the following process. First, the E. coli cells were dispersed in culture solution (10 g/L of hipolypepton, 2 g/L of yeast extract, and 1 g/L of MgSO₄∙7H₂O) in an Erlenmeyer flask and incubated overnight at 37 °C with constant shaking at 200 rpm (BR–13UM, TAITEC Corp., Japan). The tubes were then divided into conical tubes and centrifuged at 20 °C and 3000 rpm for 10 min (CAX–371, Tomy Seiko Co., Ltd., Tokyo, Japan). Finally, the supernatant was discarded and the culture solution was resuspended in 1 mL of purified water. This E. coli suspension was reserved for subsequent experiments.

Hipolypepton, yeast extract, and MgSO₄∙7H₂O were purchased from Nihon Pharmaceutical Co., Ltd., Tokyo, Japan, Nacalai Tesque Inc., Kyoto, Japan, and Hayashi Pure Chemical Ind., Ltd., Osaka, Japan, respectively.

2.3. Experimental Methods

The disinfection ability of UFBs against E. coli was evaluated for different bubbling times and pH values of the solutions. The experimental results were evaluated by counting the number of colonies. Colony counts were determined through visual observation using a colony counter (Compact Dry, Nissui Pharmaceutical Co., Ltd., Tokyo, Japan) throughout the experiments. One milliliter of the solution was inoculated with E. coli, and the colonies after incubating for 24 ± 2 h at 35 ± 1 °C, were seen as blue dots and were, therefore, counted. Tests of both bubbling time and pH were conducted in triplicate and the average colony counts were presented as the results.

2.4. Disinfection Efficiencies of UFBs after Different Bubbling Times in Neutral Solution

One milliliter of E. coli suspension was mixed with 9 mL of each UFB solution and statically cultured for 120 min at 25 °C (WB–205QM, WakenBtech Co., Ltd., Kyoto, Japan). The pH was maintained neutral and the bubbling time was varied as 0, 10, 30, and 60 min. At 0 min (i.e., before cultivation), one milliliter of the E. coli suspension in solution was extracted for obtaining the initial colony count. Colony counts were then measured at 60-min intervals by inoculating 1 mL of the mixture in the colony counter. Prior to each inoculation, the suspension was diluted to 1 × 10⁻⁸.

2.5. Disinfection Efficiencies of UFBs in Solutions with Different pH

One milliliter of the E. coli suspension was mixed with 9 mL of UFB solution and statically cultured for 90 min at 25 °C (WB–205QM, WakenBtech Co., Ltd., Japan). The solutions were bubbled for 0 or 30 min and their pH values were maintained at 5, 7, or 9 by adding HCl or NaOH. At 0 min of bubbling time (i.e., before cultivation), one milliliter of the E. coli suspension in the pH 7 solution was extracted for measuring the initial colony count. Colony counts were then measured at 30-min intervals by inoculating 1 mL of the mixture in the colony counter. Prior to each inoculation, the suspension was diluted to 1 × 10⁻⁸.

2.6. Analytical Methods

The experimental results were analyzed in terms of their fluorescence intensity, number density, mean size, and zeta potential. The fluorescence intensity of the UFB solutions was considered to indicate their ∙OH production, because the generated ∙OH radicals were expected to sterilize E. coli. Meanwhile, the number density, mean size, and zeta potential of the UFBs were used for estimating the UFB disinfection mechanism and for analyzing the effect of bubbling time and electrolyte addition on the fundamental properties of the UFBs.

2.7. Fluorescence Intensity

The fluorescence intensity was measured in a fluorescence spectrophotometer (FP–8200, Jasco Corp., Tokyo, Japan) using aminophenyl fluorescein (APF) as the fluorescent reagent. Here APF was chosen because its own fluorescence response is almost zero and its fluorescence intensity can be observed after reaction with ∙OH [14,19]. Furthermore, although the fluorescence intensity is not a direct quantitative measure of ∙OH generation, it is positively correlated with the amount of generated ∙OH [13,14].

To measure the fluorescence intensity, 5 mM of APF (SK3002–01, Goryo Chemical, Inc., Japan) was dissolved in N,N–Dimethylformamide solution and diluted with phosphate buffer (0.1 M, pH 7.4) to 10 µM. Thereafter, 9 mL of UFB solution and 1 mL of APF solution (10 µM) were mixed to obtain a final APF concentration of 1 µM. Finally, the fluorescence intensity was measured at an excitation wavelength of 490 nm and an emission wavelength of 515 nm.

Phosphate buffer was prepared by mixing 2.875 g of Na₂HPO₄ and 0.567 g of NaH₂PO₄ in 250 mL of distilled water. Both reagents were purchased from Fujifilm Wako Pure Chemical Corp., Japan.

2.8. Number Density and Mean Size

The number densities and mean sizes of the UFBs were measured to determine whether the bubbling time increased the number density of the bubbles and whether adding electrolytes altered the fundamental properties of the UFBs. The number densities and mean sizes were determined from the particle size distributions of the UFBs obtained through a nanoparticle tracking analysis (NTA) (NanoSight, Malvern Panalytical Ltd., Malvern, UK). In NTA, the Brownian-motion tracks of the nanoparticles are visualized through the scattered reflections of laser light and the size of each nanoparticle is computed [20].

2.9. Zeta Potentials

The zeta potential is strongly associated with the stability of fine bubbles. A higher absolute value of the zeta potential indicates a high stability of the fine bubbles, whereas a lower absolute value indicates unstable bubbles that will probably coalesce. Previous investigations have reported that when electrolytes are dissolved in fine bubble solutions, the ions of the electrolyte (such as H⁺ and OH⁻) largely determine the zeta potential of the fine bubbles, which are negatively charged in nature [21,22]. Hence, in the present study, the zeta potentials of the UFB solutions were measured to detect the influencing degree of the added electrolytes on the surface charges of the bubbles. Zeta potentials were measured using an electrophoretic light-scattering method (ELSZ–1TY, Otsuka Electronics Co., Ltd., Osaka, Japan).

3. Results and Discussion

3.1. Disinfection Efficienciesof UFBs after Different Bubbling Times in Neutral Solution

The number densities and mean sizes of the particles in the UFB solutions after 10, 30, and 60 min of bubbling time are shown in Figure 2, and the colony counts for each bubbling time are plotted as functions of cultivation time in Figure 3. Increasing the bubbling time increased the particle-number density but exerted no systematic effect on the mean size. The number of colonies did not significantly decrease in Figure 3, indicating that the number density of UFBs did not significantly affect the sterilization efficiency in the neutral solution. On the contrary, the addition of UFB slightly increased the number of E. coli. At a cultivation time of 120 min, colony numbers at 60 min increased by ~36% compared to the numbers corresponding to the without-UFB case.

Unlike UFBs, MBs can disinfect bacteria in the absence of stimuli and their sterilization efficiency depends on the bubbling time [10]. The different behaviors of MBs and UFBs might be explained by the high stability of UFBs in water [1,5]. Although UFBs can potentially generate ∙OH after collapse, they are more resistant to collapse than MBs [23]. Shrinkage of MBs spontaneously provides sufficient energy to generate ∙OH at the gas–water interface [18,19].

3.2. Disinfection Efficiency of UFBs in Solutions with Different pH

Figure 4 plots the colony counts as functions of cultivation time in E. coli suspensions with different pH values. Over time, the number of colonies decreased only in the acidic solutions. More specifically, the colony counts in the acidic UFB solutions significantly declined during the first 60 min. Colonies with UFB decreased by 12% at a cultivation time of 30 min and by 66% at 60 min compared to the without-UFB case. In contrast, the colony counts in the alkaline solutions (pH = 9) exceeded the detectable limit of the colony counter and were therefore excluded from Figure 4.

The reduced colony counts in the acidic UFB solutions imply that the UFBs collapsed in the acidic solutions and produced larger amounts of ∙OH than those in neutral and alkaline solutions. When UFBs collapse, their number density should decrease. Figure 5 compares the number densities and mean sizes of the particles in the UFB solutions with different pH values. The mean sizes were larger and the number densities were smaller in the acidic UFB solution than in the other solutions. However, the zeta potentials of the UFBs were less negative in acidic solutions than in neutral and alkaline solutions (Figure 6). The low number densities and small negative zeta potential indicate that rather than collapsing, the bubbles coalesced in the acidic UFB solution. This phenomenon is likely explained by the tendency of charged particles to aggregate when their absolute values of surface electrical potential are low. Meanwhile, the largish mean bubble size can be explained by coalescence and absorption of H⁺ ions at the bubble surfaces in the acidic solution. Because UFBs are generally negatively charged [6], they will attract positively charged ions such as H⁺ produced by the dissolution of HCl in solution, thereby inhibiting their collapse.

The colonies were not disinfected in the alkaline UFB solution. This finding contradicts a previous investigation, in which the alkaline environment increased the negative zeta potential of the MBs and hence their ability to decompose organic matter through ∙OH generation [17]. Although the present UFBs established a higher negative zeta potential in the alkaline solution than in the neutral and acidic solutions (Figure 6), the number density of the UFBs in the alkaline solution was close to that in the neutral solution (Figure 5), indicating that significant changes such as bubble collapses were rare. Therefore, it can be concluded that despite the increase in negative zeta potential of the UFBs in the alkaline solution, the energy was insufficient to ensure electrically and hydrodynamically stable UFBs [23,24].

Figure 7 compares the fluorescence intensities of the solutions after reacting with the fluorescence reagent, APF, as a proxy of ∙OH detection. In the solutions with and without bubbles (bubbling time 30 and 0 min, respectively), the fluorescence intensity tended to increase with pH. The alkaline solutions exhibited the highest and second–highest fluorescence intensities, and the acidic solutions exhibited the lowest and second–lowest fluorescence intensities. Moreover, the intensities were similar in the acid and neutral solutions with and without UFBs, but were higher in the alkaline UFB solution than in the alkaline solution without UFBs. These results suggest higher ∙OH content in alkaline solutions than in acidic solutions. These results contradict the expected result, namely, that UFBs exert a sterilizing effect in acidic solution.

The small fluorescence intensity of the acidic UFB solution, despite the fewer colony counts than in the neutral solution, might be explained by the removal of ∙OH radicals before the measurement was performed. Because ∙OH has an unpaired electron, it has a noticeably short lifetime (a half-life of 10⁻⁹ s) and easily reacts with other chemicals [25]. This implies that ∙OH was converted into other chemicals and was therefore undetected by the fluorescent spectrophotometer, which relies on the reaction with APF.

The alkaline UFB solution exhibited the highest fluorescence intensity but did not inhibit colony growth. This result indicates that ∙OH production from the UFBs under alkaline conditions was insufficient to disinfect the contaminating bacteria. In a previous investigation of the fluorescence response after reacting UFB solutions with APF at the same concentration under ultrasonic irradiation [13], the fluorescence intensity was much stronger than observed in the alkaline UFB solution of the present study.

3.3. Bacterial Disinfection Mechanism of UFBs

In previous investigations, ∙OH production by MBs in acid solution was mainly attributed to the increased negative zeta potential during bubble shrinkage, the breakage of oxygen molecules, or the reaction between oxygen radicals and H⁺ [3,17]. In addition, the zeta potential of the bubbles was thought to be determined by two main factors: the amount of electric charge per unit surface area and the absorption of electrolytes [3]. The shrinkage of fine bubbles increases the electric charge concentration on the bubble surface and increases the negative zeta potential. Electrolyte absorption on the bubble surface increases or decreases the zeta potential depending on the type of electrolyte used.

The high shrinkage rate increases the surface electrical charge per unit area and increases the negative zeta potential on MBs [3]. As shown in Figure 6, the UFBs insignificantly affected the negative zeta potential in acidic solutions. It was estimated that the UFBs did not significantly shrink, so the negative zeta potential did not increase. In fact, the negative zeta potential was reduced by absorption of H⁺ ions.

Unlike MBs, the present UFBs did not produce ∙OH so the bacteria were possibly disinfected by other processes. One mechanism was suggested by Wang et al. [26], who created UFBs of air, nitrogen, carbon dioxide, and hydrogen, to measure their physical properties. They showed that only the carbon dioxide UFBs effectively dissolved and decreased the pH of the solution. In our experiment, the pH surrounding the E. coli decreased because carbon dioxide was generated around the growing E. coli cells. The pH reduction does not affect E. coli in the neutral or alkaline range, but if the pH descends into the acidic range where E. coli can barely survive, the solution is eventually sterilized. However, chlorine ions contained in the solution have a strong oxidizing effect. The sterilization effect was strengthened (hastened) because contact between the chlorine ions and E. coli was likely due to a decrease in the surface tension of fine bubbles [1] and Brownian motion.

The significant growth effect of E. coli in the alkaline region is similar to the effect of plant growth promotion. Park and Kurata [27] found that, for lettuces grown under similar dissolved oxygen concentrations, the fresh weights of microbubble-treated lettuces were 2.1 times higher than those of the macrobubble-treated lettuces. They surmised that large specific surface area and surface negative electronic charges of the microbubbles may promote lettuce growth because the bubbles can attract positively charged ions dissolved in the nutrient solution. As shown in Figure 5 and Figure 6, the alkaline and neutral solutions have a larger number of bubbles and higher negative zeta potential than the acidic solution. Therefore, the amount of dissolved oxygen may have increased and the cations in the nutrients were probably attracted to the UFB, facilitating sufficient distribution of nutrients around E. coli and proliferation of the cells. The opposite scenario occurs in the acidic solution. That is, in addition to the decrease in pH, the E. coli are inefficiently surrounded by the nutrients due to the decrease in the number of bubbles and zeta potential, thereby strengthening the disinfecting effect.

4. Conclusions

The E. coli disinfection efficiency of UFBs was investigated for different bubbling times and solution pHs. First, neutral UFB solutions with different bubbling times were mixed with the E. coli suspension. Lengthening the bubbling time increased the number density of the bubbles but the UFB concentration did not elicit a bactericidal effect under neutral conditions. This finding indicates that the bubbles were too excessively stable to collapse, although they produced ∙OH voluntarily without any dynamic or hydro–chemical stimuli. In the next test, the bubbling time was fixed and UFB solutions were prepared at different pHs (5, 7, and 9) before mixing with the E. coli suspension. The colony counts were decreased only in acidic UFB solution. The acidic UFB solution reduced the number of colonies by 12% compared with the numbers corresponding to the without-UFB case at 30 min incubation time and by 66% at 60 min incubation time. There was no difference at 90 min. However, the sterilization mechanism could not be probed by measuring the fluorescence intensity as a proxy of ∙OH production, the zeta potential, the number density, or the mean bubble size. The low fluorescence intensities indicated that the ∙OH levels were too low for disinfection activity. In contrast, the colony counts increased after inoculation in alkaline UFB solutions, despite the increased negative zeta potential and strengthening of the fluorescence intensity in alkaline solution. It was estimated that the negative zeta potential was raised insufficiently to generate ∙OH at the gas/water interface of the electrically and hydrodynamically stable UFBs; accordingly, the ∙OH generation was insufficient to disinfect the bacteria.

This study revealed that the effect of UFBs on E. coli depends strongly on the pH of the solution. Moreover, the solvent should be acidified to achieve a bactericidal effect, and conversely, should be neutralized or alkalinized to enhance the growth effect. This suggests that, if the pH of the liquid is known when UFBs are used in practical applications, the type of effect exerted by these UFBs can be predicted. Therefore, the findings of this study provide useful information for the practical application of UFBs.