Manganese oxides (MnO₂) have been extensively studied due to the structural multiformity. The various structures, corresponding to different chemical and physical properties, have been widely applied in catalysis, batteries, sensors, molecular sieves, energy storage, etc. (Dawadi et al., 2020; Ye et al., 2020; Marciniuk et al., 2021; Ouyang et al., 2021; Wang et al., 2021). Particularly, δ-MnO₂ has attracted considerable attention due to its unique layered structure, where its bandgap can be tuned by filling ions between the layers (Luo et al., 2000; Wang et al., 2017). Herein, δ-MnO₂ samples with different bandgaps, doped by alkali metals such as magnesium (Mg), sodium (Na), or potassium (K), has been used as antibacterial materials.

Recently, the abuse of antibiotics is becoming a serious worldwide public health challenge, which causes antimicrobial resistance and the occurrence of superbugs (Podder et al., 2018; Teng et al., 2020). It not only prolongs treatment but also declines life expectancy due to higher morbidity/mortality risk (Podder et al., 2018; Liu et al., 2020; Estes et al., 2021). These serious public health challenges require the development of new bactericidal methods (Herman and Herman, 2014; Gatadi et al., 2021). With the development of nanotechnology, new antibacterial agents have arisen (Herman and Herman, 2014; Liu et al., 2020). These nanoscale agents may provide more effective and/or more convenient routes (Dizaj et al., 2014; Herman and Herman, 2014). Particularly, nanomaterials based on inorganic metal oxide semiconductors, i.e., CuO, ZnO, MgO, etc., have been considered as alternative antibacterial materials (Zhang et al., 2010; Bafekr and Jalal, 2018; Hong et al., 2018; Chandra et al., 2019; Ogunyemi et al., 2020; Zhang et al., 2010; Baruah et al., 2021; Haider et al., 2021; Han et al., 2021; Qian et al., 2021), whose activities can easily be tuned by altering their morphology and/or component (Prasanna and Vijayaraghavan, 2015).

Generally, reactive oxygen species (ROS) can injure biomolecules by its strong oxidation potential (Sharifi et al., 2012; Prasanna and Vijayaraghavan, 2015; Podder et al., 2018; Yao et al., 2020). The oxidant activity of metal oxide semiconductor nanomaterials usually originates from light-induced oxidative properties to generate ROS, including hydroxyl radicals (·OH), superoxide anions radicals (·O₂ ⁻), and singlet oxygen (¹O₂) (Podder et al., 2018). The generation of ROS under light exposure comes from the photo-generated electron-hole pairs excited on the appropriate band levels through the absorption of light, which interact with water and then produce ROS (Hirakawa and Nosaka, 2002; Prasanna and Vijayaraghavan, 2015). In addition to photoexcitation, the ROS can also be produced by electrons trapped by the defects at the surface of materials in the absence of light (Prasanna and Vijayaraghavan, 2015; Hao et al., 2017). However, we found that doped agents, which can provide free electrons to induce the production of ROS, also result in the antibacterial activity without external light exposure. MnO₂ has five different phases (α, β, γ, λ, and δ) and the different properties of each phase make it be extensively studied in the fields of catalysis (Suib, 2008; Truong et al., 2012; Zhou et al., 2017). Moreover, MnO₂ can effectively enhance the produce of ·OH in the aqueous solution via the excitation and formation of electron-hole pairs (Das et al., 2017; Xiao et al., 2018; Chhabra et al., 2019). Particularly, δ-MnO₂, with a unique layered crystalline structure, has aroused much investigation. By changing the quantity of filling ions between MnO₂ layers, its doping level can easily be tuned (Golden et al., 1986; Luo et al., 2008; Geng et al., 2016).

We report herein the synthesis of δ-MnO₂ nanoflowers doped by Mg, Na, and K ions to evaluate the effect of doping on the bacterial viability of Streptococcus mutans (S. mutans), a recognized cariogenic bacterium (Song et al., 2020; Afrasiabi et al., 2021). Here, the antibacterial properties of δ-MnO₂ were probed in the dark (without external illumination). It was observed that all δ-MnO₂ nanoflowers demonstrated an excellent antibacterial activity without external light exposure. Our data showed that in doped MnO₂ nanoflowers the free electrons due to doping could induce the production of ROS, resulting in the antibacterial activity. Moreover, δ-MnO₂ nanoflowers with narrower bandgap displayed a superior antibacterial ability, in which higher doping levels, providing more free electrons to induce the production of ROS, led to better antibacterial properties (following the order: K⁺>Na⁺>Mg²⁺). Furthermore, following the gradual addition of vitamin C, the δ-MnO₂ could nearly be degraded into Mn ions completely, making this materials potential for in vivo applications (Liu et al., 2021).

We started by using Mg(MnO₄)₂ as precursor to synthesize Mg-doped δ-MnO₂ materials. Figure 1A shows the SEM image of Mg-doped MnO₂, which presented a flower-like morphology and the size was ∼380 nm in diameter. Supplementary Figure S1A shows the XRD patterns of Mg-doped δ-MnO₂ nanoflowers. All the peaks correspond to the crystal planes of δ-MnO₂ (Hu et al., 2019; Li et al., 2019), and no other crystalline phases were detected. Figures 1B,C present the TEM and HRTEM images of Mg-doped MnO₂ nanoflowers. The lattice spacings of 1.42 and 2.45 Å coincide with the (110) and (101) interlayer distance in δ-MnO₂. This is also supported by the SAED patters shown in Supplementary Figure S2A.

The band gap for the Mg-doped MnO₂ nanoflowers was calculated as 1.13 eV from the UV-VIS spectrum shown in Supplementary Figure S3A. The narrower bandgap observed here relative to the previously reported (Sakai et al., 2005; John et al., 2016) might be attributed to the doping of Mg²⁺ ions between layers of MnO₂ (Luo et al., 2000; Wang et al., 2017; Hu et al., 2019; Zhu et al., 2019), which will be discussed later.

Figure 2A shows the XPS survey spectrum of Mg-doped MnO₂. It was detected that, with a ∼15 min UV exposure, surface C contamination has a ∼8 times minor intensity relative to that of Mn 2p_3/2. The binding energy (BE) values of different elements were presented in Supplementary Table S1. The Mn 2p_3/2 core-level spectrum in Figure 2D presented two components. The main peak labeled I with low BE at 642.3 eV is due to bulk-coordinated Mn, and the other peak with high BE at 645.1 eV (peak II) corresponds to MnO₂ interacted with absorbed oxygen from air (Selvakumar et al., 2015; Li et al., 2017; Hu et al., 2019; Tang et al., 2020). However, no signal corresponding to reduced Mn³⁺, which is known to occur as a result of the formation of oxygen vacancies, was observed (Selvakumar et al., 2015). Figure 2E shows the O 1s core-level spectrum, which presented three components. The main peak (BE 529.9 eV) labeled I is due to bulk-coordinated oxygen. Peak II (BE 531.5 eV) and peak III (BE 532.9 eV) are attributed to the surface component and the carbonate or hydroxyl groups chemically bound on the surface, respectively (Tang et al., 2020; Arunpandiyan et al., 2021). Figure 2F presents the Mg 1s spectrum which only has one component. The BE of Mg 1s is 1303.5 eV, which is higher than that of Mg₂Si (∼1302.3 eV) and Mg(OH)₂ (∼1303 eV) (Esaka et al., 2016; Jiang et al., 2017). The results indicate that Mg²⁺ is chemically bound to MnO₂ (Nefedov, 1977). In the Mg 1s spectrum only one component was observed, suggesting that all Mg²⁺ ions were filled between the MnO₂ layers (Hu et al., 2019).

Since the biocompatibility of materials is a prerequisite for their intended use in the human body, in vitro cytotoxicity studies of Mg-doped MnO₂ nanoflowers were carried out prior to antibacterial testing. Specifically, L929 mouse fibroblastic cells were exposed for 24 h to different concentrations (0, 100, 200, 400, and 800 μg/ml) of MnO₂ nanoflowers and then the cell viability was determined using the MTT assay, a well-established colorimetric method to evaluate the cytotoxicity of biomedical device according to ISO 10993-5:2009. After 24 h of incubation, the cell viability of L929 cells was more than 90% with the concentration of Mg-doped MnO₂ nanoflowers at 100–400 μg/ml, whereas it was reduced to 66 ± 5.7% when the dose was increased up to 800 μg/ml (Supplementary Figure S4), suggesting that the MnO₂ nanoflowers exhibited no or low cytotoxicity even at high concentration. When the concentration was higher than 400 μg/ml, an apparent decrease in cell viability was found, indicating high dose of MnO₂ nanoparticles could induce cell death, which was consistent with our previous studies showing high concentrations of silica nanoparticles could induce cell necrosis in endothelial cells (Liu and Sun, 2010). It might be attributed to the higher cellular uptake of nanoparticles, which could directly damage cell plasma membranes and thus cause cell necrosis (Liu and Sun, 2010).

Then, the antibacterial activity of Mg-doped MnO₂ sample was evaluated by detection of the bacterial viability and biofilm formation of S. mutans (Bijle et al., 2020; Daood et al., 2020; Song et al., 2020). Firstly, MTT assay was applied to evaluate the effect of Mg-doped MnO₂ sample on the bacterial viability of S. mutans. As shown in Figure 3A, Mg-doped MnO₂ can effectively inhibit bacterial viability. When the Mg-doped MnO₂ nanoflowers at 100, 200, 400, and 800 μg/ml concentration were added, the percentage of inhibition was 11.86 ± 2.18, 13.73 ± 1.45, 16.0 ± 1.83, and 20.79 ± 0.94%, respectively, reflecting that as the concentration of Mg-doped MnO₂ nanoflowers increases, the inhibition of bacterial viability shows an upward trend.

Generally, the MTT assay for assessment of antibacterial activity revealed the function of bacterial dehydrogenase system involved in metabolism (He et al., 2015). Furthermore, bacterial biofilm formation is known to increase resistance against the antibiotic and plays a critical role in the pathogenesis of infections (Banerjee et al., 2020). Then, we also examined the effect of Mg-doped MnO₂ nanoflowers on S. mutans biofilm formation by crystal violet staining (Asgharpour et al., 2019; Chan et al., 2020; Jiang et al., 2020; Song et al., 2020), and the results showed the same trend as the MTT assay (Figure 3B). In the presence of Mg-doped MnO₂, the inhibition percentage of biofilm formation was 4.1 ± 1.45% at 100 μg/ml, 21.02 ± 0.49% at 200 μg/ml, 39.86 ± 1.59% at 400 μg/ml, and 51.99 ± 1.18% at 800 μg/ml concentration, respectively, showing a dose-dependent enhanced antibacterial activity and reduced biofilm formation against S. mutans induced by Mg-doped MnO₂ nanoflowers. Interestingly, these antibacterial tests were all performed in the dark.

From the above results, it was intriguing to find that the Mg-doped δ-MnO₂ nanoflowers exhibited spontaneous antibacterial properties in the dark. Since it is established that ROS is responsible for antibacterial activity in the dark (Prasanna and Vijayaraghavan, 2015), their formation was investigated by electron spin resonance (ESR) using DMPO as a quencher without external illumination. Figures 4A,B show the characteristic DMPO-·O₂ ⁻ and DMPO-·OH signals in the case of Mg-doped MnO₂ nanoflowers at 800 μg/ml concentration (Peng et al., 2021; Zong et al., 2021). The results show that both superoxide radicals (O₂ ⁻) and hydroxyl radical ( OH ) were generated under dark conditions.

It has been reported that ZnO and MgO nanoparticles produced ROS in the dark due to the transfer of electrons trapped by the oxygen vacancies at the surface of materials (Prasanna and Vijayaraghavan, 2015; Hao et al., 2017). However, in our case, the presence of oxygen vacancies on the δ-MnO₂ surface was not detected from the XPS results (Selvakumar et al., 2015). For the Mg 1s core-level spectrum in Figure 2F, it can be observed that Mg is chemically bound to MnO₂, and thus, the valence electrons of Mg were transferred to MnO₂, becoming free electrons. It implies that the production of ROS in the dark might be produced by the free electrons in δ-MnO₂ coming from doping.

In order to further confirm this hypothesis, NaMnO₄ and KMnO₄ were employed to synthesize the Na- and K-doped δ-MnO₂ as similarly described for Mg-doped δ-MnO₂. In this case, it has been established that these δ-MnO₂ samples can enable higher doping levels relative to Mg-doped δ-MnO₂, which can lead to an increase number of free electrons (Hu et al., 2019). Figures 1D,G present their SEM images. These MnO₂ samples showed similar flower-like morphology as that of Mg-doped ones, and the sizes were ∼460 nm in diameter (Na-doped MnO₂) and ∼500 nm in diameter (K-doped MnO₂), respectively. From the XRD patterns in Supplementary Figures S1B,C, it was observed that they all presented very similar crystalline structures to Mg-doped sample. Figures 1E,H present their TEM images. The HRTEM images depicted in Figures 1F,I, as well as SAED patterns in Supplementary Figure S2B,C, display the same lattice spacings (1.42 and 2.45 Å) of MnO₂ as that of Mg-doped δ-MnO₂. The bandgaps were calculated from UV-VIS spectra (Supplementary Figures S3B,C, respectively), which were 1.06 (Na-doped MnO₂) and 0.75 eV (K-doped MnO₂), respectively.

Figures 2B,C display the XPS survey spectra of Na- and K-doped MnO₂. In Figure 2I,L, it presented only one component in both Na 1s and K 2p_3/2 spectra. The BE of Na 1s (1070.9 eV) is higher than that of NaOH (∼1069.6 eV), while the BE of K 2p_3/2 (292.5 eV) is higher than that of KF (∼292.2 eV) (Oh et al., 2013). Thus, it indicates that Na⁺ or K⁺ ions were also filled between the layers of MnO₂ (Hu et al., 2019). In Figures 2G,J, Mn 2p_3/2 core-level spectra presented 2 components, which was consistent with the Mg-doped MnO₂ sample. The BE of peak II are 644.9 (Na-doped MnO₂) and 644.7 eV (K-doped MnO₂), respectively, which are 0.2 and 0.4 eV lower than the corresponding peaks of Mg-doped MnO₂. Meanwhile, the BE of peak II and III of O 1s peak for Na-doped and K-doped samples also shift to lower BE values relative to those of Mg-doped sample. These variations demonstrate that between MnO₂ and doped ions, it occurred a charge transfer (Luo et al., 2000; Gupta et al., 2018). The reason can be explained by the different doping levels as follows.

The atomic ratios calculated from the XPS data for Mg/Mn (Mg-doped MnO₂), Na/Mn (Na-doped MnO₂), and K/Mn (K-doped MnO₂) were 1:626, 1:52, and 1:7, respectively. Details on these calculations are described in the experimental section (Hu et al., 2019; Zhu et al., 2019). This difference might be that the bigger size of ions results in larger interaction between ions and MnO₂ layers, because the ions presented the size of K⁺>Na⁺>Mg²⁺ (Hu et al., 2019). Therefore, the difference in the amounts of doping ions in MnO₂ results in the variance in the bandgap values, and meanwhile contribute to the lower BE of Mn and O peaks in Na- and K-doped samples relative to those of Mg-doped ones.

Then, MnO₂ doped by Na⁺ and K⁺ ions were employed to evaluate the effect of free electrons quantity on the bacterial viability and biofilm formation of S. mutans in the dark. In vitro cytotoxicity studies were also tested. The experimental procedure was the same as that of Mg-doped MnO₂ samples, and the results were described in Supplementary Figure S4. Consistent with the cytotoxicity results of Mg-doped MnO₂, neither Na- nor K-doped MnO₂ had a significant effect on cell viability at concentration below 400 μg/ml, while a slight reduction in cell viability was observed at 800 μg/ml. Thus, the MTT assay showed no significant cytotoxicity for MnO₂ against L929 cells when the concentration was no more than 400 μg/ml. It was noted that K-doped MnO₂ had the least cytotoxicity, while Mg-doped MnO₂ displayed higher cytotoxicity than Na-doped MnO₂ sample. According to ISO 10993-5:2009, the biocompatibility of K-doped MnO₂ was very much within the acceptable limits even at a concentration as high as 800 μg/ml.

Subsequently, we used both the MTT assay and crystal violet staining assay to evaluate the effect of MnO₂ nanoflowers with different doping on the bacterial viability and biofilm formation of S. mutans. As shown in Figure 3A, the inhibition of S. mutans bacterial viability exposed to Na-doped MnO₂ nanoflowers at 100, 200, 400, and 800 μg/ml concentration was 15.25 ± 1.11, 19.26 ± 1.39, 23.81 ± 1.83, and 27.46 ± 1.82%, respectively. Meanwhile the inhibition with K-doped MnO₂ at concentrations of 100, 200, 400, and 800 μg/ml was 16.22 ± 1.38, 23.26 ± 0.65, 27.13 ± 0.98, and 30.20 ± 0.77%, respectively, reflecting that as the concentrations of MnO₂ nanoflowers increases, the antibacterial activity shows an upward trend.

From Figure 3B, in the presence of Na-doped MnO₂, the inhibition of biofilm formation was 8.25 ± 1.48% at 100 μg/ml, 44.0 ± 2.40% at 200 μg/ml, 56.0 ± 0.46% at 400 μg/ml, and 61.57 ± 0.24% at 800 μg/ml, respectively. In the presence of K-doped MnO₂, the inhibition of biofilm formation was 35.44 ± 2.28% at 100 μg/ml, 53.56 ± 1.00% at 200 μg/ml, 62.38 ± 0.46% at 400 μg/ml, and 67.61 ± 0.61% at 800 μg/ml, respectively. Obviously, both MTT assay and crystal violet staining have shown that K-doped MnO₂ has the superior antibacterial and antibiofilm formation ability, which is better than that of Na-doped sample. Meanwhile, the Mg-doped MnO₂ had the lowest antibacterial activity. Therefore, these results confirm the hypothesis that higher doping levels could provide more free electrons, which enhanced the antibacterial properties in doped δ-MnO₂.

Figures 4A,B also show the characteristic DMPO-·O₂ ⁻ and DMPO-·OH signals in the case of Na- and K-doped MnO₂ (Peng et al., 2021; Zong et al., 2021). The results show that both O 2 − and OH produced in the dark were following the order of K⁺>Na⁺>Mg²⁺, which further confirm that the production of ROS in the dark might be produced by the free electrons in δ-MnO₂ coming from doping.

In aqueous solutions, the free electrons in δ-MnO₂ coming from doping can transfer to the water around it and form ·O₂ ⁻, while ·OH is known as a derivative of O 2 − (Morrison et al., 1988; Xu et al., 2020; Yao et al., 2020). To clarify the role of O₂ ⁻ in the mechanism of ROS production, ESR was then carried out with the addition of superoxide dismutase (SOD), a well-known superoxide scavenger (Carre et al., 2013; Piccaro et al., 2014; Jiang et al., 2016). The results in Supplementary Figure S5 demonstrate that in the dark both DMPO-·O₂ ⁻ and DMPO-·OH signal disappeared in the presence of SOD, revealing that ROS mediated through superoxide plays a major role in antibacterial activity without external illumination (Prasanna and Vijayaraghavan, 2015; Ijaz et al., 2020). Thus, the antibacterial mechanism of doped δ-MnO₂ could be proposed as in Figure 5. When the alkali metal atoms are chemically bound to MnO₂ by doping, the valence electrons of alkali metals will transfer to MnO₂. Then, the doped MnO₂ can provide free electrons to induce the production of O₂ ⁻ in aqueous solutions, which could penetrate into the bacteria cell membrane and then damage cellular components such as DNA and proteins (Xia et al., 2008; Kasemets et al., 2009; Dadi et al., 2019), resulting in the antibacterial activity without external illumination. Furthermore, by comparing the δ-MnO₂ with different doping level, the ROS generation following the doping levels of K⁺>Na⁺>Mg²⁺ can be observed. With higher doping levels, more free electrons can be transferred to MnO₂, which induce the production of more ROS, and thus present superior antibacterial activity. Compared to the antibacterial activities of MnO₂ with other iron oxide nanoparticles such as ZnO, CuO, Fe₃O₄, and Al₂O₃ (Supplementary Table S2), it was observed that MnO₂ displayed a superior antibacterial ability.

Supplementary Figure S6 presented the controllable degradation behavior of δ-MnO₂. Following the gradual addition of vitamin C into the suspension with δ-MnO₂ samples, it started to fade color, which indicated that MnO₂ could be degraded in the presence of vitamin C. When the quantity of vitamin C was 14 times over MnO₂ samples, it observed a nearly complete degradation of MnO₂ into water soluble Mn ions, which can be rapidly excreted from the body, making this materials potential for in vivo applications, presenting not only an outstanding antibacterial efficacy but also an excellent biosafety (Liu et al., 2021).

δ-MnO₂ nanoflowers doped by Mg, Na, and K ions were successful synthesized and their bandgap tunable antibacterial properties and controllable degradability mediated by vitamin C were systematically investigated. Interestingly, it was observed that all the samples showed antibacterial activity in the dark, and the antibacterial activity increased with doping levels, which was favored in the K⁺>Na⁺>Mg²⁺ order. Our data suggest that doped MnO₂ can provide free electrons to induce the production of ROS and result in the antibacterial activity in the dark. Moreover, it is shown that higher doping levels can provide more free electrons, which enhance the antibacterial activity of the δ-MnO₂ materials. Following the gradual addition of vitamin C, MnO₂ nanoflowers could nearly be degraded into water soluble Mn ions completely, indicating that these materials also display biosafety. We believe that our results shed light on the design and fabrication of antibacterial nanomaterials with tailored properties.