Self-locomotive, antimicrobial microrobot (SLAM) swarm for enhanced biofilm elimination

Abstract

Biofilm is a major cause of infections and infrastructure deterioration, largely due to molecular diffusion restrictions that hamper the antimicrobial activity of traditional antibiotics and disinfectants. Here, we present a self-locomotive, antimicrobial microrobot (SLAM) swarm that can penetrate, fracture, and detach biofilm and, in turn, nullify bacterial resistance to antibiotics. The SLAM is assembled by loading a controlled mass of manganese oxide nanosheets on diatoms with the polydopamine binder. In hydrogen peroxide solution, SLAMs produce oxygen bubbles that generate thrust to penetrate the rigid and dense Pseudomonas aeruginosa biofilm and self-assemble into a swarm that repeatedly surrounds, expands, and bursts oxygen bubbles. The resulting cavities continue to deform and fracture extracellular polymeric substances from microgrooved silicone substrates and wounded skin explants while decreasing the number of viable bacterial cells. Additionally, SLAM allows irrigating water or antibiotics to access the residual biofilm better, thus enhancing the synergistic efficacy in killing up to 99.9% of bacterial cells.

1. Introduction

According to the World Health Organization, a post-antibiotic era is approaching faster than expected due to the increased antimicrobial resistance of microorganisms [1]. In general, bacterial and fungal cells are responsible for various infection cases, such as chronic wound and urinary tract infections [2-5]. These microbial cells also result in infrastructure decay and corrosion, indirectly threatening human health and sustainability [6-9]. Various disinfectants and antibiotics are available to inhibit cell growth or kill cells, but their efficacy in treating microbial cells in biofilms is extremely low because of limited diffusion within extracellular polymer substances (EPS) [10-12]. Therefore, microbial cells in biofilms may be 10–1000 times more resistant to antibiotics than free-floating, planktonic cells [13-15] As a result, antibiotics administered to biofilm kill microbial cells only at the biofilm margin while allowing those deep inside the biofilm to remain viable [10,16].

As such, biofilm removal strategies increasingly center on administering antimicrobial agents within biofilms to enhance drug efficacy [17-22]. In most scenarios, antimicrobials are loaded into nanocarriers like liposomes [20] and polymeric nanoparticles [22] for delivery. Those nanocarriers not only can protect antimicrobials from deactivating substances in biofilms (e.g. antibiotic-inactivating enzymes) [17,20,23] but also can penetrate biofilms through the interaction between nanocarriers and EPS matrix or bacterial cells (e.g. electrostatic interaction) [21,24]. Furthermore, on-demand or selective delivery of antimicrobial agents into biofilms can be achieved by modifying the structures of nanocarriers with stimuli-responsive molecules that target endogenous enzymes and lower pH in the biofilm environment [25-27]. However, even though nanocarriers improve antimicrobial efficacy to kill bacteria, a significant fraction of the EPS remains to rebuild new biofilms. Thus, it is imperative to develop new strategies to combat biofilm by simultaneously removing EPS and killing bacterial cells.

Recently, we demonstrated that diatoms engineered to generate oxygen bubbles could invade and fracture Escherichia coli biofilm in confined space [28]. To the best of our knowledge, this is the first active matter design for biofilm removal application. Then, similar micro/-nanorobots have been reported to clean biofilms mainly formed on an open and flat surface [29-37]. However, mechanical rigidity and density of biofilm vary with microbial cell types and host structure [38]. For instance, Pseudomonas aeruginosa secrets abundant exopolysaccharides to form a more robust biofilm than E. coli, demonstrating a remarkable self-healing activity after mechanical yielding [39-41]. In addition, the biofilm that forms in a confined environment becomes sturdier and denser than that forms on a flat substrate [42]. Therefore, it is essential to engineer a more powerful microrobot that can clean P. aeruginosa biofilm in confined space.

To this end, this study demonstrates a self-locomotive, antimicrobial microrobot (SLAM) swarm that can generate oxygen bubbles at controlled rates and eliminate biofilm via collective microbubbling (Scheme 1). The SLAMs are rod-shaped, hollow diatom particles loaded with manganese oxide (MnO₂) nanocatalysts, which endow the SLAMs with the ability to self-propel in hydrogen peroxide (H₂O₂) solution by ejecting oxygen (O₂) microbubbles from the diatom’s hollow channel. We hypothesize that SLAMs doped with a critical mass of manganese oxide (MnO₂) nanocatalysts would self-assemble into a swarm due to the interaction between microbubbles from diatoms. The resulting SLAM swarm would surround a microbubble and repeat microbubble generation and burst to continuously provide cavitation energy sufficient to fracture biofilms and, in turn, decrease bacterial resistance to antibiotics and disinfectants.

Scheme 1.

Fabrication of SLAM and its mechanism for biofilm removal. (A) The SLAMs are assembled by doping MnO₂ on diatom particles through polydopamine (PDA) coating (top: MnO₂-PDA-diatom) and grafting of (3-aminopropyl)triethoxysilane (APTES) (bottom: MnO₂-diatom). (B) The mechanism of biofilm removal by SLAM: (i) The SLAMs firstly penetrate the biofilm through self-propulsion. (ii) The SLAMs further self-assemble into a swarm that keeps generating a microbubble to fracture the EPS of biofilm. (iii) The microbubble formed by SLAM swarm bursts to generate cavitation energy sufficient to abrade biofilms. (iv) The SLAM swarm repeats microbubble generation and burst to fracture biofilms continuously. (v) The SLAM swarm completely removes biofilms from the bio-fouled structure in the end.

To examine this hypothesis, first, we improved our previously developed diatom microbubbler by increasing the loading mass of MnO₂ nanocatalysts on diatoms using polydopamine (PDA) binder. Second, we examined the extent that the loading mass of MnO₂ on diatoms regulates the microbubble generation, self-propulsion speed, kinetic reaction, and swarm formation. Then, we analyzed the frequency of repeated bubble generation and burst in the swarm and proposed a mechanism that controls the swarming behavior of SLAMs. Finally, we evaluated the efficacy in cleaning P. aeruginosa biofilms covering microgrooved silicone structure and wounded porcine skin explants, along with irrigating water or antibiotics.

2. Materials and methods

2.1. Fabrication of MnO₂-PDA-diatom and MnO₂-diatom particles

The MnO₂-PDA-diatom particles were fabricated via PDA coating. First, to remove impurity and debrides, 200 mg of diatom particles (DiatomaceousEarth-food grade, amorphous silica from freshwater type) were washed with deionized water and collected by centrifugation for 3 min at 1000 rpm. The washing process was repeated three times. The resultant particle slurry was mixed with 320 mL of deionized water containing 100 mg of dopamine hydrochloride (Sigma-Aldrich) under the stirring condition at 500 rpm. After 1 h, 80 mL of Tris-HCl solution (50 mM, pH = 8.5) was added into the mixture and stirred for 24 h at room temperature to polymerize dopamine. The diatom particles were collected by centrifugation at 1000 rpm for 3 min and washed with deionized water three times. The particles, denoted as PDA-diatom, were dried overnight in a lyophilizer (Labconco). Second, 50 mg of PDA-diatom particles were mixed with 5 mL of 50 mM KMnO₄ (Sigma-Aldrich) solution and stirred at 500 rpm for 2 h. Then, the particles were collected by centrifugation at 1000 rpm for 3 min and washed with deionized water three times. The sample, denoted as MnO₂-PDA-diatom, was dried in a lyophilizer overnight.

Separately, the MnO₂-diatom particles were fabricated through grafting with (3-aminopropyl)triethoxysilane (APTES, Sigma-Aldrich). First, 2 g of purified diatom particles was mixed with 60 mL of toluene in a three-necked flask coupled with a reflux condenser under an N₂ atmosphere. Then, 0.6 mL of deionized water was added to the mixture and stirred at room temperature for 2 h. To initiate the silane grafting, 3.4 mL of APTES was added to the mixture and stirred at 60 °C. After 6 h, the mixture was cooled down and then washed with toluene, 2-propanol, and deionized water three times. The obtained particles, denoted as NH₂-diatom, were dried in a lyophilizer overnight. Second, 0.1 g of NH₂-diatoms were mixed with 1 mL of 50 mM KMnO₄ solution and sonicated for 30 min. Then, the particles were collected by centrifugation at 1000 rpm for 3 min and washed with deionized water three times. The sample, denoted as MnO₂-diatom, was dried in a lyophilizer overnight.

2.2. Characterizations of MnO₂-PDA-diatom and MnO₂-diatom particles

The morphology and elemental analysis of MnO₂-doped diatom particles were conducted using a scanning electron microscope (Hitachi S-4800 SEM) at 10 kV and 20 kV, respectively. The transmission electron microscopy (TEM) images were obtained by JEOL 2100 Cryo TEM with an accelerating voltage of 200 kV. The content of the manganese element was quantified by inductively coupled plasma mass spectrometry (ICP-MS, NexION 350D). The surface of MnO₂-doped particles was analyzed by X-ray photoelectron spectroscopy (XPS, Kratos Axis ULTRA). The motion of particles in the aqueous H₂O₂ solution was observed with an optical microscope (Leica DMIL). The speed and motion trajectory of particles were quantified and tracked by using Fiji software.

2.3. H₂O₂ decomposition kinetic analysis of MnO₂-doped diatom particles

The H₂O₂ decomposition reactivity was evaluated using batch reactors. Phosphate-buffered saline solutions were balanced with 30% H₂O₂ to give a desired H₂O₂ concentration (1 mM - 1 M) and agitated at 750 rpm. Most measurements were performed at ambient temperature, but the reactors were also placed in ice or oil baths to change the temperature for activation energy measurements, monitored with in situ thermocouples. A varying amount of catalytic material was added to the reactor to give a constant Mn concentration (1 mM) across all measurements. After the addition of the catalyst, aliquots were withdrawn every 1 min for 5 min and filtered. These aliquots were diluted in H₂O to H₂O₂ concentrations below 0.2 mM and titrated with equal volumes of an H₂O₂ indicator solution (75 v% H₂O, 25 v% CH₃CH₂OH, 4.1 mM CUSO₄, 6.0 mM neocuproine). These titrated samples were aged 30 min to allow for saturation and analyzed with a UV–Vis spectrometer (Spectronic, Genesys 20). To measure H₂O₂ concentrations, the absorbance at 454 nm for each sample was compared to a calibration curve [43] obtained using standard solutions of known concentration.

2.4. Analysis of the MnO₂-PDA-diatom swarm

The swarming clusters formed by MnO₂-PDA-diatoms in 1.5–10 wt% H₂O₂ solutions were recorded using an optical microscope (Axio Observer D1, Carl Zeiss, Inc.) coupled with a high-speed camera (Phantom Miro eX4) at 1000 fps. The time-lapse images were extracted by Fiji software. The volume of the individual O₂ bubble was calculated by measuring the diameter of the bubble. The transient volumetric change was further obtained by taking the time derivative at each time point.

2.5. Preparation of biofilm on microgrooved poly(dimethylsiloxane) (PDMS) substrate

First, the PDMS prepolymer and curing agent (SYLGARD 184 Silicone Elastomer Kit, Dow Corning) were mixed at a mass ratio of 10:1 for 3 min. The mixture was then degassed in a vacuum desiccator. After 10 min, the mixture was slowly poured onto a silicon master with a microgrooved pattern (100 μm in width) and cured at 60 °C overnight in the oven. The cured PDMS with a microgrooved pattern was gently peeled off from the silicon master and treated with plasma (Harrick plasma, plasma cleaner PDC-32G) to expose silanol groups on its surface. The microgrooved PDMS substrate was further sterilized with 70% ethanol as well as UV exposure for 30 min. Secondly, P. aeruginosa (ATCC 15442) was streaked for isolation on a Trypticase^™ Soy Agar (TSA) plate (BD BBL^™) and cultured at 37 °C overnight. Then, 3 to 5 isolated colonies were inoculated in 200 mL of Tryptic Soy Broth (TSB) broth (BD Bacto^™) for 20 h at 37 °C on a shaker at 100 rpm. The concentration of P. aeruginosa was subsequently adjusted to an optical density (OD₆₀₀) of 0.2, which corresponded to CFU/mL of 10⁸. The bacterial cell suspension (200 μL) was transferred onto the microgrooved PDMS substrate and incubated at 37 °C. After 24 h, the microgrooved PDMS substrate was gently rinsed with sterile phosphate-buffered saline (PBS, Corning) to remove planktonic bacterial cells. Then, fresh bacterial cell suspension (200 μL, OD₆₀₀ = 0.2) was added onto the microgrooved PDMS substrate and incubated at 37 °C for another 24 h. This process was repeated until the third day of biofilm growth. After 3 days, the biofilms on the microgrooved PDMS substrate were rinsed with sterile PBS before further analysis.

2.6. Biofilm removal from the microgrooved PDMS substrate

The 3-day old P. aeruginosa biofilms on the microgrooved PDMS substrate were treated with 3 wt% H₂O₂ solution, MnO₂-diatom in 3 wt% H₂O₂ solution, and MnO₂-PDA-diatom in 3 wt% H₂O₂ solution. 500 μL of H₂O₂ solution with or without particles was added to a 6-well plate in each treatment. Then, PDMS substrate was placed in each well for treatment. After 10 min, the treated PDMS substrates were gently rinsed with PBS to remove the H₂O₂ solution and diatom particle residues for further analysis.

2.7. Evaluation of biofilm removal efficacy with the microgrooved PDMS substrate

Crystal violet staining was performed to evaluate the efficacy of biofilm removal quantitively. After each treatment, the PDMS substrates with biofilms were immersed in the 0.1 wt% crystal violet (Sigma-Aldrich) solution for 30 min. The substrates were then rinsed with PBS to remove excess crystal violet, followed by immersion in 95% ethanol to dissolve crystal violet within the biofilm. The efficacy of biofilm removal for each treatment was determined by measuring the absorbance intensity of crystal violet at a wavelength of 550 nm on a microplate reader (TECAN, Switzerland).

2.8. Immunostaining of EPS in the biofilm

First, the protein in the biofilms was labeled by incubating the PDMS substrate with biofilms in 500 μL of 0.1 M NaHCO₃ buffer (pH = 9.2) containing 5 mg of fluorescein isothiocyanate (Sigma-Aldrich) for 1h [44]. Second, the substrate with biofilms was rinsed with PBS. The substrate was incubated with 500 μL of 250 μg mL⁻¹ concanavalin A, tetramethylrhodamine conjugate (Invitrogen) for 2h to label α-glucopyranosyl and α-mannopyranosyl polysaccharide residues. Again, the substrate with biofilms was rinsed with PBS and incubated in 500 μL of 300 μg mL⁻¹ brightener 28 (MP Biomedicals, LLC) to label β-linked polysaccharides. Finally, the substrate with biofilms was mounted on a glass-bottom dish for observation using a confocal laser scanning microscope (Zeiss LSM 700, Germany).

2.9. Preparation of porcine skin explants with a puncture wound

The preparation of porcine skin explants was modified from previous literature [45]. First, fresh porcine skin purchased from a local merchandiser was frozen at −80 °C freezer overnight. The frozen porcine skin was subsequently excised into explants with a diameter of 12 mm by using a biopsy punch. Then, a hand-held drill (Dremel^® 7700–1) with a high-speed cutter (Dremel^® 191) was used to consistently create a central wound site with 2 mm diameter and 1 mm depth on the explant. Second, the porcine skin explants were immersed in PBS containing 5 μL L⁻¹ Tween-80 (Sigma-Aldrich) to remove any foreign debris. After 10 min, the explants were transferred to 70% ethanol (prepared with PBS solution containing 5 μL L⁻¹ Tween-80) for another 30 min. Finally, the explants were washed with PBS containing 5 μL L⁻¹ Tween-80 for 10 min, repeating 3 times. The whole sterilization process was conducted on a shaker.

2.10. Biofilm growth in a puncture wound of porcine skin explants

P. aeruginosa (ATCC 15442) was streaked for isolation on a Trypticase^™ Soy Agar (TSA) plate (BD BBL^™) and cultured at 37 °C overnight. Then, 3–5 isolated colonies were inoculated in 200 mL of Tryptic Soy Broth (TSB) (BD Bacto^™) for 20 h on a shaker at 100 rpm. The concentration of P. aeruginosa was subsequently adjusted to an optical density (OD₆₀₀) of 0.2, which corresponded to CFU mL⁻¹ of 10⁸. Next, sterile porcine skin explants were placed on soft 0.5% agar TSA with 50 μg mL⁻¹ gentamicin in a 24-well plate. Then, 5 μL of 10⁸ CFU mL⁻¹ P. aeruginosa was inoculated in the central wound site of each explant. The 24-well plate was covered with the lid and cultured in an incubator at 37 °C for 3 days.

2.11. Biofilm removal from a puncture wound of porcine skin explants

After 3-day culture, the bottom of a skin explant with biofilm was fixed on a sterile lid of a 24-well plate by super glue. At the same time, the rinse plates were prepared by adding 3 mL of PBS in each well of a sterile 24-well plate. The testing plate was prepared by adding 3 mL of PBS, 3 wt% H₂O₂ solution, or 3 wt% H₂O₂ solution with MnO₂-PDA-diatom particles. To begin the test, the lid with explants was quickly placed onto the first rinse plate for 10 s to remove planktonic bacterial cells. Then, the lid was transferred to a testing plate for the biofilm removal test. After 10 min of treatment, the lid was removed from the testing plate and put on the second rinse plate for another 10 s to remove any biofilm debris. The control group was treated only with the first rinse plate.

2.12. Evaluation of biofilm removal efficacy from a puncture wound of skin explants

The explants were transferred in a 15 mL centrifugation tube containing 7 mL of sterile PBS with 5 μL L⁻¹ Tween-80. Catalase (Sigma-Aldrich) was added to obtain a concentration of 2 mg mL⁻¹ solution to neutralize the H₂O₂ residue if necessary. Then, a sonication probe (Fisher Scientific Model 100) was applied in each tube at 20 W in an ice bath for 30 s to separate the bacteria into suspension. The probe was disinfected with 70% ethanol between sampling. The suspension was serially diluted by 10-fold, spot plated on TSA plates, and incubated at 37 °C overnight. Colonies in a countable range of 2–20 in each spot were recorded and converted as colony-forming units per milliliter (CFU mL⁻¹), standing for the viable bacterial concentration.

2.13. SEM imaging of the biofilm in a puncture wound of skin explants

First, the explants were fixed in 0.1 M Na-Cacodylate buffer (pH = 7.4) containing 2.0% paraformaldehyde and 2.5% glutaraldehyde (both E.M. grade) in the fridge. After 4 h, the explants were rinsed with 0.1 M Na-Cacodylate buffer for 10 min on a shaker and then dehydrated with a progressively higher concentration of ethanol, namely, 37% (10 min), 67% (10 min), 95% (10 min), and 100% (10 min for 3 times). Second, the fixative explants were dried using a critical point dryer in 100% ethanol. The morphology of explants was obtained by environmental SEM (FEI Quanta FEG 450 ESEM) at 2 kV.

2.14. Optical coherence tomography (OCT) imaging of a biofilm-infected wound

Each sample was non-destructively imaged by a custom-built spectral-domain optical coherence tomography (SD-OCT) system [46]. The OCT system utilized a superluminescent diode (S5FC1325S-SP, Thorlabs, New Jersey) with a center wavelength of 1325 nm and a bandwidth of 50 nm as a light source. A spectrometer with a 1024-pixel InGaAs line-scan camera (SU-LDH2, Goodrich, North Carolina) was used as a detector that acquired spectra at a line scan rate of approximately 92 kHz. The OCT system had an axial and transverse resolution of approximately 8 μm and 16 μm, respectively. Each cross-sectional OCT image represented 3.13 mm (depth) and 4.18 mm (width). The OCT images were acquired using LabVIEW and processed using MATLAB.

2.15. Post-treatments of biofilm

After treatments with MnO₂-PDA-diatoms or controls, the biofilm residues were further treated with either irrigation or antibiotics. For irrigation, the puncture wound site was rinsed with 5 mL of PBS using a 20 mL syringe coupled with an 18-gauge needle. The explants were immediately probe-sonicated to separate the biofilm debris and spot-plated to evaluate the biofilm removal efficacy. For antibiotics treatment, each porcine skin explant was immersed in 2 mL media containing 200 μg mL⁻¹ gentamicin and incubated at 37 °C. After 24 h, the porcine skin explants were rinsed with PBS. The number of viable microbial cells was quantified with the spot-plating method mentioned above.

2.16. Statistical analysis

Three samples were analyzed per condition, and the data were presented as mean ± standard deviation unless otherwise specified. One-way ANOVA followed by Turkey’s post hoc analyses was performed for comparisons between groups to determine significance. Data were considered statistically significant with p values less than 0.05.

3. Results and discussion

3.1. Synthesis and characterizations of MnO₂ sheet-doped diatoms

The SLAM was assembled by using binding agents to incorporate MnO₂ nanosheets on the surfaces of diatom biosilica particles (Scheme 1A). The hollow cylinder-shaped diatom particles with 10 μm-diameter, 18 μm-length, and 500 nm-diameter pores (Fig. 1A) were modified with the PDA binder. The PDA layer was formed through self-polymerization of dopamine activated by a weak alkaline tris buffer solution (pH = 8.5) [47,48]. The dopamine concentration was optimized at 0.25 mg mL⁻¹ to form a uniform PDA layer on the diatom surface (Fig. 1B; Fig. S1). As shown in the transmission electron microscopy (TEM) image (Fig. 1C), the PDA layer thickness is around 50 nm compared to the TEM image of bare diatom surface (Fig. S2). Particle color was also changed from white to black because of the PDA layer (Fig. S3A, B).

Fig. 1.

Characterizations of MnO₂-PDA-diatom. (A) SEM image of a pristine diatom. Scale bar: 5 μm. (B) SEM image of PDA-diatom. Scale bar: 5 μm. (C) TEM image of PDA-diatom. Scale bar: 100 nm. (D) TEM images of MnO₂-PDA-diatom. Arrows indicate the MnO₂ deposited on the PDA layer. Inset image of (D) represents the higher magnification of MnO₂-PDA assembled around the side pore wall. Scale bar: 100 nm. (E, F) SEM and elemental mapping images of (E) MnO₂-PDA-diatom and (F) MnO₂-diatom. Scale bar: 5 μm. (G–J) XPS spectra of pristine diatoms, PDA-diatom, and MnO₂-PDA-diatom particles. (G) XPS survey spectrum of pristine diatom, PDA-diatom, and MnO₂-PDA-diatom particles. (H) Deconvolution of high-resolution XPS spectrum of PDA-diatom for N 1s. (I, J) High-resolution XPS spectrum of MnO₂-PDA-diatom particles for (I) Mn 2p and (J) Mn 3s.

The PDA-diatom particles were further loaded with MnO₂ sheets by reducing permanganate ions $({\text{MnO}}_4^{-})$ as follows [49]:

$$\begin{gathered} {\text{MnO}}_4^{-}+2{\mathrm{H}}_2\mathrm{O}+3{\mathrm{e}}^{-}(\text{electron}\text{donors:}\text{PDA})\to {\text{MnO}}_2 \\ +4{\text{OH}}^{-} \end{gathered}$$ (1)

The PDA binder on diatoms served as an agent that enhances reduction of ${\text{MnO}}_4^{-}$ into MnO₂ in situ and forms MnO₂ sheets on the diatom surface, as confirmed with the TEM image (Fig. 1D). The resulting MnO₂ changed the powder color from black to brown, indicating the MnO₂ sheets formation (Fig. S3C). The elemental mapping shows the uniform coverage of the PDA-diatoms with the MnO₂ sheets (Fig. 1E). The elemental analysis by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) further shows that the loading amount of Mn element is 18.2 wt%.

As a point of comparison, MnO₂-diatoms were prepared by modifying diatom surface to present primary amine groups using (3-aminopropyl)triethoxysilane binder (Scheme 1A) [28]. Then, ${\text{MnO}}_4^{-}$ was reduced to form the MnO₂ sheets on the diatom surface. The elemental mapping images confirmed Mn immobilized on the diatoms, but the Mn coverage was lower than MnO₂-PDA-diatoms (Fig 1F). According to ICP-AES measurement, the loading amount of Mn element was 2.5 wt% on the MnO₂-diatoms, which was 7.3-fold lower than the MnO₂-PDA--diatom. This difference is because PDA offers abundant catechol groups to trigger the reduction of ionic ${\text{MnO}}_4^{-}$ into solid MnO₂.

We also examined the chemical composition of the pristine diatom, PDA-diatom, and MnO₂-PDA-diatom surface using X-ray photoelectron spectroscopy (XPS) (Fig. 1G-J). The peaks of Si 2s (153.0 eV) and Si 2p (100.2 eV) from the pristine diatom were consistent with unmodified biosilica SiO₂ [50]. The intensify of these two peaks sharply decreased after modification, indicating the surface coverage by the PDA layer and MnO₂ sheets (Fig. 1G). We further confirmed the PDA layer on diatoms using the high-resolution N 1s XPS (Fig. 1H). The N 1s peak at 400.0 eV in PDA-diatom was fitted into three peaks at 398.6 eV, 399.9 eV, and 401.9 eV. Each of the peaks represented tertiary/aromatic (R–N = , 11.50%), secondary (R₂-N-H, 74.21%), and primary (R-NH₂, 14.29%) amine groups of PDA, respectively [51]. This N 1s signal disappeared after forming MnO₂ sheets on the PDA-diatom (Fig. S4). The two peaks in Mn 2p spectrum at 642.1 eV (Mn 2p_3/2) and 653.8 eV (Mn 2p_1/2) with spin-energy separation of 11.7 eV (Fig 1I) and the two peaks in Mn 3s spectrum with the spin-energy separation of 4.7 eV (Fig. 1J) indicate the existence of Mn⁴⁺ on the PDA-diatoms [49,52,53].

3.2. Effects of MnO₂ mass on the self-propulsion speed of MnO₂-doped diatoms

We examined the extent that MnO₂ sheets regulate the propulsion speed of diatoms. First, MnO₂-PDA-diatoms and MnO₂-diatoms were fixed on a separate glass slide immersed in the 5 wt% H₂O₂ solution to measure the O₂ bubble generation rate. Initially, a tiny O₂ bubble embryo emerged from the hollow channel of both MnO₂-PDA-diatom and MnO₂-diatom (0 ms in Fig. 2A). The O₂ bubble embryo grew over 260 ms with the MnO₂-PDA-diatom and 410 ms with the MnO₂-diatom. The O₂ bubble growth rate of the MnO₂-PDA-diatom was 1.6-fold higher than the MnO₂-diatom. As the bubble diameter reached ~12 μm, the bubble was separated from the diatom. The bubble generation and ejection continued until the H₂O₂ was consumed. As a control, the unmodified diatoms did not generate a bubble (Fig. S5).

Fig. 2.

Motion and kinetic analysis of MnO₂-PDA-diatom and MnO₂-diatom in H₂O₂ solutions. (A) Time-lapse images of microbubble generation from a fixed MnO₂-PDA-diatom and MnO₂-diatom in 5 wt% H₂O₂ solution. (B) Schematic illustration for the mechanism of bubble generation from a hollow channel of the MnO₂-doped diatom. (C) Time-lapse images of trajectories of the MnO₂-PDA-diatom and the MnO₂-diatom in 5 wt% H₂O₂ solution. (D) Self-propulsion speed of MnO₂-PDA-diatom and MnO₂-diatom particles in H₂O₂ solutions with varied H₂O₂ concentrations. (E) The dependency of the initial H₂O₂ decomposition rate on the H₂O₂ concentration for MnO₂-PDA-diatom and MnO₂-diatom. Data points represent the mean, and error bars indicate standard deviation, (n = 10 in (D) and n = 3 in (E)).

The movement of diatoms is consistent with the vast majority of O₂ bubbles forming within the hollow diatoms rather than on the diatom exterior. Previous studies [54,55] indicate that bubble nucleation energy barriers are much lower on concave than on convex surfaces, considering the cylindrical diatom morphology, agrees with the previous observation. Though MnO₂ may be distributed throughout the diatom interior and exterior, the O₂ bubble preferentially form in the interior concave channels (Fig. 2B). The 7-fold larger loading of MnO₂ within MnO₂-PDA-diatoms than upon MnO₂-diatoms leads to a proportional increase in the rate of O₂ generation, bubble growth, and ejection frequency.

The bubbles ejected from the diatom’s hollow channel generate thrust to make the particles self-propel in the H₂O₂ solution. As shown in Fig. 2C, both MnO₂-PDA-diatoms and MnO₂-diatoms self-propelled by O₂ bubbles released continuously from the channel of diatom (Movie S1-S4). The propulsion speed of the MnO₂-PDA-diatom was highly dependent on the H₂O₂ concentration, increasing from 106 ± 16 μm s⁻¹ in 2 wt% H₂O₂ solution to 270 ± 41 μm s⁻¹ in 10 wt% H₂O₂ solution (Fig. 2D). In contrast, the propulsion speed of MnO₂-diatom was less dependent on the H₂O₂ concentration, ranging from 27 ± 8 μm s⁻¹ in 2 wt% H₂O₂ solution to 98 ± 29 μm s⁻¹ in 10 wt% H₂O₂ solution.

Supplementary video related to this article can be found at https://doi.org/10.1016/j.biomaterials.2022.121610

The difference in the self-propulsion speed between MnO₂-PDA-diatom and MnO₂-diatom was related to the decomposition rate of H₂O₂ to O₂. As shown in Fig. 2E, the initial H₂O₂ decomposition rate was linearly dependent on the H₂O₂ concentration for the MnO₂-PDA-diatoms (rate ~ [H₂O₂]^0.9±0.1). In contrast, MnO₂-diatoms led to a sublinear dependency of initial H₂O₂ decomposition rate on the H₂O₂ concentration (rate ~ [H₂O₂]^0.7±0.1). These observations are consistent with the kinetics of H₂O₂ dependence on transition metal-modified silicates and suggest a bimolecular reaction between fluid-phase H₂O₂ and H₂O₂-derived Mn-complex [56-58]. On average, the H₂O₂ decomposition rate based on per gram of the MnO₂-PDA-diatoms was 11-fold higher than the MnO₂-diatoms.

We also calculated the apparent activation energy (E_a) for H₂O₂ decomposition by fitting curves of the H₂O₂ decomposition rate versus temperature to the Arrhenius Equation (Fig. S6). The E_a obtained with MnO₂-PDA-diatom was 26 ± 3 kJ mol⁻¹, while that with MnO₂-diatoms was 29 ± 3 kJ mol⁻¹. This similar E_a, together with a nearly first-order dependence upon H₂O₂ concentration suggests that both MnO₂-PDA-diatoms and MnO₂-diatoms decompose H₂O₂ by similar mechanisms. Overall, the number of H₂O₂ decomposition sites changes from MnO₂-PDA-diatom to MnO₂-diatom but not the inherent reactivity of those sites.

3.3. Self-assembly of MnO₂-PDA-diatom swarm

Next, we examined the interaction between SLAMs and their capability of forming a swarm through the microbubble interaction. To evaluate whether the concentration of SLAMs would play a role on the swarm formation, the concentration of SLAMs in 5 wt% H₂O₂ was increased from 0.032 mg mL⁻¹ to 1 mg mL⁻¹. As shown in Fig. S7, MnO₂-PDA-diatoms self-propelled individually at low particle concentrations. However, they started self-assembling to form swarms at 0.25 mg mL⁻¹ and higher concentrations. In particular, diatoms interacted more frequently at the higher particle concentration. As shown in Movie S5, the MnO₂-PDA-diatoms swarm created an O₂ bubble with 100 μm in diameter at the center. The O₂ bubble burst rapidly to let the swarm continue to create a new O₂ bubble. This repeated bubble generation and burst made the diatoms keep the swarm pattern without falling apart. Likewise, MnO₂-diatoms also formed swarms at 0.25 mg mL⁻¹ and higher particle concentration (Fig. S7). However, the O₂ bubble at the center of the swarm expanded more slowly and aggregated with neighboring bubbles instead of burst (Movie S6).

Supplementary data related to this article can be found at https://doi.org/10.1016/j.biomaterials.2022.121610.

To study the reason causing this difference, we firstly investigated the interaction between MnO₂-PDA-diatoms and the subsequent O₂ bubble generation at 0.25 mg mL⁻¹ with various concentrations of H₂O₂ using a high-speed camera (Fig. 3A; Movie S7). As shown in Fig. 3A-i and Movie S7, MnO₂-PDA-diatoms added to 10 wt% H₂O₂ solution instantaneously initiated self-propulsion. At the same time, 20 μm diameter O₂ bubbles hanging on the MnO₂-PDA-diatoms collided with each other and fused to form a single bubble within 30 ms. As a consequence, MnO₂- PDA-diatoms formed a swarm that surrounded the bubble. The bubble continued to grow with additional O₂ produced from the swarmed diatoms over the next 28 ms. The bubble finally burst at 59 ms. Even after the bubble burst, the diatoms kept the swarmed form, likely due to a local pressure depression. Then, the diatoms repeated the fresh O₂ bubble growth and burst (Movie S7). Such rapid oscillatory bubble formation and rupture caused by the MnO₂-PDA-diatom swarm was observed when H₂O₂ concentration was higher than 2 wt% (Fig. 3A-i to Fig. 3A-v). However, decreasing the H₂O₂ concentration from 10 to 2 wt% decreased the frequency of O₂ bubble growth and burst in a swarm from 16.9 to 4.2 bubbles per second. When the H₂O₂ concentration was decreased to 1.5 wt%, MnO₂-PDA-diatoms generated O₂ bubbles very slowly. Accordingly, the O₂ bubbles fused slowly into a larger bubble without bursting over 4688 ms, which acts like quasi-static bubble growth (Movie S7).

Fig. 3.

Analysis of swarming by MnO₂-PDA-diatoms. (A) Time-lapse images of a cycle of O₂ bubble growth and rupture manipulated by the MnO₂-PDA-diatoms in aqueous media with varied H₂O₂ concentrations: (A-i) 10 wt%, (A-ii) 7 wt%, (A-iii) 5 wt%, (A-iv) 3 wt%, (A-v) 2 wt%, (A-vi) 1.5 wt%. The images were captured using a high-speed camera (1000 fps). The blue arrows in images captured at 0 ms indicate individual MnO₂-PDA-diatom that moves and forms a swarm. Scale bar: 100 μm. (B) The volume of O₂ bubble (V_b) increased over time in H₂O₂ solutions. The inset plot represents the curves in the dash line box. The V_b was calculated by assuming a spherical shape $(V_b=4∕3\pi r_b^3)$. (C) The transient change of V_b with time (dV_b/dt) in H₂O₂ solutions. The inset plot represents the curves in the dash line box. (D) Schematic illustration of force balance on a growing O₂ bubble at the air-liquid interface. (E) Schematic illustration of bubble fates altered by the O₂ generation rate of MnO₂-PDA-diatoms and MnO₂-diatoms.

Supplementary data related to this article can be found at https://doi.org/10.1016/j.biomaterials.2022.121610.

We further quantified the O₂ bubble volume (V_b) increased over time before burst (Fig. 3B). In the beginning, there was a minimal increase of V_b regardless of H₂O₂ concentration because O₂ bubbles were at the stage of collision. Then, V_b increased sharply over time as diatoms formed a swarm around the O₂ bubble and supplied O₂ gas in a collective manner. Given that O₂ mass generated by MnO₂-PDA-diatoms determines V_b, we calculated the transient O₂ generation rate with time derivative (dV_b/dt) as shown in Fig. 3B. In 2 to 10 wt% H₂O₂ solutions, dV_b/dt increased over time, indicating the accelerated bubble growth (Fig. 3C). The acceleration of dV_b/dt became larger with increasing H₂O₂ concentration, while dV_b/dt remained constant over time in 1.5 wt% H₂O₂ solution. Moreover, the dV_b/dt at the time when the bubble burst was almost linearly dependent on the H₂O₂ concentration from 2 to 10 wt% (Fig. S8).

Based on the results above, we suggest that the accelerated bubble growth over time leads to the bubble burst. In general, as shown in Fig. 3D, there are three forces acting on a growing bubble at the air/water interface: buoyancy force F_B, capillary force F_σ, and contact pressure force F_CP [59,60]. Before the bubble ruptures, the sum of these forces should remain zero (Fig. 4D):

$${\stackrel{\rightharpoonup }{F}}_B+{\stackrel{\rightharpoonup }{F}}_{\sigma }+{\stackrel{\rightharpoonup }{F}}_{CP}=0$$ (2)

Fig. 4.

Evaluation of P. aeruginosa biofilm removal from a microgrooved PDMS substrate. (A) Schematic illustration of conditions used to remove biofilm: untreated, 3 wt% H₂O₂ solution, 3 wt% H₂O₂ solution with MnO₂-diatom, and 3 wt% H₂O₂ solution with MnO₂-PDA-diatom. (B) Optical images of PDMS substrate with biofilms for each treatment: (i) untreated biofilms, (ii) 3 wt% H₂O₂ solution, (iii) 3 wt% H₂O₂ solution with MnO₂-diatom, and (iv) 3 wt% H₂O₂ solution with MnO₂-PDA-diatom. The images in the first column represent the treatment at an intermediate stage (i.e., 2 min after treatment starts). The images in the second column represent the PDMS substrate with biofilm after treatment for 10 min. The images in the third column represent the biofilm stained with crystal violet (CV) after treatment for 10 min. Scale bar: 200 μm. (C) Intensity of biomass remained on the PDMS substrate after each treatment for 10 min. Bars represent the average value, and error bars indicate standard deviation. * represents significant difference between the two groups, *p < 0.05 (n = 5).

Equation (2) can be further expanded by the definition of each term:

$$2\pi r_0\sigma \sin \theta ={\rho }_{l}gV+\pi r_0^2(P_{O_2}-P_{atm})$$ (3)

where r₀ is the radius of bubble cross-section at the air-liquid interface, σ is the surface tension, θ is the instantaneous contact angle, ρ_l is the density of the liquid, g is the acceleration due to gravity, V is the bubble volume below the air-liquid interface, P_O2 is the O₂ bubble pressure, and P_atm is the atmospheric pressure. According to Equation (3), the surface tension holds the dynamic stability of a bubble while P_O2 steadily keeps increasing due to the increase of the number of O₂ molecules.

The accelerated rise of P_O2 may disrupt the dynamic stability and, finally, cause the bubble to burst. The P_O2 inside the bubble is estimated by applying kinetic measurement of MnO₂-PDA-diatom particles as follows (see Supplementary data for a detailed derivation):

$$P_{O_2}-P_{atm}=0.2425C_{diatom}{C}_{0,H_2{O}_2}RTt$$ (4)

where C_diatom is the concentration of MnO₂-PDA-diatom particles and C_0,H2O2, is the initial concentration of H₂O₂. As shown in Fig. S9, the pressure difference between bubble and atmosphere increases faster at higher H₂O₂ concentrations. However, the pressure difference between bubble and atmosphere at the moment of bubble burst converges at 350–550 Pa, independent of H₂O₂ concentrations. Therefore, O₂ bubbles burst more frequently in the higher H₂O₂ concentration. After the first O₂ bubble burst, MnO₂-PDA-diatom swarms continue to decompose H₂O₂ to O₂, thus repeating the bubble growth and burst with the same frequency. The significantly reduced O₂ bubble growth and burst frequency in the 1.5 wt% H₂O₂ solution is attributed to the delayed P_O2 increase and constant dV_b/dt. Separately, MnO₂-diatoms added into the H₂O₂ solution exhibited quasi-static bubble growth (Fig. S10) due to the 11-fold lower O₂ generation rate than the MnO₂-PDA-diatoms. The O₂ bubbles generated from MnO₂-diatoms either burst with a significantly lower frequency than MnO₂-PDA-diatoms or remained stable without burst (Fig. S10 and Fig. 3E). The quasi-static bubble growth with MnO₂-diatoms was observed at any H₂O₂ concentrations (Fig. S11).

3.4. Activities of MnO₂-doped diatoms to remove biofilm in confined space

Pseudomonas aeruginosa, as one of the prevalent pathogens, has been shown to have the ability to form robust biofilms due to its secretion of abundant exopolysaccharides [39-41]. According to our measurement, P. aeruginosa biofilm has a storage modulus more than two times greater than that of E. coli biofilm (Fig. S12). Therefore, we examined whether the swarming behavior of SLAMs could further bring a synergistic efficacy on the removal of a tougher P. aeruginosa biofilm in a microgrooved polydimethylsiloxane (PDMS) substrate (Fig. 4A). As shown in Fig. 4B-ii, biofilm exposed to the 3 wt% H₂O₂ solution generated O₂ bubbles on the surface because endogenous catalase within the biofilm decomposed H₂O₂ (Movie S8) [61]. However, most of the biofilm remained in the grooved substrate after exposure for 10 min. The biofilm residue was also stained with crystal violet [62], which binds with EPS and bacterial cells. Only about 20.5% of biofilm was removed from the PDMS substrate (Fig. 4C). In contrast, both MnO₂-diatoms and MnO₂-PDA-diatoms suspended in the 3 wt% H₂O₂ solution entered biofilm and generated microbubbles (Fig. 4B-iii, 4B-iv; Movie S9, S10). After 10 min, MnO₂-diatoms and MnO₂-PDA-diatoms removed ~66.7% and ~95.6% of biofilms, respectively (Fig. 4C). The biofilm removal efficacy was also confirmed by analyzing the residual EPS with immunostaining [44]. As shown in Fig. S13, most EPS consisting of α-glucopyranosyl/α-mannopyranosyl sugar residues (red), β-linked polysaccharides (blue), and extracellular proteins (green) remained in microgrooves after the treatment with the 3 wt% H₂O₂ solution. The mixture of H₂O₂ and MnO₂-diatoms removed a significant fraction of EPS. More strikingly, the mixture of H₂O₂ and MnO₂-PDA-diatom removed the EPS almost perfectly, as evidenced with minimal fluorescence signal.

Supplementary data related to this article can be found at https://doi.org/10.1016/j.biomaterials.2022.121610.

To understand the mechanism by which MnO₂-PDA-diatoms removed the biofilm more effectively than MnO₂-diatoms, we examined self-propulsion and O₂ generation of diatoms within the 3D biofilm microscopically. As shown in Fig. 5A and Movie S11, MnO₂-diatoms mixed with 3 wt% H₂O₂ solution penetrated the biofilm first. The O₂ bubbles generated by MnO₂-diatoms collided (0 s) and fused into a larger bubble (1.94 s). Then, the O₂ bubbles continued to expand and displaced the biofilm gradually (4.62 s). The O₂ bubbles remained stable for more than 10 s without further deforming or detaching the biofilm (Movie S11). In contrast, a larger number of MnO₂-PDA-diatoms penetrated biofilm than MnO₂-diatoms because of the higher self-propulsion speed. They continued to generate O₂ bubbles within the biofilm (Fig. 5B). Those O₂ bubbles collided (0 s) and fused to form larger bubbles surrounded by the diatom swarm (1.93 s). The resulting swarm drove O₂ bubbles to expand more rapidly than MnO₂-diatoms (4.53 s). The O₂ bubbles finally burst to create a cavity that drove dislocation and detachment of biofilm from the microgrooves (5.38 s). After the burst, the MnO₂-PDA-diatoms kept the swarmed form and repeated a cycle of bubble generation, fusion, and rupture (Movie S12).

Fig. 5.

Time-lapse images of P. aeruginosa biofilm removal process by O₂ bubbles from MnO₂-doped diatoms with a corresponding schematic illustration. (A) Treatment with MnO₂-diatoms mixed with 3 wt% H₂O₂ solution. The yellow arrows and dash line indicate the location of biofilm. The red-dash circles represent the O₂ bubbles generated by MnO₂-diatoms. (B) Treatment with MnO₂-PDA-diatoms mixed with 3 wt% H₂O₂ solution. The yellow arrow and dash line indicate the biofilm in the microgroove. The blue-dash circles represent the O₂ bubbles generated by the MnO₂-PDA-diatom swarm. Scale: 100 μm.

Supplementary data related to this article can be found at https://doi.org/10.1016/j.biomaterials.2022.121610.

We suggest the more frequently repeated growth and burst of O₂ bubbles by MnO₂-PDA-diatoms exert powerful mechanical perturbation within the 3D biofilm by deforming and cavitating EPS of the biofilm. On the other hand, MnO₂-diatoms damage biofilm less effectively than MnO₂-PDA-diatoms as the O₂ bubbles from MnO₂-diatoms expand more slowly while causing minimal cavitation energy.

3.5. The ability of MnO₂-PDA-diatoms to treat biofilm-infected wounds in skin explants

Topical treatment with 3% H₂O₂ solution has been extensively used for wound cleaning [63]. However, recent clinical studies suggest that H₂O₂ may not effectively reduce the bacterial burdens in wounds [63, 64]. To this end, we further evaluated whether incorporating MnO₂-PDA-diatoms into H₂O₂ could provide synergistic effects to clean biofilm-infected wounds. Before the test, we firstly evaluated the skin response to topical treatment with 3 wt% H₂O₂ and MnO₂-PDA-diatoms/3 wt% H₂O₂. The in vivo results indicate that neither 3 wt% H₂O₂ solution nor MnO₂-PDA-diatoms/3 wt% H₂O₂ mixture caused any irritation issue on the back skin (Fig. S14). In addition, the histology analysis also suggests both treatments did not result in any damage on the epidermal and dermal tissues on the skin (Fig. S15). Then, we created the biofilm in a punctured porcine skin explant by inoculating the wound site with P. aeruginosa for 3 days (Fig. 6A) [45,65], followed by the treatments with 3 wt% H₂O₂ and MnO₂-PDA-diatoms/3 wt% H₂O₂ mixture (Fig. S16). Optical coherence tomography (OCT) was used to evaluate the biofilm removal efficacy (Fig. 6B), allowing for non-invasive and label-free 3D tissue imaging [66-69]. The OCT image of a biofilm-free punctured skin wound shows a hollow hole with a diameter of 2 mm and a depth of 1 mm (Fig. 6B-i). Three days after the punctured wound site was inoculated with P. aeruginosa, the wound site had become filled with the growth of a 0.5 mm-thick biofilm (Fig. 6B-ii). Adding 3 wt% H₂O₂ solution to the biofilm-infected wounds generated O₂ bubbles by endogenous catalase. This treatment damaged the superficial layer of biofilm only (Fig. 6B-iii). In contrast, the MnO₂-PDA-diatoms/3 wt% H₂O₂ mixture created many deeper and larger craters, marked with negative contrast, in the biofilm (Fig. 6B-iv).

Fig. 6.

Removal of P. aeruginosa biofilm infecting the puncture wounds created in the porcine skin explant. (A) Image of 2 mm-diameter puncture wound in porcine skin explants. The arrows indicate the location of the puncture wound site on the porcine skin explant. Scale bar: 1 cm. (B, C) Optical coherence tomography (OCT) and scanning electron microscope (SEM) images of infected wound sites in porcine skin explants: (i) pristine porcine skin without biofilm, (ii) porcine skin infected by biofilm, (iii) biofilm-infected wound after treatment with 3 wt% H₂O₂ solution for 10 min, and (iv) biofilm-infected wound after treatment with MnO₂-PDA-diatom and 3 wt% H₂O₂ solution for 10 min. The asterisks in (B) indicate an area with mirror artifacts due to the limited imaging depth in SD-OCT. Scale bar: 1 mm in (B) and 10 μm in (C). (D) SEM images of MnO₂-PDA-diatom particles that penetrate biofilm. Scale bar: 10 μm. (E) Microbial cell viability level after the treatments and post-treatment with water irrigation and gentamicin. The number of viable cells was quantified by colony-forming unit (CFU) plate counting. Bars represent the average value, and error bars indicate standard deviation. * represents significant difference between the two groups, *p < 0.05 (n = 3).

In parallel, we imaged the biofilm that remained in the wounds using scanning electron microscopy (SEM) (Fig. 6C). Fig. 6C-i displays collagen fibers at the bottom part of the punctured, biofilm-free wound in the porcine skin. After 3-day culture of P. aeruginosa, the collagen fibers of skin tissue were covered with a thick layer of biofilms and planktonic bacterial cells (Fig. 6C-ii), as also shown in the OCT image. The wound treated with 3 wt% H₂O₂ solution was covered with rod-shaped bacterial cells associated with EPS (Fig 6C-iii). In contrast, biofilm-infected wounds treated with MnO₂-PDA-diatoms/3 wt% H₂O₂ mixture for 10 min revealed collagen fibers that resemble the pristine skin wound, indicating that the treatment could remove biofilm with minimal perturbation on the surrounding tissues (Fig. 6C-iv). We suggest that this successful biofilm removal results from penetration of MnO₂-PDA-diatoms into the biofilm as shown in Fig. 6D and subsequent O₂ bubble growth and rupture by diatom swarms within the biofilm. We also quantified the number of viable cells that remained in wounds using the spot plate count method [70,71]. The treatment with 3 wt% H₂O₂ solution reduced viable P. aeruginosa cells by 72.2% (Fig. 6E). In contrast, the MnO₂-PDA-diatoms/3 wt% H₂O₂ mixture decreased the number of viable cells more significantly by 99.2%.

We further examined whether the biofilm treated with MnO₂-PDA-diatoms/3 wt% H₂O₂ mixture became more susceptible to the conventional processes used to remove the biofilm in the wound. These processes include water irrigation and treatment with the antibiotic gentamicin. Irrigating the biofilm-infected wound reduced viable P. aeruginosa cells by 81.4% (Fig. 6E). Infected wound treated with 3 wt% H₂O₂ solution for 10 min and subsequent irrigation made additional 40% reduction, resulting in an 83.3% total reduction (Fig. 6E). These results combined with OCT and SEM images (Fig. 6B-iii, 6C-iii) indicated both irrigation only and irrigation following 3 wt% H₂O₂ solution removed the planktonic cells on the superficial layer of biofilm. The EPS residues held bacterial cells together, thus becoming a mechanical barrier against irrigating water. In contrast, irrigation following the treatment with MnO₂-PDA-diatoms/3 wt% H₂O₂ mixture made an additional 73.3% reduction of viable cells, leading to a 99.8% total reduction. This result confirms that biofilm damaged by MnO₂-PDA-diatoms becomes vulnerable to the external force exerted by irrigation (Fig. 6E).

Separately, biofilm in wounds was treated with the antibiotic gentamicin. Wounds treated with gentamicin reduced viable bacterial cells by 77.8% (Fig. 6E). In addition, gentamicin administered after treatment with 3 wt% H₂O₂ solution made 22.9% additional reduction, resulting in a 78.6% total reduction. Similar to irrigation, either gentamicin only or gentamicin following 3 wt% H₂O₂ solution did not remove EPS, hence limiting the transport of both gentamicin and H₂O₂ through the biofilm. However, gentamicin administered after treatment with MnO₂-PDA-diatoms/3 wt% H₂O₂ mixture resulted in additional 85.1% reduction of viable cells, leading to 99.9% total reduction. This result also indicates that MnO₂-PDA-diatoms made biofilm residues permeable to gentamicin, thus increasing the antibiotic efficacy of gentamicin.

4. Conclusion

This study presents an active antimicrobial particle (SLAM) that can remove biofilm in abiotic and biotic structures (microgrooved substrates and puncture wounds) and subsequently decrease the number of viable bacterial cells. The SLAM was assembled by doping controlled amounts of MnO₂ nanosheets to porous diatoms using different binders. In particular, increasing Mn mass percentage in the diatom to 18.2 wt% using the polydopamine binder made the resulting MnO₂-PDA-diatoms penetrate the P. aeruginosa biofilm by generating O₂ bubbles in H₂O₂ solutions and forming a swarm. The resulting diatom swarms were active to repeat O₂ bubble generation and rupture, which created a wave of cavitating energy sufficient to fracture EPS and remove biofilm from the original fouled site. Furthermore, MnO₂-PDA-diatoms removed almost 99.9% of viable cells when combined with irrigation or antibiotics. In contrast, decreasing MnO₂ mass percentage in diatoms to 2.5 wt% through the APTES grafting method reduced biofilm removal efficacy because the slower O₂ bubble growth lowered the frequency of bubble generation and rupture to further remove biofilm residues. Such powerful cleaning activities would be further tuned with other catalysts, H₂O₂ concentrations, and temperature, thus enabling the SLAM broadly applicable to clean various biofilm-fouled tissue and infrastructure.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.