Cold atmospheric plasma degrades methylene blue and shifts bacterial inactivation during photodynamic therapy

Abstract

Photodynamic therapy (PDT) and cold atmospheric plasma can inactivate bacteria via oxidative stress. Here we examined their combined action against Staphylococcus aureus (S. aureus) using methylene blue (MB)-mediated PDT with a red organic light-emitting diode (OLED) and surface dielectric barrier discharge (SDBD). Under individual treatments, S. aureus inactivation by both PDT and SDBD varied with MB concentration and exposure time. In the presence of MB, plasma rapidly degraded MB, and the effect of microbial inactivation was reduced compared with MB-free conditions. When PDT and SDBD were applied simultaneously, bacterial reduction increased from approximately 1.4 log at 10 min to 4.8 log at 20 min, while MB decreased markedly. This suggests that early inactivation was dominated by MB-mediated photodynamic action, whereas later inactivation was sustained by secondary reactive species produced from SDBD after MB depletion. Photolysis modeling suggested that OLED emission did not substantially alter plasma-generated reactive species, supporting a mechanism driven by time-dependent changes in liquid chemistry rather than gas-phase interference. Acute NIH/3T3 screening showed reduced metabolic activity, indicating the need for cytocompatible dose optimization. These results provide a mechanistic basis for time-dependent pathway shifts during combined MB-PDT and SDBD.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-43354-6.

Introduction

Antimicrobial photodynamic therapy (aPDT) is a light-based antimicrobial technique in which photosensitizers absorb visible light and transfer energy or electrons to molecular oxygen, generating reactive oxygen species that damage microbial membranes, proteins, and nucleic acids¹. Phenothiazinium compounds, including methylene blue (MB), have been extensively researched for their efficacy in managing both planktonic bacteria and biofilms. These agents are particularly relevant in fields such as dentistry, dermatology, and infectious disease treatment^2,3. MB-based PDT has also been applied in infection-related indications (e.g., infected wounds or chronic ulcers), with clinical reports supporting feasibility and generally acceptable tolerability^4,5.

For practical implementation, aPDT requires careful control of photosensitizer dose and exposure conditions to balance antimicrobial efficacy with host-tissue tolerance and to minimize residual photosensitizer levels after treatment. Because treatment outcomes can vary substantially with parameters such as photosensitizer concentration, incubation time, and light dose, efforts toward dose rationalization and standardization are increasingly emphasized in aPDT studies⁶. In this context, combination approaches have been explored as a pragmatic strategy to improve antimicrobial performance while potentially reducing the dose burden required from any single modality. For example, combining aPDT with conventional antibiotics has been reported to reduce minimum inhibitory concentrations and improve biofilm disruption compared with either treatment alone⁷. Cold atmospheric plasma (CAP) has also been investigated as a potential partner technology for aPDT⁸.

CAP is a partially ionized gas produced at atmospheric pressure and near room temperature, which delivers reactive species, photons, and electric fields to biological targets. Various plasma sources, including surface dielectric barrier discharge, plasma jets, and plasma-activated water (PAW), have been developed for biological applications such as disinfection, wound management, and biofilm control⁹. The flexibility of parameters such as the electrical waveform, feed-gas composition, treatment distance, and exposure time makes it possible to control the generation of reactive species and improve the reproducibility and standardization of plasma treatment¹⁰. In vitro and preclinical studies have shown that CAP consistently reduces the microbial load and disrupts biofilms, including those formed by antibiotic-resistant microorganisms involved in wound- and device-related infections¹¹. PAW, a secondary product of CAP, serves as a strong oxidant that reduces microbial contamination and biofilm formation^12,13. These characteristics highlight the versatility of CAP, which can be integrated with other technologies to enhance its antimicrobial efficacy and extend its applications.

Both PDT and CAP are based on oxidative processes with broad applications, making their combined use worth exploring. However, existing studies have limitations in that the device types, photosensitizers, and operating conditions are inconsistent, and the effects of PDT and plasma are not distinguished in detail. These limitations highlight the need for carefully controlled and standardized studies that clearly distinguish the individual effects of PDT and plasma. Accordingly, we systematically investigated the antimicrobial effects of combined PDT and plasma treatment against methicillin-resistant Staphylococcus aureus, with an emphasis on understanding their respective roles and time-dependent interactions.

Methods

Experimental setup

Test chamber and source geometry

Figure 1A shows the custom stainless-steel chamber (150 mm × 150 mm × 115 mm; L × W × H). The chamber accommodated both a red organic light-emitting diode (OLED) module (A9F4C0A; Konica Minolta Inc., Tokyo, Japan) (Fig. 1B) mounted on the lower deck facing upward toward a standard polystyrene 6-well plate, and a surface dielectric barrier discharge (SDBD) (Fig. 1C) plasma source mounted on the upper panel facing downward. The distance between the grounded electrode surface and liquid surface in the 6-well plate was fixed at 5 cm using a spacer, and the electrode-to-well alignment was kept constant for all experiments. For all the assays, 3.0 mL of a liquid sample (deionized water, methylene blue, or bacterial suspension) was placed in each well. The plate lid was removed during the treatment, and the chamber door was closed. The chamber was flushed with ambient air to remove residual reactive species and restore baseline conditions.

Fig. 1.

Experimental configuration of the photodynamic therapy (PDT) and surface dielectric barrier discharge (SDBD) system. (A) Stainless-steel chamber (150 mm × 150 mm × 115 mm) housing a red organic light-emitting diode (OLED) beneath and an SDBD above a 6-well plate; sources can be operated independently or simultaneously at fixed geometry. (B) OLED emission area. (C) Configuration and structure of the SDBD electrode; top view showing the electrode array pattern; side view illustrating the layered assembly of the high-voltage electrode, dielectric barrier, and ground electrode.

Photodynamic therapy (PDT)

An OLED (A9F4C0A) emitting red light with a dominant peak at 632 nm was used as the photodynamic irradiation source (Fig. 2A). The device was driven in the constant-current mode at 100 mA using a source meter (Keithley 2601A; Tektronix Inc.). Irradiance (22 mW/cm²) was determined using a spectroradiometer (CS-2000; Konica Minolta, Inc.), as previously described¹⁴. The absolute irradiance spectrum was measured using a spectrometer (HR2000+, Ocean Insight) after absolute calibration using a calibration lamp (HL-3plus-CAL, Ocean Optics). The optical fiber was positioned at the center of the space inside the 6-well plate to measure the average irradiance. This spectral value was confirmed by integrating the wavelength and comparing it with the irradiance values.

Fig. 2.

Opto-electrical characterization of sources. (A) OLED emission spectrum (λ_max = 632 nm). (B) Absorbance spectrum of methylene blue, highlighting Q-band peaks at 611 and 661 nm. (C) Voltage and current waveforms of the SDBD. (D) Optical emission spectrum of the SDBD plasma.

Methylene blue (MB) (Sigma-Aldrich, St. Louis, MO, USA) stock (1 mM in deionized water) was 0.22 μm filtered (PTFE; Jet Biofil) and stored at 4 °C in the dark. Working solutions were freshly prepared, and the absorption Q-band maximum was 661 nm (Fig. 2B).

Surface dielectric barrier discharge (SDBD)

Cold atmospheric plasma was generated by an SDBD driven by an AC power supply (POWERSOFT, Korea) at a maximum voltage of 2.2 kV and a frequency of 30 kHz (Fig. 2C). Current and voltage waveforms were measured using a current probe (Model 110; Pearson Electronics, Palo Alto, CA, USA) and high-voltage probe (P6015A, Tektronix, Beaverton, OR, USA), respectively, with a 2 Series MSO Mixed Signal oscilloscope (Tektronix, Beaverton, OR, USA). Optical emission spectra (OES) were acquired using a fiber-coupled spectrometer (Maya 2000 PRO; Ocean Insight, Orlando, FL, USA). OES revealed dominant N₂ second-positive system bands and N₂⁺ first-negative features that are characteristic of air-fed SDBD (Fig. 2D). The dielectric barrier material of the SDBD electrode was alumina (Al₂O₃). Between experiments, the test chamber was operated in a fume hood; after each treatment, the chamber was opened and a blower was used to exchange the air inside the chamber with fresh ambient air to promote convective cooling of the electrode and chamber. The electrode surface temperature was monitored using a thermal imager (TiS40, Fluke, USA), and the next run was initiated only after the electrode temperature returned to near room temperature (25 ± 2 °C).

Reactive species interaction modeling under concurrent OLED and SDBD operation

The simultaneous operation of OLED illumination and SDBD plasma can induce photolysis-driven reactions among the plasma-generated gas-phase species. To assess possible interference, we numerically estimated the photolysis of reactive species under red OLED emission.

The gas-phase species generated by SDBD were first analyzed using optical absorption spectroscopy based on the Beer–Lambert law:

where and are the absorption cross-section (cm² molecule^{− 1}) and number density (molecule cm^{− 3}) of species j, and l is the optical path length. As shown in Fig. 3A, the absorption cross sections were referenced from established databases¹⁵. Optical path length is equal to chamber size (15 cm) since the collimating lens was mounted the chamber wall as shown in Fig. 1A. The absorption signal was acquired by incident light lamp (Ocean Optics, DH-2000) and spectrometer (Ocean Optics, HR-6) through two collimating lens (Ocean Optics, UV-74). And the example of absorption signal and fitting procedure is in Fig. 3B which is obtained at 5 min during experiment. Four chemical component (O₃, NO₂, NO₃, N₂O₅) cross-section considering its contribution was synthesized as cyan line and which is well fitted to the measured data (black dot) with high R² value near 0.99. Optical depth τ(λ, t) was obtained every 10 s, and gas-phase concentrations were estimated following the procedure of Huh et al.¹⁵. Under SDBD-only conditions (OLED off), O₃ and N₂O₅ were dominant, with peak concentrations near 1100 ppm and 440 ppm, respectively, while NO₂ stabilized near 24 ppm and NO₃ gradually increased to approximately 1 ppm after 10 min (Fig. 3C).

Fig. 3.

Absorption cross sections and simulated concentration changes of atmospheric reactive species under SDBD and concurrent OLED+SDBD operation. (A) Reference absorption cross sections of major atmospheric reactive species, including O₃ (200–650 nm); panel adapted from Huh et al.¹⁵. (B) Example of the Beer-Lambert law fitting procedure at 5 min (during the experiment). (C) Simulated temporal evolution of O₃, NO₂, NO₃, and N₂O₅ concentrations during SDBD operation (20 min). (D) Modeled differential concentration (Δ) of the same species under concurrent OLED (red light) and SDBD exposure.

Direct quantification during concurrent OLED + SDBD operation was not attempted because the OLED emission band overlapped with the absorption range of nitrogen oxides, making simultaneous detection unreliable. Instead, numerical calculations were conducted to estimate the photolysis-driven redistribution of the reactive species under OLED illumination. As shown in Figs. 2A and 3A, the emission of the red OLED partially overlaps with the NO₃ absorption band, whereas the NO₂ absorption is minor and has a near-zero quantum yield¹⁷. The potential photochemical flux of each species was calculated as follows:

where (λ,T) is the quantum yield, (λ,T) the absorption cross-section, and F(λ) the measured irradiance per wavelength (Fig. 2A). Table 1 summarizes the relevant reactions (R1–R10) and parameters^16–19. NO₃ photolysis occurs via two pathways (R1–R2), generating NO and O(³P), which further participate in secondary reactions (R4–R10) involving O₃, NO₂, and N₂O₅. Therefore, we used simple ordinary differential equations (ODEs), including all reactions due to photolysis (R1–R10), to simulate the chemical concentrations under red-light exposure in Python. The initial value of each chemical species was fixed at the highest value for each species for simplicity of calculation and analysis. The simulated changes during OLED irradiation indicated small variations (Fig. 3D): NO₂ decreased by 1 ppm, O₃ by 0.2 ppm, and N₂O₅ increased by 0.5 ppm. NO₃ increased by 0.003 ppm (approximately 0.003% of the initial value). These small changes indicate that under the present experimental and modeling conditions, the red OLED emission did not measurably alter the plasma-generated NO_x composition.

Table 1.

Photolysis and subsequent reactions in the OLED-SDBD interaction model.

No.	Reactions	Quantum yield (560–750 nm)	References
R1	NO3 + hν → NO2 + O(3P)	0–1	16
R2	NO3 + hν → NO + O2	0–0.2	16
R3	NO2 + hν → NO + O(3P)	0	17

No.	Reactions	Reaction rate	References
R4	O(3P) + O2 + M → O3 + M	6.0 × 10−34 (T/300)−2.4	18
R5	O(3P) + O3 → 2O2	8.0 × 10−12 exp(−2060/T)	18
R6	O(3P) + NO2 → NO + O2	5.5 × 10−12	18
R7	NO + O3 → NO2 + O2	3.0 × 10−12 exp (−1500/T)	19
R8	NO3 + NO → 2NO2	1.5 × 10−11 exp(170/T)	19
R9	NO3 + NO2 + M → N2O5 + M	3.7 × 10−42 exp(−1600/T)	18
R10	N2O5 + M → NO3 + NO2 + M	4.92 × 10−11 exp(−10040/T)	18

Bacterial strain and culture

Methicillin-resistant Staphylococcus aureus ATCC 33591 (S. aureus) was cultured in tryptic soy broth (Difco™, Becton, Dickinson and Company, Sparks, MD, USA) at 37 °C for 18 h in a shaking incubator (160 rpm). Cells were harvested by centrifugation at 2800 × g for 15 min at 4 °C (Model 1580R; Labogene, Daejeon, Korea) and washed twice with 0.85% (w/v) saline. The resulting suspension was adjusted to 10⁷–10⁸ colony-forming unit (CFU)/mL immediately before the experiments, as verified by serial dilution and plate counting.

Single and combined treatment of PDT and SDBD

The bactericidal effects of PDT, SDBD, and their combination were evaluated against S. aureus. To determine an appropriate MB concentration for PDT experiments, MB was selected using a two-step approach. First, the dark toxicity of MB was evaluated to ensure that MB alone did not significantly reduce bacterial viability under dark conditions. For this screening, the MB solution and S. aureus suspension were mixed at a 1:1 (v/v) ratio to obtain final MB concentrations of 0, 25, 50, 75, and 100 µM, and the mixtures were incubated at 25 °C for 60 min in the dark. Based on these results, MB concentrations of ≤ 50 µM were considered suitable for subsequent PDT testing (Supplementary Fig. S1). Second, within this range, PDT was performed at final MB concentration of 0, 1, 10, and 50 µM under OLED irradiation (22 mW/cm², 20 min) to identify an optimal photosensitizer concentration. Based on the resulting log reductions, 50 µM was selected for subsequent experiments. For all assays, the MB solution and S. aureus suspension were mixed at a 1:1 ratio and incubated for 20 min at 25 °C in the dark (the control group received sterile deionized water instead of MB). Accordingly, the MB solution was prepared at 2× the target concentrations so that the final MB concentrations were achieved after 1:1 mixing. Subsequently, 3 mL of the mixture was transferred to a 6-well plate, placed at the designated position on the stand inside the test chamber, and exposed to OLED light, SDBD, or both treatments in combination. Immediately after treatment, samples were transferred to 15 mL tubes and mixed with Dey-Engley neutralizing broth (Sigma-Aldrich, St. Louis, MO, USA) at a 1:9 ratio for neutralization, and then serially diluted. Aliquots of 100 µL from each dilution were spread on tryptic soy agar (TSA; Difco™, Becton, Dickinson and Co., USA) plates, which were incubated at 37 °C for 24 h, after which colonies were enumerated.

Analytical assays

MB degradation

MB degradation during SDBD was monitored by measuring the UV-Vis absorbance at 661 nm. Samples containing 50 µM MB (final concentration after 1:1 mixing, identical to the bactericidal assays) were prepared and subjected to SDBD treatment for 0, 10, and 20 min. The experimental conditions were identical to those used in the bactericidal assays, in which 3 mL of the sample was placed in a 6-well plate, treated inside the chamber, and subsequently collected. In addition, MB quantification was performed in the PDT alone and combined PDT+SDBD treatments to perform bactericidal assays. Absorbance was measured using a UV-Vis spectrophotometer (SPECTROstar Nano, BMG LABTECH, Ortenberg, Germany). Calibration curves (0–50 µM) were generated for each day of measurements, and sample absorbance values were converted to concentrations after blank subtraction. The calibration curve for MB showed linearity (R² = 0.9958), enabling reliable quantification (Fig. S2A).

Detection of reactive species and physicochemical parameters

Reactive oxygen species were assayed under two conditions: (i) photodynamic illumination in MB-containing suspensions and (ii) SDBD exposure in dye-free deionized water (DIW) unless otherwise specified. All fluorescence assays were conducted in black 96-well plates at a final volume of 200 µL per well. Probe working solutions were prepared at 2× concentration and mixed with an equal volume of sample (1:1, v/v) to obtain the indicated final probe concentrations; samples were mixed and protected from ambient light. Singlet oxygen (¹O₂) was monitored using Singlet Oxygen Sensor Green (SOSG, Thermo Fisher) with fluorescence at 504/525 nm (SpectraMax M2e; Molecular Devices, San Jose, CA, USA). The SOSG-based ¹O₂ assay was performed with minor modification of the method described by Nie et al.²⁰. Briefly, SOSG was added with 50 µM MB suspensions (final SOSG 25 µM) and measured after 20 min illumination. Hydroxyl radicals (·OH) were detected using coumarin-3-carboxylic acid (3-CCA, 100 µM) with readout of 7-hydroxy-3-CCA at 346/452 nm with minor modifications of previously reported method²¹. Briefly, an equal volume of 3-CCA solution and sample was mixed (1:1, v/v) to yield a final 3-CCA concentration of 100 µM. For PDT, 3-CCA was mixed with MB-containing suspensions, and for SDBD, 3-CCA was mixed with DIW. Fluorescence was measured immediately after 20 min treatment (PDT illumination or SDBD exposure). Fluorescence signals were normalized to the control group, and are reported as relative fluorescence intensity (%). Hydrogen peroxide (H₂O₂) was quantified using a fluorimetric assay (MAK165, Sigma-Aldrich) per the manufacturer’s instructions (λ_ex = 540/λ_em = 590 nm) with a seven-point calibration curve (0–10 µM). Nitrite and nitrate concentration were quantified using a commercial Nitric Oxide Assay Kit (Invitrogen, EMSNO) according to the manufacturer’s instructions. For RNS measurements, SDBD treatment was applied to DIW under the same conditions as the bactericidal assays, and samples were collected immediately after 0, 10, and 20 min of SDBD treatment. Nitrite concentration was measured directly via Griess reaction at an absorbance of 540 nm. To determine nitrate levels, nitrate was reduced to nitrite by nitrate reductase, and the nitrate concentration was calculated by subtracting the initial nitrite value from the total nitrite concentration after reduction. The detection limits of the assay were 0.222 µM for nitrite and 0.625 µM for nitrate, respectively. The pH and bulk temperature were measured immediately after the treatment using a calibrated meter (Seven2Go; Mettler-Toledo, Columbus, OH, USA).

Acute cytotoxicity screening

Acute cytotoxicity was assessed using NIH/3T3 fibroblasts (ATCC CRL-1658). Cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37 °C in a humidified incubator with 5% CO₂. For the assay, 2 mL of cell suspension (1.0 × 10⁵ cells/well) was seeded into 6-well plates and pre-incubated for 24 h. Cells were then assigned to the following groups: untreated control (Con), MB only (MB), OLED-PDT (PDT), SDBD only (SDBD), and simultaneous PDT+SDBD (PDT+SDBD). MB was applied at a final concentration of 50 µM for the MB-containing groups (MB, PDT, SDBD, and PDT+SDBD), consistent with the bactericidal assays.

Cells were exposed to OLED irradiation and/or SDBD under the same exposure configuration and treatment times (10–20 min) used in the antimicrobial experiments. Immediately after exposure, the treatment solutions were removed and the cells were gently rinsed with PBS. Fresh culture medium was added, and cell viability was quantified using a Chromo-CK cell Viability Assay Kit (CH-10000) according to the manufacturer’s instructions. Briefly, 200 µL of Chromo-CK (CCK-8) reagent was added to each well, followed by incubation for 1 h at 37 °C (protected from light), and absorbance was measured at 450 nm using a microplate reader. Cell viability was calculated by normalizing background-corrected absorbance values to the untreated control.

Statistical analyses

All experiments were performed with at least three independent replicates, and results are presented as mean ± standard deviation (SD). Differences among multiple groups were analyzed using one-way analysis of variance (ANOVA), followed by Tukey’s post-hoc test for multiple comparisons. For factorial designs, two-way ANOVA including the interaction term was applied, followed by Tukey’s post-hoc test for multiple comparisons. For direct comparisons between two groups, an independent t test was used. Statistical analyses were performed using OriginPro 2022b (OriginLab Corporation, Northampton, MA, USA). Statistical significance was defined as p < 0.05.

Results

Effect of MB concentration on PDT-mediated bacterial inactivation

As shown in Fig. 4A, exposure of S. aureus to MB in the absence of light did not reduce bacterial counts, even at 50 µM, confirming that MB alone was not cytotoxic under the conditions tested. However, OLED illumination (22 mW/cm², 20 min) in the presence of MB induced a concentration-dependent reduction in cell viability (p < 0.05). Mean survival levels were 7.27, 7.27, 5.90, and 3.78 log CFU/mL at 0, 1, 10, and 50 µM MB, respectively. Among the tested conditions, 50 µM MB under illumination achieved the greatest inactivation after 20 min, and this condition was therefore selected for subsequent experiments. Correspondingly, singlet oxygen production was detected in the 50 µM MB samples following illumination (Fig. 4B), supporting a photodynamic origin for the observed antimicrobial effect. Consistently, the hydroxyl radical signal did not increase during PDT (Fig. 4C), indicating that, under these conditions, the response was dominated by singlet oxygen-type pathways rather than by radical chemistry.

Fig. 4.

Bactericidal efficacy and reactive-species generation under individual PDT and SDBD treatments. (A–C) Photodynamic therapy (PDT) using a red OLED (22 mW/cm², λ_max = 632 nm, 100 mA) in the presence of methylene blue (MB; 0–50 µM). (A) Survival of S. aureus after 20 min illumination. (B) Singlet oxygen (¹O₂) detection by singlet oxygen sensor green (SOSG) fluorescence (50 µM MB, relative intensity at 0 vs. 20 min). (C) Hydroxyl radical (·OH) detection by coumarin-3-carboxylic acid (CCA) fluorescence (0 vs. 20 min). (D–F) Surface dielectric barrier discharge (SDBD) plasma treatment with or without MB (50 µM). (D) Survival of S. aureus after 0-, 10-, or 20-min treatment. (E) Hydrogen peroxide (H₂O₂) concentration in MB-free, cell-free solution after 20 min treatment. (F) Hydroxyl radical detection by CCA fluorescence in MB-free, cell-free solution (0 vs. 20 min). Bars are mean ± SD (n = 3). Different letters indicate significant differences (two-way ANOVA, Tukey’s HSD, p < 0.05). *Unpaired two-tailed t test comparing 0 vs. 20 min; p < 0.05.

Effect of SDBD treatment on bacterial survival and MB

SDBD plasma treatment induced a time-dependent reduction in S. aureus count (Fig. 4D). In the absence of MB, the number of viable cells decreased from 7.10 log CFU/mL to 4.76 log CFU/mL after 10 min, and the counts reached the detection limit by 20 min (p < 0.05). In the presence of MB, no significant reduction was observed after 10 min, whereas a modest but significant decrease was observed after 20 min (p < 0.05). Thus, MB transiently attenuated the bactericidal effects of SDBD in the early period. Importantly, hydrogen peroxide accumulated (Fig. 4E) and hydroxyl radical generation was detected after 20-min of SDBD treatment (Fig. 4F), supporting the contribution of reactive oxygen species to plasma-mediated inactivation. In addition, SDBD treatment resulted in the progressive degradation of MB, as indicated by the decrease in absorbance at 661 nm (Fig. S2B). The MB content decreased significantly at 10 min (p < 0.05) and was nearly depleted by 20 min, as also evident from the fading blue coloration (Fig. S2B, insert).

Combined PDT and SDBD treatment

At each treatment time (10 and 20 min), the combined PDT+SDBD was compared with the corresponding single-treatment groups (Fig. 5A). At 10 min, PDT and PDT+SDBD significantly reduced viable counts relative to the control, whereas SDBD alone did not differ from the control; moreover, PDT and PDT+SDBD were not significantly different at 10 min. At 20 min, all active treatments (PDT, SDBD, and PDT+SDBD) reduced viability relative to the control, and PDT+SDBD achieved the greatest reduction, being significantly more bactericidal than either PDT or SDBD alone.

Fig. 5.

Bacterial viability and physicochemical changes during PDT, SDBD, and their combined treatment. (A) Viable counts of S. aureus (Log CFU/mL) for control, PDT, SDBD, and PDT+SDBD at 10 and 20 min. (B) Residual methylene blue (MB) concentration for the same groups/time points. (C) Bulk pH and (D) temperature measured immediately after treatment under the same conditions. Bars are mean ± SD (n = 3). Different letters indicate significant differences (two-way ANOVA, Tukey’s HSD, p < 0.05).

MB quantification (Fig. 5B) showed that MB remained unchanged under PDT alone, whereas plasma conditions (SDBD and PDT+SDBD) caused rapid MB degradation at 10 min and near-complete depletion by 20 min. Physicochemical measurements supported non-thermal operation and acidification under plasma exposure: pH decreased after SDBD and PDT+SDBD compared with control and PDT alone (Fig. 5C), and bulk temperature rose across all active treatments but remained ≤ 32 °C (Fig. 5D). Together, these data indicated a time-dependent shift from PDT-dominated to plasma-driven killing after MB depletion.

Acute cytocompatibility screening

Figure 6 summarizes the acute cytocompatibility screening results. At 10 min, only the combined PDT+SDBD condition showed a significant decrease in metabolic activity compared with the control, whereas MB, PDT, and SDBD alone did not differ from the control. At 20 min, metabolic activity decreased in all treatment groups, and PDT+SDBD showed the lowest viability among the tested conditions.

Fig. 6.

NIH/3T3 fibroblast viability after PDT and/or SDBD exposure (CCK-8). Cells were treated for 10–20 min (Con, MB, PDT, SDBD, PDT+SDBD). After treatment, cells were washed twice with PBS, replaced with fresh medium, and CCK-8 absorbance was measured at 450 nm after 1 h. Values were normalized to control (100%) and presented as mean ± SD (n = 3). Different letters indicate significant differences (two-way ANOVA, Tukey’s HSD, p < 0.05).

Discussion

This study showed that combining PDT with CAP resulted in greater and more consistent inactivation of S. aureus than either approach alone. The overall trend indicated two stages: In the early phase, bacterial inactivation was mainly driven by photodynamic oxidation, while methylene blue (MB) was still present. In the later phase, the plasma-derived reactive species became more important after MB degradation. This sequential transition is illustrated in the proposed working model (Fig. 7).

Fig. 7.

Conceptual model of the time-dependent shift from photodynamic- to plasma-driven bacterial inactivation during combined PDT and SDBD. Early (≤ 10 min): Methylene blue (MB) is abundant, and bacterial killing is primarily mediated by photodynamic action. Mid (10–20 min): MB declines and plasma influence increases, leading to a transition in the dominant inactivation pathway. Late (≥ 20 min): MB is largely depleted, and plasma-derived chemistry sustains inactivation despite ongoing illumination. Schematic is conceptual and not to scale; see Fig. 5 for supporting data.

Before analyzing these two stages in detail, we considered whether the simultaneous operation of OLED illumination and SDBD could alter the distribution of reactive species via optical photolysis. Because direct gas-phase measurements were not feasible, we modeled this interaction using irradiance data and established absorption cross sections. Under the present experimental and modeling conditions, the calculated concentration changes of O₃, NO₂, NO₃, and N₂O₅ were all within ± 1 ppm, indicating that the red OLED emission did not meaningfully perturb the plasma-generated NO_x composition in this study.

NO is one of the products of NO₃ photolysis and rapidly reacts with existing O₃ (R7) because the reaction rate is much higher than that of another NO quenching reaction (R8). The other product, O(³P), involved quick reactions (R4, R5), and through R6, NO was produced, which was also a rapid reaction of R7. This indicated that the NO₃ photolysis products reacted with O₃ and O₂ to produce NO₂. Here, the equilibrium reaction (R9, R10) was re-established within tens of seconds at room temperature (300 K) owing to increased NO₂; therefore, transient deficits in NO₃ were promptly back-filled by R10. Consequently, the 20 min net changes in NO₃ and N₂O₅ remained within the experimental variability, indicating negligible interference between light and plasma chemistry under our operating conditions, as N₂O₅ acts as a buffer.

Initially, illumination of MB with an OLED produced singlet oxygen, which caused oxidative damage to the cellular membranes, proteins, and nucleic acids²². The SOSG assay confirmed a higher fluorescence after 20 min of light exposure, whereas the hydroxyl radical signal remained unchanged. The OLED source was particularly effective in providing uniform red light over a large area with minimal heating, and its emission wavelength matched well with the absorption band of MB, thereby ensuring efficient excitation. Previous studies have suggested that MB-PDT with flexible OLEDs effectively inactivates S. aureus^14,23. This evidence supports rapid early inactivation by MB-PDT in our system.

During the later stages of the study, exposure to SDBD plasma resulted in the accumulation of hydrogen peroxide and hydroxyl radicals, accompanied by a decrease in pH levels. These findings align with established literature on plasma-liquid interactions, which indicate that long-lived reactive species can form in the aqueous phase, and that the dissolution of nitrogen oxides into the liquid contributes to acidification²⁴. To experimentally support the transfer of reactive nitrogen species into the liquid under our configuration, we additionally analyzed deionized water treated by SDBD for 10 and 20 min (Table S1). The results showed marked acidification together with nitrate accumulation, while nitrite remained below the detection limit (< 0.222 µM), consistent with NO_x dissolution and subsequent aqueous-phase chemistry leading to stable nitrate formation. This trend is consistent with plasma-activated water (PAW) studies showing that reactive nitrogen species composition in the liquid can evolve rapidly after plasma exposure, where nitrite can be depleted through secondary reactions and/or re-emission of NO/NO₂ favored by their low Henry’s law solubility, while nitrate persists as a more stable end product²⁵.

Although ozone was not measured directly in this study, an increase in peroxide and ·OH levels, and a decrease in pH provide a reasonable explanation for the oxidative activity. Importantly, SDBD decomposed MB, thereby affecting MB-mediated PDT. Previous studies have demonstrated that MB gradually degrades under plasma exposure with a loss of chromophore activity and oxidized intermediates²⁶. Similar observations from ozonation studies indicate that both the ozone and hydroxyl radical pathways can drive MB oxidation depending on the solution conditions²⁷. This may cause a decrease in the efficiency of the photodynamic action and mitigate the risks associated with residual photosensitizers (e.g., tissue staining and local phototoxicity) if the dye remains after treatment. Taken together, these findings suggest that late-stage inactivation is mainly maintained by the SDBD-derived reactive species after MB decomposition.

Another important observation was that the presence of MB influenced the microbial inactivation of the plasma. It should be noted that in our system, liquid suspensions contain both dyes and bacteria, and the reactive species often react rapidly with soluble organic matter. MB reacts rapidly with hydroxyl radicals, meaning that it can consume these short-lived reactive species, thereby competing with bacterial targets for the same reactive species²⁸. This pattern, together with the observed decrease in SDBD efficiency at 10 min when MB was still abundant, suggests temporary competition for reactive species in the solution. Similar reductions in plasma-activated liquid efficacy have been reported in the presence of organic loads^29,30.

Overall, these findings explain the following dynamic sequence: photodynamic reactions dominate in the early stages, plasma-derived reactive species sustain killing in the late stages, and dyes modulate both processes. These results suggest two important points for improving the combined treatment. First, it is important to preserve photodynamic activity before significant MB degradation, and then the plasma contributes to continuing antimicrobial action without thermal stress. Second, while plasma-driven MB degradation may reduce residual dye, the present cytotoxicity screening indicates that the antimicrobial exposure conditions are not yet optimized for mammalian cytocompatibility.

Acute NIH/3T3 screening (CCK-8) showed reduced metabolic activity under the same exposure configuration, particularly for prolonged (20 min) PDT+SDBD treatment. These findings indicate the further optimization of treatment parameters (e.g., exposure time, plasma power/duty cycle, irradiance/fluence, MB concentration, and treatment sequencing) is required to improve cytocompatibility while maintaining antimicrobial efficacy for translational applications. Future studies should consider wider ranges of variables and experimental factors. Applying specific scavengers (e.g., histidine or azide for singlet oxygen, catalase for hydrogen peroxide, and cPTIO for nitric oxide-related species) can help distinguish the individual contributions of PDT, plasma, and the different reactive species involved. Finally, it is important to evaluate the potential positive and negative effects that can occur depending on the type and properties of the photosensitizer and plasma used.

Conclusion

This study demonstrated that the combined PDT and SDBD treatment produced a sequential, time-dependent shift in the dominant inactivation pathway. Early inactivation was driven mainly by MB-mediated photodynamic actions, whereas the later phase was sustained by plasma-derived reactive species after MB degradation. These findings are limited to the specific conditions used in this study, such as the plasma device setup, operating environment, and type and concentration of the photosensitizer. The outcomes may differ from those of other plasma sources, discharge parameters, or microbial targets; therefore, the results should not be generalized without caution. Acute NIH/3T3 screening under the same exposure configuration showed reduced metabolic activity at the antimicrobial exposure conditions, indicating that cytocompatible dose optimization will be required prior to translational use. Therefore, future studies should test various conditions, including alternative photosensitizers, plasma settings, and applications in surface or biofilm models. Systematic exploration of these variables will help determine whether the sequential inactivation pattern is reproducible and define the conditions under which PDT and CAP can be most effectively combined.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

K.H.B. was involved in conceptualization, data curation, formal analysis, investigation, methodology, validation, visualization, writing—original draft. J.Y.P. and Y.Y. were involved in data curation, resources. J.C. and S.J. were involved in methodology, resources. Y.-J.C. and M.-J.C. were involved in investigation, resources. H.I.Y. and S.L. was involved in funding acquisition, project administration, resources, supervision, writing—review and editing.

Funding

This work was supported by the Fundamental Research Program (PNKA630) of the Korea Institute of Materials Science.

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.