White-Light-Activated Antibacterial Surfaces Generated by Synergy between Zinc Oxide Nanoparticles and Crystal Violet

The prevalence of hospital-acquired infections (HAIs) caused by multidrug-resistant bacteria is a growing public health concern worldwide. Herein, a facile, easily scalable technique is reported to fabricate white-light-activated bactericidal surfaces by incorporating zinc oxide (ZnO) nanoparticles and crystal violet (CV) dye into poly(dimethylsiloxane). The effect of ZnO concentration on photobactericidal activity of CV is investigated, and we show that there is synergy between ZnO and CV. These materials showed highly significant antibacterial activity when tested against Staphylococcus aureus and Escherichia coli under white light conditions. These surfaces have potential to be used in healthcare environments to decrease the impact of HAIs.


INTRODUCTION
Hospital-acquired infections (HAIs) are among the most serious complications of hospital care associated with extended duration of hospital stays, increased morbidity, and increased medical costs. According to the U.S. Center of Disease Control and Prevention, about 1 in 25 hospital patients acquires at least 1 HAI, equivalent to 722 000 infections in the U.S. in 2011, and resulting in 75 000 deaths during hospitalization. 1 The treatment of these infections costs to the U.S. government app. $35−$45 billion per year. 2 In the U.K., 300 000 healthcareassociated infections occur each year and cost over £1 billion to the National Health Service (NHS). In healthcare facilities, 80% of infections are because of contaminated touch surfaces. Therefore, new antimicrobial surfaces need to be developed to minimize the risk of spreading bacteria.
A number of studies have reported novel and efficient antimicrobial surfaces including surfaces releasing microbicides, 3,4 antibiotics, 5,6 and quaternary ammonium salts 7,8 and copper surfaces. 9,10 Photodynamic therapy (PDT) has been raised as a promising alternative to current techniques. 11 Antimicrobial photodynamic therapy, a type of PDT, is based on the use of nontoxic light-activated antimicrobial agents (LAAAs), known as photosensitizers, that are irradiated by a light source with an appropriate wavelength. 12 In this process, these agents produce reactive oxygen species (ROS) that act on bacteria via multiple attack pathways including the oxidative destruction of cellular membranes, intracellular proteins, and DNA. Given the multiple targets of ROS, the emergence of bacterial resistance is extremely unlikely. 13−15 Zinc oxide (ZnO) is an inexpensive, wide band gap (3.3 eV at room temperature) semiconducting material that has excellent optical and electronic properties, finding a wide range of applications from field effect transistors to gas sensors and in dye-sensitized solar cells. 16 Moreover, ZnO has the inherent advantage of showing antibacterial activities against a broad range of Gram-positive and Gram-negative bacteria. 17−19 The antibacterial activity of ZnO has been well exploited under irradiation with light (UV or white light) and in the absence of light. 19−23 Similar to that of LAAAs, the proposed mechanism of its antibacterial activity is thought to be via oxidative stress as a result of ROS production. 24 Regarding surface application of LAAAs, it has been previously demonstrated that photosensitizers (e.g., crystal violet (CV) and methylene blue) can be successfully encapsulated into a variety of different polymers by a facile "swell−encapsulation−shrink" technique 25−27 and these polymers demonstrate efficacious photoactivated antibacterial activity under irradiation with white light. Also, dyeimmobilized polymeric materials have demonstrated an enhanced photobactericidal activity when incorporated with different nanoparticles (NPs) (e.g., silver and gold). 26,28,29 Even though these materials show potent photobactericidal activity, they include complex, labor-intensive, time-consuming synthesis of NPs. Also, these NPs can be implemented only to soft materials such as polymers and silicones by the swell− encapsulation−shrink process.
In our previous work, poly(dimethylsiloxane) (PDMS)/ZnO NP composites with 0.36 wt % ZnO were fabricated through a solvent-aided dispersion process using chloroform as a solvent. 30 Then, ZnO-incorporated polymer squares were coated with CV and these samples induced the lethal photosensitization of Escherichia coli within 4 h and Staphylococcus aureus within 1 h under white light conditions (∼10 500 lux). 30 In the present study, a similar, but a simpler, more versatile and scalable approach is reported for fabricating highly efficacious white-light-activated self-sterilizing surfaces based on three components ZnO NPs, CV, and PDMS to overcome the aforementioned problems. The main objective of this study was to investigate whether the photobactericidal activity of CV is influenced by the ZnO content. Also, we utilized a lower white light intensity compared to that in the previous work (∼6500 lux) in antimicrobial tests. The results showed that the synergistic combination of CV and ZnO NPs led to superior antibacterial activity of the polymers.

RESULTS AND DISCUSSION
2.1. Materials Synthesis. CV and ZnO NPs were incorporated into PDMS by a simple two-step procedure to fabricate a potent bactericidal polymer−nanocomposite ( Figure  1). The first step involved mixing ∼60 nm ZnO NPs (see Supporting Information, Figure S2a) with PDMS by mechanical rabbling on the basis of a series of PDMS/ZnO NPs with varying nanoparticle content, as shown in Table 1. Then, the resultant milky polymer was degassed under vacuum to remove entrapped air bubbles. Finally, the polymer was cast on glass substrates and cured in a preheated oven at 100°C. Whereas PDMS is transparent, the ZnO-embedded polymers are white ( Figure 2).
In the second step, these modified polymers were coated with CV using a swell−encapsulation−shrink method. In this process, the polymer squares were dipped and left to swell in 750 ppm CV solutions in acetone for 24 h under dark conditions. Acetone induced polymer swelling, allowing dye penetration through the polymer matrix. After removing from the solution, the swollen polymer had shrunk to its original dimensions as the remaining solvent evaporated. This method generated dark purple polymer squares, indicating an efficacious uptake of the dye onto the polymer surfaces ( Figure 2).    Figure S3 shows the infrared absorbance spectra of the polymer films, from which no significant change was confirmed across the ZnO/PDMs composite films from 0 to 5 wt % of ZnO nanoparticle concentration before and after the swell−encapsulation−shrink process. The similarity between the untreated and treated samples can be due to the strong PDMS absorbance bands and the low concentrations of dye present in the polymer samples. Consequently, the results indicate that the incorporation of ZnO NPs and CV in the silicones had no significant influence on the internal structures of the films within the detection limit of Fourier transform infrared (FT-IR).
The presence of the ZnO NPs was confirmed by X-ray diffraction (XRD), photoluminescence (PL) spectroscopy, and functional testing of the polymer films. The X-ray diffraction of neither the 0.5ZnO nor the 1.0ZnO sample displayed any discernible reflection because of the low density of the ZnO NPs embedded within the polymers, whereas phase-pure ZnO in the wurtzite crystal structure was appeared when ZnO concentration increased to 5.0 wt % (Figure 3b) (peaks at 2θ = 31.7, 34.4, 36.2, 47.6, and 56.5°assigned to (100), (002), (101), (102), and (110) planes, respectively). Also, it can be seen that the incorporation of CV into the polymer does not impact the internal structure of the polymer film, in good agreement with the experimental results of Figure S3.
Photoluminescence measurements were made on the PDMSencapsulated materials with and without CV coating ( Figure 4). The ZnO-incorporated samples without CV displayed a clear, broad fluorescence band, peaking at ca. 520 nm, increasing in intensity with increasing ZnO loading. This is in agreement with reported values for ZnO nanoparticles of this size and with a powder sample of ZnO measured ex situ of the PDMS. 31,32 On coating with CV, the ZnO emission band is overwhelmed by CV emission.
The treated PDMS samples were also characterized using UV−vis absorbance spectroscopy ( Figure 5). The main and shoulder absorption peaks of PDMS containing CV alone are at λ ≈ 590 and 533 nm, respectively. The addition of ZnO NPs increased the maxima, intensity, and breadth of the CV peak, indicating that dye uptake was enhanced by increasing the ZnO concentration. This phenomenon has been demonstrated in another study in which gold nanoparticles enhanced the UV− visible absorption of toluidine blue. 33 However, for 5.0ZnO− CV, the absorption reached its maximum as expected because 5.0ZnO−CV was darker in color compared with any other CVcoated sample. The CV concentrations within the polymer were found to be 1.17 × 10 −4 , 6.64 × 10 −4 , 8.50 × 10 −4 , and  Moreover, a characteristic absorption peak at 375 nm for ZnO NPs suspended in water was seen in the range of UV (see Supporting Information, Figure S2b), indicating that ZnO NPs are UV-activated photocatalysts (<385 nm) and hence should not show significant antibacterial activity under white light conditions. However, when encapsulated in PDMS, there was no characteristic UV−vis peak of ZnO because of high nanoparticle concentrations.
Wetting properties of the samples were characterized by contact angle (CA) measurements using water as a test liquid.

Article
As demonstrated in Table 1, bare PDMS is inherently hydrophobic having a CA of 97.0 ± 1.8°, whereas no significant difference in the wetting properties of the treated samples was observed, varying in contact angle by ±6°.
To examine whether there was leaching of the dyeincorporated polymers, they were placed in 10 mL of phosphate-buffered saline (PBS) solution at 37°C and the CV leachates were measured spectroscopically as a function of time ( Figure 6). It can be seen that all of the CV-coated samples leached dye into PBS after 24 h and that the amount of CV leaching was correlated with increasing the nanoparticle weight. Over a period of more than 300 h, the CV concentration of all of the samples plateaus at below 8.5 × 10 −7 M and, overall, no further leaching was observed, indicating stability of the dye in the polymers. Crystal violet, a topical antiseptic, possesses antibacterial and antifungal properties. 34 It is still listed by the World Health Organization even though new drugs have superseded its use in the medical field, and clinical trials showed its efficiency to treat infected wounds, superficial skin, and methicillin-resistant S. aureus infections. 35,36 Hence, if there is significant leaching of the dye, it is not expected to show toxicity toward human cells.
In addition, a similar setup was used to detect whether the ZnO-incorporated samples release nanoparticles into PBS solution. There was no observable change in the amount of leaching, color, or conductivity of the solution (data not shown), showing that the ZnO in the polymer matrix is more stable to leaching in contrast to CV.

ACS Omega
Article source emitting an average light intensity of 6500 ± 300 lux at a distance of 25 cm from the samples (see Supporting Information, Figure S4). Also, a control sample set was tested in the dark for the same exposure times. Figure 7 demonstrates the bactericidal activities of the samples against S. aureus after 45 min of incubation in the dark and in white light. Under dark conditions, none of the polymer samples showed statistically significant reduction in S. aureus numbers. Similarly, in white light, no significant killing of S. aureus was observed on the 0.5ZnO sample surface compared to that on the control sample. However, a greater reduction in bacterial numbers was apparent on the surfaces of both the 1.0ZnO and 5.0ZnO samples (P-value < 0.01 for both). Furthermore, the 0.5ZnO−CV sample caused ∼1 log reduction in the number of viable bacteria, with enhanced killing compared to that on the sample containing CV alone (0.35 log reduction). The effect of both the 1.0ZnO−CV and 5.0ZnO−CV samples was even more remarkable, reducing bacterial counts to below the detection limit (≥4 log, P-value = 0.002) in the light, within 45 min.
The antibacterial properties of the same polymers were also investigated against E. coli under the same conditions but with a longer exposure time (Figure 8). In the dark, over a 90 min period, no statistically significant decrease in the numbers of E. coli was demonstrated by the samples containing ZnO NPs relative to that by the control polymer. On the other hand, except the polymer containing CV alone, all of the samples containing ZnO and coated with CV displayed statistically significant (P-value < 0.01) activity compared to that of the control sample.
After 90 min of white light exposure, the 0.5ZnO−CV sample and the 1.0ZnO−CV sample achieved 0.70 log and 1.33 log reductions in the numbers of E. coli (Figure 8). Moreover, the 5.0ZnO−CV sample displayed a highly significant reduction in bacterial numbers after 90 min of irradiation, with bacterial counts reduced to below the detection limit (≥4 log, P-value = 0.002).
The difference between the susceptibilities of the Grampositive (S. aureus) and the Gram-negative (E. coli) bacteria is most probably because of the differences in their cell wall structures. 37 Whereas Gram-positive bacteria possess a single thick peptidoglycan layer (20−80 nm), Gram-negative bacteria contain a thinner peptidoglycan layer (7−8 nm) but possess a second (outer) membrane, 16 reducing the permeability of radicals and making them less susceptible to dye-coated polymeric surfaces. 38,39 Overall, the results show that the polymer containing ZnO or CV alone does not possess potent antibacterial properties. Hence, it is required to combine both agents into the polymer to fabricate a superior self-sterilizing material that shows efficacious photobactericidal activity against both E. coli and S. aureus.
To evaluate the efficacy of these polymeric surfaces, their photobactericidal properties were compared with those of other antibacterial systems reported in the literature. For example, under white light conditions (40 000 lux), protoporphyrin-and zinc protoporphyrin-grafted nylon fibers resulted in 5.6% reduction in the numbers of S. aureus, whereas they did not show significant killing of E. coli. 40 Moreover, Sehmi et al. reported CV in combination with ∼18 nm ZnO NPs incorporated in polyurethane squares. Upon illumination with a white light source (6600 ± 990 lux), these polymer samples caused lethal photosensitization of both E. coli and S. aureus in 1 and 4 h, respectively. 41 In our previous work, CV was encapsulated into PDMS squares in combination with ZnO NPs used in the present study, and they demonstrated significant antibacterial activity against S. aureus and E. coli in 1 and 4 h, respectively, under white light conditions (∼10 500 lux). 42 These results validate the potent photobactericidal activity observed with the 5.0ZnO−CV surfaces reported here, achieving decreases in bacterial numbers of greater than 4 logs for both S. aureus and E. coli in only 45 and 90 min, respectively.
These antibacterial materials can also be compared with other antibacterial surfaces that showed enhanced antibacterial activity. For example, at a much lower white light intensity of ∼500 lux, significant lethal photodestruction of S. aureus and E. coli was achieved using polyurethane samples containing the combination of CV and ZnO NPs (∼2−4 nm in size). This enhanced antibacterial efficacy may be ascribed to the small size of ZnO NPs because there is a correlation between the decreasing size of ZnO NPs and their antibacterial activity. 19,43,44 However, this method involves complex, laborious, time-consuming synthesis of the NPs and is only applicable to soft materials. Therefore, the proposed method in this study outperforms its counterparts in many aspects such as its easy fabrication, the low-cost materials utilized, and the material's high versatility for being coated on a wide range of substrates (e.g. metal, glass, paper, and soft polymer plates) (see Supporting Information, Figure S5).
The antibacterial behavior of CV is well established. Upon light activation of the dye, two different photochemical reaction pathways occur simultaneously: one generates hydroxyl radicals ( • OH) and superoxide ions (O 2

•−
) through the electron transfer process (type I) and the other generates singlet oxygen ( 1 O 2 ) through energy transfer (type II). 45 Photogenerated ROS can attack bacteria via a non-site-specific multiple attack mechanism (unlike antibiotics, which usually have a singlet target) including the disruption of cellular membranes and DNA by means of oxidative damage, making bacteria unlikely to develop resistance because resistance normally occurs when a bactericide has a single target. 14,46,47 In contrast to that of CV, the killing mechanism of ZnO is still unclear, but there are various mechanisms proposed in the literature, including the production of ROS, 19,20,48−52 the release of zinc ions, 21,48,53,54 and the accumulation of NPs in the bacterial cytoplasm and on the outer membrane of Gramnegative bacteria. 19 Some studies showing significant antibacterial activity of ZnO in the dark attributed this to the release of zinc ions from a ZnO suspension, which penetrate into the bacteria, damaging DNA and resulting in cell death. 23,55,56 However, there are many studies indicating that the concentrations of released Zn 2+ ions were not high enough to cause nanotoxicity. 19,20,57,58 Lakshmi et al. proved that its antibacterial activity is mainly attributed to ROS in the dark 59 rather than release of zinc ions (zinc ion dissolution) and ZnO internalization. 22,23,60,61 The presence of surface defects on ZnO NPs can generate ROS under lower-energy-light illumination (i.e., visible light). Previous study has demonstrated that the encapsulation of ZnO in combination with CV into polyurethane can increase the photobactericidal activity of the dye via both type I and type 2 pathways, 41 which is in good agreement with the data presented here.
Regarding real-world applications, antimicrobial materials should to be durable, environmentally friendly, and not harmful to humans. 62 Therefore, it is imperative that the dye incorporated in the polymer does not leach significantly or

ACS Omega
Article rub off upon contact with hands or disinfectant wipes that are often used to reduce the risk of the spread of infection. We have demonstrated that the dye-coated samples proved to be stable to contact with aqueous solution at 37°. The amount of dye leached from the samples was relatively small relative to the concentration (20 000 ppm), resulting in dermal irradiation or sensitization over an extended period of clinical assessment. 36 Furthermore, to determine the stability of the modified polymers against standard hospital hygiene regime conditions, they were vigorously wiped with Clinell wipes that are utilized to clean hard surfaces on the wards in University College London Hospital NHS Trust hospitals. There was no visual evidence of dye removal from the modified polymer after rigorous wiping with 70% alcohol, proving its robustness against our standard hospital cleaning protocol and, hence, transmission of dye upon touching is unlikely. Similarly, the leaching of ZnO NPs and their transfer by human contact is minimal because we have shown that they are well embedded into the polymer. ZnO NPs have been widely utilized in the cosmetics and food industry, and their toxicity in humans has been extensively investigated, with the conclusion that it depends on many factors such as their size, shape, route of administration, and dosage. Studies have showed that their penetration through skin is improbable although other routes of administration of these nanoparticles can demonstrate harmful effects at high doses (>100 μg mL −1 ). 63,64

CONCLUSIONS
In this study, ZnO NPs have been successfully embedded with different nanoparticle concentrations in PDMS, followed by incorporating CV using a simple swell−encapsulation−shrink process to fabricate white-light-activated antibacterial surfaces. The ZnO/CV composites possess superior antibacterial activities that are dependent upon the concentration of zinc in the polymer films. Among them, the 5.0ZnO−CV composite is proved to be one of the most efficacious antibacterial polymer surfaces developed to date, active in light conditions relevant to U.K. hospitals and exhibiting lethal photosensitization of S. aureus in just 45 min and of E. coli in just 90 min, with a minimum of a 4 log reduction in the numbers of both bacteria. Because release or transmission of dye and nanoparticle upon using this polymer coating is improbable, it has potential applications from healthcare settings (e.g., doors and hospital bed rails) to electronic devices (e.g., tablet and mobile phone covers) to help reduce bacterial surface contamination and thus reduce the risk of the spread of infection.

MATERIALS AND METHODS
4.1. Chemicals and Substrates. All chemicals used in this study were purchased from Sigma-Aldrich Chemical Co. except for acetone (VWR, U.K.). The PDMS prepolymer and curing agent (SYLGARD 184) were purchased from Dow Corning Corporation.
4.2. Materials Synthesis. 4.2.1. Preparation of ZnO NPs/ PDMS Composites. Polymer/nanoparticle composites with various ZnO contents (0, 0.5, 1, and 5 wt %) were produced. The appropriate amount of ZnO NPs was dispersed in PDMS (monomer/crosslink = 10:1 in mass), and the mixture was homogenized by hand-stirring. Afterward, the slurry was degassed in a desiccator for 15 min, followed by casting on glass substrates and they were cured in a preheated oven (100°C ) for 45 min. After the cooling, the polymer sheets were peeled off from the substrates and were cut into smaller pieces (squares 2.0 cm × 2.0 cm).

Preparation of Polymers with Embedded Crystal
Violet. Bare PDMS and ZnO-incorporated samples were coated with CV (by placing 2 cm × 2 cm squares into 750 ppm solutions of CV in acetone for 24 h) before removal, washed, and air-dried (24 h).
4.3. Material Characterization. Transmission electron microscopy (TEM) images were taken using JEOL JEM 1200EX with a 4 megapixel Gatan Orius SC200 charge-coupled device camera at an acceleration voltage of 120 kV.
The infrared absorbance spectra of the modified polymer samples were recorded within the range of 1000−4000 cm −1 with an accumulation of 16 scans per sample using a Bruker Platinum ATR.
The photoluminescence (PL) spectra of the treated samples were recorded using a Fluoromax 4.0 Jobin Yvon Horiba spectrofluorometer.
UV−vis absorption spectra of the ZnO NP aqueous suspension and CV-coated polymer samples were measured using a PerkinElmer Lambda 950 UV−vis spectrometer over a range of 200−800 nm (full range not shown).
A single colony of either E. coli or S. aureus was cultured in BHI broth (10 mL) for 17 h, shaking at 37°C. The bacterial pellet was harvested by centrifugation (20°C, 4000 rpm, 5 min) and washed in PBS (10 mL). After centrifugation again under the same conditions, the bacteria were resuspended in PBS (10 mL). The washed bacterial suspension was then

ACS Omega
Article diluted 1 in 1000 in PBS to give an inoculum of ∼10 6 CFU mL −1 .
Duplicates of each polymer sample were inoculated with 25 μL of the inoculum and covered with a sterile cover slip (1.8 cm × 1.8 cm). The samples were then irradiated for the required time period using a white light source that emits an average light intensity of 6500 ± 300 lux at a distance of 25 cm from the samples. A further set of samples (in duplicate) was maintained under dark conditions for the duration of the irradiation time.
Post irradiation, the inoculated samples and cover slips were transferred to a tube containing PBS (5 mL) and vortexed (60 s). The neat suspension and 10-fold serial dilutions were plated on the appropriate agar. The plates were incubated aerobically for 24 h (E. coli) or for 48 h (S. aureus) at 37°C, and the were colonies enumerated to determine the number of viable bacteria. The bacterial numbers in the inoculum were also determined in each experiment by viable colony counting. Each experiment contained two technical replicates, and the experiment was reproduced three times. The Mann−Whitney U test was used to analyze the statistical significance of the data.

Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS E.O. would like to thank Dr. William Peveler for his help with PL.