**Influence of Polysaccharides' Molecular Structure on the Antibacterial Activity and Cytotoxicity of Green Synthesized Composites Based on Silver Nanoparticles and Carboxymethyl-Cellulose**

**María de los Ángeles Martínez-Rodríguez 1, Elizabeth Madla-Cruz 2, Victor H. Urrutia-Baca 3, Myriam A. de la Garza-Ramos 4, Virgilio A. González-González 1,5 and Marco A. Garza-Navarro 1,5,\***


Received: 13 May 2020; Accepted: 10 June 2020; Published: 14 June 2020

**Abstract:** In this paper we report on the influence of polysaccharides' molecular structure on the antibacterial activity and cytotoxicity of composites based on silver nanoparticles (AgNPs) immobilized into carboxymethyl-cellulose (CMC). These composites were green synthesized from the reduction of silver ions into aqueous solutions of the polysaccharide, using CMC with different degree of substitution (DS) and molecular weight (Mw). The composites were characterized by transmission electron microscopy (TEM), as well as infrared (ATR-FTIR), ultraviolet (UV-Vis), Raman, and X-ray photo-electron (XPS) spectroscopic techniques. The antibacterial activity was evaluated with minimum inhibitory concentration against *Enterococcus faecalis*. The cytotoxicity of composites was assessed against human gingival fibroblast. Experimental evidence suggests that particle size distribution and morphology of AgNPs change according to the quantity of silver precursor added to the reaction, as well as the DS and Mw of CMC used for composites preparation. This is related to the dispersion of silver precursor into aqueous solutions of the polysaccharide and the formation of Ag-O coordination bonds among AgNPs and COO− moieties of CMC. Moreover, these coordination bonds modify the ability of nanoparticles to produce and release Ag<sup>+</sup> into aqueous dispersion, adjusting their antibacterial activity and the induction of cytotoxicity into the tested biological environments.

**Keywords:** silver nanoparticles; carboxymethyl-cellulose; composite; antibacterial activity; cytotoxicity

#### **1. Introduction**

In recent years, nanotechnology has impacted the development of new functional materials based on nanostructures. Among the so-called nanomaterials, silver nanoparticles (AgNPs) have emerged as a promising specie to be used in biomedical and food packaging applications as a bactericide, fungicide, and antiviral [1–4]. There can be found studies regarding the effect of AgNPs on gram-negative bacteria, which indicate that nanoparticles with size between one and ten nanometers adhere to the surface of the bacteria cell membrane and disturb its permeability and respiration [5]. In addition, AgNPs dispersed into aqueous media can release Ag<sup>+</sup> ions that can be internalized by passive bacterial transport through the channels in the cell membrane of both Gram-negative and positive bacteria [6]. The Ag<sup>+</sup> ions inflict further damage to the bacteria due to its interaction with sulfur- and phosphorousgroups at the DNA, causing a loss in its ability to replicate; along with the deactivation of bacteria proteins, because of the interaction of Ag<sup>+</sup> ions with their thiol (R-SH) groups [7].

Nonetheless, the main concern about the use of AgNPs in biomedical and food packaging applications is their toxicity. It has been reported that AgNPs display a size-dependent cytotoxicity, related to the generation of reactive oxygen species (ROS) during their surface oxidation and subsequent release of Ag<sup>+</sup> ions into biological environments [8,9]. So, it is necessary to search for low-toxic AgNPs from methodologies that do not use nor produce toxic species. Accordingly, the "green" chemistry implies the design, development, and application of chemical products and process to reduce or eliminate the use or generation of hazardous substances to human health and the environment [10]. As has been reported in literature, the green chemistry routes for synthesis of AgNPs consider biopolymers such as chitosan, poly(lactic acid), sodium alginate, cellulose, and carboxymethyl-cellulose as both reducing and capping agents [11–16].

Among these biopolymers, the carboxymethyl-cellulose (CMC) emerges as a promising reducing and immobilization media for the green synthesis of AgNPs, due to its good chemical stability, as well as its biocompatible and biodegradable characteristics. The CMC is a semi-synthetic polysaccharide derived from the natural polymer cellulose, which undergoes the partial substitution of cellulose native hydroxymethyl (RCH2OH) groups by carboxymethyl (RCOOH) groups [17]. The degree of substitution (DS) of RCH2OH by RCOOH is reported as an average of carboxymethyl groups per monomer unit. The CMC is usually commercialized as a water-soluble sodium salt, which in aqueous solution can be loaded with metallic ions as Ag<sup>+</sup> by a simple displacement reaction of Na<sup>+</sup> [18]. Moreover, due to the abundant hydroxyl groups on its molecular structure, CMC has been successfully used as a reducing agent for the preparation of CMC-AgNPs composites [16,19]. From this approach is possible to get an outstanding particle size control and good efficiency over the silver ions reduction, without the use or generation of hazardous substances.

We previously reported on the ability of green synthesized CMC-AgNPs composites to inhibit the proliferation of Gram-positive and negative bacteria, such as *Streptococcus mutans* and *Porphyromonas gingivalis*, respectively, with a suitable cytotoxicity [20]. Nonetheless, currently, experimental evidence regarding the role of molecular structure of polysaccharides as CMC on the antibacterial activity and cytotoxicity of AgNPs-based composites does not exists. Consequently, in this work we report on the influence of polysaccharides' molecular structure on the antibacterial activity and cytotoxicity of CMC-AgNPs composites synthesized from a green chemistry route, by the use of CMC with different DS and molecular weight (Mw) as a reducing agent and immobilization media.

#### **2. Materials and Methods**

#### *2.1. Synthesis and Characterization of CMC-AgNPs Composite*

The CMC with DS = 0.7 and Mw = 90 kDa (0.7CMC), CMC with DS = 0.9 and Mw = 250 kDa (0.9CMC), CMC with DS = 1.2 and Mw = 250 kDa (1.2CMC) and silver nitrate (AgNO3) were purchased from Sigma-Aldrich Co., Edo. de México, México, and used as received without any further treatment for the synthesis of CMC-AgNPs composites. Deionized water was used for the preparation of all solutions for this investigation (Barnstead EASYpure II system with ρ = 13 MΩ-cm).

The synthesis of CMC-AgNPs composites was performed following a previously reported route, with some modifications [19]. Briefly, aqueous CMC and AgNO3 solutions were prepared at concentrations of 15 mg/mL and 0.24, 0.48, 0.94, or 1.26 mg/mL, respectively, using deionized water. Then, 20 mL of CMC was added into a round-bottom three-neck flask (reactor) and stirred for 10 min under room conditions. Later, 10 mL of AgNO3 solution was added to the reactor and the temperature was raised to 90 ◦C. The reaction was kept at this temperature for 24 h under reflux conditions. After 24 h, the resultant yellowish to reddish dispersions (depending on the concentration of AgNO3 solution added to the reaction) was poured into a previously cooled round-bottom flask, in order to rapidly lower its temperature towards room temperature. These dispersions were frozen and then lyophilized. This process was performed using aqueous solutions of 0.7CMC, 0.9CMC, or 1.2CMC at a constant concentration of 15 mg/mL; as well as AgNO3 solutions at the aforementioned concentrations of 0.24, 0.48, 0.94, or 1.26 mg/mL to obtain the composite samples 0.7/0.9/1.2Ag1, 0.7/0.9/1.2Ag2, 0.7/0.9/1.2Ag3, or 0.7/0.9/1.2Ag4, respectively. The Table 1 shows the CMC and AgNO3 weights that were added to the reactor for the synthesis of each sample. Finally, dried samples were weighted and dissolved in deionized water to prepare CMC-AgNPs composites' dispersions for their further characterization.


**Table 1.** Reagents used for the synthesis of each composite sample.

The crystalline and morphological features of CMC-AgNPs composites were examined by transmission electron microscopy (TEM) in a Field Emission Gun, FEI Titan G2 80-300 microscope, using electron microscopy techniques as bright field (BF) and Z-contrast (HAADF-STEM) imaging, as well as selected area electron diffraction (SAED). Particle size distribution of AgNPs was obtained from the measuring of at least 300 randomly selected particles in CMC-AgNPs samples using Graphic for Mac 3.1 software; and adjusting the experimental measuring data to the Gaussian statistic model in OriginPro 8.5.0 software, using tools as a descriptive statistic (frequency counts) and analysis (fitting). Ultraviolet-visible spectroscopy (UV-vis) studies of CMC-AgNPs composites as well as AgNO3 precursor solution were performed in a Perkin-Elmer, Lambda 35, spectrometer to evaluate the reduction efficiency of the proposed synthesis route. Interactions between CMC molecules and AgNPs were examined using infrared spectroscopy (ATR-FTIR). ATR-FTIR spectra of pure 0.7CMC, 0.9CMC, 1.2CMC, as well as CMC-AgNPs samples were recorded in a Frontier MIR FT-IR, Universal ATR spectrometer. In addition, Raman spectroscopy was carried out in a Thermo Scientific, DXR Raman microscope. The spectra of selected composite samples were measured after 30 s of exposure and acquisition time of 60 s, using a radiation of 532 nm. Finally, X-Ray Photoelectron Spectroscopy (XPS) was perform for the measuring of C1s, O1s, and Ag3d spectra for pure 0.7CMC, 0.9CMC, and 1.2CMC, as well as for selected CMC-AgNPs samples in a Thermo-Scientific, K-Alpha spectrometer with monochromatized AlKα radiation (E = 1.5 keV), X-ray spot of 400 μm, and flood gun for charge compensation.

#### *2.2. Antibacterial Assay*

The antibacterial activity of CMC-AgNPs composites was examined using the standard broth dilution method. The minimal inhibitory concentration (MIC) was determined from 96-well flat-bottom plates containing 50 μL of CMC-AgNPs dilutions with concentrations [AgNPs] = 60 to 3.75 μg/mL in Brain Heart Infusion (BHI) medium (Becton Dickinson Bioxon, Edo. de México, México); and 50 μL of 1.0 <sup>×</sup> <sup>10</sup><sup>8</sup> CFU/mL of *Enterococcus faecalis* (ATCC® <sup>29212</sup>™) (*E. faecalis*), up to a final volume of 100 μL per well. In addition, ampicillin at 5 μg/mL was used as a positive control for inhibition of bacterial growth; whereas BHI medium was employed as negative control. The prepared cultures were incubated at 37 ◦C for 24 h in an aerobic atmosphere. Bacterial growth was measured from the absorbance of the cultures at 595 nm using an iMark™ microplate reader (Bio-Rad laboratories, Hercules, CA, USA). Subsequently, the percentage of growth inhibition was calculated using:

$$\% \text{inhibition} = 100 - \left\{ \left| \frac{(Sample - Positive \text{ control})}{(Negative \text{ control} - Positive \text{ control})} \right| \times 100 \right\} \tag{1}$$

The MIC value was defined as the lowest concentration of CMC-AgNPs that inhibited 99% of growth bacterial.

#### *2.3. Cytotoxicity Assay*

The cytotoxicity of CMC-AgNPs composites was evaluated against human gingival fibroblast cells (ATCC®PCS-201-018™), from 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl tetrazolium (MTT) assay. The cell line was cultured in Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% FBS, 1X antibiotic-antimycotic, and 6 mM L-glutamine (complete-DMEM) at 37 ◦C for 48 h in a humidified atmosphere of 5% CO2. Later, 100 <sup>μ</sup>L of complete-DMEM containing 5 <sup>×</sup> 104 cells were placed into each well of a flat-bottom 96-well plate and grown to approximately 90% confluence. Then, 100 μL of CMC-AgNPs dilutions at [AgNPs] = 60 to 3.75 μg/mL were added to each well and incubated for 24 h. Complete-DMEM and 2% Triton X-100 were used as negative and positive control, respectively. After incubation, supernatant was discarded, and the cells were carefully washed with PBS. Later, 100 μL of MTT diluted in complete-DMEM at 0.5 mg/mL were added to the wells and the cultures were incubated for 4 h. Subsequently, the supernatant was discarded, and the resulting formazan crystals were solubilized with 200 μL dimethyl sulfoxide. Finally, the absorbance of the cultures was recorded at 570 nm using a microplate reader. The percentage of cytotoxicity was calculated using:

$$\% \text{ cytotoxicity} = 100 - \left\{ \left[ \frac{(Sample - Positive\ control)}{(Negative\ control - Positive\ control)} \right] \times 100 \right\} \tag{2}$$

#### **3. Results and Discussion**

#### *3.1. Morphological and Crystalline Features of CMC-AgNPs Composites*

Figure 1 shows HAADF-STEM images taken from CMC-AgNPs composites that were synthesized using 0.7CMC as a reducing and immobilization agent. As Figure 1a displays, nanoparticles in sample 0.7Ag1 show a quasi-spherical morphology. There is also presented the adjustment of experimental data from particle size measuring to the Gaussian statistic model. The center of particle size distribution is 13.1 nm, showing a standard deviation of 5.3 nm. Nonetheless, the statistical distributions obtained for samples 0.7Ag2 and 0.7Ag3 depict centers at a larger particle size of 26 and 24.6 nm, respectively; as well as wider dispersions, with standard deviations of 17.7 and 17.0 nm, respectively (see Figure 1b,c). In addition, the nanoparticles in samples 0.7Ag2 and 0.7Ag3 show a change on their morphology from quasi-spherical to one that displays facets. The presence of faceted nanoparticles is also observed in sample 0.7Ag4, along with a large population of small nanoparticles with a mean size of 10.6 nm and standard deviation of 5.1 nm (see Figure 1d). The change in particle size distribution and morphology

of nanoparticles could be related to a decrease in the ability of CMC to control their growth as the weight content of AgNO3 is increased (see Table 1).

**Figure 1.** HAADF-STEM images obtained from composites samples: (**a**) 0.7Ag1; (**b**) 0.7Ag2; (**c**) 0.7Ag3; and (**d**) 0.7Ag4. The particle size distribution of each sample is shown just below its HAADF-STEM image.

The Figure 2 displays the morphological features of CMC-AgNPs samples that were obtained from 0.9CMC aqueous solutions. Herein we observed that mean particle size shows a small decrease from 32.6 to 28.3 nm as the weight content of AgNO3 used for the synthesis of 0.9Ag1, 0.9Ag2, and 0.9Ag3 increases (see Table 1); as well as an increase in the standard deviation from 8.9 to 16.4 nm (see Figure 2a–c). Moreover, the nanoparticles in sample 0.9Ag4 display a remarkable increase in both mean particle size (44.1 nm) and standard deviation (28.8 nm) (see Figure 2d). The formation of aggregates from faceted nanoparticles in 0.9Ag4 is also noticeable. According to these results, the 0.7CMC reagent gives smaller mean particle size but larger standard deviation than 0.9CMC reagent when 2.41, 4.82, or 9.45 mg of AgNO3 are used for the synthesis of samples.

Figure 3 shows the morphological characteristics for samples prepared with 1.2CMC reagent. For this case, the mean particle size tends to increase from 11 to 22.3 nm for samples 1.2Ag1, 1.2Ag2, and 1.2Ag3 (see Figure 3a–c) as the weight content of AgNO3 increases (see Table 1). Nonetheless, this trend is not followed by 1.2Ag4, since it displays a mean particle size of 19.1 nm (see Figure 3d). The presence of nanoparticle aggregates that resemble those observed in the 0.9Ag4 sample is also seen in Figure 3d. Finally, the standard deviation for this experimental set varies in a direct proportion with the weight of AgNO3 added for the synthesis of samples 1.2Ag1, 1.2Ag2, 1.2Ag3, and 1.2Ag4; and covers an interval from 4 to 14.8 nm. This evidence suggests that 1.2CMC reagent provides better control on particle size distribution than 0.7CMC and 0.9CMC reagents at the CMC/AgNO3 weight ratios used for sample preparation (see Table 1). Table 2 reports the data from particle size distribution that were obtained for the synthesized composite samples.

Figure 4 resumes the crystalline features that were observed for nanoparticles prepared from aqueous solutions of 0.7CMC, 0.9CMC, and 1.2CMC. As Figure 4a,c,e display, the nanoparticles from samples 0.7Ag1, 0.9Ag1, and 1.2Ag1, respectively, depict a regular atomic arrangement, showing lattice fringes with a regular interplanar spacing of 2.4 Å. This spacing is congruent with that reported for family planes {111} of the crystalline structure of silver (see JCPDS: 04-0783). Furthermore, in the SAED patterns reported in Figure 4b,d,f we recognize diffraction rings related to family planes {111}, {200}, {220}, and {311} of the face-centered cubic (FCC) packing of silver (see JCPDS: 04-0783). This evidence confirms the formation of AgNPs in the synthesized samples.

However, in order to obtain a first approach regarding the reduction efficiency of Ag<sup>+</sup> from our synthesis route, we record UV-vis spectra for CMC-AgNPs composites and AgNO3 solution used for their synthesis. As it can be observed in Figure 5, the spectra obtained from CMC-AgNPs composites do not show the absorption band related to the Ag<sup>+</sup> at 301 nm (see Figure 5a); instead, they display a well-defined band around 425–429 nm (see Figure 5b–d). According to the literature, this band is related to the characteristic surface plasmon resonance of AgNPs [21]. This result indicates that there are no detectable traces related to Ag<sup>+</sup> ions in the analyzed samples, suggesting full reduction of added Ag<sup>+</sup> into Ag0. Accordingly, the CMC/AgNPs weight ratio could be calculated as 200, 100, 50, and 37.5 for the composites 0.7/0.9/1.2Ag1, 0.7/0.9/1.2Ag2, 0.7/0.9/1.2Ag3, and 0.7/0.9/1.2Ag4, respectively (see Table 1). Nonetheless, to get more information about the immobilization features of the different kind of CMC, it is necessary to evaluate the manner that the polysaccharide's chains interacts with the synthesized AgNPs. Consequently, we proceed to perform ATR-FTIR measures of CMC reagents and CMC-AgNPs samples.

**Figure 2.** HAADF-STEM images obtained from composites samples: (**a**) 0.9Ag1; (**b**) 0.9Ag2; (**c**) 0.9Ag3; and (**d**) 0.9Ag4. The particle size distribution of each sample is shown just below its HAADF-STEM image.

**Figure 3.** HAADF-STEM images obtained from composites samples: (**a**) 1.2Ag1; (**b**) 1.2Ag2; (**c**) 1.2Ag3; and (**d**) 1.2Ag4. The particle size distribution of each sample is shown just below its HAADF-STEM image.


**Table 2.** Data from particle size distribution obtained for composite samples.

**Figure 4.** BF images and SAED patterns obtained from samples: (**a**) and (**b**) 0.7Ag1; (**c**) and (**d**) 0.9Ag1; (**e**) and (**f**) 1.2Ag1.

**Figure 5.** UV-vis spectra measured from (**a**) AgNO3; (**b**) 0.7Ag4; (**c**) 0.9Ag4; and (**d**) 1.2Ag4.

#### *3.2. Spectroscopic Characterization*

Figure 6 shows the ATR-FTIR spectra obtained from powdered 0.7CMC, 0.9CMC, and 1.2CMC reagents, as well as those recorded from powdered CMC-AgNPs samples. Figure 6a displays the spectrum obtained for 0.7CMC, where it can recognize absorption bands related to [18,22,23]: symmetrical and asymmetrical stretching at O-H bond of hydroxyl groups (R-OH) at 3360 cm−1; asymmetrical stretching at the C-H bond of the hydroxymethyl functional groups (R-CH2OH) at 2911 cm−1; asymmetrical and symmetrical stretching of -O-C = O bonds on the carboxymethyl functional groups (R-CH2OCOO<sup>−</sup>) at 1590 cm−<sup>1</sup> and 1413 cm<sup>−</sup>1, respectively; bending of -C-CH and O-CH- bonds on the R-CH2OCOO<sup>−</sup> groups at 1321 cm<sup>−</sup>1; stretching of C-O bond on R-CH2OCOO<sup>−</sup> at 1269 and 1026 cm<sup>−</sup>1; and stretching of C-O-C bonds on R-CH2OCOO<sup>−</sup> at 1099 and 1043 cm<sup>−</sup>1. Figure 6a also shows the spectra recorded from samples 0.7Ag1, 0.7Ag2, 0.7Ag3, and 0.7Ag4. Herein we noticed a slight bathochromic shift in the position of the band related to asymmetrical stretching of O-C = O moieties at R-CH2OCOO−, from 1590 to 1586 cm−1; along with a hypsochromic one from 1043 to 1053 cm−<sup>1</sup> of the band associated to stretching on C-O-C bonds at the same functional groups.

**Figure 6.** ATR-FTIR spectra obtained from (**a**) pure 0.7CMC and its composite samples; (**b**) pure 0.9CMC and its composite samples; (**c**) pure 1.2CMC and its composite samples.

In addition, Figure 6b depicts the spectrum obtained from powdered 0.9CMC, as well as those from composites 0.9Ag1, 0.9Ag2, 0.9Ag3, and 0.9Ag4. Likewise, bands related to vibrational modes of 0.9CMC molecules can be noticed, as the asymmetrical stretching of -O-C = O moieties at 1590 cm−<sup>1</sup> and stretching of C-O-C bonds at 1044 cm−1, which display bathochromic and hypsochromic shifts towards 1587 cm−<sup>1</sup> and 1050 cm<sup>−</sup>1, respectively, in the composites spectra. Moreover, this phenomenon also occurs for the samples 1.2Ag1, 1.2Ag2, 1.2Ag3, and 1.2 Ag4, since they display a noticeable shift on the band related to the stretching of C-O-C, from 1049 to 1055 cm−1, with respect the position of this band in the spectrum recorded for 1.2CMC (see Figure 6c). These features suggest an interaction between CMC molecules and AgNPs for all the samples, which could be attributed to the adsorption of R-CH2OCOO<sup>−</sup> onto nanoparticles, as it has been reported elsewhere [24–26].

In order to corroborate the adsorption of CMC chains onto AgNPs, we record Raman spectra from selected powdered samples. Figure 7 shows the Raman spectra obtained from 0.7Ag1, 0.7 Ag4, 1.2Ag1, and 1.2Ag2. Here it is possible to identify bands associated with vibrational modes of CMC, such as stretching of C-H at 2916–2909 cm<sup>−</sup>1; as well as asymmetrical and symmetrical stretching of O-C = O at 1588–1577 cm−<sup>1</sup> and 1384–1376 cm−1, respectively [25–28]. An increase in the intensity of Raman scattering in the bands attributed to stretching vibration in O-C = O can also be noticed, which seems to be related with the increase in the weight content of AgNPs in samples. The increase in intensity of both bands was evaluated taking as reference the intensity of the band attributed to C-H stretching vibration. As it has been documented in literature, the increase in the intensity of these bands can be related to the adsorption of COO− moieties onto metal or semimetal nanoparticles; and occurs due to electric field induced surface enhanced Raman scattering (SERS) [29]. Moreover, there can be found a band at 234–228 cm−1, attributed to the stretching vibration of Ag-O bond [30,31]. These results confirm the adsorption of CMC chains onto AgNPs and suggest the formation of a bond between Ag and O in the COO− moieties of CMC, as it has been proposed elsewhere [30].

**Figure 7.** Raman spectra recorded from: (**a**) 0.7Ag1; (**b**) 0.7Ag4; (**c**) 1.2Ag1; and (**d**) 1.2Ag2.

In order to get further insight regarding the adsorption of RCH2OCOO<sup>−</sup> onto AgNPs surface, we proceed to measure C1s, O1s, and Ag3d XPS spectra from 0.7CMC, 0.9CMC, 1.2CMC reagents, as well as from some powdered samples. Peaks of the recorded XPS spectra were deconvoluted and fitted using a Gaussian approach in PeakFitV4.12 software. Accordingly, Figure 8a shows C1s and O1s spectra recorded from 0.7CMC. The C1s spectrum exhibits four peaks at 285.2, 287.0, 288.6, and 290.1 eV, which can be attributed to C in C-C, C-O, C = O, and O-C = O, respectively. Three peaks are observed in O1s spectrum that can be attributed to C = O, O-C = O and Auger electrons from Na at 531.5, 533.4, and 535.7 eV, respectively. Likewise, C1s spectra recorded from 0.9CMC and 1.2CMC were deconvoluted into four peaks at 285.0 eV (C-C), 286.7 eV (C-O), 288.2 eV (C = O), and 289.6 or

288.8 eV (O-C = O); whereas their O1s spectra show peaks at 531.4 eV (C = O), 533.1 eV (O-C = O), and 535.6 eV (Auger-Na) (see Figure 8b,c). The presence of these signals agrees with those expected from the molecular structure of these polysaccharides [32,33].

Figure 9 show C1s, O1s, and Ag3d spectra from samples 0.7Ag1, 0.9Ag1, and 1.2Ag1. C1s spectra of samples show peaks related to C-C, C-O, C = O, and O-C = O of the polysaccharide's chains, although they display changes in their binding energies with respect to that obtained from CMC reagents (see Figure 8); along with a change in the intensity of each peak (quantity of photoelectrons emitted from samples). Changes in binding energy occur into an interval from 0.2 to 0.9 eV, and are more obvious for emissions associated with C = O and O-C = O. In addition, O1s spectra show shifts in the binding energies related to emissions from C = O and O-C = O, into an interval between 0.4 and 0.9 eV. There is also observed Ag3d spectra of these samples that display peaks at 374.2–374.7 eV and 368.0–368.7 eV, which correspond to photoelectrons emitted from 3d3/<sup>2</sup> and 3d5/<sup>2</sup> states, respectively. The difference between the binding energies of such emissions is 6 eV for all cases, confirming that silver in samples is only Ag0 [34–36]. Moreover, this experimental evidence is congruent with that obtained from UV-vis spectra regarding the full reduction of Ag<sup>+</sup> into Ag0 (see Figure 5).

Likewise, Figure 10 shows the XPS spectra recorded from samples 0.7Ag2, 0.9Ag2, and 1.2Ag2. Herein, C1s and O1s spectra depict shifts in the binding energies related to C and O in C = O and O-C = O bonds that reach up 0.9 eV, with respect to those observed in Figure 8; as well as peaks around 374 and 368 eV in their Ag3d spectra, related to core emissions from 3d3/<sup>2</sup> and 3d5/2, respectively. Thus, considering that the energy of photoelectrons emitted from discrete states as 1s is quite susceptible to change depending on the bonds that elements form, the fact that peaks related to C-C, C-O, O-C = O, and C = O display shifts on their binding energies indicates that AgNPs are immobilized in CMC by coordination bonds [36,37]. Moreover, taking into account the results obtained from Raman spectra, the coordination bonds can be attributed to those Ag-O among AgNPs and COO− moieties of CMC.

**Figure 8.** C1s and O1s spectra recorded by XPS from: (**a**) 0.7CMC; (**b**) 0.9CMC; and (**c**) 1.2CMC.

**Figure 9.** C1s and O1s spectra recorded by XPS from: (**a**) 0.7Ag1; (**b**) 0.9Ag1; and (**c**) 1.2Ag1.

**Figure 10.** C1s, O1s and Ag3d spectra recorded by XPS from samples: (**a**) 0.7Ag2; (**b**) 0.9Ag2; and (**c**) 1.2Ag2.

Considering our experimental findings, the variation of the morphological features of AgNPs can be explained as follows. The CMC is capable of attracting Ag<sup>+</sup> ions to intermolecular sites nearby the negatively charged R-CH2OCOO<sup>−</sup> when both are diluted in aqueous media [18]. In these sites, the Ag<sup>+</sup> ions are reduced with the electrons realized from R–OH or R–CH2OH groups of CMC at high temperature (i.e., 90 ◦C). Accordingly, the coalescence of Ag<sup>0</sup> conduces to the nucleation and subsequent growth of AgNPs, which will depend on the quantity of silver reagent added to the solution [19]. Hence, it is possible to state that when a CMC with a given DS and Mw is used for the synthesis of AgNPs, their particle size could increase as the weight content of AgNO3 added to the reaction increases. This could explain the particle size distributions (mean size and standard deviation) obtained from almost the samples prepared from 0.9CMC and 1.2CMC aqueous solutions (see Figures 2 and 3). In addition, we observe that samples prepared with the same weight content of AgNO3 but different CMC reagent display variations in their particle size distribution and morphology. According to our experimental evidence, the AgNPs are susceptible to form coordination bonds with COO− moieties. Therefore, the particle size distribution and morphology of AgNPs will depend on the quantity of R-CH2OCOO<sup>−</sup> available for their immobilization. The quantity of R-CH2OCOO<sup>−</sup> available to restrict the size of AgNPs and avoid their secondary growth varies according to the DS of 0.9CMC and 1.2CMC (see Table 2).

This explanation seems to disagree with the morphological features of samples obtained from 0.7CMC aqueous solutions, since their particle size distributions (mean size and standard deviation) do not change in direct proportion with the weight of AgNO3 added to each reaction; and they display smaller mean particle size than samples prepared with 0.9CMC reagent at the same weight content of AgNO3 (see Table 2). In order to explain this phenomenon, we should consider the following. As Table 1 shows, all CMC aqueous solutions were prepared at the same concentration for the synthesis of samples. However, the 0.7CMC reagent has a lower Mw than 0.9CMC and 1.2CMC. It is well known that viscosity of a polymer solution varies in a direct proportion with its Mw at a given concentration. So, it is reasonable to think that, at the same weight content of AgNO3, the dispersion of Ag<sup>+</sup> ions in 0.7CMC aqueous medium differs from that in 0.9CMC or 1.2CMC solutions. Moreover, considering the low DS of 0.7CMC, it is possible that some Ag<sup>+</sup> ions do not reach a site nearby the R-CH2OCOO<sup>−</sup> groups; hence, they could be reduced elsewhere. This implies that the coalescence of Ag0, nucleation and subsequent growth of AgNPs also occur far from the R-CH2OCOO<sup>−</sup> groups. This could explain the high standard deviation obtained from samples 0.7Ag2 and 0.7Ag3, as well as the formation of large and faceted nanoparticles in sample 0.7Ag4 (see Figure 1). Nonetheless, the AgNPs are immobilized in CMC by coordination bonds, thus, it is possible to argue that these Ag-O bonds among AgNPs and COO− moieties restrict the growth for a large number of nanoparticles when 0.7CMC reagent is used for preparation of composite samples (see Figure 1). This could explain the small mean particle size obtained at low weight content of AgNO3 added for the preparation of 0.7Ag1 (see Table 2).

Therefore, it can be concluded that the key factor for the control of particle size distribution of synthesized AgNPs is the quantity of R-CH2OCOO<sup>−</sup> available for their immobilization. The quantity of these groups varies in direct proportion with the DS of CMC, which in general gives smaller particle sizes for CMC with higher DS. It is worth mentioning that the observed trends regarding the changes on standard deviation and morphology of AgNPs, as well as their plausible explanations, should be confirmed in further studies.

#### *3.3. Antibacterial Activity*

As we explain in Section 3.2, particle size distribution and morphology of the AgNPs mainly vary according to the DS of CMC used for their synthesis. This is related to the fact that nanoparticles are immobilized in CMC by the formation of Ag-O coordination bonds among AgNPs and COO− moieties of the polysaccharide's chains. Thus, in order to address the effect of these bonds on the antibacterial activity of AgNPs, we proceed to test samples with similar particle size distributions but with nanoparticles immobilized into CMC with distinct DS, 0.7Ag1 and 1.2Ag1; along with the sample

0.9Ag4, which displays a different particle size distribution for nanoparticles immobilized into CMC with close DS to 1.2CMC. The antibacterial activity assays were performed by three replicates of three independent experiments using doses with a known concentration of AgNPs, [AgNPs]. As Figure 11 shows, an important antibacterial activity of CMC-AgNPs composites was observed at [AgNPs] = 60 μg/mL for the three tested samples. At this dose no statistical difference was observed compared to 5 μg/mL ampicillin (*p* > 0.05) for 0.9Ag4 and 1.2Ag1 samples. Therefore, [AgNPs] = 60 μg/mL can be established as the MIC value for our experimental setup, except for 0.7Ag1 sample. The sample 0.7Ag1 shows an inhibitory effect of 85.5 ± 2.3% at this dose. In addition, a residual inhibitory effect was observed at [AgNPs] = 30 μg/mL in 0.7Ag1, 0.9Ag4, and 1.2Ag1 with 26.3 ± 2.9%, 36.9 ± 5.7%, and 41.4 ± 8.3%, respectively (see Figure 11).

These results can be explained as follows. The antibacterial activity of AgNPs is believed to be related to the production and release of positive charged Ag ions from their surface in aqueous media [6]. Thus, smaller particle size leads to a large surface area, hence, to produce a higher amount of Ag<sup>+</sup> in aqueous solution than larger AgNPs [38]. The Ag<sup>+</sup> ions can be internalized by passive bacterial transport through the channels in the cell membrane bacteria and inflict damage due to their binding to cellular structural elements such as enzymes and proteins, particularly to their R-SH groups [6,7]. This binding diminishes the membrane permeability and leads to cell death [39]. Specifically, MIC value against gram-positive *E. faecalis* bacteria has been found in an interval between 500 and 0.19 μg/mL, depending on synthesis route, particle size, and surface modification [39–42]. Accordingly, the Ag-O coordination bonds among AgNPs and COO− moieties of CMC do not blur the ability of nanoparticles to produce and release Ag<sup>+</sup> from their surface, thus, to display a remarkable antibacterial activity in aqueous media.

Moreover, it can be noticed that AgNPs with quite different particle size distribution but immobilized in CMC with close DS, as is the case for 0.9Ag4 and 1.2Ag1, display almost the same inhibition of bacteria growth (see Figure 11). This suggests that Ag-O coordination bonds among AgNPs and COO− moieties of these polysaccharides enhance the antibacterial activity for CMC-AgNPs composites. This feature is congruent with the antibacterial activity observed for 0.7Ag1, since this sample presents a lower inhibitory effect than 1.2Ag1, even though both have a similar particle size distribution (see Table 2).

**Figure 11.** Antibacterial activity of CMC-AgNPs against *E. faecalis* growth. The data represent the percentage mean ± the percentage deviation.

#### *3.4. Cytotoxicity*

Considering the inhibitory effect on the bacteria growth of samples 0.7Ag1, 0.9Ag4, and 1.2Ag1, we proceed to evaluate their cytotoxicity at the same tested doses for the antibacterial assays. The

cytotoxicity assays were performed by three replicates of three independent experiments for the tested samples. As can be noticed in Figure 12, the highest cytotoxic effect occurs at [AgNPs] = 60 μg/mL for the three samples of CMC-AgNPs, showing a cytotoxicity greater than 95% after 24 h of treatment. However, the cytotoxic effect decreased in a dose dependent manner. At 30 μg/mL, 0.7Ag1 shows a lower cytotoxicity (60.5 ± 9.4%) than that from 0.9Ag4 and 1.2Ag1 of 100.3 ± 1.7% and 99.7 ± 5.9%, respectively. For subsequent dilutions, the decrease in cytotoxicity was more pronounced for 0.7Ag1 compared to the other two samples (*p* < 0.001). In addition, no cytotoxic effect was observed at a dilution of [AgNPs] = 3.75 μg/mL for the tested samples, since there are not statistical differences compared to the complete-DMEM control (*p* > 0.05). It is worth mentioning that at this dose no significant antibacterial activity was observed (see Figure 11).

In order to explain these results, we should consider the following. It is well known that AgNPs-mediated cytotoxicity in mammalian cells depends greatly on the nanoparticle size, shape, surface charge, dosage, oxidation state, and agglomeration condition as well as the cell type. Moreover, it has been demonstrated that antibacterial activity of AgNPs in aqueous media is related to oxidation of their surface and subsequent release of Ag<sup>+</sup> [43]. This oxidation conduces to the formation of reactive oxygen species (ROS) which trigger several negative effects on cell structures and their functions, inducing cytotoxicity [6,44]. Accordingly, the fact that 0.7Ag1 displays lower toxicity than 0.9Ag4 and 1.2Ag1 at all tested doses suggests that generation of ROS is diminished by the use of CMC with low DS for AgNPs immobilization. This is congruent with the results obtained from antibacterial activity assays regarding the inhibition of bacteria growth.

Hence, it can be concluded that Ag-O coordination bonds among AgNPs and COO− moieties of CMC modify the ability of nanoparticles to produce and release Ag<sup>+</sup> into aqueous dispersion, adjusting their antibacterial activity and the induction of cytotoxicity into the tested biological environments. Finally, Table 3 summarize the results obtained from this work.

**Figure 12.** Cytotoxicity obtained from CMC-AgNPs composites against ATCC®PCS-201-018™ cell line. The data represent the percentage mean ± the percentage deviation.


**Table 3.** Summary of the results obtained from this work.

#### **4. Conclusions**

The influence of polysaccharides' molecular structure on the antibacterial activity and cytotoxicity of green synthesized composites based on AgNPs immobilized into CMC was reported. The experimental evidence suggests that the particle size distribution and morphology of AgNPs mainly depend on the quantity of R-CH2OCOO<sup>−</sup> groups available for their immobilization. This is related to the fact that nanoparticles are immobilized in CMC by the formation of Ag-O coordination bonds among AgNPs and COO− moieties of the polysaccharide's chains. Accordingly, the quantity of R-CH2OCOO<sup>−</sup> groups varies in direct proportion with the DS of CMC, which in general gives smaller particle size for CMC with higher DS. Moreover, the biological assays indicate that the antibacterial activity and cytotoxicity of the tested samples increase by the use of CMC with higher DS as AgNPs immobilization medium. Hence, it can be concluded that Ag-O coordination bonds among AgNPs and COO<sup>−</sup> moieties of CMC modify the ability of nanoparticles to produce and release Ag<sup>+</sup> into aqueous dispersion, adjusting their antibacterial activity and the induction of cytotoxicity into the tested biological environments. Finally, it is worth mentioning that the width of the particle size distribution and morphology of AgNPs also depends on the weight of AgNO3 added to the reaction and the Mw of CMC used for their synthesis. This could be related to the manner that silver ions are dispersed into the used CMC aqueous solutions for their reduction and subsequent nucleation and growth of AgNPs. Nonetheless, the observed trends regarding the variation of standard deviations and morphology, as well as their plausible explanations, should be confirmed in further studies.

**Author Contributions:** Conceptualization, M.A.d.l.G.-R. and V.A.G.-G.; data curation, M.A.M.-R. and E.M.-C.; formal analysis, V.H.U.-B. and M.A.G.-N.; investigation, M.A.M.-R., E.M.-C., and V.H.U.-B.; methodology, M.A.d.l.G.-R., and V.A.G.-G.; supervision, M.A.G.-N.; writing—original draft, M.A.M.-R., E.M.-C., V.H.U.-B., V.A.G.-G., and M.A.G.-N.; writing—review and editing, M.A.d.l.G.-R. and M.A.G.-N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was founded by PAICYT-UANL, project number IT686-11. In addition, authors acknowledge to CHRISTUS—LATAM HUB Center of Excellence and Innovation, S.C. (CHRISTUS CEI) for financially supporting the study.

**Conflicts of Interest:** The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **ZnO Nanostructures with Antibacterial Properties Prepared by a Green Electrochemical-Thermal Approach**

**Maria Chiara Sportelli 1,2, Rosaria Anna Picca 1,2,\*, Margherita Izzi 1, Gerardo Palazzo 1,2, Roberto Gristina 3, Massimo Innocenti 4, Luisa Torsi 1,2 and Nicola Cio**ffi **1,2**


Received: 23 January 2020; Accepted: 29 February 2020; Published: 5 March 2020

**Abstract:** Zinc oxide (ZnO) nanostructures are widely applied materials, and are also capable of antimicrobial action. They can be obtained by several methods, which include physical and chemical approaches. Considering the recent rise of green and low-cost synthetic routes for nanomaterial development, electrochemical techniques represent a valid alternative to biogenic synthesis. Following a hybrid electrochemical-thermal method modified by our group, here we report on the aqueous electrosynthesis of ZnO nanomaterials based on the use of alternative stabilizers. We tested both benzyl-hexadecyl-dimetylammonium chloride (BAC) and poly-diallyl-(dimethylammonium) chloride (PDDA). Transmission electron microscopy images showed the formation of rod-like and flower-like structures with a variable aspect-ratio. The combination of UV–Vis, FTIR and XPS spectroscopies allowed for the univocal assessment of the material composition as a function of different thermal treatments. In fact, the latter guaranteed the complete conversion of the as-prepared colloidal materials into stoichiometric ZnO species without excessive morphological modification. The antimicrobial efficacy of both materials was tested against *Bacillus subtilis* as a Gram-positive model microorganism.

**Keywords:** zinc oxide; electrochemical synthesis; BAC; PDDA; nanorod; XPS; TEM; *B. subtilis*

#### **1. Introduction**

Semiconductor metal oxide nanostructures, and specifically ZnO ones, have received a great amount of interest in many areas due to their unique physical, chemical and optical properties [1]. Moreover, ZnO nanomaterials are well-known antimicrobial agents [2], which can be employed in several fields where pathogen spreading should be inhibited [3–5]. Besides the many physical [6–8] and chemical [9–11] methods for the production of such materials, scientific research has recently focused on green and simple electrochemical approaches [12,13]. These synthetic routes have the great advantage of being completely eco-friendly, using mainly aqueous solutions in the absence of reductants or other harmful chemicals [14]. Moreover, they can be easily scaled up for industrial purposes, and are cheap and easily reproducible [12,15]. Electrochemical approaches for the production of unsupported ZnO nanostructures (ZnONSs) are divided into three categories: electrochemical deposition under oxidative conditions (EDOC), electrolysis in alcohols, and electrolysis in the aqueous phase. EDOC [16] involves three stages: oxidation of a sacrificial anode, cathodic reduction of metal

ions, and oxidation of metal nanoclusters by means of the oxygen introduced into the cell, assisted by stabilization by surfactants. Electrolysis in alcohols consists in the electrochemical degradation of a Zn electrode at high potentials in an anhydrous environment [17,18]. This process was recently updated by Dhayagude et al. [19] using tetra-butyl-ammonium bromide (TBAB) as a stabilizer in a water/ethanol electrochemical medium. Regarding the synthesis in the aqueous alkali phase, a hybrid electrochemical-thermal method was proposed in 2010, which used an aqueous solution of NaHCO3 to generate mixed carbonate and hydroxide Zn species, which were subsequently converted into ZnO by a calcination process at temperatures >300 ◦C [15,20]. In our group, we modified this method, in order to obtain a higher level of morphological control in the produced nanostructures. Both anionic [14,21] and cationic [22] stabilizers were dispersed in the electrochemical media and tested, and we demonstrated that we were able to produce various morphologies, ranging from spheroidal particles to rice-grain or rod-like structures.

Here, we propose a further upgrade of the aforementioned electrochemical process, which is run under mild galvanostatic conditions, using either benzyl-dimetylammonium chloride (BAC) or poly-diallyl-(dimethylammonium) chloride (PDDA) as a stabilizer. The first is a well-known disinfectant, and was used with the intent of producing nanostructures with enhanced antimicrobial properties [23,24]. The second one, a high charge density cationic polymer, is already used for the preparation of transistors and sensors, and as a coadjuvant in layer-by-layer deposition processes [25]. The gel-like products of the electrochemical step underwent a two-step thermal treatment: overnight desiccation at 120 ◦C and calcination at 450 ◦C for 1 h. Both the dried and the calcined samples, prepared with the two stabilizers, were characterized by transmission electron microscopy (TEM), and UV–Vis, infrared (FTIR) and *X*-ray photoelectron (XPS) spectroscopies. The critical role of temperature on the physicochemical properties of the final nanostructures was demonstrated. The antimicrobial activity of both the BAC- and PDDA-stabilized ZnONSs was demonstrated against a model Gram-positive microorganism, i.e., *B. subtilis*.

#### **2. Materials and Methods**

#### *2.1. Materials*

Zinc sheets (1 mm thick, 99.99+%) were purchased from Goodfellow Ltd and cut into <sup>2</sup> <sup>×</sup> 1 cm2 pieces. Sodium bicarbonate (NaHCO3, purum p.a., 99.0%), 2-propanol (IPA, Chromasolv® Plus, for HPLC, 99.9%), hydrochloric acid (HCl, ACS reagent, 37%), benzyl-hexadecyl-dimetylammonium chloride (BAC, pure cationic surfactant), potassium bromide (KBr, FTIR grade, >99% trace metals basis), poly-diallyl-(dimethylammonium) chloride (PDDA, with an average molecular weight MW of 200,000–350,000, 20 wt.% in H2O), and Luria-Bertani (LB) broth (Miller, pH 6.8–7.2, 2.5% solution) were purchased from Merck-Sigma Aldrich (Milan, Italy). Aluminum oxide (Al2O3, purum p.a., 99.7%), for the mechanical polishing of zinc sheets, was obtained from Fluka Chemicals (Milan, Italy). Agar powder, meat extract, and peptone (from casein, pancreatic digest) were supplied by SIFIN Diagnostics Gmbh (Berlin, Germany). Milli-Q water was used in all experiments.

#### *2.2. Synthesis of ZnONSs*

A three-electrode setup was employed, using two zinc sheets as working (WE) and counter (CE) electrodes, and an Ag/AgCl (KCl sat.) as a reference electrode (RE). First, zinc electrodes were polished using sandpaper and afterward using alumina slurries with a different granulometry. Then, they underwent sonication, alternating between MilliQ water and IPA. Electrodes were finally activated in 1 M HCl for 30 s prior to use. The process yield (in terms of the mass of produced ZnONSs) was estimated by a differential weighting of the two zinc sheets before and after the process. The electrosynthesis was carried out in mild galvanostatic conditions, applying a current density of 10 mA/cm<sup>2</sup> with a CH-1140b potentiostat-galvanostat (CH Instruments, Bee Cave, TX, USA). The process was performed under continuous stirring for 1 h, at room temperature. The electrolytic

medium was composed of either 0.01 M BAC or 1 g/L PDDA, dissolved in 30 mM NaHCO3 aqueous solution. pH was monitored at the beginning and the end of the synthesis. Afterward, the colloidal dispersion was centrifuged at 6000 rpm for 30 min, and the resulting precipitate was dried overnight at 120 ◦C. The obtained powder underwent calcination at 450 ◦C in a tubular muffle furnace for 1 h.

#### *2.3. Morphological and Spectroscopic Characterizations*

TEM microscopy was performed with an FEI Tecnai 12 (Hillsboro, OR, USA) instrument (high tension: 120 kV; filament: LaB6), by dropping NP suspensions on a Formvar®-coated Cu grid (400 mesh, Agar Scientific, Stansted, UK).

UV–Vis spectroscopy was carried out with a Shimadzu UV-1601 double beam spectrometer, equipped with a silicon photodiode detector, from 200 to 700 nm, with 1-cm Quartz Suprasil® (Hellma Analytics, Jena, DE) cuvettes.

Infrared spectra were recorded on a Perkin Elmer Spectrum-Two (Milan, IT), in the spectral range of 4000–400 cm−1, with a resolution of 1 cm−1, averaging 8 consecutive scans. The samples were analyzed as KBr pellets by grinding a proper quantity of the ZnO nanopowder in an agate mortar.

XPS surface analysis was performed on a PHI (Chanhassen, MN, USA) Versaprobe II spectrometer. A monochromatized Al-Kα source (1486.6 eV) was used. Dual-beam charge neutralization was constantly applied during the analysis. Large-area XPS was performed, operating with a sampling area of <sup>200</sup> <sup>×</sup> <sup>1400</sup> <sup>μ</sup>m2. Samples were mounted onto the sample holder by means of double-sided tape. Survey scans (pass energy = 117.4 eV, step size = 1 eV) and high-resolution regions (pass energy = 58.7 eV, step size = 0.125 eV) relevant to C1s, O1s, N1s, Na1s, Cl2p, Zn2p, and ZnLMM were investigated. Detailed spectra processing was performed by CasaXPS® (v. 2.3.18PR1.0) software. Binding Energy (BE) referred to the aliphatic component of C1s at 284.8 eV. For estimating the proper peak positions and assignments in both IR and XPS analyses, a commercial ZnO powder, obtained from Sigma Aldrich (Milan, IT), was used.

Aqueous ZnONS suspensions of 0.5 g/L were also characterized by ζ-potential measurements, using a Zetasizer-Nano ZS from Malvern Instruments (Rome, IT). The cell holder was maintained at 25 ◦C by a Peltier element; Laser-Doppler electrophoresis (LDE) exploited forward scattering at 17◦. The LDE measurements were performed in a disposable capillary cell, and the ζ-potential was evaluated from the electrophoretic mobility according to the Smoluchowski approximation.

#### *2.4. Antimicrobial Activity of ZnONSs*

BAC- and PDDA-ZnONSs were suspended in Milli-Q water at a concentration of 0.15% w/v and deposited by drop casting on sterilized circular glass slides (- 12 mm, Agar Scientific). The *B. subtilis* isolated strain (MTCC 441) was provided by the Institute of Bioscience and Biology of the University of Bari. Fifty milliliters of fresh LB broth was inoculated with 1 mL of cell suspension and incubated at 30 ◦C. The bacterial culture at the beginning of the exponential growth phase, at a concentration of 10<sup>7</sup> CFU/mL (colony forming unit), was seeded (200 μL) on LB-agar Petri dishes. Then, circular glass slides covered with ZnONSs were put in contact with the seeded plates and incubated at 30 ◦C for 40 h. Experiments were performed in 5 replicates. The average values were used for calculation of the inhibition zone area. The minimum diameter of the inhibition zone was measured in mm.

A bare glass slide was used as a control sample.

#### **3. Results and Discussion**

#### *3.1. Electrochemical Production of ZnONSs*

This work follows the steps of the electrochemical-thermal hybrid method proposed by Chandrappa et al. [20], and improved by our group with the use of both cationic [22] and anionic [14,21] stabilizers. A colloidal suspension of hydroxides and zinc carbonates was produced in the presence of PDDA and BAC. As already shown [14,21,26], in fact, the final pH of the electrolytic medium is always

highly basic (>9), and species such as Zn(OH)2, [Zn(OH)3] <sup>−</sup> and [Zn(OH)4] 2- are thermodynamically stable in these conditions. The mixture is then converted into pure ZnO by thermal treatments. For each synthesis, the electrodes were weighed before and after the process, in order to calculate the mass variation and the experimental yield (Equation (1)).

$$
\Delta m\_{\text{experimental}} = \Delta m(\text{WE}) - \Delta m(\text{CE}) \tag{1}
$$

The theoretical mass was calculated according to Faraday's law for a two-electron process, as expressed in Equation (2), where *I* is current intensity, Δ*t* is process duration, *MZn* is Zn atomic mass, *Z* is the number of electrons involved in the process, and *F* is the Faraday constant:

$$m\_{theoretical} = \frac{M\_{Z,n} \left(\frac{\mathcal{X}}{mol}\right) \times \Delta t(s) \times I(A)}{Z(\#e^{-}) \times F \frac{\mathcal{C}}{mol}} \tag{2}$$

The obtained results, for both stabilizers, are listed in Table 1.

**Table 1.** The process efficiency for the preparation of nanocolloids, expressed in terms of nanodispersed Zn masses, and % yields.


The synthesized products were first subjected to a desiccation treatment in an oven at 120 ◦C, which was carried out overnight, and a further thermal treatment at 450 ◦C for 1 h.

#### *3.2. Morphological Characterization*

Morphological analyses on both dried (120 ◦C) and annealed (450 ◦C) ZnONSs exhibited the influence of both stabilizers and thermal treatments. Figure 1 shows the TEM micrographs obtained on the ZnO nanostructures that were prepared in the presence of BAC.

**Figure 1.** Transmission electron microscopy (TEM) images of electrosynthesized ZnO-based materials in the presence of benzyl-dimetylammonium chloride (BAC). Upper panels: ZnO dried at 120 ◦C overnight; lower panels: ZnO calcined at 450 ◦C for 1 h.

Rod-like and flower-like structures (with an average length of 1–2 μm) were obtained from samples dried at 120 ◦C. This morphology was preserved after calcination at 450 ◦C. However, a certain number of spheroidal nanoparticles (with an average diameter >5 nm) appeared after calcination, mainly at the wire tips. This phenomenon can be indicative of a possible mechanism of wire growth induced by calcination. In fact, the calcined samples showed slightly longer nanowires. The role of BAC in the growth mechanism of elongated rod-like structures may be tentatively explained by considering certain similarities between the effect of BAC and that exerted by cetyltrimethylammonium bromide (CTAB), which is an asymmetric quaternary ammonium salt, as well [27]. It has been reported in the literature that CTA<sup>+</sup> ions, which have a structure of a charged tetrahedron with a long hydrophobic tail, can electrostatically interact with [Zn(OH)4] <sup>2</sup><sup>−</sup> (anion present in solution with a tetrahedral geometry, known to be the growth unit for ZnO), forming ion pairs [28–30]. In this way, the lateral growth is inhibited, whereas it is promoted along the *c*-axis ([0 0 0 1] direction), thus favoring the hexagonal (wurtzite) rod-like morphology [28]. Both CTAB and BAC promote the growth of ZnO crystallites in the form of wurtzite, the most thermodynamically stable ZnO crystalline form. Moreover, heating can favor assembly into flower-like shapes [28].

Analogously, Figure 2 reports images relevant to ZnO synthesized in the presence of PDDA.

**Figure 2.** TEM images of electrosynthesized ZnO-based materials in the presence of poly-diallyl- (dimethylammonium) chloride (PDDA). Upper panels: ZnO dried at 120 ◦C overnight; lower panels: ZnO calcined at 450 ◦C for 1 h.

The latter promoted the formation of flower-like microstructures, with an average wire length of 200–500 nm. This morphology was also preserved after calcination, although the thermal deterioration of the polymer tended to produce more aggregates. In the case of PDDA, a growth mechanism involving lamellar superimposition can be hypothesized, as shown by the growth lines running along wires' length (Figure S1). Typically, the ZnO nanophases obtained by the proposed method show a crystalline, wurtzite structure, with a significant number of defects, as assessed by selected-area electron diffraction (SAED) analysis (Figure S2). Interestingly, other results indicate that calcination is not essential to obtain ZnO when using PDDA (vide infra).

#### *3.3. FTIR and UV–Vis Characterizations*

The FTIR spectra of all samples are presented in Figure 3.

**Figure 3.** The FTIR spectra of dried (120 ◦C) and calcined (450 ◦C) samples: (**a**) ZnO-BAC; (**b**) ZnO-PDDA. The main IR modes are highlighted by arrows.

In the case of BAC (Figure 3a), IR data indicate that ZnO is already present in the 120 ◦C-dried samples, as highlighted by the characteristic IR band at 440–500 cm−<sup>1</sup> attributed to the Zn-O stretching [22]; the conversion of Zn(II) species into stoichiometric ZnO is complete with calcination at 450 ◦C. The region between 1300 and 1600 cm−<sup>1</sup> is typically associated with the presence of hydrozincite-like species (Znx(CO3)y(OH)z) [31] and lowers after calcination. Furthermore, the increase in the treatment temperature also led to the disappearance of C-H stretching (3000–2800 cm−1) signals [32]; on the contrary, hydroxyl (3500–3400 cm<sup>−</sup>1) and styrene/carbonyl bands (1700–1500 cm<sup>−</sup>1) slightly increase upon calcination, as the effect of a possible BAC degradation with temperature, forming burnt moieties.

Regarding PDDA (Figure 3b), the IR spectrum of the dried sample presented some characteristic bands of PDDA, not degraded by the thermal treatment. The absence of the two main absorptions related to the symmetric and asymmetric stretching modes (with the characteristic double-pointed shape) of the carbonates at 1389–1500 cm−<sup>1</sup> [33] and of the other characteristic signals at 1047 cm−1, 837 cm<sup>−</sup>1, and 710 cm−1, combined with the presence of the 3400-cm−<sup>1</sup> band relative to O-H stretching (H2O and hydroxides), indicates that Zn2<sup>+</sup> is mainly present as hydroxide and not as hydrozincite. On the other hand, the IR spectrum of the calcined sample showed a stronger signal due to Zn-O stretching, whereas the hydroxide band decreased. Moreover, signals related to the polymeric stabilizer were no longer visible after calcination at 450 ◦C.

The UV–Vis spectra of all of the samples are presented in Figure 4. In the case of BAC (Figure 4a), the UV–Vis spectrum of the dried sample exhibited an absorption peak ascribable to ZnO (350–400 nm). The calcination process induced an increase in the peak intensity, due to pure ZnO and to a stronger absorption. For PDDA (Figure 4b), the band ascribable to ZnO showed a slight red shift from 374 nm to 382 nm with the increase in the treatment temperature, which could be related to a nanostructure size increase (a phenomenon generally reported in the literature) [14]. Broad absorption at higher wavelengths is attributed in the literature to crystal defects, especially oxygen vacancies in the sample [34].

**Figure 4.** UV–Vis spectra of dried (120 ◦C) and calcined (450 ◦C) samples: (**a**) ZnO-BAC; (**b**) ZnO-PDDA.

#### *3.4. XPS Characterization*

XPS was useful in the univocal determination of zinc chemical speciation. Zn(II) species can be discriminated by means of the ZnL3M45M45 Auger signal (expressed in kinetic energy, KE) and of the modified Auger parameter α . In fact, the Zn2p3/<sup>2</sup> photoelectronic signal is uninformative as the chemical shift on this signal is small, compared to that on the corresponding Auger maximum. In fact, typical values for Zn2p3/<sup>2</sup> ranging between 1021 and 1023 eV are indifferently reported for zinc metal or Zn(II) [35]. All the analyzed samples presented Zn2p3/<sup>2</sup> positions in this range (Table 2). This means that the correct zinc speciation, based only on the main photoelectronic peak, is unreliable [36–38]. Considering that the typical value of α reported for ZnO is equal to 2010.2 ± 0.3 eV [14,22,26,35,39,40],

the Zn Auger signal was acquired and analyzed to assess the presence of zinc oxide in the investigated samples (Figure 5). The main component of ZnL3M4,5M4,5 was used for the further calculation of α and it is mainly ascribed to the 1G final state transition, which is known to be the most probable [37,41]. In the case of BAC, XPS data revealed that the increase in the calcination temperature led to the increment of α (Table 2), reaching values compatible with the presence of ZnO on the surface of the nanostructures. Dried samples presented a low value of α (compatible with zinc hydroxides) [39], whereas it was around 2010.0 eV for 450 ◦C-calcined samples. These results are in agreement with the IR data. The XPS data on ZnO-BAC samples treated at 450 ◦C did not show any signal ascribable to nitrogen. This is indicative of the non-persistence of the cationic stabilizer on ZnONSs after calcination, and is in agreement with the IR results.

**Figure 5.** ZnL3M4,5M4,5 XP spectra of dried (120 ◦C) and calcined (450 ◦C) samples. ZnO-BAC (**a**,**b**); ZnO-PDDA (**c**,**d**).

**Table 2.** Zn2p, ZnL3M4,5M4,5 positions, and corresponding modified Auger parameter (α ) for all of the samples, as a function of the stabilizer and of the thermal treatment.


In a similar way, the XPS results on the PDDA-based samples are corroborated by the IR findings. The modified Auger parameter for zinc was compatible with the presence of ZnO, already in dried samples at 120 ◦C. The obtainment of pure ZnO in milder conditions with respect to the current state-of-the-art [14,21] makes ZnO-PDDA a highly appealing material. It is worth pointing out that, for the ZnO-PDDA treated at 120 ◦C, the N1s signal (Figure S3) that resulted is made up of two components: the first at a binding energy of 399.2 ± 0.2 eV, and the second at 402.2 ± 0.2 eV. They were

associated with the presence of free amine groups −N(CH2)/−NH2, and protonated groups (mainly −N(CH3)3 +/−NH3 <sup>+</sup>) [35]. This evidence can be indicative that the quaternary nitrogen of the PDDA has partly been preserved and that the cationic character of the polymer has not been completely lost.

#### *3.5.* ζ*-Potential Measurements*

ζ-potential measurements (Figure 6) were performed on the most interesting samples in terms of chemical composition, morphology, and ease of preparation. From the characterizations reported above, it appears clear that ZnONSs prepared with BAC benefit from a calcination step when stoichiometric ZnO phases are desired. During this step, the stabilizing agent degrades and forms burnt carbonaceous moieties on nanostructures. This could be the reason why the calcined ZnONSs @BAC did not show any isoelectric point and had a highly negative ζ-potential in the pH range investigated here (Figure 6a).

In the case of ZnO-PDDA, a mild drying process at 120 ◦C is enough to obtain ZnO. Moreover, as demonstrated by XPS, the polyelectrolyte is retained on the ZnO surface after the thermal treatment at 120 ◦C. The persistence of the cationic NP stabilizer is consistent with the presence of a positive and almost constant ζ-potential, with a pH ranging from 5 to 11 (Figure 6b).

**Figure 6.** ζ-potential measurements as a function of pH: (**a**) calcined (450 ◦C) ZnO-BAC nanostructures; (**b**) dried (120 ◦C) ZnO-PDDA nanostructures.

#### *3.6. Agar Disk Di*ff*usion Tests*

Antibacterial activity studies were carried out on both calcined (450 ◦C) ZnO-BAC nanostructures and dried (120 ◦C) ZnO-PDDA nanostructures. They showed consistent inhibition (>10 mm) [42] and no significant variation amongst stabilizers (Figure 7). In fact, the average inhibition diameters measured on five replicates resulted equal to 10.8 ± 0.6 mm for ZnO-BAC and 10.3 ± 1.2 mm for ZnO-PDDA. Slight differences in NP morphology and opposite ζ-potential values did not seem to have any influence on the NP antimicrobial behavior against a Gram-positive bacterial strain. No bacterial growth inhibition was found on the control samples.

**Figure 7.** Evaluation of antibacterial action by an Agar diffusion test: (**a**) calcined (450 ◦C) ZnO-BAC nanostructures; (**b**) dried (120 ◦C) ZnO-PDDA nanostructures.

#### **4. Conclusions**

In summary, we proposed a green electrochemical strategy for the synthesis of ZnONSs in aqueous alkaline media with low-cost, non-toxic chemicals under mild conditions. ZnO rod- and flower-like structures were successfully synthesized by employing two different stabilizers, namely, BAC and PDDA. While for BAC a thermal treatment at temperatures ≥450 ◦C was necessary for the complete conversion of the as-prepared gel-like material into ZnO, PDDA allowed for the preparation of pure ZnONSs without the need for severe thermal treatment. In fact, both FTIR and XPS measurements confirmed the presence of stoichiometric ZnO after a simple drying step at 120 ◦C. In all cases, TEM analysis revealed the presence of elongated (or rod-like) structures, generally assembled into more complex and ordered aggregates after calcination at 450 ◦C. Both BAC- and PDDA-containing materials exhibited a consistent antimicrobial efficacy against *B. subtilis*, as demonstrated by agar diffusion tests. The approach presented here can be considered as an improvement of the current methodologies to produce elongated ZnO nanomaterials in an aqueous solution, employing cationic capping agents, thanks to higher yields and milder preparation conditions [22]. Application of these ZnO nanostructures in transistor devices (PDDA-capped) and for cultural heritage preservation (BAC-capped) is envisaged, and work is scheduled for the future in this field.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/10/3/473/s1, Figure S1: TEM image of dried (120 ◦C) ZnO-PDDA sample. Figure S2: Selected area diffraction (SAED) image obtained on ZnONSs. Figure S3: N1s XP spectrum of ZnO-PDDA dried at 120 ◦C.

**Author Contributions:** M.C.S. performed TEM and spectroscopic characterizations, along with microbiological experiments, and wrote the first draft of the paper. R.A.P. developed the electrochemical approach and performed *X*-ray spectroscopic analyses. M.I. (Margherita Izzi) performed data treatment. G.P. performed ζ-potential measurements. R.G. developed microbiological protocols and supervised biological experimental activities. M.I. (Massimo Innocenti) and L.T. contributed to protocol development. N.C. supervised experimental activities. Authorship is limited to those who have contributed substantially to the work reported. All authors have read and agreed to the published version of the manuscript.

**Funding:** Partial financial support is acknowledged from the European Union's Horizon 2020 research and innovation program under the Marie Skłodowska-Curie Grant Agreement No. 813439, and from the Italian MIUR project "E-Design" ARS01\_01158.

**Acknowledgments:** Vincenzo Villone, Thomas Centobelli, Romeo Lettini, and Marina Menga are acknowledged for their help in collecting some experimental data.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Enhanced Antifungal Activities of Eugenol-Entrapped Casein Nanoparticles against Anthracnose in Postharvest Fruits**

#### **Yang Xue, Shitong Zhou, Chenyue Fan, Qizhen Du \* and Peng Jin \***

The Key Laboratory for Quality Improvement of Agricultural Products of Zhejiang Province, The College of Agricultural and Food Sciences, Zhejianga&F University, Hangzhou 311300, China; gxgxy0205@163.com (Y.X.); zst18846494139@163.com (S.Z.); fffffcy1021@163.com (C.F.)

**\*** Correspondence: qizhendu@163.com (Q.D.); jinpeng@zafu.edu.cn (P.J.); Tel.: +86-571-63742176 (Q.D.); +86-18368187965 (P.J.); Fax: +86-571-88218710 (P.J.)

Received: 19 November 2019; Accepted: 11 December 2019; Published: 13 December 2019

**Abstract:** This study aims to improve the antifungal effects of eugenol through low-energy self-assembly fabrication and optimization of eugenol-casein nanoparticles (EC-NPs). Optimized EC-NPs (eugenol/casein ratio of 1:5) were obtained with a mean size of 307.4 ± 2.5 nm and entrapment efficiency of 86.3% ± 0.2%, and showed high stability under incubated at 20 and 37 ◦C for 48 h. EC-NPs exhibited satisfactory sustained-release effect at 20 ◦C or 37 ◦C, with remaining eugenols amounts of 79.51% and 53.41% after 72 h incubation, respectively, which were significantly higher than that of native eugenol (only 26.40% and 19.82% after the first 12 h). EC-NPs exhibited a greater antifungal activity (>95.7%) against spore germination of fungus that was greater than that of native eugenol, showed 100% inhibition of the anthracnose incidence in postharvest pear after 7 d. EC-NPs is potential as an environmental-friendly preservatives in the food industry.

**Keywords:** essential oil; encapsulation; bovine casein; antifungal activity; fruit preservation

#### **1. Introduction**

Up to now the low-cost and efficient of synthetic chemicals (i.e., inorganic salts, organic reagents, fungicides and pesticides) are widely used in the decay-controlling against postharvest diseases since the decay caused by spoilage microorganism invasion brings huge economic loss [1,2]. Considering to the food safety and human health, natural compounds that have low health or environmental impact are receiving wide attention [3,4]. Plant-derived bioactive substances have been applied as natural preservatives (e.g., essential oils, Base Natural, a commercial preservative) in the food industry due to their antifungal properties [5–7]. Eugenol (1,2-methoxy-4-(2-propenyl)-phenol), a principal component of the herbal essential oil from basil, has been classified as GRAS (generally recognized as safe) food additives, because reports indicated it has potential for the food field due to their antimicrobial properties against a wide range of microorganisms [8,9].

However, eugenol possesses high volatility and low water-solubility, which severely handicap its utilization due to the short timeliness for antibiosis [10]. In recent decades, the embedding and encapsulation has become a novel strategy to markedly offer numerous benefits for active compounds, for example, protection against oxidation, enhanced stability, retention of volatile ingredients and permeability [11]. Recently, encapsulation of eugenol showed enhanced bioactivity and stability [12,13]. However, the high-energy techniques or cumbersome processes for nano-formulation are difficult to scale-up for practical application. In addition, the preparation of some nanocarriers requires the addition of some chemical reagents, such as, dichloromethane [14], surfactant [15], which increases the risk of potential toxicity and safety that limit their applications in food industry [16]. Therefore,

eco-safe technologies for the nano-formulation of eugenol are needed to develop, so that they can be used in the food industry [17,18].

Caseins, the major component of milk proteins in bovine milk, is a low-cost and commercially available food-grade additive in food and beverage [19]. In the native state of caseins, they are very stable for high temperature and pressure (for example, treated by 100 ◦C and 100 MPa without losing their essential integrity). The natural caseins can self-assemble to the micellar structure in solution, with diameters in the approximate range 10–400 nm [20]. The amphiphilic nature of the caseins causes them to act more as block copolymers micelles of alternating charge and hydrophobicity, which are suitable for encapsulating the compounds of poor water solubility in the hydrophobic core of the micelle [21]. Therefore, caseins have been widely used as efficient nanocarriers for hydrophobic drug for controlled release [22]. Accordingly, the present study aims to prepare eugenol-entrapped casein nanoparticles through a low-energy and simple self-emulsifying technique, which will provide a promising alternate for nano-formulation of eugenol against postharvest decay of fruits and vegetables.

#### **2. Materials and Methods**

#### *2.1. Materials*

Bovine casein (>98%) and eugenol (99%) were obtained from Sigma–Aldrich, Chemical Co. (St. Louis, MO, USA). Ethanol was purchased from China National Pharmaceutical Industry Co. Ltd. (Beijing, China). All solutions were prepared by using deionized water (Millipore, Bedford, MA, USA). All other reagents were of analytical grade.

#### *2.2. Preparation of Nanoparticles*

Bovine casein used for the preparation of nanoparticles was fully dissolved in deionized water to form a final concentration of 20 mg/mL. Eugenol was dissolved into ethanol with a final concentration of 200 mg/mL. The casein solution (10 mL) in 50 mL glass beaker was stirred for 30 min at 500 rpm, and then eugenol solution was stepwisely added into the casein solution with a volume of 5 μL. After the addition of eugenol solution, the mixture solution was continued to stir for 30 min to yield eugenol-casein nanoparticle (EC-NP) dispersion. The denatured protein in the dispersion was removed through centrifugation at 12,000 rpm for 10 min. The EC-NPs samples were stored in a freezer (4 ◦C) for further use.

#### *2.3. Characterization of EC-NPs*

The size distribution of the fresh nanoparticles was determined as mentioned in our previous study [23]. In brief, 500 μL of the nanoparticles suspension were diluted into 5 mL of pre-filtered deionized water. The analysis was performed in dynamic light scattering using a Zetasizer ZS 90 instrument (Malvern Instruments, Malvern, UK) at 25 ◦C temperature, employing a nominal 5 mW He−Ne laser operating at a 633 nm wavelength and 173◦ scattering angle. The EE was obtained by determining the free eugenol in EC-NPs solution, which was separated by using an Amicon Ultra-7K centrifugal filter device (7000 MWCO, Millipore Corp., Billerica, MA, USA.). The quantitative analysis of eugenol was performed by a high-performance liquid chromatographic (HPLC) Shimadzu LC 20A system (Shimadzu, Kyoto, Japan), consisted of two LC-10A pumps, an SIL-10Avp autosampler, an SPD-M10Avp UV detector and a Symmetry C18 (5 μm, 4.6 mm × 250 mm) column. The mobile phase was composed of methanol/water (65:35) at a constant flow rate of 1 mL/min at 30 ◦C, and monitored at 282 nm [24]. All data were expressed as the mean value of three independent batches of the samples.

The entrapment efficiency (EE, in percent) of eugenol was calculated as the percentage of entrapped eugenol to total eugenol (as the following equation).

$$EE\% = \frac{\text{Total amount of euegenol} - \text{free euegenol amount}}{\text{Total amount of eugenol}} \times 100\%$$

The morphological characterization of eugenol nanoparticles was performed by transmission electron microscopy (TEM), (Hitachi, H-9500E, Tokyo, Japan) [25]. Briefly, a dispersion of NPs diluted with pure water was adsorbed onto a carbon-coated formvar film that was attached to a metal specimen grid. Excess sample was removed through blotting and the grid was covered with a small drop of staining solution (2% *w*/*v* phosphotungstic acid). The staining solution was left on the grid for a few min and then the excess solution was drained. The sample was allowed to air dry thoroughly and was then examined using a transmission electron microscope.

#### *2.4. Stability Assessments of EC-NPs*

To investigate the effects of temperatures on stability of the EC-NPs dispersions, 5 mL samples solution were placed in glass vials and stored at 4 ◦C, 20 ◦C and 37 ◦C, respectively. The entrapment efficiency (EE) of the dispersions was measured at predetermined time intervals (0, 2, 4, 8, 12, 24, 48 and 72 h) for the stability evaluation of the nanoparticles. After a treatment of the EC-NPs dispersion containing 4 mg/mL eugenol and 20 mg/mL caseins, the released eugenol and EE of EC-NPs were measured according to the above method. a mixture solution containing 4 mg/mL eugenol and 20 mg/mL caseins was used for a control. All samples were measured in triplicate.

#### *2.5. Cell Culture*

Luria−Bertani (LB) medium (g/L; 10 g tryptone, 5 g yeast extract and 10 g NaCl) added 1% glucose was used for the isolation and culture of spoilage microorganisms. Spoilage bacteria *Botrytis cinerea* [26] was isolated from decayed pear fruits. Mycelium was precultured on LB agar media at 30 ◦C for 5 d as the seed. The spore suspension was prepared as our previous description [23], by washing the 20-day seed cultures with sterile water containing 0.01% (v/v) Tween-80 and then diluting them to 1 <sup>×</sup> 105 spores/mL with the aid of a hemocytometer.

#### *2.6. Determination of Antifungal Activity*

The antifungal sensitivity of free eugenol and EC-NPs nanoparticles was determined by modified disc diffusion method [27]. Briefly, various amounts of free eugenol (1.00, 4.03, 8.04, 10.06, 15.08, 20.11, 25.13, 30.16, 35.19, 40.21 and 50.24 μg/mL) and equivalent eugenol-entrapped casein nanoparticles solution were uniformly smeared onto the LB agar media, respectively. Of the prepared spore suspension 5 μL was inoculated on the sterile filter paper at the center of the petri dish, and then incubated at 30 ◦C. The same manual tests were performed for control without eugenol. Finally, the inhibition rate of mycelium growth was calculated as a percentage of the control groups without eugenol. Minimum inhibitory concentration (MIC) is defined as the lowest concentration of eugenol or EC-NPs, which completely inhibits visible growth on solid media. To further quantitatively investigate the sustained-release effect of EC-NPs on the spore germination, the spore suspension was inoculated into 20 mL LB medium, added with an initial concentration of 40 μg/mL eugenol in both nanoencapsulated and nonencapsulated groups, respectively, and then cultured at 30 ◦C with shaking (200 rpm). The optical density at 600 nm (OD600) was used to monitor cell growth. The control was set without eugenol in the culture, and the entire experiment contained three replicates.

#### *2.7. E*ff*ects of EC-NPs on Anthracnose Disease on Pear Fruit*

The epidermis of pear fruits (a pear orchard in Hebei, China) was scrub with sterile saline, and then inoculated with 5 <sup>μ</sup>L of *Botrytis cinerea* spore suspension (1 <sup>×</sup> <sup>10</sup><sup>5</sup> spores/mL) in the middle of each fruit with a sterile injection needle. An equivalent amount of eugenol (4 mg) of free eugenol (1 mL) and EC-NPs solution (1 mL) was sprayed onto the pear fruit surface. Each group contained three replicates of 25 pear fruits, and the control group was correspondingly treated with sterile distilled water (1 mL). Subsequently, all of the treated samples were stored at 25 ◦C in a sterile room. Disease incidence and lesion diameters were recorded. The incidence was defined as the appearance of decay spots and the

color changes of flesh at the inoculation hole. All data were analyzed by using Duncan's multiple range tests in the SPSS 13.0 software (IBM Corp., Armonk, NY, USA)

#### **3. Results and Discussion**

#### *3.1. Characterization of Eugenol-Entrapped Casein Nanoparticles*

Casein nanocarriers are widely used as delivery systems for hydrophobic drugs and bioactive compounds due to its excellent emulsification [28] and self-assembled ability [22]. In this work, eugenol-loaded casein nanoparticles (EC-NPs) were prepared through the pre-formation of spherical casein micelles followed by the stirring treatment of the mixture with eugenol, resulting in assemblies of casein/eugenol complexes at the nanometric scale. The effects of the eugenol concentration on the particle size, polydispersity (PDI), zeta potential, and entrapment efficiency were further studied as shown in Table 1. The increase of the eugenol amount from 1 to 6 mg/mL resulted in formation of larger nanoparticles from 249.3 to 333.8 nm. Similar results were observed in previous studies, in which the presence of antimicrobial agent yielded a more viscous dispersed phase, resulting in larger particles [13,14]. Meantime, PDI values of casein nanoparticles was found to in the range of 0.261–0.333, which indicates moderate size distribution [29]. Size and size distribution are important characteristic indexes because it is related to the release of active compound. Generally, spherical shape and better dispersity are beneficial to the release of active compound [30]. Zeta potential is also an important parameter to reflect the physicochemical and biological stabilities of nanoparticles in dispersion [31], which helps the formulation to enhance the long-term stability [32]. As shown in Table 1, the addition of eugenol slightly increased negative charges on the particle surface from −14.47 to −21.91 mV, which indicated that the nanoparticles with eugenol concentration of 4 mg/mL exhibited ideal particle stability.


**Table 1.** Effects of the concentration of eugenol on the particle size, zeta potential, polydispersity indexes (PDI) and entrapment efficiency (EE) of eugenol loading in casein nanoparticles.

Entrapment efficiency (EE) is used to indicate the amount of compound entrapped into the polymeric matrix. The EE of the nanoparticles reached the maximum value of 91.1% when eugenol concentration was 2 mg/mL. However, with the eugenol rising from 2 to 6 mg/mL, the EE decreased from 91.1% to 67.1%. This result was in agreement with the nanoethosomes in our previous work [23]. The EC-NPs made of 2 mg/mL of eugenol exhibited a spherical shape (Figure 1a). Furthermore, to optimize the formulation of the nanoparticles, the antifungal effects of EC-NPs with consistent eugenol amount were evaluated in LB media against anthracnose strain *Botrytis cinerea*. As shown in Figure 1b, EC-NPs with 4 mg/mL eugenol showed the highest antifungal efficiency of 88.5% against anthracnose after 48 h of inoculation. There was a positive correlation between eugenol concentration and antifungal efficiency in this study. This result also showed that this increase in particle size did not compromise the antifungal action of the eugenol component-grafted EC-NPs.

**Figure 1.** Characterization of eugenol nanoparticles. (**a**) Transmission electron microscopic image of eugenol-casein nanoparticles (EC-NPs). (**b**) Effects of eugenol concentration (mg/mL) on antifungal effect of eugenol-entrapped casein nanoparticles. The mean ± SD for three replicates are illustrated.

#### *3.2. Release Assessment of Eugenol Nanoparticles*

In order to study the stability effects of encapsulated eugenol compared with that of native eugenol, the eugenol residual amount of EC-NPs and free eugenol in solution was investigated for incubation at different temperatures. As shown in Figure 2, EC-NPs was stable for 72 h when incubated at 4 ◦C or 20 ◦C, with still 88.75% and 79.51% of remaining eugenol in the EC-NPs, respectively. Even after 72 h incubation at 37 ◦C, the EC-NPs showed a satisfactory slow release trend, possessing 53.41% of remaining eugenol. However, the unencapsulated eugenol solution gradually lose eugenol at 4 ◦C, only 21.09% of remaining eugenol after 72 h. As to 20 and 37 ◦C, the native eugenol possessed remaining eugenol amounts of only 26.4% and 19.82% after 12 h, respectively, and the eugenol lose about 80% of the initial amount, and especially the eugenol disappeared almost after 48 h. This result further clearly confirmed that native eugenol usually occurs with the burst effect dissipation and resulted in the short-term existence, which greatly compromised its long-lived antifungal effect [23]. The higher amounts of eugenol remaining in solution containing EC-NPs lead to the better lasting antifungal effect, which implied that the encapsulation of eugenol into casein nanoparticles could improve the stability and produce maintain-released effect of eugenol. Furthermore, the particle size, PDI and zeta potential of EC-NPs were further investigated for incubation at different temperatures (Table 2). These results showed that EC-NPs exhibited remarkable physical and chemical stability at 20 ◦C for 48 h though the size and PDI values present slight increase.

**Figure 2.** Effects of temperature on the eugenol stability in solution and in EC-NPs dispersion.


**Table 2.** Effects of the storage temperature on the particle size, zeta potential and polydispersity indexes (PDI) of EC-NPs nanoparticles after 48 h.

#### *3.3. Antifungal Assessment of Eugenol Nanoparticles*

The antifungal activity of eugenol and its potency were quantitatively assessed by determining the MIC. Table 1 show the MIC of free eugenol and EC-NPs nanoparticles against anthracnose in vitro. The difference in the mycelium diameter of the inhibition zone indicates the sensitivity of fungal strains to various concentrations of free eugenol or EC-NPs. The MIC values of EC-NPs against anthracnose were 40.21 μg/mL, which were slightly lower than those of free eugenol (50.24 μg/mL), which indicate EC-NPs possessed higher antifungal activity than free eugenol (Table 3). These results further show the property of eugenol and EC-NPs, i.e., they are potential to inhibit the growth of anthracnose in common fruit. Compared with the free form of eugenol, the antifungal ability of EC-NPs was compromised at the initial period of antifungal reactions [33], due to the slow release profile of the nanoencapsulated eugenol. This result was consistent with our previous work on the antifungal effect of eugenol nanoethosomes [23].



<sup>a</sup> The eugenol concentration is defined as all the eugenol amount in Luria–Bertani (LB) agar media. <sup>b</sup> The inhibition ratio is defined as a relative percentage of the control mycelial diameter after 72 h culture.

Evaluation of antifungal ability of EC-NPs was also performed by observing its inhibition effect on the mycelial diameter of anthracnose on LB agarose plate. As shown in Figure 3a, after incubation for 48 h, the mycelial growth was observed with significant differences between the EC-NPs group and other groups. Compared with the control, native eugenol and casein-eugenol mixture gave 50% inhibition of the mycelial growth. In contrast, the EC-NPs showed a completely inhibition on the fungal growth. Furthermore, the antifungal effects of EC-NPs were evaluated by calculating the inhibition rate on spore germination of fruit *Botrytis cinerea* pathogen fungus. As shown in Figure 3b, after 24 h of incubation, the spore germination rate of native eugenol group and casein-eugenol mixture group already reached to 51.8% and 61.1%, respectively, while the spore germination rate of EC-NPs were only 4.3% after 48 h and 38.7% after 72 h. Obviously, the fungal inhibition effect of EC-NPs was stronger than the native eugenol or casein-eugenol mixture, indicating that the nano-formulation of eugenol could effectively improve the fungal inhibition effectiveness, which should be contributed by the maintain-release of eugenol from EC-NPs, which prolonged the action time of eugenol against fungi. In the previous work

we have determined that native eugenol is usually volatile with a burst-effect release pattern, resulted in rapid dissipation during the initial period [23], while EC-NPs gave a maintain-release during a long time. In result, EC-NPs possess the potential for the antifungal application in post-harvest fruits.

**Figure 3.** Evaluation of antifungal effects of eugenol nanoparticles. (**a**) Effects of native eugenol (E), casein-eugenol mixture (Mix) and eugenol-entrapped casein nanoparticles (EC-NPs) on the mycelial growth and (**b**) dynamic analyses of the spore germination in liquid media treated with the samples.

#### *3.4. EC-NPs as Preservative against Fruit Corruption*

EC-NPs was further determined the effectiveness on suppressing the anthracnose of fruits caused by the pathogenic invasion. Compared to the native eugenol, the rot of pear fruit caused by *Botrytis cinerea* strain was employed to test the effectiveness. After spore suspension (10 μL, <sup>1</sup> <sup>×</sup> 105 CFU/mL) was injected into epidermis of the pear fruit, the rot of pear fruit was observed by sprayed native eugenol or EC-NPs. As shown in Figure 4a, after being inoculated for 7 d, the obvious rots (black spot) were observed on the epidermis at the control group (treatment with sterilized water instead of preservatives), and minor wound decay occurred in native eugenol-treated group. In contrast, almost no any black spot and wound decay was observed in EC-NPs-treated group. After cutting open the inoculation site, we found that both the control group and native eugenol-treated group decayed in deep pulp while EC-NPs-treated group did not decay in deep pulp (Figure 4b). Obviously, EC-NPs completely suppressed the pathogen. As we known, nanoparticles possess excellent permeability for cortex [23,34]. Thus EC-NPs might get through the deeper pulp to deliver eugenol for antifungal. The disease incidences of control and native eugenol-treated groups reached 100% and 89% for 8 d incubation after inoculation, while that of EC-NPs group was only 23% (Figure 4c). In combination of the inhibition of the fruit rot and the disease incidences, we can conclude that EC-NPs significantly potentiate the antifungal efficacy of eugenol in inhibiting fruit anthracnose.

**Figure 4.** Effects of native eugenol and EC-NPs suppressing disease lesion of pear fruit inoculated with *Botrytis cinerea*. (**a**) Rot in the cuticular layer, (**b**) rot in pulp layer and (**c**) disease incidence of the pear fruit. Vertical bars represent the standard errors of the means of triplicate assays. The symbol \* (*p* < 0.05) and \*\* (*p* < 0.01) indicate a significant difference between EC-NPs and native eugenol group.

#### **4. Conclusions**

Eugenol was encapsulated in casein micelles by simple process without any other additives. a mass ratio of 5:1 of caseins/eugenol yielded the highest encapsulation efficiency and stability for eugenol-casein nanoparticles (EC-NPs). EC-NPs significantly improve the antifungal efficacy against anthracnose. These results indicate that EC-NPs nanoparticles could be used as an economical and simple-manufactured preservative for postharvest fruits against microbial spoilage.

**Author Contributions:** P.J. and Q.D. designed the experiments, analyzed the data and wrote the paper. Y.X., S.Z. and C.F. performed the experiments.

**Funding:** This work was supported by the National Natural Science Foundation of China (grant numbers 31700078), the Key R&D Program Project of Zhejiang Province, China (Grant 2019C02072), the Zhejiang Agricultural and Forestry University for Talent Program (W20170029), and the Natural Science Foundation of Zhejiang Province, China (LY17C200018).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Layer by Layer Antimicrobial Coatings Based on Nafion, Lysozyme, and Chitosan**

**Ella N. Gibbons 1, Charis Winder 2, Elliot Barron 2, Diogo Fernandes 3, Marta J. Krysmann 1,\*, Antonios Kelarakis 2,\*, Adam V. S. Parry <sup>4</sup> and Stephen G. Yeates <sup>4</sup>**


Received: 14 October 2019; Accepted: 28 October 2019; Published: 4 November 2019

**Abstract:** The study focuses on the development of a new family of layer-by-layer coatings comprising Nafion, lysozyme and chitosan to address challenges related to microbial contamination. Circular dichroism was employed to gain insights on the interactions of the building blocks at the molecular level. Quartz crystal microbalance tests were used to monitor in real time the build-up of multilayer coatings, while atomic force microscopy, contact angle and surface zeta potential measurements were performed to assess the surface characteristics of the multilayer assemblies. Remarkably, the nanocoated surfaces show almost 100% reduction in the population of both *Escherichia coli* and *Staphylococcus aureus*. The study suggests that Nafion based synergistic platforms can offer an effective line of defence against bacteria, facilitating antimicrobial mechanisms that go beyond the concept of exclusion zone.

**Keywords:** antimicrobial; layer-by-layer; coatings; Nafion; multilayers

#### **1. Introduction**

The quest for advanced antimicrobial materials is driven by the large diversity of remarkably adaptable pathogens coupled with the alarming evolution of drug resistant strains that cause serious infections to humans and the ecosystem [1–4]. Despite preventive measures and increased public awareness, contiguous bacterial colonies are found on a variety of surfaces such as foodservice equipment, water distribution pipelines, swimming pools, lakes, rivers, public transport vehicles, toilets, door handles, home appliances, and air-conditioning filters. At the same time, bloodstream infections originating from catheters, implants, and surgical tools result in enormous costs for the healthcare system [5].

Layer by layer (LbL) assemblies, based on the alternated adsorption of oppositely charged molecules or nanoparticles, is a versatile approach that affords control at the nanoscale level, generating stable and robust coatings [6–8]. Based on those principles, a wide range of LbL antimicrobial coatings comprising polymers, nanoparticles, enzymes, peptides, biological molecules, and antibiotics as building units has been reported [9]. Their antimicrobial performance relies on bioadhesion resistance, contact-killing, release-killing, or a combination of those mechanisms [9].

Along those lines, the positively charged amidated ponericin G1, a strong antimicrobial against *Staphylococcus aureus* (*S. aureus*), was incorporated to hydrolytically degradable LbL coatings based on poly (b-amino esters). The peptide was released from the film in a controlled manner and was effective in inhibiting bacteria attachment, thus demonstrating significant potential for implant materials and bandages [10]. In a contact-killing demonstration, LbL assemblies comprising poly (allylamine hydrochloride) and poly (sodium 4-styrene sulfonate) showed sufficient density of mobile cations that endowed significant antimicrobial activity [11]. In an anti-adhesion strategy, poly (L-lysine)/poly (L-glutamic acid) multilayers with the top bilayers bearing the pegylated polyanion drastically suppressed the adsorption of *Escherichia coli* (*E. coli*) [12].

In this work, we focus on LbL assemblies comprising two naturally occurring antimicrobials, namely lysozyme [13] and chitosan [14] along with Nafion, a synthetic ionomer with a robust, Teflon-like backbone bearing hydrophilic sulphonic acid groups. As a direct consequence of its chemical composition, Nafion forms proton exchange membranes with supreme structural and chemical stability that set the benchmark for fuel cell applications [15]. Moreover, Nafion has been shown to possess an exclusion zone against bacterial growth, an effect that has been attributed to repulsive forces between its negatively charged surface and the similarly charged cell membranes [16,17]. In this work, we demonstrate that, although the surface charges of Nafion have been neutralised (if not overcompensated) by the adsorption of positively charged molecules, the coatings show remarkable antimicrobial activity against *E. coli* and *S. aureus*. In that sense, our work paves the way for the development of a new family of Nafion-based nanostructured coatings with enhanced antimicrobial performance that do not necessarily rely on exclusion zone effects.

#### **2. Materials and Methods**

#### *2.1. Materials*

Nafion (DE 1021) (Chemours Company, Wilmington, DE, USA) with a total H<sup>+</sup> exchange capacity of 1.1 mequiv/g was obtained as a 10 wt% dispersion in water (Ion-Power) or as a 15 wt% dispersion in a mixture of low aliphatic alcohols (3-propanol, ethanol, and others) and water (Ion Power). Lysozyme (Buchs, Switzerland) from chicken eggs (106 U/mg) and medium molecular weight chitosan (Milwaukee, WI, USA) were obtained from Sigma Aldrich (Dorset, U.K). Chitosan was dispersed in water containing 0.1 wt% acetic acid.

#### *2.2. Quartz Crystal Microbalance with Dissipation Monitoring (QCM-D)*

The quartz crystal microbalance with dissipation monitoring (QCM-D) tests were performed using a Q-sense E1 unit (Biolin Scientific, Stockholm, Sweden) equipped with a Peltier-controlled flow cell (flow rate was set at 0.2 mL/min) with temperature accuracy of 0.02 ◦C. Au-modified crystals with a fundamental resonance frequency close to 5 MHz and diameter 150 nm were spin-coated (by depositing a drop of a 0.5 wt% Nafion solution in ethanol) and were then left at room temperature for at least seven days. All measurements included an initial equilibrium step of the crystal in the air to determine the fundamental resonant frequency, followed by an equilibrium step under constant flow of water to establish the baseline of the hydrated surface.

On the basis of Sauerbrey relation: Δ*m* = −(*C*/*N)* Δ*f*, deposition of a uniform layer with mass Δm reduces the resonant frequency of the crystal by Δ*f*, where *N* is the overtone number (herein all values reported refer to *N* = 3) and *C* is the integrated crystal sensitivity that depends upon the intrinsic properties and the thickness of the crystal [18]. The dissipation factor *D* is defined as *D* = *Ed*/(2π*Es*), where *Ed* is the energy dissipated during one period of oscillation and *Es* is the energy stored in the system [19].

#### *2.3. Contact Angle Measurements*

The contact angle of distilled water droplets (5 μL) deposited on the coated quartz crystals was determined by means of an OptoSigma (OptoSigma Corp., Santa Ana, CA, USA) optical tensiometer using the standard sessile drop technique (Digidropmeter, GBX). The photos were captured 20 s following the deposition of the droplets. A minimum of five spots on each specimen were measured.

#### *2.4. Circular Dichroism (CD) Spectrometry*

Lysozyme solutions in the presence and absence of Nafion were inserted to a quartz cuvette of 0.1 cm light path length and their circular dichroism (CD) spectra at 25 ◦C were collected using a Jasco J-815 CD spectropolarimeter (Jasco, Tokyo, Japan). Each spectrum was collected for five accumulations, with a scanning range from 260 to 180 nm, a band width of 2 nm, data pitch of 0.5 nm, digital integration time of 1 s, and a scanning speed of 100 nm/min. Values for the baseline (measured without a cuvette) and the blank solutions (ultrapure water) were subtracted from test values. The CD spectra in terms of α-helix, β-sheet, and random structures were analysed using DichroWeb [20]. All samples had total concentration of 0.01 mg/mL, but varying fLys values, where fLys stands for weight of lysozyme/weight of Nafion.

#### *2.5. Atomic Force Microscopy (AFM)*

The samples were mounted on magnetic sample holders for AFM tests. The measurements were performed on a Park XE-100 (Parksystems, South Korea) in non-contact mode using a cantilever with spring constant of approximately 40 N/m. Images were taken with a 512 by 512 pixel resolution at a scan rate of between 0.2 and 0.5 Hz.

#### *2.6. Surface Zeta Potential (*ζ*surface)*

(Naf/Lys)6, (Naf/Chi)6, and (Naf/Lys/Naf/Chi)2 coatings were deposited on Nafion-precoated aluminium foil and polypropylene substrates via standard dip-coating protocols. ζsurface measurements were recorded at 25 ◦C in a surface zeta cell apparatus, at a forward angle of detection (13◦) on a Zetasizer Nano ZS, (Malvern Panalytical, Malvern, UK). Data were recorded at different displacement distances from the surface which then allow for the surface zeta potential to be calculated, according to the equation below:

$$
\mathcal{L}\_{\text{surface}} = -\mathcal{L}\_{\text{intercept}} + \mathcal{L}\_{\text{tracer particles}} \tag{1}
$$

where ζintercept is the zeta potential at displacement 0 from the surface, calculated from a linear regression fit. Four repeat measurements were recorded at each displacement of 1.25 μm from the previous point, with a total of four displacement points.

Polystyrene beads (DTS1235, −42 mV ± 10%) were used as the tracer particles and water was selected as the dispersant. The measured electrophoretic mobilities (UE) were converted into ζ values assuming Smoluchowski approximation *F*(ka) = 1.5 for Henry's equation *U*<sup>E</sup> = 2εζ*F*(ka)/3η, where ε, η are the dielectric constant and the viscosity of the dispersant, respectively [21].

#### *2.7. Antimicrobial Testing*

A. Culturing Method. 250 mL Erlenmeyer flasks containing 25 mL nutrient broth were inoculated with a single loop of bacteria and incubated for 24 h in a SciQuip Incu-Shake MIDI orbital shaker (SciQuip Ltd, Newtown, Wem, Shropshire, UK) set to 200 rpm at 37 ◦C. Cultures were centrifuged at 4000 rpm for 10 min. Subsequently, the supernatant was discarded, 20 mL of 1/4 strength Ringer's solution was added, and the tubes were vortexed.

Tubes were centrifuged for a further 10 min at 4000 rpm and supernatant again discarded. 2 mL of <sup>1</sup> <sup>4</sup> strength Ringer's solution was added and tubes vortexed a final time. Resuspended cultures were diluted in nutrient broth to obtain an absorbance reading equivalent to a 0.5 McFarlane standard as recorded by Biochrom WPA S800 visible spectrophotometer (Biochrom Ltd, Cambridge, UK).

B. Disk testing method. Disk testing method was adapted from the literature [22]. Each disk was assigned to one of the twelve wells using a random number generator. Each of these wells was lined with sterile aluminium foil, for ease of removal of disks and prevention of bacterial run-off. An additional well was filled with 1mL sterile distilled deionized water to prevent dehydration of samples.

200 μL bacterial culture was added to each disk and was incubated for 20 h at 37 ◦C. Following incubation, each disk was transferred along with the foil into 9.8 mL 1/4 strength Ringer's solution and sonicated for 10 min. Sonicated solutions were serially diluted 100 μL sample solution into 900 μL 1/4 strength Ringer's solution. A 100 μL respective sample was used to spread onto each nutrient agar plate, each dilution being plated in triplicate. Plates were incubated for 20 h at 37 ◦C, then counted for colonies. For each type of coating five crystals were tested and the average values were determined.

#### **3. Results**

The QCM-D sensogram shown in Figure 1a describes the build-up of four Nafion/lysozyme bilayers on a Nafion precoated crystal resonator. We note that Nafion combines a hydrophobic backbone with hydrophilic perfluoroether side chains terminated with sulphonic acid groups and, thus, it undergoes microphase separation into polar and nonpolar domains. In particular, Nafion is described as a network of parallel water-filled channels held in place by the cross-linking action of the crystalline domains [23].

**Figure 1.** QCM-D sensograms at 25 ◦C monitoring the build-up of; (**a**) (Naf/Lys)4 and (**b**) (Naf/Lys/Naf/Chi)2. The letters "W", "N", "L", and "C" signify the injection of water, Nafion, lysozyme, and chitosan solution, respectively.

As seen in Figure 1a, injection of water results in a pronounced drop of the oscillating frequency coupled with a corresponding increase in the dissipation factor, indicating significant swelling of the ionic channels. The baseline recorded for the hydrated Nafion membrane remains stable, eliminating the possibility of even minor dissolution under water flow. Upon water exposure, Nafion's surface is reorganised with the sulphonic groups to be turned into the aqueous phase, a mechanism that gives rise to a large water contact angle hysteresis [24]. Owing to its amphiphilic nature Nafion has been shown to bind not only with charged polymers such as poly (oxyethylene) [25] and poly (oxypropylene) based diamines [26], but also with non-ionic surfactants via hydrogen bonding, hydrophobic interactions [27–29].

As evident by the significant drop in Δ*f* (Figure 1a), lysozyme (at pH = 6.2), a ubiquitous enzyme widely used as food preservative, is massively adsorbed on the hydrated Nafion film. Given that lysozyme bears positive charges within the pH range 1–11 [30], it is strongly attracted to the negatively charged Nafion, so that subsequent rinsing with water removes only a limited amount of weakly bound lysozyme molecules. The Nafion/lysozyme deposition cycle was repeated for three more times in an identical fashion to generate an ultrathin LbL membrane denoted hereafter as (Naf/Lys)4, while the deposition of two further bilayers led to (Naf/Lys)6. Likewise, the (Naf/Chi)6 LbL coating was assembled using chitosan, an aminopolysaccharide biopolymer with a broad antimicrobial spectrum that is extensively used to prevent bacterial contamination in food and drug packaging, as the positively charged layer. The QCM-D sensogram in Figure 1b describes the formation of a

(Naf/Lys/Naf/Chi)2 assembly as a three-component coating that relies on the attractive Nafion/lysozyme and Nafion/chitosan forces. The action of those attractive forces is further confirmed by the spontaneous precipitation that takes place upon mixing 0.1 wt% Nafion with either 0.1 wt% lysozyme (pH = 6.2) or 0.1 wt% chitosan. Note that the injection of chitosan results in a rather limited drop in Δf compared to lysozyme, presumably due to enhanced steric hindrance.

Figure 2 displays AFM images of the Nafion, (Naf/Lys)6, (Naf/Chi)6, and (Naf/Lys/Naf/Chi)2 coatings. It has been demonstrated that lysozyme adsorbed on a solid surface undergoes pronounced conformational reorganization driven by hydrophobic-hydrophobic interactions, ultimately resulting in the formation of aggregates that diffuse on the surface [31]. Chitosan molecules on a solid substrate follow similar clustering/agglomeration patterns, ultimately adopting significant levels of surface roughness [32]. At the same time, the topological characteristics of Nafion mirror the microphase separation of the bulk and are largely dependent on the relative humidity and the hydration levels [33].

**Figure 2.** AFM images of the: (**a**) Nafion, (**b**) (Naf/Lys)6, (**c**) (Naf/Chi)6, and (**d**) (Naf/Lys/Naf/Chi)2 coatings deposited on quartz crystal microbalance with dissipation monitoring (QCM-D) crystals.

As shown in Figure 3, the water contact angles for (Naf/Lys)6, (Naf/Chi)6, and (Naf/Lys/Naf/Chi)2, were found to be 45.3◦, 59.0◦, and 65.1◦, respectively, compared to 73.3◦ for a Nafion coated surface. It is generally accepted that hydrophobic surfaces are desirable for antimicrobial applications, however there is evidence to suggest that intermediate contact angles, as those found in the present systems, might also be compatible with advanced antimicrobial behaviour [34]. The above coatings were applied to polystyrene surfaces without compromising their optical transparency (Figure S1a), although the nanocoated surfaces showed enhanced UV-vis absorbance (Figure S1b). The development of transparent, yet UV blocking packaging materials with advanced antimicrobial properties are of supreme importance in the food industry and the coatings disclosed here point to this direction.

**Figure 3.** Water contact angles of: (**a**) Nafion, (**b**) (Naf/Lys)6, (**c**) (Naf/Chi)6, and (**d**) (Naf/Lys/Naf/Chi)2 coatings deposited on QCM-D crystals.

As shown in Figure 4, the (Naf/Lys)6, (Naf/Chi)6, and (Naf/Lys/Naf/Chi)2 coatings inhibit *E. coli* growth by 99.99%, 99.99%, and 99.95%, respectively, compared to 57.7% for the Nafion coated crystal. Moreover, the (Naf/Lys)6, (Naf/Chi)6, and (Naf/Lys/Naf/Chi)2 coatings all inhibit *S. aureus* growth by 99.99%, compared to 57.1% for the Nafion coated crystal. For reference, (Naf/Chi)3 and (Naf/Lys/Naf/Chi)1 coatings reduce *E. coli* by 74.6% and 88.5%, respectively, and inhibit *S. aureus* growth by 99.9% and 83.4%, respectively (Figure S2).

**Figure 4.** Reduction (in log scale) of the population of *E. coli* and *S. aureus* cultures exposed to (Naf/Lys)6, (Naf/Chi)6, (Naf/Lys/Naf/Chi)2, and Nafion coated QCM-D crystals. (For each type of coating five crystals were tested and the average values are shown in this figure).

The photos displayed in Figure 5 clearly depict the significant advantages of the nanocoatings disclosed here. When *E. coli* and *S. aureus* cultures are exposed to uncoated QCM-D crystals (blank samples), no antimicrobial effect is observed. In contrast, when the same *E. coli* and *S. aureus* cultures are exposed to, otherwise identical, (Naf/Lys)6, (Naf/Chi)6, and (Naf/Lys/Naf/Chi)2 coated QCM-D crystals, virtually all bacteria appear to be eliminated. Such a remarkable antimicrobial performance might reflect the synergistic effect of the contact-killing behaviour of lysozyme and chitosan, the bacteria-repelling behaviour of Nafion combined with contributions arising from surface roughness and wettability. Standard agar diffusion tests indicated the absence of bacteria inhibition zones around the coated crystals, confirming that the building blocks of the LbL assemblies are firmly fixed and do not diffuse into the agar.

**Figure 5.** Photos of the petri dishes containing *E. coli* (upper photos) and *S. aureus* (lower photos) cultures. The petri dishes on the left have been exposed to uncoated QCM-D crystals, while those on the right have been exposed to (Naf/Lys)6 coated discs, under otherwise identical conditions.

The antibacterial performance of lysozyme stems from its enzymatic activity to cleave 1,4 beta-linkages between N-acetylmuramic acid and N-acetyl-D-glucosamine that triggers peptidoglycan hydrolysis and, ultimately, cell lysis. Evidently, this mechanism is less effective for Gram-negative bacteria whose protective outer membranes prevents access to the enzyme [35]. Adjusting the pH of the lysozyme solution to 4 and 9 decreases the efficiency of (Naf/Lys)6 against *E. coli* but leaves its ability to combat *S. aureus* essentially intact (Figure S2). It is noted that the lytic activity of lysozyme has been found to exhibit a maximum at pH 6.2 over a broad range of ionic strengths [36].

This supreme antimicrobial performance of lysozyme is only encountered on the condition that it's secondary structure is well preserved [37]. The CD spectrum of lysozyme shown in Figure 6 is dominated by two negative bands at 208 and 222 nm and suggests the presence of 71% α-helix and 10% β-sheet, consistent with data published previously [38]. Upon mixing with Nafion at fLys = 0.9 and fLys = 0.3, the secondary structure of lysozyme is modified to a small extent, given that the α-helix content decreases to 64% and 50% and the β-sheet content increases to 16% and 23%, respectively.

**Figure 6.** Circular dichroism (CD) spectra of aqueous solutions containing: Lysozyme and lysozyme/Nafion mixtures at fLys = 0.9 and fLys = 0.3 at 25 ◦C.

#### **4. Discussion**

A number of LbL assemblies employing lysozyme as a key antimicrobial ingredient has been reported in the literature. To that end, LbL assemblies comprising lysozyme and pectin deposited on cellulose mats were shown to induce a clear zone of bacterial inhibition, the more so when lysozyme is present at the outermost layer [39]. In a similar manner, lysozyme and gold nanoparticles were deposited on cellulose nanofibrous mats improving their antimicrobial performance against Gram-positive as well as Gram-negative model bacteria [40]. In addition, mechanically robust LbL membranes based on DNA-SWNT and lysozyme-SWNT (where SWNT stands for single-walled carbon nanotubes) were shown to exhibit long-term activity against *S. aureus* and *Micrococcus lysodeikticus*, however only when lysozyme is found on the outermost layer [41]. In our study, the effect of the outermost layer is minimal, given that both (Naf/Lys)3 and (Naf/Lys)3.5 coated crystals (with lysozyme at pH = 9) exhibit identical antimicrobial performance (Figure S2). Regarding the (Naf/Lys)6, it appears that strong Nafion-protein electrostatic interactions keep lysozyme firmly fixed to the LbL assembly without compromising its secondary structure and while allowing its active sites to remain accessible by the bacteria.

The antimicrobial performance of chitosan critically depends upon intrinsic characteristics such as molecular weight and the degree of deacetylation and charge density, as well as external properties such as pH (enhanced activity at low pH), temperature (enhanced activity at higher temperatures) and ionic strength [42]. In general, three modes of action have been identified for chitosan: One arising from its polycationic nature that facilitates lysis of the microbial cell membranes, the second associated with its strong metal-chelating properties that deprives bacterial cells of essential nutrients, and the third stemming from its DNA-binding ability that inhibits protein and mRNA synthesis from the bacterial cells [14].

In analogy to lysozyme, a number of LbL antimicrobial membranes rely on the electrostatic immobilisation of chitosan. To that end, it has been reported that LbL assemblies based on chitosan/hyaluronic acid on PET show excellent activity against *E. coli* [43]. LbL assemblies based on chitosan/polyanionic lentinan sulphate on polyurethane showed 58% improvement against the opportunistic pathogen *P. aeruginosa* [44]. Moreover, LbL assemblies of chitosan and alginates on cotton samples were effective against *S. aureus* and *Klebsiella pneumonia* [45], while chitosan/lignosulphonates multilayers on cellulose fibres suppress *E. coli* growth up to 97% [46].

By comparison, only a limited body of work is centred around the antimicrobial activity of Nafion, even though Nafion coated stainless steel disks were shown to inhibit *E. coli* adhesion [22]. It is widely accepted that repulsive forces between the negatively charged Nafion and the similarly charged bacteria gives rise to a bacterial exclusion zone (EZ) at the Nafion-water interface. A recent study employed confocal laser scanning microscope to show that the EZ is a non-equilibrium phenomenon that diminishes with time, as van der Waals and acid-base forces start to dominate the Nafion-bacteria interactions [17]. Because of those forces, a significant number of cells are able to break the EZ barrier after 48 h of incubation, compared to rare and sporadic cell attachment after 24 h of incubation.

Interestingly, ζ for (Naf/Lys)6, (Naf/Chi)6, and (Naf/Lys/Naf/Chi)2 deposited on the aluminium foil was found +2.8 mV, −22.2 mV, and −30.7 mV, respectively, compared to −52.1 mV for the Nafion coated aluminium foil surface and similar trends were recorded for the multilayers deposited on a polypropylene substrate. Although, Nafion's negative charges have been overcompensated in (Naf/Lys)6 coatings, they exhibit exceptional antimicrobial behaviour that cannot be attributed solely to EZ effects. In that sense, our work demonstrates for the first time that Nafion based compositions exhibit supreme antimicrobial behaviour that goes beyond the concept of the EZ.

#### **5. Conclusions**

In conclusion, we report a systematic study on the structure-property relationships of a new series of antimicrobial coatings comprising Nafion, chitosan and lysozyme. The chemical composition, the topological characteristics and the wetting performance, are all important parameters that define

the superior antimicrobial behaviour of the ultrathin films. Our study provides solid evidence that coupling between Nafion and conventional antimicrobial agents can generate highly effective platform coatings to combat the colonization and spread of bacteria.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/9/11/1563/s1, Figure S1: (i) Photos of UV-vis polystyrene cuvettes: Uncoated (A), (Naf/Lys)6 (B), (Naf/Chi)6 (C), (Naf/Lys/Naf/Chi)2 (D), demonstrating their transparency levels. (ii) UV-vis spectra of polystyrene cuvettes coated with (Naf/Lys)6, (Naf/Chi)6, and (Naf/Lys/Naf/Chi)2. Figure S2: Reduction of the population of E. coli and S. aureus cultures exposed to QCM-D crystals coated with (Naf/Lys)6 (pH = 4), (Naf/Lys)6 (pH = 9), (Naf/Lys)3 (pH = 9), (Naf/Lys)3.5 (pH = 9), (Naf/Lys/Naf/Chi)1, and (Naf/Chi)3.

**Author Contributions:** Conceptualization, A.K.; methodology, E.N.G., M.J.K., A.K.; resources, M.J.K., A.K., S.G.Y.; data curation, e.g., C.W., E.B., D.F., A.V.S.P.; writing-original draft preparation, A.K., M.J.K.; writing-review and editing, A.K.; supervision, A.K., M.J.K., S.G.Y.; project administration, A.K., M.J.K., S.G.Y.; revision: A.K., E.N.G.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Influence of Preparation Procedure on Physicochemical and Antibacterial Properties of Titanate Nanotubes Modified with Silver**

#### **Manu Jose 1, Paulina Sienkiewicz 1, Karolina Szyma ´nska 2, Dominika Darowna 1, Dariusz Moszy ´nski 1, Zofia Lendzion-Bielu ´n 1, Kacper Szyma ´nski <sup>1</sup> and Sylwia Mozia 1,\***


Received: 7 May 2019; Accepted: 20 May 2019; Published: 23 May 2019

**Abstract:** Silver nanoparticles (NPs) are effective antibacterial agents; however, aggregation of NPs and uncontrolled release of Ag<sup>+</sup> affect their efficiency and may pose a risk to the environment. To overcome these disadvantages, immobilization of Ag onto titanate nanotubes (TNTs) was investigated. This paper describes the physicochemical and antibacterial properties of silver incorporated titanate nanotubes (Ag/TNTs) prepared using five procedures and containing different Ag amounts (0.11–30.85 wt.%). The methods were (i) sol-gel followed by a hydrothermal process; (ii) photodeposition under ambient conditions; (iii) photodeposition under an inert atmosphere; (iv) NaBH4 reduction; and (v) electroless deposition after activation of TNTs with Sn2+. Depending on the synthesis procedure, the presence of metallic Ag NPs, AgO or AgCl was observed. The electroless method led to an additional deposition of SnO2 NPs. The antibacterial properties of Ag/TNTs were analyzed as a function of Ag content and released against *Escherichia coli* and *Staphylococcus epidermidis*. The best bactericidal properties exhibited Ag/TNTs prepared through the photodeposition process due to the higher interaction of exposed Ag NPs with bacteria. An increase of Ag loading resulted in improvement of antibacterial activity of Ag/TNTs although no direct correlation between silver content or release and inhibition of bacterial growth was found.

**Keywords:** titanate nanotubes; TNT; Ag; antibacterial; *Escherichia coli*; *Staphylococcus epidermidis*

#### **1. Introduction**

The presence of large concentrations of microorganisms such as bacteria, protozoans, and viruses in the majority of the world's water resources limits the use of a major portion of them as drinking water. According to the World Health Organization (WHO), 80% of diseases arise due to the consumption of contaminated water [1]. Hence, the development of new efficient technologies for the removal of harmful microorganisms from water will require much attention in the coming decades. Recently, nanomaterials with superior photoactivity, high surface-to-volume ratios, antibacterial properties, and good hydrophilicity have been explored for the inactivation of pathogenic microorganisms. Among them, silver NPs have gained much attention for their powerful antibacterial properties [2,3] which find application in wound dressing [4], textiles [5], and self-sterilizing surfaces

in food and pharmaceutical production [6]. The Ag NPs show good antimicrobial properties towards a broad spectrum of bacterial and fungal species including antibiotic-restraint strains [7,8].

Even though Ag NPs are known as promising biocidal agents due their distinctive physicochemical properties, their wide applicability is limited mainly due to the uncontrolled release of Ag<sup>+</sup> ions from the Ag NPs, which was found to have many toxic effects on the environment [9]. The high activity of Ag NPs arises from their ultra-small size and high mobility [8]. However, mobile Ag NPs are found to aggregate easily in the medium that alters their cytotoxicity, [10] and therefore, numerous studies have been conducted to improve their dispersion. To overcome these disadvantages of Ag NPs, their immobilization onto various supporting materials such as metal oxides, activated carbon, graphene oxide, polymers, etc., have been investigated [11]. Among various hybrid NPs, TiO2 and TNTs modified with Ag NPs are being researched for their excellent antibacterial properties both in the presence and absence of light. The modification of TiO2/TNTs with Ag NPs results in changes in the physicochemical characteristics such as size, shape, stability, and oxidation state of Ag NPs, which results in enhanced antibacterial, photocatalytic, and catalytic activities [12–16]. Wang et al. [17] observed improved and long-lasting antibacterial activity for Ag–polydopamine–TiO2 nanotube composites which was attributed to the tethering of Ag NPs onto TNTs by polydopamine layers.

Important factors which affect the bactericidal activity of Ag NPs are their size, shape, surface functionalization, and stability [11]. The antibacterial properties of Ag NPs are found to increase with a decrease in their diameter, and the direct interaction of Ag NPs with bacteria mainly occurs when the diameters are around 1–10 nm [18,19]. Kubacka et al. [20] synthesized Ag/TiO2 nanocomposites through impregnation and photodeposition methods with various Ag contents and studied their photocatalytic disinfection ability against *Escherichia coli*. They observed that below 1 wt.% of Ag, the disinfection activity of the samples obtained by the two methods was comparable while at higher silver content, the photo-deposited samples displayed improved performance. Also, they found that the presence of Ag NPs helps to improve the adhesion of bacteria onto nanocomposite surfaces and the Ag lixiviation can be controlled by optimizing the amount of Ag in the nanocomposite. Similar results were reported by other researchers [7,8]. It was also found that the Ag NPs changed the surface characteristics of TiO2 such as the point of zero charge (PZC), which can influence bacterial attraction to the Ag/TiO2 surface [20]. Es-Souni et al. [21] prepared Ag/TiO2 nanocomposite coatings through a sol-gel approach and found that bactericidal actions rely on Ag<sup>+</sup> ion release, Ag NPs size, and hydrophilicity of the nanocomposites. Keleher et al. [22] observed higher antibacterial activity for Ag/TiO2 than that of Ag metal, which was ascribed to the more available surface for Ag<sup>+</sup> ion release in solution. Sotiriou et al. [23] studied the leaching of Ag<sup>+</sup> ions from Ag/SiO2 nanocomposites and reported that the amount of the released Ag<sup>+</sup> corresponded to the dissolution of 1–2 silver oxide monolayers present on the surface of Ag NPs, depending on their size. The authors also found that the reduction of silver oxide to metallic silver resulted in a significant minimization of Ag<sup>+</sup> ion leaching which was found to decrease the antibacterial activity against *E. coli*. The investigation on TNT ions exchanged with various metal ions presented by Rónavári et al. [24] revealed that only TNTs containing silver exhibited potential antibacterial and antifungal properties against different microbial species, which was ascribed to the release of ionic Ag<sup>+</sup> to the surrounding solution. In other research [14], it was observed that the controlled release of Ag<sup>+</sup> from Ag/TNTs nanocomposite through diffusion and osmosis effects provided extended antibacterial activities of this material. Rodríguez-González et al. [16], based on their research on antifungal properties of Ag/TNTs, concluded that due to the nanotubular morphology, the TNTs could easily damage cell walls and accelerate vacuolation and invagination which results in inactivation of fungi.

Silver NPs are predominantly synthesized from silver nitrate (AgNO3) and silver acetate (CH3COOAg) as the precursors [14–16,25]. Various reported methods for the preparation of Ag/TiO2 nanocomposites are photoreduction [13], sol-gel [26], chemical reduction [27], template induced and solvothermal [28] methods. The Ag/TNTs are obtained through photoreduction [29], chemical reduction [30], ion exchange followed by calcination [31], microwave-assisted methods [32], and hydrothermal processes [33]. The Ag/TNTs nanocomposites are mainly used as antibacterial nanomaterial [14], visible light photocatalyst [34], nanofiller in modified polymeric membranes [35], and as a multicolor photochromic material [36]. During the preparation of Ag-modified nanocomposites through a photoreduction approach, the physicochemical properties like size, uniformity, and density of photodeposited Ag NPs are depended upon solvent, silver precursor concentration, reaction atmosphere (ambient or inert), irradiation wavelength, and time or type of support (e.g., TiO2, TNTs, etc.) used [13,37,38]. Ma et al. [30] reported that Ag NPs in the metallic form with a size range of 3–10 nm could be deposited onto TNTs by NaBH4 reduction, whereas the reduction reaction carried out without the use of TNTs resulted in the formation of highly agglomerated Ag NPs with a size of 20–50 nm. The stability of the Ag NPs on TNTs was attributed to the strong bonding interaction between Ag NPs and the oxygen atoms of TNTs [30]. Priya et al. [27] demonstrated the synthesis of Ag2O/Ag0-loaded TiO2 NPs by an electroless coating technique in which Ag<sup>+</sup> ions were reduced onto the TiO2 using Sn2+. Lai et al. [39] produced Ag/TiO2 nanotubes by hydrothermal treatment of the sol-gel-processed Ag/TiO2 NPs. The reduction of Ag<sup>+</sup> ions to metallic Ag occurred during the thermal treatment step of the sol-gel processed TiO2 nanopowder. After the hydrothermal treatment, Ag NPs of sizes 4–8 nm were found to be well dispersed on the exterior of the nanotube surface with a small fraction of Ag NPs encapsulated in the interior of the TiO2 nanotubes [39]. To the best of our knowledge, there are no reports on the application of the electroless technique to fabricate Ag-modified TNTs.

The physicochemical and antibacterial properties of Ag-modified TNTs are expected to vary with the adopted synthesis method. In view of this, the present study is focused on the evaluation of the influence of the preparation procedure on the properties and stability of Ag-modified titanate nanotubes (Ag/TNTs). Twelve types of Ag/TNTs were prepared with different Ag incorporation approaches and Ag contents. The synthetic procedures included (i) sol-gel followed by a hydrothermal process; (ii) photodeposition under ambient conditions; (iii) photodeposition under an inert atmosphere; (iv) NaBH4 reduction; and (v) an electroless deposition process after activation of the TNTs' surface with various amounts of Sn2<sup>+</sup> ions. The physicochemical properties of the hybrid Ag/TNTs were examined and discussed in detail. Moreover, the antibacterial performance of the composites against both Gram-positive and Gram-negative bacteria were evaluated under dark conditions.

#### **2. Materials and Methods**

#### *2.1. Materials*

Titanium(IV) isopropoxide (TTIP, Sigma–Aldrich, St. Louis, MO, USA, 97%) and anatase TiO2 powder were purchased from Sigma–Aldrich Chemicals (St. Louis, MO, USA). HCl (35–38 wt.%), H2SO4 (96 wt.%), AgNO3, SnCl2, ammonia solution (25%), NaOH, Na2HPO4, and KH2PO4 were purchased from Avantor Performance Materials (Gliwice, Poland). NaBH4 was supplied by Merck, (Darmstadt, Germany). 2-propanol, KCl, and NaCl were provided by Chempur (Piekary Sl ˛ ´ askie, Poland).

Microbiological tests were carried out using Plate Count Agar (PCA) and Brain Heart Infusion (BHI) Agar (BIOMAXIMA, Lublin, Poland). Gram-negative *Escherichia coli* (strain K12, ATCC 29425, Manassas, VA, USA) and Gram-positive *Staphylococcus epidermidis* (ATCC 49461, Manassas, VA, USA) were used as model microorganisms. The initial concentration of bacteria suspension was set at 0.5 using McFarland scale (McFarland standards, bioMérieux, Marcy-l'Étoile, France).

In all experiments, pure (deionized) water (type 2, 0.066 μS cm<sup>−</sup>1) from Elix 3 (Millipore, Burlington, MA, USA) was used, unless otherwise stated.

#### *2.2. Preparation of TNTs*

Titanate nanotubes were prepared by employing alkaline hydrothermal treatment of anatase TiO2 powder. Initially, TiO2 (2 g) was ultrasonicated with 60 mL of 10 M NaOH solution for 1 h at room temperature to obtain a homogeneous dispersion. The mixture was then transferred to a Teflon-lined stainless-steel autoclave and then heated at 140 ◦C for 24 h. After being cooled down to room temperature, the product was first washed with 2 L of 0.1 M HCl and then with deionized water until the conductivity of the filtrate became ~1 <sup>μ</sup>S·cm<sup>−</sup>1. Finally, the white product was dried at 80 ◦<sup>C</sup> in an oven for 12 h and stored.

#### *2.3. Preparation of Hybrid Ag*/*TNTs*

#### 2.3.1. Preparation of Ag/TNTs by Sol-Gel Combined with Hydrothermal Process

Silver-modified nanocrystalline TiO2 powders were synthesized by sol-gel process. First, 5.53 g of TTIP was dissolved in 100 mL of 2-propanol. A second solution was prepared by dissolving 50.96 and 254.8 mg of AgNO3 (corresponding to Ag: Ti atomic ratios of 0.01 and 0.05, respectively) in a mixture of deionized water (50 mL) and 2-propanol (100 mL). Both solutions were sealed immediately and stirred thoroughly using the magnetic stirrer. The water part of the solution was then added drop-wise to the alkoxide part under continuous magnetic stirring. After the complete addition of the water part of the solution, the resulting suspension was stirred for 4 h before drying in an oven at 80 ◦C for the complete removal of residual water and the solvent. The dried powder was then ground well using a mortar and pestle and then calcined in a muffle furnace at 500 ◦C for 2 h at the heating rate of 5 ◦C min−<sup>1</sup> for the crystallization of amorphous TiO2. Such obtained Ag/TiO2 powders containing various amount of Ag were then hydrothermally treated according to the procedure described above (see Section 2.2). The resulting hybrid products were denoted as Ag/TNT-1\_SH and Ag/TNT-5\_SH, where the "1" and "5" represented the initially used atomic ratio of Ag/Ti for the sol-gel synthesis of Ag/TiO2 (Table 1).


**Table 1.** Summarizes the applied methods of Ag/titanate nanotube (TNT) synthesis, the corresponding concentrations of AgNO3, and samples nomenclature.

#### 2.3.2. Preparation of Ag/TNTs by Photodeposition

In this technique, 0.5 g of TNTs were dispersed into 50 mL of AgNO3 solution (2.5 and 100 mM) with magnetic stirring (250 rpm) for 2 h in a glass reactor. The processes were carried out under either an ambient or inert (Ar) atmosphere. Afterward, the slurry was irradiated with a low-pressure mercury vapor lamp (TNN 15/32, Heraeus Noblelight GmbH, 15 W, λmax = 254 nm) for 2 h with continuous stirring. The suspension was then collected by centrifugation and subsequently washed several times with deionized water for the complete removal of excess of Ag<sup>+</sup> ions. Finally, the products were dried at 80 ◦C in an oven for 12 h and stored. The samples were denoted later as Ag/TNT-2.5\_AM and Ag/TNT-100\_AM for ambient atmosphere, or Ag/TNT-2.5\_IN and Ag/TNT-100\_IN for inert atmosphere, where the numbers represented the concentration of AgNO3 solution (Table 1).

#### 2.3.3. Preparation of Ag/TNTs by NaBH4 Reduction

The Ag<sup>+</sup> ions were reduced onto TNTs according to the procedure described elsewhere [30]. In a typical synthesis, 0.5 g of TNTs were dispersed into 50 mL of AgNO3 solution (2.5 and 100 mM) and magnetically stirred for 2 h. The nanotubes were then separated from the solution by centrifugation at 3000 rpm. An ice-cold solution of NaBH4 (5 mL, 0.1 M) was added drop-wise to the centrifuged sample. The product was then collected and washed with deionized water before drying at 80 ◦C in an

oven for 12 h. The samples were denoted later as Ag/TNT-2.5\_NB and Ag/TNT-100\_NB, where the numbers represented the concentration of AgNO3 solution (Table 1).

#### 2.3.4. Preparation of Ag/TNTs by Electroless Reduction

First, 0.5 g of TNTs were dispersed in 30 mL of deionized water. Then, a second solution was prepared by dissolving SnCl2 (either 0.1 g or 1 g) in 20 mL of 0.2 M HCl. The two solutions were then mixed and stirred for 2 h at room temperature to obtain the surface sensitized TNTs. The suspension was subsequently centrifuged at 3000 rpm and washed three times with deionized water. The residue was then transferred to 50 mL of AgNO3 solution (2.5 and 100 mM) and the resulting solution was made alkaline by the addition of 5 drops of aqueous NH3 solution. The suspension was stirred for 1 h and the product was separated by centrifugation at 3000 rpm. After washing with deionized water, the nanomaterial was dried at 80 ◦C in an oven for 12 h. The samples were denoted later as Ag/TNT-2.5\_EL (0.1) and Ag/TNT-100\_EL (0.1) (for the samples processed using 0.1 g SnCl2), and Ag/TNT-2.5\_EL (1) and Ag/TNT-100\_EL (1) (for the samples processed using 1 g SnCl2), where the numbers 2.5 and 100 represented the concentration of AgNO3 solution (Table 1).

#### *2.4. Characterization Methods*

The morphological analysis of pure TNTs and Ag/TNTs was carried out using a transmission electron microscope (TEM), FEI Tecnai F20. The elemental composition of the samples was studied with the usage of energy dispersive X-ray spectroscopy (EDS). The samples were prepared by sonication in ethanol followed by adding a drop of the suspension on a carbon-coated copper grid (300 mesh). The phase composition of the pure TNTs and Ag/TNTs was determined based on the X-ray diffraction (XRD) method (PANalytical Empyrean X-ray diffractometer) using CuKα radiation (λ = 1.54056 Å). Raman spectra were recorded with a 532 nm laser line (Elaser = 1.58eV) with a Renishaw in Via Raman micro-spectrometer. The isoelectric point (IEP) of the Ag/TNTs nanocomposites was measured using Zetasizer Nano-ZS (Malvern Instruments Ltd. Malvern, UK) equipped with a Multi-Purpose Titrator MPT-2 and a degasser. The samples were dispersed in ultrapure water and the pH was adjusted using HCl and NaOH solutions.

The composition of the Ag/TNTs surface was analyzed with use of the X-ray photoelectron spectroscopy (XPS). Measurements were conducted with Al *Ka* (h = 1486.6 eV) radiation in a Prevac system equipped with Scienta SES 2002 electron energy analyzer operating at constant transmission energy (*Ep* = 50 eV). The spectrometer was calibrated using the following photoemission lines (with reference to the Fermi level): EB Ag 3d5/<sup>2</sup> = 368.3 eV and EB Au 4f7/<sup>2</sup> = 84.0 eV. The analysis chamber was evacuated to the pressure below 1·10−<sup>9</sup> mbar. A powdered sample of the material was placed on a stainless-steel sample holder. The quantitative analysis of the surface composition was done on the basis of the peak area intensities using the sensitivity factor approach and assuming homogeneous composition of the surface layer.

Inductively coupled plasma optical emission spectrometry (ICP-OES) analysis was carried out using an Optima 5300DV spectrometer (Perkin Elmer, Waltham, MA, USA). To determine the real load of Ag in the Ag/TNTs, the samples were prepared by dissolution in a hot solution of (NH4)2SO4 in concentrated H2SO4. After the solution cooled down, it was diluted with water.

To determine the release kinetics of Ag<sup>+</sup> ions from the different Ag/TNTs for seven days, 0.1 g of each nanocomposite was dispersed into 100 mL of deionized water and placed in a digital shaking water bath maintained at 30 ◦C. A defined number of samples was withdrawn after 1, 3, and 7 days and separated through a 0.2 μm filter. The concentration of Ag in the filtrate was analyzed using ICP-OES spectrometer and the given values were a mean from three repetitions.

#### *2.5. Microbiological Study*

#### 2.5.1. Preparation of Culture Medium

First, the BHI and PCA solutions were prepared according to the instructions given by the manufacturer. Next, the Petri dishes were filled with an adequate solution and left to be solidified. Finally, the prepared agar plates were sterilized under UVC light for 20 min and then dried in the incubator for 3 days.

A NaCl solution was prepared by dissolving 8.5 g NaCl in 1 L of distilled water and then autoclaved.

Phosphate buffered saline (PBS, pH 7.2) was obtained by dissolution of 8 g NaCl, 0.2 g KCl, 1.44 g Na2HPO4, and 0.24 g of KH2PO4 in 1 L of distilled water, and the pH was adjusted using HCl. Before application, the solution was sterilized by autoclaving.

#### 2.5.2. Antibacterial Study of Nanomaterials

A series of glass bottles filled with 100 mL of nanomaterial suspension (20 mg L−1) in NaCl or PBS solutions containing *E. coli* or *S. epidermidis*, respectively, were prepared. The number of bacteria was set at 0.5 according to the McFarland scale. The control sample was prepared in the same way, but without addition of NPs. The bottles were incubated for 24 h at 37 ◦C with continuous stirring at 250 rpm. After that, the bacteria were counted using the serial decimal dilutions in NaCl and PBS solutions, respectively. 0.3 mL of a suitable diluted solution was put in the middle of a plate containing PCA or BHI and spread using a spreader. Three repetitions for each dilution were prepared. The plates with bacteria were incubated at 37 ◦C for 24 h. After that, the visible colonies of bacteria on agar plates were calculated by the counter (LKB 2002, POL-EKO, Wodzisław Sl ˛ ´ aski, Poland). The average colony forming unit (CFU) per mL values were evaluated according to Equation (1):

$$\text{CFU/mL} = \frac{N \times Y}{Z} \tag{1}$$

where: *N*—number of bacteria colonies visible on the Petri dish, *Y*—total dilution factor, and *Z*—volume of bacteria suspension put on the agar plate (0.3 mL).

The log reduction of bacterial growth was determined with reference to the blank sample using Equation (2):

$$\text{log reduction} = \log\left(\frac{A}{B}\right) \tag{2}$$

where: *A*—number of bacteria determined in control sample, i.e., without addition of NPs (CFU/mL), *B*—number of bacteria determined in the presence of NPs (CFU/mL).

#### **3. Results and Discussion**

#### *3.1. ICP Compositional Analysis*

The amount of Ag in the different Ag/TNTs nanocomposites was evaluated based on ICP-OES analysis, and the values are presented in Table 2.

**Table 2.** Amount of Ag in different Ag/TNTs nanocomposites measured by ICP-OES.


For all preparation procedures, the weight fraction of Ag in the nanocomposite was found to be increased with an increase in the initial concentration of AgNO3. A minimum Ag content was observed for Ag/TNT-1\_SH, and a maximum Ag loading was found in the case of Ag/TNT-100\_EL (1). The amount of Ag was almost comparable for photodeposition (both inert and ambient atmosphere), and NaBH4 reduction processes, with a moderately higher Ag loading observed for the latter approach. Silver/TNTs processed using an electroless method indicate the presence of Sn, originating from the SnCl2 used as a reducing agent. The amount of Sn in the case of Ag/TNT-2.5\_EL was found to be higher compared to Ag/TNT-100\_EL samples for both 0.1 g and 1 g SnCl2 loading, which confirms the role of Sn2<sup>+</sup> ions in the reduction of Ag<sup>+</sup> to Ag0 on the TNTs.

#### *3.2. Morphological Analysis*

Transmission electron microscopy was employed in order to determine the morphology of the Ag/TNTs and the size distribution of Ag NPs deposited on TNTs. The results are presented in Figure 1.

**Figure 1.** TEM images of Ag/TNTs and size distribution (inserted) of Ag nanoparticles (NPs) on TNTs: (**a**) Ag/TNT-1\_SH, (**b**) Ag/TNT-5\_SH, (**c**) Ag/TNT-2.5\_AM, (**d**) Ag/TNT-100\_AM, (**e**) Ag/TNT-2.5\_IN, (**f**) Ag/TNT-100\_IN, (**g**) Ag/TNT-2.5\_NB, (**h**) Ag/TNT-100\_NB, (**i**) Ag/TNT-2.5\_EL (0.1), (**j**) Ag/TNT-100\_EL (0.1), (**k**) Ag/TNT-2.5\_EL (1), (**l**) Ag/TNT-100\_EL (1).

The TEM images confirmed the formation of open-ended nanotubes with lengths in the range of ~50–200 nm and diameters of ~5–10 nm for all the Ag/TNTs. The sol-gel-assisted hydrothermal method of TNT modification did not result in the formation of Ag NPs within the used concentrations of the AgNO3 solution. In the case of the photodeposition approach, the presence of the Ag NPs was observed when higher concentrations of AgNO3 solution were applied, regardless of the atmosphere of the reaction. The ICP-OES compositional measurement (Table 2) also indicated low content of Ag for Ag/TNT-1\_SH, Ag/TNT-5\_SH, Ag/TNT-2.5\_AM, and Ag/TNT-2.5\_IN. Electroless and NaBH4 reduction methods led to the creation of Ag NPs in the case of both lower and higher AgNO3 solution concentration. The particle size distribution data in the insets in Figure 1d,f–l indicate the presence of NPs with sizes between 1–10 nm on TNTs. The sizes ranging from 2 to 10 nm, 1 to 8 nm, 1 to 9 nm and 2 to 4 nm corresponded to Ag NPs present in Ag/TNT-100\_AM, Ag/TNT-100\_IN, Ag/TNT-2.5\_NB, Ag/TNT-100\_NB, respectively. In the case of Ag/TNT-2.5\_EL (0.1), Ag/TNT-100\_EL (0.1), Ag/TNT-2.5\_EL (1), and Ag/TNT-100\_EL (1), the determined particle size (2 to 10 nm, 2 to 7 nm, 2 to 8 nm, and 2 to 10 nm, respectively) refers to both Ag and SnO2 identified on the surface of the samples processed by the electroless method.

The TEM images also demonstrate the difference between physical and chemical Ag deposition processes (Supplementary Materials, Figure S1). The photoreduction approach led to creation of Ag NPs exclusively on the outer surface of TNTs, whereas chemical reduction techniques introduced Ag NPs both on the outer surface and inside the TNTs. This is because during UV irradiation, the outer region of the TNTs was more exposed to the action of the radiation, and hence, Ag<sup>+</sup> ion reduction preferably occurred in this area as observed in Figure S1a. For chemical reduction, the reducing agent had equal accessibility for the exterior and interior surfaces of the TNTs [40]. Hence, the Ag NPs could be deposited on either side of the TNTs as observed in Figure S1b,c. From particle size distribution analysis, it was observed that Ag/TNT-100\_NB (inset of Figure 1h) contained almost exclusively Ag NPs with a size of 2 nm. The small size of the Ag NPs helped in the modification of both the inner and outer sides of the TNTs for this sample. However, for electroless deposition process, both surfaces of TNTs were found to be covered with an excess of NPs. This was due to the coexistence of both Ag and SnO2 NPs with similar size making them indistinguishable from each other.

Based on TEM-EDS analysis (Figure S2), it was found that Ag/TNT-2.5\_AM (Figure S2a) was characterized by a homogenous distribution of Ag all over the TNTs, despite the absence of NPs, as was found from Figure 1c. This suggests that Ag was built into the structure of nanotubes. In the case of Ag/TNT-5\_SH (Figure S2b), an additional signal corresponding to Cl can be noticed, which indicates the presence of AgCl, possibly formed during the acid (HCl) washing step after the hydrothermal treatment. The EDS elemental mapping of Ag/TNT-2.5\_EL (1), shown in Figure S2c, reveals the presence of Sn (red color) in addition to Ag (blue color). It can also be observed that Sn was uniformly distributed all over the TNTs in contrast to much smaller amounts of Ag concentrated in particular places. It is clear from the high resolution TEM (HRTEM) image of Ag/TNT-100\_EL (1) shown in Figure S3 that the Ag and SnO2 NPs were attached to the TNTs, and the interplanar spacing of NPs with distances 0.23 nm and 0.34 nm could be attributed to the (111) planes of Ag and (110) planes of SnO2, respectively. Hence, the application of the electroless process resulted in the modification of TNTs with both Ag and SnO2 NPs. From Table 2, it was observed that the Ag/TNT-2.5\_EL (0.1) and Ag/TNT-2.5\_EL (1) exhibited higher concentrations of Sn compared to Ag, therefore, it can be concluded that majority of NPs on the surface of TNTs for these materials are SnO2 NPs.

#### *3.3. XRD Analysis*

The structural evolution from pure TNTs to different Ag/TNTs was studied by XRD measurements (Figure 2).

Figure 2a (i) shows the XRD pattern of pure TNTs, and it exhibited peaks at *2*θ~9.7◦, 24.3◦, 27.8◦, and 48◦, which indicate the formation of layered titanates such as H2Ti2O5·H2O, H2Ti3O7 or HxTi2−x/4<sup>x</sup>/4O4 (x~0.7, -: vacancy) [31,41,42]. In comparison with pure TNTs, noticeable structural changes were observed in the XRD pattern of Ag/TNTs especially at higher Ag loading. The most perceptible change with Ag loading is the disappearance of the peak at *2*θ~9.7◦ that resulted from the (100) plane of the TNTs. This was because during the synthesis of Ag/TNTs, Ag<sup>+</sup> ions first diffused to the TNTs' surface and deposited as silver hydrate intermediate (Ag(OH)n(H2O)m), which upon dehydration with surface Ti–OH groups resulted in binding to the surface by sharing with surface oxygen atoms of the TiO6 octahedron layers in the (100) planes of the TNTs. As a result, the (100) planes of the TNTs underwent drastic deformation, and hence loss of its X-ray diffraction pattern with Ag loading [43]. In addition, the X-ray diffraction peak at *2*θ~24.3◦ was found to be distorted or weakened with Ag incorporation, which also resulted from the deformation at the surface of crystal lattice caused by the modification of the layered titanate structure [44]. This effect is maximal for Ag/TNT-100\_NB, Ag/TNT-100\_EL (0.1), Ag/TNT-2.5\_EL (1), and Ag/TNT-100\_EL (1). This is due to the presence of a high concentration of Ag and/or SnO2 NPs, as shown by ICP-OES (Table 2) and TEM (Figure 1) analysis.

**Figure 2.** (**a**) XRD pattern of (i) pure TNTs, (ii) Ag/TNT-1\_SH, (iii) Ag/TNT-5\_SH, (iv) Ag/TNT-2.5\_AM, (v) Ag/TNT-100\_AM, (vi) Ag/TNT-2.5\_IN, (vii) and Ag/TNT-100\_IN. (**b**) XRD pattern of (i) Ag/TNT-2.5\_NB, (ii) Ag/TNT-100\_NB, (iii) Ag/TNT-2.5\_EL (0.1), (iv) Ag/TNT-100\_EL (0.1), (v) Ag/TNT-2.5\_EL (1), and (vi) Ag/TNT-100\_EL (1).

The XRD patterns of Ag/TNTs processed from sol-gel-derived Ag modified anatase TiO2 (Ag/TNT-1\_SH and Ag/TNT-5\_SH) indicate the presence of AgCl (JCDPS 31-1238). The EDS mapping also suggested the presence of Cl in Ag/TNT-5\_SH (Figure S2b). Furthermore, a higher fraction of Ag was found to be transformed to AgCl for Ag/TNT-5\_SH compared to Ag/TNT-1\_SH, which was directly related to the Ag content of the Ag–TiO2 precursor used for hydrothermal treatment. The analysis of Ag/TNTs processed through electroless approach and NaBH4 reduction method revealed two spikes at 2θ~32◦ and 2θ~38◦, which can be assigned to silver oxide (AgO) (JCDPS 76-1489) and elemental Ag (JCDPS 04-0783), respectively.

#### *3.4. Raman Spectra Analysis*

Figure 3 shows the Raman spectra of the prepared TNTs and Ag/TNTs. Almost identical Raman vibration patterns were observed for both TNTs and Ag/TNTs, which consisted of mainly four very broad bands centered at 275, 450, 667, and 830 cm−<sup>1</sup> that could be assigned to the protonated type of TNTs [34,45–47]. The occurrence of the peak at 149 cm−<sup>1</sup> indicates the formation of a tetrahedron structure in the nanotubes with oxygen deficiencies [48]. The presence of Raman bands at 191, 275, 450, 667, 830, and 930 cm−<sup>1</sup> confirmed the formation of H2Ti3O7 nanotubes [41,49,50]. According to previous reports, the three Raman bands at around 270, 450, and 700 cm−<sup>1</sup> are assigned to the Ag symmetric modes of Ti–O–Ti vibrations of layered titanates [41]. The bands at 830 cm−<sup>1</sup> are assigned to the Ti–O–H symmetric stretching mode with short Ti–O distance [51], and the band at 930 cm−<sup>1</sup> is due to the four coordinate Ti–O vibrations in the titanate structure [52]. Raman spectra of Ag/TNT-100\_NB, Ag/TNT-100\_EL (0.1), Ag/TNT-2.5\_EL (1). and Ag/TNT-100\_EL (1) also showed a loss characteristic vibration of TNTs due to the higher loading of NPs (both Ag and SnO2). This indicate that higher metal loading drastically alters and/or diminishes the characteristic XRD and Raman vibration pattern of TNTs.

**Figure 3.** (**a**) Raman spectra of (i) pure TNTs, (ii) Ag/TNT-1\_SH, (iii) Ag/TNT-5\_SH, (iv) Ag/TNT-2.5\_AM, (v) Ag/TNT-100\_AM, (vi) Ag/TNT-2.5\_IN, and (vii) Ag/TNT-100\_IN. (**b**) Raman spectra of (i) Ag/TNT-2.5\_NB, (ii) Ag/TNT-100\_NB, (iii) Ag/TNT-2.5\_EL (0.1), (iv) Ag/TNT-100\_EL (0.1), (v) Ag/TNT-2.5\_EL (1), and (vi) Ag/TNT-100\_EL (1).

Different from the other samples, the spectra of Ag/TNTs prepared by sol-gel-assisted hydrothermal process show low intensity peaks at 396, 516, and 638 cm−<sup>1</sup> which correspond to B1g, A1g, and E2g vibration modes of anatase TiO2, respectively [53]. Additionally, the intensity of the band at 149 cm−<sup>1</sup> was found to be comparatively higher for this Ag/TNTs. This indicates that a small fraction of anatase TiO2 remained unconverted after the hydrothermal process for these samples [32,49]. However, the presence of anatase TiO2 could not be confirmed based on the XRD analysis (Figure 2) due to (i) the overlapping of the anatase TiO2 peak with that of hydrogen titanate or (ii) too low content of anatase to be detected by this method.

Based on the XRD and Raman analyses, it can be concluded that the hydrogen titanate structure of Ag/TNTs remains unaffected at lower Ag loading and is affected when the Ag loading is higher.

#### *3.5. XPS Analysis*

The surface concentration of elements was measured with application of XPS analysis. Silver/TNT samples prepared by various methods using higher AgNO3 content were selected (Table 3). The surface of all these samples consisted of titanium, oxygen, silver, and carbon atoms. The presence of Sn atoms was proven for the samples prepared by electroless reduction, Ag/TNT-100\_EL (0.1) and Ag/TNT-100\_EL (1). Sodium atoms were present on the surface of Ag/TNT-100\_NB sample, which was confirmed by the Na KLL Auger peak. Unfortunately, the XPS Na 1s and Auger Ti LMM lines overlapped. Therefore, the XPS Na 1s peak's intensity could not be resolved and the concentration of sodium atoms was not considered in the calculation of the surface composition of Ag/TNT-100\_NB sample.

**Table 3.** The surface concentration of elements identified by XPS on the surface of selected samples.


In Table 3, the surface composition of the samples analyzed by XPS is shown. The calculations were employed with the assumption that the spatial distribution of all elements identified in a near-surface region was homogeneous.

In general, the Ti:O ratio observed for the investigated samples was close to 1:3. Therefore, the surface structure of Ti–O compounds was considered as a TiO(OH)2 type rather than TiO2 type. In samples Ag/TNT-100\_EL (0.1) and Ag/TNT-100\_EL (1), the Ti:O ratio was even smaller due to a significant concentration of oxygen atoms being a part of Sn–O compounds. Since the depth of detection for XPS was around 5–10 nm [54,55], the Ag NPs present on both the outer and inner TNTs walls were analyzed (Figure S1). This led to the similar Ag concentrations for Ag/TNT-100\_AM, Ag/TNT-100\_IN, and Ag/TNT-100\_NB (2–3 at.%). A significant enrichment of the surface with silver atoms was observed for the samples obtained by the electroless reduction, especially in the sample Ag/TNT-100\_EL (1). Considering that the surface concentration of silver w presented in atomic percent, the direct correlation of these data with the silver concentrations measured by ICP-OES method is not possible. However, a general relation of these concentrations between the samples is kept (Figure S4).

The high-resolution XPS spectra were analyzed to elucidate the chemical state of silver atoms formed by different types of preparation methods. The XPS Ag 3d spectra (Figure S5) have virtually identical positions of the maximum and a very symmetric spectrum envelope. They contained two spin-orbit components: 3d5/<sup>2</sup> and 3d3/<sup>2</sup> located at the binding energy of 368.3 eV and 374.3 eV, respectively. The full-width at half maximum (FWHM) of these components was also identical for all samples and amounted to 1.7 eV. Therefore, it is concluded that the chemical state of silver in all analyzed samples was identical. The binding energy of the maximum of XPS 3d5/<sup>2</sup> component at 368.3 eV was characteristic for metallic silver [56]. The presence of silver oxides which can be considered in the context of Ag/TNTs materials should result in the XPS features located at the binding energy region between 367.3 eV and 368 eV.

#### *3.6. Ag*<sup>+</sup> *Ion Release Measurement*

One of the critical parameters which determines the antibacterial properties of nanocomposites modified with Ag NPs is Ag<sup>+</sup> ions' release ability [51]. Figure 4 shows the percentage of Ag<sup>+</sup> released from different Ag/TNTs for a period of one, three, and seven days of immersion in deionized water. The values were calculated with reference to the initial silver content in Ag/TNTs (Table 2).

**Figure 4.** The percentage of Ag<sup>+</sup> released from different Ag/TNTs.

During a period of seven days, a very small amount of Ag<sup>+</sup> (<5%) was observed to be leached from the prepared Ag/TNTs which indicated their high stability. The amount of Ag<sup>+</sup> leaching from the NPs was found to increase with the increase in immersion time. However, for some Ag/TNTs, the amount of Ag released after the third day was slightly lower than after the first day (e.g., Ag/TNT-5\_SH). This could be explained by reincorporation of the released Ag<sup>+</sup> ions in the TNTs' structure by the ion-exchange process. The minimum Ag<sup>+</sup> release percentage was observed for Ag/TNT-1\_SH and the maximum release exhibited the Ag/TNT-100\_EL (1) which was directly related to the Ag content in the nanocomposite (Table 2). For Ag/TNTs with lower Ag content (less than 5 wt.%, Table 2), Ag/TNT-5\_SH showed the highest Ag<sup>+</sup> release percentage. This sample was prepared by sol-gel followed by a hydrothermal process and was characterized by the presence of AgCl (almost insoluble in water, Ksp for AgCl at room temperature was 1.77 <sup>×</sup> <sup>10</sup>−<sup>10</sup> [57]) except from metallic Ag. In the case of this sample, no Ag NPs were identified on the surface (Figure 1b), although a uniform distribution of silver was confirmed by the TEM-EDS mapping (Figure S2b). Therefore, its low stability can be attributed to a dissolution of a silver layer covering the TNTs, a release of silver ions from AgCl or a removal of Ag<sup>+</sup> from titanate structure, where it was possibly built-in via the ion exchange process during the synthesis step. Except for Ag/TNT-100\_NB, in the cases of all the other Ag/TNTs processed using a higher concentration of AgNO3 (100 mM), the Ag<sup>+</sup> release percentage was directly related to the initial Ag content. The sample Ag/TNT-100\_NB exhibited the maximum stability for Ag<sup>+</sup> leakage during the seven days of measurement which could be attributed to the presence of Ag NPs not only on the outer surface of the TNTs but also inside the nanotubes, as observed in Figure S1c–f. Such a structure can be regarded as a container (TNT) housing Ag NPs and serving as protection which hindered Ag+'s release.

#### *3.7. Surface Charge Measurements*

On the basis of the zeta potential measurement as a function of pH, the isoelectric point (IEP) of the NPs was evaluated, and the corresponding values are summarized in Table 4.


**Table 4.** The isoelectric point (IEP) of the TNTs and Ag/TNTs.

Zeta potential describes the electrostatic interactions between the charged surface of a particle and the bulk of a liquid. There is an electrical double layer surrounding the particle. In the inner layer, the ions are strongly bound to the particle, and in the outer layer (diffuse layer) they are less firmly attached. The boundary between those regions is called the shear or slipping plane. The zeta potential is the potential between the dispersion medium and the stationary layer of the fluid attached to the particle. The zeta potential strongly depends on pH [58]. The pH at which the surface of the NPs has zero net charge is called the IEP [59]. When the pH is above the IEP, the surface sites become negatively charged either by adsorbing hydroxyl ions or by desorbing protons and vice versa. The value of the zeta potential of the nanomaterials could also affect their interactions with other species, such as microorganisms present in a liquid. The IEP of living *E. coli* and *S. epidermidis* are around 2.4 and 1.5–2.0, respectively [60,61]. Thus, the bacteria cells have positive charge only under very acidic conditions.

Both pure TNTs and Ag/TNTs show IEP in the range of ~3.1–3.9 and exhibit negative zeta potential above these values. The IEP of Ag/TNTs was found to be at higher pH than that of pure TNTs which could be attributed to (i) the presence of silver species, either in the form of Ag NPs or as Ag<sup>+</sup> ions replacing H<sup>+</sup> in TNTs structure, and (ii) the presence of SnO2 NPs in case of the samples prepared by the electroless method. A similar trend in zeta potential was observed when the TiO2 surface was modified with metals like Cu, Fe or Co [62]. However, for the same method of preparation, the Ag/TNTs with higher Ag content have slightly lower IEP values than that of Ag/TNTs with lower

Ag content. This is possibly because at a high Ag amount, the role of Ag NPs, which have an IEP of ~2.5 in deionized water, becomes more prominent [63].

#### *3.8. Antibacterial Properties of Ag*/*TNTs*

The antimicrobial properties of various Ag/TNTs were evaluated with reference to two types of bacteria: Gram-positive (*S. epidermidis*) and Gram-negative *(E. coli*). The results are presented in Figure 5. In general, for every type of Ag/TNTs synthesis approach, the samples obtained using lower AgNO3 amount (i.e. 2.5 mM AgNO3), and thus containing lower silver loading, were less active compared to the NPs synthesized with application of higher AgNO3 concentration (i.e. 100 mM AgNO3).

**Figure 5.** Antibacterial properties of Ag/TNTs towards *E. coli* and *S. epidermidis*.

Analyzing the results shown in Figure 5, it can also be observed that *S. epidermidis* was inactivated with lower efficiency than *E. coli*, regardless of the Ag/TNTs used. This phenomenon can be related with the composition of the bacteria cell wall. Gram-positive bacteria (i.e., *S. epidermidis*) have a relatively thick (20–80 nm), continuous cell wall, composed of peptidoglycan and covalently attached to other cell wall polymers (teichoic acids, polysaccharides, peptidoglycolipids) [64]. Such a structure results in a high rigidity of the bacterial cell as well as provides a very limited number of anchoring sites for Ag NPs and makes Ag NPs and ions difficult to penetrate [65,66]. On the other hand, Gram-negative bacteria possess a thin (5–10 nm) peptidoglycan layer, which in the case of *E. coli* is probably only a monolayer thick [64]. Outside the peptidoglycan layer, there is an outer membrane (7.5–10 nm). Despite the presence of many covalent bonds between polysaccharides and lipids in the outer membrane, the strength and rigidity of Gram-negative bacteria are low. Additionally, the presence of micro-channels known as porins responsible for bilateral transport of substances can facilitate transport of Ag<sup>+</sup> ions to the inner of bacteria cells [67].

The exact mechanism of action of nano-Ag as an antibacterial agent is not fully understood; however, in general, its antibacterial behavior is explained with the help of three approaches. First, due to the high binding affinity of Ag towards sulfur, Ag NPs attach to the bacterial cell membrane due to the presence of sulfur-containing proteins in it and cause many structural and functional changes to it [68,69]. Secondly, nano-Ag undergoes oxidation and the formed Ag<sup>+</sup> ions are released to the physiological environment which upon complexation with nucleic acids leads to DNA condensation and loss of replication ability [69]. Also, the Ag<sup>+</sup> ion has high affinity towards the thiol group of the cysteine residues of protein NADH dehydrogenases and causes disorder to the respiratory chain which finally leads to cell damage [70]. Thirdly, reactive oxygen species (ROS) like hydrogen peroxide (H2O2), hydroxyl radicals (OH•) or superoxide anions (•O2 −) formed in the presence of Ag NPs also contribute to the bactericidal actions [71,72].

Based on the results presented in Figure 5, it can be observed that the method of preparation of Ag/TNTs had some influence on the antimicrobial action of the nanomaterials. Ivask et al. [73] investigated the relation between size and antibacterial activity of Ag NPs and suggested that the mechanism of action is mainly dependent upon Ag NPs size. When the Ag NP's diameter is above 10 nm, the antibacterial activity depends on the released Ag<sup>+</sup> ions, whereas when the size of Ag NP's diameter is below 10 nm, the interaction of NPs with the bacterial cell wall becomes more important. In the present investigations, the prepared Ag/TNTs contain Ag NPs with a size <10 nm, and hence, it can be expected that the antibacterial action occurs through direct interaction of Ag NPs present in the hybrid structures with the bacterial cell walls. From the morphological analysis (Figure S1), it was noticed that the Ag/TNTs processed through the photodeposition method contained Ag NPs which were mostly attached to the outer surface of TNTs. This led to the higher antibacterial activity of nanomaterials synthesized by the photoreduction approach than that of Ag/TNTs prepared by chemical reduction methods, using both NaBH4 and SnCl2 as reducing agents, even though they contained higher amounts of Ag (Table 2). Nonetheless, in the presence of SnO2, the mechanism of antibacterial action of the NPs can differ from that of TNTs containing Ag only. The bactericidal potency of SnO2 was investigated by Vidhu and Philip [74] who found formation of zones of inhibition in the presence of these nanoparticles. They attributed that to the mechanism typical for metal oxides, i.e., formation of reactive oxygen species and electrostatic interaction of nanostructures with bacterial cell walls. Furthermore, Kumar Nair et al. [75] reported a synergic antimicrobial action of Ag and SnO2 towards *E. coli*. However, the authors applied UV irradiation to induce the antibacterial action. Nonetheless, the above data show that a direct comparison of Ag/TNTs containing solely Ag and those modified with both Ag and SnO2 NPs is difficult and the explanation of the antimicrobial action of the NPs containing tin oxide needs further investigations. Nonetheless, in general, the obtained results revealed that the efficiency of inhibition of *E. coli* and *S. epidermidis* growth was lower when SnO2 was present in the samples. Another mechanism of antibacterial action can be expected for the sol-gel-derived Ag/TNTs. In the case of these nanomaterials, no Ag NPs were detected during TEM analysis (Figure 1a,b); however, the XRD measurement confirmed the presence of the AgCl phase (Figure 2). Okkyoung et al. [76] in their work demonstrated that colloidal AgCl can be as important antibacterial agent as the other Ag forms, including Ag<sup>+</sup> ions. Taking the above into consideration the antibacterial properties of the discussed nanomaterials can be linked with the presence of AgCl.

In order to evaluate if there is any correlation between silver content in the hybrid Ag/TNTs and the log reduction values, the samples were divided into three groups, i.e., containing low (0.11 wt.%), medium (2.33–4.08 wt.%), and high (≥11.98 wt.%) amounts of Ag. Figure 6 summarizes the results. It can be observed that the nanomaterial with the lowest Ag content (0.11 wt.%) was characterized by the lowest antimicrobial activity against both bacteria. Analysis of the samples modified with a medium amount of Ag revealed that the antibacterial performance of the NPs containing between 2 and 4 wt.%Ag did not differ much, especially in terms of inhibition of *E. coli* growth. The highest log reduction with reference to both microorganisms (*E. coli*: 4.5 log reduction; *S. epidermidis*: 3.6 log reduction) exhibited Ag/TNT-2.5\_AM, containing 3.77 wt.%Ag. A slightly better activity towards *E. coli* (5.1 log reduction) was observed in the case of Ag/TNT-5\_SH (3.56 wt.%Ag); however, that sample was less efficient in terms of *S. epidermidis* inactivation (3.1 log reduction). In the discussed group of the NPs with medium Ag content, the lowest amount of silver was measured for Ag/TNT-2.5\_EL (0.1) (2.33 wt.%), for which the sample exhibited the lowest activity with reference to the Gram-positive bacteria. The third group of Ag/TNTs covered samples with the highest Ag content. Amongst them, the best antibacterial performance against both *E. coli* and *S. epidermidis* revealed Ag/TNT-100\_IN (6.0 and 5.4 log reduction, respectively), containing 12.58 wt.%Ag. No correlation between Ag content and antibacterial activity of the samples assigned to the third group was found. For example, the nanomaterial containing the highest Ag amount (30.85 wt.%) was less effective than the NPs with 12.58 wt.%Ag content. Such a phenomenon can be ascribed to the presence of Sn in the samples prepared by the electroless method (Table 2). Hassan et al. [77] observed that depending on SnO2

content, the antibacterial activity of SnO2/TiO2 composites can be improved or decreased. Therefore, as was mentioned earlier, the antibacterial properties of TNTs modified with both Ag and Sn NPs need further detailed studies.

**Figure 6.** Antibacterial properties of Ag/TNTs with reference to Ag content in the hybrid nanomaterials.

Another attempt in explanation of antibacterial properties of the various Ag/TNTs was based on Ag<sup>+</sup> release from the NPs (Figure 4). The samples were again divided into three groups, i.e., characterized by low (≤0.05% versus total Ag content or <sup>≤</sup>0.2 mg L−<sup>1</sup> in solution), medium (0.07–0.55% or 0.11–0.55 mg L<sup>−</sup>1) and high (≥0.64% or <sup>≥</sup>1.13 mg L<sup>−</sup>1) Ag<sup>+</sup> leakage (Figure S6). The lowest inhibition of bacterial growth was found for Ag/TNT-1\_SH, in which case no Ag<sup>+</sup> leaching for seven days was observed. However, no correlation between Ag<sup>+</sup> release and antibacterial properties was noted. For example, the least stable Ag/TNT-100\_EL (1), for which the concentration of the released Ag<sup>+</sup> was the highest (7.32 mg L<sup>−</sup>1), exhibited a similar activity as Ag/TNT-100\_NB with approximately 66 times lower Ag<sup>+</sup> leakage (0.11 mg L<sup>−</sup>1).

The above analysis confirms that when the size of Ag NPs is less than 10 nm, the interaction of NPs with the bacterial cell is more important than Ag content or Ag<sup>+</sup> ion release. Nonetheless, in general, the samples containing higher Ag loading were more active than samples with low silver amount. Furthermore, the nanomaterials for which silver concentration in the solution was in the medium and high range were characterized by higher antibacterial activity than those exhibiting the lowest Ag<sup>+</sup> release. A comparison of the Ag/TNTs synthesis procedures analyzed in this study revealed that the photodeposition approach was the most effective technique to obtain nanomaterials with the best antibacterial properties.

#### **4. Conclusions**

Silver/TNTs with different Ag contents were synthesized through five different procedures including (i) sol-gel followed by a hydrothermal process; (ii) photodeposition under ambient conditions; (iii) photodeposition under an inert atmosphere; (iv) NaBH4 reduction; and (v) an electroless deposition process after activation of TNTs' surface with various amounts of Sn2<sup>+</sup> ions. The physicochemical characterization of various Ag/TNTs revealed the presence of Ag NPs in most samples. The NPs with size ~1–10 nm were uniformly deposited onto TNTs' surface. Moreover, the electroless deposition resulted in the additional decoration of TNTs with SnO2 NPs. The presence of Ag NPs was not confirmed in case of nanomaterials obtained by the method (i), for which the AgCl phase was detected by the XRD analysis. Furthermore, no Ag NPs were observed in samples prepared from 2.5 mM AgNO3 solution using the photodeposition approach. Nonetheless, the ICP-OES analysis confirmed that all samples contained Ag and its loading varied from 0.11 wt.% to 30.85 wt.%. For Ag/TNTs with higher Ag content, a maximum stability of the nanocomposite was shown by the sample prepared with NaBH4 reduction method. It was noted that for the photoreduction process, the Ag NPs were

specifically deposited on the outer surface of the TNTs while chemical reduction led to the introduction of Ag NPs on both inner and outer surfaces. The antibacterial activity of different Ag/TNTs against both Gram-positive (*S. epidermidis*) and Gram-negative (*E. coli*) bacteria was evaluated under dark conditions. In general, Ag/TNTs with higher Ag content exhibited higher antibacterial activity compared to the nanomaterials with lower Ag loading. Also, *S. epidermidis* was inactivated with lower efficiency compared to *E. coli*, regardless of the hybrid NPs used. The Ag/TNTs obtained by the photodeposition approach were found to exhibit moderately higher antibacterial properties compared to samples prepared by other methods due to the higher interaction of Ag NPs present on the TNTs with bacterial cell walls.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/9/5/795/s1, Figure S1: HRTEM image of (a) Ag/TNT-100\_IN, (b) Ag/TNT-100\_EL (0.1), and (c)–(f) Ag/TNT-100\_NB (The circled regions in (e) and (f) represent the Ag NPs anchored to the inner surface of the TNTs. Moreover, in Figure S1(e) the Ag NP blocking the entrance to the TNT can be observed), Figure S2: EDS elemental mapping of (a) Ag/TNT-2.5\_AM, (b) Ag/TNT-5\_SH, (c) Ag/TNT-2.5\_EL (1). Scanning transmission electron microscopy (STEM) images with red squares present the scanned area, Figure S3: HRTEM image of Ag/TNT-100\_EL (1), Figure S4: The dependence between Ag content measured by XPS and ICP methods, Figure S5: XPS spectra of (i) Ag/TNT-5\_SH, (ii) Ag/TNT-100\_AM, (iii) Ag/TNT-100\_IN, (iv) Ag/TNT-100\_NB, (v) Ag/TNT-100\_EL (0.1), and (vi) Ag/TNT-100\_EL (1), Figure S6: Antibacterial properties of Ag/TNTs with reference to Ag release from the hybrid nanomaterial.

**Author Contributions:** Conceptualization, M.J. and S.M.; Funding acquisition, S.M.; Investigation, M.J., P.S., K.S. (Karolina Szyma ´nska), D.D., D.M., Z.L.-B., K.S. (Kacper Szyma ´nski) and S.M.; Methodology, M.J., P.S., K.S. (Karolina Szyma ´nska), D.D., D.M., Z.L.-B. and S.M.; Project administration, S.M.; Supervision, S.M.; Validation, M.J., P.S., K.S. (Kacper Szyma ´nski) and S.M.; Writing—original draft, M.J., P.S., K.S. (Karolina Szyma ´nska), D.M. and S.M.

**Funding:** This work was supported by the National Science Centre, Poland under project No. 2016/21/B/ST8/00317.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
