**2. Results**

The structure and chemical composition of the silver-containing nanocomposite were studied by transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS) and extended X-ray absorption fine structure (EXAFS).

The TEM micrograph of the cotton fibre is shown in Figure 1. The photograph shows amorphous extended structure with a diameter in the range of 16–25 nm, and dark crystalline nanostructures, the density of which is higher than that of cotton. The enlarged image of the regions of ordered atoms on a surface marked with a square in Figure 1 is shown in Figure 2.

The EDS spectra obtained from dark nanostructures (not shown) contain the characteristic line AgLα = 2.98 keV, which allows attribute dark nanostructures to AgNPs. Most of AgNPs are in the form of chains and agglomerates, which is a characteristic feature of the systems prepared by MVS. The particle size distribution is narrow and monomodal. The average particle size is 1.75 ± 0.25 nm.

Simulation of the electron diffraction for the group of the atoms selected in Figure 2 is shown in Figure 3. It is seen from Figures 2 and 3 that groups of atoms form faces with interplanar distances d1 and d2. Calculating the ratio d1/d2 gives a value of 1.09. The angle between d1 and d2 is 52◦. The values 1.09 and 52◦ indicate the presence of faces (111) and (200) of a face-centered cubic (fcc) structure in silver particles (for an ideal fcc structure: d(111)/d (200) = 1.15, the angle is 55◦).

**Figure 1.** Transmission electron microscopy (TEM) micrograph of Ag-cotton system.

Based on the synthesis method and sample storage conditions, three models of the chemical composition of the particle surface can be proposed: Ag0, Ag+, and Ag<sup>0</sup> + Ag+. With a high degree of probability, the mixed composition in crystalline form can be excluded from the consideration. Otherwise, "double reflexes" should be observed in Figure 3. The fcc lattice constants (**a**) for Ag<sup>0</sup> and Ag2O are 4.08 Å and 4.76 Å, respectively. Calculation of **a** by the formula **a** = d(hkl) × (h<sup>2</sup> + k2 + l2)1/2, where h = 1, k = 1, l = 1; d (hkl) = 2.3 Å, gives a value of 3.98 Å. Consequently, there is reason to believe that the surface of the crystalline particles consists of metallic silver. However, presence of Ag+ state in amorphous form cannot be excluded.

Figure 4 shows the survey spectrum of Ag black. Along with the peaks characteristic of silver atoms there are peaks of impurity carbon and oxygen atoms.

The determination of the chemical state of silver atoms in nanoparticles by the XPS method is a complex task. This is due to the fact that the spectral characteristics of the metal particles and the oxide particles are fairly close. According to NIST XPS Database [14] the binding energies of the Ag 3d5/2 peak for Ag, Ag2O and AgO are in the ranges 367.9–368.4, 367.7–368.4, and 367.3–368.1 eV, respectively. One of the solutions to this problem is the use of the Auger parameter. However, as a rule, the concentration of silver nanoparticles in materials is low, whereas the registration of the Auger spectrum requires a significantly longer time than the recording of the photoelectron spectrum, which can lead to the reduction of silver from oxides under the action of X-ray irradiation [15].

**Figure 2.** An enlarged image of the regions of ordered atoms on a surface, marked with a square in Figure 1.

Another approach may be based on controlled differential charging, in which separation of signals from regions with good and poor conductivity is possible. If the region has a metallic conductivity and is in good electrical contact with the sample holder, then it does not accumulate a charge due to the emission of secondary electrons. In contrast, in a region with poor conductivity, a positive charge is usually accumulated, the value of which depends on the secondary emission coefficient and conductivity, and it leads to the displacement of photoelectron peaks in the region of high binding energies. When a positive bias voltage Ub is applied, the stray electrons flow to the sample surface is increased and contributes to the compensation of the surface charge. In this case, narrowing of the photoelectron peaks and their shift toward higher binding energies are observed. Photoelectron peaks corresponding to regions with good conductivity should be shifted by Ub the amount of applied bias voltage, while from areas with a worse conductivity by a smaller amount. With a negative bias voltage, the flux stray electrons decreases, it leads to increase in charging on non-conducting regions and an increase in the interval between peaks corresponding to signals from the conducting and non-conducting regions. In this case, the signal from the conductive region must also be shifted by Ub.

Taking into account the difference in the electrical conductivity of metallic silver and its oxides, we attempted to separate signals from regions containing silver atoms in a metal state and other chemical states by controlled differential charging, changing the potential on the sample holder. This approach is widely used to determine the presence of different phases in a sample [16–21].

**Figure 3.** Simulation of the electron diffraction pattern for the group of atoms circled in Figure 2 and assignment for face-centered cubic (fcc) structure.

**Figure 4.** X-ray photoelectron spectroscopy (XPS) survey spectrum of Ag black.

*Antibiotics* **2018**, *7*, 80

To discriminate regions with different conductivities, or other words with different chemical states, a bias voltage Ub = ±7 V was supplied to the sample holder. Figure 5 shows the Ag 3d spectra of the Ag black measured at different bias voltage Ub applied to the sample holder. It is seen that binding energy of the Ag 3d5/2 and Ag 3d3/2 peaks is slightly depends on Ub. Table 1 presents the characterization of the Ag 3d spectra. The full widths at high maximum (FWHM) are rather different as well. Both the binding energies of the Ag 3d peaks and their FWHM values indicate that the spectra contain some states with different conductivities. Considering that the recording the Ag2O and AgO spectra is accompanied by the surface charging [22,23], one can assume the presence of the Ag+ and/or Ag2+ state in the Ag black. To determine the characteristics of the Ag<sup>δ</sup>+ state a subtraction of the spectrum measured at Ub = 7 V from the spectrum measured at Ub = −7 V was performed under the condition of their best coincidence in the high-energy region. The difference spectrum is presented in Figure 5. The binding energies of the Ag 3d5/2 and Ag 3d3/2 peaks 367.73 and 373.71 eV correspond to Ag2O state [22]. The relative intensity of this state is no less than 0.27. This estimate is based on the fact that the spectrum measured at Ub = 7 V may contain Ag2O state.

**Figure 5.** The Ag 3d photoelectron spectra of Ag black measured at different Ub = +7 (1), −7 (2), 0 V (3), ant difference spectrum (2)–(1). The spectra are corrected for Ub.


**Table 1.** Binding energies and full widths at high maximum (FWHMs) measured by XPS.

The C 1s and O 1s spectra of Ag black show a strong dependence on Ub (Figure 6) which indicate that a large part of carbon and oxygen is has low conductivity. It should be noted that binding energy of the main C 1s peak measured at Ub = +7 V is 284.77 eV, and corresponds to that used for charge reference [14].

It follows that some of the carbon atoms have good conductivity, or in other words, is in close contact with silver atoms. To estimate this value, we use the fitting the C 1s spectrum measured at Ub = −7 V, when the best separation of the electron emission from regions with good and poor conductivity is realized. When the spectrum was fitted with some components, two restrictions were imposed: The width of the low-energy peak should describe the low-energy side by the best way, and

the energy interval between the low-energy peak and the next should not be less than that at Ub = +7 V. The second restriction is imposed because of the possible manifestation of differential charging for regions containing C-O groups. The fraction of such carbon atoms is 0.47. A similar value of 0.48 was obtained for the spectrum measured at Ub = 0 V, whereas for the spectrum measured at Ub = +7 V it is 0.77. This is due to conductivity induced with the stray electrons, which make conducting regions that are not in direct contact with silver nanoparticles. Figure 7 shows fitting the corresponding C 1s spectra, and Table 2 contains corresponding data.

**Figure 6.** The C 1s and O 1s photoelectron spectra of Ag black measured at different Ub = +7 (1), −7 (2), 0 V (3).

**Figure 7.** Fitting the C 1s photoelectron spectra of Ag black measured at Ub = −7, 0 and 7 V.

**Table 2.** Binding energies (Eb), peak widths (W) and relative intensities of the peaks deconvoluted in the C 1s and O 1s spectra of Ag black.


It should be noted that the fitting the C 1s spectra with some components measured at Ub = −7 and 0 V, presented in Figure 7 is largely conditional and do not reflect the real relative concentrations of COx groups. At the same time, applying a positive bias voltage Ub = +7 V practically compensates the surface charging, and Figure 7 (+7 V) reflects the real relative concentrations of COx groups.

It was found that the peaks in the low-energy region of the O 1s spectra, measured at Ub = −7 and 0 V (Figure 8), have close binding energies and intensities. Based on the reference data [21,22], the binding energy of these peaks of about 530.8 eV cannot be attributed to C-O bonds. Therefore, they should be attributed to the Ag-O bonds. And from the weak dependence of Eb and Irel on Ub, one can assign binding energies of 530.6 and 530.9 eV to Ag-Ag-O state. The Irel of this state, determined from the fitting the O 1s spectra measured at Ub = −7, 0 and 7 V, are 0.24, 0.18 and 0.18, respectively.

It should be stressed that relative intensities of CO groups in the O 1s spectrum measured at Ub = 7 V correspond those obtained from the relative C 1s spectrum. It means that the bias voltage of Ub = 7 V neutralize the surface charging.

**Figure 8.** Fitting the O 1s photoelectron spectrum of Ag black measured at Ub = −7, 0 and 7 V.

Figure 9 shows the survey spectrum of the Ag/bandage system. Along with the peaks characteristic of silver, carbon and oxygen there are peaks of impurity silicon atoms.

**Figure 9.** Survey spectrum of the Ag/bandage system.

In case of the Ag/bandage system the charge referencing was done using the C 1s spectrum of cellulose. The latter was simulated using the reference data [24] by considering the difference in peak resolution. The binding energy of 286.7 eV was assigned to C-OH state of cellulose. Figure 10 shows the C 1s spectrum of the Ag/bandage sample fitted with four Gaussian peaks at 286.7, 288.1, 284.8 and 283.13 eV. The first and the second peaks are assigned to cellulose [24]. The third peak is assigned to adventitious carbon, while the origin of the peak at 283.1 eV is not clear because it does not correspond to reference data for polymers [24]. However, it may be resulted either of differential

charging or low-molecular weight species. Similar C 1s spectrum was recorded for Au/bandage system. It slightly differs in relative concentration of C-C/C-H peak and peak at 283.1 eV. The O 1s spectra of Ag/bandage and Au/bandage systems are practically indistinguishable. It means that the peak at 283.1 eV may be assigned to C-C/C-H state as well.

**Figure 10.** The C 1s photoelectron spectra of Ag/bandage and Au/bandage systems.

Figure 11 shows the C 1s spectra of Ag black and Ag/bandage system. It is clearly seen that the spectra are strongly different. The C 1s peak of the Ag/bandage system shifted to high binding energy region by 0.66 eV and its FWHM is 1 eV more than that of Ag black. These differences may be assigned to the size effect in photoelectron spectra which induces both energy shift to high binding energy and the peak broadening [25,26]. The transition from AgNPs in the Ag black, to their dispersion in the bandage followed with fairy large changes in size of AgNPs, and an increase of the proportion of the Ag<sup>0</sup> state was observed. Apparently, there was a partial reduction of silver and its stabilization by a modified layer of cellulose. However, as follows from a comparison of the O 1s spectra of the Ag black and Ag/bandage system (Figure 12), the low-energy side of the latter might be assigned to the Ag-Ag-O group.

**Figure 11.** The C 1s photoelectron spectra of Ag black (1) and Ag/bandage system (2).

**Figure 12.** The O 1s photoelectron spectra of Ag black (1), Ag/bandage (2), and Au/bandage systems (3).

But, given the almost complete coincidence of the O 1s spectra of Ag/bandage and Au/bandage systems, this conclusion should be rejected. Thus, one can conclude that the basic state of Ag atoms in Ag/bandage system is Ag<sup>0</sup> state, whereas the oxidized silver is in the form of Ag-Ag-O groups, and, as follows from Figure 12 the proportion of oxidized state is small. This is in accordance with EXAFS and TEM data which indicate that silver atoms are mainly in Ag<sup>0</sup> state. It should be noted that EXAFS is not a surface-sensitive method as XPS and electron diffraction may be recorded only from the ordered regions.

Table 3 shows the number of colony-forming units (CFU) of the studied microbes along the perimeter of the bandage at a distance equal to the diameter of one colony on both sides of the edge in the form Me (Q1; Q3), together with the level of statistical significance, where Me is the median, Q1—lower quartile, Q3—upper quartile.

**Table 3.** The number of colony-forming units (CFU) of the studied microorganisms along the edge of the bandage at a distance to both sides of the edge equal to the diameter of one colony (Me (Q1; Q3)) and the level of statistical significance (*p*) between control groups and gauze with AgNPs.


Due to the fact that the data of the control groups of different strains differ, in order to compare the antibacterial effect of the AgNPs-containing bandage with respect to different microorganisms, we calculated the percentage reduction factor. Table 4 presents the results of the study of the percentage reduction of CFU.

The results of the change in the antibacterial properties of the ordinary medical gauze bandage under the influence of laser irradiation are presented in Table 5, in the form Me (Q1; Q3), where Me is the median, Q1 is the lower quartile, Q3 is the upper quartile.


**Table 4.** Percentage reduction in the number of CFU towards control.

**Table 5.** The number of CFU of the investigated microorganisms along the edge of the ordinary medical gauze bandage (Me (Q1; Q3)) and *p*—the level of statistical significance between groups.


*p*1 = the level of statistical significance between groups 1 and 2, and *p*2 = the level of statistical significance between groups 1 and 3.

Based on the data presented in Table 5, no statistically significant changes in the indices of the number of CFUs when using an ordinary medical gauze bandage without or in conjunction with laser irradiation have been identified. This indicates the absence of an antibacterial effect in laser radiation at a wavelength of 470 nm (blue region of the spectrum), a 5 mW power, and with exposure time of 5 min.

Table 6 shows the results of the change in the antimicrobial properties of the medical gauze bandage containing AgNPs, depending on the presence or absence of exposure to the laser and the time through which it was performed.

**Table 6.** The number of CFU of the investigated microorganisms along the edge of the medical gauze bandage containing AgNPs, (Me (Q1; Q3)) and *p*—the level of statistical significance between the groups.


*p*1 = the level of statistical significance between groups 4 and 5, and *p*2 = the level of statistical significance between groups 4 and 6.

According to the data presented in Table 6, laser radiation in the blue region of the spectrum did not have a statistically significant increase in the antibacterial effect of AgNPs when applied two hours after inoculating the Petri dish and placing a bandage on it. However, when the laser treatment was used four hours after the seeding an increase in the antibacterial effect of this bandage was observed with respect to all studied microorganisms. The phenomenon is statistically significant in all groups of microorganisms. The difference can be observed in Table 7.

According to the data, presented in Table 7, there is no significant diversity between difference in percentage reduction of CFU of Gram-positive and Gram-negative microorganisms. Considering the fact that they differ in the structure of the cell wall, it can be concluded that the mechanism for increasing the antibacterial effect of the AgNPs-containing bandage cannot be explained solely by the effect on it. As follows from Table 7, laser radiation, when exposed four hours after inoculating the Petri dish and the application of a AgNPs-containing bandage on it allows to increase the antibacterial properties of the bandage by 15–24%, depending on the strain of the microorganism.

**Table 7.** Difference in percentage reduction of CFU between the groups of the AgNPs-containing bandage without laser irradiation and with laser irradiation four hours after inoculating the Petri dish and placing the bandage.

