Microscale ZnO with controllable crystal morphology as a platform to study antibacterial action on Staphylococcus aureus

Zinc oxide is a well-established semiconductor material, with ubiquitous chemical elements, which is inexpensive in production, and demonstrating numerous useful applications. Among those, antibacterial and antiodor properties render micro- and nanoscale ZnO very promising for use in medicine, food packaging, and first aid.^1–4 In particular, studies of ZnO nanoparticles suggest that their antibacterial action is driven by cell membrane damage, consecutive cell infiltration, and disruption of the inner workings of the cell.^5–7 Other studies indicate that biomimetic signaling, oxidative stress, or metal ion infiltration can interfere with bioenergetics, cell metabolism, and DNA replication inside the cell.^5–8 Furthermore, high concentrations of Zn²⁺ ions can enter the cell via transport proteins and replace other metal ion cofactors used in certain enzymatic processes within the cell. This can cause mismetallation (leading to protein dysfunction) and inhibit the aforementioned enzymatic processes within the cell. Currently, the fundamental mechanisms behind the antibacterial activity of ZnO are not established. In our present work, we attempt to test the hypothesis that, as previously suggested, such antibacterial action stems from the physical and chemical interactions between the surfaces of ZnO particles and those of bacterial cells or micro-organisms.^1,9

ZnO has an anisotropic wurtzite crystal structure that yields two characteristically different types of crystallographic surfaces—polar (hexagonal, Zn- or O-terminated) and neutral nonpolar (rectangular). It is well established that those differences in surface polarity and morphology lead to manifest discrepancies in their physical and chemical characteristics. In this regard, it is quite feasible that those discrepancies may produce their own hierarchy of antibacterial mechanisms mentioned; thus, addressing this aspect may improve the overall understanding of the fundamentals behind the antibacterial action by ZnO.

As of today, most of the studies of the antibacterial action by ZnO employed primarily nanoparticles.^10–13 However, since ZnO nanoscale particles largely lack a distinct crystallographic morphology and tend to be internalized by bacterial cells, such an approach limits the understanding of the role of surface polarity. In our studies, in order to address this deficiency, we propose to utilize particles of the same size as or larger than that of bacteria. In doing so, we essentially eliminate the possibility of internalization of the particles and ensure the promotion of direct interactions between surfaces of the bacteria and the ZnO particles. Additionally, microscale crystals allow for the isolation of the distinct surface types: polar versus nonpolar. Thus, production of microparticles with a controllable array of geometries will render it a promising platform to study contributions of specific ZnO surface types to its interaction with bacteria.

These contributions of the surfaces with different polarities to the antimicrobial action dwell on the fundamental mechanisms driving this behavior. Furthermore, although polar surfaces of anisotropic metal oxides are naturally unstable, ZnO surfaces tend to self-stabilize via the introduction of certain types of defects that are not in abundance at the nonpolar surfaces.^14–18 The presence of these defect states as well as the electric charge distributions at the surface can impact interactions between ZnO and bacteria. To tackle these fundamental aspects connecting optoelectronic properties to both antibacterial action and surface morphology, we conduct additional studies of the surface charge dynamics in the investigated ZnO specimens.

The morphological criteria ZnO particles need to meet (controllable abundance of surface polarity, comparable sizes to those of bacteria) are complemented by the requirements for the growth procedures to be scalable and economical. To the best of our understanding, the most adequate approach to meet these prerequisites is to utilize a bottom-up hydrothermal method to grow ZnO microcrystals as follows. ZnO growth protocols utilized varying aqueous concentrations of hexamethylenetetramine (HMTA), 99.999% pure Zn foil, and zinc acetate dihydrate [Zn(Ac)₂]. De-ionized water was first added to a beaker. HMTA was then added to the beaker with the mix being continuously stirred with a magnetic stirrer. After 5 min of stirring, Zn(Ac)₂ was added to the solution. Specific volumes of each precursor can be found in Table I. This mixture was then stirred for 30 min. At this point, a 10 × 50 mm² strip of Zn foil was cleaned using acetone and ethanol and placed into a Teflon-lined stainless-steel reactor (Baoshishan). The solution of the chemical precursors was then added to the reactor, which was then sealed shut and placed into an AI Stabletemp forced air drying oven. Inside the oven, the reactions took place for 4–24 h at temperatures ranging from 95 to 130 °C. The times and temperatures were varied to produce particles of different sizes and abundances of polar surfaces.

Morphologically different ZnO crystals can be grown by influencing the growth rate of the crystal along its fastest growing directions $([21¯1¯0],[01¯1¯0],and[0001])$⁠.¹⁹ Crystal growth is fastest along the c axis since the nonpolar surfaces have a lower energy than the polar ones. The preferential growth direction can be changed or weighted differently by modifying the concentration of precursors, growth temperature and time, as well as the pH of the solution.

HMTA is not a supplier of Zn²⁺ ions or O²⁻ ions, but it heavily influences the morphology of hydrothermally grown crystals. Higher HMTA concentrations increase the pH of solutions and tend to yield thicker morphologies. Zinc acetate acts as a zinc supplier for the reaction and contributes to the lowering the pH of the reactant solution.

Particle size is influenced by the time and temperature of the reaction as well as the availability of precursors. Studies have shown that particle size increases with the increasing reaction time²⁰ until Zn²⁺ supply is depleted. Furthermore, higher growth temperatures contribute to the growth along the c axis. Lower temperatures, therefore, yield shorter, more polar crystals.

The preferential growth direction of these crystals is heavily influenced by the pH of the solution. More acidic solutions limit the availability of the OH⁻ ions. ZnO formation and growth along the c axis is a direct result of interactions between Zn²⁺ ions and OH⁻ ions.²¹ More alkaline solutions with a lower concentration of Zn²⁺ ions contribute to the growth of longer rods.²¹ Studies have indicated that as the pH increases, the growth along the c axis dominates forming longer hexagonal rods.²¹ 

We used Scanning Electron Microscopy (SEM) and Energy Dispersive x-ray (EDX) spectroscopy to characterize the synthesized microcrystals. This was done using a JEOL field emission scanning electron microscope. The samples were mounted inside the SEM chamber and observed under a 15 kV electron beam for morphology characterization. The EDX probe was used to verify the chemical composition of the particles.

The antibacterial action was characterized via Minimum Inhibitory Concentration (MIC) assays, which were performed using the Newman strain of Staphylococcus aureus (SA) in the Becton-Dickinson Mueller-Hinton Broth (MHB) medium. The cultures grown overnight were diluted and grown to an early log phase with an optical density at 600 nm (OD600) of 0.4. The log phase cultures were diluted 1:200 into microcentrifuge tubes containing ZnO particles at the indicated concentrations in a final volume of 1 ml. ZnO particles were either synthesized as described above or purchased (Sigma-Aldrich). In order to ensure that the ZnO particles were in a continual contact with the bacterial cells, the microcentrifuge tubes were continuously inverted for 16–20 h at 37 °C. As a control, tubes containing only ZnO particles at each concentration in 1 ml MHB were also included. Tubes were then removed from the inverter and centrifuged for 30 s at 100 rpm to separate the ZnO particles from the bacterial cells. 200 μl of the supernatant was then transferred from each tube into individual wells of a 96-well plate and OD600 was determined using a Fluostar Omega plate reader from BMG Labtech. The OD600 values from the ZnO control tubes containing no bacteria were then subtracted from the corresponding ZnO tube containing bacteria in order to control for the increased background levels caused by the increasing amounts of ZnO. All MIC assays were repeated at least three times.

To prepare for SEM imaging, the remaining cultures from the MIC assays were combined, washed once in Phosphate Buffered Saline, and then fixed in 1.6% glutaraldehyde for 1 h at room temperature. The samples were dehydrated in a series of 10-min ethanol incubations at the following concentrations: 30%, 50%, 70%, 85%, 90%, and two times at 100%. The samples were then incubated with hexamethyldisilazane overnight with the lid open until the excess liquid evaporated. The dry pellets were crushed, transferred to a metal pedestal, and sputter-coated with 8 nm of gold.

Surface Photovoltage (SPV) is a useful tool for studying surface and subsurface charge dynamics and electronic structure in materials with a bandgap. SPV allows probing the charge carrier distribution by measuring changes in surface potential due to illumination. There are two primary methods of studying SPV: SPV spectroscopy and transient SPV. SPV spectroscopy measures changes in the potential versus the energy of incident photons, therefore probing the electronic structure near the surface, whereas transient studies reveal information about the near-surface charge dynamics.²² 

The surface and near-surface electronic states are highly sensitive to the environment and the presence of adsorbed species. For this reason, in our studies, the SPV experiments were performed in a high vacuum chamber under 10⁻⁷ Torr pressure. A Besocke Kelvin Probe S and a controller were used to measure the change in the contact potential difference (CPD) for each sample during both transient and spectroscopic studies. Prior to spectral measurements, the samples were flooded with white light for a dynamic saturation of the surface states and then left in the dark for an extended period of time ranging from tens of minutes to tens of hours to relax back to the unexcited state. After the surface had been quenched, SPV behavior under white light was measured as a function of time, thus revealing the so-called transient “light-on” behavior of charges in the near-surface space charge region (SSCR). Once the sample reached a steady state of no change in CPD or it had been exposed to white light long enough so that the SSCR would not be significantly affected, the light was turned off. The transient SPV signal was then measured in the dark to obtain the “light-off” dependences. After completing these SPV transient experiments for each sample, SPV spectra were collected as follows.

An Oriel Cornerstone 130 1/8 m monochromator controlled by an NI LabView virtual instrument was used to disperse the white light from a tungsten halogen lamp for a monochromatic illumination of the surface. Bandpass filters were employed to prevent illumination of the sample by any first and second order diffraction harmonics. The same labview software was used to measure changes in the CPD and plot it as a function of the illumination energy.

Employing the hydrothermal growth approach described above, we produced a broad variety of ZnO microcrystalline morphologies, as confirmed by the SEM experiments. We successfully achieved a desirable control of the particles’ sizes, shapes, size distributions, and relative abundances of surface polarities through a systematic and detailed variation of the synthesis conditions (temperature, processing time, etc.) and the concentrations of chemical precursors. To produce the desired hexagonal crystals, we found that the 1:1 ratio of HMTA and Zn(Ac)₂ was necessary. In Table I, we provide several examples of the combination of ZnO growth parameters and the resulting sample geometries.

Table II lists the pH of the reactant solution before and after heat treatment. Note that the pH changes over time during the heat treatment process. We observed reaction solutions becoming more acidic as ZnO crystals are formed. Although the Zn²⁺ ions initially contribute to lowering the pH, as ZnO is formed, the OH⁻ ions in solution interact with the Zn²⁺ ions to form ZnO and the H⁺ ions in solution.

Figure 1 provides the corresponding SEM images of the samples listed in Table I. In Fig. 1(a), one can see that the type LR sample consists of primarily microrods with hexagonal cross sections. Obviously, in this specimen, the predominant crystallographic free surfaces are nonpolar. On the other hand, Fig. 1(b) for the type HP sample shows an assembly of rather thin hexagonal plates with a significant abundance of polar surfaces, both O- and Zn-terminated. Figure 1(c) is an SEM image of the type B sample exhibiting a morphology with a rather balanced relative abundance of polar and nonpolar surfaces. Figure 1(d) for sample type M is an example of a mixture of morphologies with varying relative abundances of polar and nonpolar surfaces. A number of other ZnO microcrystalline morphologies and relative abundances of polar versus nonpolar morphologies were also produced and characterized. The specimens shown in Fig. 1 demonstrate that the chosen synthesis protocol allowed a successful realization of our goal to have a platform to investigate antibacterial action mechanisms at ZnO surfaces with different polarities. Importantly, practically all of the crystals in the samples produced were of micrometer or supermicrometer dimensions, critical to simultaneously achieving two goals—increase the crystalline surface in contact with bacteria (compared to bulk ZnO surfaces) and prevent internalization of ZnO crystals into cell bodies.

Characterization of the stoichiometry of the samples was implemented using EDX spectroscopy described in Sec. II. EDX measurements were run in conjunction with the SEM imaging of the grown material; they consistently and unambiguously confirmed a robust well-balanced ZnO stoichiometry in practically all the specimens obtained.

The SPV measurements were used to characterize the overall quality of the ZnO crystals, to identify the surface/near-surface electronic states, and to examine the response in the charge dynamics to the variation in illumination. The time-dependent SPV experiments described above revealed, in virtually all samples, a rather complex transient behavior, with multiple characteristic time scales pointing to several channels of charge recombination in the surface/subsurface vicinity.²³ An example of SPV transients for a type B sample is shown in Fig. 2(a). Detailed investigations of the SPV transients observed in our samples will be performed in the nearest future to elucidate the underlying optoelectronic mechanisms. The SPV spectroscopy results unambiguously confirmed the n-type semiconductor behavior in all samples studied and demonstrated a high surface quality of ZnO crystals, as illustrated in Fig. 2(b) by a strong and sharp bandgap transition at ∼3.3 eV and a relatively flat spectral signal from the sub-bandgap energy range. This SPV spectrum was taken on a type B sample with a balanced abundance of polar and nonpolar surfaces. More detailed inspection of the SPV sub-bandgap spectral regions in our samples revealed primarily three different surface-trap-related SPV transitions. Generally, SPV transitions can be characterized by the nature of slope changes of CPD as a function of illumination energy. The direction of these slope changes, or “knees,” indicates the nature of the transition. For example, an “outward knee” corresponding to the slope becoming more negative indicates an electronic transition from the surface gap state to the conduction band minimum, whereas an “inward knee” corresponds to the transition from the valence band maximum to the gap state level. Sample types B and LR exhibit “knees” at ∼1.6 eV to the conduction band and at ∼2.3 eV from the valence band. We also observed a transition at ∼2.65 eV to the conduction band in sample type B. Following Refs. 24 and 25, these transitions could be attributed correspondingly to such native defects as oxygen vacancies, oxygen interstitials, and zinc vacancies. Figure 3 illustrates these transitions for sample type B. Figure 4 illustrates these transitions for sample type LR. Similar near-surface transitions are also present in sample type HP, shown in Fig. 5. Noteworthy, the SPV transitions detected in our microcrystals are consistent with those reported for commercial grade ZnO nanopowders. It should be mentioned, however, that the SPV transitions listed above had different relative intensities in samples with different morphologies. One can surmise that the relative abundance of these surface states might influence the antibacterial activity of ZnO particles.

The MIC results for the ZnO microcrystals synthesized by us indicate a strong, consistent inhibition of bacterial growth. Figure 6 illustrates these results for sample types M, B, and LR, all showing an inhibitory concentration of ∼0.625 mg/ml. This observation is crucial for answering one of the questions outlined in the Introduction, since it definitively demonstrates that cell internalization of the ZnO particles into the bacterial cell is not necessary for bacterial growth inhibition. Furthermore, this strongly suggests the importance of interactions (physical, chemical, and biochemical) between ZnO surfaces and extracellular material.

For comparison, we ran similar MIC measurements employing commercial grade ZnO nanopowders, which also demonstrated substantial and, on average, somewhat higher inhibition of bacterial growth. Thus, while internalization of the particles may contribute to more effective growth inhibition, it is not the fundamental driving force behind the antibacterial action of ZnO. Moreover, size effects play a large role in this observation, since the nanoparticles are orders of magnitude smaller than the SA bacteria and have a substantially larger surface-to-volume ratio. While this lends itself to the interpretation that internalization improves the antibacterial action of ZnO, it could simply mean that the number of particles, and therefore the intensity of interactions, is orders of magnitude greater for nanoparticles than for microparticles.

The MIC results were consistent across the various morphologies tested. However, the signal-to-noise ratio in the performed antibacterial assays is, at this point, below the level allowing to clearly distinguish the influence of the surface polarity on the antibacterial action. Alternatively, assuming that surface defects play a role and having observed similar SPV defect signatures in different morphologies, it is possible that the polarity plays a secondary role, whereas the relative abundance of surface defects is more important. In our further studies, we will focus on refining the experimental conditions to address this issue.

Some samples were examined by SEM after running the MIC assays, as described in the Experiment section, to reveal possible mechanisms of ZnO-bacterial interactions. Figure 7(a) is an SEM image of the SA cells interacting with ZnO particles of a type M sample that has a mixed abundance of morphologies. It should be noted that the bacteria are clearly in direct contact with ZnO particles. One can see that the crystalline surfaces of ZnO exhibit a certain amount of deterioration and damage especially along the crystallographic corners and edges. Figure 7(b) shows further damage to ZnO surfaces of the same sample, thus implying that interactions with bacteria lead to the degradation and possible dissolution of the ZnO crystals. Studies have shown that the polar faces and edges with a high number of oxygen dangling bonds are susceptible to dissolution in acidic (pH < ∼3.8) environments.^26,27 This, however, does not fully explain the observed behavior as the MHB has a pH of ∼7.3. In our future studies, we will address the postassay ZnO damage in a more systematic manner.

In this work, we successfully developed a growth routine to produce microscale ZnO with controllable abundance of surface polarities. SEM, EDX, and SPV measurements indicated a high quality of the grown microcrystals. MIC assays, designed to intentionally avoid intracell penetration of ZnO particles, have confirmed that internalization of the ZnO particles is not necessary for antibacterial action and, therefore, a possible surface-to-surface interaction between ZnO and SA during the antibacterial action.

This work was supported in part by the National Science Foundation (Grant No 1852267). The authors would like to thank J. Coffer (TCU) for useful discussions and the following high-school students participating in the TCU Research Apprentices Program for technical assistance: P. Ahluwalia, M. Hattarki, T. Haun, L. Le, R. Maheshwari, L. Pane, T. Ryu, and L. Simon.

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