Physicochemical and Antibacterial Properties of Bivalent Freeze and Furnace Dried Nanoscale Cerium Oxide


 Bone healing is a complex process, and if not managed, can lead to bacterial infections, non-union and compromised healing of bone. The broad overuse of antibiotics has led to the emergence of antibiotic-resistant bacteria. Due to the growing urgency to minimise the dependency on antibiotic drugs, alternative treatment strategies, including the use of nanoparticles, have attracted significant attention. Bivalent cerium oxide nanoparticles (Ce4+ and Ce3+) synthesised via a hydroxide mediated approach were calcined at 280, 385 and 815 ˚C identified using the Simultaneous Thermal Analysis technique. The resulting nanoparticles were characterised using X-ray Powder Diffraction, Fourier Transform Infrared Spectroscopy, Ultraviolet-Visible Spectroscopy, Transmission Electron Microscopy, and Electron Energy Loss Spectroscopy. The antibacterial potential of cerium oxide nanoparticles corresponds to the particle size and the presence of oxygen vacancies in the fluorite crystal structure. The antibacterial efficacy of nanoparticles was characterised at concentrations of 50, 100 and 200 µg/ml and tested against the following strains, Escherichia coli, Staphylococcus epidermis and Pseudomonas aeruginosa by determining the half-maximal inhibitory concentration (IC50). The calcination temperature was found to affect the agglomeration tendency, particle size distribution and the ratio of Ce3+:Ce4+ oxidation states. The hydroxide mediated approach yielded spherical nanoparticles of ceria with particles size ranging from 4 nm to 53 nm. The freeze-dried bivalent nanoparticles exhibited 18.5 ± 1.2%, 10.5 ± 4.4%, and 13.8 ± 5.8% increased antibacterial efficacy against Escherichia coli, Pseudomonas aeruginosa and Staphylococcus epidermis, respectively compared to nanoparticles consisting solely of Ce4+ ions, i.e. nanoparticles calcined at 815 ˚C.


Introduction
Bone infections are frequent due to the increasing incidence of trauma and elective surgeries (1)(2)(3). Sources of infectious bacteria include the operating room, surgical equipment, contaminated orthopaedic/medical device but also resident microbiota already present on the patient's skin (4). Epidemiological studies suggest between 2 and 5 % (5, 6) of all the implantrelated procedures are likely to be further complicated due to post-operative infections leading to an increase in substantial health-related costs, prolonged hospitalisation, revision surgeries but also morbidity (7)(8)(9). The treatment of infection is complex involving parenteral or systemic drug administration and in extreme cases, debridement of bone and tissue due to compromised blood circulation (10). Drug administration for infection control uses a broad range of antibiotics which has led to the emergence of multidrug-resistant bacterial strains (8,11). As a result, the treatment of diseases which were under control several years ago now require higher doses of multiple antibiotics, often leading to intolerable toxicity.
Contaminated surgical instruments or residual bacteria present on the implanted orthopaedic device introduces bacteria into the patient's body. Tissue and proteins in the blood gradually adsorb to the surface of the implanted device, as demonstrated extensively in the literature (7,(12)(13)(14)(15). Bacteria in a planktonic state adheres to the adsorbed proteins and proliferates until a colony of bacteria is formed, enabling a change in the gene expression pattern. The genes responsible for the production of bacterial extracellular polymeric substances (EPS) are activated and expressed (12). The excretion of EPS further facilitates the exponential growth of bacteria leading to biofilm formation (7,14,15). Thus, the presence of a biofilm hinders the host's immune response and antibiotic delivery; hence biofilms are one of the major causes for the development of bacterial resistance (8).
The mechanism for bacterial resistance can be attributed to evolutionary processes taking place during antibiotic therapies and via horizontal gene transfer by (i) conjunctions, (ii) transduction and (iii) transformation (i.e. commonly known as superbugs) (13). Antibiotics work by attacking at least three modes of targets; cell wall, translation mechanism and DNA replication mechanism. However, many bacteria have developed a resistant gene known as New Delhi Metallo-Beta-Lactamase-1 (NDM-1), which prevents the destruction of bacteria. Mechanisms of bacterial resistance include the expression of certain enzymes which can modify or destroy antibiotics, i.e. β-lactamases (16), modification of ribosomes and cell wall as shown with tetracycline and vancomycin resistance (7). Due to the growing urgency to minimise the dependency on antibiotic drugs, alternative treatment strategies, including the use of nanoparticles, have attracted significant attention as these types of particles manifest unique physicochemical properties. The unique properties are only apparent on the nanoscale, e.g.
surface area to volume ratio, surface charge and oxidation state when compared with bulk counterparts (17). Several types of inorganic nanoparticles (ZnO, TiO2, Ag) are emerging as novel antibacterial agents and proven their effectiveness for treating infectious diseases (18).
Antibacterial efficacy of nanoparticles is affected by factors such as the size, shape, surface charge and surface area to volume ratio (17). Nanoparticles size is essential with regards to biological functions as the morphological dimensions are comparable with (i) small biological molecules (1-10 nm), (ii) viruses (10-100 nm) and (iii) ability to attack biological entities without changing their functions (17). The antibacterial mechanism is likely to be attributed to their ability to enter biofilms; unlike antibiotics, nanoparticles directly attack the cell wall of bacteria by attaching via (i) electrostatic interaction, (ii) Van der Waals forces as well as (iii) receptor-ligand and hydrophobic interactions (8) disrupting the integrity of the bacterial cell wall leading to cell death (19).
Nanoparticles can prevent bacteria from mutating via cell death; thus could replace or reduce the use of conventional antibiotics (20). The antibacterial properties of metal and metal oxide nanoparticles such as silver (21), copper (22), zinc oxide (23) and titanium dioxide (24) demonstrated to alter the metabolic activity of Gram-positive and negative bacteria (8). Zinc oxide nanoparticles are found to inhibit Staphylococcus aureus, whereas concentrationdependent silver nanoparticles exhibited antimicrobial activity against Escherichia coli and Pseudomonas aeruginosa (25). However, specific nanoparticles, i.e. silver, are toxic to host cells even at low doses despite exhibiting antibacterial properties (26). Nanosilver intraperitoneal injection and its dispersion through blood are found to adversely affect the lungs, liver, gastroenterological tract and brain tissues (27). For this reason, the use of silver in treating internal infections is now severely curtailed. Other studies have reported gold (28), magnesium oxide (29) and copper oxide (8,30,31) based nanoparticles prevented the formation of biofilm, which is linked to the high surface area-to-mass ratio, i.e. smaller sized less than 10 nm (7,32). Additionally, triangular-shaped silver particles exhibited higher antibacterial properties as compared with spherical or rod-shaped nanoparticles (33).
Cerium oxide nanoparticles have attracted a great deal of interest as antibacterial agents due to the ability to cycle between the two valences states (Ce 3+ and Ce 4+ ) leading to the formation of oxygen vacancies in the lattice. The apparent beneficial oxygen buffering capability enables the nanoparticles to act as a catalyst for both oxidation and reduction reactions (34); hence can manifest a unique antibacterial mechanism (20). The intrinsic bivalence of cerium oxide nanoparticles induces antioxidant capabilities (35) (catalytic oxidation and reduction) protecting the cells from oxidative stress, inflammation (36) and potential radiation damage (37). Nanoscale cerium oxide can mimic an antioxidant enzyme superoxide dismutase found in all living cells (38). Superoxide dismutase, catalase and glutathione are considered the body's cellular defence as they catalyse the breakdown of potentially harmful oxygen molecules known as reactive oxygen species (ROS)/free radicals thus preventing tissue damage within the body. The primary role of antioxidants is to reduce excessive amounts of ROS/free radicals hence combating oxidative stress-related diseases (39). Nanoscale cerium presents relatively low or no toxicity to mammalian cells (40)(41)(42)(43) and is proven to decrease catalysts of chronic inflammation via nanotherapeutics (36) as well as demonstrating the ability to enhance neuroprotection (44).
Cerium oxide nanoparticles exhibit pro-oxidative behaviour, depending on the environment, i.e. oxidative stress is induced, which is directed at bacteria (45). Conversely, several studies are concluding that there is no apparent antibacterial effect of cerium oxide nanoparticles (46,47). However, other findings highlight possible adverse effects of the cerium oxide nanoparticles, where oxidative stress was induced in epithelial human lung cells (48). The range of conflicting data may be attributed to factors such as varying manufacturing processes, chemical solvents not entirely removed, irregular pH during production and increased calcination temperatures (49). The redox properties of cerium oxide nanoparticles can be tuned via materials preparation method, drying method, particle size, surface chemistry, particle shape and level of dopant materials (34). The drying method of nanoparticles is a vital aspect to consider as nanoparticles tend to agglomerate, which adversely affects the physicochemical properties of the particles (50). Thus, the procedures employed to evaluate the antimicrobial and antibacterial properties associated with cerium oxide nanoparticles may also make it challenging to form significant conclusions.
In the fluorite structure of cerium oxide, the redox equilibrium between the two valence states (Ce 3+ :Ce 4+ ) may be explained by considering the reaction in the presence of oxygen gas, as shown in equation (1): The intrinsic presence of oxygen vacancies in the CeO2 crystal structure renders CeO2 into a CeO2-x non-stoichiometric oxide with 'x' vacant oxygen sites (51). The ratio of Ce 3+ :Ce 4+ ionic states compensates against vacant oxygen sites and may be represented by the following equation: In equation (2), the oxygen vacancy is shown by ∎ where x is the fractional value of vacant sites in the fluorite structure. Oxygen deficiencies in the fluorite crystalline lattice occur when the oxygen partial pressure is less than the value predicted value from equation (1) enabling the vacant oxygen sites to act as a sink for oxygen, an essential step in redox reaction equilibrium in the cerium-oxygen system.

Reagents & Materials
Reagents and materials used for the preparation of cerium oxide nanoparticles were of

Furnace Dried Cerium Oxide Nanoparticles
The hydroxide mediated synthesised cerium oxide nanoparticles were placed into a furnace at 80 °C for 24 hours for drying. Thermal analysis of the furnace dried powder was used to determine the optimal calcination temperature. All synthesised cerium oxide nanoparticles are presented in Table 1 with sample code names, corresponding formulas, and synthesis method.

Thermal Analysis
Reactions and phase changes of freeze-dried scaffolds were investigated using the Perkin Elmer STA 8000. The thermal experiments were carried out from 30 ˚C to 600 ˚C at a heating rate of 20 ˚C/min. The Perkin Elmer STA 8000, which was used to study the phase transformation and chemical reactions covered the temperature heating range from 30˚C to 1000 ˚C. The thermal analysis characterisation of furnace dried ceria was essential for the optimisation of the calcination process without promoting the growth of nanoparticles. The isochronal heating rate of 20 ˚C min -1 was used to determine the optimal calcination temperature for the furnace dried cerium oxide nanoparticles samples for comparative studies.

Fourier Transform Infrared (FTIR) and Ultraviolet-Visible (UV-Vis) Spectroscopy
The molecular vibration spectroscopic analysis of synthesised powders (FUNP, FRNP, C285, C385 and C815) were analysed and characterised by using the Vertex 70 FTIR spectrometer in the attenuated total reflection mode (ATR). The beam splitter used was KBr, and the light source used was a NIR lamp. Each sample was scanned 32 times in the 400 cm -1 to 40000 cm -1 range. The spectral resolution was 4 cm -1 . For the characterisation of the electronic absorption spectra of the nanoparticles, the PerkinElmer ® , LAMDA 950 UV/VIS/NIR Spectrometer was used. A homogeneous clear suspension of nanoparticles in deionised water at various concentrations was used to collect the absorption spectrum between 190 nm and 600 nm.

X-ray Diffraction (XRPD)
A D8 X-ray powder diffractometer using the Kα radiation of Cu (λ = 0.15406 nm) was used to determine the crystalline structure of all prepared nanoparticles. The powder samples were analysed in the Bragg angle (2θ) scanning range of 10° to 80° at a scan speed of 5 seconds with a step size of 0.03˚. The recorded patterns were analysed using the HighScore Plus software, and the Rietveld refinement was employed for peak shape and intensity analysis for ascertaining the crystallinity of powder sample.

Transmission Electron Microscopy (TEM)
For TEM analysis, the samples were prepared by ultrasonic dispersion of the nanoparticles in methanol after which several drops were placed onto holey carbon copper TEM grids. The particles suspended in methanol were then allowed to dry using a heat lamp. The Titan Themis Cubed 300 TEM operated at 300 kV with high brightness X-FEG and Supertwin objective lens, was used for the characterisation of the nanoscale size distribution and morphological analysis.
The Bright field TEM images were collected using the Gatan OneView 16 Megapixel CMOS digital camera. Selected area electron diffraction (SAED) patterns, low magnification and darkfield (DF) TEM images were obtained for the analysis of the crystallinity of synthesised ceria.
Electron energy loss spectroscopy (EELs) using the Gatan GIF Quantum ER imaging filter at low and high energy loss spectra were collected for the characterisation of the co-existence of the two oxidation states in the calcined nanoparticles.

X-ray Photoelectron Spectroscopy (XPS)
The nanoparticles (FRNP, FUNP, C280, C385 and C815) samples were prepared for XPS by sprinkling small amounts onto adhesive conductive carbon tape mounted on to the sample holders. The samples were sprayed with aerosol air to remove any loose particles before analysis. The nanoparticles were then characterised using the UHV XPS with a SPEC Phoibos 150 analyser and a SPECS XR50-M. Samples surveys were taken at pass energy of 50Ev, and high-resolution scans were taken at 30Ev. The data collected were analysed using Prodigy, CASAxps and Originpro.

Surface Area
Micromeritics Tristar 3000 was used to characterise the Brunauer Emmett Teller (BET) surface area of the synthesised nanoparticles using the nitrogen absorption method in the powder bed.
Measured amounts of nanoparticles in glass sample tubes were placed into the FlowPrep TM 060 where both heat and inert gas was applied to the samples for 30 minutes to remove any atmospheric contaminants, i.e. water and absorbed gas.

Bacterial Culture and Antibacterial Properties
The nanoparticles tested during the bacterial experiments were sterilised via autoclaving by suspending the particles in a Brain Heart Infusion (BHI) broth and once cooled were used within the hour. Optical density and viable count (in colony-forming units, CFUs) measurements were carried out to assess the antibacterial characteristics of the nanoparticles over 48 hours. An initial Optical Density OD600 of 0.015 was selected (58) and was kept constant for all experiments to ensure reproducibility.

Optical Density Measurements Without and With Nanoparticles
Optical density (OD) measurement is a widely used method to assess the number of growing bacteria in a culture; thus, the absorbance values of bacterial suspensions can be measured using a photometer (59). The initial optical densities of each bacterium type were measured using the Jenway 6305 UV/Visible Spectrophotometer at 600 nm (OD600

Thermal Analysis
Simultaneous Thermal Analysis (Figure 1) of FUNP nanoparticles revealed gradual weight loss from 100 ˚C due to the removal of trapped moisture and atmospheric gases, i.e. water and carbon dioxide (61). The three exothermic peaks (i) 280˚C, (ii) 385˚C and (iii) 815˚C are likely to be associated with the combustion of organic residues, e.g. NaOH and C2H5OH (62) as calcination allows for the removal of any impurities leading to the formation of crystalline nanoparticles. No weight loss is observed above 600 ˚C, suggesting the formation of crystalline nanoparticles as confirmed by sample C815 X-ray diffraction pattern (Figure 4). Temperature plays a significant role with regards to nanoparticle structural changes, i.e. higher calcination temperatures often lead to larger nanoparticle sizes (63) but also affects the Ce 3+ and Ce 4+ ionic ratio. The literature mentions that antibacterial properties of nanoparticles are dependent upon ionic ratio, and particle size, i.e. the smaller the particle size, the better the overall antibacterial properties (64)(65)(66). Ostwald Ripening (62) or the coalescence of smaller particles (agglomeration) could lead to a temperature-dependent particle size increase. The structural changes occurring at the endothermic peaks are investigated by heat-treating the FUNP nanoparticles calcined at 280˚C, 385˚C and 815˚C.

Fourier Transform Infrared Spectroscopy
FTIR spectra for the synthesised cerium oxide nanoparticles are shown in

Ultraviolet-Visible Spectroscopy
Absorbance spectra for the cerium oxide nanoparticles are shown in With A being the absorption, α refers to the absorption coefficient, hʋ is the photon energy (1240/λ), and Eg relates to the bandgap. Therefore, the bandgap energies are determined from the x-axis intersections of (αhʋ) 2 vs hʋ plots (Figure 3 (b-e)).

X-ray Diffraction
The crystalline structure, i.e. composition, crystallite size and microstrain in the nanoparticles have been determined from the diffraction patterns (Figure 4). The XRD diffraction spectra for the synthesised nanoparticles exhibit eight major characterisation peaks located at 28.  Table 2. Consequently, the full width at half maximum (FWHM) of XRD peaks decreased as the crystallinity of the nanoparticles increased with temperature. The reduction of FWHM peaks confirms that the calcination temperature affects the size and morphology of synthesised nanoparticles.  (7) where 'a' refers to the FCC lattice parameter, 'd' corresponds to the crystalline face spacing and 'hkl' are the crystalline face indexes. The results are displayed in Table 2. The comparison of the crystallite sizes determined via the Scherrer and Williamson-Hall methods are also displayed in Table 2, where significant differences are observed between the values. The crystallite size noticeably increases with increasing calcination temperature, which is consistent with increasing particles size as confirmed from TEM analysis in Figure 5 and also by the decreasing Brunauer Emmett Teller (BET) surface area results as displayed in Table 2.
The intensities of the calcined nanoparticles increased with increasing calcination temperatures from 280 ˚C to 815 ˚C and can be attributed to crystallinity improvement of the nanoparticles confirmed from crystallite measurements displayed in Table 2. Consequently, the full width at half maximum (FWHM) of XRD peaks decreased as the crystallinity of the nanoparticles increased with temperature; hence the reduction of FWHM peaks confirm that the calcination temperature affects the size and morphology of synthesised nanoparticles.

Transmission Electron Microscopy
The majority of synthesised nanoparticles are spherical except RNP4 and the C815 nanoparticles. The smaller the size of the nanoparticles, the higher the agglomeration thus  Table 3. The presence of lattice fringes is an indication of the crystalline nature of the synthesised nanoparticles. The Selected Area Electron Diffraction (SAED) patterns for all samples except RNP4 and C815 nanoparticles depict continuous ring patterns (Figure 5h). The presence of discrete rings is an indication of the polycrystalline structure of the nanoparticles but also the relatively small size of the nanoparticles. All the obtained SAED rings are in good agreement with the XRD diffraction patterns, confirming the fluorite structure.

X-ray Photoelectron Spectroscopy
The XPS measurements revealed significant differences in surface chemistry of the synthesised nanoparticles likely related to the drying method (furnace or freeze-drying) and calcination temperatures. It should be noted that XPS analyses the surface charge of the nanoparticles, whereas EELs analyses' the subsurface thus, is likely to differ regarding the Ce 3+ :Ce 4+ ratio.  Table 4 displays the Ce 3d data for the FRNP, FUNP, C280, C385 and C815 cerium oxide nanoparticles. The spin-orbit doublet peaks associated with all the Ce 3d spectrums are Ce 3d5/2 located between ~870 eV to 895 eV, and Ce 3d3/2 located between ~895 eV to 915 eV.
Additionally, multiple shake-up and shake-down satellite peaks are also present (75). The XPS spectra indicate the co-existence of Ce 3+ and Ce 4+ due to the presence of satellite peaks linked to each oxidation state. Table 4

Electron Energy Loss Spectroscopy
Cerium oxide is electropositive and can exist in two oxidation modes, i.e. Ce 3+ and Ce 4+ . The EELs measurements were acquired by rastering the beam across several locations for each nanoparticle sample. EELS spectra from two controls (Ce 3+ and Ce 4+ ) and five synthesised nanoparticles are shown in Figure 6. In each case, the background was removed, and Fourier-Ratio Deconvolution routines were applied prior to fitting. From the EELs spectra, the Ce M4,5 edges of FRNP, FUNP, C280, C385 and C815 spectra are shown where the black lines represent the sample data, the red lines represent the Ce 4+ ion spectra, and the grey lines represent the Ce 3+ ion spectra.
It is evident as the size of the nanoparticles increases for calcined samples the temperature-dependent; therefore, the optimal calcination temperature must be characterised to ensure optimal Ce 3+ :Ce 4+ , which provides the highest antibacterial efficacy. Based on the XPS and EELs results, the FRNP, C385 and C815 nanoparticles are selected due to the range of particle size and Ce 3+ :Ce 4+ ratio to be further investigated.

Optical Density Measurements with Nanoparticles
To investigate the antibacterial properties of nanoparticles with varying Ce 3+ :Ce 4+ ratio and  (66). Several studies also found cerium oxide exhibited cytotoxic (19) and induced toxicity against Escherichia coli (19,80). Reduction of bacterial growth has been observed for coated nanoscale ceria which inhibited 50 % Pseudomonas aeruginosa growth (81). Furthermore; cerium oxide nanoparticles have also exhibited moderate inhibitive activity against Escherichia coli and Bacillus subtilis (82)(83)(84).
The half-maximal inhibitory concentrations (IC50) calculated from linear regression models are shown in Table 5. It is clear smaller sized nanoparticles, i.e. < 10 nm exhibit improved overall antibacterial efficacy against both Gram-positive and Gram-negative bacteria. The nanoparticles presented enhanced antibacterial activity against the Gram-negative bacteria, i.e.
Escherichia coli and Pseudomonas aeruginosa, which is likely to be attributed to the structure of the bacterial cell wall. The cell wall of Gram-positive bacteria, e.g. Staphylococcus epidermidis, is composed of a relatively thick continuous peptidoglycan containing peptides and linear polysaccharide chains that is difficult for the nanoparticles to penetrate. However, the cell wall of Gram-negative bacteria is composed of a thin layer of peptidoglycan with a lipopolysaccharide surrounding the bacteria. Therefore allowing the nanoparticles to cause more damage Gram-negative bacterial cell walls resulting in cell lysis, other studies (85,86) also found similar results.