Influence of the shaker mill in the properties of ZnO processed by high energy milling

This work evaluates how the High Energy Ball Milling (HEBM) in a shaker mill influences the optical, physical, and microstructural properties of ZnO. The procedure also combines Fe inclusion from the grinding medium with particle size reduction. ZnO powder was milled by 1, 2, 3, 4, and 5 h, which resulted in a particle size reduction to the nanometric scale with a mean size of around 50 nm and a crystallite size reduction by three times when processed from 4 h. Milling has proven to be an efficient process for obtaining nanoparticles with an incredibly short processing time and changed the morphology of the particles from random to spherical shapes. Results also indicate the processing progressively expanded the ZnO hexagonal structure due to the imposed strain and Fe inclusion, which can help to decrease the bandgap and slow down the recombination rate of the electron-hole pairs, improving the photocatalysis activity. The optical results showed no additional band appeared due to milling processes and diminished the bandgap from 3.37 to 3.21 eV. Milling also led to an increase in the c value from 5.2076 to 5.2112 Å, which is one of the most important factors for improved antibacterial activity. HEBM has proved to be a suitable process for obtaining ZnO nanoparticles with properties useful for various applications.


Introduction
ZnO is a ceramic material and II-VI semiconductor (as it is formed by a metallic IIB and a non-metallic VIA element), with n-type semiconduction. It has a wide band gap (Eg = 3.37 eV) and high exciton binding energy (60 meV) at room temperature (Fan, Sreekanth, Xie, Chang, & Rao, 2013;Samavati et al., 2021;Toporovska et al., 2020). These characteristics make it interesting for several applications in electronic and optoelectronic devices, such as lasers, sensors, photocatalysts, solar cells, transistors, and antibacterial materials, as they can be activated by UV light to release reactive oxygen species in an aqueous solution (Noman, Amor, & Petru, 2021;Otis, Ejgenberg, & Mastai, 2021;Wang, Zhu, Huang, Mei, & Jia, 2019). Among the possible crystalline defects in the structure of pure ZnO (O vacancy, Zn vacancy, interstitial Zn, interstitial O, substitutional Zn), interstitial O and Zn vacancies are the most frequent, which leads to the predominance of ntype semiconduction without the need for external doping (Fan et al., 2013). Most binary semiconductors II-VI crystallize in cubic zinc blende and/or hexagonal wurtzite structures, where the tetrahedron is formed by one anion surrounded by four cations at the corners (Morkoç & Özgür, 2009). This structure provides a mixed character between ionic and covalent, which raises the bandgap energy above that expected when only covalent bonds are present. For ZnO, the possible crystal structures are, besides the zinc blende and the wurtzite, the cubic rock-salt structure. Under normal temperature and pressure conditions, wurtzite is the most thermodynamically stable crystal structure. The zinc blend is metastable and obtainable under specific processing conditions. Similarly, rock-salt can only be obtained under high pressure, so that the interaction between the atoms is forced as much as possible to become predominantly ionic (Lee, Lai, Ngai, & Juan, 2016;Morkoç & Özgür, 2009).
The properties of ZnO are widely attributed to its particle size, the amount of the grains and grain boundaries (Aimable, Goure Doubi, Stuer, Zhao, & Bowen, 2017;QIN, SHAO, LIU, & WANG, 2005) and also the concentration of defects (B. Chen, Xia, & Lu, 2013;Lott et al., 2015) Thus, electrical and optical properties of ZnO have been improved by maintaining its nanometer-scale microstructure (Lee et al., 2016). The interest in nanostructured materials is growing due to the gain in several properties, such as optical, mechanical and catalytic, among others (Aimable et al., 2017;Phuah, Rheinheimer, Akriti, Dou, & Wang, 2021;Samavati et al., 2021). By starting the processing step with a nanoscale material, it is possible to refine the microstructure. With refinement, several interesting properties can be controlled, such as electrical properties (due to higher electron mobility), optical properties (due to access to a wider range of radiation absorption), mechanical properties, among others (Choi, Jung, Shin, & Sohn, 2015;Lee et al., 2016;Lott et al., 2015;Phuah et al., 2021;Samavati et al., 2021).
There are several techniques for obtaining semiconductor powders in nanometric scale. The existing methods include Research, Society andDevelopment, v. 10, n. 12, e476101220855, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i12.20855 chemical, thermophysical, mechanical synthesis, and mechanical milling, as in the case of high energy ball milling (HEBM) (Pawan Mayo, Hague, & Chen, 1993;Noman et al., 2021;Samavati et al., 2021). Among the high energy mills currently used in HEBM, it can be mentioned shaker mill, planetary ball mills, attritor mills, and commercial mills. They differ in capacity, milling efficiency, and additional systems for heating and cooling (Dutková et al., 2016;Suryanarayana, 2001). Shaker type mills movements cause friction between the balls inside the vial and ball-wall friction, promoting the mixing and grinding of the material under high impact forces, characterizing it as high energy (Suryanarayana, 2001). HEBM is a consolidated technique among the routes for obtaining fine powders (Dzik, Svoboda, Kaštyl, & Veselý, 2019). Also, it consists of a technique for doping ceramic oxides.
Some studies have reported applications of ball-milled powders (Bégin-Colin et al., 2009;Dias et al., 2016;Otis et al., 2021;Reddy et al., 2019;Šepelák, Bégin-Colin, & Le Caër, 2012). For ZnO semiconductors, besides other important properties such as size and surface area, the exposure surface and the surface defects play an important role in its applications (D. Chen et al., 2014;Zhang et al., 2014). ZnO nanoparticles increases the photocatalytic, antibacterial, and antimicrobial activity (Noman et al., 2021). Salah et al (2011) (Salah et al., 2011) firstly applied HEBM through a planetary mill up to 50h for reducing the particle size of ZnO to nanometric scale from a microcrystalline powder, suggesting the modifications induced in ZnO nanoparticles might prove useful as antibacterial material. Gancheva et al (2013) (Gancheva, Iordanova, Dimitriev, Avdeev, & Iliev, 2013) used the same milling to evaluate the photocatalytic activity of ZnO after milling for 10h with different milling speed rotation (Gancheva et al., 2013). Chen et al (2014) (D. Chen et al., 2014) stated that the defects promoted during HEBM highly influence the photocatalytic activity of ZnO with increased time and rotation due to the bulk defects promoted during the milling process. Mekprasart et al. (2020) (Mekprasart, Chutipaijit, Ravuri, & Pecharapa, 2020) evaluated different time of processing and milling rotation on the ZnO size reduction as a top-down process with high efficiency to achieving nanoparticles.
Additional to particle size reduction, the mechanic activation by milling processes was also applied to the incorporation of ions to improve the photocatalytic efficiency of ZnO (Kotha, Kumar, Dayman, & Panchakarla, 2021). This process can help decrease the bandgap and slow down the recombination rate of the electron-hole pairs. Güler et al (2016) (Güler, Güler, Evin, & Islak, 2016)  ZnO, no discussion concerning shaker mills has been found in the literature.
The purpose of this study was to investigate the applicability of HEBM in the particle size reduction of the ZnO powder with an increase in the number of particles at a nanometer scale. Also, it aimed to investigate if the generation of higher surface area, until then unexposed, (one of the most important properties to evaluate in samples that have undergone milling processes) will influence on the ZnO's physical and optical properties. Since the milling was carried out in steel vial and ball, the incorporation of Fe ions was also analyzed.

Methodology
The present work consists of an experimental research based on a qualitative approach and the data were discussed on the lights of descriptive-diagnostic analysis. Zinc Oxide powder (ZnO, 99% purity, Synth®) was used as the material of study for the milling procedures. ZnO commercial was characterized by X-ray diffraction. The analyses were performed with Cu Kα radiation in 0.1542 nm, at 40 kV and 30 mA, scanning between 15 and 75 º with an angular step of 0.02º in continuous mode.
The powder was dry milled in a SPEX 8000 shaker/ Mill, in which hardened steel vial and ball (5 mm in diameter).
The ball/powder mass ratio was 5:1, and the milling times were 1, 2, 3, 4, and 5 h, according to previous works (Pallone et al., 2010;Yokoyama, 2008). After finishing the milling process, the fine powder was leached with 0.5 M oxalic acid (Anidrol, analytical grade), for the removal of surface Fe from the grinding medium. The removal of the excess Fe residue will favor the analysis only of the interference of the ions incorporated in the ZnO. For the acid leaching, the ratio of 3 g of the powder to 100 mL of oxalic acid solution was used. The mixture was left under stirring and heating to approximately 100 ºC for 4 hours.
After the leaching, the solution was kept at rest for the separation of the liquid, through the decanting of the powder. Once separated, the powder was led to another leach, so that the procedure was performed three times for each sample. After leaching, the powders were washed to remove the residual acid, by stirring and heating in distilled water for 30 minutes, repeating the procedure three times with water exchange at each repetition. Isopropyl alcohol was added after the last wash and drying, and the suspension was allowed to stir again for thirty minutes. The samples were then deagglomerated in isopropyl alcohol medium in Ultrasonic Sanders bath for 25 minutes and slow dried at 50ºC. After drying, the powder was sifted through a 45 μm opening sieve. The leaching procedure was adapted according to previous works (Hernández, García, Cruz, & Luévanos, 2013;Silva et al., 2008).
The milled powders were characterized by X-ray diffraction at the same condition as previously described. The crystallite average size was determined using the Williamson-Hall method. Assuming that the microstress is uniform in all crystallographic directions according to the uniform deformation model (UDM), Equations A is used for calculus (Mote, Purushotham, & Dole, 2012;Prabhu, Rao, Kumar, & Kumari, 2014;Williamson & Hall, 1953): where β is the full width at half maximum of the diffracted peak (FWHM); θ is the Bragg diffraction angle; K is a shape value constant, dependent on the morphology of the particle; λ is the wavelength of the used radiation (Cu Kα = 0.1542 nm); ε is the component of crystalline microdeformations; D is the crystallites average size. From Equation A, when considering linear function y=ax+b, where y corresponds to β*cosθ and x to 4*sinθ, the angular coefficient of the linear adjustment performed for the experimental points will be equivalent to the microdeformation ε and the linear coefficient will correspond to (K*λ)/D. Differential Scanning Calorimetry (DSC) and Thermogravimetric Analysis (TGA) were simultaneously carried out in a Netzsch Jupter STA 449F3 equipment, between 200 and 1100 ºC, under Nitrogen atmosphere, with a heating rate of 10 K/min. Also, Fourier Transform Infrared Spectrometry (FTIR) was performed in a Perkin-Elmer Spectrum 1000, with transmittance measurements performed in the range of 4000 to 450 cm -1 , with KBr-pressed samples.
Semi-quantitative chemical analysis via X-ray fluorescence (FRX) was performed using the Axios PW 4400/40 DY electron microscopy (SEM) in a Magellan 400 L equipment.
The optical analysis was evaluated by Diffuse Reflectance Spectroscopy (DRS), performed in a Varian Cary 5G spectrophotometer between 250 nm and 800 nm. The samples' optical band gap was calculated by the Tauc's Plot as described by Dias et al (2018) (Dias, Oliveira, Renda, & Morelli, 2018). For nomenclature purposes, the samples were prefixed Z followed by the processing time number. The ZnO as purchased is named ZnOP (pattern). The samples that underwent the leaching process received the suffix L. Also, some samples that underwent heat treatment received the terminology T followed by the used temperature. Heat treatments were carried out at temperatures of 600, 800, 900 and 1000 ºC. The analysis of the particle size distribution was performed by laser diffraction (Horiba -LA-950V2), in alcoholic medium, with prior 30 min in the ultrasound, considering the number of particles as distribution basis.

Results and Discussion
The results will be presented according to the characterization steps, corresponding to the precursor powder, then to the processing, and finally to the processed material. The results for X-ray diffraction of the precursor and milled powder are presented in Figure 1. Research, Society andDevelopment, v. 10, n. 12, e476101220855, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i12.20855 According to Figure 1, the ZnOP (pattern) presented the hexagonal crystalline structure, corresponding to the mineral Wurtzite (ICDD #2003-6-1451). Figure 2 (a) shows the comparison of the diffractograms of the milled ZnO samples from 1 to 5 h and their respective precursor powder. It can be seen the main diffraction planes present in all samples. All diffractograms were identified according to the same precursor oxide card (ICDD #36-1451). As described in the literature (Dib, Trari, & Bessekhouad, 2020;Promod Kumar et al., 2020;Poornaprakash, Chalapathi, Subramanyam, Vattikuti, & Park, 2020), this crystalline structure corresponds to the active phase for most required ZnO applications. Small peaks (on the magnification of Figure 1 (a)) indicate Fe presence (ICDD # 6-696). The Fe peak intensity has increased following the increase in the milling time. The difference between the oxides before and after milling is related to the enlargement of the diffracted peaks for the milled samples. Analyzing the diffractograms of the processed samples (Figure 1(a)), as the milling time increases, there is a crystallinity loss of the samples due to the induced deformations. The peak broadening may occur as a function of crystallite size decrease, lattice distortion and crystalline defects (Mote et al., 2012;Prabhu et al., 2014;Williamson & Hall, 1953). Table I describes the crystallite average size for each sample processed by HEBM according to the W-H method, where the crystallite size contribution to the peak broadening is calculated separately to the lattice strain.   Adequate adjusts were obtained, as exemplified for ZnOP (Fig. 3a) and Z5 (Fig. 3b). The residual between the calculated and experimental data shows a good refinement was achieved, and also the Rwp and χ 2 values show high-quality and trustworthy results. The hexagonal strucure was progressively expanded according to the milling time, since the lattice parameters a, c, and the structure volume V increased after processing, as shown in Figure 2(c). The structure was schematized in Fig.3c, utilizing the crystallographic coordinates from Schreyer at al. (Schreyer, Guo, Thirunahari, Gao, & Garland, 2014).
The lattice suffered a greater expansion in the first hour. Next, the expansion continued occurring, but in a slower rate. This phenomenon can be attributed to two concomitant processes: the strain due to the impact and shear forces promoted by HEM; and the Fe cation structural inclusion from the vessel and milling spheres. The former occurred more intensively in the first hour, observed by a significant increment of the diffraction peaks broadening compared to the raw oxide. The latter is mainly attributed to the cationic structural inclusion, which can also lead to lattice expansion. Fe 2+ has a larger crystal radius than Zn 2+ in tetahedral coordination (0.77 Å versus 0.74 Å, respectively) (Saleh, Prakoso, & Fishli, 2012;Shannon, 2011), therefore, isovalent inclusions promoted the lattice expansion in a minor extent.
According to Rietveld analysis, the c value has increased from 5.2075 Å to 5.2112 Å. The increase in the c value typically leads to high amounts of H2O2 release which can inhibits bacterial growth. Salah et al (2011) were able to increase the c value of ZnO from 5.204 to 5.216 with 50h of milling in a horizontal oscillatory mill (Retsch, PM 400). Therefore, the use of a shaker mill might reduce the time of processing to produce ZnO with adequate characteristics for antibacterial applications.
FTIR was performed to identify the functional groups and chemical bonds present in the samples after leaching. Figure 3 shows the FTIR spectral referring to ZnOP and ZnO processed via HEBM and submitted to acid leaching and Table 2 shows the FTIR band assignments. 1632 νas (CO) (Cooper, 2015;Wladimirsky1 et al., 2011) 1364 νs(CO) + νs(CC) (Georgeta et al., 2015;Wladimirsky1 et al., 2011) 1320 νs(CO) + δ(OCO) (Cooper, 2015;Georgeta et al., 2015) 823 νs(CC) + δ(OCO) (Cooper, 2015;Wladimirsky1 et al., 2011)  However, for ZnO submitted to milling and leaching, it was observed the appearance of several bands, which were identified and related to the probable chemical groups responsible for such interactions, according to the FTIR spectra available in the Research, Society andDevelopment, v. 10, n. 12, e476101220855, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i12.20855 9 literature for hydrated zinc oxalate (Table 2) (Cooper, 2015;Georgeta, Marcela, Anda, Gratiela, & Alexandrina, 2015; Wladimirsky, Palacios, María C. D'Antonio, & Ana C. González-Baró and Enrique J. Baran, 2011). In table 2 one can see the characteristic wavenumber of each spectral trace present in Figure 3 and its respective assignment. According to the mentioned literature of each attributed assignment, these bands can be of stretching, antisymmetric stretching, symmetric stretching, bending or rocking vibrations for the functional groups of the zinc oxalate hydrate. From the FTIR spectra, although the Z4L sample presented different x-ray diffraction peak, according to the infrared technique it presents the same chemical groups as the others, thus confirming the presence of interactions of the oxalate group with the zinc oxide used. To reverse the formation of zinc oxalate crystals formed during the acid leaching stage, it was performed for the Z3L sample (chosen as an intermediate between the milling times) the DSC/TG, presented in Figure 4 (a). According to Figure (a), there are two significant events showed by the coupling of the mass loss and heat flow curves: the presence of an endothermic peak at approximately 150ºC accompanied by a mass loss, corresponding to the vaporization of the hydration water; the second event, at 400ºC, also characterized by an endothermic peak accompanied by a significant mass loss corresponding to the decomposition of oxalate. Zinc oxalate can be thermally converted to zinc oxide after approximately 400 º C. Aiming to reverse the effect of the acid leaching a heat treatment at 600ºC was proposed for both samples: those submitted only to milling (also seeking to mitigate the effects of crystallinity loss due to HEBM) and the leached ones. Figure 4 (b) shows the FTIR spectrum for the heat-treated samples. The ZnOP sample is presented just for comparison and shows a water band around 3383 cm -1 , which is absent on the other samples. Regarding the Zn-O bonding band, between 600 and 400 cm -1 , it is now present on the heat-treat samples and was not on the leached ones (Figure 3). It confirmed the heat treatment at 600 ºC was able to reverse the oxalic acid formation. Comparing the bands' intensity in Figure 4 (b), one can see that for both leached and heat-treated samples (Z1LT600 and Z5LT600), the Zn-O band, although present, was weaker than the just heat-treated ones. The intensity of the oxide bond was decreased according to the experimental procedures performed and had its lowest intensity for the sample leached and processed for the longest milling time. This fact corroborates the XRD analysis that the milling and the leaching processes led to deformations on the ZnO structure. Particle size distribution ( Figure   5) by laser diffraction was performed for the samples milled for 2 and 5 h, before and after heat treatment at 600 ºC. The results showed in Figure 5 demonstrated the milling process was indeed able to reduce the particle size. When increasing the milling time from 2 to 5 h, the particle size was reduced completely to a nanometric scale (green curve), with an average size of approximately 50 nm. However, the heat treatment for both milling times provided the formation of aggregates and partially reversed the milling action, since the curves for the Z2T600 and Z5T600 showed particle size distribution in a higher scale. However, even with the heat treatment, the particle size did not reach the initial value of the ZnO precursor powder. To confirm the results of the presence of Fe, the X-ray fluorescence analysis is described in Table 3 for the samples processed for 2 and 5 h (Z2 and Z5). Comparing the leached samples obtained for both processing times according to the results presented in Table 3, an increase in the percentage values of the other compounds (Al2O3 and SiO2) is observed. This fact can be explained, however, when observing the percentage values of ZnO. After acidic leaching, there was a decrease in the percentage of this oxide in the samples. This phenomenon suggests two events: the elimination of a small portion of ZnO solubilized in the acid during the leaching process and also the loss of some ZnO nanoparticles that have not been able to decant after the leaching. Therefore, Research, Society andDevelopment, v. 10, n. 12, e476101220855, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i12.20855 11 the presence of Fe in the samples is confirmed even after the leaching process. It can also be noticed the increased in the Fe content according to the milling time, since Z2 presented 1.38% and Z5 presented 2.27%, indicating that an increase in 3 h of processing led to an increase of almost 65% of iron incorporation.
The SEM images of the precursor powder and the milled and leached powder are demonstrated in Figure 6 and 7, respectively. The ZnO used as raw material consisted of particles predominantly on a submicron scale (between 100 and 600 nm, as shown in Figure 6). It can also be seen that the morphology of the particles is randomly shaped, more predominantly hexagonal prism and thin rods.  Research, Society andDevelopment, v. 10, n. 12, e476101220855, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i12.20855 13 shapes. This behavior was also reported in our work, where one can see by the SEM images the presence of spherical shapes rather than the hexagonal shapes presented by the precursor powder, which might also be useful for various applications. In order to verify the changes on the optical properties, the DRS spectra are shown in Figure 8.  (Akhundi & Habibi-Yangjeh, 2016;Shekofteh-Gohari & Habibi-Yangjeh, 2017). The high processing time could be reduced by replacing the planetary mill by the shaker mill (as used in this work), since the energy and frequency of impact is much higher in the latter (Suryanarayana, 2001).
As shown in previous works, interest in the applications of ZnO nanoparticles has grown in recent decades as it exhibits properties that can be applied to different fields, as biomedical, sensor and catalysis. The increased specific surface area and the reduced particle size leads to enhanced particle surface reactivity. Bottom-up approaches for obtaining nanoparticles (such as vapor synthesis, pyrolysis, sol-gel synthesis, and precipitation methods) are well stablished in the literature, enabling control of the particle size and morphology. However, for optimized conditions, these processes can present high complexity and costs. In this sense, top-down approaches (as mechanical milling) consist of a simple operation, coupled with low cost and possibility of high amounts of material been processed at the same time (S. Sharma, Pande, & Swaminathan, 2017). In relation to high energy ball milling, this is the first time in the literature that a shaker mill was applied to ZnO particle size reduction and physical characterization. High energy ball milling with a shaker mill proved to be suitable in achieving nanoparticles with great optical, physical and microstructure properties with a short time of processing. This could broaden the applicability of ZnO nanoparticles processed by HEBM, such as photocatalytic, antibacterial and antimicrobial applications.
Also, crystallographic orientation plays an important rule regarding antibacterial activity and the paper also showed that an increase in the c value was able to be achieve in a reduced time compared to previous work with other milling type.

Conclusion
High energy milling using a shaker mill resulted in particle size reduction for both processing times with an increase in milling time. The oxalic acid leaching process presented strong interaction with the ZnO, leading to the formation of zinc oxalate which was reversed by heat treatment. The heat-treated powder, however, presented lower interaction between the Zn-O bond. Rietveld refinement showed the hexagonal structure progressively expanded according to milling time. Also, the c value increased, which might be useful for antibacterial and antimicrobial applications. The band-gap decreased to 3.21 V with 5h of milling, which might be useful for excitation events. High-energy milling using a shaker mill proved to be an efficient technique in reducing particle size in a significantly shorter processing time than other conventional mills.
HEBM showed how the process can promote structural changes and enhance the properties of the final processed powder. This process could help researches to overcome some challenges and future prospects to the field of nanostructured materials. Further investigations aiming at visible light applications to ZnO, by adding metal dopants (such as Sn, Mn, and Co) or semiconductors (such as TiO2, SnO2, and Co3O4) to the high energy milling process to form doped ZnO or ZnO heterostructures, respectively, can be an important route to broad and reduce the costs of ZnO application.