Large Scale Synthesis of ZnO Nanostructures of Different Morphologies through Solvent-free Mechanochemical Synthesis and their Application in Photocatalytic Dye Degradation

Corresponding Author: Umapada Pal Instituto de Física, Benemérita Universidad Autónoma de Puebla, Apdo. Postal J-48, Puebla, Pue. 72570, Mexico Email: upal@ifuap.buap.mx Abstract: Well crystalline, near-stoichiometric zinc oxide nanostructures of different morphologies are fabricated in large scale, utilizing a simple, cost effective mechanochemical synthesis in absence of solvent. Effects of ionic and nonionic surfactants along with the concentration of hydrolyzing agent on the evolution of nanostructure morphology have been studied. It has been observed that while a cationic surfactant such as Cetyltrimethylammonium Bromide (CTAB) favors the nanostructures to growth along their polar c-axis, a nonionic surfactant such as Polyvinylpyrrolidone (PVP) reduces their preferred c-axis growth. Increase of hydrolyzing agent in the reaction mixture enhances the growth rate of the nanostructures. The nanostructures have been tested for photodegradation of anionic dye molecules such as Methylene Blue (MB). All the nanostructures manifest high photocatalytic performance. Apart from morphology, the specific surface area, crystal plane orientation and the concentration of basic sites at surface are seen to contribute significantly to the photocatalytic performance of the zinc oxide nanostructures.


Introduction
Recent thrust of metal oxide nanostructure research is driven by their application potentials in diverse fields such as gas sensing (Hsu et al., 2008;Kumar et al., 2015;Xu et al., 2000), optoelectronic device fabrication (Hsu et al., 2008;Olson et al., 2007;Valerini et al., 2008;Zimmler et al., 2008), catalysis (Matsamura and Ishibe, 2009;Lorenz et al., 2013;Spencer 1999) photocatalysis (Jia et al., 2015;Morales Flores et al., 2014;Ruiz Peralta et al., 2012;Yang et al., 2015) and dye adsorption (Chen et al., 2015;Khosla et al., 2015;Norman, 1970). Due to high thermal and chemical stability and reasonable biocompatibility, the application of metal oxide nanostructures extended further to biology and medicine (Hahn et al., 2012;Hilger and Kaiser, 2012;Solanki et al., 2011). While for each of these applications metal oxide nanostructures of specific nature and characteristics such as size, composition, nature of dopant and morphology are frequently preferred, apart from some basic characteristics such as band gap energy, excitonic energy, photoresponse and specific surface area, the role of other characteristics has not been associated to their performances with certainty. To elucidate the facts, we can consider a specific material, such as zinc oxide, which is a metal oxide semiconductor of high band gap energy (3.37 eV at room temperature) and high excitonic energy (60 meV). In nanostructure forms, it is very much suitable for fabricating optoelectronic and piezoelectric devices (Jalali et al., 2013;Solanki et al., 2011;Wang and Song, 2006), photocatalytic degradation of organic materials (Morales Flores et al., 2014;Ruiz Peralta et al., 2012) and catalytic supports (Lorenz et al., 2013;Matsamura and Ishibe, 2009;Spencer, 1999). ZnO nanostructures of different morphologies and sizes have been fabricated using a vast number of chemical and physical methods and the roles of their morphology, defect structure and surface area have been studied extensively, especially in the context of later applications (Jalali et al., 2013;Pan et al., 2014;Ruiz Peralta et al., 2012;Wang and Song, 2006). While through physical evaporation process, 1D ZnO nanostructures such as nanorods, nanowires and nanobelts have been fabricated He et al., 2006;Wang et al., 2005;2004a;2004b), through chemical processes such as sol-gel (Tak and Yong, 2005), hydrothermal and other solution based synthesis (Chen et al., 2000;Ni et al., 2005;Shao et al., 2008;Xie et al., 2011), ZnO nanostructures of a large variety of morphologies have been fabricated successfully. Utilizing amine-assisted nonhydrolyzing synthesis method, Zhang et al. (2007) could fabricate ZnO nanostructures of different morphologies using zinc acetate precursor. On the other hand, utilizing the simple and cheap mechanochemical technique, ZnO nanostructures of different morphologies (Anand et al., 2014;Tsuzuki and McCormick, 2001;Shen et al., 2003) could be synthesized with good control over their shape and size. In fact, a vast number of reports have been published on the morphology control of ZnO nanostructures through hard and soft chemical routs and their photocatalytic performances for organic dye degradation under UV illumination. While controlled fabrication of ZnO nanostructures of different morphologies and specific surface areas through cheap and soft chemical routs remains a challenge, understanding the roles of morphology, surface area, defect content and surface acid/base sites on their catalytic and photocatalytic performance is the other task, essential for fabricating suitable ZnO nanostructures for efficient environmental applications.
In the present article, we demonstrate the fabrication of ZnO nanostructures of different morphologies in large scale through a solvent-free soft mechanochemical process, utilizing ionic and non-ionic surfactants (structure driving agents) and NaOH as precipitating/hydrolyzing agent. Effects of morphology, surface area and surface acid/base sites on their photocatalytic dye degradation performance have been studied applying Methylene Blue (MB) as a test dye. It has been demonstrated that the surfactant and concentration of hydrolyzing agent both play important roles for defining the final morphology of ZnO nanostructures. The roles of surfactant and hydrolyzing agent on the evolution of morphology and photocatalytic performance of mechanochemically grown ZnO nanostructures have been discussed.

Preparation of ZnO Nanostructures using CTAB and PVP
ZnO nanostructures were synthetized through mechanochemical process following the procedure reported by Anand et al. (2014) with some modifications. Briefly, Zinc Acetate dihydrate (ZnAc) and CTAB or PVP in 2:1 molar ratio were mixed in a ceramic mortar for about 30 min. After that, a certain amount NaOH was added to the above mixture and manually grinded for another 30 min at room temperature. The molar ratio of ZnAc and NaOH in the mixture was maintained either 1:3 or 1:9. The obtained white pastes were dispersed in water and washed with ethanol and water several times by centrifugation (8000 rpm, 10 min). Obtained samples were dried at room temperature and annealed thereafter at 350°C for 2 h in air. Utilizing 40 mmol of zinc acetate precursor, about 3.0-3.2 g of ZnO nanostructures were obtained for each of the samples. The synthesis conditions used for fabricating different ZnO nanostructures and their nomenclatures are given in Table 1. The samples were prepared at Zinc Acetate (ZnAc) and NaOH molar ratios of 1:3 and 1:9 for a fixed molar content of surfactants CTAB and PVP to study the effects of surfactant and NaOH concentration on the morphology and photocatalytic behaviors of the fabricated nanostructures. As the growth rate of ZnO nanostructures and their morphology are known to be driven by the surface energy of their crystal facets or planes, which generally get modified by surfactants apart from the pH of the reaction (Xie et al., 2009;Jang et al., 2009), we used CTAB and PVP as ionic (cationic) and non-ionic surfactants, respectively, to observe the effect of surfactant nature on morphology evolution of ZnO nanostructures.

Characterization of the ZnO Nanostructures
The crystallinity and structural phase of the nanostructures were analyzed through powder X-Ray Diffraction (XRD, Bruker D8 Discover) and micro Raman spectroscopy (Horiba Yovin Lab RAM HR). While the XRD patterns of the samples were recorded in 10-70° range utilizing CuKα (λ = 1.5406 Å) radiation in 0.02° 2θ steps, their room temperature Raman spectra were recorded using the 633.8 nm line (10 mW, 1.0 mm spot size) of a He-Ne laser as excitation source and a thermoelectrically cooled charge-coupled device as detector. For morphology and elemental analysis, a JEOL JSM-7800F Field-Emission Scanning Electron Microscope (FESEM) with Oxford X-Max analytical accessory was used. A JEOL JEM 2100F field emission transmission electron microscope (FE-TEM) operating at 200 kV was utilized to study the crystal structure and growth induced defects in the nanostructures. To estimate the specific surface area and average pore size in the nanostructures, their adsorption-desorption isotherms at 77 K were recorded in a Belsorp-Mini II (BEL Japan, Inc.) analyzer. The samples were degassed at 250°C for 5 h under vacuum prior to their analysis. The specific surface area and average pore diameter in the samples were estimated utilizing Brunauer-Emmett-Teller (BET) procedure (Brunauer et al., 1938).

Photocatalytic Evaluation of the ZnO Nanostructures
The photocatalytic responses of the nanostructures were recorded in a Shimadzu UV-3101 PC UV-Vis-NIR double-beam spectrophotometer. The tests were performed in aqueous solution using methylene blue (MB, C 16 H 18 ClN 3 S•3H 2 O, Sigma-Aldrich) as a test contaminant. A cylindrical jacketed glass mini-reactor of 500 mL capacity with 50 mm internal diameter and about 120 mm of height was utilized to perform batch reactions. Room temperature MB photodegradation rate of the catalysts was monitored by dispersing 20 mg of powder ZnO sample in 40 mL of MB solution (10 ppm).
The catalyst-containing reaction mixture under magnetic agitation was illuminated by a 10 W ultraviolet LED lamp (model JX-10UV9X1B365) of 365 nm emission. The lamp was placed at the top of the reactor, about 13 cm above the test solution surface. The temperature of the reaction mixture was maintained fixed at 25°C by circulating cold water through the jacket of the reactor. Oxygen was bubbled through the reaction solution under UV illumination.
The MB concentration in the reaction mixture was monitored by a computer-controlled spectrophotometer at regular intervals, withdrawing each time about 3.5 mL of the solution (aliquot) from the reactor. To separate the dispersed catalyst from the aliquot, a reusable syringe filter holders (z268410) with nitrocellulose membrane filter of 0.22 mm pore size was utilized. From the absorbance spectra of MB recorded at different time intervals, the concentrations of MB in the reaction solution were estimated using a pre-calibrated concentration curve. The MB concentration in the reaction mixture was determined from the intensity of the most prominent absorption band of MB around 664.5 nm. Figure 1 presents the X-ray diffraction patterns of the samples prepared at different ZnAc:NaOH ratios in absence and presence of surfactants (Table 1). As can be seen, all the samples revealed well resolved, high intensity diffraction peaks correspond to crystalline ZnO in hexagonal wurtzite phase (JCPDS card no. 79-0207). The relative intensity of the main diffraction peaks like (100), (002) and (101) varied a little from sample to sample. No significant shift in the position of the diffraction peaks was observed for the samples prepared at different precursor-NaOH molar ratios or for different surfactants. Utilizing the most intense (101) diffraction peak, average grain size (P) in the ZnO nanostructures was estimated using the Debye Scherrer equation (Cullity and Stock, 1956):

Results and Discussion
where, K is the shape factor considered to be 0.9, λ is the X-ray excitation wavelength (λ = 1.5406 Å), β is the FWHM of the (101) peak and θ is the Bragg angle.
Estimated average grain size values for the ZnO nanostructures are presented in Table 1. In general, the presence of surfactant (of either nature) enhances the average grain size of the nanostructures. However, apparently the nature of surfactant does not affect their crystallite size. A decrease of average grain size for the nanostructures prepared with higher NaOH content (higher pH of the reaction) is probably due to higher growth rate of the nanostructures at higher reaction pH.  .). On the other hand, the effect of reaction pH on morphology evolution of ZnO nanostructures is very prominent for both the surfactants. As can be seen from the SEM images of the samples ZnO-3 and ZnO-5 prepared at 1:9 precursorhydrolyzer molar ratio using CTAB and PVP as surfactant, respectively, the pH of the reaction affects the morphology of ZnO nanostructures severely. While the former sample revealed echinoid like (flower consisting nanorods emanating from a common point) morphology, the later sample revealed rose-like (flower consisting petals) morphology.
The evolution of very distinct morphologies in these two samples clearly demonstrates that both the pH of the reaction and the nature of surfactant define the final morphology of ZnO nanostructures in mechanochemical synthesis.
The observed morphologies for the samples prepared at different reaction conditions can be understood considering the polar nature of ZnO, roles of hydrolyzing agent and surfactant on the modification of surface energy of its different facets or crystal planes.
The commonly accepted growth process of ZnO crystallites in basic aqueous medium can be described through the reactions (Huang et al., 2010 ). Through a selfassembly or oriented attachment of the nuclei, elongated ZnO nanostructures are formed with preferred c-axis orientation. In the case of solvent-free mechanochemical synthesis, absence of a solvent inhibits the formation of metal ions and hence the formation of Zn(OH) 2 . The reaction in highly basic condition favors a direct formation of Zn(OH) 4 2− and its fast dehydration to form ZnO nuclei. Although the growth habit of ZnO crystal is mainly driven by its intrinsic structure, external parameters such as the presence of surfactant and pH of the reaction have significant effects (Bai et al., 2010). Addition of cationic CTAB in the reaction solution not only reduces its surface tension, reducing the energy of formation of ZnO, induces the formation of CTA + -Zn(OH) 4 2− ion-pair, enhancing the transport of Zn(OH) 4 2− units (Yan et al., 2001) and hence the growth rate of ZnO nanostructures. As we see in the Fig. 2 (sample ZnO-2), incorporation of CTAB in the ZnAc-NaOH mixture of 1:3 molar ratio results ZnO nanorods of longer dimension and smaller diameters than the elongated nanoparticles formed in its absence (sample ZnO-1). Incorporation of additional NaOH in the reaction mixture results the formation of longer ZnO nanorods emanating from a common point of nucleation as seen for the sample ZnO-3. In fact, addition of excess NaOH in the reaction mixture enhances the hydrolysis and condensation rates in presence of water molecules (whatever adsorbed by NaOH during mechanochemical processing). Such uncontrolled and unselective hydrolysis/condensation leads to the formation of highly branched or flower-like structures, irrespective of the nature of added surfactant. In the case of PVP, as it is a nonionic surfactant, its addition in reaction mixture reduces surface energy of all the crystal faces of ZnO nuclei due to nonselective adherence, producing nanostructures of spherical morphology as seen for the sample ZnO-4. Again, the incorporation of additional NaOH in the reaction mixture causes a reduction of surface tension, enhancing the growth rate of ZnO. However, in contrast to the earlier case, where CTAB was utilized as surfactant, due to nonionic nature of PVP, no ionpair is formed. Instead, the excess OH − ions get attached to the polar surface of positive charge, i.e. the <001> face, probably forming a negatively charged dielectric double-layer, hindering its further growth. As a result, the growth of ZnO along its caxis reduces and the planer or flatter nanostructures of plate-like or sheet-like morphology evolve. Again, an uncontrolled hydrolysis/condensation at high NaOH concentration leads to the formation of flower-like aggregates of thin sheets. In fact, both the evolution of rod-like and plate-like ZnO nanostructures induced by surfactant (Li et al., 2008) and the pH of reaction mixture (Wahab et al., 2009) have been reported in the literature for hydrothermal and precipitation synthesis, respectively.
To study the crystallinity and lattice defects, high resolution TEM images of the typical rod-shaped (ZnO-2), spherical (ZnO-4) and petal-shaped (ZnO-5) ZnO nanostructures have been recorded (Fig. 3). As can be seen from Figure 3, all the nanostructure are well crystalline with occasional presence of lattice distortion and point defects (indicated by red arrows). Fast Fourier Transform (FFT) of selected areas of the HRTEM images of all three sample (insets of Fig. 3) revealed well defined points patterns correspond to their crystalline nature. It is interesting to note that the rod-shaped nanostructures (sample ZnO-2) contain higher structural defects than the spherical (sample ZnO-4) and petalshaped (sample ZnO-5) nanostructures. The higher defect content in the ZnO-2 sample is also revealed in its FFT pattern where the diffraction spots are elongated due to structural disorder.
The crystallinity and lattice defects of the nanostructures were studied further recording their room temperature Raman spectra presented in Fig. 4. To understand the effect of surfactant and hydrolyzing agent (NaOH) separately, the results are presented in two groups: Samples prepared using CTAB (Fig. 4a) and the samples prepared with PVP (Fig. 4b).
Nanostructures of both the groups revealed well resolved Raman bands corresponding to fundamental and multiphonon modes of crystalline ZnO. The intense bands revealed around 100 and 439 cm −1 correspond to the E 2L and E 2H modes of nonpolar optical phonons, respectively, characteristic of wurtzite hexagonal ZnO (Morkoç and Özgür, 2008). The Raman bands appeared around 381 and 582 cm -1 are the transverse optical (TO) and longitudinal optical (LO) modes of polar A 1 mode. Raman spectra of all the samples revealed two weak bands around 204 and 1150 cm −1 , which correspond to second-order processes, attributed to 2E 2L and 2A 1 (LO) phonon modes of ZnO, respectively. Moreover, all the Raman bands associated to common multi-phonon processes in ZnO such as E 2H -E 2L (at 333 cm −1 ), 2(E 2H -E 2L ) (at 665 cm −1 ) and A 1 (LO)+E 1 (TO)+E 2L (at 1090 cm −1 ) were evolved in the ZnO nanostructures, demonstrating clearly their high crystalline quality. For the samples synthetized using PVP as surfactant, there appeared two weaker bands around 538 and 143 cm −1 , which did not appear for the samples prepared with CTAB. While the earlier band is associated to E 2H +E 2L multiphonon in ZnO lattice, the later band has been found at low temperature microRaman spectra of ZnO and assigned to intrinsic defects associated with synthesis process (Cuscó et al., 2007).
As one of the aims of this study is to utilize fabricated ZnO nanostructures as photocatalyst, their surface area and texture have been estimated form nitrogen adsorption-desorption isotherms at 77 K. The adsorption-desorption isotherms of the samples presented in Fig. 5 are typical type III isotherms in Brunauer classification, with no limiting adsorption at high relative pressure. The hysteresis loops can be ascribed to type H3 of mesoporous material according to the IUPAC classification (Lu et al., 2012). Estimated BET specific surface areas of the nanostructures are presented in Table 2. As we see, the estimated specific surface area of the samples vary in between 6.9 to 16.35 m 2 g −1 ; highest for the sample ZnO-1 (prepared with no surfactant, ZnAc:NaOH = 1:3) and lowest for ZnO-5 (prepared with PVP, ZnAc:NaOH = 1:9). On the other hand, the average pore size in the nanostructures estimated using BET analysis varied in between 30.1 (sample ZnO-2) and 80.5 nm (sample ZnO-4). As it can be noticed, the estimated average pore size of most of the samples is beyond mesoporous limit (2-50 nm). Such high average pore size values estimated for the nanostructures clearly demonstrate that they correspond to inter-particle spaces, rather than the porosity of individual nanostructure. In fact, the pore size distribution of all the samples through Barrett-Joyner-Halenda (BJH) (Barrett et al., 1951) analysis revealed a prominent peak at around 1.64 nm pore radius, along with some broader distributions at higher values (insets of Fig. 4).
As it is well known, on illumination with high energy light, the valance band electrons of ZnO get promoted to its conduction band, creating electron-hole pairs (Equation 5), which diffuse to the surface of the semiconductor: The photo-generated holes near the nanostructure surface react with surface OHgroups and H 2 O molecules to produce HO • radicals (Equation 6 and 7), which get adsorbed at the catalyst surface. The surface adsorbed HO • radicals are strong oxidizing agents, which react with surface adsorbed organic dye molecules such as MB, mineralizing them partially or completely following the reaction (Equation Photocatalytic degradation of most of the organic compounds in a heterogeneous system follows first-order kinetics (Serrano et al., 2004;Yadav et al., 2007), for which the steady-state photocatalytic rate can be expressed through Langmuir-Hinshelwood relation. The photocatalytic degradation of MB by the ZnO nanostructures grown at different reaction conditions obeyed the pseudo-first-order kinetics. The photo degradation rate of MB for low initial concentration can be expressed as: Or: where, k is the pseudo-first-order rate constant, C 0 is the initial concentration of MB in reaction solution and C is the MB concentration in reaction solution after time t of UV illumination.
The reduction of MB content in catalytic solution upon UV irradiation for a typical sample (ZnO-2), the plots of C/C 0 versus irradiation time and ln(C/C 0 ) versus irradiation time for all the nanostructures are shown in Fig. 6. The linear relationship between ln(C/C 0 ) and the UV illumination time (Fig. 6C) clearly demonstrates the photo-degradation process follows first-order kinetics. The values of the pseudo-first-order rate constant k estimated from the linear regression curves (linear fits), the estimated values of t 1/2 (dye half-life) and corresponding correlation coefficients (R 2 ) for the ZnO nanostructures are presented in Table 2. We can see that the sample ZnO-2 prepared using CTAB at ZnAc:NaOH = 1:3 molar ratio has highest rate constant (k = 4.71×10 −2 min −1 ) and lowest t 1/2 (17.5 min) among the nanostructures, indicating their excellent photocatalytic activity for MB degradation. From the data presented in Table 1 and 2, we can see that all the nanostructures have near-stoichiometric composition with minor variation of average crystallite size (32-38 nm). However, due to morphological diversity, the BET surface area and BET pore size in the nanostructures varied. As apart from surface area, pore size and crystallite size, the photocatalytic activity of metal-oxide nanostructures depend on the density of basic (OH -) or acid sites at their surfaces, a quick and fast quantitative measurement of basic sites was performed following the method described by Corro et al. (2014). As the nanostructures were fabricated in highly basic conditions, the density of basic sites at their surface was very high (Table 2), which is responsible for their enhanced dye adsorption capacity. While all the nanostructures fabricated in our soft mechanochemical process present high efficiency for MB degradation under UV illumination, the sample ZnO-2 prepared at ZnAc:NaOH = 1:3 in presence of CTAB revealed best MB degradation performance. The nanostructures of sample ZnO-2 contain highest surface basic sites, bigger average crystallite size, smaller average pore diameter and moderate BET surface area (Table 1 and 2). Therefore, it is very difficult to associate their higher photocatalytic performance to any one of these texture parameters specifically, as have been practiced in a vast number of published works. Considering the morphology and growth habit of all the nanostructures fabricated in this study, it seems, the high photocatalytic performance of the well dispersed rod-like nanostructures of ZnO-2 sample is due to the high exposure of their non-polar facets, which do not adsorb MB molecules as strongly as polar surfaces. A strong adsorption of cationic dye over polar surface of ZnO probably blocks the penetration of exciting UV radiation, responsible for generating electron-hole pairs at semiconductor surface. On the other hand, the electron mobility in flater ZnO surface such as nanopetals is expected to be higher than other surfaces due to their directional and uninterrupted conduction channel. The high electron mobility of semiconductor surface would increase the recombination rate of photo-generated electrons and holes, which is unfavorable for photcatalytic process.

Conclusion
In summary, ZnO nanostructures of different morphologies could be fabricated in large scale and controlled manner through solvent-free mechanochemical process by controlling the precursor to hydrolyzer molar ratio and suitable choice of surfactant. Apart from its cheapness, the process can produce nearstoichiometric ZnO nanostructures with enhanced photocatalytic performance for cationic dye degradation under UV illumination. The high concentration of basic sites at the surface of the nanostructures is the principal responsible for their high photocatalytic activity. However, due to polar nature of ZnO crystals, the orientation of crystal facets exposed to the reaction mixture also plays a significant role on their photocatalytic performance. The ZnO nanostructures fabricated by our soft mechanochemical synthesis process can mineralize 50% of MB molecules in aqueous solution as fast as in 17.5 min. The large scale production ability, low cost processing and high dye degradation ability of the fabricated nanostructures suggest the solvent-free mechanochemical process is extremely attractive for fabricating ZnO nanostructures for dye removal from industrial effluents.