Room Temperature syntheses of ZnO and its structure

ZnO has many technological applications which largely depend on its properties that can be tuned by controlled synthesis. Ideally, the most convenient ZnO synthesis is carried out at room temperature in aqueous solvent. However, the correct temperature values are often loosely defined. In the current paper we performed synthesis of ZnO in aqueous solvent, by varying reaction and drying temperature by 10°C steps and monitored the synthesis products primarily by XRD. We found out that a simple direct synthesis of ZnO, without additional surfactant, pumping of freezing, required both a reaction (TP) and a drying (TD) temperature of 40°C. Higher temperatures also afford ZnO, but lowering any of the TP or TD below the threshold value results either in the achievement of Zn(OH)2 or in a mixture of Zn(OH)2/ZnO. A more detailed Rietveld analysis of the ZnO samples reveals a density variation with the synthesis temperature and an increase of the nanoparticles average size also verified by SEM images. The optical properties of ZnO obtained by UV-Vis reflectance spectroscopy indicate a red shift of the band gap by ~0.1 eV.

The synthetic strategy and the careful control of synthesis parameters play a key role in determining the ZnO properties. In addition, environmentally friendly syntheses must be privileged, also on account of the extent of ZnO applications and consequent sizeability where TP is the temperature of the precipitation reaction and TD is the drying temperature. More in detail, a 0.1 M Zn(NO3)2 solution was first prepared in deionized water and stirred for a few minutes until completely dissolved. Then, an equal volume of a 0.1 M NaOH solution was added dropwise to the solution normalized in oil bath at the selected temperature and stirred 24h.
In first instance, in the selection of the synthesis parameters, the ratio between Zn(NO3)2 and NaOH was kept at 1:1, i.e. in excess of Zn 2+ . This implies an initial pH of 6, that corresponds to a minimum of ZnO solubility in water (at 25°C) [54], to ensure that formed ZnO does not redissolve. In a typical synthesis, the reaction is carried out at a given temperature (TP), the slurry digested for two hours, centrifuged (3500 rpm for 10 min), washed with distilled water multiple times and then dried at the chosen temperature (TD) in an oven for 72 h. A white powder was, then, achieved, which was further characterized. The summary of the syntheses conditions is reported in Table 1.

Sample
Reaction Temperature Drying Temperature  P30D30  30°C  30°C  P30D40  30°C  40°C  P40D40  40°C  40°C  P50D30  50°C  30°C  P50D50  50°C  50°C  P60D60 60°C 60°C X-ray diffraction (XRD) patterns of the synthesized samples were collected with an X'pert pro X-ray diffractometer by Philips, using CuK-Alpha radiation. They were, then, analyzed by a Rietveld procedure with the GSAS-II suite of programs [55]. Scanning Electron micrographs (SEMs) were collected with a Zeiss Auriga 405, Field Emission-Scanning Electron Microscope instrument, mounting a Gemini column and operating at 7 kV.
Optical measurements in the ultraviolet (UV), visible (Vis), and near infrared (NIR) spectral regions were obtained by a diffuse reflectance setup from Avantes BV (The Netherlands). The latter comprises a combined deuterium-halogen radiation source (AvaLight-DH-S-BAL) connected via an optical fiber to a 30 mm-diameter Spectralon® coated integrating sphere (AvaSphere-30-REFL used to illuminate the samples (sampling port diameter 6 mm) and collect the radiation diffusely reflected from all angles. . The ZnO powders were placed in the well of a sample holder with a depth such as to be able to consider the reflectance spectra as those of an infinitely thick sample ( ). The integrating sphere was connected through another optical fiber to a spectrometer (AvaSpec-2048x14-USB2). This configuration allows applications in the 248-1050 nm range with a 2.4 nm spectral resolution. A laptop is used for spectrometer control and data recording, whereas factory calibrated Spectralon® (Labsphere, USA), is used as reflectance reference. Each reflectance spectrum was obtained by averaging 5 acquisitions lasting 5 seconds each.

Results and Discussion
The solvo-synthesis of ZnO can be depicted as a two-stages process, i.e. precipitation and drying. The achievement of the ZnO, rather than Zn(OH)2 or a combination of the two moieties, strictly depends on the operational modalities of two stages.
In the present paper we opted for determining the best match in terms of temperature of each of the two steps, aiming at finding the lowest possible ones, which would guarantee no detectable presence of Zn(OH)2. The verification of the composition was primarily carried out by XRD. Additional information on the fine structure of ZnO were determined by Rietveld refinement. A first synthesis was carried at TP=TD=30°C, hence the closest, in our experimental setup, to the classical definition of RT=25°C. In this setup we obtained nearly 100% Zn(OH)2 samples ( Figure 1). We, then, increased the drying temperature by 10°C to achieve the sample P30D40 and, subsequently, also the reaction temperature by 10°C (sample P40D40). The former setup still affords a mixture of ZnO and Zn(OH)2, whereas the XRD of the latter indicates the solely presence of ZnO. We also verified that a temperature of 40°C was strictly necessary to achieve ZnO, by varying the reaction temperature to 50°C and using either 30°C or 40°C for the drying procedure. Once more, the former yields a mixture of ZnO and Zn(OH)2 whereas the latter is pure ZnO. Finally, we performed a synthesis at TP=TD=60°C obtaining pure ZnO. Therefore, in general, 40°C is the threshold temperaturenecessary both for the rection and for drying in order to achieve ZnO.
Since RT ZnO syntheses reported in literature are typically carried out in basic conditions, with Zn(NO3)2 and NaOH ratio 1:5 to 1:50 [51][52][53][54][55][56][57][58] and pH≥12, we verified whether basic conditions would allow lower temperature conditions and carried out a synthesis using a Zn(NO3)2:NaOH ratio of 1:10, at Tp=Td=30°C. Also in this case we achieved Zn(OH)2 (XRD not shown). The powder X-Ray diffraction spectra of all the samples P30D30, P30D40, P40D40, P50D30, P50D40 and P60D60, were first visually compared ( Figure 1) and then analyzed with Rietveld refinement using the GSAS-II software [55]. P40D40, P50D40 and P60D60 proved to be composed of pure ZnO, and the difference of the structural parameters was investigated, whereas P30D30, P30D40, and P50D30, on the contrary, turned out to contain both the target product (ZnO) and the precursor Zn(OH)2. For this reason, the relative composition of the two solid phases in the mixture was assessed besides the structural parameters. The preliminary qualitative comparison of the spectra pointed out that in P60D60, P50D40, P40D40-the first three patterns reported in Figure 1 -the peak positions are similar, the only notable difference being their intensity that is highest for P60D60 and comparable in the other two systems. A common crystal structure can therefore be envisaged for them. A more accurate analysis, though, (see Figure 2), shows that the most intense peaks from P60D60 fall almost in the middle between P40D40 peaks (lower angle) and P50D50 ones (higher angle); the latter two sets of peaks are shifted by 0.1 degrees, on average. This issue suggests that a slight change of some crystal parameter(s) occurs among the three structures. Regarding P30D40, P30D30, and P50D30, the patterns are characterized by a large number of new peaks that show up in all angle ranges, especially the very intense ones before 30 degrees, a portion of the spectrum where no peaks are found in P60D60, P50D40 and P40D40. The crystal structures of the mineral zincite (100% ZnO, space group P63mc) [56] and of Wulfingite (pure zinc hydroxide, P212121) [48] were then employed in the fitting procedure. Both single-crystal phase data were downloaded from the American Mineralogist Crystal Structure Database [ 57 ] as Crystallographic Information Files (CIF). The simulated powder pattern for each sample was obtained, for CuK wavelength (1.54 Å), modelling the peak shapes with a convolution of modified asymmetric pseudo-Voigt functions [58]. The background was modelled with a 3-term Chebyshev polynomial of the first kind. P60D60 profile was successfully fitted on 3100 observations, reaching the weighted-profile R factor wR = 11.76%, R=8.20%, χ 2 = 5420.14 at convergence. The same approach was adopted for the other systems, obtaining wR = 14.96%, R=10.83%, χ 2 = 6382. 69 Supplementary Information figures SI1 through SI6. The sample P30D30 will be from now on considered as pure hydroxide for simplicity. The fitting procedure also returns the optimized structural parameters for the cell(s) and the atom positions; additional fitted parameters were the domain size and microstrain of the crystallites, as well as the sample position and the instrumental broadening parameters U, V, W. The most important parameters are reported in the tables below (Table 2 for pure ZnO, Table  3 for the mixtures).  P40D40, the densest structure, having structural correlations that occur at smaller distance on average, gives origin to diffraction peaks that fall at larger angles (according to the scattering vector and Bragg's law definitions Q=4πsinθ/λ and d=2π/Q [59,60])while the distances between the atoms contained in the least dense sample (P50D40) are typically larger and the resulting peaks fall at smaller angle. Interestingly, the cell dimensions are proportional to the average (fitted) dimension of the crystallites, signaling that smaller nanoparticles (P40D40) occupy the smallest volume, i.e. they are more efficiently packed. When compared with the single-crystal cell dimensions (a=3.24950 c=5.2069), or with some recently reported ZnO nanosheets (a= 3.245 c=5.198, [ 61 ], the first two nanoparticle samples appear to be significantly less dense, while P40D40 values are much closer. The calculated density values confirm this trend and satisfactorily comply with the experimental density value of 5.61 g cm -3 [62]. The modulation of the structural features observed demonstrates that our synthesis protocol, by a simple tuning of the temperature., is capable of leading, easily and successfully, to final products with different properties. The synthesized samples display different types of morphology, depending on the temperature adopted in the different phases of the synthesis. Examples of the morphology of the samples are reported in Figure 3. The P30D30 sample, mostly composed of Zn(OH)2, has micrometric structure of aggregated foils. In the P30D40 sample the micrometric arrangements become more compact to form blocks and some nanoparticles appear on their surface highlighted in Figure 3b with light orange circles. Once pure ZnO is formed, i.e. when both reaction and drying temperatures are, at least, 40°C, the samples are characterized by the shear presence of nanoparticles. The nanoparticle morphology of ZnO is kept also for higher synthesis temperature, though the average size is slightly larger. In particular, the average size of the P40D40, P50D40 and P60D60 are 53±10 nm, 65±12 nm and 69±11 nm, respectively, thus in agreement with the crystallite size detected by XRD, within the experimental error. In addition, there is a good agreement between estimated crystallite size and the samples morphology, for mixed Zn(OH)2/ZnO samples with larger average size associated to Zn(OH)2 as compared to ZnO. XRD allows a separate assessment of the average size of each phase. Though this is not possible by SEM imagining, since the morphology cannot quite distinguish the crystallographic phase, it can be hypothesized that the small nanoparticles on the surface of the blocks in the P30D40 sample are actually ZnO deriving from the Zn(OH)2 dehydration.  The band gap of the ZnO nanoparticles samples was estimated by diffuse reflectance R∞ spectra, where is the reflectance of an infinitely thick layer of sample, converted into absorption spectra by using the Kubelka-Munk function [63,64]:

Preprints
where K and S are the Kubelka-Munk absorption and scattering coefficient, respectively... The K absorption coefficient is related to the intrinsic absorption coefficient of the particles α by α = K/2, in case of diffuse light distribution as in our case [65]. The steady preparation of ZnO powder sample for optical measurements also allowed us to consider S constant from sample to sample, then the ratio K/S ≈ . In order to estimate the band gap from the absorption spectra we can, then, apply the Tauc relation [66]: Where n=1/2 for indirect gap and n=2 for direct gaps, such as the case of ZnO, and C1 is a constant.  The ZnO band gap including the quantum confinement as a function of the average particle size is obtained from the effective mass model given by [67,68]:

Preprints
In this equation all terms are in eV, is the bulk ZnO band gap (3.35 eV), ℎ is the Planck constant, r is the particle radius, e is the charge of the electron, is the reduces mass of electron and holes = * * * * , where * and * are the effective mass of the electrons and holes in ZnO, respectively, * = 0. 19 and * = 0.8 , where is the free electron mass [69]; (3.7) is the relative permittivity, is the permittivity of free space.
The Eg values of 3.25(0), 3.25(8) and 3.26(2) eV (Table 4) injected in the previous equation all yields nanoparticles sizes (2r) of about 15 nm, providing an estimation of the smaller size in the synthesized ZnO nanoparticles. Compared to a band gap of 3.35 eV in bulk ZnO [70], P40D40, P50D40, P60D60 exhibit a red shift by ~0.1 eV. ZnO nanoparticles, in general, may undergo both a blue and a red shift, many factors concurring to the actual band gap value. As shown by Eq. 1, below 14 nm diameter threshold, the size effect is dominant and a blue shift is observed [71,72]. However, also the shape plays a role and both red and blue shift, depending on the particles morphology, as determined by the employment of different solvents during the synthesis [73]. In RT synthesis followed by lyophilization, a red shift by ~0.1 eV has been

Sample
Band Gap (eV) observed between synthesis at 20°C and at 50°C and it has been associated to a slight size increase of the nanoparticles [51]. However a simultaneous shape change was also observed from elliptical to conical which may also contribute to the red shift. P40D40, P50D40, P60D60 have similar morphologies and the comparable band gap values would indicate similar average sizes. Comparison with XRD results, which give domain sizes ranging from 46.9 to 84.8 nm, suggests a shape anisotropy for the P40D40, P50D40 and P60D60 nanoparticles, which can be also observed from SEM images (Figure 3 c and d). It must be added that the average particles size in band gap corresponds to the pathway for an electronic transition through a gap, which can be equivalent for anisotropic samples, with similar minimum size pathway.

Conclusions
Room temperature synthesis of ZnO nanoparticles in aqueous solution has been achieved. The synthesis has been carried out by controlling reaction and drying temperature, without the aid of any surfactant, pumping or freezing. We found out that the minimum temperature for achieving ZnO is 40°C for both synthetic phases. Lowering any of the two temperatures results in the achievement of Zn(OH)2 or a mixture Zn(OH)2/ZnO. Beyond the threshold value of 40°C, ZnO nanoparticles are still obtained, though with larger average size and different density. The morphology is similar for all ZnO nanoparticles as well as the optical band gap. The morphology of lower temperature samples indicates the presence of aggregated foils, which merge to form blocks, from which ZnO nanoparticles eventually detach. The results obtained from SEM experiments comply quite well with X-Ray diffraction experiments and optical analysis of the band gap, where it was found that the scattering patterns observed for the samples synthesized at temperatures larger than 40°C show a shape anisotropy and an evident shift towards smaller angles, thus indicating systems of larger average dimensions and crystallites size.