ZnO Nano-Particles Production Intensification by Means of a Spinning Disk Reactor

Zinc Oxide is widely used in many industrial sectors, ranging from photocatalysis, rubber, ceramic, medicine, and pigment, to food and cream additive. The global market is estimated to be USD 3600M yearly, with a global production of 10 Mt. In novel applications, size and shape may sensibly increase the efficiency and a new nano-ZnO market is taking the lead (USD 2000M yearly with a capacity of 1 Mt and an expected Compound Annual Growth Rate of 20%/year). The aim of this work was to investigate the possibility of producing zinc oxide nanoparticles by means of a spinning disk reactor (SDR). A lab-scale spinning disk reactor, previously used to produce other nanomaterials such as hydroxyapatite or titania, has been investigated with the aim of producing needle-shaped zinc oxide nanoparticles. At nanoscale and with this shape, the zinc oxide particles exhibit their greatest photoactivity and active area, both increasing the efficiency of photocatalysis and ultraviolet (UV) absorbance. Working at different operating conditions, such as at different disk rotational velocity, inlet distance from the disk center, initial concentration of Zn precursor and base solution, and inlet reagent solution flowrate, in certain conditions, a unimodal size distribution and an average dimension of approximately 56 nm was obtained. The spinning disk reactor permits a continuous production of nanoparticles with a capacity of 57 kg/d, adopting an initial Zn-precursor concentration of 0.5 M and a total inlet flowrate of 1 L/min. Product size appears to be controllable, and a lower average dimension (47 nm), adopting an initial Zn-precursor concentration of 0.02 M and a total inlet flow-rate of 0.1 L/min, can be obtained, scarifying productivity (0.23 kg/d). Ultimately, the spinning disk reactor qualifies as a process-intensified equipment for targeted zinc oxide nanoparticle production in shape in size.


Materials and Methods
All of the reagents were purchased from Carlo Erba (Milan, Italy) and were of analytical grade, whereas all of the solutions were prepared with deionized water. The following reagents have been adopted in the experiments: Zinc Sulphate Heptahydrate (ZnSO 4 ·7H 2 O, purity > 99.95%, M = 287.49 g/mol) was used as a Zn(II) precursor, whereas KOH (purity > 99.00%) was used as the base, for the formation and precipitation of Zn(OH) 2 , according to the Equation (1): The molar ratio used in the production experiments was fixed as Zn(II)/OH = 0.25 mol/mol, according to the optimal results obtained in preliminary experiments (data not showed).
The SDR used in this work is schematically reported in Figure 1, to allow insight into the main part of this equipment, which would not be possible by displaying a photo (since it is a boxed equipment).
The equipment consists of an inner disk made of Teflon, with a 4.25 cm radius inside an external cylinder (stator part). The feed streams are injected onto the disk surface, which rotates by means of an electric motor. The distance from the injection point to the disk center can be varied, hereafter called r i [cm], from 3.0 cm down to 1.5 cm at 0.5 cm steps, whereas the rotational velocity ω [rpm] can be varied from 0 rpm up to 1400 rpm. The reagent solutions have been injected, maintaining constancy at 25 • C-the temperature of the water batch. The internal diameter of the feed stream tubes was 3 mm.
A total of 25 experimental runs were performed. The production experiments were performed varying the following parameters as reported in brackets: Zn(II) initial concentration (0.02-2.00 M, and, as a consequence, KOH molar concentration varied in the range 0.08-8.00 M), reagent solution inlet flowrates (Q Zn and Q KOH [L/min], 0.025-0.500 L/min), ω (400-1400 rpm) and r i (1.5-3.0 cm). Referring to the data reported in Table 1, the pH of the produced solution was in the range 12. 51-12.62 in the first 16 runs. Then, at higher Zn(II), i.e., KOH molarity, the pH increased: in runs 17-19 it The solid residue was washed three times with ethanol and was placed in the oven for 24 h at 105 °C. The obtained powder was then characterized and the size distribution and average diameter, d [nm], were measured by the Dynamic Light Scattering method, using a Zetasizer Nano ZS (Malvern Panalytical, Westborough, MA, USA), and the pH was measured using a Crison pH-meter (Barcellona, Spain). The pH value during measurements was almost the same for all of the samples, and equal to 6.0. The yield of ZnO was measured as the ratio between ZnO produced mass and Zn(II) initial mass. The most convincing productions were characterized using a SEM-EDS (Zeiss, Oberkochen, Germany).   Initially, the Zn(II) and KOH molar concentrations were set equal to 0.02 and 0.08 M, and the maximum rotational velocity (1400 rpm) value was fixed, according to the optimal results obtained on metallic iron nanoparticles in a previous work [34], varying both the reagent solution flowrate (0.025, 0.050, 0.075 and 0.100 L/min) and r i (1.5, 2.0, 2.5 and 3.0 cm). Once the optimal reagent flowrate and r i values were found, all of the operating parameters were fixed, varying ω (200, 400, 800, 1200 rpm). Therefore, once the ω optimal value was found, the Zn(II) and KOH molar concentrations were The productivity P [kg/d] was calculated as the multiplication of the total flowrate (converted in l/d), initial Zn concentration (converted in kg/L), and yield of ZnO.

Results
In this section, the experimental results of the 25 runs will be reported. A summary table (Table 1) of the investigated operating conditions is reported. Figure 2 repots the average diameter of nZnO as a function of different Q Zn (i.e., Q KOH ) and r i values. Table 2 reports the detail of the results obtained in the first 12 runs.   Table 2 reports the detail of the results obtained in the first 12 runs. Table 2. Results obtained in the first 12 runs.     Table 3 reports the detail of the results obtained in the remaining 13 runs.    Figure 4 displays the SEM and EDX results obtained of the nZnO produced in runs: 23 (a, Q Zn = 0.05 L/min, ω = 1400 rpm, r i = 2.5 cm and Zn(II) = 0.02 M) and 8 (c, Q Zn = 0.5 L/min, ω = 1400 rpm, r i = 2.5 cm and Zn(II) = 0.5 M). The first one was selected as it exhibited the highest productivity still characterized by a unimodal particle size distribution; the second because it performed at the same operating conditions, but at 1/10 of the reagent concentrations compared with run 23.  Figure 4 displays the SEM and EDX results obtained of the nZnO produced in runs: 23 (a, QZn = 0.05 L/min, ω = 1400 rpm, ri = 2.5 cm and Zn(II) = 0.02 M) and 8 (c, QZn = 0.5 L/min, ω = 1400 rpm, ri = 2.5 cm and Zn(II) = 0.5 M). The first one was selected as it exhibited the highest productivity still characterized by a unimodal particle size distribution; the second because it performed at the same operating conditions, but at 1/10 of the reagent concentrations compared with run 23.

Discussion
The influence of r i on the nZnO particles and total flowrate on the rotating disk was clearly visible in Figure 2. Indeed, the optimal values were 2.5 cm for r i and 0.05 L/min for Q Zn , i.e., 0.1 L/min as the total flowrate. These two parameters strongly influenced the fluid dynamic conditions established in the rotating liquid film, which was generated onto the surface of the disk [46]. The residence time τ [s] on the disk can be calculated as [37]: where Q is the total inlet flowrate [m 3 /s], ν is the kinematic viscosity [m 2 /s], assumed equal to that of water being the diluted solution, and r d [m] is the disk radius. Figure 5 summarizes the influence of r i on residence time. The increase of r i and Q caused a decrease in the reagents residence time on the rotating disk surface; therefore, τ being too low, or Q too large, may hinder completion of the reaction. Usually, these kind of precipitation reactions performed by the SDR are very rapid [38], with induction times lower than 1-1.5 ms. Therefore, the influence of the analyzed parameters on the reaction yield was quite limited within the investigated range. Indeed, at fixed r i , an increase in Q caused a yield reduction of approximately 1-3% (see Table 2), whereas when maintaining Q as constant, an increase in r i led to a yield reduction of 1-2%. These parameters also influenced the average dimension of the obtained nanoparticles, mainly due to the different specific power values dispersed in the rotating liquid film ε [W/kg] and mixing time τ mix [s], as already observed in previous works [10,24]: where v r [m/s] is the average radial velocity calculated according to a simplified centrifugal model [35]. Figure 6 displays the influence of ε (a) and of τ mix (b) on the nZnO particles mean diameter.
visible in Figure 2. Indeed, the optimal values were 2.5 cm for ri and 0.05 L/min for QZn, i.e., 0.1 L/min as the total flowrate. These two parameters strongly influenced the fluid dynamic conditions established in the rotating liquid film, which was generated onto the surface of the disk [46]. The residence time τ [s] on the disk can be calculated as [37]: where Q is the total inlet flowrate [m 3 /s], ν is the kinematic viscosity [m 2 /s], assumed equal to that of water being the diluted solution, and rd [m] is the disk radius. Figure 5 summarizes the influence of ri on residence time. The increase of ri and Q caused a decrease in the reagents residence time on the rotating disk surface; therefore, τ being too low, or Q too large, may hinder completion of the reaction. Usually, these kind of precipitation reactions performed by the SDR are very rapid [38], with induction times lower than 1-1.5 ms. Therefore, the influence of the analyzed parameters on the reaction yield was quite limited within the investigated range. Indeed, at fixed ri, an increase in Q caused a yield reduction of approximately 1-3% (see Table 2), whereas when maintaining Q as constant, an increase in ri led to a yield reduction of 1-2%. These parameters also influenced the average dimension of the obtained nanoparticles, mainly due to the different specific power values dispersed in the rotating liquid film ε [W/kg] and mixing time τmix [s], as already observed in previous works [10,24]: where vr [m/s] is the average radial velocity calculated according to a simplified centrifugal model [35]. Figure 6 displays the influence of ε (a) and of τmix (b) on the nZnO particles mean diameter. The mean particle's diameter trend with respect to ε and τ mix was, of course, analogous, as the latter parameter depends on the former, and showed that the same minimum occurred for r i = 2.5 cm, at all Q values. The order of magnitude of ε and τ mix were in line with those obtained in a previous study, working at similar operative conditions (Q = 0.1-0.2 L/min) [34]. It has already been demonstrated that when the mixing time of the SDR is in the order of 0.1-1 ms, the obtained particles can reach dimensions lower than 100 nm, as obtained in the present study [35].
The subsequent runs were performed by fixing the Q to 0.1 L/min and the r i to 2.5 cm, according to the lower mean diameter obtained. The influence of the Zn(II) precursor molar concentration on d was quite limited, compared with that of ω (see Figure 3). This can be explained by considering that when the SDR is used, the average size of the produced nanoparticles is mainly influenced by ε and of τ mix (i.e., by ω and fluid dynamic conditions), whereas the initial reagents concentration can influence d only to some extent, as reported in previous works, where classical synthesis pathway and equipment were also used.
The results obtained by the SDR are well comparable with those reported in literature by Liu and Zeng [47], Gao et al. [48] and Wirunmongkol et al. [49]. In this work, the mean hydrodynamic diameter of the obtained particles was always in the range of 50-80 nm, with a lower OH − /Zn(II) molar ratio in comparison with those reported in the aforementioned studies, where the most commonly used ratio value was 10/1. In detail, in the first study, ZnO nanorods of an effective diameter of approximately 50 nm, were obtained by means of hydrothermal synthesis at 180 • C for 20 h, using ethylenediamine (EDA) at an EDA/Zn(II) molar ratio of 50/1, in addition to the adoption of a Zn(II)/OH − molar ratio of 20/1. Therefore, the synthesis required a higher energy demand and a larger quantity of reagents to obtain a homogeneous rod-like structure for the product, if compared with the SDR equipment. Regarding the second mentioned study [45], the authors adopted a OH − /Zn(II) molar ratio equal to 10/1 and produced ZnO nanorods with an average diameter of approximately 100 nm, assembled into sphere-like structures, by means of hydrothermal synthesis at 95 • C for 5 h, using hexamethylenetetramine (HMT) at a HMT/Zn(II) molar ratio equal to 1/1. Finally, the last study [46] reported ZnO nano-rods assembled in flower-structures, similar to those reported in the present study (Figure 4a,b), with a mean diameter in the range of 30-80 nm, using hydrothermal synthesis (60 • C for 6 h) and adopting a OH − /Zn(II) molar ratio equal to 10/1. The main influence of the OH − concentration was on the ZnO nanoparticles morphology variation: at a lower OH − /Zn(II) molar ratio (2/1), they were of spherical shape, and at a higher molar ratio (10/1-20/1), they became rod/wire-shaped, when the classical hydrothermal synthesis was adopted. Another important consideration reported in the aforementioned studies was the influence of the initial Zn(II) concentration: an increase in the precursor molar concentration usually led to an increase in the average diameter and a change in the morphology, causing an increase in the mean diameter and the occurrence of rod/wire-shaped particles. Analogous considerations were made in another study, where a continuous hydrothermal synthesis by supercritical water was reported [50]. Analogous results have been reported in the present study, where the flower-shaped particles obtained at a Zn(II) concentration of 0.5 M (Figure 4b), they changed their morphology to wire/rod-shaped when the Zn(II) concentration decreased down to 0.02 M (Figure 4a). The initial reagents concentration also influenced the shape of the Particle Size Distribution (PSD): indeed, bimodal PSD occurred when the Zn(II) molar rate was larger than 0.5 M-both at low and high inlet flowrates. This can be explained by considering that at higher reagent concentration and at fixed or lower liquid film volume values (as at higher inlet flowrate), the volume concentration of the ions quickly increased, leading to more possibilities for nucleation reactions, but also crystal growth and aggregation phenomena [51][52][53]. The mean particle's diameter trend with respect to ε and τmix was, of course, analogous, as the latter parameter depends on the former, and showed that the same minimum occurred for ri = 2.5 cm, at all Q values. The order of magnitude of ε and τmix were in line with those obtained in a previous study, working at similar operative conditions (Q = 0.1-0.2 L/min) [34]. It has already been demonstrated that when the mixing time of the SDR is in the order of 0.1-1 ms, the obtained particles can reach dimensions lower than 100 nm, as obtained in the present study [35]. Overall, the production of the chemical precipitation of nano ZnO by means of a spinning disk reactor, does not appear to require particular precautions to work continuously for a long period of time. Scaling in the reactor was absent after the experimental campaign, and all of the particles were suspended in the liquid, which did not evaporate at the adopted temperature values.

Conclusions
In conclusion, the SDR demonstrated itself to be a suitable equipment for the intensification of nano-ZnO particles production, as it was able to obtain good performances in terms of average size (approximately 50 nm), high yield (>97%), and unimodal particle size distribution. The same could be achieved by adopting a larger inlet flowrate and initial Zn(II) precursor concentration, in order to increase the production rate. Indeed, the last four runs of the reported experimental set showed that it was possible to keep the modal particle size below 60 nm using a Zn(II) molar concentration <1 M. In these conditions, productivity values were higher than 50 kg/d, making the overall production process useful for industrial application. Higher inlet flowrates were not suggested, as these operating conditions led to a sensible increase in the average dimension of particles and the size distribution, changing from a unimodal to a bimodal one. However, further studies should be performed to investigate the possibility of achieving larger production rates (up to 1000 kg/d), without exiting the range of the desired product's characteristics.