**Tuning the Morphology of ZnO Nanostructures with the Ultrasonic Spray Pyrolysis Process**

**Elif Emil 1,2, Gözde Alkan 3,\*, Sebahattin Gurmen 1, Rebeka Rudolf 4, Darja Jenko <sup>5</sup> and Bernd Friedrich <sup>3</sup>**


Received: 6 June 2018; Accepted: 19 July 2018; Published: 24 July 2018

**Abstract:** Nanostructured zinc oxide (ZnO) particles were synthesized by the one step Ultrasonic Spray Pyrolysis (USP) process from nitrate salt solution (Zn(NO3)2·6H2O). Various influential parameters, from Zn(NO3)2·6H2O concentrations (0.01875–0.0375 M) in the initial solution, carrier gas (N2) flow rates (0.5–0.75 L/min) to reaction temperature (400–800 ◦C), were tested to investigate their role on the final ZnO particles' morphology. For this purpose, Scanning Electron Microscopy (SEM), High Resolution Transmission Electron Microscopy (HRTEM) and (Selected Area Electron Diffraction) SAED techniques were used to gain insight into how the ZnO morphology is dependent on the USP process. It was revealed that, by certain parameter selection, different ZnO morphology could be achieved, from spherical to sphere-like structures assembled by interwoven nanoplate and nanoplate ZnO particles. Further, a more detailed crystallographic investigation was performed by XRD and Williamson-Hall (W-H) analysis on the ZnO with unique and non-typical planar morphology that was not reported before by USP synthesis. Moreover, for the first time, a flexible USP formation model was proposed, ending up in various ZnO morphologies rather than only ideal spheres, which is highly promising to target a wide application area.

**Keywords:** ZnO; ultrasonic spray pyrolysis; influential parameters; formation mechanism; structure; morphologies; characterization; TEM; HRTEM

#### **1. Introduction**

ZnO in a nanosized form is an indispensable candidate for electronic, optical, and gas sensors, as well as in catalysis applications, owing to its band gap value of 3.37 eV, large exciton binding energy of 60 meV and high electron mobility [1–5]. In the last decade, ZnO with various morphologies, including flower-like [6,7], nanodisc [1], nanobelt [8], and nanotube [9], targeting various application areas, have been investigated using different synthesis methods, e.g., sol-gel [10], hydrothermal [11], the microwave-assisted method [12] and spray pyrolysis [13]. Kajbafvala et al. reported synthesis of spherical and flower-like ZnO via the microwave-assisted method for organic dye photo-degradation via UV lamp irradiation. It was revealed that the degradation efficiency of the spherical particles is better than that of those with flower-like morphology due to their higher surface area, which provides the absorption of oxygen molecules and OH-ions, resulting in an increment of H2O2 and OH−

radicals' formation rate [14]. In another study, Sin et al. [15] reported synthesised self-assembled ZnO microsphere formed by nanoplanar crystal units to improve photocatalytic performance for various endocrine-disrupting chemicals under UV irradiation by the chemical solution route. The enhancement in the photocatalytic performance was attributed to the combined effects of the hierarchical surface structure and the large surface area, which suppresses the recombination of photo-generated electrons (e−) and holes (h+), and expedites the diffusion of electrons [15]. In another study by Li et al., carbon-doped and coated spherical ZnO particles were synthesized for use as anode material in high power Zn-Ni batteries. ZnO microspheres exhibit excellent cycling stability and superior high-rate performance [16].

These previous studies revealed that ZnO finds wide application areas, e.g., planar nano ZnO favors catalytic properties owing to its higher surface area, while granular nano ZnO is preferred in optical applications due to better absorption behavior, and battery applications due to structural stability and anti-corrosion capability [13]. Considering these morphology dependent utilizations of ZnO, adjusting a synthesis method to end up with varying morphologies in a controlled way would be advantageous. There have already been some studies reported, that control ZnO morphology by wet chemical methods is possible; however, these used reactant agents, such as polyethylene glycol (PEG) and cetyltrimethylammonium bromide (CTAB). In our previous study focusing on Ag/ZnO core shell nanostructured materials for photocatalytic applications, under some synthesis conditions, variation from the typical spherical morphology Ultrasonic Spray Pyrolysis (USP) into entangled plates was observed in ZnO particles without any additives [17]. These findings raised the question of whether it is possible to tune ZnO particles morphology by the USP process with simplicity, good process control, high flexibility, and good scale-up potential, without any additives, obtaining high-purity products [17]. For this purpose, varying influential USP process parameters like Zn(NO3)2·6H2O concentration in the initial solution, carrier gas (N2) flow rate and reaction temperature (400–800 ◦C) were examined. Based on this, the aim was to synthesize ZnO nanoparticles and to investigate the newly formatted ZnO morphology through the USP process. With detailed ZnO morphological investigations by High Resolution Transmission Electron Microscopy (HRTEM) and Scanning Electron Microscopy (SEM), the role of each USP parameter was determined. A more detailed crystallographic investigation was performed via XRD and W-H analysis only on the non-conventional nanostructured ZnO particles. Moreover, the reason behind the different morphologies of ZnO were explained, and a reaction progress model was proposed considering the USP process and thermodynamic conditions. There has been no study in the literature utilizing USP to synthesize ZnO nanoparticles with the aim of obtaining different morphologies.

#### **2. Experimental Procedures**

#### *2.1. Synthesis of Zinc Oxide Particles*

For the synthesizing of ZnO particles with USP, an aqueous solution of zinc nitrate hexahydrate (Zn(NO3)2·6H2O, purity > 99.9%) was used, purchased from MERCK Chemical GmbH (Darmstadt, Germany) Zinc nitrate hexahydrate, at the determined amount given in Table 1, was dissolved in distilled water and stirred for 30 min by a magnetic stirrer. This solution represents the so-called precursor, which was atomized by an ultrasonic nebulizer at 1.7 MHz. The formed aerosol was transferred into a one-step USP device to the pre-heated furnace (Nabertherm, R 50/250/12, Lilienthal, Germany) through a quartz tube (0.7 m length and 0.02 m diameter) using N2 gas. The experiments were conducted with 0.5 L/min and 0.75 L/min N2 flow rate at 400, 600 and 800 ◦C reaction temperature for 3 h. The thermal decomposition of Zn(NO3)2·6H2O under air atmosphere during the spray pyrolysis has been reported previously [17,18]. The chemical balance equations were given in 3 reaction steps:

$$\text{Zn(NO}\_3\text{)}\_2\cdot6\text{H}\_2\text{O} \rightarrow \text{Zn(NO}\_3\text{)}\_2\cdot\text{H}\_2\text{O} + 5\text{H}\_2\text{O}\uparrow\tag{1}$$

$$\text{R}3\text{Zn(NO}\_3\text{)}\_2\text{-H}\_2\text{O} \rightarrow \text{Zn(NO}\_3\text{)}\_2\cdot2\text{Zn(OH)}\_2 + \text{H}\_2\text{O} \uparrow + 4\text{NO}\_x\uparrow\tag{2}$$

$$\text{Zn(NO}\_3\text{)}\_2\text{-}2\text{Zn(OH)}\_2 \rightarrow 3\text{ZnO} + 5\text{H}\_2\text{O}\uparrow + 2\text{NO}\_3\uparrow\tag{3}$$

ZnO is the final product of all intermediate product decomposition of zinc nitrate. Synthesized ZnO particles were collected in a washing bottle filled with ethanol. Details of the synthesis procedure can be found elsewhere [19,20]. To characterize nanostructured ZnO particles, ethanol was evaporated in a drying oven at 70 ◦C for 300 s. A summary of the process parameters is illustrated in Table 1.

**Table 1.** Process parameters of nanostructured ZnO particles synthesized with various solution concentrations, reaction temperature and N2 gas flow rate.


#### *2.2. Characterization of Zinc Oxide Particles*

The morphology of the ZnO nanoparticles (size and shape) was examined by Field Emission Scanning Electron Microscopy (FE-SEM, JSM 700F, JEOL, Tokyo, Japan), operating at 5 kV. During SEM sample preparation, the SEM holder was grinded, the particles were dispersed in ethanol, and then the suspension was added dropwise onto the SEM holders, and, afterwards, a conductivity Pt coating was added to prevent charging of the particles by Sputter Coater (Polaron Range SC7620, Quorom Technologies, East Sussex, UK). A Transmission Electron Microscope (TEM, JEM-2100 HR, JEOL, Tokyo, Japan) with integrated Selected Area Electron Diffraction (SAED) pattern analysis, operating at 200 kV was used. Samples of ZnO nanoparticles in demineralized water for TEM analyses were drop cast onto a copper TEM grid covered with a carbon support film, dried, and then used for investigations.

The as-synthesized ZnO nanoparticles were analyzed using a X-ray diffractometer operating at 40 mA and 40 kV with Cu-Kα radiation (λ = 0.154051 nm). The diffraction spectra were recorded 2θ in the range between 10◦ and 90◦ in 2θ steps of 0.02 degrees. Phase composition was determined with a digital library of crystallographic cards JCPDS. Based on the Full Width at Half Maximum (FWHM), zinc oxide crystallite size was evaluated by Williamson–Hall (W-H) analysis and the Debye-Scherrer (DS) method.

#### **3. Results and Discussion**

#### *3.1. Effect of Precursor Concentration*

SEM micrographs of nanostructured ZnO particles are given in Figure 1, revealing the effect of precursor concentration on final morphology.

**Figure 1.** SEM micrographs of nanostructured ZnO particles (**a**) S1; (**b**) S2; and (**c**) S3, at 800 ◦C, 0.5 L/min of N2 gas flow rate.

As can be observed, a relatively lower zinc nitrate concentration (0.01875 mol/L) exhibited dominant plates and accompanying sphere-like morphology, assembled by interwoven nanoplates, which is not commonly achieved via USP. A gradual change in the microstructure was revealed by increases in zinc nitrate concentrations. An increase to 0.02875 mol/L resulted in a sphere-like morphology assembled by nanoplates. A further increase to (0.03750 mol/L) resulted in dominant granular accompanied particles. Although there are many studies dealing with ZnO synthesis with USP, such a unique morphology has not been reported and explained previously. In order to better understand its crystallography and to see if this morphology difference was due to a different phase formation, a detailed crystallographic investigation was performed only on that sample. The corresponding XRD pattern is given in Figure 2.

**Figure 2.** Indexed XRD pattern of sample S1.

As shown in Figure 2, reference sample S1 exhibited a phase pure zinc oxide hexagonal structure with a space group P63mc (186), unit cell of a = 3.2533 Å and c = 5.2073 Å corresponding to JCPDS Card 00-036-1451. The peaks at 2θ = 31.80◦, 34.45◦, 36.32◦, 47.56◦, 56.65◦, 62.89◦, 66.38◦, 67.98◦ and 69.10◦ are assigned to (100), (002), (101), (102), (110), (103), (200), (112) and (201) diffraction planes, respectively. No characteristic peaks of zinc nitrate salt were detected in the diffraction pattern, implying termination of decomposition reaction. Based on the most intense peaks (101) diffraction planes, the crystallite size was calculated using the DS method, where the contribution of crystallite size and lattice strain to total peak broadening were not considered. The *β* parameter was corrected due to the effect of instrumental broadening on total peak broadening in DS, and the crystallite size of S1 was calculated as 24 nm by DS. However, the contribution of lattice strain on total peak broadening should be noted, since lattice strain is induced by a rapid heating/cooling rate and short residence time of the USP process, as reported previously [21]. This relation is expressed in Equation (4):

$$
\beta\_{\text{hkl}} = \beta\_{\text{D}} + \beta \text{s} \tag{4}
$$

Williamson Hall (W-H) analysis was preferred to evaluate the crystallite size and lattice strain. Details of the W-H analysis are explained elsewhere [22]. The intense peaks corresponding to (100), (002), (101), (102), (110), (103) and (112) planes were selected to conduct the W-H analysis. The strain-induced broadening was connected with crystal distortion, and is defined in Equation (5):

$$
\kappa \approx \beta \text{s} / \tan \Theta \tag{5}
$$

DS relation, Equation (4), can be reformulated with Equation (5) which results in Equation (6):

$$(\beta\_{\rm hkl}\cos\Theta = (k\lambda/D) + (4\varepsilon\sin\Theta) \tag{6}$$

where *β*hkl is the total peak broadening, *k* is the shape factor, *λ* is the wavelength of Cu-Kα radiation (λ = 0.154051 nm), D is the crystallite size of the synthesized particles, *ε* is the lattice strain, *β* is FWHM of the peak, and θ are the Bragg angles. The individual contribution of crystallite size and lattice strain were evaluated by the W-H method integrated with Uniform Deformation Model (UDM). Graphics of (4sin θ) versus Bhklcosθ were drawn in Figure 3. Crystallite size and lattice strain respectively were estimated from the intercept on the y-axis and slope of linearly fitted data.

**Figure 3.** Williamson-Hall (W-H) analysis integrated with Uniform Deformation Model (UDM) model of S1.

Based on the broadening X-ray diffraction peak (FWHM), the crystallite sizes of the synthesized particles from 0.01875 mol/L were estimated at 38 nm and 24 nm using the W-H method integrated with UDM, and the DS method, respectively. This fine crystallite size and the XRD pattern given in Figure 2, which is consistent with the literature, imply the formation of secondary particles by coalescence of these primary crystals. Due to the neglection of the intrinsic strain at the lattice, the values of the crystallite sizes calculated using the W-H method differed from that evaluated by DS analysis, as expected. According to the plots of (4sin θ) vs. Bhklcos θ of particles, the intrinsic strain at the lattice

for S1 was found to be 9.462 × <sup>10</sup><sup>−</sup>4. It was proved that DS analysis calculates a smaller crystallite size than the size calculated using the W-H method in the presence of a tensile lattice strain. Moreover, these strain and crystallite size values were found to be similar to values reported previously for spherical ZnO particles [23,24]. Since the crystallographic findings do not differ dramatically from the previously reported data for spherical ZnO, the origin of this unique morphology was searched for in the USP formation mechanism.

In order to highlight microstructural features in detail, HRTEM was performed on the samples synthesized from varying precursor concentrations. Corresponding HRTEM images with electron diffraction patterns can be found in Figure 4.

**Figure 4.** High Resolution Transmission Electron Microscopy (HRTEM) micrographs and electron diffraction ring pattern of S1 (**a**); HR-TEM micrographs and electron diffraction spot pattern S1 (**b**); HRTEM micrographs and electron diffraction ring pattern of S2 (**c**) and S3 (**d**).

The TEM Bright-Field (BF) image, and the corresponding Selected Area Electron Diffraction (SAED) pattern of S1 solution concentration, showed fine particle sizes in the range of 10–70 nm. The planar morphology was also revealed, and even some hexagonal particles were observed. The HR-TEM image (Figure 2b), with a 2-D Fast Fourier Transform (FFT), shows a lattice of two ZnO nanoparticles synthesized from 0.01875 mol/L. The bigger particle is oriented in the (100) direction, and the attached neighboring particle on the left of this particle is oriented in the (002) direction. The measured distance from the HR-TEM image between the lattice planes is 2.8 Å in the (100) direction and 2.6 Å in the (002) direction, which is in very good agreement with the cell parameters of ZnO (cell parameters a = 3.2498 Å, c = 5.2066 Å). In parallel with the XRD findings, with an increase in concentration, a gradual change in morphology was observed in the HRTEM images. Planar fine crystals were transformed into spherical ZnO nanostructured spheres (500–700 nm) with increasing concentrations. The SAED of all ZnO nanoparticles showed characteristic diffraction of a ring pattern, with some brighter and more distinct spots in the rings, which indicated the presence of some larger crystallites, though the rings were still relatively continuous, which meant that the crystallites were small, in the nm range, and in a random orientation. The electron diffraction spots can be described by a hexagonal crystalline-structured zinc oxide with a space group P63mc (JCPDS card No. 00-00-036-1451) with indices as shown in the inset of Figure 4a and in accordance with the XRD spectrum in Figure 2. Although the microstructure changed from planar to spherical with increasing solution concentrations, they exhibited the same crystallite structure [23]. In previous

studies, it was reported that the decomposition of zinc nitrate hexahydrate into zinc oxide takes place via step-wise reaction with the formation of a Zn(NO3)2·2Zn(OH)2 intermediate compound. However, before the decomposition takes place, the initial salts melt and reaction takes place in the liquid phase [17,18]. Since particle formation in the melt phase is driven by nucleation and growth processes, the final morphology can be adjusted by controlling nucleation and growth rates via process parameters. The different morphologies achieved in this study at various temperatures, concentrations, and flow rates can be explained by this fact.

In the case of a higher precursor concentration, supersaturation induces higher nuclei rates, yielding more nuclei formation and less growth rate, and resulting in the formation of spherical-like ZnO particles. Similarly, when the temperature increased, the supersaturation degree increased again. This also increased the nucleation rate and, therefore, at 800 ◦C, almost all samples exhibited spherical-like morphologies. In the USP process, temperature change and flow rate also act parallel to concentration in terms of supersaturation degree. Increased temperatures and flow rate results in higher temperature gradients and increases the supersaturation rate. In order to demonstrate the change, the supersaturation degree plays a crucial role in the final morphology; the optimal concentration (0.02875 mol/L) was selected and the temperatures and flow rates were varied to observe their individual effects.

#### *3.2. Effect of Reaction Temperature and N2 Gas Flow Rate*

USP reaction temperature and gas flow rate together determine the residence time of droplets/particles in the heating zone, and therefore they also play a significant role in the nucleation step and growth mechanism. In previous synthesis conditions, reaction temperature was fixed at 800 ◦C. Temperature was changed to 600 ◦C/400 ◦C while ensuring complete decomposition of zinc nitrate. SEM micrographs of the formatted ZnO particles can be found in Figure 5.

**Figure 5.** SEM micrographs of nanostructured ZnO particles synthesized at various reaction temperatures at S5 (**a**), S6 (**b**) and S7 (**c**), where concentration (0.02875 mol/L) and flow rate (0.5 L/min) were constant.

To begin with, it is worth emphasizing that all samples synthesized at different temperatures exhibited complete conversion to ZnO. With increasing synthesis temperature, similar to the change in concentration, a gradual morphological change was observed from plates to spheres. When the reaction temperature was high, the supersaturation rate and mobility of particles were expected to be high. Higher supersaturation degree yielded higher nucleation rates, while a higher mobility increased the collision rates and growth rates of individual particles. For low temperatures, the driving force of the reaction in the system and nucleation rates are low. Therefore, there can be an accumulation on already existing nuclei and directed growth may be observed, as shown in Figure 5, in parallel with low precursor concentrations (see Figure 1a). However, when a particle synthesized at the highest temperature was considered, there was no observed coarsening, which implies that the increased nucleation rate was more dominant.

In order to assess flow rate effect in morphology, a high temperature (800 ◦C) was used to ensure complete conversion and ZnO was synthesized at a relatively lower flow rate. The SEM micrographs presented in Figure 6 reveal the effect of flow rate.

**Figure 6.** SEM micrographs of nanostructured ZnO particles synthesised from 0.02875 mol/L at 600 ◦C; (**a**) 0.75 L/min N2 gas flow and (**b**) 0.5 L/min N2 gas flow rate.

Figure 6 reveals that an obvious morphology change occured when the flow rate was varied. Although the particle size remained similar, when the flow rate decreased from 0.75 to 0.5 L/min, the granular morphology was replaced with plates and spheres assembled by these plates. In the case of higher flow rates, a higher gradual temperature was achieved within the droplet, which also increased the supersaturation and driving force of the reaction.

#### *3.3. Formation Mechanism*

In the USP process, the aerosol droplets undergo evaporation/drying, precipitation and thermolysis in a single-step process and under extreme synthesis conditions (high droplet/particle heating rate and high surface reaction), as presented in Figure 7. Within the short reaction time (2–3 s), intraparticle transport, solute nucleation and growth take place. The morphological findings presented in Figures 1, 5 and 6, revealing the effect of temperature, precursor concentration and flow rate, are consistent, implying that lower supersaturation degrees decrease the driving force of the reaction and, hence, nucleation rates. During the USP process, as summarized in Figure 7, in the case of lower nucleation rates, growth occurs from already existing nuclei in favored crystallographic directions, ending up with planar growth. In the case of a higher driving force, higher nucleation rates ensure homogenous nucleation, and precipitation takes place in the determined volume and shape of spherical droplets.

**Figure 7.** Schematic diagram of spherical and sphere-like structures assembled with interwoven nanoplate ZnO formation through Ultrasonic Spray Pyrolysis (USP).

Keeping in mind that morphological features of nanostructured ZnO particles are directly related to their functional properties, USP can be utilized as a suitable method targeting various application areas of ZnO. It has already been reported that higher surface areas of ZnO plates are favorable in photocatalysis applications [24]. For such applications, a synthesis strategy can be utilized dealing with low concentrations, temperatures and flow rates. On the other hand, the photocatalysis application necessitates the sintering of particles and UV-blockage properties, as spherical particles are more favorable owing to their good sinterability and good absorbance with less scattering of light [25,26]. For such applications, USP should be utilized at higher concentrations, temperatures and flow rates. A basic and empirical model can be found in Figure 7, which summarizes recipes to synthesize spherical or planar ZnO particles using the USP method.

#### **4. Conclusions**

The synthesis of pure and nanostructured ZnO particles was accomplished by USP. A lower concentration of Zn(NO3)2·6H2O in the initial solution, and the reaction temperature and flow rate of N2 resulted in lower Zn-saturation and, therefore, in lower nucleation rates. All these facts lead to the formation of nanostructured ZnO particles with a planar morphology, which is not typical for USP. Moreover, it was explained by the schematic representation of the formation mechanism, which showed that it is possible to control nanostructured ZnO particle morphologies, from spheres assembled by plates to plates via altering the USP parameters. This pioneering study aimed to explain the formation mechanisms of ZnO nanostructures and will contribute to targeting various applications in our next study.

**Author Contributions:** E.E., G.A., R.R., B.F. and designed the research, performed the experiments and analysed the data. D.J. provided the analysis tools and performed the characterisation. G.A., E.E. and R.R. wrote the article.

**Acknowledgments:** The authors would like to thank Gültekin GÖLLER and Technician Hüseyin SEZER for the SEM studies. The authors acknowledge the financial support from the Slovenian Research Agency (Research Core Funding No. P2-0120, P2-0132, BI-DE/17-19-12). The responsible proof reader for the English language is Shelagh Hedges, Faculty of Mechanical Engineering, University of Maribor, Slovenia.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### *Metals* **2018**, *8*, 569

#### **Abbreviations**


#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Characteristics of Ti6Al4V Powders Recycled from Turnings via the HDH Technique**

#### **Mertol Gökelma 1,\*, Dilara Celik 2, Onur Tazegul 2, Huseyin Cimenoglu <sup>2</sup> and Bernd Friedrich <sup>3</sup>**


Received: 28 March 2018; Accepted: 7 May 2018; Published: 9 May 2018

**Abstract:** The objective of this research is for Ti6Al4V alloy turnings, generated during the machining of implants, to produce powders for the fabrication of Ti base coating via the cold spray method. In order to decrease the cost of powder production and increase the recycling rate of the turnings, the hydrogenation-dehydrogenation (HDH) process has been utilised. The HDH process consists of the following sequence: surface conditioning of the turnings, hydrogenation, ball milling (for powder production), and dehydrogenation. Afterwards, the properties of the recycled powder were analysed via phase, chemical, and morphological examinations, and size and flowability measurements. Usability of the powder in additive manufacturing applications has been evaluated via examining the characteristics of the deposit produced from this powder by the cold spray method. In short, promising results were obtained regarding the potential of the recycled powders in additive manufacturing after making minor adjustments in the HDH process.

**Keywords:** Ti6Al4V; HDH; powder metallurgy; powder synthesis

#### **1. Introduction**

Ti6Al4V alloy (Ti, 6 wt. % Al, 4 wt. % V) is α + β phase titanium alloy, which occupies about 50% of the total titanium market [1,2]. It is very commonly used in aerospace, automotive, and medical industries, due to its high strength/weight ratio, corrosion resistance, biocompatibility, and low thermal expansion characteristics. Although Ti6Al4V alloy is a high-cost material it has a loss of about 70–80% as scrap while manufacturing of engineering components [3]. Owing to the high oxygen affinity, recovery of titanium-based scraps by re-melting is a difficult and costly process. Furthermore, re-melting may cause an imbalance in alloy composition. From this point of view recovery of titanium scraps in the form of powder by following the powder metallurgy route can be attractive especially for the recently-growing additive manufacturing market (such as 3D printing and cold-spray applications, etc.). Thus, powder metallurgy of Ti6Al4V alloy can be an alternative method either to use powder as the starting material or to recycle the new scrap which comes from semi-fabrication and manufacturing operations as the powder with improved purity [4,5].

The additive manufacturing industry has been growing very quickly in recent years and reached US\$5.1 billion in 2015 [6]. This increasing trend requires more input material and new production methods fulfil the requirements for Ti6Al4V powders. There are many additive manufacturing technologies in the market and the main methods are stereolithography, selective laser sintering, direct metal laser sintering, fused deposition modelling, and 3D printing [7]. 3D printing has been improved with many ongoing studies and it is expected to be used more commonly in the future [8].

It has been documented that, owing to the high oxidation tendency of atomization, a titanium melt can be made under vacuum or inert gas atmosphere at high production costs by utilising special technologies, such as electrode induction melting gas atomization, vacuum arc melting and cold hearth melting, etc. [9]. As an alternative method hydrogenation/dehydrogenation (HDH) can be an economical option for the recycling of titanium based scraps [10]. The HDH method benefits from room temperature embrittlement of the hydrate and produces powder-sized hydrate which must be dehydrated to produce the final powder product. The balance pressure and the temperature of the system are the main parameters for hydride transformation. The typical reaction can be written as *M* + *H*<sup>2</sup> ↔ *MH*<sup>2</sup> and can be reversed by the balance pressure of hydrogen gas. If the pressure is above the equilibrium pressure, hydrogen atoms enter the lattice to form metal hydrate; if it is below that level, hydrogen atoms diffuse out from the metal to from hydrogen gas [8,9]. During the hydration, some hydrogen atoms (H) enter the crystal lattice under the critical temperature and pressure. Then, while cooling under room temperature, hydrogen atoms are diffused into the crystal lattice and occupy the tetrahedral sites [11]. When titanium-based scraps are of concern, this mostly occurs in the β-Ti phase, which has more hydrogen solubility than the α-Ti phase [10–13]. Enrichment of the material by hydrogen (hydrogenation) imposes embrittlement leading to fracturing under mechanical loads (i.e., during ball milling), subsequent dehydrogenation provides some ductility as the result of outward diffusion of hydrogen. After dehydrogenation, powders must be kept under the protective atmosphere because of the oxidation affinity of the titanium. There are many studies that sintered the hydrogenated powder and performed the dehydrogenation step after the final shape was given by sintering [2,14–17].

The current work aims to present the preliminary results of Ti6Al4V powder production via the HDH method. Powder characteristics are analysed and its performance as a coating material has been evaluated. This study presents an innovation beyond the state of the art which is based on the usage of recycled Ti6Al4V scrap as a material for coating and/or additive manufacturing.

#### **2. Experimental Methodology**

In this study turnings produced during manufacturing of Ti6Al4V–ELI (0.20 wt. % of oxygen) implants were used as the scrap. Figure 1 presents flow chart for production of Ti6Al4V powder, which starts with an etching process to clean and activate the surface of the turnings. The turnings were then hydrogenated to form a brittle titanium hydrate for easy milling down to a size of less than 100 μm. Then powders of Ti6Al4V were dehydrogenated as a final step. The produced powder is characterized at the end to evaluate the usability of the product for different targets. Additionally, deposition performance of the recycled powder was analysed by low-pressure cold spray equipment.

Acid cleaning of the turning was made in H2SO4:H2O solution (having a ratio of 1:6) to activate their surfaces by removing any organic or inorganic impurities. After washing with deionised water and drying at 60 ◦C for about 45 min turnings were placed into the chamber of the reactor for hydrogenation. Hydrogenation was performed under 2.5 bar hydrogen atmosphere and the turnings were heated at a rate of 400 ◦C/h up to about 700 ◦C. After reaching a peak temperature of 700 ◦C the reactor was switched off and the turnings were allowed to cool to room temperature in the reactor. Diffusion of hydrogen into the turnings (activation) started around 470 ◦C according to Figure 2, representing the applied hydrogenation process as a "temperature vs. time" plot. The activation can be observed by the increasing temperature with a higher rate. After switching off the reactor temperature remained at about 600 ◦C for a certain interval as a result of the exothermic reaction imposed by the diffusion of hydrogen atoms. Then hydrated turnings were milled at room temperature in the steel ball milling equipment (MM 301, Retsch GmbH, Haan, Germany) with the frequency of 20 s−<sup>1</sup> for up to four minutes and sieved under 100 μm.

**Figure 1.** Experimental procedure of powder production via the HDH process.

**Figure 2.** Temperature measurement time during hydrogenation of the Ti6Al4V turnings.

The sieved powder was placed again in the same reactor for dehydrogenation under continuous vacuum. By the time, the system was cooled down under vacuum (1 mbar) to room temperature. Figure 3 presents the variation of temperature and pressure with respect to dehydrogenation duration. While the heating rate was gradually increasing to the peak temperature of about 700 ◦C pressure increased after 95 min as the result of the start of dehydrogenation. Pressure gradually increased for a duration of about 110 min, above which it sharply increased to over-pressure (1000 mbar is the detection limit of the unit) due to the high amount of hydrogen release. After switching off the reactor pressure sharply reduced while temperature gradually decreased. The dehydrogenised powder was taken out from the reactor in a glovebox with the oxygen concentration of <30 ppm. It was noticed that the powder was loosely sintered during the dehydrogenation process. Therefore, the dehydrated powder was milled once again in the steel-ball milling unit under an argon atmosphere. After a final sieving, the powder under 45 μm is sent for characterization.

**Figure 3.** Temperature measurement time during dehydrogenation of the hydrate powders.

The recycled powder was characterized by the following methods: elemental analysis by atomic absorption (for Al, V, Fe) solid state infrared absorption (for oxygen) and thermal conductivity (for hydrogen), phase analysis by XRD (X-ray Diffraction, BrukerTM D8 Advance Series 35 kV 40 mA, Billerica, MA, USA), flowability measurement by AOR (angle of repose), morphology by SEM (scanning electron microscope, JEOLTM JCM-6000Plus NeoScope, Peabody, MA, USA) and particle size distribution by laser diffraction (Mastersizer 2000, Malvern Panalytical, Worcestershire, UK). Additionally, deposition efficiency of the powder on commercial purity titanium substrates by low pressure (600 kPa) cold gas dynamic spray equipment (Rusonic Model K201, Rus Sonic Technology, Arcadia, CA, USA) having a converging-diverging tubular nozzle was evaluated. Characteristics of the deposits were determined by SEM examinations and XRD analysis.

#### **3. Results**

Figures 4 and 5 present the size distribution and morphology of the powder recycled from Ti6Al4V-ELI turnings, respectively. Fifty percent of the powder is under 13 μm and 90% is under 37 μm. The morphology of the particles is irregular, as expected, with the circularity of 0.72 ± 0.04 and the form factor of 1.42 ± 0.15 (max-axis/min-axis) [18]. The flowability of the powder was measured as 38.93 ± 1.55, which is accepted as "fair-aid not needed" or "some cohesiveness" [19].

The elemental composition of the powder is given in Table 1 along with the standard composition of Grade 5 and Grade 23 Ti6Al4V alloys. In general, the concentrations of the alloying elements are in an acceptable range according to the standards of ASTM B988-13 Grade 5, except the hydrogen content. Additionally, the average oxygen content of the dehydrogenated powder is slightly higher than that prescribed by ASTM B988-13 Grade 5. Moreover, oxygen and hydrogen contents of dehydrogenated powder are not in the range of Grade 23, which is called extra low interstitial (ELI).

**Figure 4.** Size distribution of Ti6Al4V particles synthesized by the HDH method.

**Figure 5.** Electron microscopy picture of powder synthesized by the HDH method.

**Table 1.** Elemental analysis of dehydrogenated powder and relevant standards.


XRD patterns of the as-received turnings, and powders (in hydrogenised and dehydrogenised states) are shown in Figure 6. The turnings consisted of α and β-Ti phases as expected. On the XRD patterns of the powders peaks of titanium hydrides (in the form of TiH2 and TiH1.5) appeared. As compared to the hydrogenised state, dehydrogenation caused domination of α and β-Ti phase peaks along with peaks of titanium hydrides by considering the intensity ratios of the Ti peaks with

those of TiH1.5 and TiH2. However, the applied dehydrogenation process did not completely remove titanium hydrides from the powders.

**Figure 6.** X-ray diffraction analysis of the raw material (**a**), after hydrogenation (**b**), and after dehydrogenation (**c**).

In order to evaluate the deposition characteristics of the recycled powder, some attempts have been made by the cold spray method which is based on the acceleration of particles in a process gas over the supersonic velocity through the convergent-divergent type nozzle [24,25]. In this system, particles of the feedstock powder are deposited as they impact on the surface of the substrate [24]. The bonding of the cold sprayed powder is a result of the mechanical interlocking between cold sprayed particles and the substrate. Increasing of the coating thickness is provided by the deposition of the following particle on the previous ones. In this process air, argon, helium, or nitrogen can be used as the process gas. In terms of acceleration capability and applicability, helium is the most favourable for high deposition efficiency. On the other hand, air appears as the best option when the cost of the coating is of concern. For those reasons, helium and air have been chosen as the process gases in this study to evaluate the deposition characteristics of the recycled powder with the parameters listed in Table 2. It should be emphasized that recycled powder was not successfully deposited with the utilized low pressure cold spray equipment over commercial purity titanium (Cp-Ti) with air as the process gas, unlike He. This observation can be associated with approximately three times higher acceleration capability of helium as compared to that of air due to its low molecular weight [24]. In order to overcome this problem 3 wt. % Al has been added into the feedstock as a binder according to our experience when using air as the process gas [26]. Additionally, the traverse speed has been reduced five times to further assist the binding of the feedstock in air.


**Table 2.** Low-pressure coating parameters of cold spray process.

A cross-section of the coatings deposited on the Cp-Ti substrate by cold spraying of the powder produced is shown in Figure 7. When He was used as the process gas powders were deposited on the substrate without remarkable discontinuities at the coating/substrate interface. However, a high amount of porosity was detected within the coating showing lower integration among powder particles. When air was used as the process gas successful deposition was not obtained unless addition of aluminium powder (at concentration of 3 wt. %) into the feedstock was done. Thus, Al acted as a binder for successful deposition of the powder on Cp-Ti (without noticeable porosities in the coating and discontinuities at the coating/substrate interface) when air was used as the process gas. In this respect, the dark coloured regions in the coating shown in Figure 7b were identified as Al particles by EDX (Energy-dispersive X-ray spectroscopy) analysis conducted during SEM surveys.

**Figure 7.** SEM micrographs of cold spray deposits produced by utilization of (**a**) He and (**b**) air as process gases (deposition parameters are listed in Table 3).

XRD patterns of the deposits shown in Figure 7 are depicted in Figure 8. Peaks of TiH1.5 and Ti were detected on the XRD pattern of the deposit formed by using He as the process gas (Figure 8a). Since Al powder has been added into the feedstock, additional Al peaks appeared on the XRD pattern of the deposit produced by utilization of air as the process gas (Figure 8b). It should be noted that the peaks of TiH1.5 which were detected in the XRD pattern of the recycled powder (Figure 6c) also remained after the cold spray process.

**Figure 8.** X-ray diffraction patterns of the coatings deposited by using (**a**) He and (**b**) air as the process gas.

#### **4. Discussion**

Ti6Al4V powder recycled from turnings via the HDH method have a mostly irregular morphology, which is also shown in this study. In spray-based deposition processes, like the cold spray method, irregular powders favour higher particle velocities than spherical ones because of their higher drag force [27–30]. In this respect, irregular powder particle morphology has some advantage on the final product of the cold-spray method [31]. In contrast, spherical particle morphology is more desirable than irregular ones for 3D printing, in terms of flowability requirements of the process [32–34]. Flowability is mostly related to the particle shape of the powder. Irregular shapes generally have lower flowability as compared to spherical ones. Although, the smaller-sized particles would have an acceptable flowability range, decreasing the particle size under the critical level could affect the productivity of the dehydrogenation process [34]. Apart from there being no special requirement of flowability for the cold spray process, the mean particle size, and the oxygen and nitrogen contents of the powder are more important factors for good coating properties and deposition efficiency [27]. When the average particle size is of concern, the HDH method provided an average particle size (d0.5) of 13.7 μm, which is within the acceptable range for almost every coating method and additive manufacturing processes [18]. The interstitials, such as oxygen and nitrogen in the powder, deteriorate the deposition efficiency by reducing the plastic deformation capacity of the Ti powders [35].

Even after dehydrogenation TiH1.5 still remains in the Ti6Al4V powder. This result matches with the research of Bhosle et al. [34], where they reported that the dehydrogenation occurs in a two-step process: TiH2 → TiHx → α − Ti. The finer hydride particles contain lower hydrogen in the second hydride phase (TiHx) and they are thermally more stable at higher temperatures than the TiH2 phase, which may result in the remaining hydrates even after the dehydrogenation process. This is also caused by a poor vacuum setup. In Table 3 the parameters of vacuum, temperature, and duration are listed for successful dehydrogenation of titanium powder from previous study, suggesting that higher vacuum or longer duration would provide complete dehydrogenation of the powder obtained by recycling. Moreover, it should be noted that the XRD peaks shifted to lower angles as compared to

that of the turnings due to the heavy deformation imposed during the ball milling process applied after hydrogenation. Additionally, the oxygen content of the recycled powder should be minimized according to the relevant standards [20–23]. The oxygen content is slightly higher than the desired range for the standard B265-15/B348-13/B381-13, but it is an acceptable range for the standard B988-13. After the HDH process, Ti6Al4V alloy powder should be within the chemistry limits of international standards and correspond to the industrial specification, which include medical, aerospace, and defence applications [12]. One of the important parameters to be considered to reduce the oxygen content is the vacuum level applied during the hydrogenation stage. Oxidation of very fine hydrate particles during milling can also be a source of oxygen due to very high surface area [36]. Therefore, it is suggested that adjusting the vacuum level and processing time of the dehydrogenation stage could also provide a solution for reducing the oxygen level of the recycled powder, which the process needs during dehydrogenation.


Deposition characteristic of the recycled powder have been analysed by using the low-pressure cold spray method. Usage of helium as a process gas generated a deposit with relatively high porosity content. It is possible to improve the quality by using high-pressure cold spray systems generating higher velocities as compared to low-pressure cold spray systems. Moreover, reducing the hydrogen and oxygen content of the recycled powder, which imposes brittleness, would lead to the generation of denser deposits even by utilizing air as the process gas (without the addition of Al powder). Irregular morphology of the recycled powder also affected the deposition capacity. Although these powders can reach higher velocity than spherical ones [29,31], it is not possible to have a dense coating because of the irregularity. In summary, the coating trials gave promising results about the usability of the Ti6Al4V powder recycled from scrap via the HDH process. However, some adjustments are necessary to increase the success of this powder in additive manufacturing processes.

#### **5. Conclusions**

The results of the current work can be summed up in the following points:


• Low pressure cold spray deposition studies revealed the potential of the powders obtained by the HDH process for coating. In order to evaluate the usability of this powder produced from turnings scrap, the characteristics of the deposits from this powder are being investigated and the results will be presented in a follow-up publication.

**Author Contributions:** M.G. and D.C. planned the content of the paper and conducted the experimental work. O.T. and D.C. performed the characterisation of samples and coating. B.F. and H.C. were supervisors of the research work and participated in the writing and assessment of results.

**Funding:** This research received no external funding.

**Acknowledgments:** Dilara Celik expresses her deep thanks to European Commission for the provision of student researcher mobility at the University of RWTH Aachen under the Erasmus program.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Metals* Editorial Office E-mail: metals@mdpi.com www.mdpi.com/journal/metals

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18