**E**ff**ect of Gradient Energy Density on the Microstructure and Mechanical Properties of Ti6Al4V Fabricated by Selective Electron Beam Additive Manufacture**

#### **Ta-I Hsu 1, Yu-Ting Jhong <sup>2</sup> and Meng-Hsiu Tsai 3,\***


Received: 25 February 2020; Accepted: 23 March 2020; Published: 26 March 2020

**Abstract:** Selective Electron Beam Additive Manufacturing (SEBAM) is a promising powder bed fusion additive manufacturing technique for titanium alloys that select particular area melting in different energy density for producing complexly shaped biomedical devices. For most commercial Ti6Al4V porous medical devices, the gradient energy density is usually applied to manufacture in one component during the SEBAM process which selects different energy density built on particular zones. This paper presents gradient energy density base characterization study on an SEBAM built rectangular specimen with a size of 3 mm × 20 mm × 60 mm. The specimen was divided into three zones were built in gradient energy density from 16 to 26.5 J/mm3. The microstructure and mechanical properties were investigated by means of scanning electron microscopy, X-ray diffraction, transmission electron microscopy and mechanical test. The α martensitic and lack of fusion were observed in the low energy density (LED) built zone. However, no α phase and no irregular pores were observed both in overlap energy density (OED) and high energy density (HED) built zones located at the middle and bottom of the specimen respectively. This implies the top location and lower energy density have positive effects on the cooling rate but negative effects on densification. The subsequence mechanical properties result also supports this point. Moreover, the intermetallic Ti3Al found in the bottom may be due to the heat transfer from the following melting layer. Furthermore, the microstructure evolution in gradient energy built zones is discussed based on the findings of the microstructure and thermal history correlation analysis.

**Keywords:** selective electron beam additive manufacture; Ti6Al4V ELI alloy; phase transformation; spatial; gradient energy density; martensitic decomposition; Ti3Al intermetallic compound; fracture analysis

#### **1. Introduction**

Selective electron beam additive manufacture (SEBAM) is one of powder bed fusion technology that builds metallic parts with complex geometries by using the electron beam as the heat source to selectively preheat and melt over the metallic powder bed via layer by layer based on 3D CAD model input. SEBAM has attracted much attention over the past few years for its advantages, such as high material utilization, low porosity and fast production rate without post heat treatment in particular

cases, and so on. Furthermore, SEBAM is an advanced manufacturing technology that has widely fabricated titanium, nickel base, CoCrMo and difficult forming alloys [1,2].

Ti6Al4V is commonly used in the SEBAM process. Its excellent corrosion resistance, mechanical property and biocompatibility are widely applied to medical applications such as orthopedic [3], dental [4] and spinal implants [5]. Porous medical metallic parts by SEBAM process, which possess low elastic Young's modulus matching to bone tissue and are capable of providing space for in growth of bony tissue to achieve better fixation, have been widely used for medical implants due to reduced stress shielding effect [6] and increased bone osseointegration [7,8]. Therefore, most researchers adjusted the porosity and geometric to improve the benefit biocompatible reticulated meshes and foams with an interconnected porous structure which has porosity open cell and foams (porosity between 55%~90%) [9–11]. As mentioned above, previous studies have fabricated porous complex structure by SEBAM method in low energy density parameters. The reasons for the low energy built could be the post treatment of the porous structure easily, i.e., it is commonly seen that the support structure is connected state plate for heat conduction, lower thermal deformation in normal energy density built [12], but it is difficult to remove the support of the porous structure. Therefore, the low energy density was applied to build the porous structure without the support structure. The effect of microstructure and mechanical property were reported during SEBAM process by in changing various parameters [13–18]. Above the line energy density 0.18 J/m built, the minimum relative density of 99.5% was observed [13]. Optimized overlap distance were 0.25–0.75 mm without fused line defects and the average value of microhardness was 360 HV [14]. However, over this range, it decreased ~340 HV on average. Different built orientations in XY, ZX, ZY, XY 30◦ and XY 60◦ inclination to the start plate, the microhardness, nearly 330 HV in the ZX-P specimen, was higher than the value ~275 HV in ZY [15]. With increasing build height and thickness, the tensile strength was decreased 3% in the top [16] and the thicker one had a lower microhardness of 320 HV than 362 HV of the thin sample [17]. The impeller was built with double influence factors to effective the microhardness with more than 10% change was reported [18], the microhardness between ~340 HV to ~360 HV depended on the locations of impeller. The unique microstructure may be due to complex cyclic thermal process in three main stages as rapid cooling from the molten state to the layer temperature, followed by a near isothermal stage at the local temperature until completion of the build, and finishing with a slow cooling to the room temperature [19,20]. In summary of microstructure evolution as mentioned above, micro size β grain transformation and α formation in the first stage, followed by the α martensitic being fully or partially transformed into an α + β structure. Moreover, the previous melted layer experiences a thermal history for each subsequent layer which will result in different microstructure depends on phase transformation path during the cooling. However, in previous research showed one parameter with one working process or the many samples compared with one specific parameter. In our experiments demonstrated that 3 parameters in one sample to indicative one workpiece made by more reasons to modify the finally commercial products manufacture [13–15,18,21].

It is worth noting that most efforts were made to successfully develop optimized porous structures for medical metallic implants with suitable combination process parameters in order to obtain better bone osseointegration and avoid the stress shielding effect. For example, as shown in Figure 1, the geometry of orthopedic implants is divided into solid and porous parts in gradient energy density built for specified purpose by using SEBAM process. The solid structure built in high density built for dense and strength purpose, but built in low energy built for non-support demands. Transition zone between solid and porous structure, namely of overlap zone with built in mixed high energy and low energy density. Therefore, gradient energy has been widely used for porous medical metallic implant production. To the best of the authors' knowledge, no study has been published to date on the overall characterization of gradient energy built parts, though, there are medical industrial demands for such parts fabricated by SEBAM. The specific research question this study addressed concerns the comprehension process of the effect on the microstructure and mechanical properties in gradient energy density built by SEBAM process. In this study, gradient energy density including three different of

energy density was selectively built in different zones (top, middle and top) of one rectangular sample by SEBAM method using Ti6Al4V (ELI) alloy powders on the effect of the microstructure and mechanical properties. It is supposed to be of significance to evaluate the performance of SEBAM-built parts. More importantly, a thorough understanding of gradient energy on microstructure and mechanical properties obtained from this study will aid to further propel the practice application of SEBAM-built metallic parts.

**Figure 1.** Schematic diagram of commercial porous fusion cage structure was fabricated by selective electron beam additive manufacture (SEBAM) method using gradient energy density on three different zones. Solid and porous parts were made independently by higher and lower energy density. The overlap site with 0.2 mm in width between solid and porous parts was made sequentially by higher and lower energy density.

#### **2. Materials and Methods**

#### *2.1. Sample Preparation*

In order to investigate the microstructure and micro-harness evolution of one SEBAMed (Kaohsiung, Taiwan) metallic part was built by SEBAM method in gradient energy density. A rectangular sample which 3 mm (thickness) × 20 mm (width) × 60 mm (height) in size was used and selectively in setting three different energy density namely high energy density (HED), low energy density (LED) and overlap energy density (OED) built on bottom (0 ≤ z ≤ 20), middle (20 ≤ z ≤ 40) and top 40 ≤ z ≤ 60 of the specimen respectively along the z axis building direction as shown in Figure 2. The energy density parameters in detail were described in the next section.

**Figure 2.** Schematic of SEBAMed samples made by higher, lower and overlap energy density from the bottom to top which labeled as high energy density (HED), overlap energy density (OED) and low energy density (LED) along build direction fabricated by electron beam additive manufacture. HED, LED built zones independently 26.5 and 16 J/mm3. OED built zone in 26.5 J/mm3 first and build sequentially 16 J/mm3.

#### *2.2. Selective Electron Beam Additive Manufacturing*

The Extra Low Interstitial Ti-6Al-4V (Grade 23) powder supplied by Arcam AB company was used; the powder size distribution was quoted as 45 to 100 μm. The chemical composition of Ti-6Al-4V powder was supplied as showed in Table 1 followed as ASTM F3001. Recycling of non-melted powder and/or sintered powder (Figure A1 in Appendix A) was achieved via powder recovery system and a mechanical vibrating sieve which mesh size ≤ 150 μm. All SEBAM parts were fabricated on the Arcam Q10 machine (Producer: Arcam EBM GE Additive company, Gothenburg, Sweden). The SEBAM process was implemented in the vacuum chamber with pressure below 5 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar in the beginning and finished with a pressure of 2 <sup>×</sup> <sup>10</sup>−<sup>5</sup> mbar. Each layer was setting preheated 730 ◦C by fast scanning with the defocused electron beam and layer thickness of 50 μm. Gradient energy density in building parameters was set as follows: the standard parameter of the beam current from 3 mA to 15 mA and scan speed from 1500 mm/s to 4530 mm/s with fixed layer thickness 50 μm, beam diameter of 100 μm and hatch spacing of 150 μm were performed. The energy density can be approximated as:

$$\text{E(J/mm3)} = \frac{\text{V(kV)} \times \text{C(mA)}}{\text{t} \, (\text{mm}) \times \text{h} (\text{mm}) \times \text{S} (\text{mm}/\text{s})} \tag{1}$$

where energy density E, operation voltage V, beam current C, layer thickness t, hatch distance h and scan speed S in Table 2. The high energy density (HED) resulting from the reduced scan speed and increased beam current was therefore 63% higher compared to low energy density (LED) setups. For the transition overlapping 2 mm in width was applied overlap energy density (OED), i.e., HED in the first scan and LED in the after.


**Table 1.** The chemical composition (in wt %) of Ti-6Al-4V powder.


**Table 2.** Main process parameters with fixed layer thickness 50 μm, beam diameter of 100 μm and hatch spacing of 150 μm used for the SEBAM process.

#### *2.3. Microstructure Observation*

Optical microscopy and scanning electron microscopy (SEM) of SEBAMed samples prepared by grinding and polishing- diamond abrasive disc with water used as coolant in the grinding process and polished with silicon carbide and 0.3 μm Al2O3 suspension. Optical and SEM samples were etched in 2.5% HNO3 + 5% HCl + 92.5% ethanol reagent. Secondary electron image SEM was carried on JEOL JSM-6380 (Tokyo, Japan) at 15 kV. Phase analysis was conducted using X-Ray Diffraction (XRD) (Panalytical B.V., Almelo, The Netherlands) and transmission electron microscopy (TEM) (PHILPLIES CM200 and JEOL 2000 FX, Tokyo, Japan). The characteristic CuKα radiation (λ = 1.5412 Å) in the 2θ range from 20 to 100 degrees having voltage 40 keV and current 40 mA was used. TEM samples were prepared using the following procedures: thin samples sections were manually ground to 0.06 ± 0.02 mm in thickness with silicon carbide paper and then the thin foils for TEM were electro polish using Automatic Twin-Jet Electropolisher Model 110 (Yokohama, Japan) at 25 V in a solution bath consisting of 12% perchloric acid, 15% glacial acetic acid and 75% ethanol reagent.

#### *2.4. Mechanical Properties*

Vicker microhardness (500 g, 15 s hold) of 15 individual measurements in each HED, OED and LED were performed on the metallographic samples using an Akashi MVK-H100 (Osaka, Japan) machine. The tensile tests on an MTS-10t (Eden Prairie, MN, USA) Tester using cylinder specimen with gauge length of 30 mm of specimen 3 (Figure A2 in Appendix B) according to ASTM-E8M specifications. The specimens were subjected to impact testing with the use of a Tinius Olsen model IT 504 polymeric impact testing machine (Tinius Olsen, Horsham, PA, USA). All tests were conducted at a pendulum capacity of 15 J, a drop height of 609.6 mm and velocity of 3.46 m/s.

#### **3. Results**

#### *3.1. E*ff*ect of Gradient Energy Density on the Defects*

Figure 3 shows representative specimens of the gradient energy from high energy density to low energy density built zones along the z axis building direction. The irregular pores which size exceeds 100 μm was only observed in the LED built specimen (Figure 3a) which is caused by the lack of fusion. Above line energy density more than 100 J/m is necessary for full densification of Ti6Al4V and below 100 J/m SEBAMed specimens contain more than 1% porosity as earlier reported [13]. This means that irregular pores did not obviously exist both in the HED and OED built zones, implying that highly dense sample can be fabricated by the SEBAM process if the energy density were more than 100 J/m. The irregular pores would greatly reduce yield stress, tensile stress and ductility [13].

In contrast the typical spherical pores several μm in size were found in all HED, OED and LED built samples in Figure 3b–d. It means no relation between the spherical pores and energy density and is mainly caused by the entrapped argon gas inside the Ti6Al4V powder during the production of plasma wire gas atomization [22]. The presence of limited small pores cannot be eliminated by hot isostatic press treatment [23] and did not significantly affect tensile properties. It might be decreased fatigue life.

**Figure 3.** Side view of specimen from bottom to top along the build direction (BD) denoted in red arrow of (**a**) showing the high visible pores which range from several hundred μm to several mm in size denoted by red arrow in LED built zone. SEM secondary electron images shows smaller pores several μm in size denoted by blue arrows in (**b**) LED, (**c**) OED and (**d**) HED built zones. Besides, α+β lamella and grain bound (GB) of α structure denoted by white arrows both were observed in (**c**) OED and (**d**) HED built zones.

#### *3.2. E*ff*ect of Gradient Energy Density on the Microstructure Evolution*

In order to understand the influence of the gradient energy density on three different sites, the microstructure was studied along the build direction. Figure 3b shows partially acicular α martensitic and α + β were developed by LED built zone where is at the top of the specimen. At the middle of the specimen built in the OED parameter, a lamella mixture of α+β structure was observed (Figure 3c). A coarser α+β lamella structure in the prior β grain was also observed at the bottom where is built in HED parameter (Figure 3d). No obvious difference in SEM analysis was found both in OED and HED built zones. However, the acicular α martensitic only was observed in LED built zone, which indicates cooling rate in LED zone was higher than OED and HED built zones.

Figure 4 shows the XRD analysis of the gradient energy density built by SEBAM method using Ti6Al4V powder. The characteristic peaks of dominant α phase with (101), (002) and small fraction of β phase with (110) were observed in all energy density built zones. Minor fraction of β phase in both OED and LED built zones compared to HED built zones. It is difficult to distinguish α and α phase by XRD analysis. Based on the mentioned above, the acicular α martensitic phase was only found in the LED built zone where the present characteristic peaks of α phase overlapped with α . Interestingly, the intermetallic phase Ti3Al was observed in HED built zone. Further TEM analysis in Figure 5 revealed that the nano-size intermetallic Ti3Al dispersed near the α grain boundary, and a fine lamella α+β structure. Moreover the crystallographic relationship between α, Ti3Al and β phase as [001]<sup>α</sup> // [001]Ti3Al and [011]<sup>α</sup> // [001]<sup>β</sup> are similar results to what Barriobero-Vila reported of selective laser melting with intrinsic heat treatments of Ti6Al4V [24]. The analysis of the formation of nanosize Ti3Al without the post heat treatment or repetitive melting (intrinsic heat treatment) during the SEBAM process will be discussed in the following section.

**Figure 4.** The X-ray diffraction spectrum of HED, LED and OED built samples. The α-Ti6Al4V characteristic (100), (002) and (101) and β-Ti6Al4V characteristic (110) peaks were observed in all samples. The characteristic peaks of α phase are overlapped with α martensitic phase only in LED built sample. Minor Ti3Al phase of its characteristic peak (110) was observed only in HED built sample.

**Figure 5.** (**a**) Bright field image of ultrafine α+β lamella microstructure found at the center of HED sample. (**b**) Select area diffraction patter from the cycle region (I) in (**a**) showing the present Ti3Al within an α lamella matrix with crystallographic relationship [001]α//[001]Ti3Al (**c**) Select area diffraction patter from the cycle region (II) in (**a**) showing the ultrafine α + β lamella microstructure with crystallographic relationship [001>]α // [011]β.

#### *3.3. E*ff*ects of Gradient Energy Density on the Mechanical Properties*

A continuous microhardness on the HED, OED and LED built zones were measured in Figure 6. Apparently, it was divided into two groups via gradient energy density built. The microhardness values in average 320 to 350 HV also revealed no obvious difference between HED and OED built zones. Nevertheless, a relatively high microhardness (from 375 to 420 HV) was observed in the LED built zone. The higher microhardness in the LED built zone was caused by the acicular α martensitic microstructure (Figure 3b) and it is consistent with those reported for selective laser melting built Ti6Al4V alloy [25,26]. Moreover, the measure average α lath width in the HED and OED built zones were determined to be 0.69 ± 0.04 μm and 0.59 ± 0.14 μm respectively. The result is in good agreement with no obvious difference in mircohardness in the two zones. However, the microhardness in the OED built (double melting) zone showed the finer α+β lamella and grain bound of α structure than in the HED built zone. The result was in contrast to the paper showed the coarsening microstructure was caused by double melting in the same build height [14]. This implies that double melting is not only seen to be relevant to grain size. The effect of thermal history from the different building height might have a powerful impact on grain size [18].

**Figure 6.** Vickers microhardness profile was measured from bottom to top (black arrow) of the cross section specimen which were built by gradient energy. HED, OED and LED built zones are indicated by black, red and blue symbols.

Stress-strain curves and impact energy of three different energy density built samples were shown in Figure 7. The ultimate tensile strength (UTS) estimated 228 MPa of the LED built sample lower obviously compared to the OED built and HED built samples in Table 2. The presence of the obvious plastic deformation both in OED and HED curves were distinct with LED curve which almost no plastic deformation was observed. Both OED and HED curves had no obvious difference from the UTS and elongation, however in the finally the HED curve was slightly reduced the strain. Higher elastic modulus of OED and HED built samples compared to LED built sample. Interestingly, the highest hardness was in the LED built zone due to rapid cooling leading to acicular α martensitic but it had the smaller UTS and elongation compared to OED built and HED built sample. To further understand the variation of the mechanical properties, the fracture morphology was observed in Figure 7b–d. The transgranular ductile dimple tearing resulting from the coalescence of microvoids fracture surface was observed in both OED and HED built samples. Crack propagation of fine dimples at the tensile fracture indicated the extent of plastic deformation. Besides, according to the Hall-Petch strength mechanism, smaller grain size provided more grain boundaries, which can impede the movement of the dislocation, demonstrated as OED built sample has slightly higher tensile stress and elongation compared to HED built sample. In contrast, the fracture surface included cleavage, unmelted and partially melted powder were shown in LED built sample. This implied insufficient energy density that was unable to melt the deeper layer had a negative effect on tensile stress and elongation properties. This result is consistent with others reported [13,27].

**Figure 7.** (**a**) Tensile stress-strain curves of gradient energy built samples. LED, OED and HED built samples were respectively denoted by red, blue and black curves. Representative fracture surface of tensile test of different density built in (**b**–**d**). (**b**) Unmelted powder and brittle cleavage in LED built sample, (**c**,**d**) both tensile fracture morphologies show fine and deeper dimples in HED and OED built samples. Representative fracture surface of impact test of different density built in (**e**–**g**). (**e**) Unmelted powders and voids were observed in LED samples, (**f**,**g**) both impact fracture morphologies show fine and deeper dimples in HED and OED built samples.

The average fracture toughness properties in different energy density were shown in Table 3. Both OED and HED built samples showed the higher impact energy compared to the LED built sample which had the lower fracture toughness properties (2 to 3.2 J). These are similar results to the tensile stress in the fracture surface. Both fracture surfaces as transgranular ductile dimple fractures from the coalescence of microvoids in OED and HED built samples were observed in Figure 7b–f. However, the lower value of impact energy of fracture surface in LED built sample exhibited pores, unmelted powder and partial melt region (See Figure 7d,g). This means lower energy density may contributed to large area of pores and unmelted powder [13,27].


**Table 3.** Mechanical properties of SEBAMed Ti6Al4V in varying energy density.

#### **4. Discussion**

#### *4.1. Formation Gradient Microstructure on the Gradient Energy Density Built Zones*

The gradient microstructure was built by gradient energy density from the bottom to top of the specimen. In this work, the microstructure of the particular zones in different energy density built will get the gradient microstructure (α + α + β in LED built zone, slightly finer α + β lamella structure in the OED built zone and α+β lamella with minor fraction of nanosize Ti3Al intermetallic in HED built zone were observed in Figures 3–5). Moreover, all the specimens showed the microhardness values between 304 to 420 HV. The highest microhardness was contributed to the presence of α martensitic phase (Figure 3b) in LED built zone. However, no α phase was observed and only fine α+β lamella were observed both in the OED and HED built zones where the average microhardness of 340 HV was shown in Figure 6. The presence of limited small amount Ti3Al did not significantly affect the microhardness in the HED built zone compared to the microhardness in the OED built zone. In the present study, the thermal events were more complex due to a combination of the building height and gradient energy density.It is showed that the difference of previously report as varying microstructure only in changing building direction [15] process parameters [28,29], building high [18,30], overlap distance [14], heat treatment [19], intrinsic heat treatment [24]. Therefore, the analysis of formation of gradient microstructure/phase in gradient built zones will be discussed in the following section.

#### 4.1.1. Formation α Martensitic in LED Built Zone

The thermal gradient, G (K/m), and solidification velocity, R (m/s) can be used to predict solidification microstructure where G/R controls the mode of solidification (morphology) and G × R (cooling rate) in SEBAM process [31]. For an increasing energy density, the cooling rate became slower as was shown in solidification map for SEBAM Ti6Al4V melted [31]. Moreover, the cooling rate of the top was higher than the bottom also reported in SEBAMed Ti6Al4V parts [18]. In the present study, the SEM investigations in Figure 3 revealed that formation of acicular α martensitic in LED-built zone where at the top of the specimen along the build direction. Both the low energy built and the top of the building height caused a high cooling rate. Moreover, during the SEBAM process, as the electron beam scanned over the Ti-6Al-4V powder, a melt pool formed and then was rapidly solidified into β grains. The prior β grains underwent a rapid cooling (a critical cooling rate of >410 ◦C/s) transforming into martensitic α phase. Moreover, it its known that the build temperature of 600~650 ◦C in the chamber was below martensitic start temperature (Ms) 800 ◦C. Thus, α martensitic phase should be formed after melting.

#### 4.1.2. Formation α+β in OED Built and HED Built Zones

The observation of the α+β lamella structure inside OED built (Figure 3c) and HED built (Figure 3d) samples also proves the aforementioned process. Following the above mentioned formation of martensitic α , β phase was transformed into an α (β → α ) reaction on cooling. In addition, α martensitic will decompose into α and β phases, i.e., α → α + β, when it is subjected to isothermal annealing in the α + β two phase field. It was reported that α in the thicker sample would be able to decompose into α/β phases due to higher thermal mass causing the slower cooling. During the EBM process, an electron beam was constantly scanning over the Ti6Al4V powder bed and the heat was mainly transferred from top to bottom through the built parts, the temperature of the newly deposited layers must be higher than 650 ◦C in the middle and bottom of the specimen that was sufficient to enable α decomposition.

#### 4.1.3. Formation Ti3Al in HED Built Zones

The formation of the intermetallic Ti3Al phase along the building direction of the bottom of the specimen was revealed using X-ray diffraction by the presence of (001) characteristic peak as shown in Figure 4 as well as by select area diffraction carried out during TEM investigations (Figure 5). According to previous investigations of SLMed Ti6Al4V and SLMed Ti6Al4V with intrinsic heat treatment, the driving force for Ti3Al formation due to the fast solidification during the SLM process, concentration of aluminum and oxygen [24]. Moreover, Al depletion at the interface during EBM process was reported [29]. In this work, diffraction pattern evidence of Ti3Al was only observed in HED built zone (the bottom of the specimen) mostly affected by heat transfer from the following layer. Since precipitation of this phase can occur during aging at 500–600 ◦C for several hours [31], it is reasonable to conclude that this range of temperatures was reached by the scanning of following layers, lead to the precipitation of an intermetallic Ti3Al.

#### *4.2. Cooling Path in Gradient Energy Density and Subsequence Heat Transfer*

From the microstructure point of view as mentioned above in three different zones. The α + α + β in the LED built zone where it is the top of the specimen, α + β lamella structure in OED built zone where it is the middle of the specimen and α + β lamella with small amount of Ti3Al in HED built zone where it is the bottom of the specimen, were observed due to different thermal history. It implied the cooling rate of the top (LED built zone) was faster than the middle (OED built zone) and the bottom (HED built zone) of the specimen. Further the evidence of slight smaller width of α lath in the middle in Figure 3 implied slight higher cooling rate in the middle compared to the bottom. Based on the literature reported the complex thermal events in the EBM process can be simplified in three main stages [19]. The first step is a rapid cooling from the molten state to the layer temperature, followed by a quasi-isothermal stage at the local cooling until completion of the build, and finishing with a slow cooling to room temperature. Gradient energy density and different locations causing different thermal events were predicted in Figure 8 base on the microstructure analysis as mentioned above. Moreover, the heat coming from the following layer melting will transfer to the previous layer causing the similar effect as aging heat treatment was observed only at the bottom. Those observed results proved the effect combined gradient energy density and different locations on microstructure and phase transformation due to different thermal history.

**Figure 8.** Prediction cooling path and phase transformation based on the literature [19] of LED, OED and LED parameters built in top, middle and bottom of the specimen respectively. The difference in cooling rate result from different energy density and spatial location thus contribute to different microstructure. i.e., α + α + β in LED built zone, α + β lamella structure in the OED built zone and α + β + Ti3Al intermetallic phase in the HED built zone were observed.

#### **5. Conclusions**

In this work, gradient energy density was carried out on a specimen from the bottom to the top by SEBAM technology. The effect of energy density on the microstructure and mechanical properties of SEBAMed Ti6Al4V was investigated by SEM, X-ray, TEM, microhardness, tensile and impact tests. Based on the experimental results, the following conclusions can be drawn:


**Author Contributions:** T.-I.H. and M.-H.T. wrote the main manuscript text and prepared all figures. Y.-T.J. carried out the SEM experiments and prepared all the sample. T.-I.H. performed the TEM analysis. M.-H.T. designed all the experimental and edited the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Ministry of Science and Technology, grant number 108-2218-E-992-320-MY2.

**Acknowledgments:** The authors gratefully acknowledge the sponsorship from Ministry of Science and Technology of Taiwan, ROC, under the project No. MOST 108-2218-E-992-320-MY2. Meanwhile, the authors are thankful for the EBAM lab equipment and financial support from Metal Industry Research & Development Centre (MIRDC). **Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A Ti6Al4V Powder Morphology**

**Figure A1.** The secondary electron image of Ti6Al4V powder in (**a**) fresh powder and (**b**) recycle powder. Both types of powder were spherically shaped.

#### **Appendix B The Test Sample of Tensile Test**

**Figure A2.** The cylinder test sample according to ASTM E8 standards with standard size, printed and machining size in (**a**). (**b**) The machined built tensile test samples which were built in LED parameter.

#### **References**


Nickel-Base Superalloy Fabricated by Electron Beam Melting. *Metall. Mater. Trans. A* **2011**, *42*, 3491–3508. [CrossRef]


© 2020 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/).

**Jan Palán 1,\*, Radek Procházka 1, Jan Džugan 1, Jan Nacházel 1, Michal Duchek 1, Gergely Németh 2,3, Kristián Máthis 2, Peter Minárik <sup>2</sup> and Klaudia Horváth 2,3**


Received: 21 November 2018; Accepted: 10 December 2018; Published: 11 December 2018 -

**Abstract:** This paper describes the mechanical properties and microstructure of commercially pure titanium (Grade 2) processed with Conform severe plastic deformation (SPD) and rotary swaging techniques. This technology enables ultrafine-grained to nanocrystalline wires to be produced in a continuous process. A comprehensive description is given of those properties which should enable straightforward implementation of the material in medical applications. Conform SPD processing has led to a dramatic refinement of the initial microstructure, producing equiaxed grains already in the first pass. The mean grain size in the transverse direction was 320 nm. Further passes did not lead to any additional appreciable grain refinement. The subsequent rotary swaging caused fine grains to become elongated. A single Conform SPD pass and subsequent rotary swaging resulted in an ultimate strength of 1060 MPa and elongation of 12%. The achieved fatigue limit was 396 MPa. This paper describes the production possibilities of ultrafine to nanocrystalline wires made of pure titanium and points out the possibility of serial production, particularly in medical implants.

**Keywords:** titanium; ECAP; Conform; continuous extrusion; wire; medical implants

#### **1. Introduction**

Much work has recently been done in the field of severe plastic deformation (SPD) processing of pure titanium [1]. These processes have shown great potential as they have delivered a more than two-fold improvement in mechanical properties [2–5]. With improved mechanical properties, pure titanium has become suitable for applications that require high strength levels (dental implants, bone screws, etc.). Besides that, implants with a smaller cross-section can be designed. For instance, implants with smaller load-bearing cross-sections are suitable for children [6]. In terms of biocompatibility, pure titanium is associated with increased osteoblast proliferation and better osseointegration [6,7]. Kim and Park [7] established a model explaining enhanced surface cell attachment on nanocrystallized commercially pure titanium (CP–Ti). The reason for this is that the number of nodules at the triple-point junctions of the grain microstructure increases as the grain size is refined through equal-channel angular pressing (ECAP). These properties facilitate a connection between the implant and the human tissue [6–10].

Commercialisation of ultrafine-grained materials prepared by SPD techniques is hindered by the fact that these techniques are rarely suitable for production outside the laboratory [1,11,12]. These techniques

were generally intended for microstructural evolution observations and assessment of the properties of the investigated materials. Hence, the next step is to transfer laboratory procedures to industrial practice. Some attempts to develop processes for the continuous production of ultrafine-grained materials have taken place recently. The attempts included wire production methods such as equal-channel angular swaging (ECAS) [13] and Conform ECAP [14–16]. Their commercialisation is under way, particularly in the field of pure titanium wire processing. This development is associated with a research group headed by R. Z. Valiev [17], which is responsible for the development of Conform ECAP. Conform ECAP processing is a relatively new modification of the conventional ECAP technique. In this process, the principle used to generate the frictional force to push a work-piece through an ECAP die is similar to the Conform process, while a modified ECAP die design is used so that the work-piece can be repetitively processed to produce ultrafine-grained (UFG) structures. The angle of channel intersection is 120◦. The input stock for this method is a square bar of 11 × 11 mm<sup>2</sup> cross-sectional area [14]. To obtain the desired properties, six passes at 200–400 ◦C are typically carried out. The resultant semi-finished product is transferred to a wire-drawing machine, where it is worked at an elevated temperature [17,18]. This process is known as the thermomechanical treatment (TMT) of pure titanium. Wire drawing leads to intensive strengthening. Products of the Conform ECAP method have equiaxed grains with high-angle boundaries. During the subsequent drawing operation, these fine grains become elongated, their dislocation density steeply increases and the material gradually loses its plastic deformation capability. For instance, processing of Grade 4 commercially pure (CP) titanium through six passes in Conform ECAP at 200 ◦C and subsequent drawing at 200 ◦C leads to a strength of 1267 MPa [17].

An alternative for the continuous production of ultrafine-grained wires is the Conform SPD technique (Figure 1) [19]. This technique is being explored by the company COMTES FHT. Its principle is similar to that of Conform ECAP, as the feedstock is continuously forced through a chamber along an angled channel. Unlike Conform ECAP, however, Conform SPD uses round stock and the angle of its channel intersection is 90◦ [2–4,19]. The principle of forming titanium using Conform SPD technology to obtain ultrafine-grained products is shown in Figure 1. The feedstock is guided by the coining roll to the gap between the driving wheel and the shoe. High friction forces cause the feedstock to move along the groove in the driving wheel all the way to the abutment. Once the material hits the abutment, it changes direction and exits the Conform SPD machine through the chamber die [20–24].

This paper gives a description of the processing of CP Grade 2 titanium by a combination of Conform SPD and cold rotary swaging, and includes an overview of the final properties. It summarises information which is expected to facilitate the implementation of this material in the serial production of medical implants.

**Figure 1.** Schematic representation of the Conform severe plastic deformation (SPD) technique.

#### **2. Materials and Methods**

This study concerns commercially pure Grade 2 titanium (ASTM B348 Gr2). Its chemical composition is given in Table 1. The composition was determined by means of the Bruker Q4 Tasman optical emission spectrometer. The initial diameter of the wire was 10 mm. The initial material was delivered with respect to the ASTM B348 standard.


**Table 1.** Chemical composition of feedstock, wt. %.

The wire was processed by the Conform SPD technology. Figure 2 shows the distribution of the strain rate, material velocity and the temperature in the die chamber, respectively. They were calculated with the DEFORM-3D software. Detailed conditions for these calculations are described in the publication [25]. The calculations described in this publication are for information only. The feedstock had a round cross-section of 10 mm diameter. At the start of the processing, it was at room temperature and no external heating was provided. Up to three passes were completed (1 pass, 2 passes, 3 passes). The driving wheel ran at 0.5 rpm. The A route was used, i.e., the feedstock orientation was identical in all passes.

**Figure 2.** Numerical modelling of the Conform SPD process: (**a**) Strain rate distribution (s−1); (**b**) Velocity distribution (mm/s); (**c**) Temperature distribution (◦C).

In order to further increase the mechanical properties of the material, cold rotary swaging was applied in the next step. Rotary swaging was performed on the as-received material and on the material after a single pass through Conform SPD. The purpose was to compare the final properties after rotary swaging combined with Conform SPD, and after rotary swaging alone. The diameter of the initial feedstock was the same as for Conform SPD—10 mm. One operation involved the reduction of the cross-sectional area by 20%. The total reduction in cross-sectional area was up to 80%.

Metallographic specimens were prepared using a standard procedure involving grinding and subsequent polishing. The microstructure was revealed by etching with Kroll reagent. A Carl Zeiss—Observer, the Z1m optical microscope, and bright-field illumination were used for microstructure observation.

Fractographic examination was performed using the JEOL JSM 6380 scanning electron microscope. Fracture surfaces of fatigue test specimens were observed using secondary electron imaging.

Thin foils were prepared for transmission electron microscope (TEM) observations. The final electrolytic thinning was performed with the use of a Tenupol 5 device, using a solution of 300 mL CH3OH + 175 mL 2-butanol + 30 mL HClO4 at −10 ◦C and a voltage of 40 V. The TEM analysis was performed in a JEOL 200CX instrument, with an acceleration voltage of 200 kV. Selective electron diffraction was used for determining phases. The grain size was measured using the linear intercept method.

The electron-backscatter diffraction (EBSD) observations were performed on the cross-section of the samples using an FEI Quanta 200 FX scanning electron microscope (SEM, Thermo Fisher Scientific, Brno, Czech Republic). The EBSD measurements were conducted at a working distance of 13 mm using

a step size of 50 nm at 10 kV acceleration voltage. The exact sizes of the scanned areas are indicated in the captions of particular figures. Prior to the EBSD measurement, the surfaces of the samples were ground on SiC papers (from 320 to 1200 grit) and subsequently polished for 24 h in a three-step vibratory polishing procedure. The surfaces of the specimens were finally ion-beam-polished on a Leica EM RES102 system (Leica Mikrosysteme, Wetzlar, Germany) before observation. In order to observe the microstructure of the rotary swaged sample transmission, the Kikuchi diffraction (TKD) method was used. The TKD method is also referred to as "transmission EBSD". For ultra-fine grained materials with grain size from a few tens to hundreds of nanometers, the resolution of conventional EBSD is not sufficient since the interaction volume is comparable to the grain size. Moreover, in heavily deformed materials, EBSD analysis is obscured by high dislocation density, residual strains and lattice rotations. Transmission EBSD was measured in a Zeiss Auriga Compact FIB-SEM (Jena, Germany) using a step of 10 nm at 30 kV acceleration voltage on a standard TEM foil.

Tensile testing was carried out in an electromechanical testing machine under quasi-static loading conditions, at a constant strain rate of 0.0002/s at room temperature. Round bar tensile specimens with a diameter of 3 mm and 5 mm, respectively, were employed. Samples of 5 mm in diameter were used for wires after Conform processing. Samples of 3 mm in diameter were used for wires after Conform processing and rotary swaging. A mechanical extensometer was used for strain measurement for both testing specimen geometries.

The conventional fatigue testing procedure consists of cyclic load application on sub-sized round-bar specimens of an hourglass shape of 1.5 mm in diameter, with sinusoidal waveform at constant amplitude in the tension-compression mode. Specimens were intended to be tested under cyclic loading, with stress ratio R = −1 at room temperature. Due to the nature of bulk ultrafine-grained (UFG) materials, it was possible to use a test frequency up to 50 Hz without signs of self-heating in the high cyclic fatigue (HCF) regime verified by thermography measurements. Tests were conducted on an INOVA servo-hydraulic testing machine with a loading capacity of 15 kN. Tests were run until failure (Nf) or up to 10 million cycles (red circles in graphs). The test results are plotted on semi-log scale in Figure 8. Due to the data scattering, a statistical approach using a 90% two-sided confidence interval was added to the data plot. The fatigue limit of 396 MPa was estimated based on two specimens surviving 10 million cycles without failure.

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

#### *3.1. Microstructure Evolution in Pure Titanium after Conform SPD and Rotary Swaging*

The microstructure of the feedstock consisted of equiaxed recrystallized grains with annealing twins (Figure 3). These annealing twins can be seen in Figure 3a. They are a characteristic feature in materials with a hexagonal crystallographic structure due to their limited potential for deformation slip at low deformations [26]. The twins are broad and lenticular.

**Figure 3.** (**a**) Micrograph of the as-received structure in the transverse direction; (**b**) detailed micrograph of the as-received structure.

Figure 4 shows transmission electron micrographs of the material processed by Conform SPD with various numbers of passes. Table 2 lists average grain sizes after various numbers of passes through the Conform SPD machine. Figure 4a,b show a general view of the substructure in the transverse and longitudinal direction after the first pass. There are polyhedral grains with non-uniform dislocation density (Figure 4a) and regions with elongated grains with low-angle boundaries (Figure 4b). Variations in the grain morphology were found over the volume and the microstructure is not homogenous after the first pass. The variations in grain morphology are an indication of the non-uniform thermomechanical conditions (temperature distribution, stress state, and strain rate) within the material—the non-homogeneous character of the process is clearly visible in Figure 2.

The second and third pass produced equiaxed grains (Figure 4, from c to f). The third pass even resulted in a larger mean grain size of 420 nm in the longitudinal direction (Table 2). This increase can be attributed, in part, to non-uniform deformation and to the high surface activity of the UFG structure. High surface activity and dislocation density (strain magnitude) lower the temperatures of softening processes [4]. Certain grain growth can be expected to occur during formation due to deformation heat and the heat retained in the die chamber. This effect proves that the deformation heat (Figure 2c) is very likely to be a contributing factor in the recrystallization (post dynamic recrystallization) of the refined microstructure. The generation and effects of deformation heat on softening processes are often ignored in SPD applications. Furthermore, the decrease in grain size after the two passes is smaller than after the first pass. The refinement effect of Conform SPD seems to be limited when the initial microstructure is in the sub-micron scale [27]. A recent study [28] has shown that small grains may cause low stability when the metals are heavily deformed, and additional straining may occur as Conform SPD does not further reduce the grain size because of the intrinsic instability of nano-sized (below 100 nm) and submicrometre-sized (between 100 and 1000 nm) grains [11]. The dynamic balance of grain refinement between structure refinement and recovery at ambient temperature occurred, which was already proven in the stated study [27]. Thus, the grain refinement phenomena of Conform SPD exhibit a limit, and with additional passes, the grain size and dislocation density will reach their ultimate values. Conform SPD processing leads to substantial grain refinement and to mostly equiaxed grains in the work-piece, even in a single pass. These grains are so small that they only rotate in the subsequent passes, after which the microstructure becomes more homogeneous [29]. The mechanism by which the subgrains rotate is not so well understood. Wu et al. [11] describe a process in which dislocation motion becomes restricted due to the small subgrain size and grain rotation becomes more energetically favourable [30]. Mishra et al. [29] proposed a slightly different explanation, in which the rotation is aided by diffusion along the grain boundaries (which is much faster than through the grain interior) [29]. Conform SPD therefore reaches its grain refinement limit in just one pass. Further passes mostly lead to nothing more than homogenization of the microstructure. However, according to some studies, the maximum (limit) grain refinement is only achieved after multiple passes [11,17,18].

**Figure 4.** *Cont*.

**Figure 4.** (**a**) Substructure in the transverse direction after the first pass; (**b**) substructure in the longitudinal direction after the first pass; (**c**) substructure in the transverse direction after the second pass; (**d**) substructure in the longitudinal direction after the second pass; (**e**) substructure in the transverse direction after the third pass; (**f**) substructure in the longitudinal direction after the third pass.

**Table 2.** Mean grain size for different processing steps.


In order to further decrease the grain size, cold rotary swaging was employed on the material which had already been processed using the Conform device. The product of a single pass through the Conform SPD machine was rotary-swaged with 80% total reduction in the cross-sectional area. Small grains thus extended in the longitudinal direction and their dislocation density steeply increased (darker areas). This is illustrated in Figure 5.

**Figure 5.** (**a**) Substructure in the longitudinal direction after one pass through the Conform SPD machine and rotary swaging; (**b**) substructure in the longitudinal direction after one pass and rotary swaging.

#### *3.2. Texture Evaluation*

The results of the EBSD measurements conducted in the centres of samples after one, two and three Conform SPD passes are shown in Figure 6. With increasing numbers of passes, significant grain refinement occurs (please note the different scales of the images). The texture evolution is also shown in the inverse pole figure (IPF) maps in Figure 6. After the first pass, an intensive maximum can be observed close to the (0001) pole. With further passes, this maximum weakens and a new component in the inclined area between the horizontal and vertical axes is observed. This new component is dominant in the sample after three Conform SPD passes. A similar texture evolution was observed in ECAPed Mg alloys containing rare earth elements, where the intensity of this maximum was inclined by ≈ 55◦ from the processing direction [31]. The reason for the appearance of this component is most probably given by the activation of non-basal slip systems.

**Figure 6.** Cross-section orientation image maps (OIM) and inverse pole figure (IPF) maps for samples after (**a**) first pass (50 × 50 μm), (**b**) second pass (20 × 20 μm), and (**c**) third pass (15 × 15 μm) through Conform SPD. In the figure on the left, the scale of texture intensities is shown as multiples of the random density (m.r.d.) from 0–8.000. The maximum value of each texture is listed below the IPF maps. The orientation triangle for the electron-backscatter diffraction (EBSD) maps is shown in the right-hand corner of the figure.

Figure 7 presents the transmission EBSD map of the sample after one SPD pass followed by rotary swaging. The microstructure, similarly to that after one SPD pass, is not homogeneous. It contains some coarse grains with some low-angle grain boundaries near the recrystallized area. The sample has a very intensive fibre texture with basal planes oriented parallel to the SPD direction. This texture is not only observed in the coarse grains but also in the recrystallized ones.

**Figure 7.** Transmission EBSD map of the sample after one Conform SPD pass and rotary swaging (3 × 3 μm). High-angle grain boundaries (>15◦) are marked by black lines, while the low-angle grain boundaries (between 4◦ and 15◦) are marked by white lines. The texture intensities are carried out as the multiples of random density (m.r.d.) from 0–8.000, with maximum at 12.409 m.r.d.

#### *3.3. The Effect of Conform SPD and Rotary Swaging on Tensile Properties*

Table 3 lists the tensile tests results. Materials in several differing conditions were tested after various numbers of passes through the Conform SPD machine and after rotary swaging with a reduction in cross-sectional area. The first pass through the Conform SPD machine led to an increase in Ultimate Tensile Strength (UTS) from 480 MPa for the as-received material to 580 MPa, whereas only a minimum decrease in A5 elongation occurred. Further passes did not bring any significant strengthening. After the second pass, the strength was 600 MPa and after the third pass it reached 623 MPa. These increases in strength were not accompanied by decreases in A5 elongation. These results are in agreement with the trends in grain size, since further passes caused no additional refinement and therefore there was no additional increase in strength. The cross-sectional area of a wire processed with one pass through the Conform SPD machine was further reduced by rotary swaging. It led to a notable increase in ultimate strength and yield strength. The ultimate strength was 1060 MPa, which is twice as high as in the as-received material. On the other hand, elongation was a mere 12%, in contrast to the 25% achieved with just the Conform SPD processing. This can be attributed to a much higher dislocation density and the elongation of the initially equiaxed grains [4]. The reduction in area was greater than in the initial material. This finding is in agreement with the study reported in [4]. This improvement in the reduction of area was observed for all post-Conform SPD conditions. The rotary-swaged specimen identified as "rotary swaging 80% area reduction" had a strength of 964 MPa. This is approximately 100 MPa less than in the specimen after one pass through the Conform SPD machine and rotary swaging to the same reduction level of the cross-sectional area. The elongation was 3% lower and the reduction in area was almost 24% lower in the first specimen than in the latter one. These values suggest that ductility increases for Conform SPD-processed specimens.

**Table 3.** Mechanical properties after Conform SPD processing and after Conform SPD + rotary swaging. Ultimate tensile strength (UTS); offset yield (OYS); reduction in area (RA); elongation (A5).


#### *3.4. Evaluation of Fatigue Properties*

Figure 8 shows fatigue data for the specimen that underwent one pass in Conform SPD and was then rotary-swaged to an 80% reduction in cross-sectional area. This type of processing gives high values of mechanical properties while maintaining the process productivity. Its fatigue strength was 396 MPa. The fatigue strength of Grade 2 titanium in the ASTM B348 condition is approximately 240 MPa [32]. Comparing those values indicates a demonstrable increase in fatigue strength and the impact of the grain size on fatigue. Similar results were reported by other authors [32,33]. The results show that the fatigue strength of ultra-fine to nano-grained titanium at 10<sup>7</sup> cycles is 60 MPa higher than conventional CP titanium Grade 4 but does not exceed that of the Ti-6Al-4V alloy, which has 530 MPa. It is clear that increasing the fatigue strength of CP titanium depends on tensile strength, this relationship being characteristic of titanium, as opposed to wavy-slip fcc materials [5]. The explanation is that the cross-slip of dislocations is more difficult in the hcp lattice. Therefore, the fatigue strength of Ti depends on the parameters of the size and shape of the grains and the type of boundaries. The twinning mechanism does not play a key role in the cyclic deformation of UFG titanium, and the fatigue mechanisms are likely related to the grain boundaries. Figure 9 shows images of the fracture surface of a ruptured fatigue test sample. It is obvious that the fatigue area covers about 60% of the fracture surface. The arrow points in the crack propagation direction. The fracture surface has been strongly smoothed out [5,27].

**Figure 8.** S–N curve for the sample after Conform SPD one pass + rotary swaging (80% area reduction).

**Figure 9.** Fracture surface of the sample (Conform SPD one pass + rotary swaging (80% area reduction)) after (**a**) fatigue testing and (**b**) fracture initiation area.

In summary, ultra-fine to nano-grained Ti (Grade 2) can replace conventional Ti Grade 4 with the assumption of increased component life, as the fatigue strength is roughly 60 MPa higher. The feasibility of replacing Ti-6Al-4V is still in question because its fatigue strength is about 130 MPa higher. On the other hand, the ultimate strength is about 200 MPa higher. Another advantage of commercially pure Ti is its enhanced biocompatibility as compared to the Ti-6Al-4V alloy. Recent studies, in particular, point out the toxic effects of Al and V after release of these elements into the human body. An assessment of fatigue strength of a dental implant made from ultra-fine to nano-grained titanium Grade 2 is already being evaluated by COMTES FHT and will be the subject of further articles and studies.

#### **4. Conclusions**

This paper presents the successful processing of commercially pure Grade 2 titanium with Conform SPD and rotary swaging techniques. The proposed route enables high-strength wires with an ultrafine to nanocrystalline microstructure to be produced in a continuous process. The article primarily describes the characteristics of the high-strength wire. The following conclusions can be stated:

• Processing with Conform SPD already leads to dramatic grain refinement in the first pass. The average grain size was 320 nm. Subsequent rotary swaging further reduced the average grain

size. Grains were preferentially elongated in the longitudinal direction and the sample has a very intensive fibre texture, with basal planes oriented parallel to the longitudinal direction.


**Author Contributions:** J.P. and M.D. conceived and designed the technological experiments and conducted the final evaluation of the results. R.P. and J.D. conceived and designed the fatigue tests. J.N. carried out the metallographic evaluation. G.N., P.M., K.H. and K.M. conceived and designed the texture evaluation.

**Funding:** This research was developed under the project entitled Development of West-Bohemian Centre of Materials and Metallurgy No.: LO1412, which is financed by the Ministry of Education of the Czech Republic.

**Acknowledgments:** This paper was developed under the project entitled Development of West-Bohemian Centre of Materials and Metallurgy No.: LO1412, which is financed by the Ministry of Education of the Czech Republic. K.M. is grateful for support from the Czech Science Foundation (GB 14-36566G). G.N. and K.H. acknowledge the support of the Operational Programme Research, Development and Education, The Ministry of Education, Youth and Sports (OP RDE, MEYS) [CZ.02.1.01/0.0/0.0/16\_013/0001794].

**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/).

### *Article* **Bioactive Sphene-Based Ceramic Coatings on cpTi Substrates for Dental Implants: An In Vitro Study**

**Hamada Elsayed 1,2, Giulia Brunello 3,4, Chiara Gardin 5,6, Letizia Ferroni 5,6, Denis Badocco 7, Paolo Pastore 7, Stefano Sivolella 4, Barbara Zavan 5,6 and Lisa Biasetto 3,\***


Received: 27 September 2018; Accepted: 5 November 2018; Published: 9 November 2018 -

**Abstract:** Titanium implant surface modifications have been widely investigated to favor the process of osseointegration. The present work aimed to evaluate the effect of sphene (CaTiSiO5) biocoating, on titanium substrates, on the in vitro osteogenic differentiation of Human Adipose-Derived Stem Cells (hADSCs). Sphene bioceramic coatings were prepared using preceramic polymers and nano-sized active fillers and deposited by spray coating. Scanning Electron Microscopy (SEM) analysis, surface roughness measurements and X-ray diffraction analysis were performed. The chemical stability of the coatings in Tris-HCl solution was investigated. In vitro studies were performed by means of proliferation test of hADSCs seeded on coated and uncoated samples after 21 days. Methyl Thiazolyl-Tetrazolium (MTT) test and immunofluorescent staining with phalloidin confirmed the in vitro biocompatibility of both substrates. In vitro osteogenic differentiation of the cells was evaluated using Alizarin Red S staining and quantification assay and real-time PCR (Polymerase Chain Reaction). When hADSCs were cultured in the presence of Osteogenic Differentiation Medium, a significantly higher accumulation of calcium deposits onto the sphene-coated surfaces than on uncoated controls was detected. Osteogenic differentiation on both samples was confirmed by PCR. The proposed coating seems to be promising for dental and orthopedic implants, in terms of composition and deposition technology.

**Keywords:** implant; titanium; osseointegration; biocompatibility; bioactive ceramic coatings; sphene

#### **1. Introduction**

Commercially pure titanium (cpTi) and its alloys represent the materials of choice for manufacturing orthopedic and dental implants because of their high mechanical properties, good corrosion resistance and excellent biocompatibility [1–4].

The successful outcome of the dental implants mainly depends on the osseointegration at bone-implant level. In recent years, several surface treatments have been applied to Ti implants with the aim of increasing the speed and success rate of osseointegration [5,6]. Surface modifications included acid etching, isothermal oxidation, hydrothermal synthesis method, and combination of sandblasting and acid etching (SLA) [5,6].

Bioactive coatings appear to be promising materials favoring a faster and more enhanced osseointegration [1]. Several different coatings were reported in the literature, including hydroxyapatite, bioglass, proteins, polysaccharides, drugs, calcium phosphate, and calcium silicate [7–11].

Hydroxyapatite (HA), an osseoconductive material, can induce direct formation of bone tissue around implants, thanks to its unique chemical composition similar to that of the inorganic mineral part of bone tissue [9,12]. Despite its excellent biological properties, there are concerns for using HA-coated Ti-6Al-4V implants. However, the mismatch of their thermal expansion coefficients may cause delamination at coating-substrate interface, compromising the long-term success of osseointegrated implants [13,14].

In addition, CaO–SiO2-based bioglasses have been investigated as coating materials. However, they have demonstrated poor bonding strength, due to the higher thermal expansion coefficient compared to that of Ti-6Al-4V [15,16], and high degradation rate [17–21].

Cp-Ti possess an average CTE (Coefficient of Thermal Expansion) of 8.9 (10−<sup>6</sup> C<sup>−</sup>1), Ti alloys an average CTE of 9.2 (10−<sup>6</sup> C<sup>−</sup>1). Hydroxyapatite shows CTE in the range between 10 and 14 (10−<sup>6</sup> C<sup>−</sup>1), while sphene has a CTE of 6 (10−<sup>6</sup> C<sup>−</sup>1). SiO2-CaO bioactive glasses show CTE ranging between 8.5 and 19 (10−<sup>6</sup> C<sup>−</sup>1).

However, the clinical applications of bioglasses are limited by their low bending strength, high brittleness, low fracture toughness and workability. To overcome the major drawbacks related to the use of bioglass, several silicate-based ceramics were produced and tested as bioactive materials for bone tissue regeneration. Moreover, when compared to CaSiO3 ceramics and tricalcium phosphate TCP (TriCalcium Phosphate), sphene (CaTiSiO5) ceramics drove higher bone-derived cell attachment, spreading, proliferation and differentiation. CaTiSiO5 ceramics have shown significantly higher chemical stability. Also, they have shown better chemical and biological properties compared to HA ceramics.

Sphene (CaTiSiO5) bioactive coatings on titanium substrates were produced using various methods, such as plasma-spraying [14], sol-gel method [22], a hybrid technique of microarc oxidation coupled with heat treatment [23], and airbrush spray coating [24]. These studies have proved their chemical stability, excellent bonding strength (improved compared to HA coatings), bioactivity and cellular biocompatibility. Taken together, these results suggest that sphene ceramics may be potential candidates for bioactive coatings for orthopedic and dental implants. This reduced difference between CTE of cpTi and CaTiSiO5 compared to that of cpTi and HA, is known to be the main property responsible for their improved adhesion [14,22–26].

Recently, our research group deposited sphene coatings from a preceramic polymer precursor "Silicone" onto Ti plates by the airbrush spray technique. Coatings exhibited excellent adhesion to the substrates and homogeneous deposition mode [24,27].

The novelty of this procedure compared to those reported in the literature consists both of the development of sphene synthesis via preceramic polymers route and nano-sized precursors, as well as in the deposition technique. The main advantages of the proposed technology can be summarized as follow:


• The optimized composition of the bioceramic coating together with its improved adhesion to the substrate is the result of the synergic effect of synthesis of sphene via preceramic polymer route and the deposition technique we used.

The primary objective of the current research consists of evaluating the effect of polymer-derived sphene-coated Ti substrates on the osteogenic differentiation of Adipose-Derived Stem Cells (ADSCs) in vitro. The secondary aim consists of assessing sphene coating composition, its in vitro chemical stability, and its micrometer-scale profile- and areal-topography.

#### **2. Materials and Methods**

#### *2.1. Samples Preparation and Coating Deposition*

Sphene (CaTiSiO5) bioactive ceramic coating on Titanium plates (cpTi) was developed by using a preceramic polymer "silicone" approach as previously reported [24,27]. Preceramic silicone "polymethylsiloxane" (commercially available and known as "SILRES® MK", Wacker-Chemie GmbH, München, Germany), isopropanol, nano-sized CaCO3 powder (PlasmaChem, Berlin, Germany, 90 nm) and nano-sized TiO2 powders (Evonik Degussa GmbH, Essen, Germany, 21 nm) were used as starting materials.

Briefly, to prepare the coating suspension, SILRES® MK silicone resin powder was used as SiO2 precursor. The MK silicone was placed in isopropyl alcohol. Active fillers of nanoparticles of CaCO3, and TiO2 were mixed with the MK silicone under magnetic stirring. The total solid content load in the suspension was ~48 vol%. The addition of active fillers was in stoichiometric molar ratio that allows developing the sphene bioceramic as a final ceramic product after sintering at relative lower temperature (i.e., 950 ◦C).

The suspension was homogenized by sonication for 15 min and then transferred into an automatic airbrush (Prona-RA-C2, PronaTools, Toronto, Canada M3J 3A1) for spray coating deposition. During the coating process, the silicone-fillers mixture was kept homogenous by magnetic stirring.

Rectangular plates of 13 × <sup>13</sup> × 3 mm3 size of commercially pure Ti grade 2 (Torresin Titanio s.r.l, Padova, Italy) were used as substrates. The chemical composition of the cpTi plates, as given by the producer (wt.%), was: Fe 0.060, O 0.140, N 0.004, H 0.003, C 0.016, Ti 99.78. Before coating deposition, the cpTi plates were ultrasonically cleaned with acetone, alcohol, and deionized water for ten minutes for each passage, and finally the titanium substrates were dried using compressed air. The use of cpTi instead of Ti alloys as substrate was driven by the choice of restricting the field of study to dental implants.

After many preliminary tests, the processing setup was optimized. The distance between the substrate and nozzle distance was fixed to be 350 mm. The diameter of nozzle of the airbrush was of 1 mm. The air inlet was set at three bars and a deposition time of 1 s was chosen. After deposition process, the plates were left to dry in air at room temperature. Then, the coated samples were heat-treated in static air at 950 ◦C for 1 h (using 5 ◦C/min as heating rate).

#### *2.2. Surface and Coating Characterization*

The morphology of the surface of both uncoated (cpTi) and sphene-coated (Sphene) cpTi plates was investigated by Field Emission Gun Scanning Electron Microscopy (FEG-SEM, Quanta 250 Fei, Eindhoven, The Netherlands).

X-ray diffraction (XRD) patterns of the coatings, after thermal treatment, were collected with an X-ray diffractometer (XRD Bruker D8 Advance, Milano, Italy) operated with Cu-Kα radiation 40 mA and 40 mV. The Rietveld refinement was subsequently applied to quantitatively analyze the XRD data using MAUD (Materials Analysis Using Diffraction) software and COD (Crystallography Open Database).

Surface topography was investigated using a stylus profilometer (Form Talysurf i-Series, Taylor Hobson Ltd., Leicester, UK), a high-resolution system for both profile- and areal-topography measurements. For 2D profile measurements (Ra and Rz), two plates of each group (cpTi and Sphene) were analyzed and a total of 3 scans (evaluation length equal to 5 mm) were extracted from the surface of each sample. For 3D areal measurements, at least 4 squared areas of 0.5 mm × 0.5 mm, extracted from the same surface, in at least two samples per group (cpTi and Sphene) were examined by 3D scanning. Areal parameter Sa and Sz were selected to describe the surface topography. Profile and areal data were filtered to eliminate waviness as well as form components of surface topography by applying a Gaussian filter with a sampling length equal to 0.8 mm, in accordance with ISO 4288 and ISO 25178. Data evaluation was performed with Talymap analysis software (Taylor Hobson Ltd., Leicester, UK).

#### *2.3. Chemical Stability*

The chemical stability of the sphene-coated cpTi substrates was evaluated in Tris-HCl buffer solution. The coated samples were immersed in the buffer solution of tris-(hydroxymethyl)-aminomethane (Tris, (CH2OH)3CNH2) and hydrochloric acid (HCl) with a pH value of 7.4 and were examined at 1, 3 and 7 days. The ratio of plate surface area to Tris-HCl solution volume was 0.1 cm2/mL.

Before and after soakings, dried coated samples were analyzed by means of a stylus profilometer (Form Talysurf i-Series) to evaluate changes in surface roughness. The roughness values of 3 different tracks (5 mm evaluation length) and 4 square areas (0.5 mm × 0.5 mm) per sample were recorded. The following profile roughness parameters, Ra and Rz, and areal parameters, Sa and Sz, were selected.

The variation of Ca, Si and Ti concentrations in the Tris-HCl solution was measured by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES; iCAP 7400 ICP-OES, Thermo Fisher Scientific, Cambridge, UK) and confirmed by Inductively Coupled Plasma Mass Spectrometry (ICP-MS, Agilent Technologies 7700 × ICP-MS system, Agilent Technologies International Japan, Ltd., Tokyo, Japan). The ICP-MS was tuned daily using a tuning solution containing 10 μg L−1 140Ce, 7Li, 205Tl, and 89Y (Agilent Technologies, UK). A 100 μg L−<sup>1</sup> solution of 45Sc and 115In (Aristar, BDH, UK) prepared in HNO3 1.4% was used as internal standard through addition to the sample solution via a T-junction.

All calibrations of Ca, Si and Ti were obtained in Tris-HCl solution using the multi-standard CCS-5 (Inorganic-Ventures) for Si and Ti (100 mg L−1), and multi-standard IMS-120 (Ultra Scientific Multistandard) for Ca (1000 mg L<sup>−</sup>1). All regressions were linear with a determination coefficient (R2) larger than 0.999. Five samples for each time point were taken and measured five times. Data were given as mean ± standard deviation. The dissolution kinetics of Ca, Si and Ti were obtained from the released concentrations.

#### *2.4. Human ADSCs Isolation and Cell Culture*

Human ADSCs (hADSCs) were isolated from the adipose tissue of healthy patients undergoing cosmetic surgery procedures. Tissue collection protocol received a favorable ethical opinion by the Local Bioethical Committee, Padova University and all participants provided written consent. In addition, all experiments performed with human-derived materials were conducted in accordance to the relevant guidelines and regulations. The tissues were digested, and the cells isolated and expanded as described in Gardin et al. [28].

#### *2.5. Seeding of hADSCs*

hADSCs were seeded on both uncoated and sphene-coated samples, in a 12-well plate, with a density of 2 × <sup>10</sup><sup>4</sup> cells/sample. The cells were cultured in DMEM (Dulbecco's Modified Eagle's Medium) High Glucose or Osteogenic Differentiation Medium, at 37 ◦C and 5% CO2, for 21 days. Osteogenic Differentiation Medium was made of DMEM High Glucose supplemented with 10 nM dexamethasone, 10 mM b-glycerophosphate, and 10 ng mL−<sup>1</sup> of basic fibroblast growth factor. Both media were completed with 10% fetal bovine serum and 1% penicillin/streptomycin. After 21 days of culture, Methyl Thiazolyl-Tetrazolium (MTT) test, Scanning Electron Microscopy (SEM)

analysis, immunofluorescent staining with phalloidin, Alizarin Red S staining and quantification, and real-time PCR were carried out.

#### *2.6. MTT Assay*

To assess the proliferation rate of hADSCs grown on both uncoated and coated samples, a colorimetric mitochondrial viability assay was performed as described by Denizot and Lang [29] with minor modifications. After removing the culture medium, the samples were incubated in 1 mL of 0.5 mg/mL MTT solution in phosphate buffered saline (PBS) for 3 h at 37 ◦C. The MTT solution was then removed, and each sample was extracted with 0.5 mL of 10% dimethyl sulfoxide in isopropanol for 30 min at 37 ◦C. For each sample, Optical Density (O.D.) values, at 570 nm, were recorded in duplicate on 200 μL aliquots using a multilabel plate reader (Victor 3, Perkin Elmer, Milan, Italy).

#### *2.7. SEM Analysis*

For SEM imaging, hADSCs grown on both uncoated and sphene-coated samples for 21 days were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer for 1 h, then progressively dehydrated in ethanol. Cell spreading and morphology were evaluated using FEG-SEM (Quanta 250 Fei, Eindhoven, The Netherlands).

#### *2.8. Immunofluorescence*

Cell adhesion to the scaffolds surface was evaluated by immunofluorescent staining with phalloidin. Briefly, cells were fixed in 4% paraformaldehyde in PBS for 10 min, then permeabilized with 0.1% triton X-100 in PBS for 30 min at room temperature. 5 mg/mL phalloidin were then used for fluorescent staining of actin filaments, whereas nuclear staining was performed with 2 μg mL−<sup>1</sup> Hoechst H33342 solution for 5 min. Image acquisition was obtained with an inverted optical microscope DMI4000 B (Leica Microsystems, Wetzlar, Germany).

#### *2.9. Alizarin Red S Staining and Quantification*

The formation of extracellular mineral deposits onto uncoated and sphene-coated samples was detected by Alizarin Red S staining and quantification. Cells were fixed in 4% paraformaldehyde in PBS for 10 min at room temperature. Then, cells were stained adding 0.5 mL of 40 mM freshly prepared Alizarin Red S solution (pH 4.2) for 20 min at room temperature. After washing with double-distilled water, the Alizarin Red S stained areas were extracted with 0.5 mL of 10% cetylpyridinium chloride solution for 1 h at room temperature under gentle agitation. For each sample, O.D. values at 570 nm were recorded in duplicate on 200 μL aliquots using a Victor 3 plate reader.

#### *2.10. Real-Time PCR*

Total RNA was extracted from hADSCs, cultured on both uncoated and sphene-coated samples for 21 days, with a Total RNA Purification Plus Kit (Norgen Biotek Corporation, Thorold, ON, Canada). For the first-strand cDNA synthesis, 1 μg of total RNA of each sample was reverse transcribed with a SensiFAST™ cDNA Synthesis Kit (Bioline, London, UK), following the manufacturer's protocol. Human primers were selected for each target gene using a Primer 3 software. Real-time PCRs were run using the chosen primers at a concentration of 400 nM and SensiFAST™ SYBR No-ROX Kit (Bioline) on a Rotor-Gene 3000 (Corbett Research, Sydney, Australia). The thermal cycling conditions were as follow: 2 min denaturation at 95 ◦C; 40 cycles of 5 s denaturation at 95 ◦C; annealing for 10 s at 60 ◦C; and 20 s elongation at 72 ◦C. Differences in gene expression were calculated by normalizing to the expression of the Transferrin Receptor (TFRC) housekeeping gene.

#### *2.11. Statistical Analysis*

One-way analysis of variance (ANOVA) was used for data analysis. *t*-tests were used to determine data significance (*p* < 0.05). All tests were performed using SPSS 16.0 software (SPSS Inc., Chicago, IL, USA) (licensed to the University of Padua, Padova, Italy).

#### **3. Results**

#### *3.1. Surface Characterization*

FEG-SEM images of uncoated cpTi and sphene-coated surfaces are shown in Figure 1a,b, respectively. Sphene coating is characterized by a homogenous grey substrate, composed of tetragonal crystals <1 μm, and "white islands" growing on it. These white agglomerates appear to be composed of vertically aggregated spherical particles. As reported in previous work [27], EDS (Energy Dispersive X-ray Spectrometry) analysis identified Ti, O, Si and Ca in areas corresponding to these white spherical structures. Instead, in the grey phase Ti and O were mainly present.

**Figure 1.** Surface morphology (SEM-FEG) of (**a**) as-received uncoated cpTi and (**b**) sphene-coated cpTi.

The coating composition was further investigated by XRD analysis and Rietveld refinement (Figure 2) that showed the presence of rutile (TiO2, 60 wt.%) followed by sphene (CaTiSiO5, 31 wt.%) and perovskite (CaTiO3, 9 wt.%). The as-received TiO2 powders, used as a filler to develop the sphene ceramic, are characterized by a weight percentage ratio of 80 to 20 between anatase and rutile, as claimed in the manufacturer's datasheet. Anatase and rutile are the two main polymorphs of titanium oxide. They both exhibit tetragonal crystalline structure but obey different space groups. Anatase is a metastable phase, and the transformation to the stable rutile structure occurs as irreversible phase transformation in the range between 600 ◦C and 700 ◦C [30]. The treatment at 950 ◦C, performed in air, may have induced the nucleation and growth of rutile crystals before sphene synthesis, to yield a final weight percentage ratio of 60 wt. % rutile. The higher stability of rutile compared to anatase may have inhibited the reaction to produce sphene at this temperature. It has indeed been reported in the literature that the formation of sphene by the sol-gel method is complete at 1300 ◦C [31]. As previously reported [24], the choice of keeping the temperature at 950 ◦C was driven by the need for preserving the structure of the cpTi plate without affecting the bonding of the coating to the substrate. Further studies are ongoing aimed at increasing the amount of sphene produced by the preceramic polymer precursor route.

**Figure 2.** XRD pattern of sphene-coated cpTi after heat treatment at 950 ◦C in air for 1 h.

Surface roughness was investigated by means of a profilometer. 2D profile measurements (Ra and Rz) and 3D areal measurements (Sa and Sz) of uncoated (cpTi) and sphene-coated (Sphene) cpTi substrates are reported in Table 1 and Figure 3.

A clear difference in the values of Ra and Rz between the cpTi substrate and the sphene-coated substrate can be observed in the 2D measurements. However, this difference is smoothed in the 3D areal measurements. Indeed, the surface of the cpTi substrate is characterized by a flatter morphology (Figure 3a) than the sphene-coated sample (Figure 3b). The sphene coating presents a high number of peaks homogeneously distributed, while the cpTi substrate presents a limited number of peaks but characterized by a wider area. The yellow and red peaks in Figure 3b well correspond to the edges of the white agglomerates detected by SEM observations (Figure 1b).

#### *3.2. Chemical Stability*

The 2D roughness data after Tris-HCl immersion test show a slight increase after 1 day of soaking time. However, this increase in roughness values shows an almost constant trend after 3 and 7 days. In accordance to these data, 3D maps of sphene after immersion in Tris-HCl solution (Figure 3c,d) demonstrated a reduction in the numbers of peaks after the dissolution test, with a similar pattern of peak and valley distributions at 1, 3 and 7 days after soaking. This is confirmed by Sa and Sz values that were reported in Table 1.


**Figure 3.** 3D maps of: (**a**) uncoated cpTi; (**b**) sphene-coated cpTi; (**c**) sphene-coated cpTi after 1 day in Tris-HCl buffer solution; (**d**) sphene-coated cpTi after 7 days in Tris-HCl buffer solution.

Figure 4 shows ICP analysis of Ca, Si and Ti ion release of the sphene coating in Tris-HCl solution at different time points. The average concentrations of Ca and Si ions, after 1 day of soaking, are 0.069 mM and 0.045 mM, respectively. Similar concentration levels were found after 3 and 7 days of immersion in Tris-HCl solution. Only trace amount of Ti ions released from the coating was detected (0.001–0.002 mM).

These findings are consistent with the above mentioned areal-topography measurements and suggest that the dissolution occurs only in the first hours after soaking.

**Figure 4.** Ca, Ti and Si ion release of sphene-coated samples, after 1, 3, and 7 days of soaking time.

#### *3.3. Cell Proliferation*

The MTT assay proved that hADSCs were able to proliferate on both sphene-coated and uncoated samples (Figure 5). The O.D. values recorded for cells loaded on coated samples were found to be significantly (*p* < 0.05) higher to those observed for the controls in both the experimental conditions. Moreover, from MTT results, it was shown that hADSCs' proliferation was higher when hADSCs were cultured in Osteogenic Differentiation Medium compared to in DMEM High Glucose, in particular for the sphene-coated samples.

**Figure 5.** MTT assay of hADSCs cultured for 21 days on uncoated cpTi samples (*cpTi*) and on sphene-coated (*Sphene*) samples in DMEM High Glucose or Osteogenic Differentiation Medium. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

#### *3.4. Cell Adhesion and Morphology*

The SEM analyses showed how hADSCs anchored to the surface of the specimens (Figure 6a–d). The cells were extremely flat with the typical star morphology associated with the osteoblastic-like

phenotype and their distribution was similar, independent of the culture medium used (Figure 6a,c). After 21 days of culture onto the sphene-coated surfaces, cells showed short and thin filopodia when grown in DMEM High Glucose (Figure 6b); whereas they appeared flat and overlapped when cultured in the presence of the Osteogenic Differentiation Medium (Figure 6d). In both cases, the sphene coating remained intact during the culturing time, thus demonstrating its chemical stability.

**Figure 6.** SEM images (1000× magnification) of hADSCs grown for 21 days on (**a**) uncoated cpTi in DMEM High Glucose; (**b**) sphene-coated cpTi in DMEM High Glucose; (**c**) uncoated cpTi in Osteogenic Differentiation Medium; (**d**) sphene-coated cpTi in Osteogenic Differentiation Medium.

#### *3.5. Cytoskeletal Organization*

Staining with fluorescent phalloidin showed that cells were able to attach and spread on both the cpTi (Figure 7a,c) and sphene-coated samples (Figure 7b,d). After 21 days of culture, hADSCs had completely colonized the surfaces and formed a continuous layer with no significant differences between the experimental conditions. In vivo, such as in periosteum, osteoblasts are aligned with the collagen they produce and also on system cells are aligned to the collagen substrate. Cytoskeletal stresses and tension increase with increasing ECM (ExtraCellular Matrix) stiffness. The cytoskeleton provides a structural frame for the cell through Focal adhesions. Focal adhesion forms when adapter proteins link the cytoskeleton to integrins, which permits cells to adhere to the substrate. A variety of signaling proteins are also associated with focal adhesions, including focal adhesion kinase (FAK), an important mediator of signaling at these centers. Forces are also transmitted to the substrate

at these sites. Osteogenic differentiation can also be affected by the substrate properties where cells are seeded. On our samples, we can confirm that the shape of a cell follows the steps that occur in vivo.

**Figure 7.** Immunofluorescent staining of the actin filaments with phalloidin (in red). Cell nuclei are counterstained with Hoechst (in blue). hADSCs seeded for 21 days on: (**a**) uncoated cpTi in DMEM High Glucose; (**b**) sphene-coated cpTi in DMEM High Glucose; (**c**) uncoated cpTi in Osteogenic Differentiation Medium; (**d**) sphene-coated cpTi in Osteogenic Differentiation Medium.

#### *3.6. In Vitro hADSC Osteogenic Differentiation*

The in vitro osteogenic differentiation of hADSCs was first evaluated by marking of extracellular calcium deposition with Alizarin Red S staining and quantification at 21 days of culture (Figure 8). A difference in the presence of calcium deposits on the sphene-coated surfaces and the cpTi controls was evidenced for hADSCs cultured in the Osteogenic Differentiation Medium. Calcium deposits were significantly (*p* < 0.05) higher in the first experimental condition. The accumulation of calcium deposits by hADSCs was not observed when cells were seeded on the sphene-coated or uncoated samples in DMEM High Glucose.

**Figure 8.** Alizarin Red S quantification of calcium deposits produced by hADSCs seeded onto *cpTi* or *Sphene* surfaces in DMEM High Glucose or Osteogenic Differentiation Medium for 21 days. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

#### *3.7. Real-Time PCR*

Real-time PCR was used to quantify the relative expression level of several osteogenic differentiation markers. In particular, expression of runt-related transcription factor 2 (RUNX2), osterix (OSX), collagen type I (COL1A1), osteocalcin (OC), osteonectin (ON), osteopontin (OPN), alkaline phosphatase (ALPL), and receptor activator of nuclear factor kappa-B ligand (RANKL) has been investigated. Runx2 and Osterix are the main transcription factors involved on osteoblastic commitment. Collagen is the principal component of the bone matrix. Alkaline Phosphatase is involved on the calcification of the Extracellular matrix.

Osteonectin is a protein which role is related to the connection between extracellular matrix component and osteoblast. It is mostly produced during the first phases of bone regeneration or during bone remodeling [32–34].

As evident in Figure 9a,c, the relative expression of RUNX2, OSX, OC, OPN, and RANKL did not change significantly between the uncoated and sphene-coated surfaces, both when cells were grown in DMEM High Glucose or in Osteogenic Differentiation Medium. Similarly, there was no noticeable difference in ALPL (Figure 9a), COL1A1 and ON (Figure 9b) mRNA expression between cpTi and sphene-coated samples in DMEM High Glucose; instead, a slight down-regulation of ALPL (Figure 9c) and COL1A1 (Figure 9d) expression was identified in cells seeded onto the sphene surfaces in the presence of Osteogenic Differentiation Medium.

**Figure 9.** Real-time PCR analysis of the osteogenic differentiation markers runt-related transcription factor 2 (RUNX2), osterix (OSX), collagen type I (COL1A1), osteocalcin (OC), osteonectin (ON), osteopontin (OPN), alkaline phosphatase (ALPL), and receptor activator of nuclear factor kappa-B ligand (RANKL). Data are normalized to the expression of the transferrin receptor (TFRC) internal reference. Gray bars indicate the relative expression level of the selected genes in the hADSCs seeded onto *cpTi* samples in DMEM High Glucose (**a**,**b**) or Osteogenic Differentiation Medium (**c**,**d**) for 21 days. Black bars represent the gene expression level of the same markers in the hADSCs seeded onto *Sphene* samples in DMEM High Glucose (**a**,**b**) or Osteogenic Differentiation Medium (**c**,**d**) for 21 days.

#### **4. Discussion**

In the present study, the sphene-based coatings on cpTi substrates showed a rough morphology, with roughness Ra in the range of 3.25 to 4.75 μm. The morphology was characterized by small crystals made of TiO2 and CaTiO3 (valley of Figure 4) and islands of unregular shape that were detected by EDS analysis to be composed of sphene [27]. The quantitative XRD analysis performed on the coated substrate showed the reaction efficiency and the amount of sphene that was about 31 wt. %.

Micro-roughened surfaces were demonstrated to have a positive effect on early blood cell/implant interactions [35] and osteoblast proliferation [36]. In addition, surface topography was found to influence osteoconduction in both animal studies [37,38] and human retrieval study [39], with good results shown by rough implant surfaces. Sphene coatings examined in the present study possessed a moderately rough surface. The surface roughness quantified by 3D areal measurements showed a good correspondence with 2D measurements in both cpTi and coated cpTi.

Even though the release of Ca and Si ions has been demonstrated to favor osteoblastic proliferation and differentiation [18,19,40], one of the major drawbacks of biosilicate ceramic coatings, such as CaSiO3 coatings, consists of their poor chemical stability and high degradation rate that might compromise their long-term stability [41,42]. Therefore, a limited release of these ions would enhance coating bioactivity and, on the other hand, would not interfere with the integrity of the coating itself [43].

To assess the chemical stability of the coatings, Tris-HCl was chosen due to the absence of Ca, Si or Ti ions in its composition. The results of dissolution tests in Tris-HCl showed that the dissolution happened only during the first hours of soaking and then remained constant after 1, 3 and 7 days. This behavior agrees with the results reported by Wu et al. [22], thus proving that the developed coating is comparable in terms of chemical properties to standard HA. At the same experimental conditions of the present study, a lower released Ca concentration (2.74 ppm) after 7 days of soaking in Tris-HCl was found in the sphene-coated samples here investigated as compared to the values of around 20 ppm and >110 ppm of plasma-sprayed sphene ceramic coatings and HA coatings, respectively [14]. Moreover, as found by Wang et al. [43], a higher concentration of Ca ions released to the solution was measured compared with the concentration of Si ions released after 7 days of immersion. In addition, only an irrelevant amount of Ti ions was detected by Wang et al. [43] in the Tris-HCl after one week of soaking.

The roughness of samples, after dissolution tests slightly increased compared to the starting samples. This is mainly observable in the areal measurements where a sharp increase of Sz was measured from unetched sphene to 1-day soaking time in Tris-HCl. In agreement with ICP analysis, no significant trends on roughness at increasing soaking time can be observed. Thus, this proves that little dissolution of Ca, Ti and Si, happened during the first 24 h.

The chemical composition of implant surfaces profoundly influences cell adhesion, spreading, proliferation, and differentiation [44,45]. Sphene-based substrates better supported hADSCs attachment and proliferation, compared with uncoated cpTi samples, as revealed by the MTT assay. By SEM analysis, a uniform layer of well-spread cells could be observed when cells were cultured onto the sphene-coated surfaces in Osteogenic Differentiation Medium. The results of Alizarin Red S quantification suggested that sphene coating favorably affects the accumulation of mineralized calcium phosphate when cells are cultured in the presence of osteogenic factors. Moreover, cells seeded onto sphene-coated surfaces expressed similar levels of osteoblasts-related genes to those expressed by cells cultured on uncoated cpTi in both culture conditions. We can hypothesize that the combination of the coating and the Osteogenic Differentiation Medium promoted a higher calcium deposition on sphene-coated surfaces than on uncoated cpTi substrates. These findings agree with Wang et al. [43], who demonstrated that sphene glass-ceramic coatings were able to support human osteoblast-like cells (HOBs) attachment, proliferation and differentiation and enhance the expression level of bone-related genes. Furthermore, it was demonstrated that sphene ceramic coatings significantly enhanced HOBs proliferation and ALP activity compared with plasma-sprayed HA coating and uncoated Ti-6Al-4V substrates [14]. Pure dense sphene ceramic disks were found to promote human bone-derived cells attachment and to significantly improve their proliferation and differentiation, as compared with α-CaSiO3 ceramic disks [25]. Pure sphene ceramic disks were not only found to modulate HOBs, but also osteoclasts and endothelial cells [26]. As compared with the study of Ramaswamy et al. [26], in the current work, mesenchymal stem cells (MSCs) were preferred to human osteoblast-like cells, human osteoclasts and human microvascular endothelial to in vitro investigate the osteoconductive properties of the biomaterial, since stem cells are the main actors enrolled during the early stages of tissue regeneration. MSCs are stem cells involved on tissue regeneration, which can migrate to the damage site thanks to their receptor and to coordinate all the phase of tissue regeneration [46]. Hypoxia, which usually occurs at the injured sites due to a disruption in blood supply, has been reported to largely contribute to MSC mobilization and homing. The hypoxia-inducible factors (HIFs) are key regulators of the cellular response to decrease in the local oxygen level, also regulating the expression of genes involved in MSC recruitment and migration to the damaged tissues [32–34,46–48].

In a good agreement with previously published work [14,25,26,43], our results suggest that the ionic environment, determined by the dissolution of Ca and Si ions from the sphene biomaterial, may contribute to stimulate cell proliferation and differentiation. In addition, the in vivo assessment revealed enhanced osseointegration of both sphene-coated and HA-coated Ti-6Al-4V implants, compared to uncoated implants, confirming the in vitro observations and the potential use of sphene coatings in clinical application [26].

#### **5. Conclusions**

In conclusion, sphene-based crack-free coatings were successfully produced starting from a preceramic polymer and nano-sized fillers and deposited using an automatic airbrush. The coatings presented good chemical stability, as a minimal dissolution occurred only during the first hours of soaking in Tris-HCl solution and then remained stable over the period considered. The coatings also supported hADSCs attachment, proliferation, and differentiation in vitro. Taken together, our findings indicate a potential for use of sphene ceramics as coating materials for orthopedic and dental implants. In vivo studies should be performed to confirm in vitro observations. Further studies are ongoing to increase the reaction efficiency.

**Author Contributions:** Conceptualization, H.E., B.Z. and L.B.; Data curation, H.E., G.B., C.G., L.F., D.B. and L.B.; Formal analysis, H.E., G.B., C.G. and L.F.; Supervision, S.S., B.Z. and L.B.; Validation, L.B.; Writing—original draft, H.E. and G.B.; Writing—review and editing, H.E., G.B., C.G., L.F., D.B., P.P., S.S., B.Z. and L.B.

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

**Acknowledgments:** The authors would like to thank Prof. Roberto Meneghello of Department of Management and Engineering for his contribution in surface characterization.

**Conflicts of Interest:** The authors declare no competing interests.

#### **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* **Plasma-Induced Crystallization of TiO2 Nanotubes**

#### **Metka Benˇcina 1, Ita Junkar 1,\*, Rok Zaplotnik 1, Matjaz Valant 2,3, Aleš Igliˇc <sup>4</sup> and Miran Mozetiˇc <sup>1</sup>**


Received: 21 January 2019; Accepted: 18 February 2019; Published: 20 February 2019

**Abstract:** Facile crystallization of titanium oxide (TiO2) nanotubes (NTs), synthesized by electrochemical anodization, with low pressure non-thermal oxygen plasma is reported. The influence of plasma processing conditions on TiO2 NTs crystal structure and morphology was examined by X-ray diffraction (XRD) and scanning electron microscopy (SEM). For the first time we report the transition of amorphous TiO2 NTs to anatase and rutile crystal structures upon treatment with highly reactive oxygen plasma. This crystallization process has a strong advantage over the conventional heat treatments as it enables rapid crystallization of the surface. Thus the crystalline structure of NTs is obtained in a few seconds of treatment and it does not disrupt the NTs' morphology. Such a crystallization approach is especially suitable for medical applications in which stable crystallized nanotubular morphology is desired. The last part of the study thus deals with in vitro biological response of whole blood to the TiO2 NTs. The results indicate that application of such surfaces for blood connecting devices is prospective, as practically no platelet adhesion or activation on crystallized TiO2 NTs surfaces was observed.

**Keywords:** TiO2 nanotubes; crystallization; gaseous plasma; biological response

#### **1. Introduction**

Self-aligned TiO2 nanotubes synthesized by electrochemical anodization of Ti foil are highly promising materials for biomedical applications due to their advanced biocompatibility in comparison to commonly used plain metal surfaces, and as such can be employed as orthopaedic and dental implants, vascular stents, antibacterial devices or surfaces and smart drug-delivery platforms [1–8]. Improved biological response to the TiO2 NTs is, among others, the result of increased surface roughness and larger surface area. Moreover, the tunable morphology of NTs that can be controlled by altering the anodization parameters such as voltage and time, allows for selective response of biological material. For instance, proteins, platelets and cell adhesion, cell morphology, proliferation and differentiation are highly affected by the TiO2 NTs diameter [9]. This is especially important in applications, in which the adhesion of certain cell types is preferred, as in the case of vascular stents.

As-anodized amorphous TiO2 NTs are, therefore, often subjected to crystallization process in order to enhance their bio-performance. It has been shown that annealed TiO2 NTs layers with anatase crystal structure have a much better corrosion resistance and bioactivity than amorphous NTs; moreover, they tend to induce more, and faster, hydroxyapatite deposition, which is crucial for successful bone bonding

(osteointegration) ability of the body implants [10–12]. In addition, the proliferation and mineralization of osteoblasts cultured on anatase or a mixture of anatase⁄rutile TiO2 NTs was significantly higher than on the surface of as-anodized amorphous TiO2 NTs [13]. Crystallized TiO2 NTs have also good bacterial resistance; it was found that such layers inhibit the growth of *S. aureus* and *P. aeruginosa* [5], which is important for designing medical devices, since implant-associated infections present a serious health care concern. In addition, the enhanced biocompatibility of crystalline TiO2 NTs could also be linked with increasing Ti-OH functional groups, responsible for enhancing wettability and apatite deposition [14,15] and increasing stability in bioliquids [16].

Crystallization of TiO2 NTs induced by annealing in a conventional furnace is a well-established process; the atom rearrangements occur due to elevated temperatures. However, such crystallization requires high temperatures and is time consuming in comparison with the plasma-induced crystallization presented in this contribution. For instance, in order to achieve anatase, mixture of anatase/rutile and rutile crystal structure, TiO2 NTs should undergo annealing in the furnace for at least 2 h at 450 ◦C, 550 ◦C [10] and 800 ◦C [17], respectively. Besides, changes in the morphology of TiO2 NTs present a significant drawback of such crystallization process; the annealing in a furnace at 800 ◦C results in the transition of anatase NTs to a dense rutile layer [17]. By contrast, hydrothermal treatment presents an intriguing low-temperature crystallization method that is catalysed by pressure and mineralizing agents [18]. Liu et al. [19] reported hydrothermal reaction in an autoclave at a temperature of less than 180 ◦C for 4 h, which led to the amorphous-to-ananase transition of TiO2 NTs. In spite of low temperature requirements, hydrothermal crystallization adversely alters TiO2 NTs morphology and leads to the formation of aggregated anatase nano-particles on the surface of NTs at about 200 ◦C [20]. Lamberti et al. [14] demonstrated another attractive near-room temperature (50 ◦C) crystallization process of amorphous TiO2 NTs, which led to anatase crystal phase formation after only 30 min of sample exposure to water vapour [14,20]. However, such a crystallization method also initiates the formation of crystals at the outer and inner walls of the NTs, which finally leads to transformation of nanotubes to nanorods. For this reason, there is an indispensable need to develop novel crystallization processes that would preserve the morphology of the materials.

In the present research, the transformation of rather amorphous TiO2 NTs to anatase or mixture of anatase/rutile crystal structure by a non-thermal oxygen plasma process, is, according to our best knowledge, reported for the first time. Fast amorphous-to-anatase or -anatase/rutile transition and unaltered NTs morphology are the major advantages of such crystallization before conventional processes, for which the preservation of morphology is still a challenge. Although such a time-saving method is convenient for inducing crystallinity of the nanomaterials/thin films [21–23], there is a lack of information about the mechanisms involved in the crystallization process. In this contribution, the influence of plasma processing conditions on the crystal structure and morphology of the TiO2 NTs is therefore examined. In addition, the biological aspects of plasma treated TiO2 NTs are presented. It has already been shown that surfaces treated by highly reactive oxygen plasma improve biological response. For instance, Junkar et al. [24] showed that plasma treatment allows for TiO2 NTs surface functionalization and enhancement of osteoblast-like cell responses. Authors also confirmed cleaning and sterilizing effects of plasma treatment [25]. Moreover, the wettability of ZrO2 was noticeably improved by an oxygen plasma treatment, which promoted the attachment, proliferation and differentiation of human osteoblast-like cells [26]. Therefore, simultaneous plasma treatment and crystallization of the material with already tunable morphology offers highly needed improvements for devices used in biomedical applications.

#### **2. Materials and Methods**

#### *2.1. Materials*

Titanium foil (Advent, 0.1 mm thickness, 99.6%), ethylene glycol (Fluka, ≥99.5%), ammonium fluoride—NH4F (Sigma Aldrich, 28.0–30.0%), hydrofluoric acid—HF (Sigma Aldrich, ≥40%) acetone (Honeywell Riedel–de Haen, 99.5%), ethanol (Sigma Aldrich, 96%), phosphate-buffered saline—PBS (Sigma Aldrich), and glutaraldehyde solution (Sigma Aldrich, 25% in H2O) deionized water (miliQ).

#### *2.2. Synthesis of TiO2 NTs by Electrochemical Anodization*

The TiO2 NTs were fabricated by an electrochemical anodization method as described in Refs. [24,25]. The synthesis was carried out at room temperature (≈20 ◦C) in a two electrode system (Pt/Ti) with the size 10 × 10 mm<sup>2</sup> and working distance of 15 mm. The thickness of the Ti electrode was 0.10 mm. Before anodization, the Ti foil was ultrasonically cleaned in acetone, ethanol and deionized water for 5 min in each and further dried under a nitrogen stream. The electrolyte used in this step was composed of ethylene glycol and NH4F (0.35 wt.%) and H2O (1.7 wt.%). The nanotubular layer grown in this step was then detached from the substrate with a successive ultrasonication in H2O, acetone and ethanol in order to obtain a pre-dimpled surface, which led to enhanced homogeneity of the NTs' surface. In the second step of anodization process, the pre-treated Ti foils were used as a substrate to grow NTs. The electrolyte based on the ethylene glycol, containing water and HF was used and the detailed synthesis parameters of these steps are presented in Table 1. The as-synthesized NTs were kept in ethanol for 2 h in order to remove components from the electrolyte and then dried under a nitrogen stream. The as-prepared NTs were used for further plasma processing (output power 200–800 W for different times) or annealed at 450 ◦C for 2 h in a furnace with annealing/cooling rate 8 ◦C/min.

**Table 1.** Influence of synthesis conditions on the diameter and length of the TiO2 nanotubes (NTs).


#### *2.3. Oxygen Plasma Treatment*

Treatment of TiO2 NTs was performed by oxygen plasma in the plasma reactor designed in house. The system was evacuated with a two-stage oil rotary pump with a nominal pumping speed of 80 m3 h−1. The discharge chamber was a Pyrex tube with a length of 80 cm and an inner diameter of 3.6 cm. Gaseous plasma was created with an inductively coupled radiofrequency (RF) generator (CESAR 1310, Advanced Energy, Fort Collins, Colorado, USA), operating at a frequency of 13.56 MHz and a nominal power of 1000 W. Generator powers between 200–800W were used. Commercially available oxygen was leaked into the discharge chamber and the pressure was measured with an absolute vacuum gauge. The pressure during the plasma treatment was fixed at 50 Pa which allows for the highest degree of dissociation of oxygen molecules for this particular plasma reactor. The samples were placed on an object glass and treated for different periods of time and with different output powers, one at a time. Details about the behaviour of reactive gaseous species versus nominal power of the RF generator are presented elsewhere [27].

#### *2.4. Characterization*

#### 2.4.1. X-ray Diffraction (XRD) Spectroscopy Analysis

X-ray diffraction (XRD) was performed using MiniFlex 600 Benchtop X-ray diffractometer (Rigaku, Tokyo, Japan) equipped with Cu K-α radiation (1.541 Å) over the 2θ range 10–70◦, with a step size of 0.017◦, divergence slit of 0.218◦ and counting step time of 25 s in continuous scanning mode. Carbon tape was used to mount the samples on the glass holder.

#### 2.4.2. Scanning Electron Microscope (SEM) Analysis

Morphology of the materials was analysed with a JSM 7100F scanning electron microscope (SEM, JEOL Ltd., Tokyo, Japan). For biological evaluation of platelets on the surface, the samples were coated

with gold/palladium and examined by SEM at an accelerating voltage of 15 kV. The test was done in triplicates and only representative images are shown in this paper.

#### 2.4.3. Temperature Measurements

During the exposure of the samples to plasma, the temperature was measured with a custom-made K-type thermocouple). Chromel and alumel wires (Goodfellow Cambridge Ltd., Huntingdon, UK) were spot welded on the rear of the samples so they were not influenced by plasma. A Keithley Model 2100 digital multimeter (Keithley Instruments/Tektronix, OH, USA) and custom-made software were used to record the time evolution of the samples temperature.

#### *2.5. In Vitro Biological Response—Interaction With Whole Blood*

All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of Slovenia (56/03/10). Whole blood was obtained from healthy volunteers via vein puncture. The blood was drawn into 9 ml tubes already coated with trisodium citrate anticoagulant. The material samples—Ti foil, amorphous and annealed (anatase) TiO2 NTs of 100 nm in diameter (7 × 7 mm2)—were incubated with the 250 <sup>μ</sup>L of whole blood for 45 min at room temperature in the 24-well cell culture plates. Afterwards, 250 μL of PBS was added to the incubated samples. The blood with PBS was then removed and the samples were rinsed 3 times with 250 μL of PBS in order to remove weakly adherent platelets and other biological material. The adherent platelets were further fixed by 400 μL of 0.5% glutaraldehyde solution for 2 h at room temperature. Then the materials were rinsed with PBS and dehydrated by using a graded ethanol series (50 vol.%, 70 vol.%, 80 vol.%, 90 vol.%, 100 vol.% and again 100 vol.% of ethanol) for 5 min and in the last stage (100 vol.% ethanol) for 10 min. Afterwards the samples were dried with liquid nitrogen and stored in vacuum.

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

#### *3.1. Crystal Structure Analysis*

The TiO2 NTs were subjected either to annealing or highly reactive oxygen plasma treatment at different input powers and at different treatment times and the influence of such treatments was further studied in terms of altered surface morphology and crystallization. It has been shown in our previous work [25] and recently also in Ref. [28] that TiO2 NTs were deformed after exposure to elevated temperatures and that NTs' stability depended on diameter (larger diameter NTs are more resistant to thermal degradation). In the present study, the alteration in crystallinity of TiO2 NTs was studied by X-ray diffraction spectroscopy, which revealed that the treatment with highly reactive oxygen plasma at different powers significantly influences the TiO2 NTs crystal structure. The as-anodized TiO2 NTs samples are amorphous and remain uncrystallized even after 3 min of exposure to highly reactive oxygen plasma at 200 W (Figure 1). The XRD data show that the Ti foil, which is used as a substrate for TiO2 NTs fabrication, is detected by XRD although NTs of 100 nm in diameter have a length of about 2.5 μm. The higher power of plasma, more than 400 W, seems to already induce crystallization if samples are exposed to plasma for more than 10 s. Actually a mixture of anatase and rutile crystal phase (hardy observable peak at 2θ = 36.5◦ as seen in Figure 1) appears after the treatment of TiO2 NTs for 10 s in plasma with the power of 400 W and 600 W. Interestingly, the 1 s plasma exposure at these conditions is not enough to initiate the change of crystal structure, since no anatase nor rutile peaks were detected. However, the plasma treatment at 800 W for 1 s and 10 s results in anatase and a mixture of anatase/rutile crystal phases, respectively (Figures 1 and 2). For comparison, the annealing of TiO2 NTs in a conventional furnace at 450 ◦C requires 2 h to induce the transition of amorphous phase to anatase crystal structure.

**Figure 1.** X-ray diffraction (XRD) patterns of TiO2 NTs of 100 nm diameter measured after plasma treatment at different powers; 200, 400, 600 and 800 W for different times and TiO2 NTs of 100 nm diameter after annealing in a furnace at 450 ◦C for 2 h.

**Figure 2.** Magnified X-ray diffraction (XRD) patterns of TiO2 NTs of 100 nm diameter measured after plasma treatment at different powers; 200 and 800 W for different times, in order to show the difference between amorphous, anatase and rutile crystal phase of TiO2 NTs. A = characteristic XRD peaks for anatase crystal structure, R = characteristic XRD peaks for rutile crystal structure and Ti = characteristic XRD peaks for Ti foil.

In order to study the effect of synthesized TiO2 layer thickness on the crystallization in plasma, TiO2 NTs of 15 nm in diameter and length of about 0.2 μm were treated with oxygen plasma at 800 W for 1 s and 10 s. Only one anatase-phase peak at 2θ = 25◦ (101) was detected for the TiO2 NTs sample treated with plasma at 800 W for 1 s, indicating that the sample is still mainly amorphous (Figure 3). Interestingly, for TiO2 NTs of 100 nm in diameter the anatase crystal phase is already observed after 1 s of treatment at 800 W (Figures 1 and 2). However, the rutile crystal phase without any evidence of anatase prevails when plasma treatment at 800 W is prolonged to 10 s for TiO2 NTs of 15 nm in diameter (Figure 3). For comparison, TiO2 NTs of 100 nm in diameter crystallize in mixture of anatase/rutile phase after prolonged plasma treatment (10 s). The XRD spectra of plain Ti foil, which is used as a substrate for TiO2 NTs fabrication, are presented in Figure 3; it can be observed that XRD pattern of the plasma-treated sample at 800 W for 1 s exhibit small rutile and anatase peaks at 2θ = 36.5◦ and

37.5◦, respectively. However, an intense rutile-phase peak at 2θ = 27◦ is confirmed after the exposure of the Ti foil to plasma of power 800 W for 10 s. These results show that although smaller-sized nanoparticles require less intensive plasma conditions to induce crystallization [21,29], the thermal effects on TiO2 NTs with 2.5 μm length and 0.21 μm length may be different. It is also noteworthy that plasma treatment induce the formation of crystalline oxide layer on the surface of the Ti foil.

**Figure 3.** X-ray diffraction (XRD) patterns of untreated Ti foil and Ti foil and TiO2 NTs of diameter 15 nm (NT15) measured after plasma treatment at 800 W for 1 s and 10 s. A = characteristic XRD peaks for anatase crystal structure, R = characteristic XRD peaks for rutile crystal structure and Ti = characteristic XRD peaks for Ti foil.

#### *3.2. Morphology Analysis*

The morphology of as-anodized TiO2 NTs (100 nm in diameter) is shown in Figure 4, while the morphology of TiO2 NTs exposed to oxygen plasma is shown in Figure 5. It is noteworthy that TiO2 NTs formed on the Ti foil are stable at plasma conditions used in the present study, except the NTs treated with plasma of power 800 W for 10 s. SEM analysis of TiO2 NTs treated at 800 W for 10 s showed that the NTs structure is destroyed as can be seen from Figure 5, where the comparisons between plasma-treated samples at 800 W for 10 s, 800 W for 1 s and 400 W for 10 s are presented. The tops of TiO2 NTs treated at 800 W for 10 s are partially closed, most probably due to collapsing of NTs walls (Figure 5A) and increase in stoichiometric oxide on the surface. It is also evident form Figure 5A that the bottom layer of the NTs is destroyed or even that there is a formation of oxide layer as already observed in Ref. [30], which could be caused by a longer exposure to high temperatures. Mazare et al. [5] observed the same loss of nanotubular structure after annealing at high temperatures (i.e., 550 ◦C, 650 ◦C and 750 ◦C). It was previously reported that the formation of a rutile oxide layer is initiated at the metal-nanotube interface during the annealing [30], and therefore changes in crystal structure firstly occur at the bottom layer of the NTs. A similar structure, although not destroyed, was formed when the TiO2 NTs were treated at 800 W for 1 s (Figure 5B). However, the morphology of TiO2 NTs treated at 400 W for 10 s is not destroyed and the tubes are well defined and opened (Figure 5C).

SEM analysis revealed that TiO2 NTs of 15 nm in diameter have open tops and their structure is not destroyed after treatment with plasma at 800 W for 1 s (Figure 6A). However, NTs are completely destroyed after plasma treatment at 800 W for 10 s (Figure 6B). Ti foil morphology is unaltered after 1 s plasma treatment at 800 W (Figure 6C), while changes in the morphology of Ti foil after plasma treatment at 800 W for 10 s can be observed; the surface is covered with nanostructures (fused oxide particles) as seen in Figure 6D.

**Figure 4.** Scanning electron microscope (SEM) images of untreated as-anodized TiO2 NTs with diameter of 100 nm; (**A**) top view, (**B**) side view.

**Figure 5.** SEM images (above: top view and below: cross sectional view) of TiO2 NTs of 100 nm in diameter treated with oxygen plasma at (**A**) 800 W for 10 s, (**B**) 800 W for 1 s and (**C**) 400 W for 10 s. The red arrow indicates destroyed bottom layer of NTs.

**Figure 6.** SEM images (top view) of TiO2 NTs of 15 nm in diameter after plasma treatment at (**A**) 800 W for 1 s, (**B**) 800 W for 10 s and Ti foil after plasma treatment at (**C**) 800 W for 1s and (**D**) 800 W for 10 s.

#### *3.3. Temperature Measurements*

Crystallization of the materials in a furnace is governed by the change in temperature. Several studies report on light-induced [31–33], laser-induced [34,35], and microwave-induced [36] crystallization, which are also correlated with sample heating. In the present study we used a plasma treatment and the temperature of the samples during the treatments in correlation with plasma output power were measured. As evident from Table 2, the temperature of the sample increases with increasing output power of the plasma reactor in the time range up to about 10 s. It has been previously reported that plasma used in the present study operates in two different modes, E and H; The E-mode prevails at a low input power and is defined by a relatively low electron density, high electron temperature and low light emission, while the H-mode at higher input power represents a higher electron density, somehow lower electron temperature and higher light emission [37]. Zaplotnik et al. [27] measured the transition from E- to H–mode in the same plasma system as used in this study. It was found that the E–H transition occurs at the generator power of about 400 W at the pressure of 50 Pa. Below the output power of 400 W, the power transmission to plasma is not ideal; for example, at generator forward power of 200 W, the forward minus reflected power is about 65 W, while for example at 400 W, there is much lower reflected power which means the transmission of power is much better. Exposure of TiO2 NTs to H-mode plasma subjects the sample, among others, to more intense electron density and light emission. Therefore, the samples exposed to H-mode for a sufficient time could crystallize in seconds, while exposure of samples to E-mode, even for 3 min, does not initiate the crystallization.

**Table 2.** Effect of radiofrequency (RF) forward (output) power and plasma processing time on the temperature and the crystal structure of TiO2 NTs of 100 nm in diameter.


Although the process of nanoparticles synthesis by plasma treatment has already been reported [21], there is a lack of reports and detailed mechanisms about the crystallization upon treatment with non-thermal plasmas. Ohsaki et al. [23] claim that the crystallization of sol-gel derived TiO2 thin films can be achieved within a few minutes by a non-thermal plasma processing. The authors suggest that crystallization of the materials should be derived from the excitation by the RF electromagnetic field and not plasma itself. Similarly, An et al. [38] speculate that formation of the crystalline grains of BaTiO3 is due to the energy provided by ion bombardment. Contrary, Kramer et al. [21] suggested that the silicon nanoparticles exceed the gas temperature in non-thermal plasmas (370–430 K/97–157 ◦C) to the point of sufficient temperature for crystallization (up to 700 K/427 ◦C), although measurements showed that the average temperature of the silicon nanoparticles exposed to plasma was close to the gas temperature. According to the authors [21], crystallization of the nanoparticles by non-thermal plasma occurs due to the electron ion recombination and reactions of radicals on the nanoparticle's surface. Similarly, Lopez et al. [22] demonstrated that plasma exposure allowed for crystallization of silicon nanoparticles due to the heating of the nanoparticles to the temperature of 1100 K/827 ◦C.

In present study, the correlations between TiO2 NTs crystal structure and plasma processing conditions have been studied. It has been shown that TiO2 NTs crystallize to anatase and at appropriate conditions also to rutile phase when exposed to highly reactive oxygen plasma for sufficient time and at appropriate plasma power. It was shown that the mixture of anatase/rutile crystal structure was achieved after 10 s at the power of 400 W, where the measured temperature of the sample was above 1210 ◦C. These operating conditions allow more intense physical processes which occur during interaction of plasma species with the sample. We presume that these interactions contribute to the alteration of the sample's crystal structure, mainly due to rapid heating of the sample induced by surface recombination of neutral oxygen atoms and ion neutralization. However, the contribution of these two mechanisms to the floating sample heating in electrodes low pressure oxygen plasma are not of the same order. The densities of the ions are a few orders of magnitude lower than the densities of the neutral oxygen atoms and, therefore, the heating of the sample due to the ion neutralization can be neglected [39]. Therefore, the heating of samples (*PH*) can be expressed with the formula (1):

$$P\_H = \gamma \cdot \mathbf{j} \cdot \mathbf{W\_D} / 2,\tag{1}$$

where *γ* is the recombination coefficient of the oxygen atoms for the specific surface, *j* is the flux of the oxygen atoms onto the sample surface and *WD* is the dissociation energy of oxygen molecules. At the same time the sample is being cooled, where two most important mechanisms are radiation and convection. The radiation part (*Pr*) is described with the Stefan–Boltzmann law (2):

$$P\_r = A \cdot \varepsilon \cdot \sigma \cdot T^4,\tag{2}$$

where *A* is the surface area of the sample, ε is the emissivity, σ is the Stefan–Boltzmann constant and *T* is the sample temperature. The convection part (*Pc*) is a linear function of temperature and can be described as (3):

$$P\_{\mathcal{L}} = h \cdot A \cdot T\_{\prime} \tag{3}$$

where *h* is the convection coefficient. The sample immersed in plasma reaches steady temperature when the heating part is equal to the sum of the cooling parts (4):

$$P\_H = P\_r + P\_{c\prime} \tag{4}$$

From here we can see that larger flux of the oxygen atoms means a higher temperature of the sample. In Figure 7 the flux of oxygen atoms versus the steady temperature of the samples of NTs of 100 nm in diameter are presented. In this figure the measure points are those presented in Table 2, and the corresponding fluxes were measured with a cobalt catalytic probe. The detailed measurement technique is described elsewhere [27]. It should be noted that in this case only the measured points of samples that reached the steady temperature were used. The theoretical fit is that derived from previous formulas for cooling and heating of samples in plasma [27] taking into account the coefficient *γ* = 1.

**Figure 7.** Flux of oxygen atoms and the steady temperature of the NTs samples of 100 nm in diameter.

Destruction of the NTs was observed at higher power plasma (800 W) at prolonged treatment (more than a second). The crystallization mechanisms of TiO2 in plasma can be correlated with RF electromagnetic field, vacuum ultraviolet radiation, interaction with radicals and ions which induce sample heating, and chemical reactions. However, the prevailing influence of each parameter alone and its possible synergistic influence is not yet understood and definitely beyond the scope of this paper. Further studies in this direction should be conducted. The results of this study confirmed that the temperature, which was measured on the bottom surface of the samples during the treatments, plays an important role in TiO2 NTs crystallization.

#### *3.4. In Vitro Biological Response—Interaction with Whole Blood*

The approach to rapidly alter the crystal structure of TiO2 NTs without influencing their morphology is an intriguing way for surface modification of NTs used in bio-applications. As mentioned before, the crystal structure of a material affects its performance in biological applications. For instance, it has been shown that platelet adhesion and activation on the crystallized TiO2 NTs depends of the annealing temperature and crystal phase [40]. More precisely, authors showed that a certain amount of pure anatase in the TiO2 NTs leads to better platelet adhesion, which is beneficial for successfully integration of dental implants. By contrast, rutile-phase TiO2 NTs reduced platelet adhesion and activation, perhaps due to the formation of a hydration layer, which reduced intimate contact area between a platelet and the surface [28]. This particular case is advantageous for vascular stents, for which the inhibition of adhesion and activation of platelets is preferred, since platelets' aggregation leads to a blood clot formation and stent thrombosis. Results of platelets performance on the surface of TiO2 NTs of 100 nm in diameter prepared in present study are shown in Figure 8. Platelets adhere to the surface of Ti foil with filopodia, while adhesion on the amorphous TiO2 NTs is done via lamelopodia and fillopodia. This indicates that the amorphous TiO2 NTs provide a better environment for platelet adhesion and activation. By contrast, platelets could not be found on the surface of annealed TiO2 NTs with anatase crystal structure. No biological material was detected on these surfaces, as platelets seemed not to adhere on these surfaces. These results indicate that anatase crystal structure (obtained by plasma or annealing) of TiO2 NTs does not provide appropriate conditions for platelets adhesion, which is beneficial from the perspective of vascular stent applications. In previous publications we showed increased hydrophilicity and reduced electrolyte residual (fluorine) on the surface of annealed TiO2 NTs [26]. Since fluorine is toxic to cells, its lowered content could be the reason for the better performance of platelets. It should be also considered that besides crystal structure also surface chemistry and wettability (surface charge), as well as morphology of the material play important role in platelet response, therefore further studies of TiO2 NTs interactions with platelets are needed.

**Figure 8.** Platelets adhesion and activation on (**A**) Ti foil, (**B**) amorphous and (**C**) annealed TiO2 NTs with 100 nm in diameter.

#### **4. Conclusions**

Non-thermal oxygen plasma induced crystallization of TiO2 NTs synthesized by electrochemical anodization of Ti foil. The transition of the NTs' amorphous phase to anatase and/or rutile crystal structure was obtained within a few seconds of exposure to oxygen plasma without changing the desired morphology of NTs. Although the gas temperature of non-thermal plasma was well below the thermal crystallization temperature, the sample temperature rose above 1300 ◦C at a certain plasma treatment condition e.g. the output power of 800 W and 10 s plasma treatment time. Results provide evidence that a rapid temperature increase of samples treated in plasma is correlated with crystallization of TiO2 NTs. Good correlation between the flux of oxygen atoms and increase in temperature of the sample was found; however, other parameters in plasma may also play a vital role in crystallization and should be further studied. It should be emphasized that crystallization in plasma is also influenced by the NTs diameter, as NTs with 15 nm in diameter seem to crystalize prevalently to the rutile phase, which was at our experimental conditions not the case for NTs of 100 nm in diameter. Moreover, the in vitro biological response of whole blood with crystalline TiO2 NTs showed that the crystallization reduces adhesion and activation of blood platelets, which is of particular interest for designing medical devices that are likely to contact blood, such as vascular stents.

**Author Contributions:** Methodology, formal analysis and investigation: M.B., I.J., R.Z., M.M. and M.V., writing—original draft preparation: M.B., I.J., writing—review and editing: I.J., R.Z, M.M. and A.I, conceptualization: M.M.

**Funding:** This research was funded by the Slovenian Research Agency for financial support [grant numbers Z3-4261, J3-9262, J1-9162, J2-8166, J2-8169, J5-7098, P2-0232] and Slovenian Ministry of Education, Science and Sport [grant "Public call for encouraging young investigators at the beginning of their career 2.0", No. 5442-15/2016/18].

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

#### **References**


© 2019 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/).
