1 Introduction

Titanium (Ti) presented a variety of unique characteristics which is applicable in many filed over the last 50 years [1] especially as implantable materials. A number of reports have shown that the surface structure of Ti is critical for determining the success or failure of clinical Ti implantation for the purpose of bone, joint, or tooth replacements [2,3,4,5]. In the past, numerous studies on implant surface modifications have been studied to optimize the surface geometry that best-fit cell interactions for adequate bone growth [6,7,8]. Recently efforts have been made to improve cell-stimulating, biomimetic activities by designing new surface geometries at the nanoscale [9, 10].

Surface modification can be derived by transformation of structure, morphology and material surface composition without leaving the bulk mechanical properties. The aim of Ti surface modification is to produce a fine porous layer of biomaterials. Specifically, cavities and high surface area of precursor adsorption and anchoring were exploited [3, 11,12,13]. To further improve Ti bioactivity, biocompatibility, the interface between bone and implant, and implant anchorage to the bone, different surface modification methods have been explored. Typically, two different strategies have been developed. In the first approach by incorporating inorganic phases such as calcium phosphate on or into TiO2 layer interface chemically was improved. In this inorganic chemical modification, bone regeneration is stimulated and biochemical interlocking between bone matrix proteins and surface materials increases. Conversely, biochemical surface modification is differed from the first strategy and refers to organic molecule incorporation such as protein, enzymes or peptides to persuade specific cell and tissue responses [14,15,16,17,18,19].

For the second approach, the interface is improved physically and geometrically by surface topography architecture. At the micrometer stage, a rough surface creates a higher surface area interaction rather than a smooth surface. This rough surface increases bone anchorage and reinforced the biomechanical interlocking of the implant with bone up till a certain level of roughness. At the nanometer stage, the roughness increases surface energy to improve protein matrix adsorption, bone cell migration, proliferation and osseointegration [17]. To date, anodization of Ti has been investigated because can easily create biological-inspired nanometer roughness. Therefore, optimized geometrical properties of an implant surface is crucial, especially on surface characteristic (e.g., chemistry, topography, surface energy and morphology [20,21,22].

In the past decade, Kubota have suggested that TiO2 nanotube arrays grows on Ti metal would be the best candidate as an implant material. The excellent biocompatibility appears to depend on the presence of a passive oxide layer (TiO2 layer) formed on the surface [23]. Several anodization variables such as voltage, anodization period, pH and temperature of electrolyte bath contribute to the TNA nanostructure formation [24]. Up to now, several generations of nanotubes have been brought forward [25, 26] to further improve its dimensionality. This work studied on the growth behavior of TNA including anodization voltage, period, current density, electrolytes aging condition with cellular adherence profiles. These investigations would provide beneficial knowledge on TNA geometrical advantages for biomedical applications.

2 Materials and Methods

2.1 Materials

The titanium (Ti) foil of 0.13 mm thickness with 99.7% purity was purchased from STREM Chemicals, USA. Prior use the foil was cleaned with acetone (Merck, USA) in ultrasonication for 15 min, then were rinsed with deionised water and dried at room temperature (RT). Ammonium fluorides (NH4F) salt were purchased from Merck, USA. The study materials were sterilised in 70% (v/v) ethanol followed by cleaning with sterile deionised water before introducing for autoclaving at 121 °C for 20 min.

2.2 Synthesis Parameters

Self-organised highly ordered TNA was successfully achieved by electrochemical anodization following the method by Effendy et al. [27, 29,30,31,32]. TNAs were prepared by electrochemical anodisation (EA) and monitored using Leios Material Analysis software. The pH of the electrolyte was maintained at pH 7 throughout the work. Anodic potential at 30 V with sweep rate of 1 V/s for different duration (10 s, 1 min, 2 min, 5 min, 10 min, 20 min, 30 min, 1 h, 2 h, 3 h and 6 h) was implemented in the whole process. The anodized sample was annealed to achieve the anatase crystalline phase as described previously [28,29,30,31].

2.3 Cell Cultures

Human osteoblast cell line SaOS-2 (HTB-85TM), human embryonic kidney cell line HEK 293 (ATCC CRL-1573), and human colon cancer cell line HT 29 (HT-29—HTB-38—ATCC) were all purchased from ATCC. Professor George Tsao of the University of Hong Kong provided HK-1. C666-1, meanwhile, was a gift from Professor Kwok-Wai Lo of the Faculty of Medicine, Department of Anatomical & Cellular Pathology, The Chinese University of Hong Kong. HEK 293 and HT 29 were maintained in DMEM supplemented with 1% (v/v) Penstrep, 1% (v/v) L-Glutamine, 10% (v/v) foetal bovine serum and 12.5 g/L of HEPES in a humidified incubator at 37 °C, 5% CO2. HK-1, C666-1 and SaOS-2 were maintained in a complete medium of RPMI with the same formulated supplements. 1 × 105 cells were seeded on 1 cm × 1 cm TNA surface using a 24-well plate for 48 h culture period.

Prior to FESEM analysis, cells on the top side of TNA samples were fixed in a fixative agent. Briefly, the TNA samples were gently rinsed with PBS, thrice then subsequently treated with 2.5% (v/v) cold glutaraldehyde for half an hour. Then, all samples were rinsed once with PBS before dehydrating with increasing concentrations of ethanol (50, 70, 90 and 99% (v/v)) for 5-min per concentration. Each sample was separately stored at − 20 °C before FESEM analysis was conducted. Gold palladium coating sputter was performed to improve the conductivity for clear and concise images. All experiments were performed in triplicate.

3 Results

3.1 The Electrolyte and Voltage Parameters

In general, anodisation voltage with 30 V produces between 66.149 ± 13.6 to 76.994 ± 9.87 of average TNA diameter. On the other hand, anodisation voltage with 60 V produces between 121.752 ± 18.5 to 169.53 ± 14.6 of average TNA diameter. Apart from voltage, the preparation of electrolytes and the growth of TNA was studied as shown in Fig. 1. Table 1 shows the average diameter of the TNA after anodization in electrolytes aged at different period. The longer the mixing period, the more soluble the fluoride salt, thereby making the release and uptake of ions easy for TNA growth. FESEM analysis of TNA in the electrolyte at 1 and 5 h shows the disorganised formation of TNA compared with TNA anodised in electrolyte aged overnight (16 h). However, the electrolyte aging did not clearly influence by the anodisation voltage condition for TNA diameter.

Fig. 1
figure 1

FESEM images of TNA with different anodisation voltage and electrolyte aging buffers

Full size image
Table 1 The average diameter of TNA in different anodisation voltage and electrolytes aging periods
Full size table

3.2 Current Density

Furthermore, the typical current density recorded during anodization of Ti at 30 V in glycerol containing 0.5 wt% NH4F solution is displayed in Fig. 2. There is correlation between the current transient behaviour and changes in surface morphology of the oxide layer formed on Ti foil during anodization. As can be seen in Fig. 2, the current density decreased drastically from 1.8 to about 1.1 mA within first few minutes due to the formation of compact oxide layer.

Fig. 2
figure 2

Current density vs. time during anodization of Ti in glycerol containing 0.5 wt% NH4F solutions at 30 V

Full size image

Oxidation of Ti was according to the Eqs. 1–4. The electrode reaction was confirmed by formation of white bubble surrounding the cathode electrode at the beginning of anodization due to O2 evolution. Details formation of the TNA was been discussed concisely by Sreekantan et al. and Indira et al. [32, 33].

$${\text{2H}}_{{2}} {\text{O}} + {\text{4H}}^{ + } \to {\text{O}}_{{2}} + {\text{8H}}^{ + }$$
(1)
$${\text{4OH}}^{ - } \to {\text{2H}}_{{2}} + {\text{O}}_{{2}} + {\text{4e}}^{ - }$$
(2)
$${\text{Anode}}:{\text{Ti}} + {\text{O}}_{{2}} \to {\text{TiO}}_{{2}}$$
(3)
$${\text{Cathode}}:{\text{4H}}^{ + } + {\text{4e}}^{ - } \to {\text{2H}}_{{2}}$$
(4)

3.3 Anodization Duration

Figure 3 shows the formation of nanoporous oxide layer at different anodization duration. From Fig. 3b and c, random pits formation can be seen to take place up to 2 min. Interaction of surface Ti4+ ions with oxygen ions in the electrolyte formed an initial oxide layer during the first 10 s of anodization (Fig. 3a) and compact layer of TiO2 formed on Ti foil can be seen. There were large cracked area of porous oxide film and craters at the interface film substrate could provide easy pathways for ions penetration from electrolyte. A few patches of un-anodized Ti still present after 1 min of anodization (Fig. 3b) and in (Fig. 3c), there is a complete transformation from Ti to TiO2 porous layer on the sample surface after 2 min of anodization. The pores are not self-ordered at this stage and small pits that originated from localized oxide were observed due to the dissolution of the chemical process induced by F− ions where it acted as a pore-forming centre. Hence, the electrolyte ions will easily penetrate into the interface and form pores.

Fig. 3
figure 3

Evolution of the nano-tubular oxide layer during anodisation of Ti in glycerol + 0.5 wt% NH4F solutions at 30 V for a 10 s, b 1 min, c 2 min, d 5 min, e 10 min, f 20 min, g 30 min, h 1 h, i 2 h, j 3 h and k 6 h

Full size image

Anodization for 5 min (Fig. 3d) resulted in the separation of individual tubes from the nano-pores to larger pores where the thickness of outer oxide layer is about 225 nm. As time increase, distinct nanoporous morphology is observed (Fig. 3e) on the sample surface after 10 min of anodization. The breakdown sites act as seeds to disordered worm-like structure growth (Fig. 3e–k). The dissolution of TiO2 from Ti resulted in pH gradient at top and bottom pore of the tube where lower pH could be accelerating the dissolution and pore penetration into Ti substrate resulted in pore transition from irregular to regular formation. To form uniform nanotube morphology, different pore will compete with each other for total available current and if the current is sufficient, the pore will survive and retain the morphology of the tube [34]. Small increase of current density around 30 min indicated the nucleation of nanopores and nanotubular oxide formation due to the random breakdown of barrier oxide film on Ti surface and dissolution of F− ions. Dissolution of TiO2 to form pores occurred according to the Eq. 5.

$${\text{TiO}}_{{2}} + {\text{12F}}^{ - } + {\text{4H}}^{ + } \to {\text{TiF}}_{{6}}^{{{2} - }} + {\text{2H}}_{{2}} {\text{O}}$$
(5)

The appearance of sample anodized for 20 min (Fig. 3f) was better than 10 min because the debris on anodized surface was less. Thus, it is clear that from 5 to 20 min, the pit transforms to tubular structure. However, compact oxide partially covering the surface. Clear self-ordered TNA structures could be observed after 30 min of anodization (Fig. 3g). The original oxide layer clearly dissolved. Anodization of Ti for 1 h, 2 h and 3 h (Fig. 3h–j) has similar morphology as sample anodized for 30 min. The TNA inner diameters were constant with time. However, by increasing the anodization time beyond 30 min, TNA length were found to increase [35]. At this stage, the predominant formation of the oxide layer is maintained, resulting in the formation of a thick oxide layer along with a slow chemical dissolution process. As the experiment continues, the chemical dissolution process actively proceeds according to Eqs. 1 and 2. Even though the length of TNA keeps increasing with time, the TNA started to collapse and peel off from Ti foil after 6 h (Fig. 3k) of anodization. Therefore, anodization time need to fix within 30 min to 3 h to have adequate well-developed nanotubes.

Figure 4 show the correlation of TNA length and inner diameter with regard of anodization time. From the graph, the thickness of TNA porous layer was influenced by anodization time. It was noticed that three different stage of inner diameter growth rate. At early stage (stage I) of anodization (within 2 min to 30 min), the length of nanotubes drastically increased. Table 2 shows the growth rate of TNA formation. The growth rate was greatly decreased after 1 h to 2 h of anodization. The longest TNA was produced after 6 h of anodization. However, long TiO2 nanotubes (~ 5 μm) tend to peel off. So, it is not suitable to be used for cell interaction studies. Therefore, TNA anodized for 3 h with 2.2 μm length was selected to undergo a cell interaction study.

Fig. 4
figure 4

Correlation of TNA nanotubes length and inner diameter with regard of anodization time. The triangles indicate the growth rate of TNA nanotubes at three different stages indicate by stage I, II and III

Full size image
Table 2 Growth rate of TNA formation
Full size table

3.4 The Crystalline Properties

Figure 5 showed XRD patterns recorded at room temperature of un-annealed TNA and annealed TNA noted as (a) and (b). The outcome of the analysis manifested smaller anatase crystal size of un-annealed compared to annealed TNA. Crystallite size of annealed TNA showed slight increase compared to un-annealed TNA. Two new peaks were observed (red ring) showing crystalline of sample occurred at position 25θ and 48θ. The anatase peak of TNA can be identified at 25.5, 38.5, 54.5 which correspond to (101), (004), (200), and (105), respectively.

Fig. 5
figure 5

XRD analysis of un-annealed (A) and annealed (B) TNA

Full size image

3.5 FESEM Adhesion Profiles

The initial stages of cell adhesion are critical for biomaterial integration. Therefore, in the present study, the adherence response of various human cell line models on TNA was studied using FESEM. The cellular protrusion adheres on the nanotube’s surface as shown in Fig. 6a–e. Detailed observation found possible filopodia interaction with the nanosurface as indicated by the red and yellow arrows. Figure 6a, c, d, and e showed clear cell spreading and proliferation activities (indicated by cell numbers).

Fig. 6
figure 6

Cells attachment studies on TNA surface by FESEM. The yellow arrow indicates the presence of filopodia anchoring the TNA surfaces

Full size image

4 Discussion

Presently, arrays consisting of highly ordered TNA fabricated by a simple electrochemical approach have attracted more researchers [36,37,38] to participate in this research area. Several factors have been discussed affecting the geometry of TNA formation [39,40,41,42,43]. Therefore, this work described in detail on the electrolyte, voltage, current density, anodization duration, and crystal structure of TNA-associated respective geometrical properties.

There are three possible major reactions depending on anodization parameters involving electrolytes condition; first is the formation of the metal ions (Mn+) from dissolving the metal in the solution, second is metal oxide formation due to the interaction of metal ions with oxygen ions from water molecules in the electrolyte and this metal oxide can be found on the surface in the bulk layer where the metal oxide products are insoluble in the electrolyte, lastly is oxide formation and oxide dissolution reactions in specific conditions will compete with each other to form porous in the metal oxide layer [44].

Findings from this study found that the electrolyte aging did not clearly influence the anodisation voltage for TNA formation. However, aged electrolytes could lead to better TNA growth. It is clear that the anodizing voltage has an important effect on the anodization of Ti. Gulati et al. [23] discovered the benefit of applying recycled, aged electrolytes, which improved the topography of the TNA and resulted in a smoother surface. Based on the result from Fig. 3, we can identify two basic mechanisms to form nanoporous structures; (1) electrochemical oxidation and (2) chemical dissolution. An initial TiO2 layer or barrier layer is formed on a metal surface under the application of an electric field. Then pore initiation occurred in the presence of fluoride (F−) ions at preferred sites on the oxide layer due to electric field-assisted dissolution. These oxidation and chemical dissolutions were active at the pore bottom. The barrier layer was reduced by pore formation while concentrated electric field intensity beneath the barrier layer ensures further growth of pores [45].

In this study, the anodic parameters of 30 V with a sweep rate of 1 V/s for 30 min produce well-aligned TNA on the oxide layer. In general, regular nanotubes were formed in shorter duration of anodization and irregular nanotubes were formed in longer duration of anodization as a result of excessive dissolution occurring along the whole tube length. Previous studies reported that the length of the nanotubes increases dramatically by extending anodization reaction time. The oxide-growth process at the TiO2 nanotubes bottom and the oxide-dissolution process at the TiO2 nanotubes mouth are two important processes responsible for the nanotubes anodic growth. The TiO2 nanotube length increases gradually before the dissolution rate becomes equal to the TiO2 nanotube growth rate.

From the graph in Fig. 4, the growth rate TNA was detected at three different regions. At stage I, TiO2 dissolution at the bottom of the pits was accelerated making the barrier layer at the bottom relatively thin [46]. In turn, the electric field intensity increases resulting in further pore growth. In meanwhile, electrochemical etching by F− ions also takes place and converts the tubes into larger tubes. It can be seen that TNA formation became slower at the second 2 stage. It is because the F− became less for electrochemical to take place and not much changes in TiO2 nanotubes inner diameter. The TNA formation growth rate decreases due to the equal rate of chemical dissolution and electrochemical etching at stage II. At the final stage (stage III), TNA inner diameter hardly varied because electrochemical etching was slower than in stage 2. It is due to the absence of F− ion. Thus, the chemical dissolution suppresses the electrochemical etching resulting higher rate of TNA length growth.

The outcome from this study found smaller anatase crystal size of un-annealed compared to annealed TNA. The crystallite size of annealed TNA showed a slight increase compared to unannealed TNA. The crystalline phase of TNA plays a crucial point in promoting cell interaction to the nanosurface [46]. Lai et al. [47] mentioned that anatase phase TNA can be achieved at 400 °C for 2-h which had been applied in our work. The predominant anatase phase was seen on the sample treated at 400 °C compared to the un-annealed sample. In the reports by Yu et al., the crystallinity of TNA also influenced cell adhesion [48, 49].

Anatase phase TNA is more biocompatible than amorphous TNA due to changes in wettability and surface energy [50, 51]. The properties of the TNA surface in promoting cell attachment are studied, such as the presence of gas during the annealing stage. Argon gas enhances the surface energy of the TNA nanosurface [52, 53] to become more positive, thereby attracting the negatively charged cells to fix and proliferate on the TNA surface compared with normal air. However, the main idea of cells attached to the nanosurface was due to the roughness of the TNA surface [54]. TNA growth promotes a larger surface area, providing space for cell attachment and filopodia anchoring. The cell structure is also maintained during cell proliferation, which supports the idea of TNA as a biocompatible material, particularly in implantation [55].

Further observation via FESEM and supported by our previous work on cellular interaction on TNA found various cellular protrusion activities on the surface of the nanotubes which mediates cell adhesion and provides a cell survival signal [56,57,58,59]. Adherence and spread of cells cultured on TNA might be due to the extremely larger surface area, the distinct up-and-down vertical topography contributing to the locked-in cell arrangement, and the existence of the body fluid channel in-between TiO2 nanotubes. These interactions might involve mechano‐cytoskeleton remodeling activities as described in previous work [60, 61]. Previous works [62, 63] also supported that stated that cells that stretched on a material surface with more anchorage could provide good adhesion properties that are curial for medical implant surface technology.

5 Conclusion

The geometrical formation of TNA by anodization is directly affected by the electrolyte aging condition, anodization duration, and crystal structures. In this study, the anodic potential of TNA was successfully maintained at 30 V with a sweep rate of 1 V/s for 30 min, and the structure of TNA was well aligned on the oxide layer. The TNA geometrical advantages may provide better cellular adhesion profiles that beneficial for medical implant surface technology.