ANovelMethodology for Economical Scale-Up of TiO2Nanotubes Fabricated on Ti and Ti Alloys

.e prospective use of nanotechnology for medical devices is increasing. While the impact of material surface nanopatterning on the biological response is convincing, creating a large surface area with such nanotechnology remains an unmet challenge. In this paper, we describe, for the first time, a reproducible scale-up manufacturing technique for creating controlled nanotubes on the surfaces of Ti and Ti alloys. We describe an average of approximately 7.5-fold increase in cost and time efficiency with regards to the generation of 20, 50, and 100 nm diameter nanotubes using an anodisation technique. .ese novel materials have great potential in the medical field through their influence on cellular activity, in particular, protein absorption, focal adhesion, and osteoinduction. In this paper, we provide a step-by-step guide to optimise an anodisation system, starting with design rationale, proof of concept, device upscaling, consistency, and reproducibility check, followed by cost and efficiency analysis. We show that the optimised device can produce a high number of anodised specimens with customisable specimen shape at reduced cost and time, without compromising the repeatability and consistency. .e device can fabricate highly uniform and vertically oriented TiO2 nanotube layer with desired pore diameters.


Introduction
Nanotubular films on titanium (Ti) and Ti alloys are used in a significant number of applications including biomedical devices [1][2][3][4], dye-sensitised solar cells [5][6][7][8][9][10][11][12], and photocatalysis [13][14][15][16][17]. Figure 1(a) shows the increasing trend in the number of publications focusing on titanium oxide (TiO 2 ) nanotubes within the last ten years; Figure 1 For the biomedical field, a wide range of target applications have been reported regarding TiO 2 nanotubes grown on Ti and its alloys, for instance, drug delivery [18][19][20], antibacterial [21][22][23], biosensors [24][25][26], and dental and bone implants [22,[27][28][29]. In particular, the increasing demand for dental and hip implants is becoming a prime turnover in the orthopaedic industry [30][31][32]. In vitro and in vivo investigations for implant applications often require a high number of specimens for assessing the biocompatibility of the biomaterials (typically a minimum of 30 specimens for fundamental biocompatibility assays). Currently, the production rate of such surfaces is slow and limited to the production of a single specimen at a time [33,34], hence the rationale for an improved scale-up methodology was used that allows the rapid, reproducible production of specimens. e high surface-to-volume ratio of TiO 2 nanotubes grown on both pure titanium and titanium alloys have been demonstrated to promote excellent protein adsorption [41][42][43] and to o er a platform for cell adhesion, proliferation, and di erentiation, leading to the enhancement of osseointegration [22,[44][45][46]. Research has demonstrated that a small change in the nanotube pore size can signicantly a ect cellular behaviour [22,29,[47][48][49], while nanotube length does not have a major impact on cellular response [50]. It is suggested that pore size changes the arrangement and strength of the focal contacts made by cells. Cells grown on nanoporous or nanotubular surfaces exhibit an upregulation in the integrin receptors that mediate the focal contacts with resulting improvements in cytoadherence, mechanotransduction, and hydroxyapatite formation [45,51,52].
Electrochemical anodisation is an electrolytic method used to increase the thickness of the natural oxide layer on metal surfaces [71] and to create biocompatible micro-or nanoporous TiO 2 coatings, with increased surface energy and roughness [72,73]. During anodisation, aqueous or organic electrolytes with uoride ions are generally employed to produce the nanotubes. Some examples of the electrolytes include aqueous electrolyte solutions with acids,   e.g., H 2 SO 4 or H 3 PO 4 [51,74]; salts, e.g., (NH 4 ) 2 SO 4 or Na 2 SO 4 [75]; and organic electrolytes with glycerol or ethylene glycol [76,77]. Anodisation experiments are usually carried out in a two-electrode or three-electrode electrochemical system, with Ti or a Ti alloy as the anode, inert platinum foil as the cathode, and in the case of a three-electrode system, an Ag or AgCl electrode as the reference electrode [2]. TiO 2 nanotubes can be obtained either under a constant potential (potentiostatic) or constant current (galvanostatic) mode. Anodisation can be performed using specimens of a variety of shapes and sizes, depending on the maximum allowable load of the power source [72]. e thickness and morphology of the oxide layers formed are usually determined by the applied potential, the duration of the anodisation process, and the chemical composition of the electrolyte used [71]. Previous works have concluded that, at a given time point using the same electrolyte, the nanotube pore diameter, interpore distance, and nanotube length are directly proportional to the applied potential [67,[78][79][80][81].
A detailed growth mechanism of TiO 2 nanotube using anodisation techniques is described in the literature [33,74,82,83]. Briefly, the stages of nanotube formation in an aqueous electrolyte under constant potential can be monitored by the changes of current over the anodising time. As the anodising potential is initially applied, the current rapidly decays to a minimum due to the formation of a high-resistance oxide layer. Subsequently, the current rises to a maximum with the development of pore nucleation and the formation of a porous structure. e current will eventually attain an approximately constant value when an equilibrium state is achieved, i.e., the rate of oxide formation is equaled by the rate of dissolution.
Conventional anodisation systems are usually limited for production of in vitro specimens by the following factors: (i) anodising is limited to a specific specimen rather than the entire surface, leading to material wastage; (ii) trimming of specimens for in vitro studies causes specimen damage; (iii) usually only one specimen can be accommodated in the system, and high volume production is typically expensive in cost and time; and (iv) consistency and reproducibility cannot be guaranteed from specimens to specimens due to small variations in the anodising conditions. Figure 2 shows the schematic diagrams of typical conventional anodisation systems, in which the desired anodised area is defined by an O-ring on the specimen holder. e limitations above highlight the need for a tailored anodisation device design to produce fully anodised surfaces and to increase the productivity and repeatability of the anodisation process. In the present work, we designed an optimised anodisation setup to reduce fabrication costs and to maximise the efficiency rate of anodising titanium discs, for the first time, with the following objectives: (i) to anodise the entirety of specimen surfaces so as to reduce excessive material waste; (ii) to increase the number of specimens per anodised batch; (iii) to anodise batches of specimens with consistent results; (iv) to anodise specimens of desired TiO 2 nanotube pore diameters; (v) to customise the specimen shape for actual practical application, in our case, for in vitro biocompatibility tests; (vi) to evaluate the current cost per complete anodised specimen associated with the optimised device.

Materials and Methods
Pure Ti foils (thickness 0.25 mm; purity 99.5%; Alfa Aesar) and Ti-6Al-4V foils (thickness 0.4064 mm; titanium grade 5 ASTM B265; William Gregor Ltd.) were cut to the desired shape using an iPG laser ytterbium fiber laser 1 kW cutting tool, or AgieCharmilles CUT 200 Sp EDM wire-cut machine.
Before electrochemical treatment, the titanium foils were ultrasonically degreased in equal volumes of acetone, ethanol, and deionised (DI) water for 5 minutes, followed by drying under a cool air stream. An additional pickling process was introduced to Ti-6Al-4V foils prior to anodisation using a pickling mixture (Sigma-Aldrich) containing hydrofluoric acid, nitric acid, and DI water for five minutes to remove the naturally formed oxide layer.
For the electrochemical experiments, an anodising device with a three-electrode configuration was used. Specimens were mounted onto a copper rod wrapped with heatshrink polytetrafluoroethylene (PTFE) tubing; areas other than the circular working area of the specimens were masked with lacquer (Stopper 45 MacDermid) to prevent exposure to the electrolyte. A platinum foil (thickness 0.1 mm; purity 99.99%; Alfa Aesar) served as the counter electrode, placed at 40 mm distance from the working electrode. A saturated calomel Hg/Hg 2 Cl 2 (1M KCl) electrode was used as a reference electrode, connected to the setup by a salt bridge placed close to the working electrode.
Constant potentials were applied using an EG&G Instruments Scanning Potentiostat (Model 362) connected to a Ministat MKIV Sycopel Scientific signal amplifier; the in situ current-time responses (per anodisation process) were recorded using Labview software. e current density was then plotted using the current-time data per total surface area of the working specimens.
For pure Ti specimens, the electrolyte consisted of a mixture of 1M H 3 PO 4 and 0.25 wt% HF. For Ti-6Al-4V titanium alloy specimens, 1M H 2 SO 4 containing 0.1 wt% HF was used. All anodisation experiments were carried out at room temperature for 60 minutes. After the electrochemical treatment, the specimens were rinsed with DI water and then ultrasonically cleaned for 10 minutes and further dried under a cool air stream.
In the following, the specimens are designated according to the material type and anodising potential; for example, TiNT-2.5V refers to Ti specimens anodised at 2.5V to create nanotubes (NT), while Ti64NT-20V refers to Ti-6Al-4V specimens anodised at 20V. As shown in Figure 3, Version A of the anodisation setup used only one specimen, while Version B accommodated 10 specimens and a larger platinum foil as the counter electrode. e specimen was designed to fit into in vitro tissue culture plates and with a tag to assist handling of specimens using tweezers. To verify that the anodisation setup was applicable to alloy, the present work was being reproduced using Ti-6Al-4V foils. e preparation time and costs for both versions were recorded and analysed.
Journal of Nanotechnology e anodised specimens were characterised using a ZEISS Ultra 55 eld emission scanning electron microscope (FE-SEM) equipped with an energy dispersive X-ray spectrometer (EDS), as reported previously [85]. e FE-SEM was operated under the InLens detector mode with 5 kV EHT in a high vacuum environment. ImageJ image analysis software was used to measure the nanotube pore diameters, interpore distances, and nanotube lengths of the anodised specimens.
Statistical signi cance was analysed using analysis of variance (ANOVA) and counter-con rmed using the twotailed paired Student's t-test in Microsoft Excel. Data are presented as mean ± standard deviation (S.D.). Probability (p value) less than 0.05 was considered to be signi cant.

Design Rationales.
While designing the new anodisation device, the design rationales of the device setup were initially evaluated, followed by proof of concept data, leading to further upscale and optimisation. e design rationales included the following: (i) To create a scalable device design that was suitable for multiple specimens per batch of the anodisation process (ii) Maintain constant electrode distances for process control (iii) Ensure device setup that was compatible with the use of hydro uoric acid (HF) (iv) Allow an even distribution of potential and current to all specimens (v) Allow monitoring of the potential of the working electrode during the anodisation process (vi) Include a counter electrode that was at least similar or larger area to that of the working electrode (vii) Allow no contamination from copper parts, which deleteriously a ects the anodising process and in vitro cell and tissue culture tests (viii) Allow manufacture of a specimen shape that maximises the surface area for in vitro tests (ix) Allow ease of use when transferring the specimens during in vitro assays post manufacture A PTFE beaker was employed to contain the electrolyte as PTFE is resistant to HF, as compared with glass or other plastic materials like low-density polyethylene (LDPE), high-density polyethylene (HDPE), polypropylene (PP), polymethylpentene (PMP), and styrene acrylonitrile (SAN).
e PTFE beaker can also be autoclaved under high temperature if needed. e holder for the working electrode was a copper rod (shielded by PTFE and lacquer masking to prevent copper from exposing to the electrolyte) to which specimens were attached. Anodisation was able to take place over the entirety of the specimen, and the specimen shape could be varied according to the application. e platinum counter-electrode was also attached to a shielded copper rod holder. e counter-electrode was positioned 40 mm from the working electrode holder. Figure 3 shows a detailed schematic of the device setup. In Version B, as the surface area of the working electrode was increased, the area of the counter-electrode was increased to be at least similar to that of the working electrode. Two square platinum foils (50 mm × 50 mm) were welded to provide an electrode of area approximately 94 mm × 50 mm. e reference electrode was placed close to the working electrode to monitor the changes of potentials across the systems. Lacquer was used to seal connections of the platinum and specimens to the copper rods. e top cover jig of the device permitted the electrode arrangements to be preserved between batches. An opening in the cover jig allowed access of air.
As the bottom well diameter of a commonly used 24 circular well plates is 15.54 mm, as shown in Figure 3(c), a disc-shaped specimen was adopted with a diameter of 14 mm and area 3.08 cm 2 . In Version A of the device design, a single specimen was anodised. In Version B, 10 wire-cut electrical discharge machining (EDM) connected specimens  were anodised. e specimens could be later readily separated by cutting the connecting strips between the discs provided by the metal sheet cutting procedure. e resulting tags on each disc could be folded into L shape to ease the transference of specimens without damaging the specimen surface and living cell culture during in vitro tests. e   Journal of Nanotechnology number of specimens was limited to 10 to be compatible with the output of the power supply. Figure 3, was used for initial proof of concept. e schematic of the full anodisation setup is displayed in Figure 3(e). e current density response recorded during anodising of a titanium specimen at 20V is shown in Figure 4. After the first few seconds of anodisation, the current fell and decayed to approximately 1.0 mA cm −2 due to the initial formation of a barrier layer and subsequent development of the nanotube layer. A uniform array of nanotubes was observed on the anodised specimens; the top view FE-SEM micrograph is displayed in Figure 5(e). e pore diameters (∼100 nm), as shown in Figure 6(a) TiNT-20V, are comparable to that reported by Bauer et al. [80].

Upscaling and Optimisation.
e full 3D render of the optimised device (Version B) can be seen in Figure 3(d), and the schematic of the setup is displayed in Figure 3(e). e schematic compares the upscaling of the anodising surface in Version B as compared with Version A. To a greater extent, the actual anodising surface of Version B was two times as shown in the schematic, as there were two connected strips of five specimens (total n � 10) per anodisation batch. As the geometry of the working electrode and the anodisation system is symmetrical, the current passing through the system was considered evenly distributed.
When determining the maximum number of specimens allowed in the system, it is necessary to consider the maximum allowable load and current for the anodisation power supply with respect to the applied potential and total anodising surface area. In theory, considering that the current flows homogeneously across the system, the potentiostat used in this work would allow more than 20 specimens per batch if the anodising potential was under or about 20V. However, due to practical safety and constraints, the number of specimens was set to n � 10 per batch so to ensure that the power supply was not overloaded. e current density responses for each experiment using the 10-specimen setup (Version B) are presented in Figure 4.
ese results are consistent with those of other studies [86][87][88] and suggest that current density increased with an increase in anodising potential.
As presented Figures 5 and 6, nanotube dimensions measured from the FE-SEM top view of the pure titanium specimens indicated that increasing potential led to an increase in the mean nanotube pore diameter, interpore distance, and nanotube length. In addition to the pore diameter and nanotube length, the interpore distance is also a crucial indicator of the consistency of the pore nucleation and dissolution rate in an anodisation process [33,83]. e EDS analysis showed that elements of titanium and oxygen were found on the anodic layer, with traces of carbon and fluorine. e data presented above were proportional to the previously reported studies [48,69,80]. e nanotube dimensions, morphology, and EDS analysis of the anodised Ti-6Al-4V titanium alloy specimens showed a similar trend to those reported in previous publications [79,[89][90][91][92]. e EDS analysis revealed that the anodic layer was composed of titanium, aluminium, vanadium, oxygen, and traces of carbon and fluorine. As the applied potential (20V) was the same as that applied to pure titanium specimens (TiNT-20V), the nanotube pore diameter of the anodised Ti-6Al-4V specimens (Ti64NT-20V) was almost similar. However, the nanotube length of the anodised Ti-6Al-4V specimens appeared to be shorter than that of the anodised pure titanium specimens. It could be due to the different electrolyte and base material composition used, as well as the effect of the additional pickling process before the anodisation is carried out to remove the naturally occurred oxide layer. ese phenomena have been previously demonstrated in other comparable studies [90,93,94]. In addition to that, the literature has also shown that changes in nanotube length are the least significant for the modulation of cellular behaviour [50], unlike the pore size and interpore distance, which are important. e overall results substantiated that the imperative dimensions of the nanotube were exquisitely controlled.

Consistency and Reproducibility
Check. Current density curves were monitored during every anodisation batch to ensure the consistency of the electrochemical treatment. After anodisation, the morphology and elemental composition of the anodised specimens were characterised using FE-SEM and EDS to confirm the absence of contaminating elements, irregular distribution, or any inconsistency of the nanotube array.
As the anodisation outcomes using both Version A and Version B were expected to be similar, the nanotube dimensions, including pore diameter, interpore distance, and nanotube length, of the anodised specimens using both versions were measured and compared. e morphologies of subsequent batches of anodised specimens were also examined to evaluate the reliability and reproducibility using the optimised device (Version B). Data analysis presented in Figure 6 concluded that the multiple measurements on different batches were consistent as there was no statistically significant difference, i.e., a p value greater than 0.05, between Version A and each of the four batches of Version B  anodised at their respective anodising potentials, regardless of the position of the specimen in the specimen holder.

Cost Analysis.
As the specimens were of customised shape, the use of cutting devices was included in the cost analysis. Two commonly used precision automated disc cutting techniques were compared: ber laser cutting (Method 1) and wire-cut EDM (Method 2). e corresponding costs are displayed in Table 1.
Both cutting methods o er an accurate and precise cutting result, although some burr or burnt edges can be found following Method 1. Method 1 was favoured in terms of sta costs and turnaround time. However, the consumable cost of Method 2 was slightly lower because the sheet metal can be stacked into a pile and cut using EDM, whereas Method 1 can only be used on one sheet at a time, leading to increased use of consumables. Method 2 also has a significantly lower rejection rate. As no dangerous chemicals, gases, or heat dissipation were involved in Method 2, it was generally less hazardous being conducted in an enclosed chamber, and less supervision was required during the operation. Method 2 (wire-cut EDM) was chosen over Method 1 due to its overall advantages, machining quality, and lower price.
In calculating the cost, we assumed that the cutting and anodising operations were conducted by trained personnel costed at a rate of £20 per hour if in the UK. Rejection of pieces could arise due, for example, to scratches, imprecise cutting, alteration of shape, or severe burning. e costs of electricity, apparatus, and chemicals (including cutting machine energy consumption, potentiostat power supply, electrolyte, anodising device construction, etc.) have not been considered. e time and cost for producing anodised specimens were evaluated, as shown in Table 2. As mentioned, the wirecut EDM technique was used for the calculations. For Version B, a one-off cost for an extra piece of platinum plus spot welding process was not included. e preparation time prior to biocompatibility test was also included to provide an actual estimation of cost when applied in real applications. As no previous literature has estimated the cost and efficiency of an anodisation setup, Version A was assumed as an approximate replica of the conventional anodisation system because it was capable of producing single anodised specimen at a given time.
Based on the analysis, the total labour time per anodised specimen for Version A was approximately 8.5 times more than that of Version B, owing to anodising of only one specimen in the former. Combining the specimen cutting and labour cost for anodising the specimens, Version B has substantially reduced the cost per anodised specimen by a factor of about 7. Overall, the significant increase in time and cost efficiency as displayed in Table 2 has made the optimised device (Version B) a promising and economical option for anodisation.

Conclusions
Our results and cost analysis displayed a promising optimisation of an anodisation device that has a potential for use in various applications and industries, for example, laboratory researches, small-medium enterprises (SMEs), and  cross-disciplinary experiments. e design and methodology used in this study concluded that the new anodisation device is capable of producing a highly ordered uniform TiO 2 nanotube layer with tuneable pore diameters for use in, but not limited to, biocompatibility in vitro tests. Not only significantly reducing the overall time and costs associated with high volume production of anodised specimens, but also the optimised device successfully provided consistent anodisation results with increased specimen throughput, customisation of the specimen shape, and high reproducibility.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.