Effect of Anodization Time on the Adhesion Strength of Titanium Nanotubes Obtained on the Surface of the Ti–6Al–4V Alloy by Anodic Oxidation

Abstract

In this work, titanium oxide nanotubes (TNTs) were formed by anodic oxidation on the surface of a Ti–6Al–4V alloy. An electrolyte based on ethylene glycol (EG) and ammonium fluoride (NH₄F) was used. Different anodizing periods (10, 20, 30, 40, 50, and 60 min) with a constant potential of 60 V were established. The morphology of the TNT array was observed by scanning electron microscopy (SEM). The adhesion strength of the TNTs to the Ti–6Al–4V surface was evaluated using nanoscratch tests. The critical load of the different TNT layers was determined from the analysis of the groove of the nanoindenter path. It ranged from 0 mN for the samples exposed to 10 min of anodization to 47.0 ± 3.0 mN for samples exposed to 50 min. These results indicate the TNT layers formed at 50 min presented the best substrate adhesion among the specimen tested.

1. Introduction

The Ti–6Al–4V alloy has been studied extensively because its α/β microstructure, grain size, and processing method provide adequate mechanical properties for different applications. Likewise, the changes in its mechanical properties affect its industrial and commercial importance [1]. This alloy is one of the most popular materials for fabricating hip joint prosthesis’ femoral stems due to its high corrosion resistance and easy machining [2,3]. Moreover, titanium and its alloys may be protected against corrosion by forming a generally amorphous titanium oxide film, which naturally appears in contact with the environment.

Titanium has various oxidation states due to its high reactivity with oxygen. In the initial stages of oxidation, small amounts of oxygen can be incorporated into the structure, forming a solid solution. The oxygen randomly occupies interstitial sites in the hexagonal lattice, with a concentration of less than 40%. A wide variety of species with different physical characteristics can be found. These structures include notable ones such as Ti₆O, Ti₃O, Ti₂O, and Ti₃O₂ [4]. In particular, crystalline titanium dioxide may exhibit three crystal phases: rutile (tetragonal structure), anatase (octahedral structure), and brookite (orthorhombic structure) [5]. Therefore, it is of utmost importance to identify the studied structure to understand its properties before and after surface modification, as explored in this study and previous work.

Additionally, the elastic modulus of titanium and its alloys is closer to that of bone tissues than the elastic modulus of other biometals such as stainless steel, Cr–Co alloys, etc. Therefore, it promotes a better load distribution between bone and implant, avoiding mechanical instability in the implant–bone interface [6]. Nevertheless, the elastic modulus of Ti–6Al–4V (~136 GPa) is still higher than that of bone. Thus, it is necessary to apply a mechanical adapter to achieve osteointegration. A possible solution for promoting osteointegration could be related to forming porous layers such as nanotube arrays [7].

Titanium oxide nanotubes have been considered for applications such as biomaterials, catalysis, sensors, and solar cells [8]. Nanotube oxide layers can be formed on titanium and its alloys surfaces by anodic oxidation to enhance bone cell proliferation due to their ordered morphologies and vertically oriented arrangement [9]. Nevertheless, some issues must be considered, such as the difficulty of forming a uniform TNT array on a substrate of Ti alloys because most of them have α- and β-phase. Hence, TNTs have frequently been developed in pure titanium because their homogeneity facilitates its formation [10]. Moreover, the nanotube layer usually curls spontaneously after drying due to weak interfacial adhesion [11]. Thus, the adhesion of the TNT layer on the Ti metal is important because of the increased drug delivery from the implant–bone system [12]. In contrast, weak interfacial adhesion is not a problem in some instances, such as solar cells or electrodes for supercapacitor applications [13]. Even though the adhesion of TNTs to the Ti surface is of great importance for interacting with tissue, there is not enough information about the adhesion force and how it can be estimated.

The mechanical behavior of TNTs depends on the material’s mechanical properties and the characteristic dimensions of the nanotube arrays, such as layer thickness and nanotube diameter. In many cases, mechanical loads in the order of tenths of mN may be helpful to limit the effect of substrate properties, thus responding more indicatively of the TNT properties and their adhesion to the substrate. In that sense, nanoscratch is an important nanomechanical testing method for analyzing the adhesion between the layers and the substrate. It could be applied to different areas of the surface. The process involves using a crescent force to a spherical tip that slides over the layer [14]. Critical loads may be defined as the normal force that causes specific damage in the layer–substrate system.

Finally, this research proposes analyzing the critical load for adhesion during the scratch tests of Ti–TNTs systems to estimate the modified surface’s adhesion strength [15]. Although considering a process previously reported in the literature [16], this work emphasizes points related to mechanical behavior and adhesion, which remain open questions for the materials analyzed in this work.

2. Materials and Methods

2.1. Anodizing Treatment

Titanium oxide nanotubes were formed by anodic oxidation on cylindrical samples of Ti–6Al–4V (ASTMB348 Gr5, Química Islas, Mexico City, México). The samples’ dimensions were 19 mm in diameter and 4 mm in thickness. The anodizing time was set as 10, 20, 30, 40, 50, and 60 min, applying a constant work potential of 60 V in all cases. The used electrolyte was based on ethylene glycol (EG) of 99.8% purity, with 1 wt.% of distilled water, and 0.5 wt.% of NH₄F of 99.98% purity (conductivity and pH 1.23 mS cm⁻¹ and 7.4, respectively), (Sigma-Aldrich, Toluca, Mexico). The experimental characteristics for this work were chosen based on the best conditions reported in previous work [16]. First, samples were prepared by a metallographic procedure consisting of sequential polishing with 80–1000 SiC paper (EXTEC CORPORATION, Enfield, CT, USA). After this metallographic finishing, samples were cleaned in an ultrasonic bath for 5 min in a mixture of ethanol and distilled water (50/50). The samples were placed in an electrolytic cell in which the samples acted as the anode, and a graphite electrode was used as the counter electrode. The electrodes were separated at a constant distance of 10 mm in all the treatment conditions. After the anodic oxidation treatment, the samples were ultrasonically cleaned for 5 min (the procedure was repeated three times) with deionized water and dried in hot air.

2.2. Surface Characterization

Scanning electron microscopy (SEM), (JSM-6010LA, JEOL Ltd., Akishima, Japan), was used to evidence the presence and morphology of the nanotube layers. SEM also allows for the measurement of the diameter and length of the TNTs. Moreover, the structure of the modified surfaces was analyzed by X-ray diffraction (XRD) using D8 FOCUS diffractometer (Bruker, Billerica, MA, USA) with Cu-Kα radiation (1.5418 Å). The results were analyzed using the graphical method developed by Hull and Davey, which provides a specific mathematical approach for non-cubic diffractograms.

2.3. Nanoscratch Test

Three nano scratch tests were performed in different regions of each modified sample. A triboindenter with a conical diamond scratch tip with a radius of 100 nm and cone angle of 90° (TI 950 Triboindenter, Bruker, Minneapolis, MN, USA) was used in the analyses. The running length was established at 500 μm. The nanoscratch test aimed to find the critical load of the TNT layer by applying an increasing load of up to 50 mN at a constant velocity of 17 μm/s (Figure 1). After the nanoscratch test, images were obtained by SEM. In addition, backscattering electrons (BSE) were used to identify differences between the substrate and the TNT layer.

3. Results

3.1. TNT Morphology

The most critical observation of the specimen structure, visible in the representative SEM images of Figure 2, is the morphology and diameter of the nanotubes for each anodizing time. Figure 2 shows that the formation of a uniform nanotube layer on the Ti–6Al–4V alloy surface was not obtained due to the β-phase. However, Chernozem et al. reported that the β-phase was essential in promoting the surface’s mechanical biocompatibility because it increases the elastic modulus [17]. Thus, it is important for bone cell proliferation. Besides, patches of detached material may be associated with the high solubility of the vanadium oxide [18], indicated by yellow arrows.

The TNTs obtained after 10 min of anodization (see Figure 2a) had the largest diameter (95 ± 12 nm). Moreover, no intertubular spaces can be observed in samples exposed to 10 min of anodization. On the other hand, the nanotube array obtained on samples anodized for 60 min showed a better distribution with a diameter of 53 ± 6 nm (Figure 2f). The behavior of the TNTs’ diameter as a function of the anodization time was widely explained and discussed in previous work (see Figure 4 in Ref. [16]). Interestingly, this sample showed increased viability and proliferation of bone cells in comparison to the non-treated Ti–6Al–4V samples, based on results reported in previous work [16]. So, it is feasible to say that the viability and proliferation of bone cells on the TNTs are strongly related to the anodization time and the nanotube diameter.

3.2. XRD Analysis

Figure 3 presents the XRD pattern of the 60 min anodized specimen. The lattice parameters of titanium are a = b = 2.95 Å and c = 4.68 Å in its α-phase as a hexagonal close-packed structure [19]. The XRD patterns shown in a previous study described the presence of Ti₂O on the modified surfaces [16]. In this case, each anodized sample’s XRD patterns were analyzed to obtain the lattice parameters of Ti₂O (

Table 1. Lattice parameters in each anodizing time.

Lattice	Anodizing Time (min)
Parameters (Å)	10	20	30	40	50	60
a = b	2.934	2.88	2.971	2.789	2.871	2.96
c	4.858	4.852	4.895	4.929	4.767	4.833

). The 60 min sample shows a = b = 2.960 Å and c = 4.833 Å as lattice parameters, which are close to the values reported for Ti₂O in the references, namely a = b = 2.960 Å and c = 4.830 Å [4]. Although Ti₂O has a structure similar to Ti, incorporating interstitial oxygen increases lattice parameters, resulting in a larger variation of the c-parameter [4]. However, little is known about Ti₂O due to its unstable nature and polymorphism.

3.3. Nanoscratch

The critical load refers to the applied load when a defined adhesion failure occurs in the substrate/coating system [20]. Figure 4a shows that the layer anodized for 10 min practically did not exhibit critical load (critical load = 0 mN), so the adherence strength almost does not exist in samples exposed to 10 min of anodization. However, the TNT layer has adhered to the metallic surface. A probable explanation for the value of the critical load in the case of the sample exposed to 10 min of anodization (0 mN) could be that the adherence strength of the TNT layer obtained in this sample is under the limit of the test tool. This assumption should be considered for future analysis.

Figure 4b details the nanoscratch test groove generated on the same sample. This figure indicates that the TNT layer was detached during the nanoscratch test without leaving a trace around it. Therefore, the protrusions of the substrate related to the β-phase were exposed, indicated by yellow arrows. In addition, the EDS analysis showed an increase in the vanadium peak. This peak represents a change from a value of 1.68 wt.% within the section enclosed by the green box to 3.02% within the section surrounded by the black box on the characteristic protrusions of the β-phase. These results can be compared to Chen et al.’s research, showing Ti–6Al–4V images with β-phase distributed on α-phase [21]. This phase is easily distinguished in the following images when the critical load is reached.

Figure 4b also indicates that spallation and delamination were the most common failures. In those cases, the coating is detached by the large compressive stress in front of the nanoindenter tip to minimize the elastic energy [22]. Moreover, independently of the anodizing time, there are collapsed nanotubes alongside the scratch groove due to the cohesive failure, defined as damage or cracking of the layer without detachment [23]. From the sample exposed to 20 min (Figure 5a,b), a visible adhesive failure occurs due to the separation and differentiation of the layer–substrate system, as observed through SEM BSE images. However, the TNT layers present plastic strain and cohesive buckling before the initial delamination; hence, it is not considered for the critical load [24].

Similarly, the conformal cracks noticed at lower loads can be ascribed to the coatings trying to conform to the shape of the sliding tip in each sample. This phenomenon is common in thin films [4]. In this case, the second lowest critical load of 12 ± 2 mN belongs to the sample modified at 20 min (Figure 4c).

Figure 5a exhibits a plot of indenter penetration depth as a function of lateral displacement and, consequently, of applying a normal load. This plot provides additional information regarding the mechanical behavior of the systems. For example, for the 20 min specimen, represented by the green curve, when a displacement of 136 µm is reached, a jump to a greater depth of the nanoindenter is observed, consistent with the reported critical load. In Figure 6b, the example of the length of the nanotube layer in the 60 min anodized sample can be observed. Furthermore, in Figure 6c, the relationship between the anodizing time and the size of each nanotube layer is shown, revealing that the 10 min layer, being the shortest, was the first one to detach. In this case, the depth of the scratch reached by the nanoindenter remains constant, as the layer no longer exists during the test, resulting in direct contact with the substrate.

Figure 5c shows the groove generated on the sample exposed to 30 min of anodization, which provided a critical load of 37 ± 5 mN. For this sample, chipping and cracks were observed propagating in the material outside the loading zone before the delamination of the TNT layer (Figure 5d).

A complete removal of the layer, representing gross spallation, can be observed in the sample exposed to 40 min of anodization (Figure 7a). In this case, the critical load was 20 ± 2 mN. Furthermore, as shown in Figure 7b, the compressive stress is observable until the layer separates, revealing a TNT length of 3.92 µm (Figure 7c). Generally, the coating gradually thins until the substrate is uncovered, as occurs in soft coatings on hard substrates [25]. A possible explanation for this behavior can be that as the anodization time increases, the TNTs increase in length to a maximum size, then, a new layer of TNTs begins to grow under the first, thus facilitating the removal of this, indicating a cyclic formation of TNT layer.

The critical load achieved on the TNTs obtained on the sample exposed to 50 min of anodization was 47 ± 3 mN. This critical load value relates to the purple curve shown in Figure 6a, where the nanoindenter suddenly jumped in depth after 437 µm. Therefore, gross spallation can be observed only for the higher loads of the nanoindenter (see Figure 7d).

As indicated in Figure 4c, the sample exposed for 60 min provided the largest dispersion regarding the critical load value. Figure 8 shows how the TNTs were compacted during one of the nanoscratch tests applied to the sample exposed to 60 min. In this case, the area representing the substrate in the BSE image (Figure 8b) is not as bright as the previous ones. Thus, a close-up was made to observe the morphology in this area. The groove was analyzed thoroughly, presenting a layer below the material delamination zone. In this layer, which was not detached, the random oxidation pits represent the β-phase of the Ti–6Al–4V alloy (Figure 8c) due to the chemical dissolution of the vanadium oxides.

Figure 8d presents another of the grooves on the 60 min sample, which is associated with an adhesion failure with a load of 24 mN. This value is significantly lower than those obtained with the other two scratches in this sample. For the scratch in Figure 8d, the gross spallation exhibits a TNT length of at least 4.55 µm (Figure 8e). It can be assumed that the higher critical loads obtained during the other two grooves on this sample result from a layer below the nanotubes, which work as a cushion, improving the adhesive failure. This layer originated from the fluoride ions accumulation in the metal–oxide interface during the TNTs’ growth. For this reason, the dissolution of this layer led to the film detachment, and the anodizing formed a new layer [25,26].

Various extrinsic factors affect the response of the nanotube layer during the nanoscratch test, such as the substrate and coating properties. For example, the critical load decreases with the decreasing of the coating thickness [27]. Nanotube arrays formed by anodic oxidation frequently present poor adhesion in the coating–substrate system. Bai et al. demonstrated that crystalline phase NTs exhibit stronger adhesion to the substrate [28]. Several studies advocate heat treatment after the anodizing process to transform the amorphous titanium oxide into crystalline [29]. Although the present work involves crystalline titanium oxide, it is in the form of Ti₂O. The Ti₂O increases the c/a ratio, translating into the titanium’s crystal lattice distortion. Thus, the titanium’s crystal lattice distortion provides greater resistance to the propagation of atomic plane dislocations. This behavior is related to increased hardness, brittleness, and poor adhesion of the TNTs on the substrate [20]. Moreover, the adhesion strength is directly related to the coating’s durability [8]. This is possibly associated with forming a layer between the oxide film and Ti substrate due to fluoride ions migration during the growth of the oxide film [30].

4. Conclusions

Based on the results obtained in this work, it is possible to draw the following conclusions:

• The anodic oxidation technique makes it possible to form TNTs on the surface of the Ti–6Al–4V alloy—however, their length, diameter, and adhesion change as a function of the anodization time.
• The samples exposed to 50 and 60 min of anodization exhibited the highest critical loads, which could be related to the TNTs’ growth and the depth of the layer.
• The results show an evident influence of the anodization time on the TNTs’ morphology, especially on the diameter, where the sample exposed to 10 min showed TNTs with a diameter of 95.0 ± 12.0 nm, while the samples exposed to 60 min of anodization exhibited TNTs with a diameter of 53.0 ± 6.0 nm on average.
• According to the behavior exhibited for the sample exposed to 40 min of anodization, it is possible to consider that the TNT layers formed on the surface of the Ti–6Al–4V by anodic oxidation seem to follow a cyclic behavior where the coating adhesion is weak after a particular time. A new TNT layer starts to grow under the previously formed layer. This behavior lets us suppose the possibility of producing TNTs continuously by anodic oxidation. This assumption is an fascinating topic, so it will be interesting to continue investigating to provide more information in future research.