Electrochemical and in vitro biological behaviors of a Ti-Mo-Fe alloy specifically designed for stent applications

Highlights

• •Ti-8Mo-2Fe was resistant to pitting corrosion in PBS, a clear advantage over 316 L.
• •Ti-8Mo-2Fe exhibited similar corrosion rate in PBS than pure ti and 316 L.
• •Ti-8Mo-2Fe alloy did not generate acute hemolysis.
• •Blood coagulation was retarded on ti-8mo-2fe surface in comparison to pure Ti.
• •The proliferation rate of ECs and SMCs on ti-8mo-2fe was lower than on 316 L.

Abstract

There is a deep interest in developing new Ni-free Ti-based alloys to replace 316 L stainless steel and Co-Cr alloys for endovascular stent application, mainly because the release of Ni can generate toxicity and allergenicity. Interactions of Ti alloy biomaterials with bone cells and tissues have been widely investigated and reported, while interactions with vascular cells and tissues, such as endothelial cells (ECs) and smooth muscle cells (SMCs), are scarce. Therefore, this study focused on the relationship among the surface finishing features, corrosion behavior and in vitro biological performances regarding human ECs, SMCs and blood of a newly developed Ti-8Mo-2Fe (TMF) alloy, specifically designed for balloon-expandable stent applications. The alloy performances were compared to those of 316 L and pure Ti, prepared with the same surface finishing techniques, which are mechanical polishing and electropolishing. Surface properties were investigated by scanning electron microscopy (SEM), atomic force microscopy (AFM), contact angle (CA) and x-ray photoelectron spectroscopy (XPS). The corrosion behavior was assessed with potentiodynamic polarization (PDP) and electrochemical impedance spectroscopy (EIS) tests in phosphate buffered saline (PBS) solution. No significant differences were observed regarding the corrosion rate measured with PDP analyses, which was of the order of 2 × 10⁻⁴ mm/y for all the studied materials. Moreover, similarly to pure Ti, TMF exhibited an advantage over 316 L for biomedical applications, namely remarkable resistance to pitting corrosion up to high potentials. The results evidenced a good cytocompatibility and hemocompatibility, making this group of alloy a potential candidate for cardiovascular implants. In fact, both ECs and SMCs proliferated on TMF surfaces showing a 7-day viability similar to that of pure Ti. Regarding hemocompatibility, TMF did not cause hemolysis, and blood coagulation was delayed on its surface in comparison to pure Ti. When compared to 316 L, TMF showed similar hemocompatibility.

1. Introduction

Despite the excellent mechanical properties of 316 L stainless steel and of Co-Cr alloys for endovascular stent applications, there is a concern about the release of Ni and Cr from these materials [1]. In addition to being cytotoxic, Ni is also allergenic, carcinogenic, genotoxic and mutagenic [2]. There is an interest in developing new Ni-free alloys to replace 316 L stainless steel and Co-Cr alloys for these applications. Biodegradable metals (Mg, Fe and Zn-based alloys) have been proposed for making stents that would fully degrade after the treated vessel is remodeled, avoiding long-term complications that often occur with permanent stents [3]. Several studies have been performed to overcome the specific limitations of each one of these biodegradable metals [4]. Regarding Mg-based alloys, known for their excellent biocompatibility, for example, researchers have been working on the improvement of their corrosion resistance, since some of them degrade too fast and non-uniformly [4,5]. Parallel to biodegradable metals, more biocompatible permanent metallic materials have been developed for stents. Ti-based alloys, especially β-type ones, have been targeted [6], [7], [8], [9], [10]. In fact, Ti-based alloys have found their niche as biomaterials in the manufacture of orthopedic and dental implants mainly due to their excellent cytocompatibility towards bone-related cells [11]. For example, for orthopedic implants, the (α + β)-type Ti-6Al-4V and Ti-6Al-7Nb alloys have been widely used, while for dental endoprosthesis, grade 2 pure titanium (α-type) has been the main choice [11]. The development of new low elastic modulus β-type Ti alloys containing elements that are more cytocompatible and less hazardous to the human body (when compared to Al, V, Cr and Ni) to replace the (α + β) and α-type alloys aforementioned for bone-contact implants have been widely explored [2,11]. The success of titanium alloys for bone-contact implants has motivated the search for Ni-free Ti-based alloys to be used in the manufacturing of endovascular stents. However, the Ti alloys already available on the market do not meet the mechanical requirements for balloon-expandable stent applications. Indeed, the main drawbacks for the present group of alloys are their disadvantageous strength-ductility trade-off and low work hardening rate [7,12]. In the past ten years, several β-type Ti alloys combining different plastic deformation mechanisms have been developed [13]. The combination of dislocation slip and mechanical twinning [14], as well as the combination of mechanical twinning and stress-induced phase transformation [7,13,[15], [16], [17], [18]], have been explored. The design goal for materials with these deformation mechanisms resulted in β-type Ti alloys with better strength-ductility compromise and a suitable work hardening rate [13], so that they could be potential candidates to replace 316 L stainless steel and Co-Cr alloys.

Bortolan et al. [19] previously reported new β metastable Ti-Mo-Fe alloys (Ti-8Mo-2Fe and Ti-10.5Mo-1Fe), combining mechanical twinning and stress-induced phase transformations (martensite α" and ω) as plastic deformation mechanisms. These alloys exhibited high strength, ductility and work hardening rate [19], representing promising candidates for balloon-expandable stent applications. In addition to the mechanical properties, the interactions of metallic biomaterials with body fluids and cells, such as their corrosion resistance, hemocompatibility and cytocompatibility towards vascular cells (namely ECs and SMCs) are of great importance for endovascular stent applications [6,[20], [21], [22]]. In the alloys developed, the selection of Mo as the main β stabilizer was based on its biocompatibility, in particular its bio-corrosion resistance and cytocompatibility [2]. Indeed, Ti-Mo alloys, with a Mo amount up to 15 wt%, were reported to have a better corrosion resistance and electrochemical stability for biomedical applications when compared to pure Ti, even in chloride-containing solutions [23]. In terms of cytocompatibility, several works have reported that Ti-Mo based alloys exhibit in general a good material response towards typical cell lines of bone tissues [24], [25], [26], [27], [28]. Data on vascular tissue line cells, such as endothelial [6] and smooth muscle ones, are scarce. Although Eisenbarth et al. [29] reported that the viability of bovine aortic endothelial cells on pure molybdenum was reduced and suggested that this element should be used in small amount in Ti-based materials to avoid adverse cellular effects, Ion et al. [6] showed that the presence of 8 wt% of Mo in a Ti-Nb-Mo alloy did not cause any adverse effect to human umbilical vein endothelial cells (HUVECs). Regarding hemocompatibility, pure titanium is known to have problems with thrombosis and platelet adhesion [20,30,31]; on the other hand, data on the hemocompatibility of Ti-Mo alloys are not promptly available. Moreover, the presence of Fe as a second alloying element in these alloys developed, even in small amount, could have an impact on the corrosion resistance [32,33] and, possibly, on their biological performances. As already evidenced for the family of alloy studied, Fe was chosen due to its low cost, strong β stabilization effect and effectiveness in solid-solution strengthening [19].

Besides the chemical composition, it is known that the properties of the oxide layer on the surface of a biomaterial and its topographical features (especially roughness) have an impact on both corrosion properties and biological performances [34,35]. Air-exposed oxide layers formed spontaneously on the surface of biomaterials, such as 316 L stainless steel and Ti-based materials, are generally thin, non-chemically homogeneous, contaminated, and defective [34,36]. Surface treatments including mechanical polishing and electropolishing are performed for a series of reasons, and also in order to obtain a less contaminated, smoother and more protective oxide layer on the surface of biomaterials [34], [35], [36], [37], [38], [39], which generally improves corrosion properties and affect biological performances. Regarding mechanical polishing, the addition of the oxidizer H₂O₂ in the polishing slurry (alumina or silica-based) has been shown to result in protective oxide layers on the surface of Ti-based materials, without contaminated and reacted layers, and with better smoothness [36,40,41]. Electropolishing, on the other hand, has been shown to improve the corrosion resistance of 316 L stainless steel, the attachment of endothelial cells to its surface and its thromboresistance [35]. Moreover, electropolishing is a more suitable surface treatment than mechanical polishing for devices with complex geometries such as stents, being widely applied to achieve a surface quality suitable for this application [3].

Therefore, the present work aimed at investigating the corrosion behavior and the biological performance in vitro when in contact with human blood and human endothelial and smooth muscle cells of the new Ti-8Mo-2Fe alloy. In particular, its corrosion and in vitro biological performances were compared to those of 316 L stainless steel [42] and pure titanium grade 2 [43]. Two surface conditions of Ti-8Mo-2Fe alloy were evaluated, that is mechanical polished one using a 50 nm average diameter silica-based slurry containing H₂O₂, and an electropolished one. Surface features such as roughness, wettability and chemical composition of all materials/conditions studied were characterized with a series of techniques to assess the morphological, physical, electrochemical and biological properties of the alloy.

2. Materials and methods

2.1. Materials and samples preparation

The materials used in the present work were commercially pure titanium grade 2 (Ti-CP, Titanium Industries Inc., Canada) and AISI 316 L stainless steel (316 L, Goodfellow Inc., England) in the form of sheets (thickness of 0.6 mm for pure Ti and of 0.5 mm for AISI 316 L), and the non-commercial Ti-8Mo-2Fe alloy (TMF) in the form of 12.7 mm diameter cylinders. Their chemical compositions are shown in Table S1 of the supplementary data document. For all materials, samples of 12.7 mm diameter disks were prepared.

The Ti-8Mo-2Fe alloy was prepared as follows: cylindrical ingots 50 mm in diameter and 250 mm in length were produced from pure metals in an induction vacuum furnace (Consarc, USA) under the protection of a high-purity argon atmosphere. The ingots were hot isostatic pressed at 140 MPa for 2 h at 900 °C in a HP1030 press (American Isostatic Presses Inc., USA), machined to 33 mm diameter ingots and then rotary swaged at 900 °C under atmospheric conditions, with a cross-sectional area reduction of 75%. A solution heat treatment was performed at 900 °C for 1 h under an argon atmosphere, followed by water quenching, in order to retain a fully β microstructure at room temperature [19].

The surfaces of the 316 L stainless steel samples were prepared by electropolishing in order to approach as maximum as possible the surface condition of the 316 L stainless steel stents on the market. In fact, electropolishing is a surface preparation method widely applied for biomedical applications and already consolidated for AISI 316 L [37,44,45]. Before electropolishing, the samples were firstly washed with industrial soap (Alconox Citranox®), rinsed with water and dried with medical grade compressed air. Subsequently, a further cleaning procedure was applied, consisting of immersion of each sample in a series of three ultrasonic baths (acetone, de-ionized water and methanol as solvent) for 10 min each, and drying with medical grade compressed air after each bath. The electropolishing process of AISI 316 L samples was carried out in an electrolyte composed of ACS grade phosphoric acid, sulfuric acid and de-ionized water (60:30:10% v/v) at 70 °C. The cathode was another 316 L sample and a constant current of 3A was applied for 6 min. After electropolishing, the samples were exposed to an acid bath (acid dipping process) containing de-ionized water, nitric acid and hydrofluoric acid (80:10:2% v/v) at 50 °C for 30 s [45]. Finally, they were rinsed with de-ionized water and dried with compressed air. This condition was named throughout the manuscript as 316 L EP.

The surfaces of the pure titanium samples were prepared with a mechanical polishing procedure, comprising grinding up to 1200 grit abrasive paper and polishing using a mixture of 0.05 µm colloidal silica suspension (OP-S, Struers) and hydrogen peroxide (67:33% v/v). After polishing, the samples were washed with water and dried with compressed air. This condition was designated as Ti-CP MP.

The surfaces of Ti-8Mo-2Fe samples were prepared with two different procedures, namely (i) mechanical polishing and (ii) electropolishing. (i) The mechanical polishing procedure was the same as described above for pure titanium samples. This surface condition was designated as TMF MP. (ii) As a preliminary step before the electropolishing procedure, the samples were grounded with 800 grit abrasive paper and subjected to the same cleaning procedure described for 316 L stainless steel samples. The electropolishing procedure was carried out with an electrolyte composed of acetic acid, sulfuric acid and hydrofluoric acid (55:35:10% v/v). Ten drops of glycerol were added for each 170 mL of electrolyte, which was cooled until the temperature was in the range of 20–25 °C. The cathode was a 316 L stainless steel sample and a constant current of 1.5 A was applied in two runs of 3 min. After each run, the sample was rinsed with de-ionized water and dried with medical grade compressed air. After electropolishing, the samples were placed in an ultrasonic bath of de-ionized water for 15 min, rinsed with de-ionized water and dried with compressed air. This condition was named as TMF EP.

The samples were stored inside a desiccator until further use.

2.2. Surface characterization

The surface morphology was assessed by a FEI QUANTA 250 (Oregon, USA) scanning electron microscope (SEM) equipped with a tungsten filament and operated in the high-vacuum mode with an acceleration voltage of 15 kV.

Topography and roughness analyses were performed using the tapping mode of a Veeco Dimension™ 3100 (New York, USA) atomic force microscope (AFM) with an etched silicon tip (model NCHV, tip radius = 10 nm, Bruker). Five different regions of randomly selected samples from each condition were analyzed at four scan sizes 50 × 50, 20 × 20, 5 × 5 and 1 × 1 µm². Root mean square roughness (R_q) was determined using NanoScope Analysis software provided by Bruker Corporation.

Surface wettability was assessed by static contact angle using an AST Products Inc. VCA optima XE device (Massachusetts, USA). Ten acquisitions for randomly selected samples from each condition were carried out using 1 µL distilled water droplets.

The surface atomic composition was assessed by a Physical Electronics PHI 5600-ci (Minnesota, USA) x-ray photoelectron spectrometer (XPS) and a PHI PC-access software. A total of six surveys and six high-resolution spectra acquisitions were carried out for each condition. Survey spectra were acquired using the K_α line of a standard achromatic Al x-ray source (1486.6 eV, 300 W), while high-resolution spectra (C1s, O1s, Ti2p, Mo3d, Fe2p and Cr2p) were acquired using the K_α line of a standard achromatic Mg x-ray source (1254.6 eV, 300 W). The detection was carried out at a detection angle of 45° to the surface normal. CasaXPS software was used for the analysis of the spectra and Shirley background was used for subtraction. Gauss-Lorentzian (30) and Lorentzian asymmetric line shapes were used for the deconvolution of the components.

2.3. Electrochemical characterization

Potentiodynamic polarization (PDP) and electrochemical impedance spectroscopy (EIS) tests were performed with a three-electrode cell (saturated calomel as reference electrode, graphite as counter electrode and the sample under evaluation as a working electrode) connected to a VersaSTAT potentiostat/galvanostat system, controlled by VersaStudio software by AMETEK Princeton Applied Research (Tennessee, USA). For both PDP and EIS, at least three samples from each condition were tested. The tests were carried out in a phosphate buffered saline (PBS) solution at pH 7.4 (Sigma-Aldrich, USA) and 37 °C. The surface area exposed to the solution was 0.096 cm². Prior to the PDP and EIS tests, the open circuit potential (OCP) was monitored for 2 h. The PDP tests were performed at a scan rate of 1 mV s⁻¹ for a potential range between −0.3 V with respect to OCP and 1.7 V with respect to the reference electrode. The corrosion rate was determined based on the ASTM G59 standard. The EIS tests were carried out over a frequency range from 100 kHz to 0.01 Hz, applying an amplitude of 10 mV. They were recorded at OCP after 2 h of immersion in PBS. Fitting of Nyquist plots was performed using EC-Lab software provided by BioLogic Science Instruments Ltd.

2.4. In vitro biological characterization

In vitro biological tests were carried out with 12.7 mm-diameter circular samples (316 L EP, Ti-CP MP, TMF MP and TMF EP) previously sterilized by UV irradiation. Briefly, each side of the samples underwent two 15-min cycles of UV (254 nm) irradiation. Samples were stored in a sterile 24 multi-well plate inside a desiccator until use.

2.4.1. Direct cell viability assay

Human umbilical vein endothelial cells (HUVECs) and human umbilical artery smooth muscle cells (HUASMCs) were used for direct cell viability assays. They were isolated from human umbilical cord samples. A detailed description of the protocols of cell isolation and culture can be found elsewhere [46]. M199 and Dulbecco's Modified Eagle's Medium were the culture media used for HUVECs and HUASMCs, respectively. Both were purchased from Gibco (USA) and supplemented with 10% fetal bovine serum (Gibco), 1% penicillin/streptomycin (Gibco), 2 ng mL⁻¹ of basic fibroblast growth factor (Gibco), 0.5 ng mL⁻¹ of epidermal growth factor (Invitrogen, USA), 1 µg mL⁻¹ of l-ascorbic acid (Sigma, USA), 5 µg mL⁻¹ of human insulin solution (Santa Cruz Biotechnology, USA), 1 µg mL⁻¹ of hydrocortisone (Sigma), 90 µg mL⁻¹ of porcine heparin sodium salt (Grade I-A, Sigma), and 1 µg mL⁻¹ of endothelial cell growth supplement (Becton Dickinson, Canada) [46]. All experiments were performed in conformity with the Canadian Tri-Council Policy Statement, as described in the Ethical Conduct for Research Involving Humans and institutional CHU de Québec - Laval University guidelines. The Ethics Committee of the CHU de Quebec Research Center approved the protocol (CER #S11–03–168).

The effect of different materials/surface conditions on the HUVECs and HUASMCs viabilities was assessed by direct viability assays. HUVECs and HUASMCs were tested individually. Standard culture plates were used as control (CTRL). Cells in culture media were seeded onto the sterile samples (20,000 cells/cm²) and incubated at 37 °C in a saturated atmosphere (5 vol.% CO₂). After 1, 3 and 7 days, the medium was removed, and the cells were incubated for 4 h with a 1X resazurin sodium salt solution. The resorufin product obtained was collected and fluorescence intensity at a wavelength of 545 nm_ex/590 nm_em was measured with a SpectraMax i3x Multi-Mode Plate Reader (Molecular Devices, California, USA). Twelve measurements (four for each of the three randomly selected samples) were performed for each condition at each time point.

2.4.2. Hemocompatibility assay

Hemolysis test, clotting time assay and platelets adhesion test were performed to assess the hemocompatibility of the materials/surface conditions studied in the present work. Whole human blood from a healthy donor, collected in citrate-containing blood collection tubes, was used for the hemocompatibility tests.

For the hemolysis test, PBS 1X and de-ionized water were used as negative (CTRL Neg) and positive (CTRL Pos) control, respectively. Each sample was placed in a 15 mL tube and 10 mL of sterile PBS 1X was added to each tube. They were incubated at 37 °C for 30 min. After incubation, 200 µL of diluted citrated blood (4 parts of citrated blood to 5 parts of PBS 1X) were added to each tube, carefully mixed by inversion and incubated at 37 °C for 1 h. After 30 min, the tubes were carefully mixed by inversion. After incubation, the tubes containing the controls and the samples were subjected to a centrifugation step at 800 g for 5 min. 100 µL aliquots of the supernatant were collected and placed in a 96-well plate. The absorbance (OD) at a wavelength of 540 nm was measured with a SpectraMax i3x Multi-Mode Plate Reader. Twelve measurements (four for each of the three randomly selected samples) were performed for each condition. Hemolysis was calculated according to the following equation:

$$\mathit{Relative}\mathit{hemolysis}=\frac{OD\mathit{sample}-OD\mathit{CTRL}\mathit{Neg}}{OD\mathit{CTRL}\mathit{Pos}-OD\mathit{CTRL}\mathit{Neg}}$$ (1)

where OD sample is the sample absorbance, OD CTRL Neg is the negative control absorbance and OD CTRL Pos is the positive control absorbance.

For the clotting time assay, the samples were placed in the wells of 12-well multi plates. One plate was used for each of the six evaluated time points (0, 5, 10, 15, 20 and 40 min). Culture-treated plastic was used as control. 100 µL of citrated blood were placed on the surfaces of the different samples. Immediately afterwards, 20 µL of calcium chloride (CaCl₂) were added to the blood in order to activate the coagulation cascade. Samples were incubated at 37 °C for the selected time points. At each time point, 2 mL of de-ionized water were added to each sample in order to lysate the erythrocytes not entrapped in the blood clot formed. The aqueous solutions containing the free hemoglobin were then transferred to a 96-well plate. The absorbance at a wavelength of 540 nm was measured with a SpectraMax i3x Multi-Mode Plate Reader. Three samples were tested for each time point for each condition.

For the platelet adhesion test, the human blood collected was transferred to 15 mL tubes, which were centrifuged at 1000 rpm for 10 min in order to separate the plasma from the corpuscular fraction. The platelet-rich plasma was collected and placed in fresh tubes. Platelets number was assessed by counting using the Bioanalytic GmbH Thrombo-TIC kit (Freiburg, Germany) following the supplier's instructions. The platelet-rich plasma was diluted in PBS 1X in order to achieve a final concentration of 3 × 10⁸ platelets/mL. 100 µL of plasma solution were pipetted on each sample (n = 3 for each condition), which were incubated at 37 °C in a cell culture incubator for 1 h. After incubation, samples were rinsed three times with PBS 1X, fixed in glutaraldehyde 1% solution (Sigma-Aldrich, Ontario, Canada) at room temperature for 30 min, and finally rinsed three times with de-ionized water. They underwent a dehydration procedure, comprising immersion in solutions with increasing concentrations of alcohol (20%, 50%, 90% and 100%); they were immersed twice in each solution for 5 min. Samples were then sputtered with gold for SEM analyses in a FEI QUANTA 250 (Oregon, USA) scanning electron microscope. The microscope was operated in high-vacuum mode with an acceleration voltage of 15 kV.

2.5. Statistical analyses

Statistical analyses were performed using GraphPad software (California, USA). Comparison between groups was evaluated by one-way ANOVA with post hoc Tukey test to correct for multiple comparisons. Significance was retained when p < 0.05.

3. Results

3.1. Surface characterization

Fig. 1(a) and (b) show SEM micrographs at 100 x and 2000 x magnification, respectively, of 316 L EP, Ti-CP MP, TMF MP and TMF EP. Flat surfaces with smooth finishing were obtained for all of them. Grains and grain boundaries were clearly visible for all conditions, with smaller grain size for 316 L and Ti-CP in comparison to TMF alloy. When comparing Ti-CP MP and TMF MP, prepared by the same mechanical polishing procedure, a smoother (qualitatively seen in Fig. 1 and quantitatively confirmed in Fig. S1 of the supplementary data document) and more homogeneous surface was observed for the TMF alloy. The clear distinction of grains and the smooth surfaces obtained by mechanical polishing using a mixture of colloidal silica and hydrogen peroxide have already been reported in the literature for pure Ti [40]. When comparing TMF MP and TMF EP surfaces, the only difference noticed in the SEM micrographs was the presence of small cavities in TMF EP (whiter spots in Fig. 1(a), better visualized in Fig. 1(b)). They may be pitting formed during the electropolishing process due to gas development [47,48].

Fig. 1.

SEM micrographs of electropolished 316 L stainless steel (316 L EP), mechanically polished pure titanium (Ti-CP MP), mechanically polished Ti-8Mo-2Fe alloy (TMF MP) and electropolished Ti-8Mo-2Fe alloy (TMF EP) at (a) 100 x and (b) 2000 x magnifications. Two-dimensional AFM images of these material surfaces in scan sizes of 50 × 50 µm² (c), 20 × 20 µm² (d), 5 × 5 µm² (e) and 1 × 1 µm² (f).

Fig. 1 (c-f) show typical two-dimensional AFM images of the conditions studied at scan sizes of 50 × 50, 20 × 20, 5 × 5 and 1 × 1 µm². For 316 L EP, alloy microstructure features such as grain boundaries (Fig. 1 (d-f)) and twins (Fig. 1(d)) were evident in the two-dimensional AFM images. The presence of small holes, which are possibly pitting, was also observed on the surface of 316 L EP (darker spots in Fig. 1(c) and (d), also seen in the SEM image in Fig. 1(b)). For Ti-CP MP, in the larger scale acquisitions, 50 × 50 µm² (Fig. 1(c)) and 20 × 20 µm² (Fig. 1(d)), some grains were visible and exhibited greater height when compared to their surroundings. For TMF MP, a smooth surface was observed, even for larger scale acquisitions; only some traces corresponding to shallow scratches coming from polishing procedure were present. For TMF EP, a wavy-like surface was observed at the scan size of 50 × 50 µm². For the smaller scale acquisitions, 5 × 5 and 1 × 1 µm², a ripple-like structure was observed. Similar ripple-like structures were reported by Asgari et al. [49] on pure Ti surface after electropolishing with a different acidic electrolyte from the one used in the present work for TMF alloy.

R_q roughness values measured from larger (50 × 50 and 20 × 20 µm²) and smaller (5 × 5 and 1 × 1 µm²) scale acquisitions are shown in Fig. S1 (a) and (b) of the supplementary data document, respectively. For the largest scan size (50 × 50 µm²), TMF EP exhibited the roughest surface (R_q = 13.9 ± 0.9 nm), which may be a result of its wavy-like “primitive” surface morphology. Thus, TMF EP condition showed a rougher surface than 316 L EP (R_q = 7.6 ± 2.7 nm for 50 × 50 µm² scan size). For this scan size, no significant differences were observed between the roughness of Ti-CP MP (R_q = 5.3 ± 1.8 nm) and TMF MP (3.2 ± 1.6 nm) surfaces. However, for a scan size 20 × 20 µm², Ti-CP MP showed a higher roughness (R_q = 3.7 ± 2.1 nm) than TMF MP (R_q = 1.1 ± 0.1 nm), which may be explained by the presence of grains with a height greater than their surroundings in Ti-CP MP, as observed in the two-dimensional AFM images (Fig. 1(d)). For smaller scan sizes (5 × 5 and 1 × 1 µm²), TMF EP also exhibited the roughest surface (R_q = 2.4 ± 0.3 nm for 5 × 5 µm² and R_q = 2.1 ± 0.3 nm for 1 × 1 µm²), which may be a result of its ripple-like structure. The other materials/surface conditions showed roughness R_q lower than 1.4 nm.

Static contact angle (CA) measurements taken for all conditions are shown in Fig. S1 (c) of the supplementary data document. All surfaces were hydrophilic (between 0° and 90°) [20]. TMF MP exhibited the lowest CA value (73 ± 2°), while Ti-CP MP exhibited the highest CA value (86 ± 5°). No significant difference was observed between the CA value of TMF EP and 316 L EP surfaces, both being 78°

XPS measurements were carried out to evaluate the chemical composition of the surfaces. The survey spectra and atomic concentrations are shown in Fig. 2(a) and (b). Significant amounts of oxygen and carbon were detected in all samples. The Ti-CP MP, TMF MP and TMF EP samples exhibited similar amount of oxygen (17.2–18.2 at.%). The carbon content on these surfaces was around 10–13 at.%. The 316 L EP sample showed higher concentrations of oxygen (41.3 at.%) and carbon (41.8 at.%). Titanium was present on the surfaces of Ti-CP MP, TMF MP and TMF EP; the concentration was higher on the Ti-CP MP surface (71.9 at.%) when compared to the TMF surfaces (66.8–66.9 at.%). The TMF MP and TMF EP surface layers also contained a small amount of molybdenum (1.2–1.3 at.%). No iron was detected on the TMF surfaces. In addition to carbon and oxygen, the 316 L EP sample showed 10.1 at.% chromium, 3.8 at.% iron, and a small amount of molybdenum (0.8 ± 0.1 at.%) on its surface. A small amount of fluorine, 1.0 ± 0.2 at.%, was observed on both 316 L EP and TMF EP surfaces. Its presence is attributed to the use of hydrofluoric acid in the acid dipping and electropolishing step of the 316 L EP and TMF EP sample preparation, respectively.

Fig. 2.

XPS survey spectra (a) and atomic concentrations (b) for all surfaces. * p < 0.05. High-resolution (c-e) O1s, (f-h) Ti2p and (i-k) Mo3d spectra of Ti-CP MP, TMF MP and TMF EP, respectively.

To evaluate the amount of each carbon species on the surface, high-resolution C1s spectra were analyzed, and the results are shown in Fig. S2 of the supplementary data document. The C1s peak for all surfaces was formed by four components. The first component, in the range of 284.8–285.2 eV, could be attributed to the presence of alkyl carbon (C—C, C—H) [47]. The second, in the range of 286.1–286.7 eV, was related to the presence of alcohol (C—OH), ester (C—O-C) and/or C—O-R groups [47]. The third (287.5–288.8 eV) and fourth (288.8–290.3 eV) could be attributed to the presence of C = O and O—C = O groups, respectively [50]. The alkyl component was the main one for all conditions studied, accounting for 81.9, 74.2, 56.8 and 73.9 at.% of the total C found on the 316 L EP, Ti-CP MP, TMF MP and TMF EP surfaces, respectively. The atomic concentration of the C—OH/C—O-C/C—O-R components on these surfaces was of 7.3, 13.6, 31.5 and 14.6%, respectively. For all surfaces, the C = O and O—C = O components together accounted for less than 12.2 at.% of the total C.

The high-resolution O1s spectra of Ti-CP MP, TMF MP and TMF EP are shown in Fig. 2 (c-e), while that of 316 L EP sample is presented in Fig. S3 (a) of the supplementary data document. All O1s spectra were fitted by three components associated with metallic oxides (located at 529.9–530.2 eV) [51], hydroxides and/or hydroxyl compounds (located at 530.8–531.5 eV) [51] and adsorbed water (located at 532.5–533.4 eV) [47,50]. The main component for all titanium surfaces was the metallic oxides, accounting for 50.6, 56.2 and 46.5 at.% of the total oxygen found on the Ti-CP MP, TMF MP and TMF EP surfaces, respectively. The TMF EP surface presented higher atomic concentration of hydroxides/hydroxyl groups (41.6%) when compared to the Ti-CP MP (35.1%) and TMF MP (30.3%) surfaces. The adsorbed water accounted for 14.3, 13.5 and 11.9 at.% of the total oxygen found on the Ti-CP MP, TMF MP and TMF EP surfaces, respectively. The atomic percentages of oxygen corresponding to metallic oxides, hydroxides and/or hydroxyl groups and adsorbed water on the 316 L EP surface were, respectively, 43, 47.1 and 9.9%.

The Ti2p spectra (Ti2p_3/2 and Ti2p_1/2) of Ti-CP MP, TMF MP and TMF EP (Fig. 2 (f-h)) were fitted by components associated with metallic Ti (located at 453.4 eV-453.6 eV for Ti2p_3/2 and at 459.2–459.4 eV for Ti2p_1/2), TiO (located at 455.5 eV-455.7 eV for Ti2p_3/2 and at 461.0–462.0 eV for Ti2p_1/2), Ti₂O₃ (located at 457.2 eV for Ti2p_3/2 and at 462.6–462.9 eV for Ti2p_1/2) and TiO₂ (located at 458.8 eV for Ti2p_3/2 and at 464.5–464.6 eV for Ti2p_1/2) components [51]. The same components have been reported in the literature for an electropolished titanium surface [52] and for mechanically polished pure titanium and β Ti-29Nb-13Ta-4.6Zr alloy [53]. The concentration of Ti components differed slightly between samples. The main Ti component for all surfaces was TiO₂, accounting for 67.8–70.2 at.% of the entire Ti2p peak area. Metallic Ti was the second main component (18.7–20.1 at.%), followed by Ti₂O₃ (7.9–9.1 at.%) and TiO (2.8–3.6 at.%).

The Mo3d spectra of Ti-CP MP, TMF MP and TMF EP samples are shown in Fig. 2 (i-k). The Mo3d spectra of TMF MP and TMF EP (Fig. 2 (j-k)) were fitted by components associated with metallic Mo (located at 227.0 eV-227.1 eV and at 229.8–230.4 eV) [38,54], MoO₂ (located at 228.6 eV-229.1 eV and at 231.5–232.3 eV) [38,55] and MoO₃ (located at 233.1 eV-234.1 eV and at 235.0–236.2 eV) [38]. The concentration of metallic Mo in TMF MP and TMF EP was 37.3 at.% and 33.5 at.%, respectively. The most notable differences were related to MoO₂ and MoO₃ concentrations. The atomic percentage of Mo corresponding to MoO₂ was lower in TMF MP (33.2%) when compared to TMF EP (48.1%). The opposite was observed regarding the atomic percentage of MoO₃ (29.5% for TMF MP and 18.5% for TMF EP).

The high-resolution Cr2p and Fe2p spectra corresponding to 316 L EP samples are shown in Fig. S3 (b) and (c) of the supplementary data document. Details on peak components can also be found in the supplementary data document. Although a small amount of metallic Cr and Fe were detected in the 316 L EP samples, these elements were mainly present in the oxidized state. Cr was mostly present as Cr₂O₃, while Fe was present as FeO and Fe₂O₃. The presence of Cr(OH)₃ and FeOOH was also observed. These results were in agreement with the work of Wang and co-workers [56], for instance.

3.2. Electrochemical characterization

Fig. 3(a) shows representative potentiodynamic polarization curves recorded after 2 h immersion in PBS. The electrochemical properties obtained from the potentiodynamic tests, including open circuit potential (OCP), corrosion current density (I_corr), calculated corrosion rate (C_rate) and pitting potential (E_pit) are shown in Table S2 of the supplementary data document. Regarding OCP values, no significant difference was observed between TMF MP and Ti-CP MP, whose trend was significantly nobler than the TMF EP and 316 L EP ones (similar OCP values).

Fig. 3.

(a) Potentiodynamic polarization curves recorded after 2 h immersion in PBS. (b-c) Bode EIS plots (-phase angle vs. log Frequency in (b) and log |Z| (impedance modulus) vs. log Frequency in (c)), (d) Nyquist EIS plots, and (e) the equivalent electric circuits used to fit the impedance data.

From the anodic branch of the PDP curves (Fig. 3(a)), for all conditions the current density stabilized when the scanning potential value exceeded 0.65 V, which indicates a full passivation [57]. No difference was observed between them in terms of passive current density (I_pass). For Ti-CP MP, TMF MP and TMF EP, an extensive passive plateau was observed in their PDP curves up to 1.7 V. Therefore, no pitting corrosion and transpassivation occurred in these materials [58]. However, for 316 L EP, the current density started to increase when the potential value exceeded 1 V. This increase became much faster when the potential value exceeded 1.3 V, suggesting the breakdown of the passive layer and the occurrence of pitting corrosion [58]. Thus, the pitting potential (E_pit) for 316 L EP was of 1312 ± 12 mV. No significant difference was observed between the materials/surface conditions studied in the present work in terms of corrosion current density (I_corr) and corrosion rate (C_rate), as observed in Table S2.

The results obtained from electrochemical impedance spectroscopy are shown in Fig. 3 (b-e). Bode plots, phase angle versus frequency and impedance modulus versus frequency, are shown in Fig. 3(b) and (c), respectively. Fig. 3(d) and (e) show, respectively, the Nyquist plots (imaginary impedance versus real impedance) and the equivalent circuits used to fit the Nyquist plot data. For all the conditions, at high frequency, the phase angle was close to zero degrees (Fig. 3(b)) and the impedance modulus was 10 Ω.cm² (Fig. 3(c)), meaning that, in this frequency range, the impedance was governed by the electrolyte resistance [59] (see the values of R_s, electrolyte resistance, in Table S3 of the supplementary data document).

From Fig. 3(b), 316 L EP and Ti-CP MP were found to be the systems with a time constant (only one maximum in their Bode-phase diagrams). In the intermediate and low frequency ranges, a plateau was observed in their Bode-phase diagrams at a phase angle around −80 and −85° (Fig. 3(b)), respectively; on the other hand, the impedance modulus increased linearly (slope close to 1) with decreasing frequency (Fig. 3(c)). This behavior was typical of near-capacitive materials [23,59,60]. Moreover, the impedance modulus values in the range of 10⁵ Ω.cm² (Fig. 3(c)) at low frequency were typical of a compact oxide layer with high corrosion resistance [61]. The EIS data of 316 L EP and Ti-CP MP were satisfactorily fitted using a simple Randles’ circuit (Fig. 3(e)), composed by R_s (resistance of the electrolyte), and by R₁ and Q₁ which are, respectively, the resistance and the capacitance of the compact oxide layers on the surfaces of 316 L EP and Ti-CP MP. A constant phase element (CPE) was used instead of a pure capacitor. This was commonly done for passive oxide films on the surface of metallic materials, since the CPE model took into account the heterogeneities present in these films such as defects and roughness [57,59,61]. The fitted electrical parameters, including the exponent n that indicates the deviation of the CPE from the ideal capacitive behavior, are shown in Table S3.

For TMF MP and TMF EP, two time constants (two maxima in their Bode-phase diagrams) were clearly observed in Fig. 3(b), one in the high-intermediate frequency range and one in the low frequency range. Similar to 316 L EP and Ti-CP MP, the high impedance modulus values (10⁵ Ω.cm²) at low frequency, for both TMF MP and TMF EP, are typical of a compact oxide layer with high corrosion resistance [61]. Therefore, the time constant in the low frequency range was due to the impedance processes occurring in this compact inner layer. The additional time constant in the high-intermediate frequency range could be due to the processes occurring in a porous outer oxide layer, which has already been reported in the literature for other β Ti-Mo alloys, such as Ti-12Mo, Ti-20Mo and Ti-40Mo [61]. Due to the existence of two time constants, the EIS data of TMF MP and TMF EP were satisfactorily fitted using an equivalent circuit composed of two RC components apart from the electrolyte resistance (R_s) (Fig. 3(e)). In this circuit, R₁ and Q₁ represent the resistance and the capacitance of the compact inner oxide layer, and R₂ and Q₂ represent the resistance of the solution inside the pores and the capacitance of the porous outer oxide layer. Similar to 316 L EP and Ti-CP MP, CPEs were used instead of pure capacitors and the fitted parameters are shown in Table S3.

Summarizing the data shown in Table S3, the compact inner oxide layer on TMF EP was as protective as the compact oxide layer on Ti-CP MP; no significant difference was observed between their R₁ values. TMF MP exhibited, however, the lowest R₁ value among the conditions studied, indicating that its compact inner oxide layer was the least protective. On the other side, 316 L EP exhibited the most protective compact oxide layer (highest R₁ value) among all conditions. Although TMF EP had a more protective compact inner oxide layer than TMF MP (resistance R₁ of TMF EP almost twice that of TMF MP), the resistance of the solution inside the pores of its outer oxide layer (R₂) was lower than that of TMF MP.

3.3. In vitro biological characterization

Fig. 4(a) and (b) show the results of the direct viability assay using, respectively, HUVECs and HUAMSCs. The results were normalized against the day 1 CTRL condition.

Fig. 4.

Direct viability assay results using HUVECs (a) and HUAMSCs (b). Cell viability was measured after 1, 3 and 7 days of incubation through resazurin salt solution assays. The graphs show the mean fluorescence ± standard deviation of resorufin product recorded from HUVECs (a) and HUAMSCs (b) treated with the different experimental conditions. Results were normalized against the day 1 CTRL condition. * p ≤ 0.05.

Regarding the HUVECs viability (Fig. 4(a)), after 1 day of incubation, the CTRL and Ti-CP MP conditions showed higher viability compared to 316 L EP, TMF MP and TMF EP. No significant difference was detected between CTRL and Ti-CP MP. Moreover, after 1 day of incubation, TMF EP showed higher viability compared to TMF MP; both showed similar viability compared to 316 L EP. After 3 days of incubation, CTRL and Ti-CP MP still exhibited higher viability than 316 L EP, TMF MP and TMF EP. However, after this incubation period, Ti-CP MP exhibited lower viability compared to the CTRL condition. Moreover, after 3 days of incubation, both TMF MP and TMF EP showed lower viability compared to 316 L EP. After 7 days of incubation, once again, the CTRL condition showed higher viability than all materials/surface conditions studied. Additionally, Ti-CP MP, TMF MP and TMF EP exhibited lower viability than 316 L EP. No significant differences were observed between TMF MP and Ti-CP MP after 7 days of incubation. TMF MP showed higher viability than TMF EP. Moreover, for TMF MP, as well as for CTRL and 316 L EP, the HUVECs viability increased linearly from day 1 to day 7, which was not the case for TMF EP and Ti-CP MP.

Regarding the HUASMCs viability (Fig. 4(b)), after 1 day of incubation no significant differences were detected between TMF MP and TMF EP, which exhibited lower viability when compared to CTRL, 316 L EP and Ti-CP MP. After 3 days of incubation, TMF MP still showed lower viability compared to CTRL, 316 L EP and Ti-CP MP. Moreover, its viability was lower compared to TMF EP. After 7 days of incubation, CTRL, Ti-CP MP, TMF MP and TMF EP showed lower viability compared to 316 L EP.

The hemocompatibility assay results are shown in Fig. 5. Fig. 5(a) shows the hemolysis test, Fig. 5(b) the clotting time assay and Fig. 5 (c-e) the platelets adhesion test results. The relative hemolysis of each condition was lower than 0.01, therefore, all the alloys studied were not hemolytic. In fact, for a material to be considered hemolytic, it must induce hemolysis higher than 5% [62]. No significant differences were observed between the alloys in terms of hemolysis.

Fig. 5.

Hemolysis test (a) and clotting time assay results (b). The graphs show the mean absorbance ± standard deviation recorded from the different experimental conditions. Results were normalized against the CTRL condition. * p ≤ 0.05. (c-e) SEM micrographs of platelets adhered to the surface of the different experimental conditions after 1 h of contact with blood. (c), (d) and (e) correspond to 500 x, 2000 x and 5000 x magnification, respectively.

Regarding the clotting time assay results (Fig. 5(b)), immediately after the addition of calcium chloride to activate the coagulation cascade (0 min time point), as expected, the blood did not coagulate. Moreover, no significant differences were observed between the conditions. Therefore, the value at this time point was used as reference for the test (100% relative hemocompatibility or 100% of free hemoglobin, that is, maximal hemocompatibility). After 5 min of incubation, the free hemoglobin amount in TMF MP was lower than in CTRL and 316 L EP. After 10 min of incubation, no significant difference was observed between the conditions. After 15 min, the free hemoglobin amount in TMF EP was higher than in 316 L EP. After 20 and 40 min, the free hemoglobin amount in Ti-CP MP was lower compared to CTRL, 316 L EP, TMF MP and TMF EP. Therefore, the clotting time results suggested that TMF, with MP and EP surfaces, exhibited better hemocompatibility (thromboresistance) compared to Ti-CP MP. In other words, the clot formation was delayed on both TMF surfaces when compared to Ti-CP MP.

From the micrographs of the platelets adhered to the surface of the different experimental conditions (Fig. 5 (c-e)), a higher amount of platelets adhered to the surface of Ti-CP MP was observed when compared to the other materials. Moreover, many aggregations of adhered platelets with pseudopods were observed on the surface of Ti-CP MP. In TMF MP, TMF EP and 316 L EP, the adhered platelets seemed to be more dispersed and without pseudopods, suggesting that platelets were less activated on these surfaces when compared to Ti-CP MP, corroborating the results of the clotting time.

4. Discussion

4.1. Electrochemical behavior

For permanent metallic endoprosthesis such as stents, the biocompatibility of the material is determined in part by its corrosion resistance. In other words, the more corrosion resistant, the more biocompatible the metallic material [63,64].

It was observed, in the present work, that TMF MP and Ti-CP MP exhibited significantly nobler OCP compared to TMF EP and 316 L EP. The nobler OCP values of MP surfaces might be related to the addition of hydrogen peroxide to the polishing slurry. Indeed, H₂O₂ is a strong oxidizer that has been shown to enhance the formation of the passivation layer on the surface of Ti-based materials when added to polishing slurries [41,65]. Another evidence suggesting that H₂O₂ was responsible for the nobler OCPs of MP samples is that the OCP value (measured under the same conditions as in the present study), reported by Sotniczuk et al. [57] for mechanically polished grade 2 pure titanium (without addition of H₂O₂ to colloidal silica), was −180 mV. The OCP value found by these authors was much less noble than that obtained for Ti-CP MP (−33 mV).

The potentiodynamic polarization results indicated that the corrosion resistance in the PBS solution of the newly developed TMF alloy with mechanically polished surface (MP) and electropolished surface (EP) was similar to that of mechanically polished pure titanium (Ti-CP MP). XPS analyses showed that the main titanium oxide present on all these surfaces was TiO₂, known to be resistant to corrosion [33,39], more than Ti₂O₃ [39], which was also present on their surfaces, but in smaller atomic fraction. Moreover, the corrosion rates of Ti-CP MP, TMF MP and TMF EP were similar to that of 316 L EP. In general, the corrosion rate of 316 L stainless steel is higher in comparison to Ti-based materials [63]. The electropolishing process applied to 316 L stainless steel could explain this similarity. In fact, this procedure, especially when combined with a post acid dipping treatment, has been shown to improve the corrosion resistance of 316 L for several reasons [34]. Firstly, electropolishing plus acid dipping causes the enrichment of the surface with Cr and the increase of Cr/Fe atomic concentration ratio [34,44] (calculated at 2.7 for the 316 L EP surface in the present study). Chromium oxide is known among the metallic oxides formed on the surface of 316 L as responsible for the corrosion resistance of this material [34]. Additionally, the electropolishing procedure results in smoother surfaces (even smoother when a post acid dipping treatment is applied), improving the corrosion behavior [34]. Finally, a thicker oxide layer is formed on the surface of 316 L through electropolishing, resulting in higher corrosion resistance [35,37].

Moreover, the PDP results showed that the pitting resistance of the TMF alloy (both MP and EP) in PBS solution, similarly to Ti-CP MP, was superior to that of 316 L EP. In fact, for 316 L EP, the breakdown of the passive layer and pitting corrosion occurred, while no pitting was observed for Ti-based materials. The passive oxide layers formed in Ti-based materials are known to be very protective [63]. In general, no breakdown and pitting potentials are observed in their PDP curves in saline physiological solutions (such as PBS) [58,63]. This is not the case for 316 L stainless steel, for which breakdown of the passive layer and pitting corrosion in saline solutions are generally reported [34,58,63]. Therefore, the results obtained are in agreement with the literature.

The EIS results of the TMF alloy (with both MP and EP surfaces) are in agreement with the general assumption that two layers form the passive films on most surfaces of titanium-based materials: an outer one, which is porous, and an inner one, which is more compact [59,61]. The existence of only one time constant in the Bode-phase diagram of Ti-CP MP does not necessarily mean that the passive film on this surface was formed only by a compact oxide layer. In fact, the porous layer contribution to the overall electrochemical behavior can be very small, with the impedance response being dominated by the response of the compact inner oxide layer [59,66]. Thus, it is reasonable to think that the passive film on the Ti-CP MP surface had indeed a duplex structure; however, the outer porous layer had a negligible contribution to the impedance response of this material. The resistance of the inner compact oxide layer on the Ti-based materials surfaces studied in the present work (R₁ values in Table S3) was of the same order of magnitude, ∼10⁵ Ω.cm², reported for other Ti-Mo-based alloys [28,59]. The Bode-phase diagram of 316 L, electropolished with electrolytes different from the present study, in PBS [67] and in NaCl saline solution [35], presented two time constants, differing from the results of the present work. Even though only a single layer formed the passive film on 316 L EP, its corrosion resistance was of the same order of magnitude (10⁶ Ω.cm²) that was reported in [67].

Although no significant difference was observed between the conditions studied in terms of corrosion rate through PDP analysis, from the EIS results the compact oxide layer on 316 L EP was found to exhibit the highest general corrosion resistance. As already mentioned, the electropolishing process of 316 L increases its general corrosion resistance due to the formation of a more compact and uniform oxide layer with an increased Cr/Fe ratio [67]. Likewise, the electropolishing procedure applied to the TMF alloy was shown to improve its general corrosion resistance. Its total resistance was almost twice that of TMF MP. A mixture of titanium and molybdenum oxides formed the oxide layer on the TMF surfaces, with TiO₂ being the prevalent oxide and the main responsible for the corrosion resistance. XPS results showed that the atomic percentage of TiO₂ was similar on the MP and EP surfaces. The main difference in chemical composition was related to the ratio between the amount of oxygen corresponding to the hydroxides/hydroxyl groups and the amount of oxygen corresponding to metallic oxides (OH⁻/O_{Metallic oxides} ratio). From O1s deconvolution, it was observed that, for the TMF MP surface, the OH⁻/O_{Metallic oxides} ratio was lower (0.5) than that of the TMF EP surface (0.9). In other words, on the MP surface, the percentage of O forming metallic oxides was twice the percentage of O forming hydroxyl groups. As described in the materials and methods section, the final step of mechanical polishing was performed using a mixture of colloidal silica and hydrogen peroxide. In the presence of hydrogen peroxide, a strong oxidizer, the conversion of titanium atoms to titanium dioxide is faster [36]. The same can be rationalized for the conversion of molybdenum atoms into molybdenum oxides. This could explain the higher percentage of O forming metallic oxides than OH⁻ on the TMF MP surface.

In addition to the chemical composition, the thickness of the oxide layer can affect the corrosion resistance. Although the thickness was not evaluated in the present study, it should not be higher than several nanometers for all surfaces studied. Apart from the oxides, the metallic forms of all elements (Ti for all Ti-based materials, Mo for the TMF surfaces and Cr and Fe for 316 L EP) were also detected through XPS analysis, indicating a small thickness of the passive layers [57]. According to the literature, the electropolishing of a β Ti-13Nb-13Zr alloy [39], for example, resulted in an oxide layer thickness of 89 nm. On the other hand, the thickness of the oxide layer formed on the surface of pure titanium after mechanical polishing with colloidal silica without and with addition of hydrogen peroxide was estimated to be around 2 nm [57] and 7 nm [40], respectively. According to Okawa et al. [40], the polishing of pure titanium with slurries containing hydrogen peroxide resulted in a greater weight loss due to both TiO₂ removal and dissolution (TiO₂ + OH⁻ → HTiO₃⁻). Therefore, it is reasonable to think that the oxide layer on the TMF MP surface is thinner than that of TMF EP, contributing to its lower resistance. Moreover, hydrogen peroxide has been shown to reduce the charge transfer resistance of pure titanium and of a β Ti-29Nb-13Ta-4.6Zr alloy [57]. It is noteworthy that, in that study, hydrogen peroxide was added to the PBS solution to simulate inflammation conditions during PDP and EIS experiments [57], and not to the last step of the polishing procedure as in the present study.

The EIS results showed that the compact oxide layers on the TMF surfaces were less corrosion resistant in PBS than those on Ti-CP MP. The same has been reported in the literature for a β Ti-29Nb-13Ta-4.6Zr (TNTZ) alloy [57]. According to Sotniczuk et al. [57], pure titanium should present a higher tendency to oxidation and to the formation of a stable passive oxide layer, because the metal-metal bond strength in this material is lower than in β Ti alloys such as TNTZ and TMF. The alloying elements atoms in these β Ti alloys tend to create a strong covalent bond with their neighbors by sharing d-level electrons [57]. The lower coordination number for the bcc structure (β phase, main phase in TNTZ and TMF) in comparison to the hcp structure (α phase that constitutes pure titanium) and the higher electron to atom ratio (e/a) of TMF alloy (e/a = 4.15) than pure titanium (e/a = 4) result in this effect [57]. Moreover, according to Sotniczuk et al. [57], refined grains (higher density of grain boundaries) accelerate the passivation and promotes a more compact passive layer. Therefore, the smaller grains in Ti-CP MP could have resulted in a more compact oxide layer in comparison to the TMF alloy, contributing to its higher corrosion resistance.

Therefore, although the EIS results showed that the TMF alloy (MP and EP) had lower general corrosion resistance than pure titanium and 316 L stainless steel, no significant difference was observed regarding the corrosion rate through PDP analyses. Moreover, TMF alloy has an advantage over 316 L stainless steel for biomedical applications since, by resisting to pitting corrosion in physiological solutions up to high potentials (1.7 V), it prevents metal ions release and cracking that are prone to occur in pitting sites [58].

4.2. In vitro biological behavior

Cell viability

For blood contact implants such as cardiovascular stents, the ideal biomaterial would be rapidly covered by the cardiovascular endothelium once the device is deployed in the damaged artery. In other words, the ideal biomaterial would not inhibit or, better still, would promote endothelialization on its surface. This would reduce the risk of implant failure by preveting thrombus formation, inflammation and intimal hyperplasia (caused by the migration and proliferation of smooth muscle cells) [22]. Therefore, the evaluation of the capacity of a biomaterial developed for stents to support HUVECS and HUASMCs viability and proliferation is of great importance.

The direct cell viability results of day 1 provided information on adhesion (initial attachment) of cells to surfaces, while further evaluations over time (after 3 and 7 days of incubation) gave hints on cell proliferation. When comparing the materials/surface conditions on day 1, it was observed that the adhesion and viability of endothelial cells on TMF surfaces (MP and EP) were similar to those for 316 L stainless steel, but worse compared to those for pure titanium and the control condition. Several surface aspects of biomaterials, such as roughness, surface energy (wettability), chemical composition and even grain size, can affect adhesion, viability and proliferation of endothelial cells [29,68,69]. Although it is difficult to directly point out the surface aspects that may justify the worse adhesion and viability of endothelial cells on day 1 on TMF surfaces compared to pure titanium, some considerations can be made. The main difference in chemical composition between the TMF surfaces and the pure titanium surface was the presence of Mo in the former, as observed by the XPS analyses. Eisenbarth et al. [29] reported that the viability and proliferation of bovine endothelial cells on pure molybdenum surfaces were compromised, and they suggested that this metal should be used in small amounts as alloying element in Ti-based biomaterials [29]. Thus, the presence of Mo on the TMF surfaces might have contributed to the worse adhesion and viability (on day 1) of endothelial cells and even smooth muscle cells in comparison to pure titanium. Moreover, a study performed for 316 L stainless steel showed that the grain size affected endothelial cells adhesion; smaller grains favored adhesion due to an increased grain boundary area [69]. The smaller grains of pure titanium in comparison to the TMF alloy (SEM images in Fig. 1) might have contributed to the improved adhesion of endothelial cells to its surface. The higher roughness of TMF EP and its ripple-like structure observed in small-scale AFM acquisitions (5 × 5 and 1 × 1 µm²) might have contributed to the improved adhesion and viability of endothelial cells on its surface on day 1 in comparison to TMF MP [68]. However, the positive effect of these topographical features on endothelial cells adhesion was not enough to overcome the negative impact of the presence of Mo and the larger grain size of TMF EP, in comparison to pure titanium.

Although the adhesion of endothelial cells was improved on the pure titanium surface, a weak proliferation was observed after 3 days of incubation, which was not maintained after 7 days (the viability on day 7 was statistically the same as on day 1 and day 3). Interestingly, for the TMF alloy, although the viability of endothelial and smooth muscle cells was not high on day 1, it increased with time, equaling that of pure titanium on day 7. However, the results showed a decreased proliferation rate of endothelial cells on TMF and pure titanium surfaces compared to 316 L stainless steel, which would be a disadvantage for the envisaged application. Chai et al. [70] reported that endothelial cell proliferation on the pure titanium surface was improved compared to 316 L stainless steel. These contradictory results could be due to differences in the endothelial cells line tested and/or in the surface polishing method. Similar to endothelial cells, the proliferation rate of smooth muscle cells on TMF and pure titanium surfaces was reduced compared to 316 L stainless steel, which could be considered as an advantage for stents. It would be interesting to perform co-culture tests [68] or in vivo tests in order to evaluate the competition between endothelial and smooth muscle cells, and also perform analyses to evaluate the function of the endothelial cells (through the nitric oxide release test, for example [6]) grown on the materials/surface conditions studied.

Hemocompatibility

The result of hemolysis test showed that, like pure titanium and 316 L stainless steel, the TMF alloy (both MP and EP) met one of the requirements for blood contact implants such as stents, i.e., it did not cause acute hemolysis. Pure titanium [62,71] and other β Ti alloys [71] have also been reported not to cause hemolysis.

The TMF alloy, in both MP and EP conditions, is generally more hemocompatible than pure titanium, which is an advantage for blood contact implants. It exhibited delayed coagulation in comparison to pure titanium. Moreover, platelets on TMF surfaces were less aggregated and activated than on the pure titanium surface. Several aspects affect hemocompatibility, such as surface chemical composition, roughness, wettability and the crystallinity of TiO₂ [72,73]. The crystallinity of TiO₂, the main oxide on the surfaces of the Ti-based materials studied, was not evaluated in the present study. According to Hung et al. [74], amorphous TiO₂ is formed on the surface of pure titanium after mechanical polishing. Therefore, on both the TMF MP and Ti-CP MP surfaces, the TiO₂ layer should be amorphous and thus crystallinity could not explain the delayed coagulation on the TMF surface. Regarding roughness, rougher surfaces could be a disadvantage for hemocompatibility. Rougher surfaces have larger surface areas, which provides more sites for the adsorption of proteins that induce platelets adhesion and blood coagulation [73] such as fibrinogen [75]. However, the increased roughness of the TMF EP surface in comparison to the other surfaces studied did not negatively affect its hemocompatibility. In fact, surface roughness below 50 nm (the case of TMF EP) has been considered suitable for blood contact implants [73]. The higher hydrophilicity of TMF surfaces in comparison to pure titanium might have contributed to their better hemocompatibility. Improved hydrophilicity increases the surface resistance to protein (such as fibrinogen) adsorption [72] and maintains the conformations of fibrinogen, which is beneficial for the anticoagulant capacity [73]. Another factor that might have contributed to the improved hemocompatibility of the TMF alloy in comparison to pure titanium is related to the chemical composition of the oxide layer. XPS analysis showed that a mixture of Ti and Mo oxides formed the oxide layer on TMF surfaces, while only Ti oxides formed the layer on pure titanium. The relative permittivity (dielectric constant) of the oxide layer affects the adsorption of proteins, such as albumins, which inhibit platelets aggregation. The lower the relative permittivity of the oxide layer, the larger the electrostatic force it supports and the higher the adsorption of albumins [30]. Mo oxides are known to have a lower relative permittivity than TiO₂ [76], and thus the presence of Mo oxides on TMF surfaces should result in an oxide layer of lower relative permittivity in comparison to the pure titanium surface, facilitating the adsorption of albumins and, consequently, inhibiting platelets adhesion and aggregation. Finally, the TMF alloy exhibited similar hemocompatibility to stainless steel, with coagulation being slightly delayed on the TMF EP surface in comparison to 316 L EP.

Although Ti-8Mo-2Fe alloy presented relatively good biological responses, surface modification techniques could be envisaged to improve both hemocompatibility and endothelialization. Even though Ti-8Mo-2Fe did not cause acute hemolysis, the blood still clotted on its surface. Moreover, the cytocompatibility results raised some doubts about the capability of this alloy to promote re-endothelialization, since the proliferation rate of endothelial cells on the studied alloy was lower when compared to 316 L stainless steel. Several surface treatments have been proposed to improve the surface properties of titanium-based alloys to prevent thrombus formation [20]. Among them, a promising approach is the development of multifunctional surfaces that mimic the endothelial environment [20,[77], [78], [79]] or that direct the cardiovascular cells fate [80], acting not only in the prevention of thrombus formation, but also in the stimulation of endothelialization.

5. Conclusions

This study comprised a new β metastable Ti-8Mo-2Fe alloy developed for balloon-expandable stent applications. Its electrochemical and in vitro biological behaviors were compared to those of 316 L stainless steel and pure titanium. The following are the conclusions that show that Ti-8Mo-2Fe alloy is a biocompatible material for blood contact implants such as stents:

• (1)Although EIS results showed that Ti-8Mo-2Fe alloy exhibited lower general corrosion resistance than 316 L stainless steel and pure titanium, no significant differences were observed regarding the corrosion rate measured from PDP analyses, which was of the order of 2 × 10⁻⁴ mm/y for all the studied materials. In addition, the Ti-8Mo-2Fe alloy had an advantage over the 316 L stainless steel for biomedical applications, which was the resistance to pitting corrosion in physiological solutions up to high potentials (1.7 V), preventing the release of metal ions and cracks that could occur if triggered by pitting sites.
• (2)Adhesion and viability of endothelial and smooth muscle cells at day 1 of incubation on Ti-8Mo-2Fe surfaces were worse compared to pure titanium. The presence of Mo on the surface of Ti-8Mo-2Fe and its coarse grains may have contributed to this outcome. However, both endothelial and smooth muscle cells proliferated on Ti-8Mo-2Fe surfaces, with the viability at day 7 being similar to that of pure titanium. When compared to 316 L stainless steel, a decrease in the proliferation rate of both endothelial and smooth muscle cells was observed for the Ti-8Mo-2Fe alloy. While the reduced proliferation rate of smooth muscle cells is an advantage for the application envisaged, the reduced rate of endothelial cells is a disadvantage. Further investigations using co-cultures or in vivo tests would be interesting to evaluate the competition between endothelial and smooth muscle cells.
• (3)Regarding hemocompatibility, like 316 L stainless steel and pure titanium, the Ti-8Mo-2Fe alloy did not cause acute hemolysis. Moreover, the blood coagulation was delayed on Ti-8Mo-2Fe surfaces, and the platelets were less aggregated and activated over them in comparison to pure titanium, which is an advantage for blood contact implants. When compared to 316 L stainless steel, Ti-8Mo-2Fe showed similar hemocompatibility, with the coagulation being slightly delayed on the surface of electropolished Ti-8Mo-2Fe in comparison to electropolished stainless steel.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.