Biological Activity and Thrombogenic Properties of Oxide Nanotubes on the Ti-13Nb-13Zr Biomedical Alloy

The success of implant treatment is dependent on the osseointegration of the implant. The main goal of this work was to improve the biofunctionality of the Ti-13Nb-13Zr implant alloy by the production of oxide nanotubes (ONTs) layers for better anchoring in the bone and use as an intelligent carrier in drug delivery systems. Anodization of the Ti-13Nb-13Zr alloy was carried out in 0.5% HF, 1 M (NH4)2SO4 + 2% NH4F, and 1 M ethylene glycol + 4 wt.% NH4F electrolytes. Physicochemical characteristics of ONTs were performed by high-resolution electron microscopy (HREM), X-ray photoelectron spectroscopy (XPS), and scanning Kelvin probe (SKP). Water contact angle studies were conducted using the sitting airdrop method. In vitro biological properties and release kinetics of ibuprofen were investigated. The results of TEM and XPS studies confirmed the formation of the single-walled ONTs of three generations on the bi-phase (α + β) Ti-13Nb-13Zr alloy. The ONTs were composed of oxides of the alloying elements. The proposed surface modification method ensured good hemolytic properties, no cytotoxity for L-929 mouse cells, good adhesion, increased surface wettability, and improved athrombogenic properties of the Ti-13Nb-13Zr alloy. Nanotubular surfaces allowed ibuprofen to be released from the polymer matrix according to the Gallagher–Corrigan model.


Introduction
Each implant introduced into a living organism is initially treated as a foreign body and causes a defensive reaction of the immune system in the form of allergy or inflammation. The material that is to be used to produce a biomedical implant should have many features that will prove high biocompatibility, called biotolerance. It should show very good corrosion resistance and no adverse impact of degradation on the surrounding body cells. Ensuring high biocompatibility has become the reason for a far-reaching reduction of the content of harmful elements in metallic biomaterials [1][2][3][4][5][6][7]. The literature discusses in detail the influence of the excess of some biocompatible elements in the human body on

Material Characterization
The structure of formed ONTs layers was examined using high-resolution electron microscopy (HREM) technique using a JEOL JEM-3010 Transmission Electron Microscope (TEM, JEOL Ltd., Tokyo, Japan) operating at 30 kV equipped with 2k × 2k Orius TM 833 SC200D Gatan CCD camera. To produce TEM samples, the anodized alloy under investigation was suspended in isopropanol and sonicated for 60 min in an ultrasonic bath. The resulting liquid was dropped on standardized Cu grid with an amorphous carbon film, and after evaporation under normal conditions, samples were studied using TEM. The bright field images were recorded. Chemical states of the Ti-13Nb-13Zr alloy surface in the initial state and after anodizing were studied using X-ray photoelectron spectroscopy (XPS) by means of a Prevac photoelectron spectrometer with a VG SCIENTA R3000 hemispherical analyzer (Pleasanton, CA, USA). Monochromatic X-rays, characterized by energy of 1486.74 eV and power of 400 W, were applied.

Scanning Kelvin Probe Measurements
Contact potential difference (CPD) maps of the Ti-13Nb-13Zr alloy surface covered with ONTs layers were registered using SEW-PAR Model 370 device (Princeton Applied Research, Oak Ridge, TN, USA). A tungsten Kelvin probe was used (KP, ø150 µm, Princeton Applied Research, Oak Ridge, TN, USA). The surface area of 1 × 1 mm 2 was scanned. The sample was placed at a distance of ca. 80 µm from the micro-probe. Statistical analysis of the CPD maps allowed us to determine histograms as well as the height and spatial parameters that quantitatively characterized CPD magnitudes and their distribution on the surface of tested material. The arithmetic average (CPD av ), the root mean square deviation (CPD rms ), the skewness (CPD sk ), the excess kurtosis (CPD ku ), and the autocorrelation length (CPD al ) parameters were determined.

Wettability Measurements
Water contact angle measurements were carried out using an OCA 15EC goniometer (Future Digital Scientific Corp., Westbury, NY, USA) with an accuracy of ±0.01 • by the sitting airdrop method. Moreover, 10 images of water drops with a volume of about 5 µL placed on the tested surface were recorded within 10 s. On the basis of the received images, the mean values of the contact angle ( J. Funct. Biomater. 2023, 14, x FOR PEER REVIEW mean square deviation (CPDrms), the skewness (CPDsk), the exces the autocorrelation length (CPDal) parameters were determined.

Wettability Measurements
Water contact angle measurements were carried out using an (Future Digital Scientific Corp., Westbury, NY, USA) with an acc sitting airdrop method. Moreover, 10 images of water drops with a placed on the tested surface were recorded within 10 s. On the basis the mean values of the contact angle (Ѳ) were determined. measurements conducted in different parts of the tested surface values of the Ѳ.

In Vitro Hemocompatibility Test
Hemocompatibility of the ONTs layers on the Ti-13Nb-13Z investigated in accordance with the ASTM F756-17 [38]. Healthy h donors of blood (Regional Blood Donation and Blood Treatm Poland). Based on the International Committee for Standardization the hemoglobin (Hb) calibration curve meeting the requirements w buffered saline (PBS) was used for preparation of dilutions to the H g L −1 . Specimens with the determined surface area were applied at area to 1 mL of test blood solution. In case of samples with produce ml human blood was applied. Exposition of the final samples in hu h at 37(1) °C. After incubation for 4 h, the fluid was transferred centrifuged at 3000 rpm for 15 min in a standard clinical centrifug removed, being careful not to disturb any button of erythrocytes in solution absorbance was investigated using a spectrophotometer nm. Calculation of % hemolysis was based on the following Formu tu in ion concentrat hemoglobin total super in released hemoglobin of ion concentrat hemolysis % 

Cell Culture and Cytotoxicity Assays
Cytotoxicity tests were performed in accordance with I Mammalian cell culture monolayer consisting of L-929 mouse F Aldrich, St. Louis, MO, USA, Lot: 10i019) was used. Extraction o performed by incubating the material with MEM suppleme Penicillin/Streptomycin, and GlutaMAX at 37(1) °C (humidified) in Quadruplicate monolayers of L-929 mouse fibroblast (passage no.: 1 1×, 2×, 3× and 4× dilutions of the extract and incubated at 37(1) °C in CO2 for 24 (1) h. Following the incubation, 50 µL of the MTT solutio use, was dispended in each well and incubated for 120 (15) min at 3 5(1) % CO2. Following the incubation, the MTT solution was isopropanol and incubated for 10 min at 37(1) °C (humidified) in p The percent viability for the test article and control article were blanks. Reduction of the number of living cells caused a decrease in in the sample. That decrease was directly connected with the formazan, whose presence was confirmed by the optical density at 5 filter of 650 nm.
Negative Control was HDPE (Granulat G. Motloch, LOT: C7 supplemented with 10% FBS, Penicillin/Streptomycin, and GlutaM was MEM supplemented with 10% FBS, Penicillin/Streptomycin DMSO. Blank Control was MEM supplemented with 10% FBS, P and GlutaMAX. ) were determined. The average of five measurements conducted in different parts of the tested surface was taken as the final values of the J. Funct. Biomater. 2023, 14, x FOR PEER REVIEW 4 of 24 mean square deviation (CPDrms), the skewness (CPDsk), the excess kurtosis (CPDku), and the autocorrelation length (CPDal) parameters were determined.

Wettability Measurements
Water contact angle measurements were carried out using an OCA 15EC goniometer (Future Digital Scientific Corp., Westbury, NY, USA) with an accuracy of ±0.01° by the sitting airdrop method. Moreover, 10 images of water drops with a volume of about 5 µL placed on the tested surface were recorded within 10 s. On the basis of the received images, the mean values of the contact angle (Ѳ) were determined. The average of five measurements conducted in different parts of the tested surface was taken as the final values of the Ѳ.

In Vitro Hemocompatibility Test
Hemocompatibility of the ONTs layers on the Ti-13Nb-13Zr alloy substrate was investigated in accordance with the ASTM F756-17 [38]. Healthy human volunteers were donors of blood (Regional Blood Donation and Blood Treatment Center, Katowice, Poland). Based on the International Committee for Standardization in Hematology (ICSH), the hemoglobin (Hb) calibration curve meeting the requirements was prepared. Phosphate buffered saline (PBS) was used for preparation of dilutions to the Hb concentration of 10(1) g L −1 . Specimens with the determined surface area were applied at a ratio of 3 cm 2 surface area to 1 mL of test blood solution. In case of samples with produced ONTs layers, 0.57(02) ml human blood was applied. Exposition of the final samples in human blood lasted for 4 h at 37(1) °C. After incubation for 4 h, the fluid was transferred to the test tubes and centrifuged at 3000 rpm for 15 min in a standard clinical centrifuge. The supernatant was removed, being careful not to disturb any button of erythrocytes in the test tube. Then, the solution absorbance was investigated using a spectrophotometer at a wavelength of 540 nm. Calculation of % hemolysis was based on the following Formula (1): (1)
Following the incubation, the MTT solution was replaced with 100 µL isopropanol and incubated for 10 min at 37(1) °C (humidified) in presence of 5(1) % CO2. The percent viability for the test article and control article were determined from the blanks. Reduction of the number of living cells caused a decrease in the metabolic activity in the sample. That decrease was directly connected with the amount of blue-violet formazan, whose presence was confirmed by the optical density at 570 nm with differential filter of 650 nm.

In Vitro Hemocompatibility Test
Hemocompatibility of the ONTs layers on the Ti-13Nb-13Zr alloy substrate was investigated in accordance with the ASTM F756-17 [38]. Healthy human volunteers were donors of blood (Regional Blood Donation and Blood Treatment Center, Katowice, Poland). Based on the International Committee for Standardization in Hematology (ICSH), the hemoglobin (Hb) calibration curve meeting the requirements was prepared. Phosphate buffered saline (PBS) was used for preparation of dilutions to the Hb concentration of 10(1) g L −1 . Specimens with the determined surface area were applied at a ratio of 3 cm 2 surface area to 1 mL of test blood solution. In case of samples with produced ONTs layers, 0.57(02) ml human blood was applied. Exposition of the final samples in human blood lasted for 4 h at 37(1) • C. After incubation for 4 h, the fluid was transferred to the test tubes and centrifuged at 3000 rpm for 15 min in a standard clinical centrifuge. The supernatant was removed, being careful not to disturb any button of erythrocytes in the test tube. Then, the solution absorbance was investigated using a spectrophotometer at a wavelength of 540 nm. Calculation of % hemolysis was based on the following Formula (1): % hemolysis = concentration of hemoglobin released in supernatant · 100% total hemoglobin concentration in tube . (1)

Cell Culture and Cytotoxicity Assays
Cytotoxicity tests were performed in accordance with ISO 10993-5:2009 [39]. Mammalian cell culture monolayer consisting of L-929 mouse Fibroblast cells (Sigma-Aldrich, St. Louis, MO, USA, Lot: 10i019) was used. Extraction of the test material was performed by incubating the material with MEM supplemented with 10% FBS, Penicillin/Streptomycin, and GlutaMAX at 37(1) • C (humidified) in 5(1)% CO 2 for 24 (2) h. Quadruplicate monolayers of L-929 mouse fibroblast (passage no.: 15) cells were dose with 1×, 2×, 3× and 4× dilutions of the extract and incubated at 37(1) • C in the presence of 5(1) % CO 2 for 24(1) h. Following the incubation, 50 µL of the MTT solution, prepared just before use, was dispended in each well and incubated for 120(15) min at 37(1) • C (humidified) in 5(1) % CO 2 . Following the incubation, the MTT solution was replaced with 100 µL isopropanol and incubated for 10 min at 37(1) • C (humidified) in presence of 5(1) % CO 2 . The percent viability for the test article and control article were determined from the blanks. Reduction of the number of living cells caused a decrease in the metabolic activity in the sample. That decrease was directly connected with the amount of blue-violet formazan, whose presence was confirmed by the optical density at 570 nm with differential filter of 650 nm.

Qualitative In Vitro Cell Adhesion Assay
Duplicated test articles, with the indicated side up, were placed individually in a well of a six-well plate. The L929 mouse Fibroblast cells (Sigma-Aldrich, St. Louis, MO, USA, Lot: 10i019, passage no.:8) were harvested using trypsin and counted. To each well, 80,000 cells were added with enough MEM (supplemented with 10% FBS, Penicillin/Streptomycin, and GlutaMAX) to cover the discs. The plates were incubated at 37(1) • C (humidified) in 5(1) % CO 2 for 24 (2) h. After the incubation period, the cells were fixed in 10% PFA for 15 min, followed by the Hematoxylin and Eosin staining. PFA was replaced with Hematoxylin and incubated at room temperature for 5 min. Each well was washed 3 times with tap water. Then, Eosin was added, and plates were incubated for 5 min at room temperature, followed by 3 washes with tap water. The stained cells were visualized using Delta Optical stereomicroscope SZ-630T. The photos were captured using ScopeImage 9 software and HDCE-X5 camera.

Blood Donation and Platelet-Reach Plasma Preparation
Blood was collected for the anticoagulant CPDA-1 from healthy human volunteers after making sure that they did not take any anticoagulants, including acetylsalicylic acid, clopidogrel, or warfarin/acenocumarol over 14 days. Storing the collected blood lasted up to 24 h at 2-7 • C. Blood count (hematology automaton BC 2800 VET, Mindray, Shenzhen, China), platelets aggregation under the adenosine diphosphate (ADP) and arachidonic acid (impedance aggregometer Multiplate, Roche, Switzerland, with ADPtest and ASPItest tests) and concentration of plasma-free hemoglobin fHB (spectrophotometer Plasma/Low Hb, Hemocue AB, Ängelholm, Sweden), were controlled before starting further measurements. Blood was qualified for further tests, provided that the blood count was normal, the platelet count was higher than 1.2 × 10 5 L µL −1 , the ADP test result was above 122 AUC, the ASPItest result was higher than 136 AUC, and the fHB below 0.2 g dL −1 . Platelet-rich plasma (PRP) [40] was prepared directly before the tests by centrifugation of whole blood at 100 G for 10 min at room temperature. Plasma morphology was investigated before the use of PRP in the test. Examined samples were put into polypropylene test tubes with a flat bottom of Falcon type with a volume of 50 mL and a diameter of 30 mm. Each sample was put into a separate test tube so that the sample area, which was not subjected to final processing and polishing, was at the bottom. Test tubes were filled with 10 mL of PRP, and the studied samples were incubated with PRP for 60 min at the temperature of 37(1) • C. The test tubes used for the incubation were placed on the hematology cradle at deflection ±5 • and frequency 10 cycles/minute, ensuring even coverage of the samples with platelets and their aggregates. Before the incubation ended, the examined samples were withdrawn from the test tubes and carefully washed with PBS, and next fixed in 4% buffered formalin. The material prepared in such a way was then the subject of imaging tests.

Qualitative Thrombogenicity Scale
The analysis of thrombogenicity test was carried out on 39 specimens, including 12 specimens of the 1G ONTs, 9 specimens of the 2G ONTs, 9 specimens of the 3G ONTs, and 9 specimens of the non-anodized Ti-13Nb-13Zr alloy. Neither literature data nor standards (e.g., ISO 10993-5:2009 [39]) give quantitative method of measurement thrombogenicity of biomaterials. Therefore, in the studies described in this paper, a qualitative method was used. In order to make the measurement more reliable, the assessment was performed by two experts who used a six-level ordinal scale (Table 1) [41]. As the experts, a biologist with 25 years of experience in cell testing and a bioengineer with 20 years of experience in thrombogenicity and medical statistics were chosen. Table 1. Ordinal thrombogenicity scale together with description and exemplary images of surface of the sample belonging to a given class [41].

Class
Description Example 0 contains samples characterized by minimal thrombogenicity: separated platelets of low diverse level, no platelet aggregates This material is used in medical devices devoted to long-term contact with blood. The material samples were prepared in the form of disks of 8 mm diameter and 0.8 mm thickness by high-pressure injection. The surface roughness described by an arithmetic mean deviation of the roughness profile was Ra = 0.16. Class 5 constituted positive control. As a reference material, glass disks covered by collagen (Neuvitro Corporation, Vancouver, WA, USA) were chosen. Collagen is a natural fibrous protein that exists in extracellular matrix and has strong ability to activate platelets. This material is used in medical devices devoted to long-term contact with blood. The material samples were prepared in the form of disks of 8 mm diameter and 0.8 mm thickness by high-pressure injection. The surface roughness described by an arithmetic mean deviation of the roughness profile was Ra = 0.16. Class 5 constituted positive control. As a reference material, glass disks covered by collagen (Neuvitro Corporation, Vancouver, WA, USA) were chosen. Collagen is a natural fibrous protein that exists in extracellular matrix and has strong ability to activate platelets. This material is used in medical devices devoted to long-term contact with blood. The material samples were prepared in the form of disks of 8 mm diameter and 0.8 mm thickness by high-pressure injection. The surface roughness described by an arithmetic mean deviation of the roughness profile was Ra = 0.16. Class 5 constituted positive control. As a reference material, glass disks covered by collagen (Neuvitro Corporation, Vancouver, WA, USA) were chosen. Collagen is a natural fibrous protein that exists in extracellular matrix and has strong ability to activate platelets.

. Statistical Analysis Methods
Both experts evaluated the same material in six-level ordinal scale. In order to study the degree of compliance of the two experts the alpha-Krippendorf coefficient was used. The coefficient has values from 0 to 1. It is agreed that if its value is >0.80, the compliance of assessments is very good. The Kruskal-Wallis test was used for that data analysis because the number of classes was higher than 2, variable describing thrombogenicity was on ordinal scale, and the experimental model was independent. The Kruskal-Wallis test is nonparametric equivalent of variance analysis. Multiple comparison test was applied for post-hoc analysis.

Drug Delivery System
The Ti-13Nb-13Zr samples with ONTs layers were subjected to surface functionalization heparine-dopamine (Hep-DOPA) conjugate by mixing 40 mg mL −1 heparin, 19.06 mg mL −1 EDAC (N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide, hydrochloride) and 11.5 mg mL −1 NHS (N-Hydroxysuccinimide) and reacted with 10 mL of MES buffer pH = 4.5 for 10 min (solution 1). Then, 102.2 mg mL −1 dopamine was mixed with 1 mL MES buffer pH = 4.5 (solution 2). Solution 1 and solution 2 were mixed and reacted for 12 h in the dark. The Ti-13Nb-13Zr alloy with ONTs layers was placed into the mixture prepared for 12 h. The material obtained in this way was dried, then ibuprofen was immobilized on the surface of the heparinized melt by adding 10 mg mL −1 of 5 contains samples characterized by very high-degree thrombogenicity: platelets are highly differentiated and form numerous aggregates, which are connected with each other without possibility of separating and counting objects 3 contains samples characterized by averagedegree thrombogenicity: biological material mainly comprises aggregates larger than individual blood platelets 4 contains samples characterized by highdegree thrombogenicity: sample is covered with highly differentiated biological material, and the individual objects are connected with each other without possibility of separating them 5 contains samples characterized by very high-degree thrombogenicity: platelets are highly differentiated and form numerous aggregates, which are connected with each other without possibility of separating and counting objects

Statistical Analysis Methods
Both experts evaluated the same material in six-level ordinal scale. In order to study the degree of compliance of the two experts the alpha-Krippendorf coefficient was used. The coefficient has values from 0 to 1. It is agreed that if its value is >0.80, the compliance of assessments is very good. The Kruskal-Wallis test was used for that data analysis because the number of classes was higher than 2, variable describing thrombogenicity was on ordinal scale, and the experimental model was independent. The Kruskal-Wallis test is nonparametric equivalent of variance analysis. Multiple comparison test was applied for post-hoc analysis.

Drug Delivery System
The Ti-13Nb-13Zr samples with ONTs layers were subjected to surface functionalization heparine-dopamine (Hep-DOPA) conjugate by mixing 40 mg mL −1 heparin, 19.06 mg mL −1 EDAC (N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide, hydrochloride) and 11.5 mg mL −1 NHS (N-Hydroxysuccinimide) and reacted with 10 mL of MES buffer pH = 4.5 for 10 min (solution 1). Then, 102.2 mg mL −1 dopamine was mixed with 1 mL MES buffer pH = 4.5 (solution 2). Solution 1 and solution 2 were mixed and reacted for 12 h in the dark. The Ti-13Nb-13Zr alloy with ONTs layers was placed into the mixture prepared for 12 h. The material obtained in this way was dried, then ibuprofen was immobilized on the surface of the heparinized melt by adding 10 mg mL −1 of The class 0 constituted negative control. As a reference material, polyurethane based on polyesters Bionate 55D (DSM Biomedical Inc., Geleen, The Netherlands) was chosen. This material is used in medical devices devoted to long-term contact with blood. The material samples were prepared in the form of disks of 8 mm diameter and 0.8 mm thickness by high-pressure injection. The surface roughness described by an arithmetic mean deviation of the roughness profile was Ra = 0.16.
Class 5 constituted positive control. As a reference material, glass disks covered by collagen (Neuvitro Corporation, Vancouver, WA, USA) were chosen. Collagen is a natural fibrous protein that exists in extracellular matrix and has strong ability to activate platelets.

Statistical Analysis Methods
Both experts evaluated the same material in six-level ordinal scale. In order to study the degree of compliance of the two experts the alpha-Krippendorf coefficient was used. The coefficient has values from 0 to 1. It is agreed that if its value is >0.80, the compliance of assessments is very good. The Kruskal-Wallis test was used for that data analysis because the number of classes was higher than 2, variable describing thrombogenicity was on ordinal scale, and the experimental model was independent. The Kruskal-Wallis test is nonparametric equivalent of variance analysis. Multiple comparison test was applied for post-hoc analysis.

Drug Delivery System
The Ti-13Nb-13Zr samples with ONTs layers were subjected to surface functionalization heparine-dopamine (Hep-DOPA) conjugate by mixing 40 mg mL −1 heparin, 19.06 mg mL −1 EDAC (N-(3-Dimethylaminopropyl)-N -ethylcarbodiimide, hydrochloride) and 11.5 mg mL −1 NHS (N-Hydroxysuccinimide) and reacted with 10 mL of MES buffer pH = 4.5 for 10 min (solution 1). Then, 102.2 mg mL −1 dopamine was mixed with 1 mL MES buffer pH = 4.5 (solution 2). Solution 1 and solution 2 were mixed and reacted for 12 h in the dark. The Ti-13Nb-13Zr alloy with ONTs layers was placed into the mixture prepared for 12 h. The material obtained in this way was dried, then ibuprofen was immobilized on the surface of the heparinized melt by adding 10 mg mL −1 of ibuprofen to 10 mL of Tris-HCl at pH = 8. The reaction was conducted for 24 h at room temperature with gentle shaking. The release kinetics of ibuprofen from the nanotubular layer was investigated by immersing the sample in 15 mL PBS with pH = 7.45 at 37(1) • C for 24 h. The kinetics of the released drug was studied for 60 min, taking the solution for analysis every 10 min. Each time 1 mL of the solution was withdrawn by adding fresh solution. Using UV-VIS spectroscopy, the amount of substance released from the nanotubular oxide layer was determined. The absorbance value was measured at the wavelength λ = 257 nm, determining the absorbance value of the PBS in the first step and then the absorbed solution. Based on the following Formula (2), the percentage amount of released drug was calculated:

FE-SEM and TEM Characterization
The selected Ti-13Nb-13Zr alloy to obtain 1G, 2G, and 3G ONTs layers is a new type of vital titanium alloy. The addition of Nb and Zr in the amount of 13 wt.% provides appropriate properties for application in regenerative medicine. Additionally, it ensures better biocompatibility and higher corrosion resistance compared to titanium [11][12][13][14]. Anodization of the Ti-13Nb-13Zr alloy was conducted using the anode made of the tested sample and the Pt cathode at a distance of 25 mm in a face-to-face position ( Figure 1). ibuprofen to 10 mL of Tris-HCl at pH = 8. The reaction was conducted for 24 h at room temperature with gentle shaking. The release kinetics of ibuprofen from the nanotubular layer was investigated by immersing the sample in 15 mL PBS with pH = 7.45 at 37(1) °C for 24 h. The kinetics of the released drug was studied for 60 min, taking the solution for analysis every 10 min. Each time 1 mL of the solution was withdrawn by adding fresh solution. Using UV-VIS spectroscopy, the amount of substance released from the nanotubular oxide layer was determined. The absorbance value was measured at the wavelength λ = 257 nm, determining the absorbance value of the PBS in the first step and then the absorbed solution. Based on the following Formula (2), the percentage amount of released drug was calculated:

FE-SEM and TEM Characterization
The selected Ti-13Nb-13Zr alloy to obtain 1G, 2G, and 3G ONTs layers is a new type of vital titanium alloy. The addition of Nb and Zr in the amount of 13 wt.% provides appropriate properties for application in regenerative medicine. Additionally, it ensures better biocompatibility and higher corrosion resistance compared to titanium [11][12][13][14]. Anodization of the Ti-13Nb-13Zr alloy was conducted using the anode made of the tested sample and the Pt cathode at a distance of 25 mm in a face-to-face position ( Figure 1). The obtained FE-SEM images of surface morphology for 1G, 2G, and 3G ONTs layers on the Ti-13Nb-13Zr substrate with corresponding TEM images are presented in Figure 2. From the observed regions, the selected area electron diffraction (SAED) patterns were recorded. SAED patterns (see insight in Figure 2) clearly indicate that all layers of 1G, 2G, and 3G ONTs exhibit amorphous structures. From selected areas of the FE-SEM images of three generations of ONTs layers obtained via anodization, the morphological parameters of the nanotubes were determined and summed up in Table 2. The obtained FE-SEM images of surface morphology for 1G, 2G, and 3G ONTs layers on the Ti-13Nb-13Zr substrate with corresponding TEM images are presented in Figure 2. From the observed regions, the selected area electron diffraction (SAED) patterns were recorded. SAED patterns (see insight in Figure 2) clearly indicate that all layers of 1G, 2G, and 3G ONTs exhibit amorphous structures. From selected areas of the FE-SEM images of three generations of ONTs layers obtained via anodization, the morphological parameters of the nanotubes were determined and summed up in Table 2.
Based on the results shown in Table 2, it can be seen that each generation of ONTs is characterized by different morphological parameters. For 1G ONTs, the smallest diameter of the nanotubes is obtained. In the case of 2G ONTs, an increase of 20% in the length of the outer diameter compared to 1G ONTs is observed. More than four and a half times larger outer diameters of 3G ONTs compared to 1G ONTs is confirmed. The longest oxide nanotubes are obtained for 3G ONTs of the order of almost 10 µm. The shortest ones are observed for 1G ONTs. The proposed anodic oxidation conditions for 2G ONTs allow to obtain nanotubes almost 4 µm long. Based on the results shown in Table 2, it can be seen that each generation of ONTs is characterized by different morphological parameters. For 1G ONTs, the smallest diameter of the nanotubes is obtained. In the case of 2G ONTs, an increase of 20% in the length of the outer diameter compared to 1G ONTs is observed. More than four and a half times larger outer diameters of 3G ONTs compared to 1G ONTs is confirmed. The longest oxide nanotubes are obtained for 3G ONTs of the order of almost 10 µm. The shortest ones are observed for 1G ONTs. The proposed anodic oxidation conditions for 2G ONTs allow to obtain nanotubes almost 4 µm long. Based on the literature, the influence of anodizing conditions on the diameters and lengths of oxide nanotubes obtained for different generations on the Ti-13Nb-13Zr alloy can be confirmed. The influence of voltage change on the diameter of 2G ONTs on the Ti-13Nb-13Zr alloy is presented in the paper [42]. In the voltage range of 10-45 V, an increase in nanotube diameter from 55 to 250 nm can be seen. For 3G ONTs on the same substrate  Based on the literature, the influence of anodizing conditions on the diameters and lengths of oxide nanotubes obtained for different generations on the Ti-13Nb-13Zr alloy can be confirmed. The influence of voltage change on the diameter of 2G ONTs on the Ti-13Nb-13Zr alloy is presented in the paper [42]. In the voltage range of 10-45 V, an increase in nanotube diameter from 55 to 250 nm can be seen. For 3G ONTs on the same substrate at the voltage of 15-35 V, the outer diameter of the nanotube increased from 104 (13) to 230(30) nm [43].

Roughness Profile Measurements
Since the roughness of the contact surface affects the adhesion and, thus, the strength of the connection, an important element for obtaining a good adhesive connection is the proper development of the contact surface of biomaterials. For titanium implants such as orthopedic and dental, it is advisable to obtain a permanent connection between the tissue and the material; therefore, the aim is to obtain an optimum surface roughness of Ra between 1 and 3 µm, while in the case of surgical tools and implants intended, among others, for implants prepared for contact with blood surfaces with as little roughness as possible them [43][44][45].
The geometric structure of the surface (GSS) of the Ti-13Nb-13Zr alloy without and with ONTs layers of three generations was characterized by surface micro-geometry measurements in a two-dimensional (2D) system. The discussion of the influence of anodizing conditions on the GSS was based on the selected profile height parameter Ra. Figure 3 presents the exemplary roughness profiles after alignment for the 1G, 2G, and 3G ONTs layers on the Ti13-Nb-13Zr alloy. All obtained surface profiles were symmetrical. A similar value of Ra = 0.16(01) and Ra = 0.13(01) parameters for the 1G and 2G ONTs layers was determined, respectively. The 3G ONTs were characterized by the highest Ra parameter of 1.34(04). at the voltage of 15-35 V, the outer diameter of the nanotube increased from 104(13) to 230(30) nm [43].

Roughness Profile Measurements
Since the roughness of the contact surface affects the adhesion and, thus, the strength of the connection, an important element for obtaining a good adhesive connection is the proper development of the contact surface of biomaterials. For titanium implants such as orthopedic and dental, it is advisable to obtain a permanent connection between the tissue and the material; therefore, the aim is to obtain an optimum surface roughness of Ra between 1 and 3 m, while in the case of surgical tools and implants intended, among others, for implants prepared for contact with blood surfaces with as little roughness as possible them [43][44][45].
The geometric structure of the surface (GSS) of the Ti-13Nb-13Zr alloy without and with ONTs layers of three generations was characterized by surface micro-geometry measurements in a two-dimensional (2D) system. The discussion of the influence of anodizing conditions on the GSS was based on the selected profile height parameter Ra. Figure 3 presents the exemplary roughness profiles after alignment for the 1G, 2G, and 3G ONTs layers on the Ti13-Nb-13Zr alloy. All obtained surface profiles were symmetrical. A similar value of Ra = 0.16(01) and Ra = 0.13(01) parameters for the 1G and 2G ONTs layers was determined, respectively. The 3G ONTs were characterized by the highest Ra parameter of 1.34(04). The Ra for the 3G ONTs increased over 8 times in comparison with the 1G ONTs and 10 times in relation to the 2G ONTs due to the differences in the morphological parameters of the obtained nanotubes on the Ti-13Nb-13Zr alloy. It can be seen that the most developed surface and the thickest layer of nanotubes were obtained for the 3G ONTs layer ( Table 2). Based on the obtained results, it can be expected that the 3G ONTs layers will allow for better stability at the implant-bone interface, improved osseointegration, and reduced risk of metal ion release as products of corrosion processes and biotribological wear. In our earlier work [43], it was reported that deviations of the roughness profile from the mean line at the micro-scale were not observed for the smooth surface of the Ti-13Zr-13Nb alloy before anodizing, for which Ra was only 0.10 µm. The positive role of surface roughness was reported in studies on stainless steel, Co-Cr alloy, titanium, Ti-6Al-4V, and Ti-6Al-7Nb alloys using human osteoblasts [46,47]. The authors studied the activity of bone tissue after implanting titanium implants with a smooth and rough surface on rabbits. Better contact of the implant with the bone and a larger volume of bone tissue was obtained for rough surfaces compared to smooth surfaces. The Ra for the 3G ONTs increased over 8 times in comparison with the 1G ONTs and 10 times in relation to the 2G ONTs due to the differences in the morphological parameters of the obtained nanotubes on the Ti-13Nb-13Zr alloy. It can be seen that the most developed surface and the thickest layer of nanotubes were obtained for the 3G ONTs layer (Table 2). Based on the obtained results, it can be expected that the 3G ONTs layers will allow for better stability at the implant-bone interface, improved osseointegration, and reduced risk of metal ion release as products of corrosion processes and biotribological wear. In our earlier work [43], it was reported that deviations of the roughness profile from the mean line at the micro-scale were not observed for the smooth surface of the Ti-13Zr-13Nb alloy before anodizing, for which Ra was only 0.10 µm. The positive role of surface roughness was reported in studies on stainless steel, Co-Cr alloy, titanium, Ti-6Al-4V, and Ti-6Al-7Nb alloys using human osteoblasts [46,47]. The authors studied the activity of bone tissue after implanting titanium implants with a smooth and rough surface on rabbits. Better contact of the implant with the bone and a larger volume of bone tissue was obtained for rough surfaces compared to smooth surfaces.

XPS Study of Chemical States
The XPS studies concerned the determination of atomic concentration, and chemical states analysis of the obtained ONTs components. Detailed analysis of the shape and position of Ti2p, Nb3d, and Zr3d photoemission lines for the 1G and 2G ONTs in Figure 4 a-f was conducted and compared with the results of our previous XPS research on the Ti-13Nb-13Zr alloy reported in the earlier work [13].
characteristic of the Ti 4+ oxidation state as observed in anatase [48]. A weak contribut of oxidized titanium characteristic for trivalent titanium is visible at binding energy 45 eV [49]. Both chemical states were previously observed in the Ti-13Nb-13Zr alloy, which 91.64% of titanium atoms were in the 4 + valence state and 8.36% in the 3 + [13]. T amount of the trivalent titanium in the 1G and 2G ONTs is lower than in mention reference sample; for the 1G ONTs, Ti 3+ atoms account for only 1.81% of all titanium atom whereas for the 2G ONTs, it is only 6.35%.   (Figure 4c,d) characteristic of the Nb2O5 oxide [50]. The same result for niobium was observed in 3 ONTs in our previous studies [13]. Analysis of the zirconium lines in Figure 4e,f reveale some similarities between the surfaces of the 1G and 2G ONTs and previously reporte electrochemically modified Ti-13Nb-13Zr. The positions of deconvoluted peaks indica The titanium in electrochemically obtained 1G and 2G ONTs occurs mainly as the TiO 2 . The binding energy of the prominent peak (Figure 4a) is located at 458.8 eV and is characteristic of the Ti 4+ oxidation state as observed in anatase [48]. A weak contribution of oxidized titanium characteristic for trivalent titanium is visible at binding energy 457.2 eV [49]. Both chemical states were previously observed in the Ti-13Nb-13Zr alloy, for which 91.64% of titanium atoms were in the 4 + valence state and 8.36% in the 3 + [13]. The amount of the trivalent titanium in the 1G and 2G ONTs is lower than in mentioned reference sample; for the 1G ONTs, Ti 3+ atoms account for only 1.81% of all titanium atoms, whereas for the 2G ONTs, it is only 6.35%.
The deconvolution of the lines Nb3d of the 1G and 2G ONTs indicates the presence of Nb in one chemical state-the position of the Nb3d 5/2 at about 207.4 eV (Figure 4c,d) is characteristic of the Nb 2 O 5 oxide [50]. The same result for niobium was observed in 3G ONTs in our previous studies [13]. Analysis of the zirconium lines in Figure 4e,f revealed some similarities between the surfaces of the 1G and 2G ONTs and previously reported electrochemically modified Ti-13Nb-13Zr. The positions of deconvoluted peaks indicate the presence of ZrO 2 with a peak at 182.3 eV and Zr in another oxidation state with a peak at 182.9 eV or Zr(OH) [51]. The amount of Zr in a particular chemical state is different for 1G and 2G ONTs and differs even more when compared with the previously analyzed 3G ONTs sample [13].
The results of the contribution of the particular chemical state for the electrochemically obtained 1G, 2G, G3 ONTs layers, and reference alloy are summarized in Figure 4g. The data points represent the calculated atomic concentration for each atom in each chemical state. The general observation suggests that Nb and Zr bonded with oxygen or OH group in anodic ONTs layers are relatively stable-their amount does not depend on the applied procedure of electrochemical modification. In the case of Ti, applied anodization generally influences the amount of titanium but also titanium chemical states. This is worth noting that applied procedures for electrochemical oxidation of Ti-13Nb-13Zr in all three generations lead to changes in the relative ratios of the elements under consideration (Figure 4h). Those changes are related to titanium as only the relative ratios of Ti/Zr and Ti/Nb are changed when compared with the reference sample. Based on the 1G and 2G ONTs layers studies, it can also be observed that even if the relative amount of sample constituents are almost identical, as observed for the 1G and 3G [13] ONTs layers, their chemical composition is different.

Electronic Properties
Statistical analysis of the contact potential difference maps in Figure 5 allowed us to determine the arithmetic average (CPD av ), the root mean square deviation (CPD rms ), the skewness (CPD sk ), the excess kurtosis (CPD ku ), and the autocorrelation length (CPD al ), i.e., parameters that quantitatively describe electronic properties of the Ti-13Nb-13Zr alloy surface covered with ONTs. The values of the above parameters are presented in Table 3. Table 3. Statistical parameters calculated using CPD maps of the Ti-13Nb-13Zr substrate and for the 1G, 2G, and 3G ONTs; CPD av is the arithmetic average, CPD rms is the root mean square deviation, CPD al is the autocorrelation length, CPD sk is the skewness, and CPD ku is the excess kurtosis; V KP is the voltage measured versus Kelvin probe.

Parameter
Ti  The root mean square deviation (CPDrms) determines the deviation of CPD heights in relation to the arithmetic average, whereas autocorrelation length (CPDal) characterizes the spatial distribution of CPD heights on the material surface. In other words, CPDal is a measure of the distance over the surface by which one would find a surface feature (CPD in this case) that is statistically different from the one that can be found in the original location. The CPDrms parameters obtained for the investigated surfaces indicate increasing heights of CPD in the order Ti-13Nb-13Zr, 1G ONTs, 2G ONTs, and 3G ONTs. Simultaneously values of the CPDal parameter show that the correlation length is the highest for the Ti-13Nb-13Zr substrate and decreases for surfaces covered with the nanotubes. Generally, both parameters i.e., CPDrms and CPDal indicate that the Ti-13Nb-13Zr substrate is characterized by the most uniform surface of all the samples tested and also that the production of nanotubes on the Ti-13Nb-13Zr surface increases its electrical heterogeneity. The increase in CPDrms and CPDal determined for the surfaces covered with ONTs can be explained by increasing the dimensions of outer diameters of the nanotubes from 87(10) nm for the 1G ONTs, 103(10) nm for the 2G ONTs, and 342(34) nm for the 3G ONTs. Skewness (CPDsk) and excess kurtosis (CPDku) describe the shape of the CPD distribution and give additional information on the CPD peaks/valleys profile. It was found that for all investigated samples CPDsk and CPDku parameters change in the range ±0.3 (see Table 3); thus, one can state that for each material tested, the CPD distribution follows a Gaussian (normal) distribution. It was found that the substrate surface covered with ONTs is characterized by the higher value of the CPD av in comparison with the biomedical Ti-13Nb-13Zr alloy, regardless of the solution composition from which nanotubes were obtained. Moreover, it was stated that the 3G ONTs are characterized by the highest CPD av value (ca. −387 mV KP ) among all the investigated biomaterials. This result may be related to the different chemical compositions of solutions and hence different concentrations of fluorine ions in electrolytes (Table 2), from 0.5% F − for the 1G ONTs, 2% for the 2G ONTs, and 4% for the 3G ONTs.
The root mean square deviation (CPD rms ) determines the deviation of CPD heights in relation to the arithmetic average, whereas autocorrelation length (CPD al ) characterizes the spatial distribution of CPD heights on the material surface. In other words, CPD al is a measure of the distance over the surface by which one would find a surface feature (CPD in this case) that is statistically different from the one that can be found in the original location. The CPD rms parameters obtained for the investigated surfaces indicate increasing heights of CPD in the order Ti-13Nb-13Zr, 1G ONTs, 2G ONTs, and 3G ONTs. Simultaneously values of the CPD al parameter show that the correlation length is the highest for the Ti-13Nb-13Zr substrate and decreases for surfaces covered with the nanotubes. Generally, both parameters i.e., CPD rms and CPD al indicate that the Ti-13Nb-13Zr substrate is characterized by the most uniform surface of all the samples tested and also that the production of nanotubes on the Ti-13Nb-13Zr surface increases its electrical heterogeneity. The increase in CPD rms and CPD al determined for the surfaces covered with ONTs can be explained by increasing the dimensions of outer diameters of the nanotubes from 87(10) nm for the 1G ONTs, 103(10) nm for the 2G ONTs, and 342(34) nm for the 3G ONTs. Skewness (CPD sk ) and excess kurtosis (CPD ku ) describe the shape of the CPD distribution and give additional information on the CPD peaks/valleys profile. It was found that for all investigated samples CPD sk and CPD ku parameters change in the range ±0.3 (see Table 3); thus, one can state that for each material tested, the CPD distribution follows a Gaussian (normal) distribution.
Histograms of the CPD values with fitted Gaussian distributions are shown in Figure 6. The comprehensive description of the histogram preparation and the determination of the parameters can be found elsewhere [52][53][54].
1G, 2G, and 3G ONTs; CPDav is the arithmetic average, CPDrms is the root mean square deviation, CPDal is the autocorrelation length, CPDsk is the skewness, and CPDku is the excess kurtosis; VKP is the voltage measured versus Kelvin probe.

Parameter
Ti- Histograms of the CPD values with fitted Gaussian distributions are shown in Figure  6. The comprehensive description of the histogram preparation and the determination of the parameters can be found elsewhere [52][53][54]. Approximation of the histograms using the Gaussian function allows determining the arithmetic average CPDav and the standard deviation, which in this case corresponds to the CPDrms parameter. Values of CPDsk and CPDku also indicate that the CPD heights are symmetrically distributed around the average as well as that the investigated surfaces do not show inordinately high peaks/deep valleys. Generally, one can state that the entire surface of the Ti-13Nb-13Zr alloy is evenly covered with ONTs irrespective of the different production processes (solution composition, anodization parameters).

Surface Wettability
The nature of the interaction of the biomaterial with the tissue, related to each of the key properties of the surface layer mentioned above, is also determined by surface wettability. It determines protein adsorption, blood coagulation, and biological response. Most of the commonly used biomaterials are hydrophobic and have a high affinity for many proteins. Immediately after implantation, the biomaterial is covered with a layer of proteins. These are mainly albumin, fibrinogen, fibronectin, immunoglobulins, and von Willebrand factors. In shaping the surface properties of implants, wettability should also be taken into account, which is important for the osetoinduction process preceding osseointegration and affects the absorption of molecules favoring the adhesion of fibroblasts and/or bacteria [55,56]. Hydrophilic surfaces have better biological activity in Figure 6. CPD histograms determined for the Ti-13Nb-13Zr alloy and for the 1G, 2G, and 3G ONTs; solid lines-fit of the Gaussian function.
Approximation of the histograms using the Gaussian function allows determining the arithmetic average CPD av and the standard deviation, which in this case corresponds to the CPD rms parameter. Values of CPD sk and CPD ku also indicate that the CPD heights are symmetrically distributed around the average as well as that the investigated surfaces do not show inordinately high peaks/deep valleys. Generally, one can state that the entire surface of the Ti-13Nb-13Zr alloy is evenly covered with ONTs irrespective of the different production processes (solution composition, anodization parameters).

Surface Wettability
The nature of the interaction of the biomaterial with the tissue, related to each of the key properties of the surface layer mentioned above, is also determined by surface wettability. It determines protein adsorption, blood coagulation, and biological response. Most of the commonly used biomaterials are hydrophobic and have a high affinity for many proteins. Immediately after implantation, the biomaterial is covered with a layer of proteins. These are mainly albumin, fibrinogen, fibronectin, immunoglobulins, and von Willebrand factors. In shaping the surface properties of implants, wettability should also be taken into account, which is important for the osetoinduction process preceding osseointegration and affects the absorption of molecules favoring the adhesion of fibroblasts and/or bacteria [55,56]. Hydrophilic surfaces have better biological activity in contact with body fluids and, thus, provide better osseointegration [57]. The surface wettability tests of the produced ONTs layers showed that the contact angle measured for the Ti-13Nb-13Zr alloy without and with vertically oriented 1G, 2G, and 3G ONTs took the following J. Funct. Biomater. 2023, 14, x FOR PEER REVIEW mean square deviation the autocorrelation leng

Wettability Measurem
Water contact angl (Future Digital Scientifi sitting airdrop method. placed on the tested surf the mean values of t measurements conducte values of the Ѳ.

In Vitro Hemocompa
Hemocompatibilit investigated in accordan donors of blood (Regi Poland). Based on the In the hemoglobin (Hb) cal buffered saline (PBS) wa g L −1 . Specimens with th area to 1 mL of test blood ml human blood was ap h at 37(1) °C. After inc values in sequence: 62.9(9) • , 25.3(5) • , 11.9(1) • , and 14.7(2) • , respectively (Figure 7). contact with body fluids and, thus, provide better osseointegration [57]. The surface wettability tests of the produced ONTs layers showed that the contact angle measured for the Ti-13Nb-13Zr alloy without and with vertically oriented 1G, 2G, and 3G ONTs took the following Ѳ values in sequence: 62.9(9) °, 25.3(5) °, 11.9(1) °, and 14.7(2) °, respectively (Figure 7). The tests performed confirmed the favorable hydrophilic nature of the produced ONTs layers. Investigation of the contact angle of the individual sample surface showed that they were hydrophilic in nature both before and after anodization (Figure 7a-d). A reduction in Ѳ value was observed for the samples after anodization, which indicates an increase in hydrophilicity. The most advantageous hydrophilic properties showed the 2G ONTs, then 3G ONTs, 1G ONTs, and the smallest Ti-13Nb-13Zr alloy. The obtained wettability results of nanotubular surfaces on titanium are similar [58]. Using anodization as a method of electrochemical modification enhances the surface wettability of titanium and its alloys.

In Vitro Hemocompatibility Study
Currently, the most popular method of determining hemocompatibility is the hemolysis test. The process of hemolysis is based on the transition of hemoglobin into the cytoplasma due to the rupture of erythrocytes. According to the ASTM F756-17 standard [38], the biocompatibility of the 1G and 2G ONTs examinations were performed ( Figure  8). The tests performed confirmed the favorable hydrophilic nature of the produced ONTs layers. Investigation of the contact angle of the individual sample surface showed that they were hydrophilic in nature both before and after anodization (Figure 7a-d). A reduction in J. Funct. Biomater. 2023, 14, x FOR PEER REVIEW 4 of 24 mean square deviation (CPDrms), the skewness (CPDsk), the excess kurtosis (CPDku), and the autocorrelation length (CPDal) parameters were determined.

Wettability Measurements
Water contact angle measurements were carried out using an OCA 15EC goniometer (Future Digital Scientific Corp., Westbury, NY, USA) with an accuracy of ±0.01° by the sitting airdrop method. Moreover, 10 images of water drops with a volume of about 5 µL placed on the tested surface were recorded within 10 s. On the basis of the received images, the mean values of the contact angle (Ѳ) were determined. The average of five measurements conducted in different parts of the tested surface was taken as the final values of the Ѳ.

In Vitro Hemocompatibility Test
Hemocompatibility of the ONTs layers on the Ti-13Nb-13Zr alloy substrate was investigated in accordance with the ASTM F756-17 [38]. Healthy human volunteers were donors of blood (Regional Blood Donation and Blood Treatment Center, Katowice, Poland). Based on the International Committee for Standardization in Hematology (ICSH), the hemoglobin (Hb) calibration curve meeting the requirements was prepared. Phosphate buffered saline (PBS) was used for preparation of dilutions to the Hb concentration of 10(1) g L −1 . Specimens with the determined surface area were applied at a ratio of 3 cm 2 surface area to 1 mL of test blood solution. In case of samples with produced ONTs layers, 0.57(02) ml human blood was applied. Exposition of the final samples in human blood lasted for 4 h at 37(1) °C. After incubation for 4 h, the fluid was transferred to the test tubes and centrifuged at 3000 rpm for 15 min in a standard clinical centrifuge. The supernatant was removed, being careful not to disturb any button of erythrocytes in the test tube. Then, the solution absorbance was investigated using a spectrophotometer at a wavelength of 540 nm. Calculation of % hemolysis was based on the following Formula (1) (1)

Cell Culture and Cytotoxicity Assays
Cytotoxicity tests were performed in accordance with ISO 10993-5:2009 [39]. The percent viability for the test article and control article were determined from the blanks. Reduction of the number of living cells caused a decrease in the metabolic activity value was observed for the samples after anodization, which indicates an increase in hydrophilicity. The most advantageous hydrophilic properties showed the 2G ONTs, then 3G ONTs, 1G ONTs, and the smallest Ti-13Nb-13Zr alloy. The obtained wettability results of nanotubular surfaces on titanium are similar [58]. Using anodization as a method of electrochemical modification enhances the surface wettability of titanium and its alloys.

In Vitro Hemocompatibility Study
Currently, the most popular method of determining hemocompatibility is the hemolysis test. The process of hemolysis is based on the transition of hemoglobin into the cytoplasma due to the rupture of erythrocytes. According to the ASTM F756-17 standard [38], the biocompatibility of the 1G and 2G ONTs examinations were performed ( Figure 8). The value of the hemolytic index lower than 2 indicates that the biomaterial is nonhemolytic. The obtained results of in vitro hemocompatibility study shown in Figure  8 revealed the value of the hemolytic index of 0.05 for both the 1G and 2G ONTs. We found the hemolytic index for the 3G ONTs to be 0.00 [13]. The highest hemolytic index of 0.30 The value of the hemolytic index lower than 2 indicates that the biomaterial is nonhemolytic. The obtained results of in vitro hemocompatibility study shown in Figure 8 revealed the value of the hemolytic index of 0.05 for both the 1G and 2G ONTs. We found the hemolytic index for the 3G ONTs to be 0.00 [13]. The highest hemolytic index of 0.30 was reported for the non-anodized Ti-13Nb-13Zr alloy [13]. The obtained values prove that the proposed electrochemical modification ensures outstanding hemocompatibility. In the case of the 3G ONTs, it is even possible to completely eliminate hemolysis.

Cytotoxity and Cell Adhesion Assay
The interaction of the substrate material with a nanotubular structure with the biological environment can also be assessed during the use of cell cultures and cellular response on the surface prepared in this way. In the past, in order to evaluate the cytotoxicity of biomaterials for medicine, in vitro mammalian cell culture tests were used. At present, morphology, chemistry surface, and type of cells, especially concentration and time of exposure, determine the cytotoxicity of nanomaterials. The test article showed no cytotoxic potential to L-929 mouse fibroblast cells for the non-anodized Ti-13Nb-13Zr alloy and with the obtained 1G, 2G, and 3G ONTs layers (Table 4). In the case of oxide nanotubes produced on the vital Ti-13Nb-13Zr alloy by the anodizing method, the obtained cytotoxicity for mouse fibroblast cells is similar. The adhesion, proliferation, and migration of cells forming bone tissue are related to the diameters of nanotubes. This is confirmed by numerous studies showing that the ONTs surface is conducive to the growth of cells. It should be noted that different research groups used different cell colonies, which affected the results of the study [59][60][61]. Based on the prepared samples for cytotoxicity testing, cell adhesion to layers obtained via anodization and to the starting material was also tested. The cells adhere with different efficiency, which can be seen in Table 5. The obtained results of the adhesion test showed a different degree of adhesion depending on the tested oxide layer. For the self-passive oxide layer present on the Ti-13Nb-13Zr alloy and for the 1G ONTs, high cell adhesion was obtained. For the 2G and 3G ONTs, the medium adhesion was revealed. It can be correlated with the diameter and length of the obtained nanotubes in dependence of generation. Cell adhesion decreased with the growth and length of the ONTs. Table 5. Estimated cell adherence to the Ti-13Nb-13Zr alloy before and after formation of 1G, 2G, and 3G ONTs. ONTs layers are able to increase the bioactivity of the implant surface. First of all, they increase the surface roughness at the nano-scale, forming a biomimetic nanostructure, which is similar to natural rough bone tissue. Bone-forming cells (osteoblasts) have been proven in numerous studies [62,63] to adhere to a surface that is morphologically and chemically similar to natural bone tissue. The interaction of the nanotube-structured substrate material with the biological environment can also be assessed during the use of cell culture. Cell adhesion is related to the diameter of the nanotubes; research is shown in [64]. The authors showed that oxide nanotubes with a smaller diameter had a more stimulating effect on cell growth and differentiation.

Thrombogenicity Test
The alpha Kappendorf coefficient was 0.88, which proved very good compatibility with the thrombogenicity assessment carried out by the two experts. The relationship between the evaluations is shown in Table 6. Numbers in the table cells represent the number of samples. The result of the Kruskal-Wallis test was significant (p = 0.0000), which means that at least one group differs from the rest. Results of the post-hoc test are shown in Table 7 and Figure 9.  The result of the Kruskal-Wallis test was significant (p = 0.0000), which means that at least one group differs from the rest. Results of the post-hoc test are shown in Table 7 and Figure 9.  Class 0 meant very small thrombogenicity comparable with negative control. Class 5 meant very high thrombogenicity comparable with positive control. The meaning of a given class is described in Table 1, and each description was exemplified by images of the biomaterial surfaces (Bionate 55D polymer, PEEK, Ti6Al7Nb alloy with a layer of TiN, Figure 9. Thrombogenicity of the studied Ti-13Nb-13Zr alloy before and after formation of 1G, 2G, and 3G ONTs (in six-level ordinal scale). Class 0 meant very small thrombogenicity comparable with negative control. Class 5 meant very high thrombogenicity comparable with positive control. The meaning of a given class is described in Table 1, and each description was exemplified by images of the biomaterial surfaces (Bionate 55D polymer, PEEK, Ti6Al7Nb alloy with a layer of TiN, PEEK polymer, ZrO 2 +Y 2 O 3 ceramics, and oraz glass covered by collagen) dotted by thrombocytes and their aggregates. Usage of the interval scale limited the possibility of applying advanced statistical methods and made worse the quality of inference but, on the other hand, made carrying out a statistical proof possible.
In order to study how the expert assessments differ, the alfa Kappenrorf coefficient was used. Taken values from 0 to 1 where 0 means total discrepancy of the expert assessments and 1 means their total agreement. It is accepted in the literature data that a coefficient value larger than 0.8 means a very good agreement in the assessments. The alpha Kappendorf coefficient for the presented studies was 0.88; thus, the expert's opinions were practically equal. Further analysis was carried out with the use of the Kruskal-Wallis tests with the post-hoc multiple comparison method.
Thrombogenicity of the 1G and 3G ONTs layers is practically the same, and those biomaterials strongly unfavorably influence the activation of coagulation factors. They were classified into classes 4 and 5, respectively. Class 5 contains samples of positive control, so the 3G ONTs biomaterial activates the coagulation system to the same degree as strongly thrombogenic collagen. Athorombogenic properties of the 1G ONTs biomaterial are only slightly better than these of the 3G ONTs, but one should remember that there are no statistically important differences between those two materials (p = 0.1520-see Table 4). Thrombogenicity of the 2G ONTs and Ti-13Nb-13Zr does not differ; however, these biomaterials show much better athrombogenic properties than the 1G and 3G ONTs. The 2G ONTs and Ti-13Nb-13Zr were classified into classes 1 and 3, respectively. Class 1 contained biomaterials slightly worse than the positive control (Table 1). Thus, the 2G ONTs biomaterial has a small effect on the coagulation system. The Ti-13Nb-13Zr alloy is situated in class 3; however, there is no statistical proof that it differs from the 2G ONTs (p = 0.4582, compared to Table 4).

Drug Release Kinetics
To check the possibility of the potential use of ibuprofen-loaded Hep-DOPA/ONTs as carriers for the controlled release of the drug was compared the ibuprofen release profiles for three generations of oxide nanotubes on the Ti-13Nb-13Zr alloy obtained via anodization. The analysis of the obtained results showed a rapid increase in the released ibuprofen in the first 10 min for the 1G ONTs ( Figure 10). A continuous increase in drug content is observed for up to 30 min. After this time, the ibuprofen content began to decrease, and after 60 min, the drug content dropped to 0. In the first 10 min, ibuprofen is released from 16.25% for the 2G ONTs and 10.96% for the 3G ONTs. Between 20 and 30 min, there is an increase in drug release to about 35 and 36%. From 40 min, the release of the drug decreases to less than 10% at 60 min. The high dose released from the 1G ONTs in the first minutes proves the possibility of rapid anti-inflammatory and analgesic effects by releasing the substance from the nanotubular coating, which is necessary to relieve pain in peri-implantation tissues and reduce inflammation without using oral analgesic therapy [65]. In addition, the material to which the drug is applied must be biocompatible. Oxide nanotubes belong to a group of materials characterized by a developed and diversified surface, and thus a greater possibility of application of therapeutics [71]. In the case of ONTs of the 1G, 2G, and 3G with a single-walled structure, the drug is applied to the inside of the nanotube by attaching the drug to the surface with covalent amide or ester bonds [72,73]. Ibuprofen from the nanotubes is released through the so-called "burst effect", i.e., a rapid release of the drug within the first 15 min [71,74]. The primary mechanism of drug release is the Gallagher-Corrigan model. This model describes a twostage drug release kinetics. According to the Gallagher-Corrigan model, the first phase of drug release is its burst from the polymer Hep-DOPA matrix. The second stage is slow release, determined by matrix degradation. The rate of the drug release process in this type of drug delivery system is high in the initial period, then it decreases [71].

Conclusions
The obtained FE-SEM and TEM results confirmed the presence of the 1G, 2G, and 3G oxide nanotubes on the Ti-13Nb-13Zr alloy surface obtained via anodization in different conditions. The morphological parameters of the obtained oxide nanotubes increased with the increase in the fluorine ions content in the solution causing the increase in the surface roughness. The Ra parameter for the 3G ONTs was over 8 times higher in comparison with the 1G ONTs and 10 times higher in relation to the 2G ONTs. XPS studies showed that these nanotubular layers of three generations were composed of TiO2, Nb2O5, ZrO2, and ZrOx. The electronic properties of the 1G, 2G, and 3G ONTs tested by the SKP method depended on the content of F − ions. Statistical analysis of the contact potential difference maps showed that the quantitative parameters in the form of the arithmetic mean and the square root of the height irregularity increased with the increase in the content of fluorine ions in the solution from 0.5% to 4%. The 2G nanotubular oxide layers show the best hydrophilic properties. Biological assessment of hemocompatibility of the obtained ONTs revealed no hemolytic effects confirming the fulfillment of the requirements for clinical use. The best adhesion was shown by the 1G ONTs and the alloy before surface modification. The thrombogenicity of the 2G ONTs and Ti-13Nb-13Zr alloy does not differ from each other, but they have much better athrombogenic properties than the 1G and 3G ONTs. The ibuprofen release profile from the Hep-DOPA matrix from three types of oxide nanotube generation was investigated. A two-stage Gallagher-Corrigan model was selected to describe the kinetics of the ibuprofen release mechanism. Ibuprofen (C 16 H 18 O 2 ) is a non-steroidal anti-inflammatory drug, a derivative of propionic acid [66]. It has anti-inflammatory, analgesic, and antipyretic properties. It inhibits platelet aggregation. The action mechanism of ibuprofen is based on the inhibition of the activity of the COX-1 and COX-2 cyclooxygenase enzymes [66]. COX-2 enzyme activity increases rapidly in inflamed tissues. The inhibition of this enzyme by ibuprofen reduces the inflammatory process from the first moments of action. Orally administered ibuprofen has a number of side effects. Therefore, the administration of ibuprofen in the form of a drug release system reduces the risk of side effects without irritating the gastrointestinal mucosa, which is especially important in ulceration and bleeding from the stomach and duodenum [67]. Ibuprofen is a drug that cannot be used long-term, like antibiotics [67]. Therefore, the rapid release of a dose of ibuprofen in the first moment after implantation can effectively inhibit the action of the COX-2 enzyme and thus limit the development of the first inflammation in the tissues surrounding the implant. Orally administered ibuprofen disintegrates very rapidly. The disintegration time of ibuprofen in the liver is about 2 h; therefore, ibuprofen is an excellent substance for application in controlled drug release systems [68][69][70]. The drug immobilized inside the nanotubes must be soluble in water environments so that they dissolve in body fluids, in particular in blood. This condition must be met to achieve full biodistribution in the human body. The advantage of using ibuprofen is that it is 90% bound to plasma proteins [70].
In addition, the material to which the drug is applied must be biocompatible. Oxide nanotubes belong to a group of materials characterized by a developed and diversified surface, and thus a greater possibility of application of therapeutics [71]. In the case of ONTs of the 1G, 2G, and 3G with a single-walled structure, the drug is applied to the inside of the nanotube by attaching the drug to the surface with covalent amide or ester bonds [72,73]. Ibuprofen from the nanotubes is released through the so-called "burst effect", i.e., a rapid release of the drug within the first 15 min [71,74]. The primary mechanism of drug release is the Gallagher-Corrigan model. This model describes a two-stage drug release kinetics. According to the Gallagher-Corrigan model, the first phase of drug release is its burst from the polymer Hep-DOPA matrix. The second stage is slow release, determined by matrix degradation. The rate of the drug release process in this type of drug delivery system is high in the initial period, then it decreases [71].

Conclusions
The obtained FE-SEM and TEM results confirmed the presence of the 1G, 2G, and 3G oxide nanotubes on the Ti-13Nb-13Zr alloy surface obtained via anodization in different conditions. The morphological parameters of the obtained oxide nanotubes increased with the increase in the fluorine ions content in the solution causing the increase in the surface roughness. The Ra parameter for the 3G ONTs was over 8 times higher in comparison with the 1G ONTs and 10 times higher in relation to the 2G ONTs. XPS studies showed that these nanotubular layers of three generations were composed of TiO 2 , Nb 2 O 5 , ZrO 2 , and ZrO x . The electronic properties of the 1G, 2G, and 3G ONTs tested by the SKP method depended on the content of F − ions. Statistical analysis of the contact potential difference maps showed that the quantitative parameters in the form of the arithmetic mean and the square root of the height irregularity increased with the increase in the content of fluorine ions in the solution from 0.5% to 4%. The 2G nanotubular oxide layers show the best hydrophilic properties. Biological assessment of hemocompatibility of the obtained ONTs revealed no hemolytic effects confirming the fulfillment of the requirements for clinical use. The best adhesion was shown by the 1G ONTs and the alloy before surface modification. The thrombogenicity of the 2G ONTs and Ti-13Nb-13Zr alloy does not differ from each other, but they have much better athrombogenic properties than the 1G and 3G ONTs. The ibuprofen release profile from the Hep-DOPA matrix from three types of oxide nanotube generation was investigated. A two-stage Gallagher-Corrigan model was selected to describe the kinetics of the ibuprofen release mechanism.