Materials Science and Engineering C 58 (2016) 918–926
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Investigating the structure and biocompatibility of niobium and titanium oxides as coatings for orthopedic metallic implants D. Pradhan, A.W. Wren, S.T. Misture, N.P. Mellott ⁎ Inamori School of Engineering, Alfred University, Alfred, NY 14803, USA
a r t i c l e
i n f o
Article history: Received 4 May 2015 Received in revised form 3 September 2015 Accepted 13 September 2015 Available online 14 September 2015 Keywords: Coatings Ceramics Biocompatibility
a b s t r a c t Applying sol gel based coatings to orthopedic metallic implant materials can significantly improve their properties and lifespan in vivo. For this work, niobium (Nb2O5) and titanium (TiO2) oxides were prepared via solution processing in order to determine the effect of atomic arrangement (amorphous/crystalline) on bioactivity. Thermal evaluation on the synthesized materials identified an endotherm for Nb2O5 at 75 °C with 40% weight loss below 400 °C, and minimal weight loss between 400 and 850 °C. Regarding TiO2 an endotherm was present at 92 °C with 25% weight loss below 400 °C, and 4% between 400 and 850 °C. Phase evolution was determined using High Temperature X-ray Diffraction (HT-XRD) where amorphous-Nb2O5 (450 °C), hexagonal-Nb2O5 (525 °C), orthorhombic-Nb2O5 (650 °C), amorphous-TiO2 (275 °C) and tetragonal TiO2 (500 °C) structures were produced. Simulated body fluid (SBF) testing was conducted over 1, 7 and 30 days and resulted in positive chemical and morphological changes for crystalline Nb2O5 (525 °C) and TiO2 (500 °C) after 30 days of incubation. Rod-like CaP deposits were observed on the surfaces using Scanning Electron Microscopy (FE-SEM) and Grazing Incidence-X-ray Diffraction (GI-XRD) shows that the deposits were X-ray amorphous. Cell viability was higher with the TiO2 (122%) samples when compared to the growing cell population while Nb2O5 samples exhibited a range of viability (64–105%), partially dependent on materials atomic structure. © 2015 Elsevier B.V. All rights reserved.
1. Introduction In the recent past, metallic materials such as stainless steel have been employed for load-bearing orthopedic applications for the repair or replacement of damaged or diseased joints such as total hip replacement (THR) and total knee replacement (TKR). However, a persistent concern relates to the interfacial response between the host tissue and the metallic component, which typically does not bio-integrate and can result in loosening and subsequent implant removal. Recently however, techniques have been developed to improve integration between the host tissue and the metallic component such as surface roughening and etching, thermal oxidation treatments and also through applying coatings to the surface of the metallic implant. Metal/alloys such as Ti6Al4V are known to be bioactive, [1,2] however, other widely used metals such as stainless steel are considered to be bioinert (no tissue integration), [3,4] or induces an allergic reaction (nickel, cobalt–chromium) upon implantation [5,6]. The biofunctionality of the surface of metallic materials can be altered through thermal oxidation treatments [7]. These thermal treatments can increase the performance of implants by inducing the formation of a passive oxide layer, improving the corrosion and wear resistance of coatings. In addition, the biocompatibility of thermally ⁎ Corresponding author at: 2 Pine Street, Alfred University, Alfred, NY 14802, USA. E-mail address: [email protected] (N.P. Mellott).
http://dx.doi.org/10.1016/j.msec.2015.09.059 0928-4931/© 2015 Elsevier B.V. All rights reserved.
oxidized metals, including titanium alloys, can also induce biocompatibility [8,9]. Coating metals with transition metal oxides can also increase the materials biocompatibility which can be applied to both orthopedic surgery and oral and maxillofacial restoration [10–12]. Traditionally, implant functionality and success was largely dependent upon the surface chemistry, mechanical stability, and biological compatibility of the implant material [12–13]. Oxide surface coatings can be employed to promote bioactivity or improve biocompatibility of the host implant material and given that the oxide coating is in direct contact with the host tissue, it has a profound effect on the chemical and biological processes which occur at that interface [14–15]. In particular, the oxide coatings atomic arrangement (crystalline, amorphous), dissolution chemistry and stability can affect not only the resistance to fatigue and high tensile stresses, but also biological effects such as cell adhesion, osteoblast proliferation, the inflammatory response and biomineralization. It has been proposed that if the properties of an oxide coating could be characterized, controlled and predicted, then enhanced properties can be engineered into an implant material [14]. This would facilitate the ability to tailor an implants bioactivity in addition to controlling the physico-chemical/biological processes at the coating-tissue interface. Depositing oxide coatings on metallic implants can be achieved using a variety of techniques including physical vapor deposition (PVD), chemical vapor deposition (CVD), and various sol–gel based
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
deposition techniques [11–13,16,17]. In particular, sol–gel dip coating followed by calcination has advantages that include the ability to control the coatings purity and chemistry. In addition, this processing method affords scalability and the ability to coat irregular shaped implant surfaces. Titanium (TiO2) and niobium oxide (Nb2O5) systems have shown particular promise as candidate materials for bioactive coatings on metallic implants [13,18,19]. In comparison to TiO2, Nb2O5 has been less extensively studied [13,20], however, the properties exhibited by this oxide, such as high corrosion resistance and thermodynamic stability, supports its potential as a coating for medical implants [14,21]. Eisenbarth et al. conducted studies which determined that osteoblasts preferentially adhered to the roughness of nanosized Nb2O5 coatings on polished CP Titanium [20]. A study by Ramirez et al. on coating stainless steel dental implants with crystalline (c-NbN) and amorphous (a-Nb2O5) also supports this work where they determined enhanced viability, proliferation and attachments of cementoblasts. The coatings were found to improve the surface hardness, corrosion resistance and the overall biological response of the stainless steel implants [12]. It has been reported previously that specific properties can be induced by crystal structure and atomic arrangement and that surface crystallinity plays an important role in structure–bioactivity relationships of oxide systems [14]. Ramirez et al. demonstrated that crystalline Nb2O5 shows greater CaP deposition when incubated in simulated body fluid (SBF) than its amorphous counterpart [14]. Studies by Maeda et al. demonstrated that apatite formation is crystal phase dependent, where more prominent apatite deposition was evident on anatase-TiO2 rather than rutile-TiO2 [22] and Eisenbarth et al. also demonstrated that osteoblast response to Nb2O5 is a function of crystal structure [20]. With respect to this study, TiO2 and Nb2O5 oxides are prepared in powder form with controlled atomic arrangement. Powders were used in lieu of coatings for ease of characterization. However, results here are directly analogous to coatings prepared in the same manner, as coatings could be applied through spray-based deposition of dispersed powders or through spin or dip coatings of sols prior to drying and heating. It is worth mentioning here that for the utilization of these powders for spray deposition, care must be taken in both the understanding and minimization of the effect of the coating deposition process on both implant and powder structure and morphology. Powders were characterized and evaluated using High-Temperature X-ray Diffraction (HT-XRD) and Differential Thermal Analysis (DTA). The biocompatibility of both amorphous and crystalline analogs of TiO2 and Nb2O5 is also investigated with respect to time in a synthetic biological fluid that resembles the ionic composition of human blood plasma (simulated body fluid — SBF). SBF is considered to be a strong indicator of a materials' bioactivity as precipitation of calcium and phosphate ion from the SBF on the materials' surface is considered to be a precursor to bone bonding in vivo. Any changes in surface chemistry and morphology, attributed to materials atomic arrangement, were identified using Field Emission Scanning Electron Microscopy (FESEM) and Grazing Incidence X-ray Diffraction (GI-XRD). Finally, the cytocompatibility of the materials was investigated using a model cell line, L929 Mouse Fibroblasts, to determine if TiO2 and Nb2O5 atomic arrangement influences cytotoxicity. This effect was also conducted with respect to maturation in an aqueous environment. 2. Materials and methods 2.1. Material synthesis Titanium and niobium based sols were prepared individually using titanium (IV) isopropoxide (Ti (OC3H7)4, 2.96 ml) 99.99% and niobium (V) ethoxide (Nb (OC2H5)5, 2.50 ml) 99.95% (Sigma-Aldrich) respectively, denatured ethyl alcohol (C2H5OH, Fisher Scientific) and glacial acetic acid (CH3COOH, Fisher Scientific). In each case, ethanol (58.39 ml) was taken in a beaker and glacial acetic acid was added
919
(5.72 ml) while stirring. After 15 min the titanium and niobium precursors (2.96 ml and 2.50 ml respectively), were added to the mixture dropwise. The solutions were covered with perforated parafilm and magnetically stirred until solutions were clear and transparent (~2 h). The stable solutions were then heated at 60 °C with simultaneous stirring until the liquid evaporated and powders were obtained (36– 48 h). The powders were dried at 150 °C for 24 h in a conventional oven and then crushed using an agate mortar and pestle. Dried powders were then segregated into two separate batches; (1) the first batch was stored, as-dried, in a vacuum desiccator for DTA/TGA and in-situ HTXRD analysis and (2) the second batch was calcined from 200 to 650 °C with a heating rate of 20 °C/min and a 90 min dwell period every 25 °C in a Thermo Scientific Thermolyne furnace (model no. F48025-6080). Powders were collected at the end of each dwell period and stored in separate containers for analysis. 2.2. Material characterization 2.2.1. Differential Thermal Analysis/Thermogravimetric Analysis (DTA– TGA) This analysis was performed using an SDT 2960 simultaneous DSC– TGA. The measurements were obtained from 30 °C to 850 °C under flowing air at a heating rate of 20 °C/min. 2.2.2. High Temperature X-ray Diffraction (HT-XRD) Diffraction patterns were obtained using a Bruker D8 advance diffractometer (Cu-Kα radiation) equipped with a Vantec detector with Gobel mirror, 1.0 mm divergence slit and Anton Paar high HTK 1200 temperature furnace attachment. Measurements were made over 10 to 140° 2θ with a step size of 0.03° 2θ and a scan rate of 10° 2θ/min. The high temperature experiments were carried out using a heating rate of 20 K/min, with dwell period of 90 min after every 25 K increment. The measurements were taken at the end of each dwell period and Jade 9 software (Materials Data Inc., USA) was used for analysis. 2.2.3. Raman spectroscopy The spectra were obtained from a Witec Confocal Raman Microscope CRM200 equipped with Si detectors, green laser (excitation wavelength of 532 nm and power of 70 mW), and a dispersion grating of 600 l/mm. 2.3. Evaluation of material bioactivity and cytocompatibility 2.3.1. Simulated body fluid (SBF) Simulated body fluid testing was conducted in accordance with the procedure outlined by Kokubo [23]. The reagents and their respective amount were dissolved in 700 ml of deionized water using a magnetic stirrer and the solution was maintained at 36.5 ± 0.5 °C. The pH of the solution was adjusted to 7.4 by adding Tris (hydroxymethyl) aminomethane and 1 M-HCl [23]. The oxide powders were first crushed and sieved to obtain uniform sized particles below 30 μm. Selected powders were then pressed into pellets of dimensions (8.06 mm × 1.3 mm) by applying pressure of 1000 psi with holding time of 30 s. The disk specimens were soaked in SBF for 1, 7, and 30 days. The volume of SBF for immersion of pellets was calculated on the basis of the surface area (diameter and height) of the disks. After soaking in SBF for the required time period, the disks were removed and gently rinsed with deionized water and stored in incubator at 37 °C for 24 h. 2.3.2. Scanning Electron Microscopy (SEM) The analysis was performed on disks reacted in SBF solution for 0, 1, 7, and 30 days using an FEI Quanta 200F equipped with a field emission gun. Analyses were performed at an accelerating voltage of 12.5 kV and beam current of 26 nA. In addition, semi-quantitative analysis was performed using energy-dispersive X-ray spectroscopy (EDX).
920
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
Fig. 1. DTA/TGA plots for primary a.) Nb2O5 and b.) TiO2 powders.
2.3.3. Grazing Incidence X-ray Diffraction (GIXRD) GIXRD analyses were performed on selected disks using a Siemen's Kristalloflex diffractometer using Copper Kα X-rays and a BraggBrentano Goniometer. The pellets were mounted on the sample holder using X-ray amorphous putty. Data was collected from 5° to 70° 2θ in 0.02° steps, 10 s each. Jade software (v 6.0.3, Materials Data Inc.) was used for analysis. 2.3.4. Specific surface area (SSA) The selected powders were measured via N2 gas adsorption according to the BET method with a micromeritics Tristar 3000. Prior to analysis powders were degassed in a micromeritics flow prep 060 sample degas system at 150 °C for 24 h. The instrument was standardized prior to and after analysis using silica-alumina rod standard (201 ± 5 m2/g). 2.3.5. Cytotoxicity analysis The experiments were performed using mouse fibroblast cell lines L929 (American Type Culture collection CCL 1, NCTC clone 929; ISO10993 part 5). Cells were initially cultured in Hyclone medium 199/EBSS (Thermo Scientific) supplemented with 10% fetal bovine serum (FBS).
Cells were maintained on regular feeding in a CO2 incubator at 37 °C/ 5% CO2 atmosphere/100% relative humidity. Media was changed every 2 days. At confluence, cells were detached using a trypsin treatment. The cells were centrifuged and re-suspended in a media containing 10% FBS. Finally cells were seeded onto a 24 well plate at density of 10,000 cells per well and incubated for 24 h prior to testing. The sample extracts were collected by dissolving the weight of powder equivalent to 1 m2 BET surface area (not shown) in 10 ml deionized sterile water. MTT assay was used to estimate the cell viability. 100 μl of sterile sample extracts collected after 1, 7 and 30 days (n = 3) was added into wells containing L929 cells in culture medium (1 ml). The prepared well plates were incubated for 24 h at 37 °C/5% CO2. The MTT assay was then added in an amount equal to 10% of the culture medium volume/ well. The cultures were then re-incubated for a further 4 h (37 °C/5% CO2). The cultures were removed from the incubator and the resultant formazan crystals were dissolved by adding an amount of MTT Solubilization Solution (10% Triton X-100 in acidic isopropanol (0.1 N HCI) equal to the original culture medium volume. Once the crystals were fully dissolved, the absorbance was measured at a wavelength of 570 nm. Aliquots (100 μl) of tissue culture water were used as controls, and cells were assumed to have metabolic activities of 100%.
Fig. 2. In-situ HT-XRD patterns for a.) Nb2O5 b.) TiO2 powders (indexing of individual phases can be observed in Fig. 3).
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
921
Table 1 X-ray diffraction characteristics of crystalline structures. Temp (°C)
PDF#
Chem. ID
Symmetry
a (A°)
b (A°)
c (A°)
Space grp.
450 °C 525 °C 650 °C 275 °C 500 °C
– 00-028-0317 00-030-0873 – 04–006–9240
– Nb2O5 Nb2O5 – TiO2
Amorphous Hexagonal Orthorhombic Amorphous Anatase
– 3.607 6.175 – 3.784
– 3.607 29.175 – 3.784
– 3.925 3.930 – 9.495
– P Pbam (55) – I41/amd (I41)
2.4. Statistical analysis One-way analysis of variance (ANOVA) was employed to compare the cell viability of the experimental materials in relation to 1.) maturation (1, 7, 30 days) and 2.) sample composition. Comparison of relevant means was performed using the post hoc Bonferroni test. Differences between groups was deemed significant when p ≤ 0.05.
3. Results and discussion 3.1. Material characterization and phase identification Material characterization was conducted to determine each composition's thermal profiles and properties, and to accurately identify suitable heat treatment profiles for the powdered Nb2O5 and TiO2 samples. Thermal analysis (DTA/TGA) plots are presented in Fig. 1 for the sol gel derived niobium oxide and titanium oxide powdered samples. Regarding niobium-based powders (Fig. 1a) an endotherm was observed at 75 °C followed by exotherms at approximately 279, 344 and 579 °C. The TGA curve showed approximately 40% weight loss between 27 and 400 °C, with minimal weight loss observed between ~ 400 and 850 °C. For titanium oxide powders (Fig. 1b) an endotherm was observed at 92 °C followed by exotherms at approximately 340, 412 and 522 °C. The TGA curve showed an approximately 25% weight loss between 27 and 400 °C, with an additional 4% weight loss observed between 400 and 850 °C. The first endotherm observed for both samples is attributed to elimination of adsorbed water [24]. This is further evidenced by the large weight loss observed for both sample systems, between room temperature and ~ 250 °C. Exothermic peaks observed at both 279 °C and 344 °C for the niobium based sample and at 340 °C for the titanium based sample are associated with the removal of organics [25–26]. This is also consistent with the additional weight loss observed up to ~400 °C [27]. It has been reported that the organic species are typically compounds of alcohols including C3H7, C4H8, and
adsorbed acetic acid used in the preparation of sols [18,28,29]. Upon further heating of the niobium oxide powders a sharp exothermic peak is observed at 579 °C, associated with the crystallization to hexagonalNb2O5 [17]. Upon further heating of the titanium oxide based samples, exothermic peaks are observed at 412 °C and 522 °C, presumably associated with the crystallization to anatase-TiO2 followed by either (1) a phase transformation to rutile-TiO2, or (2) removal of residual organics [30,31]. It is well understood that the exact crystallization temperatures of these oxides are a strong function of the chemistry of the sol, particle shape, as well as the heating rate of the sample, however as noted the crystallization temperatures here are in general agreement with those previously reported of similar systems [17,30,31]. High-Temperature X-ray Diffraction (HT-XRD) was employed to identify and synthesize particular phases/atomic structures and the results are presented in Fig. 2. The HT-XRD patterns in Fig. 2a present the phase evolution of as-prepared niobium oxide powders heated insitu from 30 to 650 °C. The niobium oxide powders remained X-ray amorphous up to 475 °C, with initial crystallization occurring between 475 and 500 °C. A mixture of amorphous and crystalline phases was observed after heat treatment at 500 °C which was evident from the sharp peaks as well as an uneven hump within the background of the pattern. For temperatures above 500 °C, patterns exhibit sharp, well defined peaks suggesting highly crystalline Nb2O5. The three most intense peaks at 22.7°, 28.6° and 36.7° 2θ remained unchanged between 500 and 650 °C. However, at ~600 °C additional peaks were observed at 2θ 17.4° and also at 32.6°, 42.7°, and 45.3° 2θ with intensity of peaks increasing with calcination temperature. Phase ID HT-XRD patterns, show that the crystal phase observed from 500 to 575 °C is hexagonalNb2O5 (TT-Nb2O5; PDF# 00-028-0317). This transformation temperature is in agreement with what has been reported in the literature [15, 32]. Upon heating above 575 °C, a phase transformation from a hexagonal to an orthorhombic-Nb2O5 (T-Nb2O5; PDF#-000-030-0873) was observed. The crystallite size of TT-Nb2O5 increased with temperature from 23 ± 4 nm (500 °C) to 37 ± 6 nm (550 °C) while the crystallite size of T-Nb2O5 increased from 43 ± 3 nm (575 °C) to 74 ± 3 nm
Fig. 3. Individual phase transformation of In situ HT-XRD patterns for a.) Nb2O5 and b.) TiO2 powders.
922
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
Fig. 4. Raman spectra of a.) Nb2O5 powders and b.) TiO2 powders showing different phase developments with temperature.
(650 °C), as determined by Rietveld refinement. The HT-XRD patterns in Fig. 2b present the phase evolution of as-prepared titanium oxide powders heated in-situ from 29 to 650 °C. The powders remained X-ray amorphous up to 325 °C. Upon crystallization, the three most intense peaks were observed at 25.6°, 48.3° and 48.3° 2θ. All peak positions for the crystalline phases remained at approximately the same position with calcination from 350 to 650 °C. However, with increasing temperature, the observed peaks become sharper and more well-defined. Upon phase ID analysis, it was determined the crystalline phase observed at 350 °C was tetragonal-TiO2 (anatase; PDF# 04-006-9240), with this phase remaining stable until 650 °C. The initial crystallization temperature for anatase, between 325 and 350 °C observed here is in agreement with the reports which suggest that this transformation typically occurs between 300 and 400 °C [29,33,34]. The crystallite size of tetragonalTiO2 increased with temperature from 22 ± 1 nm (350 °C) to 45 ± 1 nm (650 °C) as determined by Rietveld refinement. Three primary phases were observed via HT-XRD when heated to 650 °C for niobium oxide: amorphous-Nb2O5, hexagonal-Nb2O5, and orthorhombic-Nb2O5, while two phases were observed for titanium oxide: amorphous-TiO2 and tetragonal-TiO2; as summarized in Table 1. Individual HT-XRD patterns associated with these samples are shown in Fig. 3a &b respectively. These five structures were then chosen for further structural and biocompatibility analysis and will be referred to from this point forward as; Nb2O5-450 °C, Nb2O5-525 °C, Nb2O5-650 °C, TiO2-275 °C, and TiO2500 °C respectively. Raman spectra for the selected samples are shown in Fig. 4. Nb2O5450 °C showed spectra with broad bands centered at ~ 220 cm−1 and ~670 cm−1 with a shoulder observed at the low wavelength side of the 200 cm−1 band and high wavelength sides of the ~670 cm−1 band, all associated with Nb–O vibrational modes [25,32]. The broad bands are consistent with amorphous, disordered Nb2O5 [35]. While one could argue that these spectra may be indicative of a poorly crystalline Nb2O5, given the complete lack of crystallinity observed in the X-ray Diffraction patterns associated with this sample (Fig. 3a), the authors believe that it is appropriate to conclude that this sample is amorphous. Nb2O5-525 °C showed sharper more defined bands centered at ~ 147, 233, 278, 290 and 691 cm−1 along with weak shoulders on the high frequency side of the ~290 and ~691 cm−1 band (Fig. 4a.) consistent with crystallization. Bands are assigned to the bending of Nb–O–Nb, stretching of niobia polyhedra and stretching of Nb_O modes of crystalline (both hexagonal and orthorhombic) Nb2O5 [28,36]. In particular, the crystallization of powders to pseudohexagonal phase (TT-phase) at 525 °C is evidenced by the narrowing of broad peaks and simultaneous shifting of bands from 670 to 691 cm−1 due to increased bond order of niobia
polyhedra [33]. Further, the absence of a shoulder at the high frequency side of the ~691 cm−1 band also confirmed the presence of the crystalline phase [25]. Bands associated with Nb2O5-650 °C are equal in both position and shape to those observed for Nb2O5-525 °C with the possible addition of a band at ~450 cm−1. It is well understood that the Raman shift position is sensitive to changes in atomic structure/arrangement, therefore the lack of shift observed here between Nb2O5-525 °C and Nb2O5-650 °C is consistent with the previously reported observation that the two structures are very similar [33,37]. TiO2-275 °C showed broad, weak bands at ~ 183, 206, 421, 526, 609 and 957 cm−1. These broad, undefined bands and their positions are consistent with a disordered, amorphous TiO2 [35]. Furthermore, the weak shoulder at 957 cm−1 observed for amorphous TiO2 could be associated with the presence of adsorbed surface species which tends to disappear at higher calcination temperature. In comparison, TiO2-500 °C exhibited sharp, well defined bands at 147, 197, 394, 518 and 639 cm−1 (Fig. 4b). These band positions are in good agreement with those shown to be Raman active modes associated with tetragonal-TiO2; 144, 197, 397, 518, and 640 cm−1 [37,38]. The first three bands corresponded to O–Ti–O bending mode and the latter two were assigned to Ti–O stretching modes [37,38]. Characterization and identification of the structural evolution of the niobium and titanium oxide powders, with increasing temperature, was studied by complimentary analytical techniques including DTA/ TGA, HT-XRD, and Raman spectroscopy. The results are consistent with one another, as well as being consistent with previous reports of similarly prepared systems [12,35,36]. It was shown that the niobium oxide powders were initially amorphous and remained so until 500 °C. Between 500 and 600 °C a hexagonal crystal phase was observed (TT-Nb2O5) followed by an orthorhombic phase (T-Nb2O5), observed between 600 and 650 °C. It was shown that titanium oxide powders were initially amorphous and remained so until 350 °C. Between 350 and 650 °C a tetragonal crystal phase was observed (anatase). It is worth noting that the phase transformation between hexagonalNb2O5 and orthorhombic-Nb2O5 is not evidenced in the DTA/TGA plots via a sharp exotherm, however this is expected, as hexagonalNb2O5 is thought to be a modification of orthorhombic-Nb2O5 with low crystallinity [36,40]. Furthermore, HT-XRD results suggest that the DTA exotherm observed at 522 °C (Fig. 1b) is due to removal of residual organics and not an anatase to rutile transformation. 3.2. Determination of material biocompatibility Preliminary evaluation of the materials' bioactivity was conducted using simulated body fluid (SBF) testing on the samples Nb2O5-450 °C,
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
923
Fig. 5. SEM images and corresponding EDX spectra for Nb2O5 obtained after immersion in SBF for 30 days for a.) Nb2O5-450 °C, b.) Nb2O5-525 °C and c.) Nb2O5-650 °C.
Nb2O5-525 °C, and Nb2O5-650 °C (Fig. 5) and TiO2-275 °C and TiO2-500 °C (Fig. 6). SBF testing has become a widely recognized method of investigating a materials' bioactivity as it is a solution that contains the ionic composition analogous to human blood plasma, and surface precipitation reactions in SBF are regarded as a precursor to bone bonding in vivo [23,41,42]. This method aims to determine changes in surface chemistry and morphology when samples are incubated in SBF as a function of time. SBF testing was conducted on pressed disks of the selected samples and was incubated for 1, 7, and 30 days. After each time period, each composition/structure was analyzed for
morphological (FE-SEM) or chemical (EDX) evidence of calcium and/ or phosphorous (CaP) containing precipitates. No evidence of precipitation is observed for any disks reacted for 1 or 7 days. After 30 days of immersion in SBF, Nb2O5-450 °C, Nb2O5-650 °C, and TiO2-275 °C remain free of CaP precipitation. However, both morphological evidence and chemical evidence of precipitation are observed for Nb2O5-525 °C and TiO2500 °C after immersion for 30 days in SBF. Nanometer to micron size particles are clearly observed via SEM with both Ca and P detected on the surface via EDX (Fig. 5 & Fig. 6). When observed at higher resolution (50 kx) it is clear that the morphology of the precipitates is different, as
924
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
Fig. 6. SEM images and corresponding EDX spectra for TiO2 obtained after immersion in SBF for 30 days a.) TiO2-275 °C and b.) TiO2-500 °C.
Nb2O5-525 °C is larger and is plate like crystals where TiO2-500 °C crystals are smaller rod-shaped precipitates (Fig. 7). GI-XRD analysis is presented in Fig. 8 and was performed on materials with precipitates to determine if any new crystalline phases were evident after 30 days of incubation in SBF. However, diffraction patterns show only evidence of crystalline Nb2O5 or TiO2 and not any additional phases, which suggests that this surface precipitant is amorphous CaP, the precursor to crystalline apatite. Cytotoxicity testing was conducted using L929 Mouse Fibroblasts to determine the cell viability after exposure to each of the samples, and is presented in Fig. 9a & b. The MTT assay determined changes in cell
growth and the metabolic activity when exposed to the experimental materials' condition medium when compared to a healthy growing cell population. Cell viability testing for TiO2 is presented in Fig. 9a. In comparison to the Control cell population, which was assumed to have a metabolic rate of 100%, cell viability for TiO2-275 °C showed an increase (106–122%) over 1–30 days, however, this increase was not found to be significant when statistically compared to the Control cell population (p = 0.292). For TiO2-500 °C, cell viability increased initially to 110% after 7 days and the reduced to 83% after 30 days, which was also not significantly reduced when compared to the Control cell population (p = 1.000). Cell viability testing for Nb2O5 is presented in Fig. 9a.
Fig. 7. SEM images of CaP layer deposition on the surface of a.) hexagonal-Nb2O5 (525 °C), and b.) TiO2-tetragonal (anatase, 500 °C) after immersion in SBF for 30 days.
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
Fig. 8. GI-XRD of a.) hexagonal-Nb2O5 (525 °C), and b.) TiO2-tetragonal (anatase, 500 °C) after soaking in SBF for 30 days.
Regarding Nb2O5-450 °C exhibiting hexagonal phase, there was no significant change in viability observed after 30 days (p = 1.000). For Nb2O5-525 °C a significant decrease in cell viability was observed at 1 (p = 0.003) and 7 days (p = 0.010) at 64% and 68% respectively. However, at 30 days the metabolic activity recovered to 71% which was not considered significant when compared to the Control cells (p = 0.063). Nb2O5-650 °C exhibited a decrease in cell viability at 1 day (70%, p = 0.024) but increased to 100% after 30 days which is not significantly different than the Control cells (p = 1.000). Although the amorphous materials presented slightly higher cell viability than the crystalline analogs, (with the exception of Nb2O5-650 °C), the crystallographic phases did not overall present significant differences in cell viability compared to their crystalline counterparts. Also, there was no significant difference in cell viability when considering the effect of maturation between each material at 1, 7 and 30 days (p = 1.000). It was shown that within both Nb2O5 and TiO2 oxide systems that the atomic arrangement influences the materials' bioactivity. Within the TiO2 system the tetragonal phase exhibited bioactivity while the amorphous did not. This is in agreement with previous studies by Ghaith et al., where crystalline TiO2 (anatase) structures were found to promote CaP deposition over the amorphous counterpart [19]. In fact, it has been shown that it is not just the presence of crystallinity that affects bioactivity, but the actual crystal structure; anatase TiO2 has been shown to be more bioactive than rutile TiO2 [23,43]. We observe the
925
same phenomena here within the Nb2 O 5 system, as hexagonalNb 2 O 5 exhibited bioactivity and both amorphous-Nb 2 O 5 and orthorhombic-Nb 2O 5 did not. However, no correlation between oxide structure and cell viability was observed, with all oxides performing equally to the control cell population, which indicates that no toxicity was evident. Future work will include depositing oxide coatings of equal composition and structure on metallic substrates (stainless steel) to test their chemical durability and biofunctionality in various biologically relevant solutions. Coatings will be deposited through the spray coating of dispersed oxide powders, as prepared here, or via spin coating of the sols described here onto the stainless steel substrate followed by calcination. In our case it would be advantageous to deposit dispersed powders, assuming that a deposition method which will not change the structure or composition of the powders can be developed. Otherwise, we will develop a processing method, in which we deposit the sol on stainless steel followed by spin coating and calcination. This will require the development of a calcination maximum temperature and ramp rate for the coatings which will result in the same structure and composition as the powders reported here.
4. Conclusions In this work, Nb2O5 and TiO2 were prepared via wet chemistry. Structural evolution with calcination temperature was determined and amorphous-Nb2O5, hexagonal Nb2O5, orthorhombic-Nb2O5, amorphous TiO2, and tetragonal TiO2 were then tested for bioactivity and cytocompatibility. After immersion in SBF-solution for 30 days only TiO2-tetragonal and Nb2O5-hexagonal exhibited precipitation of CaPbased materials, a positive indicator of bioactivity. Cytocompatibility results show that none of the oxides tested had a negative or toxic effect on cell viability, and that the amorphous TiO2 and Nb2O5 were more cytocompatible than the crystalline analogs. Results from this work not only shed light on the structure–bioactivity properties of these oxides but also can serve as template to aid in the preparation of bioactive Nb2O5 and TiO2 materials.
Acknowledgment The authors would like to acknowledge the Pennsylvania State University Materials Characterization Laboratory for Raman analyses. This work was partially supported by the 3M Non-Tenured Faculty Award (NPM).
Fig. 9. MTT assay of all Nb2O5 and TiO2 structures extracts in L929 Fibroblasts after 1, 7 and 30 days of incubation.
926
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
References [1] H. Takadama, et al., XPS study of the process of apatite formation on bioactive Ti– 6Al–4V alloy in simulated body fluid, Sci. Technol. Adv. Mater. 2 (2001) 389–396. [2] Y.-T. Sul, et al., The bone response of oxidized bioactive and non-bioactive titanium implants, Biomaterials 26 (2005) 6720–6730. [3] K. Hayashi, et al., Re-evaluation of the biocompatibility of bioinert ceramics in vivo, Biomaterials 13 (4) (1992) 195–200. [4] S. Kumar, D. Simpson, R.S.C. Smart, Plasma processing for inducing bioactivity in stainless steel orthopaedic screws, Surf. Coat. Technol. 202 (4–7) (2007) 1242–1246. [5] Hussain, O.T., S. Sah, and A.J. Sidebottom, Prospective comparison study of one-year outcomes for all titanium total temporomandibular joint replacements in patients allergic to metal and cobalt–chromium replacement joints in patients not allergic to metal. Br. J. Oral Maxillofac. Surg. 52(1): p. 34–37. [6] H.F. Hildebrand, C. Veron, P. Martin, Nickel, chromium, cobalt dental alloys and allergic reactions: an overview, Biomaterials 10 (8) (1989) 545–548. [7] H. Güleryüz, H. Çimenoğlu, Effect of thermal oxidation on corrosion and corrosion– wear behaviour of a Ti–6Al–4V alloy, Biomaterials 25 (16) (2004) 3325–3333. [8] M.C. Garcia-Alonso, et al., In vitro corrosion behaviour and osteoblast response of thermally oxidised Ti6Al4V alloy, Biomaterials 24 (1) (2003) 19–26. [9] C.-M. Lin, S.-K. Yen, Biomimetic growth of apatite on electrolytic TiO2 coatings in simulated body fluid, Mater. Sci. Eng. C 26 (1) (2006) 54–64. [10] S. Nagarajan, V. Raman, N. Rajendran, Synthesis and electrochemical characterization of porous niobium oxide coated 316L SS for orthopedic applications, Mater. Chem. Phys. 119 (3) (2010) 363–366. [11] A. Ochsenbein, et al., Osteoblast responses to different oxide coatings produced by the sol–gel process on titanium substrates, Acta Biomater. 4 (5) (2008) 1506–1517. [12] G. Ramírez, et al., Niobium based coatings for dental implants, Appl. Surf. Sci. 257 (7) (2011) 2555–2559. [13] R. Olivares-Navarrete, et al., Biocompatibility of niobium coatings, Coatings 1 (1) (2011) 72–87. [14] G. Ramírez, et al., Amorphous niobium oxide thin films, J. Non-Cryst. Solids 356 (50–51) (2010) 2714–2721. [15] D.A.B. de Filho, et al., Niobia films: surface morphology, surface analysis, photoelectrochemical properties and crystallization process, J. Mater. Sci. 33 (10) (1998) 2607–2616. [16] G.-R. Gu, et al., Investigation on the structure of TiO2 films sputtered on alloy substrates, Vacuum 71 (4) (2003) 487–490. [17] M. Ristić, S. Popović, S. Musić, Sol–gel synthesis and characterization of Nb2O5 powders, Mater. Lett. 58 (21) (2004) 2658–2663. [18] S.B. Deshpande, et al., Room temperature synthesis of mesoporous aggregates of anatase TiO2 nanoparticles, Mater. Chem. Phys. 97 (2–3) (2006) 207–212. [19] E.-S. Ghaith, et al., Apatite formation on CO2 laser irradiated titanium oxide films, Mater. Lett. 60 (2) (2006) 194–197. [20] E. Eisenbarth, et al., Nanostructured niobium oxide coatings influence osteoblast adhesion, Journal of Biomedical Materials Research Part A 79A (1) (2006) 166–175. [21] W.J. Nascimento, et al., Nanostructured Nb2O5–natural hydroxyapatite formed by the mechanical alloying method: a bulk composite, Mater. Chem. Phys. 130 (1–2) (2011) 84–89. [22] H. Maeda, T. Kasuga, M. Nogami, Apatite formation on titania–vaterite powders in simulated body fluid, J. Eur. Ceram. Soc. 24 (7) (2004) 2125–2130.
[23] T. Kokubo, H. Takadama, How useful is SBF in predicting in vivo bone bioactivity? Biomaterials 27 (15) (2006) 2907–2915. [24] M. Maček, B. Orel, Electrochromism of sol–gel derived niobium oxide films, Sol. Energy Mater. Sol. Cells 54 (1–4) (1998) 121–130. [25] A. Pawlicka, M. Atik, M.A. Aegerter, Synthesis of multicolor Nb2O5 coatings for electrochromic devices, Thin Solid Films 301 (1–2) (1997) 236–241. [26] M.C. Advincula, et al., Osteoblast adhesion and matrix mineralization on sol–gelderived titanium oxide, Biomaterials 27 (10) (2006) 2201–2212. [27] S.-S. Hong, et al., Synthesis of nanosized TiO2/SiO2 particles in the microemulsion and their photocatalytic activity on the decomposition of p-nitrophenol, Catal. Today 87 (1–4) (2003) 99–105. [28] B.X. Huang, et al., Characterization of oxides on niobium by Raman and infrared spectroscopy, Electrochim. Acta 44 (15) (1999) 2571–2577. [29] H. Zhang, J.F. Banfield, Kinetics of crystallization and crystal growth of nanocrystalline anatase in nanometer-sized amorphous titania, Chem. Mater. 14 (10) (2002) 4145–4154. [30] R. Parra, et al., Reaction pathway to the synthesis of anatase via the chemical modification of titanium isopropoxide with acetic acid, Chem. Mater. 20 (1) (2007) 143–150. [31] M.R. Mohammadi, D.J. Fray, A. Mohammadi, Sol–gel nanostructured titanium dioxide: controlling the crystal structure, crystallite size, phase transformation, packing and ordering, Microporous Mesoporous Mater. 112 (1–3) (2008) 392–402. [32] K. Yoshimura, et al., Characterization of niobium oxide electrochromic thin films prepared by reactive d.c. magnetron sputtering, Thin Solid Films 281–282 (0) (1996) 235–238. [33] R.S. Sonawane, S.G. Hegde, M.K. Dongare, Preparation of titanium(IV) oxide thin film photocatalyst by sol–gel dip coating, Mater. Chem. Phys. 77 (3) (2003) 744–750. [34] D.J. Kim, et al., Influence of calcination temperature on structural and optical properties of TiO2 thin films prepared by sol–gel dip coating, Mater. Lett. 57 (2) (2002) 355–360. [35] R.M. Pittman, A.T. Bell, Raman studies of the structure of niobium oxide/titanium oxide (Nb2O5·TiO2), The Journal of Physical Chemistry 97 (47) (1993) 12178–12185. [36] J.M. Jehng, I.E. Wachs, ChemInform abstract: structural chemistry and Raman spectra of niobium oxides, ChemInform 22 (15) (1991) (no-no). [37] N. Mahdjoub, et al., SEM and Raman study of thermally treated TiO2 anatase nanopowders: Influence of calcination on photocatalytic activity, J. Photochem. Photobiol. A Chem. 211 (1) (2010) 59–64. [38] W.-X. Xu, et al., The structure of TiOx thin film studied by Raman spectroscopy and XRD, Appl. Surf. Sci. 148 (3–4) (1999) 253–262. [40] M.A. Aegerter, M. Schmitt, Y. Guo, Sol–gel niobium pentoxide coatings: applications to photovoltaic energy conversion and electrochromism, International Journal of Photoenergy 4 (1) (2002) 1–10. [41] T. Kokubo, et al., Development of bioactive materials based on surface chemistry, J. Eur. Ceram. Soc. 29 (7) (2009) 1267–1274. [42] T. Kokubo, Design of bioactive bone substitutes based on biomineralization process, Mater. Sci. Eng. C 25 (2) (2005) 97–104. [43] M. Uchida, et al., Structural dependence of apatite formation on titania gels in a simulated body fluid, Journal of Biomedical Materials Research Part A 64A (1) (2003) 164–170.
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Investigating the structure and biocompatibility of niobium and titanium oxides as coatings for orthopedic metallic implants D. Pradhan, A.W. Wren, S.T. Misture, N.P. Mellott ⁎ Inamori School of Engineering, Alfred University, Alfred, NY 14803, USA
a r t i c l e
i n f o
Article history: Received 4 May 2015 Received in revised form 3 September 2015 Accepted 13 September 2015 Available online 14 September 2015 Keywords: Coatings Ceramics Biocompatibility
a b s t r a c t Applying sol gel based coatings to orthopedic metallic implant materials can significantly improve their properties and lifespan in vivo. For this work, niobium (Nb2O5) and titanium (TiO2) oxides were prepared via solution processing in order to determine the effect of atomic arrangement (amorphous/crystalline) on bioactivity. Thermal evaluation on the synthesized materials identified an endotherm for Nb2O5 at 75 °C with 40% weight loss below 400 °C, and minimal weight loss between 400 and 850 °C. Regarding TiO2 an endotherm was present at 92 °C with 25% weight loss below 400 °C, and 4% between 400 and 850 °C. Phase evolution was determined using High Temperature X-ray Diffraction (HT-XRD) where amorphous-Nb2O5 (450 °C), hexagonal-Nb2O5 (525 °C), orthorhombic-Nb2O5 (650 °C), amorphous-TiO2 (275 °C) and tetragonal TiO2 (500 °C) structures were produced. Simulated body fluid (SBF) testing was conducted over 1, 7 and 30 days and resulted in positive chemical and morphological changes for crystalline Nb2O5 (525 °C) and TiO2 (500 °C) after 30 days of incubation. Rod-like CaP deposits were observed on the surfaces using Scanning Electron Microscopy (FE-SEM) and Grazing Incidence-X-ray Diffraction (GI-XRD) shows that the deposits were X-ray amorphous. Cell viability was higher with the TiO2 (122%) samples when compared to the growing cell population while Nb2O5 samples exhibited a range of viability (64–105%), partially dependent on materials atomic structure. © 2015 Elsevier B.V. All rights reserved.
1. Introduction In the recent past, metallic materials such as stainless steel have been employed for load-bearing orthopedic applications for the repair or replacement of damaged or diseased joints such as total hip replacement (THR) and total knee replacement (TKR). However, a persistent concern relates to the interfacial response between the host tissue and the metallic component, which typically does not bio-integrate and can result in loosening and subsequent implant removal. Recently however, techniques have been developed to improve integration between the host tissue and the metallic component such as surface roughening and etching, thermal oxidation treatments and also through applying coatings to the surface of the metallic implant. Metal/alloys such as Ti6Al4V are known to be bioactive, [1,2] however, other widely used metals such as stainless steel are considered to be bioinert (no tissue integration), [3,4] or induces an allergic reaction (nickel, cobalt–chromium) upon implantation [5,6]. The biofunctionality of the surface of metallic materials can be altered through thermal oxidation treatments [7]. These thermal treatments can increase the performance of implants by inducing the formation of a passive oxide layer, improving the corrosion and wear resistance of coatings. In addition, the biocompatibility of thermally ⁎ Corresponding author at: 2 Pine Street, Alfred University, Alfred, NY 14802, USA. E-mail address: [email protected] (N.P. Mellott).
http://dx.doi.org/10.1016/j.msec.2015.09.059 0928-4931/© 2015 Elsevier B.V. All rights reserved.
oxidized metals, including titanium alloys, can also induce biocompatibility [8,9]. Coating metals with transition metal oxides can also increase the materials biocompatibility which can be applied to both orthopedic surgery and oral and maxillofacial restoration [10–12]. Traditionally, implant functionality and success was largely dependent upon the surface chemistry, mechanical stability, and biological compatibility of the implant material [12–13]. Oxide surface coatings can be employed to promote bioactivity or improve biocompatibility of the host implant material and given that the oxide coating is in direct contact with the host tissue, it has a profound effect on the chemical and biological processes which occur at that interface [14–15]. In particular, the oxide coatings atomic arrangement (crystalline, amorphous), dissolution chemistry and stability can affect not only the resistance to fatigue and high tensile stresses, but also biological effects such as cell adhesion, osteoblast proliferation, the inflammatory response and biomineralization. It has been proposed that if the properties of an oxide coating could be characterized, controlled and predicted, then enhanced properties can be engineered into an implant material [14]. This would facilitate the ability to tailor an implants bioactivity in addition to controlling the physico-chemical/biological processes at the coating-tissue interface. Depositing oxide coatings on metallic implants can be achieved using a variety of techniques including physical vapor deposition (PVD), chemical vapor deposition (CVD), and various sol–gel based
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
deposition techniques [11–13,16,17]. In particular, sol–gel dip coating followed by calcination has advantages that include the ability to control the coatings purity and chemistry. In addition, this processing method affords scalability and the ability to coat irregular shaped implant surfaces. Titanium (TiO2) and niobium oxide (Nb2O5) systems have shown particular promise as candidate materials for bioactive coatings on metallic implants [13,18,19]. In comparison to TiO2, Nb2O5 has been less extensively studied [13,20], however, the properties exhibited by this oxide, such as high corrosion resistance and thermodynamic stability, supports its potential as a coating for medical implants [14,21]. Eisenbarth et al. conducted studies which determined that osteoblasts preferentially adhered to the roughness of nanosized Nb2O5 coatings on polished CP Titanium [20]. A study by Ramirez et al. on coating stainless steel dental implants with crystalline (c-NbN) and amorphous (a-Nb2O5) also supports this work where they determined enhanced viability, proliferation and attachments of cementoblasts. The coatings were found to improve the surface hardness, corrosion resistance and the overall biological response of the stainless steel implants [12]. It has been reported previously that specific properties can be induced by crystal structure and atomic arrangement and that surface crystallinity plays an important role in structure–bioactivity relationships of oxide systems [14]. Ramirez et al. demonstrated that crystalline Nb2O5 shows greater CaP deposition when incubated in simulated body fluid (SBF) than its amorphous counterpart [14]. Studies by Maeda et al. demonstrated that apatite formation is crystal phase dependent, where more prominent apatite deposition was evident on anatase-TiO2 rather than rutile-TiO2 [22] and Eisenbarth et al. also demonstrated that osteoblast response to Nb2O5 is a function of crystal structure [20]. With respect to this study, TiO2 and Nb2O5 oxides are prepared in powder form with controlled atomic arrangement. Powders were used in lieu of coatings for ease of characterization. However, results here are directly analogous to coatings prepared in the same manner, as coatings could be applied through spray-based deposition of dispersed powders or through spin or dip coatings of sols prior to drying and heating. It is worth mentioning here that for the utilization of these powders for spray deposition, care must be taken in both the understanding and minimization of the effect of the coating deposition process on both implant and powder structure and morphology. Powders were characterized and evaluated using High-Temperature X-ray Diffraction (HT-XRD) and Differential Thermal Analysis (DTA). The biocompatibility of both amorphous and crystalline analogs of TiO2 and Nb2O5 is also investigated with respect to time in a synthetic biological fluid that resembles the ionic composition of human blood plasma (simulated body fluid — SBF). SBF is considered to be a strong indicator of a materials' bioactivity as precipitation of calcium and phosphate ion from the SBF on the materials' surface is considered to be a precursor to bone bonding in vivo. Any changes in surface chemistry and morphology, attributed to materials atomic arrangement, were identified using Field Emission Scanning Electron Microscopy (FESEM) and Grazing Incidence X-ray Diffraction (GI-XRD). Finally, the cytocompatibility of the materials was investigated using a model cell line, L929 Mouse Fibroblasts, to determine if TiO2 and Nb2O5 atomic arrangement influences cytotoxicity. This effect was also conducted with respect to maturation in an aqueous environment. 2. Materials and methods 2.1. Material synthesis Titanium and niobium based sols were prepared individually using titanium (IV) isopropoxide (Ti (OC3H7)4, 2.96 ml) 99.99% and niobium (V) ethoxide (Nb (OC2H5)5, 2.50 ml) 99.95% (Sigma-Aldrich) respectively, denatured ethyl alcohol (C2H5OH, Fisher Scientific) and glacial acetic acid (CH3COOH, Fisher Scientific). In each case, ethanol (58.39 ml) was taken in a beaker and glacial acetic acid was added
919
(5.72 ml) while stirring. After 15 min the titanium and niobium precursors (2.96 ml and 2.50 ml respectively), were added to the mixture dropwise. The solutions were covered with perforated parafilm and magnetically stirred until solutions were clear and transparent (~2 h). The stable solutions were then heated at 60 °C with simultaneous stirring until the liquid evaporated and powders were obtained (36– 48 h). The powders were dried at 150 °C for 24 h in a conventional oven and then crushed using an agate mortar and pestle. Dried powders were then segregated into two separate batches; (1) the first batch was stored, as-dried, in a vacuum desiccator for DTA/TGA and in-situ HTXRD analysis and (2) the second batch was calcined from 200 to 650 °C with a heating rate of 20 °C/min and a 90 min dwell period every 25 °C in a Thermo Scientific Thermolyne furnace (model no. F48025-6080). Powders were collected at the end of each dwell period and stored in separate containers for analysis. 2.2. Material characterization 2.2.1. Differential Thermal Analysis/Thermogravimetric Analysis (DTA– TGA) This analysis was performed using an SDT 2960 simultaneous DSC– TGA. The measurements were obtained from 30 °C to 850 °C under flowing air at a heating rate of 20 °C/min. 2.2.2. High Temperature X-ray Diffraction (HT-XRD) Diffraction patterns were obtained using a Bruker D8 advance diffractometer (Cu-Kα radiation) equipped with a Vantec detector with Gobel mirror, 1.0 mm divergence slit and Anton Paar high HTK 1200 temperature furnace attachment. Measurements were made over 10 to 140° 2θ with a step size of 0.03° 2θ and a scan rate of 10° 2θ/min. The high temperature experiments were carried out using a heating rate of 20 K/min, with dwell period of 90 min after every 25 K increment. The measurements were taken at the end of each dwell period and Jade 9 software (Materials Data Inc., USA) was used for analysis. 2.2.3. Raman spectroscopy The spectra were obtained from a Witec Confocal Raman Microscope CRM200 equipped with Si detectors, green laser (excitation wavelength of 532 nm and power of 70 mW), and a dispersion grating of 600 l/mm. 2.3. Evaluation of material bioactivity and cytocompatibility 2.3.1. Simulated body fluid (SBF) Simulated body fluid testing was conducted in accordance with the procedure outlined by Kokubo [23]. The reagents and their respective amount were dissolved in 700 ml of deionized water using a magnetic stirrer and the solution was maintained at 36.5 ± 0.5 °C. The pH of the solution was adjusted to 7.4 by adding Tris (hydroxymethyl) aminomethane and 1 M-HCl [23]. The oxide powders were first crushed and sieved to obtain uniform sized particles below 30 μm. Selected powders were then pressed into pellets of dimensions (8.06 mm × 1.3 mm) by applying pressure of 1000 psi with holding time of 30 s. The disk specimens were soaked in SBF for 1, 7, and 30 days. The volume of SBF for immersion of pellets was calculated on the basis of the surface area (diameter and height) of the disks. After soaking in SBF for the required time period, the disks were removed and gently rinsed with deionized water and stored in incubator at 37 °C for 24 h. 2.3.2. Scanning Electron Microscopy (SEM) The analysis was performed on disks reacted in SBF solution for 0, 1, 7, and 30 days using an FEI Quanta 200F equipped with a field emission gun. Analyses were performed at an accelerating voltage of 12.5 kV and beam current of 26 nA. In addition, semi-quantitative analysis was performed using energy-dispersive X-ray spectroscopy (EDX).
920
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
Fig. 1. DTA/TGA plots for primary a.) Nb2O5 and b.) TiO2 powders.
2.3.3. Grazing Incidence X-ray Diffraction (GIXRD) GIXRD analyses were performed on selected disks using a Siemen's Kristalloflex diffractometer using Copper Kα X-rays and a BraggBrentano Goniometer. The pellets were mounted on the sample holder using X-ray amorphous putty. Data was collected from 5° to 70° 2θ in 0.02° steps, 10 s each. Jade software (v 6.0.3, Materials Data Inc.) was used for analysis. 2.3.4. Specific surface area (SSA) The selected powders were measured via N2 gas adsorption according to the BET method with a micromeritics Tristar 3000. Prior to analysis powders were degassed in a micromeritics flow prep 060 sample degas system at 150 °C for 24 h. The instrument was standardized prior to and after analysis using silica-alumina rod standard (201 ± 5 m2/g). 2.3.5. Cytotoxicity analysis The experiments were performed using mouse fibroblast cell lines L929 (American Type Culture collection CCL 1, NCTC clone 929; ISO10993 part 5). Cells were initially cultured in Hyclone medium 199/EBSS (Thermo Scientific) supplemented with 10% fetal bovine serum (FBS).
Cells were maintained on regular feeding in a CO2 incubator at 37 °C/ 5% CO2 atmosphere/100% relative humidity. Media was changed every 2 days. At confluence, cells were detached using a trypsin treatment. The cells were centrifuged and re-suspended in a media containing 10% FBS. Finally cells were seeded onto a 24 well plate at density of 10,000 cells per well and incubated for 24 h prior to testing. The sample extracts were collected by dissolving the weight of powder equivalent to 1 m2 BET surface area (not shown) in 10 ml deionized sterile water. MTT assay was used to estimate the cell viability. 100 μl of sterile sample extracts collected after 1, 7 and 30 days (n = 3) was added into wells containing L929 cells in culture medium (1 ml). The prepared well plates were incubated for 24 h at 37 °C/5% CO2. The MTT assay was then added in an amount equal to 10% of the culture medium volume/ well. The cultures were then re-incubated for a further 4 h (37 °C/5% CO2). The cultures were removed from the incubator and the resultant formazan crystals were dissolved by adding an amount of MTT Solubilization Solution (10% Triton X-100 in acidic isopropanol (0.1 N HCI) equal to the original culture medium volume. Once the crystals were fully dissolved, the absorbance was measured at a wavelength of 570 nm. Aliquots (100 μl) of tissue culture water were used as controls, and cells were assumed to have metabolic activities of 100%.
Fig. 2. In-situ HT-XRD patterns for a.) Nb2O5 b.) TiO2 powders (indexing of individual phases can be observed in Fig. 3).
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
921
Table 1 X-ray diffraction characteristics of crystalline structures. Temp (°C)
PDF#
Chem. ID
Symmetry
a (A°)
b (A°)
c (A°)
Space grp.
450 °C 525 °C 650 °C 275 °C 500 °C
– 00-028-0317 00-030-0873 – 04–006–9240
– Nb2O5 Nb2O5 – TiO2
Amorphous Hexagonal Orthorhombic Amorphous Anatase
– 3.607 6.175 – 3.784
– 3.607 29.175 – 3.784
– 3.925 3.930 – 9.495
– P Pbam (55) – I41/amd (I41)
2.4. Statistical analysis One-way analysis of variance (ANOVA) was employed to compare the cell viability of the experimental materials in relation to 1.) maturation (1, 7, 30 days) and 2.) sample composition. Comparison of relevant means was performed using the post hoc Bonferroni test. Differences between groups was deemed significant when p ≤ 0.05.
3. Results and discussion 3.1. Material characterization and phase identification Material characterization was conducted to determine each composition's thermal profiles and properties, and to accurately identify suitable heat treatment profiles for the powdered Nb2O5 and TiO2 samples. Thermal analysis (DTA/TGA) plots are presented in Fig. 1 for the sol gel derived niobium oxide and titanium oxide powdered samples. Regarding niobium-based powders (Fig. 1a) an endotherm was observed at 75 °C followed by exotherms at approximately 279, 344 and 579 °C. The TGA curve showed approximately 40% weight loss between 27 and 400 °C, with minimal weight loss observed between ~ 400 and 850 °C. For titanium oxide powders (Fig. 1b) an endotherm was observed at 92 °C followed by exotherms at approximately 340, 412 and 522 °C. The TGA curve showed an approximately 25% weight loss between 27 and 400 °C, with an additional 4% weight loss observed between 400 and 850 °C. The first endotherm observed for both samples is attributed to elimination of adsorbed water [24]. This is further evidenced by the large weight loss observed for both sample systems, between room temperature and ~ 250 °C. Exothermic peaks observed at both 279 °C and 344 °C for the niobium based sample and at 340 °C for the titanium based sample are associated with the removal of organics [25–26]. This is also consistent with the additional weight loss observed up to ~400 °C [27]. It has been reported that the organic species are typically compounds of alcohols including C3H7, C4H8, and
adsorbed acetic acid used in the preparation of sols [18,28,29]. Upon further heating of the niobium oxide powders a sharp exothermic peak is observed at 579 °C, associated with the crystallization to hexagonalNb2O5 [17]. Upon further heating of the titanium oxide based samples, exothermic peaks are observed at 412 °C and 522 °C, presumably associated with the crystallization to anatase-TiO2 followed by either (1) a phase transformation to rutile-TiO2, or (2) removal of residual organics [30,31]. It is well understood that the exact crystallization temperatures of these oxides are a strong function of the chemistry of the sol, particle shape, as well as the heating rate of the sample, however as noted the crystallization temperatures here are in general agreement with those previously reported of similar systems [17,30,31]. High-Temperature X-ray Diffraction (HT-XRD) was employed to identify and synthesize particular phases/atomic structures and the results are presented in Fig. 2. The HT-XRD patterns in Fig. 2a present the phase evolution of as-prepared niobium oxide powders heated insitu from 30 to 650 °C. The niobium oxide powders remained X-ray amorphous up to 475 °C, with initial crystallization occurring between 475 and 500 °C. A mixture of amorphous and crystalline phases was observed after heat treatment at 500 °C which was evident from the sharp peaks as well as an uneven hump within the background of the pattern. For temperatures above 500 °C, patterns exhibit sharp, well defined peaks suggesting highly crystalline Nb2O5. The three most intense peaks at 22.7°, 28.6° and 36.7° 2θ remained unchanged between 500 and 650 °C. However, at ~600 °C additional peaks were observed at 2θ 17.4° and also at 32.6°, 42.7°, and 45.3° 2θ with intensity of peaks increasing with calcination temperature. Phase ID HT-XRD patterns, show that the crystal phase observed from 500 to 575 °C is hexagonalNb2O5 (TT-Nb2O5; PDF# 00-028-0317). This transformation temperature is in agreement with what has been reported in the literature [15, 32]. Upon heating above 575 °C, a phase transformation from a hexagonal to an orthorhombic-Nb2O5 (T-Nb2O5; PDF#-000-030-0873) was observed. The crystallite size of TT-Nb2O5 increased with temperature from 23 ± 4 nm (500 °C) to 37 ± 6 nm (550 °C) while the crystallite size of T-Nb2O5 increased from 43 ± 3 nm (575 °C) to 74 ± 3 nm
Fig. 3. Individual phase transformation of In situ HT-XRD patterns for a.) Nb2O5 and b.) TiO2 powders.
922
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
Fig. 4. Raman spectra of a.) Nb2O5 powders and b.) TiO2 powders showing different phase developments with temperature.
(650 °C), as determined by Rietveld refinement. The HT-XRD patterns in Fig. 2b present the phase evolution of as-prepared titanium oxide powders heated in-situ from 29 to 650 °C. The powders remained X-ray amorphous up to 325 °C. Upon crystallization, the three most intense peaks were observed at 25.6°, 48.3° and 48.3° 2θ. All peak positions for the crystalline phases remained at approximately the same position with calcination from 350 to 650 °C. However, with increasing temperature, the observed peaks become sharper and more well-defined. Upon phase ID analysis, it was determined the crystalline phase observed at 350 °C was tetragonal-TiO2 (anatase; PDF# 04-006-9240), with this phase remaining stable until 650 °C. The initial crystallization temperature for anatase, between 325 and 350 °C observed here is in agreement with the reports which suggest that this transformation typically occurs between 300 and 400 °C [29,33,34]. The crystallite size of tetragonalTiO2 increased with temperature from 22 ± 1 nm (350 °C) to 45 ± 1 nm (650 °C) as determined by Rietveld refinement. Three primary phases were observed via HT-XRD when heated to 650 °C for niobium oxide: amorphous-Nb2O5, hexagonal-Nb2O5, and orthorhombic-Nb2O5, while two phases were observed for titanium oxide: amorphous-TiO2 and tetragonal-TiO2; as summarized in Table 1. Individual HT-XRD patterns associated with these samples are shown in Fig. 3a &b respectively. These five structures were then chosen for further structural and biocompatibility analysis and will be referred to from this point forward as; Nb2O5-450 °C, Nb2O5-525 °C, Nb2O5-650 °C, TiO2-275 °C, and TiO2500 °C respectively. Raman spectra for the selected samples are shown in Fig. 4. Nb2O5450 °C showed spectra with broad bands centered at ~ 220 cm−1 and ~670 cm−1 with a shoulder observed at the low wavelength side of the 200 cm−1 band and high wavelength sides of the ~670 cm−1 band, all associated with Nb–O vibrational modes [25,32]. The broad bands are consistent with amorphous, disordered Nb2O5 [35]. While one could argue that these spectra may be indicative of a poorly crystalline Nb2O5, given the complete lack of crystallinity observed in the X-ray Diffraction patterns associated with this sample (Fig. 3a), the authors believe that it is appropriate to conclude that this sample is amorphous. Nb2O5-525 °C showed sharper more defined bands centered at ~ 147, 233, 278, 290 and 691 cm−1 along with weak shoulders on the high frequency side of the ~290 and ~691 cm−1 band (Fig. 4a.) consistent with crystallization. Bands are assigned to the bending of Nb–O–Nb, stretching of niobia polyhedra and stretching of Nb_O modes of crystalline (both hexagonal and orthorhombic) Nb2O5 [28,36]. In particular, the crystallization of powders to pseudohexagonal phase (TT-phase) at 525 °C is evidenced by the narrowing of broad peaks and simultaneous shifting of bands from 670 to 691 cm−1 due to increased bond order of niobia
polyhedra [33]. Further, the absence of a shoulder at the high frequency side of the ~691 cm−1 band also confirmed the presence of the crystalline phase [25]. Bands associated with Nb2O5-650 °C are equal in both position and shape to those observed for Nb2O5-525 °C with the possible addition of a band at ~450 cm−1. It is well understood that the Raman shift position is sensitive to changes in atomic structure/arrangement, therefore the lack of shift observed here between Nb2O5-525 °C and Nb2O5-650 °C is consistent with the previously reported observation that the two structures are very similar [33,37]. TiO2-275 °C showed broad, weak bands at ~ 183, 206, 421, 526, 609 and 957 cm−1. These broad, undefined bands and their positions are consistent with a disordered, amorphous TiO2 [35]. Furthermore, the weak shoulder at 957 cm−1 observed for amorphous TiO2 could be associated with the presence of adsorbed surface species which tends to disappear at higher calcination temperature. In comparison, TiO2-500 °C exhibited sharp, well defined bands at 147, 197, 394, 518 and 639 cm−1 (Fig. 4b). These band positions are in good agreement with those shown to be Raman active modes associated with tetragonal-TiO2; 144, 197, 397, 518, and 640 cm−1 [37,38]. The first three bands corresponded to O–Ti–O bending mode and the latter two were assigned to Ti–O stretching modes [37,38]. Characterization and identification of the structural evolution of the niobium and titanium oxide powders, with increasing temperature, was studied by complimentary analytical techniques including DTA/ TGA, HT-XRD, and Raman spectroscopy. The results are consistent with one another, as well as being consistent with previous reports of similarly prepared systems [12,35,36]. It was shown that the niobium oxide powders were initially amorphous and remained so until 500 °C. Between 500 and 600 °C a hexagonal crystal phase was observed (TT-Nb2O5) followed by an orthorhombic phase (T-Nb2O5), observed between 600 and 650 °C. It was shown that titanium oxide powders were initially amorphous and remained so until 350 °C. Between 350 and 650 °C a tetragonal crystal phase was observed (anatase). It is worth noting that the phase transformation between hexagonalNb2O5 and orthorhombic-Nb2O5 is not evidenced in the DTA/TGA plots via a sharp exotherm, however this is expected, as hexagonalNb2O5 is thought to be a modification of orthorhombic-Nb2O5 with low crystallinity [36,40]. Furthermore, HT-XRD results suggest that the DTA exotherm observed at 522 °C (Fig. 1b) is due to removal of residual organics and not an anatase to rutile transformation. 3.2. Determination of material biocompatibility Preliminary evaluation of the materials' bioactivity was conducted using simulated body fluid (SBF) testing on the samples Nb2O5-450 °C,
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
923
Fig. 5. SEM images and corresponding EDX spectra for Nb2O5 obtained after immersion in SBF for 30 days for a.) Nb2O5-450 °C, b.) Nb2O5-525 °C and c.) Nb2O5-650 °C.
Nb2O5-525 °C, and Nb2O5-650 °C (Fig. 5) and TiO2-275 °C and TiO2-500 °C (Fig. 6). SBF testing has become a widely recognized method of investigating a materials' bioactivity as it is a solution that contains the ionic composition analogous to human blood plasma, and surface precipitation reactions in SBF are regarded as a precursor to bone bonding in vivo [23,41,42]. This method aims to determine changes in surface chemistry and morphology when samples are incubated in SBF as a function of time. SBF testing was conducted on pressed disks of the selected samples and was incubated for 1, 7, and 30 days. After each time period, each composition/structure was analyzed for
morphological (FE-SEM) or chemical (EDX) evidence of calcium and/ or phosphorous (CaP) containing precipitates. No evidence of precipitation is observed for any disks reacted for 1 or 7 days. After 30 days of immersion in SBF, Nb2O5-450 °C, Nb2O5-650 °C, and TiO2-275 °C remain free of CaP precipitation. However, both morphological evidence and chemical evidence of precipitation are observed for Nb2O5-525 °C and TiO2500 °C after immersion for 30 days in SBF. Nanometer to micron size particles are clearly observed via SEM with both Ca and P detected on the surface via EDX (Fig. 5 & Fig. 6). When observed at higher resolution (50 kx) it is clear that the morphology of the precipitates is different, as
924
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
Fig. 6. SEM images and corresponding EDX spectra for TiO2 obtained after immersion in SBF for 30 days a.) TiO2-275 °C and b.) TiO2-500 °C.
Nb2O5-525 °C is larger and is plate like crystals where TiO2-500 °C crystals are smaller rod-shaped precipitates (Fig. 7). GI-XRD analysis is presented in Fig. 8 and was performed on materials with precipitates to determine if any new crystalline phases were evident after 30 days of incubation in SBF. However, diffraction patterns show only evidence of crystalline Nb2O5 or TiO2 and not any additional phases, which suggests that this surface precipitant is amorphous CaP, the precursor to crystalline apatite. Cytotoxicity testing was conducted using L929 Mouse Fibroblasts to determine the cell viability after exposure to each of the samples, and is presented in Fig. 9a & b. The MTT assay determined changes in cell
growth and the metabolic activity when exposed to the experimental materials' condition medium when compared to a healthy growing cell population. Cell viability testing for TiO2 is presented in Fig. 9a. In comparison to the Control cell population, which was assumed to have a metabolic rate of 100%, cell viability for TiO2-275 °C showed an increase (106–122%) over 1–30 days, however, this increase was not found to be significant when statistically compared to the Control cell population (p = 0.292). For TiO2-500 °C, cell viability increased initially to 110% after 7 days and the reduced to 83% after 30 days, which was also not significantly reduced when compared to the Control cell population (p = 1.000). Cell viability testing for Nb2O5 is presented in Fig. 9a.
Fig. 7. SEM images of CaP layer deposition on the surface of a.) hexagonal-Nb2O5 (525 °C), and b.) TiO2-tetragonal (anatase, 500 °C) after immersion in SBF for 30 days.
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
Fig. 8. GI-XRD of a.) hexagonal-Nb2O5 (525 °C), and b.) TiO2-tetragonal (anatase, 500 °C) after soaking in SBF for 30 days.
Regarding Nb2O5-450 °C exhibiting hexagonal phase, there was no significant change in viability observed after 30 days (p = 1.000). For Nb2O5-525 °C a significant decrease in cell viability was observed at 1 (p = 0.003) and 7 days (p = 0.010) at 64% and 68% respectively. However, at 30 days the metabolic activity recovered to 71% which was not considered significant when compared to the Control cells (p = 0.063). Nb2O5-650 °C exhibited a decrease in cell viability at 1 day (70%, p = 0.024) but increased to 100% after 30 days which is not significantly different than the Control cells (p = 1.000). Although the amorphous materials presented slightly higher cell viability than the crystalline analogs, (with the exception of Nb2O5-650 °C), the crystallographic phases did not overall present significant differences in cell viability compared to their crystalline counterparts. Also, there was no significant difference in cell viability when considering the effect of maturation between each material at 1, 7 and 30 days (p = 1.000). It was shown that within both Nb2O5 and TiO2 oxide systems that the atomic arrangement influences the materials' bioactivity. Within the TiO2 system the tetragonal phase exhibited bioactivity while the amorphous did not. This is in agreement with previous studies by Ghaith et al., where crystalline TiO2 (anatase) structures were found to promote CaP deposition over the amorphous counterpart [19]. In fact, it has been shown that it is not just the presence of crystallinity that affects bioactivity, but the actual crystal structure; anatase TiO2 has been shown to be more bioactive than rutile TiO2 [23,43]. We observe the
925
same phenomena here within the Nb2 O 5 system, as hexagonalNb 2 O 5 exhibited bioactivity and both amorphous-Nb 2 O 5 and orthorhombic-Nb 2O 5 did not. However, no correlation between oxide structure and cell viability was observed, with all oxides performing equally to the control cell population, which indicates that no toxicity was evident. Future work will include depositing oxide coatings of equal composition and structure on metallic substrates (stainless steel) to test their chemical durability and biofunctionality in various biologically relevant solutions. Coatings will be deposited through the spray coating of dispersed oxide powders, as prepared here, or via spin coating of the sols described here onto the stainless steel substrate followed by calcination. In our case it would be advantageous to deposit dispersed powders, assuming that a deposition method which will not change the structure or composition of the powders can be developed. Otherwise, we will develop a processing method, in which we deposit the sol on stainless steel followed by spin coating and calcination. This will require the development of a calcination maximum temperature and ramp rate for the coatings which will result in the same structure and composition as the powders reported here.
4. Conclusions In this work, Nb2O5 and TiO2 were prepared via wet chemistry. Structural evolution with calcination temperature was determined and amorphous-Nb2O5, hexagonal Nb2O5, orthorhombic-Nb2O5, amorphous TiO2, and tetragonal TiO2 were then tested for bioactivity and cytocompatibility. After immersion in SBF-solution for 30 days only TiO2-tetragonal and Nb2O5-hexagonal exhibited precipitation of CaPbased materials, a positive indicator of bioactivity. Cytocompatibility results show that none of the oxides tested had a negative or toxic effect on cell viability, and that the amorphous TiO2 and Nb2O5 were more cytocompatible than the crystalline analogs. Results from this work not only shed light on the structure–bioactivity properties of these oxides but also can serve as template to aid in the preparation of bioactive Nb2O5 and TiO2 materials.
Acknowledgment The authors would like to acknowledge the Pennsylvania State University Materials Characterization Laboratory for Raman analyses. This work was partially supported by the 3M Non-Tenured Faculty Award (NPM).
Fig. 9. MTT assay of all Nb2O5 and TiO2 structures extracts in L929 Fibroblasts after 1, 7 and 30 days of incubation.
926
D. Pradhan et al. / Materials Science and Engineering C 58 (2016) 918–926
References [1] H. Takadama, et al., XPS study of the process of apatite formation on bioactive Ti– 6Al–4V alloy in simulated body fluid, Sci. Technol. Adv. Mater. 2 (2001) 389–396. [2] Y.-T. Sul, et al., The bone response of oxidized bioactive and non-bioactive titanium implants, Biomaterials 26 (2005) 6720–6730. [3] K. Hayashi, et al., Re-evaluation of the biocompatibility of bioinert ceramics in vivo, Biomaterials 13 (4) (1992) 195–200. [4] S. Kumar, D. Simpson, R.S.C. Smart, Plasma processing for inducing bioactivity in stainless steel orthopaedic screws, Surf. Coat. Technol. 202 (4–7) (2007) 1242–1246. [5] Hussain, O.T., S. Sah, and A.J. Sidebottom, Prospective comparison study of one-year outcomes for all titanium total temporomandibular joint replacements in patients allergic to metal and cobalt–chromium replacement joints in patients not allergic to metal. Br. J. Oral Maxillofac. Surg. 52(1): p. 34–37. [6] H.F. Hildebrand, C. Veron, P. Martin, Nickel, chromium, cobalt dental alloys and allergic reactions: an overview, Biomaterials 10 (8) (1989) 545–548. [7] H. Güleryüz, H. Çimenoğlu, Effect of thermal oxidation on corrosion and corrosion– wear behaviour of a Ti–6Al–4V alloy, Biomaterials 25 (16) (2004) 3325–3333. [8] M.C. Garcia-Alonso, et al., In vitro corrosion behaviour and osteoblast response of thermally oxidised Ti6Al4V alloy, Biomaterials 24 (1) (2003) 19–26. [9] C.-M. Lin, S.-K. Yen, Biomimetic growth of apatite on electrolytic TiO2 coatings in simulated body fluid, Mater. Sci. Eng. C 26 (1) (2006) 54–64. [10] S. Nagarajan, V. Raman, N. Rajendran, Synthesis and electrochemical characterization of porous niobium oxide coated 316L SS for orthopedic applications, Mater. Chem. Phys. 119 (3) (2010) 363–366. [11] A. Ochsenbein, et al., Osteoblast responses to different oxide coatings produced by the sol–gel process on titanium substrates, Acta Biomater. 4 (5) (2008) 1506–1517. [12] G. Ramírez, et al., Niobium based coatings for dental implants, Appl. Surf. Sci. 257 (7) (2011) 2555–2559. [13] R. Olivares-Navarrete, et al., Biocompatibility of niobium coatings, Coatings 1 (1) (2011) 72–87. [14] G. Ramírez, et al., Amorphous niobium oxide thin films, J. Non-Cryst. Solids 356 (50–51) (2010) 2714–2721. [15] D.A.B. de Filho, et al., Niobia films: surface morphology, surface analysis, photoelectrochemical properties and crystallization process, J. Mater. Sci. 33 (10) (1998) 2607–2616. [16] G.-R. Gu, et al., Investigation on the structure of TiO2 films sputtered on alloy substrates, Vacuum 71 (4) (2003) 487–490. [17] M. Ristić, S. Popović, S. Musić, Sol–gel synthesis and characterization of Nb2O5 powders, Mater. Lett. 58 (21) (2004) 2658–2663. [18] S.B. Deshpande, et al., Room temperature synthesis of mesoporous aggregates of anatase TiO2 nanoparticles, Mater. Chem. Phys. 97 (2–3) (2006) 207–212. [19] E.-S. Ghaith, et al., Apatite formation on CO2 laser irradiated titanium oxide films, Mater. Lett. 60 (2) (2006) 194–197. [20] E. Eisenbarth, et al., Nanostructured niobium oxide coatings influence osteoblast adhesion, Journal of Biomedical Materials Research Part A 79A (1) (2006) 166–175. [21] W.J. Nascimento, et al., Nanostructured Nb2O5–natural hydroxyapatite formed by the mechanical alloying method: a bulk composite, Mater. Chem. Phys. 130 (1–2) (2011) 84–89. [22] H. Maeda, T. Kasuga, M. Nogami, Apatite formation on titania–vaterite powders in simulated body fluid, J. Eur. Ceram. Soc. 24 (7) (2004) 2125–2130.
[23] T. Kokubo, H. Takadama, How useful is SBF in predicting in vivo bone bioactivity? Biomaterials 27 (15) (2006) 2907–2915. [24] M. Maček, B. Orel, Electrochromism of sol–gel derived niobium oxide films, Sol. Energy Mater. Sol. Cells 54 (1–4) (1998) 121–130. [25] A. Pawlicka, M. Atik, M.A. Aegerter, Synthesis of multicolor Nb2O5 coatings for electrochromic devices, Thin Solid Films 301 (1–2) (1997) 236–241. [26] M.C. Advincula, et al., Osteoblast adhesion and matrix mineralization on sol–gelderived titanium oxide, Biomaterials 27 (10) (2006) 2201–2212. [27] S.-S. Hong, et al., Synthesis of nanosized TiO2/SiO2 particles in the microemulsion and their photocatalytic activity on the decomposition of p-nitrophenol, Catal. Today 87 (1–4) (2003) 99–105. [28] B.X. Huang, et al., Characterization of oxides on niobium by Raman and infrared spectroscopy, Electrochim. Acta 44 (15) (1999) 2571–2577. [29] H. Zhang, J.F. Banfield, Kinetics of crystallization and crystal growth of nanocrystalline anatase in nanometer-sized amorphous titania, Chem. Mater. 14 (10) (2002) 4145–4154. [30] R. Parra, et al., Reaction pathway to the synthesis of anatase via the chemical modification of titanium isopropoxide with acetic acid, Chem. Mater. 20 (1) (2007) 143–150. [31] M.R. Mohammadi, D.J. Fray, A. Mohammadi, Sol–gel nanostructured titanium dioxide: controlling the crystal structure, crystallite size, phase transformation, packing and ordering, Microporous Mesoporous Mater. 112 (1–3) (2008) 392–402. [32] K. Yoshimura, et al., Characterization of niobium oxide electrochromic thin films prepared by reactive d.c. magnetron sputtering, Thin Solid Films 281–282 (0) (1996) 235–238. [33] R.S. Sonawane, S.G. Hegde, M.K. Dongare, Preparation of titanium(IV) oxide thin film photocatalyst by sol–gel dip coating, Mater. Chem. Phys. 77 (3) (2003) 744–750. [34] D.J. Kim, et al., Influence of calcination temperature on structural and optical properties of TiO2 thin films prepared by sol–gel dip coating, Mater. Lett. 57 (2) (2002) 355–360. [35] R.M. Pittman, A.T. Bell, Raman studies of the structure of niobium oxide/titanium oxide (Nb2O5·TiO2), The Journal of Physical Chemistry 97 (47) (1993) 12178–12185. [36] J.M. Jehng, I.E. Wachs, ChemInform abstract: structural chemistry and Raman spectra of niobium oxides, ChemInform 22 (15) (1991) (no-no). [37] N. Mahdjoub, et al., SEM and Raman study of thermally treated TiO2 anatase nanopowders: Influence of calcination on photocatalytic activity, J. Photochem. Photobiol. A Chem. 211 (1) (2010) 59–64. [38] W.-X. Xu, et al., The structure of TiOx thin film studied by Raman spectroscopy and XRD, Appl. Surf. Sci. 148 (3–4) (1999) 253–262. [40] M.A. Aegerter, M. Schmitt, Y. Guo, Sol–gel niobium pentoxide coatings: applications to photovoltaic energy conversion and electrochromism, International Journal of Photoenergy 4 (1) (2002) 1–10. [41] T. Kokubo, et al., Development of bioactive materials based on surface chemistry, J. Eur. Ceram. Soc. 29 (7) (2009) 1267–1274. [42] T. Kokubo, Design of bioactive bone substitutes based on biomineralization process, Mater. Sci. Eng. C 25 (2) (2005) 97–104. [43] M. Uchida, et al., Structural dependence of apatite formation on titania gels in a simulated body fluid, Journal of Biomedical Materials Research Part A 64A (1) (2003) 164–170.
