IaX

NRa

b

c

a

ARRAA

KPBTT

1

cewoo

oasocrcs

h0

Applied Surface Science 307 (2014) 372–381

Contents lists available at ScienceDirect

Applied Surface Science

journa l h om epa ge: www.elsev ier .com/ locate /apsusc

n situ characterization of the effects of Nb and Sn on thenatase–rutile transition in TiO2 nanotubes using high-temperature-ray diffraction

athália C. Verissimoa, Alessandra Cremascoa,b, Christiane A. Rodriguesc,odnei Bertazzoli a, Rubens Carama,∗

School of Mechanical Engineering, University of Campinas, Campinas, SP, BrazilSchool of Applied Science, University of Campinas, Limeira, SP, BrazilDepartment of Exact and Earth Sciences, Federal University of São Paulo, São Paulo, SP, Brazil

r t i c l e i n f o

rticle history:eceived 8 January 2014eceived in revised form 4 April 2014ccepted 6 April 2014vailable online 15 April 2014

eywords:hase transformationsiomaterials

a b s t r a c t

New metastable �-type Ti alloys for biomedical applications containing biocompatible alloying elementssuch as Nb can present remarkable mechanical behavior. Whenever the performance of an implantproduced from �-type Ti alloys is considered, it is crucial to take into account their surface proper-ties because they are intimately associated with osseo-integration. The osseo-integration of orthopedicimplant devices made from CP–Ti to �-type Ti alloys depends directly on the properties of the oxide layerformed on their surface. The aim of this study was to investigate the formation of self-organized TiO2

nanotubes by an anodization process on CP–Ti and Ti–35Nb and Ti–35Nb–4Sn alloys (wt.%) and analyzethe effects of Nb and Sn additions to CP–Ti on the amorphous–anatase and anatase–rutile phase transfor-

itanium alloysiO2 nanotubes

mations in TiO2 nanotubes using glazing-angle high-temperature X-ray diffraction. The results obtainedsuggest that the crystallization of TiO2 formed on CP-Ti occurs at 225 ◦C, whereas the anatase–rutile tran-sition occurs at 400 ◦C. As Nb was added to Ti, the temperatures at which these phase transformationsoccur increased. When Sn was added to Ti–35Nb alloy, the kinetics of the phase transformations appearedto decrease.

. Introduction

An interesting combination of properties, such as enhanced bio-ompatibility, high corrosion resistance, good processability and,specially, superior mechanical behavior, makes Ti and its alloysidely employed in biomedical applications. Ti alloys represent

ne of the best classes of metallic materials used in the fabricationf orthopedic implant devices [1,2].

The mechanical behavior of a particular Ti alloy depends directlyn its microstructure. By adding specific alloying elements andpplying proper heat treatments, it becomes possible to achieve auitable combination of types, volume fractions and morphologiesf phases and hence mechanical properties for a given appli-ation [3]. The manufacture of implant devices for hard tissue

epair requires materials with specific properties, including highorrosion resistance, enhanced biocompatibility, high mechanicaltrength and low elastic modulus. Although the needs related to

∗ Corresponding author. Tel.: +55 19 3521 3314; fax: +55 19 3289 3722.E-mail addresses: caram@fem.unicamp.br, rubenscaram@gmail.com (R. Caram).

ttp://dx.doi.org/10.1016/j.apsusc.2014.04.040169-4332/© 2014 Elsevier B.V. All rights reserved.

© 2014 Elsevier B.V. All rights reserved.

biocompatibility, corrosion resistance and mechanical strength arewell known, the requirement of materials with a low elastic mod-ulus arises from the ability to allow for the elastic deformationof the bone tissue surrounding implant devices [1]. When sub-mitted to mechanical stress, bone tissues suffer changes in theirinternal architecture and their external shape. The absence or eventhe reduction of mechanical loading might induce bone loss ordegeneration [4]. This phenomenon suggests that biomaterials tobe employed in hard tissue repair need to present a low Young’smodulus [1].

With respect to biomaterials designed for implants devices, inthe past few years, most efforts associated with the developmentof metallic materials to be used in hard tissue repair have focusedon metastable �-type titanium alloys, prepared by the additionof biocompatible alloying elements, particularly alloys contain-ing Nb [5,6]. Depending on the composition and processing routesemployed, alloys of the Ti–Nb–Sn system can exhibit remarkable

mechanical behavior [7].

In addition to the intrinsic bulk material characteristics, theperformance of an implant device designed for hard tissue repairdepends strongly on the surface features of the device material,

N.C. Verissimo et al. / Applied Surface Science 307 (2014) 372–381 373

) of ho

wcneabsat

otmaatc

rtaa

aibpi

between the implant and the bone tissue provided by osteoblastcells [15]. Inefficient mechanical interaction between regenera-ting tissue and an implant surface may lead to the loosening ofthe implant device [10]. Consequently, implant success is directly

Fig. 1. (a) X-ray diffraction pattern and optical micrographs (OM

hich are directly related to the implant osseo-integration pro-ess [8]. Cells in biological tissues, whose dimensions are on theanometer scale, directly interact with and create nanostructuredxtra-cellular matrices. Interactions between implant biomateri-ls and biological tissues are favored by a nanostructured surfaceecause such a surface may stimulate cell growth and therefore tis-ue regeneration [9,10]. Popat and co-authors have confirmed thebility of nanostructured surfaces to promote osteoblast differen-iation and bone matrix production [10].

When Ti or Ti alloys are exposed to oxygen, a passivation layerf TiO2 is rapidly formed on their surfaces. The controlled forma-ion of TiO2 by potentiostatic anodization, which leads to nano- or

icro-roughness surfaces, provides bioactivity enhancement andlso increases the probability of osseo-integration [11,12]. Thisnodic layer exhibits a high surface-to-volume ratio, and litera-ure results indicate a strong relationship between implant surfaceharacteristics and bone growth [8].

The surface features of the titanium dioxide layer are directlyelated to its polymorphism. It is well known that TiO2 exhibitshree different naturally occurring polymorphs, i.e., brookite, rutilend anatase, whose formation depends on the synthesis procedurend post-processing conditions employed [13].

According to Uchida and co-authors [14], compared to rutile,natase is more effective in providing good conditions for nucleat-

ng hydroxyapatite on Ti implants. The reason for this more suitableehavior is possibly related to the lattice features of the anatasehase and the hydroxyapatite phase. Implant osseo-integration

nvolves a competition between fibroblast cells and osteoblast cells

t-forged (b) CP–Ti, (c) Ti–35Nb and (d) Ti–35Nb–4Sn substrates.

[15]. Whereas smooth surfaces improve fibroblast cell growth, nan-otopographical surfaces favor osteoblast cell fixation and growth[8,10]. The growth of fibrous tissue on an implant surface reducesthe possibility of forming a well-established and strong bond

Fig. 2. Current density vs. anodizing time for CP–Ti, Ti–35Nb and Ti–35Nb–4Snsubstrates anodized using 0.1 vol.% HF and a constant potential of 20 V.

374 N.C. Verissimo et al. / Applied Surface Science 307 (2014) 372–381

Fig. 3. (a) X-ray diffraction pattern of the amorphous oxide nanotubes formed on substrates of Ti–CP, Ti–35Nb and Ti–35Nb–4Sn alloys. Top-view and cross-sectional SEMimages of nanotube layers formed by anodization in 0.1 HF at potentials of 20 V for 1 h on substrates of (b) CP–Ti, (c) Ti–35Nb and (d) Ti–35Nb–4Sn alloys.

605550454035302520

Tiα R R

RRA

Tiα

A

A

A

Ti α

Ti α

R

R

RA

A

A

Ti α

Tiα

Tiα

Tiα

Tiα

Ti α

Ti α

Ti α

710°C

470°C

400°C

360°C

225°C

Inte

nsity

(a.u

.)

2θ (degree s)

25°C

A

Fig. 4. X-ray diffraction patterns obtained during continuous heating and crystal-lization of nanotube layer formed on CP–Ti.

605550454035302520

Tiα

Ti α

Tiα

Tiα

Nb 2

O5

Tiα

Ti α

Tiα

Tiα

Tiβ

Tiβ

Nb 2

O5

Tiβ

Tiβ

Tiβ

Tiβ

Tiβ

Tiβ

A

Nb 2

O5

Inte

nsity

(a.u

.)

2θ (deg ree s)

930°C

560°C

480°C

295°C

25°C

665°C

Nb 2

O5

R

Nb 2

O5

A R

Nb 2

O5

A

R A

Tiβ

A

Tiβ

Fig. 5. X-ray diffraction patterns obtained during continuous heating and crystal-lization of nanotube layer formed on Ti–35Nb substrate.

N.C. Verissimo et al. / Applied Surface Science 307 (2014) 372–381 375

20 25 30 35 40 45 50 55 60

Nb 2

O5

Nb 2

O5

Nb 2

O5

Nb 2

O5

Tiβ

Tiα

Tiα

Tiα

Tiα

Nb 2

O5

Nb 2

O5

RTi β

Tiβ

Ti β

A

R

R

R

A

A

A

Ti β

Tiβ

Tiβ

925°C

775°C

605°C

585°C

390°C

Inte

nsity

(a.u

.)

2θ (degrees )

25°C

A

Fl

rwshrac

tCSu

2

wnaTiaTmtSad

ittsa

200205210215220

Nb3d(5/2)

Nb3d(3/2)

Nb2O5

Inte

nsity

(a.u

.)

Binding energy (eV)

(a)

200205210215220

Nb3d(5/2)Nb3d(3/2)

Nb2O5

Inte

nsity

(a.u

.)

Binding energ y (eV )

(b)

480484488492496500

(c)

Sn3d(3/2)Sn3d(5/2)

SnO2In

tens

ity (a

.u.)

Binding energ y (eV )

Fig. 7. XPS high-resolution spectra of Nb3d tested on the surface of the nanotube

ig. 6. X-ray diffraction patterns obtained during continuous heating and crystal-ization of nanotube layer formed on Ti–35Nb–4Sn substrate.

elated to the implant’s ability to stimulate osseo-integration,hich may be improved by the formation of a layer of nano-

tructured TiO2 on the implant surface [16]. These nanostructuresave features that enhance osteoblast activity (mononucleate cellsesponsible for bone formation), giving rise to cell differentiationnd extracellular matrix production and enhancing proliferation,ellular adhesion and, consequently, calcium deposition.

Therefore, the aim of this study was to investigate the forma-ion of self-organized TiO2 nanotubes by an anodization process onP–Ti, Ti–Nb and Ti–Nb–Sn alloys and analyze the effects of Nb andn additions to Ti on the anatase–rutile transition in TiO2 nanotubessing high-temperature X-ray diffraction.

. Experimental procedure

Ingots of CP–Ti and Ti–35Nb and Ti–35Nb–4Sn alloys (wt.%)ere arc melted using a water-cooled copper hearth and aon-consumable tungsten electrode in an argon atmosphere. Tossure high chemical homogeneity, the ingots of Ti–35Nb andi–35Nb–4Sn alloys were remelted 10 times, submitted to a chem-cal homogenization heat treatment at 1000 ◦C for 12 h undern inert atmosphere and then furnace cooled. The CP–Ti andi–35Nb and Ti–35Nb–4Sn samples were hot swaged at 1000 ◦C andachined to obtain ∅10-mm cylindrical bars. These bars were cut

o produce discs measuring 1 mm in thickness, polished by usingiC sandpaper (1500 grit) and then cleaned with isopropyl alcoholnd demineralized water in an ultrasonic bath. The samples wereried using a dry nitrogen stream.

Highly ordered TiO2 nanotubes were grown using an anodiz-ng process, which was carried out in an electrochemical cell with

wo electrodes, Pt foil as the counter electrode and a working elec-rode consisting of samples of the CP–Ti and Ti alloys. The anodizingolution consisted of a stirred aqueous solution of 0.1 vol.% HF,nd anodizing was carried out for 1 h at room temperature. The

layer of (a) Ti–35Nb and (b) Ti–35Nb–4Sn substrates and of Sn3d tested on thesurface of nanotube layer of (c) Ti–35Nb–4Sn substrate.

anodizing voltage was increased from 0 V to 20 V at a constantsweep rate of 2 V/min using an Autolab potentiostat/galvanostat,model PGSTAT 30, coupled to a voltage multiplier, model ET-2042 C.The voltage was then held constant at 20 V for 1 h.

Microstructure features of the substrates were analyzed by opti-cal microscopy (OM) (Olympus BX60M). The TiO2 layers wereevaluated by scanning electron microscopy (SEM) (Zeiss EVO 15),which allowed for the analysis of the layers’ morphological aspectsand nanotube length. X-ray diffraction (XRD) analyses were per-formed to examine the substrate microstructure and to controlthe crystallization process of the TiO2 nanotube amorphous layer.Samples were heat-treated in a heating device coupled to an X-raydiffractometer (Panalytical X’Pert diffractometer operating at 40 kVand 30 mA, CuK� radiation, � = 0.15406 nm and Pixcel fast detec-tor). To determine the crystallographic details of the TiO2 layer,glazing-angle high-temperature X-ray diffraction (GAHTXRD) mea-surements were performed with a glazing angle of 2◦. The samples

used in the GAHTXRD experiments were the same as those usedin the anodizing process (discs measuring 1 mm in thickness and10 mm in diameter). They were heated continuously in a Pt heater

376 N.C. Verissimo et al. / Applied Surface Science 307 (2014) 372–381

700600500400300200

0102030405060708090

100

Anatase Rutile

f R.1

00%

Temper ature (° C)

1

2

3

(a)

900800700600500400

0102030405060708090

1003

f R .1

00%

Tempera ture ( °C)

Anatas e Ruti lo

1

2

(b)

900800700600500400

0102030405060708090

100

f R. 1

00%

Temper ature (°C)

Anatase Rutil e

1

2

(c)

Fig. 8. Evolution of the anatase-to-rutile transformation as a function of tempera-ta

f2htEsps(se

20 25 30 35 40 45 50 55 60

Tiα

Tiα

Nb 2

O5

Nb 2

O5

Ti β

Tiβ

A

A

A

Tiα

Ti α

Tiα

Ti α

360°C

480°C

Ti-3

5Nb-

4Sn

Ti-3

5Nb

Ti c

.p.

Inte

nsity

(a.u

.)

2θ (degrees)

585° C

A

Fig. 9. X-ray diffraction patterns of nanotube layers containing approximately 100%

ure in the TiO2 nanotubes formed on (a) Ti CP, (b) Ti–35Nb and (c) Ti–35Nb–4Snlloys.

rom room temperature to 1000 ◦C at a heating rate ranging from to 10 ◦C/min, depending on the temperature range. Because theeater and sample temperatures typically differ slightly, an addi-ional thermocouple was used to monitor the sample temperatures.ach GAHTXRD scan covered a 2� range from 20◦ to 60◦. Eachample’s temperature in each scan was taken as the average tem-erature during that scan. The chemical binding energy on theample surface was analyzed by X-ray photoelectron spectroscopy

XPS) (VSW Scientific Instrument—HA100) in constant transmis-ion mode, resulting in a line width for Au 4f7/2 of 1.6 eV. Electronxcitation was carried out using Al K� radiation (1486.6 eV).

anatase on CP–Ti, Ti–35Nb and Ti–35Nb–4Sn after annealing heat treatment.

3. Results and discussion

The TiO2 nanotubes were obtained by electrochemical anodiza-tion on the CP–Ti, Ti–35Nb and Ti–35Nb–4Sn substrates. Thenanotube crystal structure, morphology and dimensional param-eters were analyzed by X-ray diffraction and scanning electronmicroscopy. High temperature X-ray diffraction experiments wereemployed to determine the phase transformation temperatures ofthe nanostructured TiO2 layer. X-ray photoelectron spectroscopytechnique was applied to determine the chemical composition ofthe nanotube layers.

Optical micrographs and X-ray patterns of the CP–Ti, Ti–35Nband Ti–35Nb–4Sn substrates after hot forging are shown in Fig. 1.In the CP-Ti samples, the � phase was observed, demonstrat-ing a serrated morphology. The micrographs of the Ti–35Nb andTi–35Nb–4Sn alloys revealed large deformed � grains, which areconsistent with the X-ray results obtained. The amount of � sta-bilizer used as well as the applied rate of cooling from the initialtemperatures in the � phase field were not sufficient to produce� phase precipitates detectable by the X-ray diffraction analyses.However, optical microscopy evaluation confirmed that a smallvolume fraction of � phase was formed in the Ti–35Nb alloy.

Fig. 2 shows changes in the current density as a function of theanodizing time. The current density presented high values duringthe initial stage of immersion. It appears that there was a signifi-cant increase in the layer thickness, corresponding to passivationbehavior of the substrate (metal) [17]. The increase in the voltage

produced a high electric field effect that induced the formation ofpits on the surface of the oxide. These pits acted as pore nucleation

N.C. Verissimo et al. / Applied Surface Science 307 (2014) 372–381 377

Fig. 10. SEM micrographs showing topographic and cross-section views of TiO2 nanotube layers with anatase structure on (a and b) CP–Ti, (c and d) Ti–35Nb and (e and f)T

cv

weflwrtoe

rcFauF

sos

i–35Nb–4Sn alloys.

enters to produce the nanotubes and caused the current densityalues to increase again [18].

The current density then decreased and reached a plateau, athich the thickness remained relatively unchanged, reaching an

quilibrium. This behavior can be explained by the presence ofuoride ions in the electrolyte during TiO2 nanotube formation,hich resulted in two competing reactions. The first reaction cor-

esponds to the hydrolysis of metal to allow for the formation ofhe TiO2 layer. The other reaction is chemical dissolution at thexide/electrolyte interface. Once an ordered nanotube structure isstablished, the current density stabilizes [19].

After the anodizing process, the samples were submitted to X-ay diffraction analysis using the glazing-angle mode, a methodommonly employed in the evaluation of thin films [20,21].ig. 3 depicts the X-ray diffraction patterns and the morphologicalspects of the nanotubes formed on CP–Ti and its alloys observedsing SEM. According to the X-ray diffraction patterns shown inig. 3(a), only substrate diffraction peaks were observed.

The topographic images shown in Fig. 3(b–d), regardless of theubstrate used, evidence the formation of nanotubular structuresn the surfaces of the CP–Ti and Ti–35Nb and Ti–35Nb–4Sn sub-trates. However, the addition of Nb and Sn as alloying elements

caused changes in the spacing between the tubes walls. Accordingto Jang and co-authors, the growth of nanotubes is affected by theaddition of alloying elements to CP–Ti, and therefore, the compo-sition of the oxides formed may be a function of these additions[22]. The use of substrates with different compositions leads to theformation of a distinct metallic oxide, which exhibits differences indissolution rates. It is likely that the difference in nanotube densityis affected by the dissolution rate.

The average length of nanotubes formed on the CP–Ti substrate,approximately 200 nm, was smaller than that of the nanotubesformed on the Ti alloys. The nanotube oxide layer formed on theTi–35Nb and Ti–35Nb–4Sn substrates showed lengths between 820and 970 nm. Factors such as the electrolyte composition of theanodizing solution directly affect the shape and thickness of thenanotube layers. The influence of these factors was evaluated byLee and co-authors [23]. They observed that the use of fluoride-containing electrolytes increases the current density during theanodization process.

Choe and co-authors [24] verified that the phase features ina substrate’s microstructure affect the nucleation and growth ofnanotubes. The authors investigated the effect of a substrate’smicrostructure containing �, � or �” (orthorhombic martensite)

378 N.C. Verissimo et al. / Applied Surface

605550454035302520

Nb 2

O5

Tiα

Ti α

Tiα

Tiα

Tiα

A

Nb 2

O5

Tiβ

TiβN

b 2O

5

Ti β

R

R

A

Ti α

Tiα

Ti α

775°C

665°C

Ti-3

5Nb-

4Sn

Ti-3

5Nb

CP

Ti

Inte

nsity

(a.u

.)

2θ (degrees)

470°C

A

A

R

Fr

pitomacttsCaach

oaantce

f[tiaoa

ig. 11. X-ray diffraction patterns of nanotube layers with a mixture of anatase andutile on CP–Ti, Ti–35Nb and Ti–35Nb–4Sn alloys after annealing heat treatment.

hases on nanotube formation on Ti–30Nb–xZr alloys immersedn fluoride-containing electrolytes and observed that nanotubeshat formed on the � phase exhibited a regular array, whereasn the � phase, nanotube formation was irregular. Finally, on �”artensite, the authors observed nanotubes with elongated shapes

nd random diameters. Similar results were observed by Kim ando-authors [25] regarding the effect of microstructure on nano-ube formation. In this study, the thickness differences observed inhe TiO2 nanotube layers were determined to be associated withubstrate features. For instance, although the microstructure ofP–Ti was fully formed in the � phase, in the case of the Ti–Nbnd Ti–Nb–Sn alloys, the � phase was the predominant phase. Inddition, heat treatments could change the phase combination, andrystal defects may have assisted nanotube formation by acting aseterogeneous nucleation sites [26].

However, in a recent study on the microstructural modificationf a CP–Ti substrate by the roll-bonding process, Tsuchiya and co-uthors [27] observed that the diameter and length of nanotubesnd the degree of ordering of the nanostructured oxide layer wereot affected by the grain size or the number of crystal defects. Onhe other hand, they observed that the variation in the chemicalomposition and/or the anodic oxidation parameters had a strongerffect on the nanotube layer features [27,28].

The nanostructured oxide layer formed on the substrate sur-ace is due to interactions between the metal and O2− or OH− ions29]. Initially, this oxide layer exhibits an amorphous structure andhus shows only diffraction peaks corresponding to the substrate

n X-ray diffraction patterns, as depicted in Fig. 2a. Heat treatmentspplied to this amorphous oxide layer result in the crystallizationf the TiO2 thin layer, transforming the amorphous layer into thenatase or rutile phase [30]. To study the effects of temperature

Science 307 (2014) 372–381

and the Sn and Nb additions on the phase transformations of TiO2,HTXRD experiments were carried out. The in-situ X-ray diffrac-tion patterns of the CP–Ti samples submitted to heat treatmentare shown in Fig. 4. The patterns demonstrate that the nucleationof the anatase polymorphic structure occurred at 225 ◦C, whereasthe TiO2 nanotubes formed on the Ti–35Nb and Ti–35Nb–4Sn sub-strates showed crystallization temperatures of 295 ◦C and 390 ◦C,respectively, as shown in Figs. 5 and 6.

The polymorphic phase transformation of anatase into rutilewas observed at temperatures close to 400 ◦C on CP–Ti, 560 ◦Con Ti–35Nb alloy and 605 ◦C on Ti–35Nb–4Sn alloy. Comparedto the CP–Ti substrate, the Nb and Sn additions led to the for-mation of microstructures composed essentially of the � phase,affected the growth of the nanotube layer and changed the tem-perature of the anatase–rutile phase transformation. The resultsobtained from the HTXRD experiments indicate that Nb has ananatase-stabilizing behavior; specifically, Nb led to an increase inthe anatase–rutile transition temperature. These results also showthat Sn addition improves the stabilizing effect of Nb, also increas-ing the anatase–rutile phase transformation temperature.

In this work, the formation of Nb2O5 was detected by thehigh-temperature X-ray diffraction experiments performed on theTi–35Nb and Ti–35Nb–4Sn alloys, which was confirmed by chem-ical composition analysis using X-ray photoelectron spectroscopy,as shown in Fig. 7. On the Ti–35Nb alloy, the formation of Nb2O5was confirmed by the Nb3d3/2 peak at 209.6 eV and Nb3d5/2 peakat 206.8 eV (Fig. 7a), whereas on the Ti–35Nb–4Sn, the formationof the same oxide was verified by the Nb3d3/2 peak at 209.6 eVand Nb3d5/2 peak at 206.9 eV (Fig. 7b) [31,32]. The formation ofSnO2 was also confirmed by the Sn3d5/2 and Sn3d3/2 peaks at 486.5and 494.8 eV, respectively (Fig. 7c) [31,33]. Based on the HTXRDpatterns, the amount of Nb oxide formed was greatly reduced,which is attributed to the oxide’s high chemical stability in fluo-ride electrolytes compared to TiO2 [34]. According to Minagar andco-authors, the Nb2O5 layer is able to improve the wear resistanceof Ti-based alloys [28].

During annealing using HTXRD, the diffraction peaks corre-sponding to the rutile phase showed an increase in intensity,whereas those corresponding to the anatase phase continuouslydecreased in intensity. The complete transformation into rutile wasonly observed in nanotubes formed on the CP–Ti and Ti–35Nb alloysubstrates. The volume fraction of anatase and rutile during thisphase transformation can be determined by using the diffractionpeak intensities corresponding to 2� = 25◦ for anatase and 2� = 27◦

for rutile and by applying the equations suggested by Slimen andco-authors [35]:

fA = KA × IA[KA × IA + IR]

(1)

fR = IR[KA × IA + IR]

(2)

where fA and fR are the volume fractions of the anatase andrutile polymorphic phases, respectively, IA and IR correspond tothe diffraction peak intensities of anatase (1 0 1) and rutile (1 1 0),respectively, and KA is a constant equal to 0.886. Fig. 8 shows theevolution of the anatase and rutile volume fractions, and the pos-itions (1) and (3) correspond to the extremes: 100% anatase and100% rutile, respectively.

The evolution of the volume fractions of anatase and rutileformed on the CP–Ti (Fig. 8a), Ti–35Nb (Fig. 8b) and Ti–35Nb–4Sn

(Fig. 8c) substrates shows that a mixture of 50% anatase and 50%rutile, position 2, occurred at 425 ◦C on CP–Ti, at 625 ◦C on Ti–35Nballoy and at 800 ◦C on Ti–35Nb–4Sn alloy. In the nanotube layerformed on CP–Ti, 100% rutile was observed at 560 ◦C and 850 ◦C

N.C. Verissimo et al. / Applied Surface Science 307 (2014) 372–381 379

F iO2 nanotube layers containing a mixture of anatase and rutile on (a and b) CP–Ti, (c andd

on

rpaira

aaslwnrta8vi

tnsaof

ig. 12. SEM micrographs showing the topographic and cross-section views of the T) Ti–35Nb and (e and f) Ti–35Nb–4Sn alloys.

n Ti–35Nb alloy. However, 100% rutile was not observed in theanotube layer formed on Ti–35Nb–4Sn alloy up to 925 ◦C.

To investigate the morphological features of the nanostructu-ed layer as a function of the content of the anatase and rutilehases, new samples were prepared based on the evolution of thenatase–rutile transformation presented in Fig. 8. Samples contain-ng approximately 100% anatase, a mixture of 50% anatase and 50%utile and nearly 100% rutile were obtained. The samples were heldt the corresponding temperature for a period of 30 min.

The X-ray diffraction patterns of samples containing 100%natase are shown in Fig. 9. Patterns corresponding to the Ti–35Nbnd Ti–35Nb–4Sn substrates also show Nb2O5 oxide peaks. Fig. 10hows topographic and cross-sectional SEM images of the nanotubeayers with nearly 100% anatase formed on the three substrates

ith their respective thicknesses. The morphological aspects of theanotube layers in the annealed samples are very similar to thoseelated to the amorphous layers. Regarding the layer thickness,he values measured are also similar to those obtained before thennealing heat treatment, close to 240 nm on CP–Ti and between60 and 990 nm on the Ti–35Nb and Ti–35Nb–4Sn alloys. It waserified that the regularity of the nanotubes formed on CP–Ti wasmproved after annealing.

A mixture of anatase and rutile phases was obtained by heatreatments carried out at 470, 665 and 775 ◦C applied to theanotubes layer formed on the CP–Ti, Ti–35Nb and Ti–35Nb–4Sn

ubstrates. The respective XRD patterns are presented in Fig. 11. Thennealing heat treatment employed to crystallize the amorphousxide layer did not modify the morphology of the TiO2 nanotubesormed on the Ti–35Nb and Ti–35Nb–4Sn alloys, as shown in Fig. 12.

Fig. 13. X-ray diffraction patterns of nanotube layers containing nearly 100% rutileon CP–Ti, Ti–35Nb and Ti–35Nb–4Sn alloys after annealing heat treatment.

380 N.C. Verissimo et al. / Applied Surface Science 307 (2014) 372–381

F 2 nano(

TwNmit

ttsoolf

t9rpwtN

ig. 14. SEM micrographs showing topographic and cross-section views of the TiOe and f) Ti–35Nb–4Sn alloys.

he reason for why this morphology was maintained is associatedith the stabilization of the anatase phase due to the additions ofb and Sn. In the case of the CP–Ti substrate, it appears that the for-ation of rutile in the oxide layer was followed by a slight increase

n nanotube diameter, suggesting the onset of the deterioration ofhe nanotube morphology.

Regarding the layer thickness, it was observed that the nano-ube layers formed on the CP–Ti and Ti–35Nb alloys increased inhickness to approximately 500 nm and 1400 nm, respectively, ashown in Fig. 12b and d. The same phenomenon was not observedn the Ti–35Nb–4Sn substrate. When Sn was added, the formationf the rutile phase was suppressed or retarded and the nanotubeayer thickness was maintained. This result also suggests that theormation of rutile occurs at the nanotube-substrate interface [36].

To promote the complete anatase–rutile phase transformation,he nanotube layers were annealed for 30 min at 710 ◦C on CP–Ti, at30 ◦C on Ti–35Nb and at 925 ◦C on Ti–35Nb–4Sn. According to theesulting XRD patterns shown in Fig. 13, the rutile structure was

redominately observed in the nanotube layer formed on CP–Ti,hereas the anatase phase remained stable in the layers formed on

he Ti–35Nb and Ti–35Nb–4Sn substrates. In addition, pronouncedb2O5 peaks were detected by XRD analysis.

tubes layers containing nearly 100% rutile: (a and b) Ti–CP, (c and d) Ti–35Nb and

Fig. 14 shows the micrographs of the TiO2 surface layers formedpredominantly by rutile on the CP–Ti, Ti–35Nb and Ti–35Nb–4Snsubstrates. The oxide layer containing rutile that formed on CP-Tiexhibited peculiar aspects. Although its nanotube arrangement wascompletely destroyed and replaced by a layer of fragmented crys-tals, its thickness was apparently reduced due to the increase inthe rutile volume fraction. In the case of the Ti–35Nb substrate,the annealing treatment initiated the deterioration process dueto the high content of rutile. However, the layer’s thickness wasmaintained at approximately 1500 nm. Finally, the morphology ofthe layer on the Ti–35Nb–4Sn remained comparable to that of thelayer before the heat treatment, suggesting that formation of rutilecauses a loss of nanotube regularity and modifies the layer’s thick-ness.

4. Conclusions

The effects of Nb and Sn additions to CP–Ti on the formation

of nanostructured TiO2 layers and on their phase transformationswere investigated using high-temperature X-ray diffraction exper-iments and scanning electron microscopy. According to the resultsobtained, the morphology and size of the TiO2 nanotube layers

urface

aftp

fttraprTtTwtioNfT

pttmrtn

A

iRTSsU

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

N.C. Verissimo et al. / Applied S

re very sensitive to the composition of the substrates. Substrateeatures play an important role in controlling the kinetics andemperatures of the amorphous-to-anatase and anatase-to-rutilehase transformations.

Under annealing heat treatment, the crystallization of TiO2ormed on CP-Ti occurred at 225 ◦C, whereas the TiO2 formed onhe Ti–35Nb and Ti–35Nb–4Sn substrates showed crystallizationemperatures of 295 ◦C and 390 ◦C, respectively. The anatase-to-utile phase transformation temperature was also affected by theddition of Nb and Sn to CP–Ti. Although the TiO2 formed on CP–Tiresented the anatase–rutile transition at 400 ◦C, the addition of Nbesulted in an increase of approximately 120 ◦C in that temperature.he addition of Sn to the Ti–35Nb alloy was observed to improvehe Nb effect on the stability of the anatase phase. Compared to theiO2 layer formed on CP–Ti, an increase of approximately 240 ◦Cas detected in the anatase–rutile phase transformation tempera-

ure of the oxide layer formed on Ti–35Nb–4Sn alloy, a significantncreasing compared with that observed upon the addition of Nbnly. X-ray photoelectron spectroscopy confirmed the formation ofb2O5 on the Ti–35Nb and Ti–35Nb–4Sn substrates, whereas the

ormation of SnO2 was also confirmed in the anodic layer of thei–35Nb–4Sn substrate.

The formation of rutile on CP–Ti substrate caused the com-lete deterioration of the nanotube arrangement. In the case ofhe Ti–35Nb substrate, the rutile phase that formed resulted inhe partial deterioration of the nanotube structure. Finally, the

orphology of the anodic layer on the Ti–35Nb–4Sn substrateemained comparable to that observed before the annealing heatreatment, which suggests that rutile formation leads to a loss ofanotube regularity.

cknowledgments

The authors gratefully acknowledge the Brazilian research fund-ng agencies FAPESP (Grant 2011/23942-6) (State of São Pauloesearch Foundation), CNPq (National Council for Scientific andechnological Development) and CAPES (Federal Agency for theupport and Evaluation of Graduate Education) for their financialupport of this work and Dr. Richard Landers, Institute of Physics,niversity of Campinas, for providing the XPS results.

eferences

[1] M. Geetha, A.K. Singh, R. Asokamani, A.K. Gogia, Ti based biomaterials, the ulti-mate choice for orthopaedic implants—a review, Prog. Mater. Sci. 54 (2009)397–425.

[2] M. Niinomi, Recent research and development in titanium alloys for biomed-ical applications and healthcare goods, Sci. Technol. Adv. Mater. 4 (2003)445–454.

[3] X. Tang, T. Ahmed, H.J. Rack, Phase transformations in Ti–Nb–Ta andTi–Nb–Ta–Zr alloys, J. Mater. Sci. 35 (2000) 1805–1811.

[4] A.C. Ahn, A.J. Grodzinsky, Relevance of collagen piezoelectricity to Wolff’s Law:a critical review, Med. Eng. Phys. 31 (2009) 733–741.

[5] A. Cremasco, P.N. Andrade, R.J. Contieri, E.S.N. Lopes, C.R.M. Afonso, R.Caram, Correlations between aging heat treatment, � phase precipitationand mechanical properties of a cast Ti–Nb alloy, Mater. Des. 32 (2011)2387–2390.

[6] E.S.N. Lopes, A. Cremasco, C.R.M. Afonso, R. Caram, Effects of double aging

heat treatment on the microstructure, Vickers hardness and elastic modulusof Ti–Nb alloys, Mater. Charact. 62 (2011) 673–680.

[7] B.L. Wang, Y.F. Zheng, L.C. Zhao, Effects of Sn content on the microstructure,phase constitution and shape memory effect of Ti–Nb–Sn alloys, Mater. Sci.Eng., A 486 (2008) 146–151.

[

Science 307 (2014) 372–381 381

[8] S. Oh, S. Jin, Titanium oxide nanotubes with controlled morphology forenhanced bone growth, Mater. Sci. Eng., C 26 (2006) 1301–1306.

[9] L. Zhang, T.J. Webster, Nanotechnology and nanomaterials: promises forimproved tissue regeneration, Nano Today 4 (2009) 66–80.

10] K.C. Popat, L. Leoni, C.A. Grimes, T.A. Desai, Influence of engineered titaniananotubular surfaces on bone cells, Biomaterials 28 (2007) 3188–3197.

11] L.L. Guehennec, A. Soueidan, P. Layrolle, Y. Amouriq, Surface treatments oftitanium dental implants for rapid osseointegration, Dent. Mater. 23 (2007)844–854.

12] Y.Q. Hao, S.J. Li, Y.L. Hao, Y.K. Zhao, H.J. Ai, Effect of nanotube diameters onbioactivity of a multifunctional titanium alloy, Appl. Surf. Sci. 268 (2013) 44–51.

13] U. Diebold, The surface science of titanium dioxide, Surf. Sci. Rep. 48 (2003)53–229.

14] M. Uchida, H-M. Kim, T. Kokubo, S. Fujibayashi, T. Nakamura, Structuraldependence of apatite formation on titania gels in a simulated body fluid, J.Biomed. Mater. Res. A 64 (2003) 164–170.

15] N. Hori, F. Iwasa, T. Ueno, K. Takeuchi, N. Tsukimura, M. Yamada, M. Hattori, A.Yamamoto, T. Ogawa, Selective cell affinity of biomimetic micro-nano-hybridstructured TiO2 overcomes the biological dilemma of osteoblasts, Dent. Mater.26 (2010) 275–287.

16] V.V.D. Rani, L. Vinoth-Kumar, V.C. Anitha, K. Manzoor, M. Deepthy, V.N.Shantikumar, Osteointegration of titanium implants is sensitive to specificnanostructure morphology, Acta Biomater. 8 (2012) 1976–1989.

17] Z. Lockman, S. Sreekantan, S. Ismail, L. Schmidt-Mende, J.L. MacManus-Driscoll.,Influence of anodisation voltage on the dimension of titania nanotubes, J. AlloysCompd. 503 (2010) 359–364.

18] K. Yasuda, J.M. Macak, S. Berger, A. Ghicov, P. Schmuki, Mechanistic aspects ofthe self-organization process for oxide nanotube formation on valve metals, J.Electrochem. Soc. 154 (2007) C472–C478.

19] G.A. Crawford, N. Chawla, K. Das, S. Bose, A. Bandyopadhyay, Microstructure anddeformation behavior of biocompatible TiO2 nanotubes on titanium substrate,Acta Biomater. 3 (2007) 359–367.

20] W.G. Liu, L.B. Kong, L.Y. Zhang, X. Yao, Study of the surface layer of lead titanatethin film by X-ray diffraction, Solid State Commun. 93 (1995) 653–657.

21] O.K. Varghese, D. Gong, M. Paulose, C.A. Grimes, E.C. Dickey, Crystallizationand high-temperature structural stability of titanium oxide nanotube arrays, J.Mater. Res. 18 (2003) 156–165.

22] S-H. Jang, H-C. Choe, Y-M. Ko, W.A. Brantley, Electrochemical characteristics ofnanotubes formed on Ti–Nb alloys, Thin Solid Films 517 (2009) 5038–5043.

23] B.G. Lee, J.W. Cho, S-E. Lee, Y-S. Jeong, H-J. Oh, C-S. Chi, Formation behavior ofanodic TiO2 nanotubes in fluoride containing electrolytes, Trans. NonferrousMet. Soc. China 19 (2009) 842–845.

24] H-N. Choe, W-G. Kim, Y-H. Jeong, Surface characteristics of HA coatedTi–30Ta–Zr and Ti–30Nb–xZr alloys after nanotube formation, Surf. Coat. Tech-nol. 205 (2010) S305–S311.

25] W-G. Kim, H-C. Choe, Y-M. Ko, W.A. Brantley, Nanotube morphology changesfor Ti–Zr alloys as Zr content increases, Thin Solid Films 517 (2009) 5033–5037.

26] C.A. Grimes, G.K. Mor, TiO2 Nanotube Arrays: Synthesis, Properties and Appli-cations, Springer, London, UK, 2009.

27] H. Tsuchiya, T. Akaki, J. Nakata, D. Terada, N. Tsuji, Y. Koizumi, Y. Minamino, P.Schmuki, S. Fujimoto, Metallurgical aspects on the formation of self-organizedanodic oxide nanotube layers, Electrochim. Acta 54 (2009) 5155–5162.

28] S. Minagar, C.C. Berndt, J. Wang, E. Ivanova, C. Wen, A review of the applicationof anodization for the fabrication of nanotubes on metal implant surfaces, ActaBiomater. 8 (2012) 2875–2888.

29] B. Luo, H. Yang, S. Liu, W. Fu, P. Sun, M. Yuan, Y. Zhang, Z. Liu, Fabricationand characterization of self-organized mixed oxide nanotube arrays by elec-trochemical anodization of Ti-6Al-4V alloy, Mater. Lett. 62 (2008) 4512–4515.

30] B. Yang, C.K. Ng, M.K. Fung, C.C. Ling, A.B. Djurisic, S. Fung, Annealing study oftitanium oxide nanotube arrays, Mater. Chem. Phys. 130 (2011) 1227–1231.

31] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-Ray Photo-electron Spectroscopy, Perkin-Elmer, Minnesota, MN, 1992.

32] M.Z. Atashbar, H.T. Sun, B. Gong, W. Wlodarski, R. Lamb, XPS study of Nb-dopedoxygen sensing TiO2 thin films prepared by sol–gel method, Thin Solid Films326 (1998) 238–244.

33] P.Y. Liu, J.F. Chen, W.D. Sun, Characterizations of SnO2 and SnO2:Sb thin filmsprepared by PECVD, Vacuum 76 (2004) 7–11.

34] X.J. Feng, J.M. Macak, S.P. Albu, P. Schmuki, Electrochemical formation of self-organized anodic nanotube coating on Ti–28Zr–8Nb biomedical alloy surface,Acta Biomater. 4 (2008) 318–323.

35] H. Slimen, A. Houas, J.P. Nogier, Elaboration of stable anatase TiO2 through

activated carbon addition with high photocatalytic activity under visible light,J. Photochem. Photobiol., A: Chem. 221 (2011) 13–21.

36] J. Yu, B. Wang, Effect of calcination temperature on morphology and photo-electrochemical properties of anodized dioxide nanotube arrays, Appl. Catal.94 (2010) 295–302.