Materials Science and Engineering C 29 (2009) 1942–1949

Contents lists available at ScienceDirect

Materials Science and Engineering C

j ourna l homepage: www.e lsev ie r.com/ locate /msec

Thermal oxidation of CP-Ti: Evaluation of characteristics and corrosion resistance as afunction of treatment time

Satendra Kumar a, T.S.N. Sankara Narayanan a,⁎, S. Ganesh Sundara Raman b, S.K. Seshadri b

a National Metallurgical Laboratory, Madras Centre, CSIR Complex, Taramani, Chennai-600 113, Indiab Department of Metallurgical and Materials Engineering Indian Institute of Technology Madras, Chennai-600 036, India

⁎ Corresponding author. Tel.: +91 44 2254 2077; fax:E-mail addresses: sat_nml@rediffmail.com (S. Kuma

(T.S.N.S. Narayanan).

0928-4931/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.msec.2009.03.007

a b s t r a c t

a r t i c l e i n f o

Article history:Received 23 October 2008Received in revised form 12 February 2009Accepted 4 March 2009Available online 13 March 2009

Keywords:Thermal oxidationCP-titaniumCorrosion behaviourX-ray diffractionBiomedical applicationElectrochemical characterization

Commercially pure titanium (CP-Ti) samples were subjected to thermal oxidation (TO) treatment at 650 °Cfor 8, 16, 24 and 48 h. The morphological features, structural characteristics, microhardness and corrosionresistance in Ringer's solution of thermally oxidized samples were compared with that of the untreated one,to ascertain the suitability of thermally oxidized sample as a bio-implant. The thickness, morphologicalfeatures and phase constituents of the oxide film formed during thermal oxidation (TO) exhibit a strongdependence on the treatment time. Samples oxidized for 48 h lead to the formation of oxide grains alongwith a thick oxide film consisting of rutile and TiO phase. Samples oxidized for 24 h lead to the formation ofoxide grains with thinner oxide layer at the grain boundary. Almost a 3 fold increase in hardness is observedfor samples oxidized for 48 h compared to that of the untreated sample. Based on the corrosion protectiveability, the untreated and thermally oxidized samples can be ranked as follows: {TO 48 h}N {TO 16 h}N{TO 8 h}≈{TO 24 h}Nuntreated. From corrosion protection point of view, TO for 48 h is a promising surfacetreatment and it can be a suitable alternative to the untreated CP-Ti as a bio-implant.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Titanium and titanium alloys are widely used as bio-implantmaterials, particularly for orthopedic and osteosynthesis applicationsdue to their low density, excellent biocompatibility, corrosionresistance and mechanical properties [1–4]. One of the importantreasons for choosing these materials is their ability to develop anaturally formed passive oxide layer, typically of 4–6 nm thickness,which is comprised of either amorphous or poorly crystallized non-stoichiometric TiO2. This passive oxide layer is highly stable and itsneutral behaviour in corrosive medium provides excellent corrosionprotection. However, the stability of the passive oxide layer could bealtered under in vivo conditions. Mu et al. [5] have identifiedaccumulation of metal ions on adjacent tissues when the implantswere retrieved for analysis. The main reason for this occurrence is dueto the inferior mechanical properties of the native forms of TiO2 filmsthat can be disrupted at very low shear stresses, even by rubbingagainst soft tissues [6]. Due to the inherent property of the titaniumand its alloys, the passive oxide layer could subsequently form uponreaction with the local environment. However, the wear debris andthemetal ions released during fracture of the passive layer could causeadverse tissue reactions. Fretting and sliding wear conditions couldalso lead to fracture of the passive oxide layer [7–10]. Under extreme

+91 44 2254 1027.r), tsnsn@rediffmail.com

ll rights reserved.

conditions, the effects of fretting and sliding wear might lead toloosening and eventual failure of the implant, causing suffering to thepatients and warrants re-surgery. The above limitations preclude theuse of Ti and its alloys for articulating surfaces.

To meet the challenging demands and, to circumvent the problemsdue to the poor mechanical property of the naturally formed oxidefilms on titanium implants, numerous surface modification methodssuch as, chemical treatment (acid and alkali treatment) [11,12],electrochemical treatment (anodic oxidation) [13], sol–gel coatings[14], chemical vapour deposition [15], physical vapour deposition [16],plasma spray deposition [17], ion implantation [18], thermal oxidation[19], etc. have been explored. Among them, thermal oxidation (TO)was found to be a cost-effective method to deliberately generate abarrier oxide layer of relatively higher thickness (~20–30 μm) ontitanium compared to the naturally formed oxide layer (typically of 4–6 nm). TO treatment of titanium is aimed to produce in situ ceramiccoatings, mainly based on rutile, in the form of a thick, highlycrystalline oxide film, which is accompanied by the dissolution ofoxygen beneath them. The thermally formed oxide layer enables anincrease in hardness, wear resistance and corrosion resistance oftitanium and its alloys [20–24]. However, the improvement in thesefunctional properties becomes limited when the treatment tempera-ture and time are increased beyond a certain limit. TO treatment oftitanium ≥800 °C and for prolonged time duration results inthickening of the oxide layer, which eventually spalls-off from thesurface [25,26]. Dearnley et al. [26] have studied the oxidationbehaviour of the CP-Ti at 800 °C for 36 h. They have reported that the

1943S. Kumar et al. / Materials Science and Engineering C 29 (2009) 1942–1949

developed thicker TiO2 layer (~25 µm) fails along the oxide-oxygendiffused zone interface which is supported by the formation of manycracks on the oxide layer. The large volume ratio of rutile to Ti (1.73)[27], large lattice mismatch and the large difference in coefficient ofthermal expansion between rutile and titanium are considered to beresponsible for the spallation of the oxide layer from the substrate[28]. Hence, it is important to optimize the treatment temperature andtime so as to prepare an adherent, homogeneous and thick surfaceoxide layer, preferably with rutile structure so that an improvement inhardness, wear resistance and corrosion resistance could be realized.In this perspective, the present paper aims to study the effect oftreatment time (viz. 8,16, 24 and 48 h) on the thermal oxidation of CP-Ti at 650 °C to optimize the treatment condition. Corrosion resistanceis an important property of a biomaterial as it determines thebiocompatibility of the material. The corrosion resistance of TOtitanium in simulated body fluids has been reported in the literature[20,29]. However, there are limited results available on the corrosionresistance of CP-Ti thermally oxidized as a function of treatment time,which is rather limited. Hence, the corrosion resistance of TO CP-Ti inRinger's solution (simulated body fluid environment) was alsoevaluated to identify the optimum treatment time.

2. Experimental details

CP-Ti (grade-2) (chemical composition inwt.%: N: 0.01; C: 0.03; H:0.01; Fe: 0.20; O: 0.18 and Ti: Balance) of 2 mm thickness was used asthe substrate. The CP-Ti samples were mechanically polished usingvarious grades of SiC paper, rinsed in deionized water and dried using

Fig. 1. SEM images of surface morphology of samples thermally oxidized

a stream of compressed air. TO of samples was carried out in air in arectangular furnace (LENTON make) at 650 °C for different periods oftime viz. 8, 16, 24 and 48 h. The rate of heating was kept at 5 °C/min. inall the cases. After oxidation treatment the samples were allowed tocool in the furnace itself. The temperature employed for oxidationwasfixed at 650 °C based on the available literature. Debonding andspalling of the oxide layer were noticed when the thermal oxidation ofCP-Ti was done at temperatures higher than 800 °C and for longertime [25–27]. In view of this it was decided to fix the temperature notclose to 800 °C but at 650 °C.

The phase constituents of untreated and thermally oxidizedsamples and the nature of the oxide film formed on the surfacewere determined by X-ray diffraction (XRD) (D8 DISCOVER, Bruckeraxs) using Cu–Kα radiation. The thickness of the oxide layer wasdetermined using scanning electron microscopy (SEM). For measur-ing the thickness of the thermally oxidized layer, the cross-section ofthe treated sample was analyzed using SEM. The thermally oxidizedsample was cut using a slow speed cutter (Buehler) using a very lowload, followed by mounting and polishing using various grades of SiCpaper and 0.3 µm alumina paste to a mirror finish, rinsed withdeionized water, etched with Kroll's reagent for 10–15 s. Surfacemorphology of oxide was also assessed by SEM. The microhardness ofthe untreated and thermally oxidized samples was measured at thesurface using a Leica VMHTMOT microhardness tester at a load of200 gf applied for 15 s. Seven indentations were made on each sampleand the values were averaged out.

Thecorrosion resistanceof untreated andthermallyoxidized sampleswas evaluated by potentiodynamic polarization and electrochemical

at 650 °C for various time periods; (a) 8, (b) 16, (c) 24 and (d) 48 h.

1944 S. Kumar et al. / Materials Science and Engineering C 29 (2009) 1942–1949

impedance spectroscopic (EIS) studies using apotentiostat/galvanostat/frequency response analyzer of ACM instruments (model: Gill AC).Ringer's solution, having a chemical composition (in g/l) of 9 NaCl, 0.24CaCl2, 0.43 KCl and 0.2 NaHCO3 (pH: 7.8), that chemically simulates thephysiological medium of the human body, which is also referred assimulated body fluid (SBF), was used as the electrolyte solution. All thecorrosionexperimentswereperformedat roomtemperature (27±1°C).Before performing the corrosion studies, the untreated sample wasmechanically polished using various grades of SiC paper followed byalumina polishing to get mirror finish, rinsed with deionized water,pickled usingamixture of 35 vol.%, HNO3, 5 vol.%HFandbalancewater at40 °C for 60–70 s as per the guidelines specified by Schutz and Thomas[30]. The pickling of untreated sample was performed to remove thepassive film that is already present so that it can develop a fresh passiveoxidefilm just before subjecting themto corrosion studies. After pickling,the untreated sample was thoroughly rinsed in deionized water anddried using a stream of compressed air. The thermally oxidized sampleswere rinsed in deionized water and dried using a stream of compressedair. The cleaned samples, either untreated or thermally oxidized, formedthe working electrode while a saturated calomel electrode (SCE) and agraphite rod served as the reference and auxiliary electrodes, respec-tively. These electrodes were placed in a flat cell in such a way that only1 cm2 area of the working electrode was exposed to the electrolytesolution.

Potentiodynamic polarization tests were carried out at a scan rateof 100 mV/min from −250 to +3000 mV vs. SCE with respect to theopen circuit potential (OCP). A very high anodic potential (+3000mVvs. SCE) was chosen to study the effect of oxidation conditions and

Fig. 2. SEM images of cross-section of thermally oxidized sample fo

consequently the oxide layer thickness on the passivity and break-down of the oxide layer in Ringer's solution, if any. The corrosionpotential (Ecorr) and corrosion current density (icorr) were determinedfrom the polarization curves using Tafel extrapolation method. Thepassive current density (ipass) was determined from the anodic regionof the polarization curve at +1000 mV vs. SCE. Electrochemicalimpedance spectroscopy (EIS) study of untreated and thermallyoxidized samples was conducted by applying a sinusoidal signalamplitude of 32 mV and the electrode response was analyzed in thefrequency range between 10,000 Hz and 0.1 Hz, in Ringer's solution, attheir respective OCP's. The charge transfer resistance (Rct) and doublelayer capacitance (Cdl) were determined from the Nyquist plot byfitting the data using Boukamp software. The potentiodynamicpolarization and EIS studies were repeated at least 3 times so as toensure reproducibility of the test results.

3. Results and discussion

3.1. Surface morphology and thickness of the oxide film

The surface morphology of thermally oxidized samples clearlyreveals the presence of oxide scales throughout the surface withoutspallation, irrespective of the time of oxidation (Fig. 1a–d). The oxidescales are relatively smooth for samples oxidized for 8 and 16 hwhereas they become relatively rougher when the oxidation time isincreased to 24 and 48 h. After 8 h of oxidation, the oxide scale coversthroughout the surface (Fig.1(a)). Obviously, the duration of oxidationis considerably high to achieve a good surface coverage with the oxide

r various time periods: (a) 8 h, (b) 16 h, (c) 24 h and (d) 48 h.

Fig. 3. XRD patterns of untreated and thermally oxidized samples.

1945S. Kumar et al. / Materials Science and Engineering C 29 (2009) 1942–1949

scale. Garcia-Alonso et al. [29] have reported that the surface of Ti–6Al–4V alloy is fully covered with oxide scales in just one hour when thesamples are oxidized at 700 °C. Agglomeration of the oxide begins after16 h of oxidation (Fig. 1(b)). However, the surface is fully covered withthe oxide scale at this stage. The extent of agglomeration is increasedafter 24 h of oxidation, resulting in the formation of bigger grains,which thereby reduce the thickness of oxide film at the grain boundary(Fig. 1(c)). When the oxidation time is increased to 48 h, there isconsiderable grain growth and each oxide grain is attached with theneighbouring grains so that the surface is homogeneously coveredwith a thicker layer of oxide (Fig. 1(d)). The evolution of surfacemorphology of the oxide film as a function of oxidation time suggeststhe mechanism of oxide film formation. Accordingly, the nucleation of

Fig. 4. Schematic representation of XRD measurement of thermally oxidized

oxide takes place throughout the surface when it immediately comesin contact with oxygen whereas the growth mode involves theformation of a thin oxide scale followed by its agglomeration andgrowth, to completely cover the surface. The SEM of cross-sectionalview of CP-Ti sample oxidized at 650 °C for different durations of timesuch as 8, 16, 24 and 48 h reveals the presence of a thick andhomogeneous oxide layer and thickness of the oxide layer increaseswith the increase in treatment time (Fig. 2). The thickness of the oxidelayer is increased from3.5 to 19 µmwith the increase in treatment timefrom 8 to 48 h.

3.2. Structural characteristics of untreated and thermally oxidizedsamples

The XRD patterns of untreated and thermally oxidized CP-Tisamples are shown in Fig. 3. The untreated CP-Ti sample is entirelycomprised of hexagonal α-phase (denoted as ‘Ti’ in Fig. 3). The XRDpatterns of thermally oxidized samples exhibit the presence of rutile(denoted as ‘R’ in Fig. 3) and oxygen diffused Ti (denoted as ‘TiO’ inFig. 3) as the predominant phases along with a small amount of α-Ti.The presence of α-Ti peaks is clearly evident in thermally oxidized CP-Ti samples having a thinner oxide layer, since the Cu–Kα radiationcould penetrate to a depth of 10–20 µm, much larger than thethickness of the oxide film. When the CP-Ti samples are oxidized for 8and 16 h, α-Ti and TiO become the predominant peaks whereas peakspertaining to the rutile phase are relatively small. However, when theoxidation time is increased to 24 and 48 h, peaks pertaining to therutile phase become dominant with a corresponding decrease in theintensity of the α-Ti peaks. Similar observations were also made byother researchers [20,23,29]. Siva Rama Krishna et al. [23] havereported the formation of TiO at temperatures less than 700 °C andrutile as the dominant phases at and above 800 °C, respectively.Guleryuz and Cimenoglu [19,31] have reported the formation ofanatase phase when Ti–6Al–4V alloy is oxidized at 600 °C for 24 and48 h whereas rutile becomes the dominated phase when it is oxidized

sample for various time periods: (a) 8 h, (b) 16 h, (c) 24 h and (d) 48 h.

Table 1Surface microhardness and corrosion parameters of untreated and thermally oxidized CP-Ti samples at 650 °C as a function of treatment time viz., 8, 16, 24 and 48 h.

Treatment condition Surface microhardness (HV0.2) Ecorr (mV vs. SCE) icorr (µA/cm2) ipass (µA/cm2)a Rct (Ω cm2) Cdl (F)

Untreated CP-Ti 178±6 −474 11.7×10−1 23.85 3.1×104 9.9×10−5

CP-Ti thermally oxidized at 650 °C for 8 h 365±9 −296 6.7×10−3 0.012 1.4×106 3.3×10−7

CP-Ti thermally oxidized at 650 °C for 16 h 396±15 −140 1.8×10−3 0.012 4.4×106 3.4×10−9

CP-Ti thermally oxidized at 650 °C for 24 h 443±20 −245 4.9×10−3 0.440 2.0×106 8.3×10−9

CP-Ti thermally oxidized at 650 °C for 48 h 679±43 −87 4.2×10−4 0.016 3.2×107 5.2×10−9

a Measured at 1000 mV vs. SCE.

1946 S. Kumar et al. / Materials Science and Engineering C 29 (2009) 1942–1949

at 650 °C for 48 h. However, no anatase phase is found during the TO ofCP-Ti samples at 650 °C at 8, 16, 24 and 48 h, in the present study.

A comparison of the intensity of α-Ti, TiO and rutile phases insamples oxidized for different durations of time suggests that thethickness of the oxide layer increases with the increase in oxidationtime. The presence of α-Ti and TiO phases for samples oxidized at650 °C for 8 and 16 h indicates the formation of a thin oxide filmwhilethe formation of rutile as the dominant phase for samples oxidized for24 and 48 h suggests the formation of a thicker oxide film. Thepresence of TiO as the strongest peak for samples oxidized for 8 and16 h suggests that rutile formation is minimal under these conditions,which would otherwise lead to the absence of the TiO peaks [32]. Thedependence of the intensity of peaks pertaining to various phases onthe thickness of the oxide layer is represented in Fig. 4. When theoxidation time is less, it is obvious to expect a thinner oxide layerthroughwhich the X-rays can penetrate and reach the substrate (Fig. 4(a) and (b)). However, when the thickness of the oxide layer increaseswith increase in oxidation time, the depth of penetration of the X-raysis mostly covered by the rutile phase and as a consequence thecontribution from the substrate is significantly reduced (Fig. 4(c) and(d)).

3.3. Microhardness of untreated and thermally oxidized samples

The microhardness of untreated and thermally oxidized CP-Tisamples measured on the surface of the oxide layer is given in Table 1.The microhardness of untreated CP-Ti is 178±6 HV0.2. Compared tothe untreated sample there is a significant increase in the hardness ofthermally oxidized samples, due to the formation of hard oxide layersand the strain evolved during the dissolution of oxygen beneath theoxide layer of the substrate [33,34]. Almost a three fold increase inhardness is observed for the sample oxidized for 48 h compared tothat of the untreated CP-Ti sample. The extent of scattering inmicrohardness appears to increase with increase in oxidation time.This may be due to the microstructural inhomogenities generatedduring the growth of oxide islands and their orientation on thesurface. It has been reported elsewhere that thermal oxidation wouldlead to an increase in surface roughness of the oxide film due to thedifferential oxidation rate of individual grains of the polycrystalline

Fig. 5. Schematic representation of hardness measurement of thermally oxidizedsamples (a) for lesser time duration (like 8 h) and (b) for higher time duration (like48 h).

alloy [26,35]. The increase in surface roughness could also be acontributing factor for the observed scattering in the hardness values.

A schematic representation, depicting the relationship betweenthe measurement of microhardness at the surface and the surfaceroughness of the oxide layer for samples oxidized at 650 °C for 8 and48 h, is shown in Fig. 5. When the thickness of the oxide layer isincreased with increase in oxidation time, the surface roughness islikely to increase. Under such conditions, there is a chance that the tipof the indenter might enter in between the grooves of twoneighbouring grains, causing a higher deviation in their microhard-ness. In contrast, when the oxide film is thin and the surfaceroughness is low, the extent of deviation is also less.

3.4. Potentiodynamic polarization studies of untreated and thermallyoxidized samples

The potentiodynamic polarization curves of untreated and ther-mally oxidized CP-Ti samples in Ringer's solution in the potentialrange of −250 mV to +3000 mV vs. SCE with respect to their OCPsare shown in Fig. 6. The corrosion potential (Ecorr), corrosion currentdensity (icorr) and passive current density (ipass) of the untreated andthermally oxidized CP-Ti samples are compiled in Table 1. Comparedto the untreated sample, oxidized samples exhibit a shift in Ecorrtowards the noble direction and a significant decrease in icorr,irrespective of the oxidation time. The improvement in corrosionresistance observed for thermally oxidized samples is believed to bedue to the coverage of the surface of the sample by the oxide scale,which serves as a barrier layer, physically separating the substrate andthe corrosive medium. This attribute is validated by the excellentcorrosion resistance offered by samples oxidized for 48 h (Table 1).Increase in oxidation time from 8 to 16 and 48 h enables an increase incorrosion resistance. However, sample oxidized for 24 h exhibitrelatively lower corrosion resistance. This may be due to the growth ofthe oxide layer, leading to the formation of oxide islands and grain

Fig. 6. Potentiodynamic polarization curves of untreated and thermally oxidizedsamples.

Fig. 7. (a): Nyquist plots of untreated and thermally oxidized samples; (b): Bode impedance plots of untreated and thermally oxidized samples; (c): Bode theta phase angle plots ofuntreated and thermally oxidized samples.

1947S. Kumar et al. / Materials Science and Engineering C 29 (2009) 1942–1949

1948 S. Kumar et al. / Materials Science and Engineering C 29 (2009) 1942–1949

boundaries, which is supported by the morphological features of theoxide scales (Fig. 1(c)).

The anodic branch of the polarization curve exhibits an active–passive transition for untreated as well as thermally oxidized CP-Tisamples. However, the active region of the polarization curves of thethermally oxidized samples is extended towards lower current region,suggesting the ability of these samples to offer a better corrosionresistance. The passive current density (ipass) of all the materials wasmeasured at 1000 mV vs. SCE and the results are compiled in Table 1.Compared to untreated CP-Ti sample, there is a significant decrease inipass of TO samples, irrespective of the oxidation time. The ipass ofsample oxidized for 48 h (0.016 µA/cm2) is about 1500 times lowerwhen compared to that of untreated sample (23.85 µA/cm2). Thevalue of ipass of CP-Ti samples thermally oxidized for 8, 16 and 48 h isquite similar whereas the one oxidized for 24 h exhibits a breakdownof the passive film at around 1650 mV vs. SCE, which might be due tothe breakage of the thin oxide film at the grain boundary of oxideislands. However, the passivity for the sample is re-established withinanother 100 mV in the anodic direction and the passivity is retaineduntil the highest anodic potential range is studied.

The shift in Ecorr towards the noble direction and decrease in icorrand ipass of the thermally oxidized samples suggest an improvementin their corrosion protective ability compared to that of the untreatedsample. Several researchers have also confirmed the fact thatthermally oxidized CP-Ti and Ti–6Al–4V alloy offer a better corrosionresistance in a variety of environments. The ipass of thermally oxidizedCP-Ti sample of the present study, in Ringer's solution, is quite similarto that of Ti–6Al–4V alloy thermally oxidized at 500 and 700 °C for 1 h[29]. Guleryuz and Cimenoglu [19] have reported that Ti–6Al–4V alloythermally oxidized at 600 °C offers excellent corrosion resistance thanthat of untreated alloy when they are subjected to acceleratedcorrosion test in 5 M HCl solution. According to them, no loss inweight is observed for thermally oxidized sample even after 36 h dueto the formation of a thick and stable oxide film on the surface of theTi–6Al–4V alloy sample. Bloyce et al. [24] have compared thecorrosion resistance of untreated, plasma nitrided (PN), thermallyoxidized (TO), palladium-treated thermally oxidized (PTO) CP-Ti in3.5% NaCl solution at room temperature by potentiodynamicpolarization studies. Both TO and PTO CP-Ti samples exhibit a shiftin Ecorr towards the noble direction and a decrease in icorr compared tothat of PN and untreated CP-Ti. They have also evaluated the corrosionresistance of these samples by the accelerated corrosion test thatinvolves immersion in boiling 20% HCl. Both TO and PTO CP-Ti samplesincreased the lifetime by a factor of about 13 and 27, respectively,when compared to that of PN CP-Ti. The results of the present studyfurther confirm the fact that thermally oxidized CP-Ti samples couldoffer excellent corrosion resistance compared to untreated CP-Ti.

3.5. EIS studies of untreated and thermally oxidized samples

The Nyquist plots of untreated and thermally oxidized CP-Tisamples in Ringer's solution, at their respective OCPs, are shown inFig. 7(a). The untreated CP-Ti sample exhibits only a single semicirclein the entire frequency range, samples oxidized for 8 and 16 h exhibit asemicircle in the high frequency region followed by a loop in the lowfrequency region, while samples oxidized for 24 and 48 h exhibit asemicircle in the high frequency region, followed by a Warburgdiffusion tail in the low frequency region. TheWarburg diffusion tail issuppressed in case of sample oxidized for 48 h. There is a significantvariation in the diameter of the semicircle of untreated and thermallyoxidized samples. Since the diameter of the semicircle of untreatedmaterial was very small in comparison to the treated one, individualgraphs is made for better clarity and understanding Fig. 7(a). Theobserved variation in the curves suggests a strong dependence of thecorrosion resistance on the nature and thickness of the oxide layer.

The charge transfer resistance (Rct) and double layer capacitance(Cdl) values, determined after fitting the data using Boukamp software,are compiled in Table 1. The increase in Rct and decrease in Cdl valuessuggest an increase in the corrosion protective ability of thermallyoxidized samples and support the observations made by potentiody-namic polarization studies. TheRct of sample thermally oxidized for 48 his almost 1000 times higher compared to that of the untreated sample,which confirms the ability of thermally oxidized samples to offer anexcellent corrosion resistance. Based on theRct value,which is ameasureof corrosion protective ability, the untreated CP-Ti and those thermallyoxidized for different durations of time can be ranked as follows:

{TO 48 h}N{TO 16 h}N{TO 8 h}≈{TO 24 h}Nuntreated.

The reduction in diameter of the semicircle of CP-Ti sample oxidizedfor 24 h can be explained due to the diffusion of the electrolyte ionsthrough the thin oxide layer along the grain boundary of theagglomerated oxide grains, which is further confirmed by the SEMsurface topography as shown in Fig. 1(c). However, sample oxidized for48 h, the oxides grow, leading the formation of bigger grains along withthe development of thick oxide layer at the grain boundary. Conse-quently, the thick layer of oxide on the surface hinders the diffusion ofelectrolyte ions which is reflected in the EIS curve. A similar type ofresult was also observed in potentiodynamic polarization curve.

In order to get a better insight about the mechanism, Bodeimpedance (Fig. 7(b)) and Bode phase angle (Fig. 7(c)) plots weremade. The Bode impedance plot reveals that the impedance valueincreases with increase in oxidation time from 8 to 48 h (Fig. 7(b))and further confirms the observations made from the Nyquist plot.The Bode impedance plot of the sample oxidized for 24 h exhibithigher impedance at higher frequency level but declines to lowervalue at lower frequency range and ends with the lesser impedancevalue than those oxidized for 16 h and almost equal to the impedancevalue of the sample oxidized for 8 h.

The Bode phase angle plots (Fig. 7(c)) exhibit much variation,which provides an understanding of the corrosion mechanism.Untreated sample exhibits a phase angle maximum of −65° at anintermediate frequency of around 8 Hz, which suddenly declines to−40° at 0.1 Hz. The Bode phase angle plot of the sample oxidized for8 h exhibits a phase angle maximum of −40° at higher frequency(10,000 Hz) that further increases to relatively higher phase anglemaximum of about −80° at 500 Hz, which remains constant infrequency range from 500 Hz to 50 Hz, before it declines to 0°.However, the Bode phase angle plot of the sample oxidized for 16 hexhibits a phase angle maximum (−70°) at higher frequency(10,000 Hz) higher than those treated for 8 h that further increasesimmediately to higher phase angle maximum of about −80° at1300 Hz, which remains constant up to 500 Hz, before it declines to−13°. The higher phase angle maximum observed at 10,000 Hzfrequency suggests the impenetrability of the electrolyte ions throughthe oxide layer present on the surface. The occurrence of a constantphase angle maximum over a wide range of frequency is typical ofpassive surfaces, which indicates a near capacitive behaviour. Such anoccurrence also describes the dielectric properties of the oxide film;the difficulty encountered in charge transfer process and, ascertains agood corrosion protective ability of the oxide film. The phase angleplots of the sample oxidized for 24 and 48 h exhibit a maximum(−85°) at the start frequency of 10,000 Hz and continuously declineto −14° and −4° respectively, at the lowest frequency 0.1 Hz. Thecontinuous decrease of the phase shift in the entire frequency range(10,000 to 0.1 Hz) can be explained using a transmissive boundarymodel, which considers the fact that only the bottom of the hole isconducting whereas its wall is isolating [36].

The occurrence of a single semicircle for untreated sampleindicates the involvement of a single time constant. The occurrenceof a single phase angle maximum in the phase angle plot further

1949S. Kumar et al. / Materials Science and Engineering C 29 (2009) 1942–1949

confirms the same. Garcia-Alonso et al. [29] have also reported a singletime constant for untreated as well as thermally oxidized Ti–6Al–4Valloy at 500 °C for 1 h. The occurrence of two inflections in the phaseangle plots of the CP-Ti samples oxidized for 8 and 16 h indicates theinvolvement of two time constants. The involvement of two timeconstants has also been reported earlier by Garcia-Alonso et al. [29] forTi–6Al–4V alloy oxidized at 700 °C for 1 h. The occurrence of aWarburg diffusion tail in the low frequency region of the Nyquist plotindicates that the corrosion process is diffusion controlled; the loopobserved at the low frequency region of the sample oxidized for 24and 48 h indicates diffusion of electrolyte through the pores in theoxide film. The low frequency dispersions, observed in the phase angleplots of these samples, confirm the involvement of a diffusioncontrolled process and diffusion through pores.

4. Conclusions

The thermal oxidation behaviour of CP-Ti at 650 °C for 8, 16, 24 and48 h was studied in terms of the surface morphology, microhardnessand structural characteristics. The corrosion resistance of thermallyoxidized CP-Ti in Ringer's solution was evaluated by potentiodynamicpolarization and EIS studies and compared with that of the untreatedone. The study leads to the following conclusions:

• The surface morphology of CP-Ti thermally oxidized at 650 °Creveals the presence of oxide scales throughout the surface withoutspallation, irrespective of the treatment time. The mechanism ofoxidation involves nucleation of a thin layer of oxide followed by itsagglomeration and growth to completely cover the surface.

• XRD measurements reveal the formation of rutile, TiO and α-Tiphases on CP-Ti thermally oxidized at 650 °C. The volume fraction ofthe rutile phase increases with increase in oxidation time.

• Almost a three fold increase is observed for samples oxidized for48 h, compared to the untreated one.

• The ipass of the CP-Ti sample oxidized for 48 h (1.6×10−2 µA/cm2) isabout 1500 times lower compared to that of the untreated one(238.5×10−1 µA/cm2), which is also reflected in the significantincrease in the Rct value of this sample (from 3.1×104 to 3.2×107 Ωcm2), suggesting its excellent corrosion resistance.

• Based on the corrosion protective ability, the untreated andthermally oxidized CP-Ti samples can be ranked as follows: {TO48 h}N{TO 16 h}N{TO 8 h}≈{TO 24 h}Nuntreated.

• From corrosion protection point of view, TO of CP-Ti at 650 °C for48 h is a promising surface treatment and it can be a suitablealternative to the untreated CP-Ti as a bio-implant.

Acknowledgement

The authors express their sincere thanks to Prof. S. P. Mehrotra,Director, National Metallurgical Laboratory, Jamshedpur, for his keeninterest and permission to publish this paper.

References

[1] P. Kovacs, J.A. Davidson, in: S.A. Brown, J.E. Lemons (Eds.), Medical Applications ofTitanium and Its Alloys: The Materials and Biological Issues, Chemical andelectrochemical aspects of the biocompatibility of titanium and its alloys, vol.1272,1996, p. 163, ASTM STP.

[2] J.A. Hunt, M. Stoichet, Curr. Opin. Solid State Mater. Sci. 5 (2001) 161.[3] D.F. Williams, in: D.F. Williams (Ed.), Biocompatibility of Clinical Implant

Materials, Titanium and titanium alloys, vol. II, CRC Press, 1981, p. 9.[4] M.J. Long, H.J. Rack, Biomaterials 19 (1998) 1621.[5] Y. Mu, T. Kobayashi, M. Sumita, A. Yamamoto, T. Hanawa, J. Biomed. Mater. Res. 49

(2000) 238.[6] P.A. Lilley, P.S. Walker, G.W. Blunn, Transactions of the 4th Word Biomaterials

Congress, Berlin, April 24–28, 1992, p. 227.[7] R. Thull, M. Schaldach, Springer-Verlag, New York, 1976, p. 242, /Heidelberg/

Berlin.[8] S.A. Brown, P.J. Hughes, K. Merrit, J. Orthop. Res. 6 (1988) 572.[9] D.W. Hoeppner, V. Chandrasekaran, Wear 173 (1994) 189.[10] L.M. Rabbe, J. Rieu, A. Lopez, P. Combrade, Clin. Mater. 15 (1994) 221.[11] C. Sittig, M. Textor, N.D. Spencer, M. Wieland, P.H. Vallotton, J. Mater. Sci., Mater.

Med. 10 (1) (1999) 35.[12] H.B. Wen, J.G. Wolke, J.R. Wijn, Q. Liu, F.Z. Cui, K. de Groot, Biomaterials 18 (1997)

1471.[13] X. Nie, A. Leyland, A. Matthews, Surf. Coat. Technol. 149 (2002) 245.[14] T. Brendel, A. Engel, C. Russel, J. Mater. Sci., Mater. Med. 3 (3) (1992) 175.[15] P. Andreazza, M.I. De Barros, C. Andreazza-Vignolle, D. Rats, L. Vandenbulcke, Thin

Solid Films 319 (1998) 62.[16] M. Uchida, N. Nihira, A. Mitsuo, K. Toyoda, K. Kubota, T. Aizawa, Surf. Coat. Technol.

177 –178 (2004) 627.[17] S.W.K. Kweh, K.A. Khor, P. Cheang, Biomaterials 23 (2002) 775.[18] T. Hanawa, Y. Kamiura, S. Yamamoto, T. Kohgo, A. Amemiya, H. Ukai, K. Murakami,

K. Asaoka, J. Biomed. Mater. Res. 36 (1) (1997) 131.[19] H. Guleryuz, H. Cimenoglu, Biomaterials 25 (2004) 3325.[20] M.F. Lopez, J.A. Jimenez, A. Gutierrez, Electrochim. Acta 48 (2003) 1395.[21] D. Velten, V. Biehl, F. Aubertin, B. Valeske, W. Possart, J. Breme, J. Biomed. Mater.

Res. 59 (2002) 18.[22] M.L. Escudero, J.L. Gonzalez-Carrasco, C. Garcıa-Alonso, E. Ramırez, Biomaterials 16

(1995) 735.[23] D. Siva Rama Krishna, Y.L. Brama, Y. Sun, Tribol. Int. 40 (2007) 329.[24] A. Bloyce, P.-Y. Qi, H. Dong, T. Bell, Surf. Coat. Technol. 107 (1998) 125.[25] H. Dong, A. Bloyce, P.H. Morton, T. Bell. Titanium'95 Science and Technology; 1995.

p. 1999.[26] P.A. Dearnley, K.L. Dahm, H. Çimenoglu, Wear 256 (2004) 469.[27] G.P. Burns, J. Appl. Phys. 65 (5) (1989) 2095.[28] C. Boettcher, Surf. Eng. 16 (2) (2000) 148.[29] M.C. Garcia-Alonso, L. Saldana, G. Valles, J.L. Gonzalez-Carrasco, J. Gonzalez-

Cabrero, M.E. Martinez, E. Gil-Garay, L. Munuera, Biomaterials 24 (2003) 19.[30] R.W. Schutz, D.E. Thomas, Corrosion of Titanium and Titanium Alloys, ASM

Handbook, Corrosion, Vol.13, ASM International, 1987, pp. 671.[31] H. Guleryuz, H. Cimenoglu, Surf. Coat. Technol. 192 (2005) 164.[32] C.C. Ting, S.Y. Chen, D.M. Liu, Thin Solid Films 402 (2002) 290.[33] S. Weissman, A. Shrier, in: R.I. Jaffee, N.E. Promisel (Eds.), The Science, Technology

and Application of Titanium, Pergamon Press, London, 1968, p. 441.[34] F. Borgioli, E. Galvanetto, F.P. Galliano, T. Bacci, Surf. Coat. Technol. 141 (2001) 103.[35] P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, Essex, 1988.[36] Yves Mueller, Roger Tognini, Joerg Mayer, Sannakaisa Virtanen, J. Biomed. Mater.

Res. 82A (2007) 936.