Surface treatment and its effect on the electrochemical behavior of Ti–15Mo–3Nb–3Al alloy

Abstract

The Ti–15Mo–3Nb–3Al alloy was subjected to four different surface treatments involving alkaline and acidic solutions. HNO₃, HF–HNO₃, NaOH etching and cathodic cleaning using NaOH were performed. The effect of these etchants on the surface morphology and the surface chemical composition was studied. Energy dispersive spectroscopy (EDS) and X-ray photoelectron spectroscopy (XPS) studies revealed the loss of oxygen content on the surface which indicated the partial breakdown of the surface oxide layer. Potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) curves showed the poor protection of the treated alloys against corrosion in 3.5% NaCl. Furthermore, these surfaces were coated with a silica–titania hybrid sol–gel film and the corrosion protection of the coated surfaces was assessed and compared. The results showed that sol–gel coating restored the corrosion resistance of all the treated surfaces.

---

1. Introduction

Titanium and its alloys are used in a wide range of applications from the biomedical field to aerospace industries because of their unique properties such as high tensile strength and toughness, light weight, good corrosion resistance, excellent biocompatibility and ability to withstand high temperatures. The aerospace industry is the largest market for titanium products primarily due to these exceptional properties. They are mostly utilized in jet engines, airframe components and other critical structure parts.¹ Their basic advantages like high reliability during performance and good corrosion resistance make them a top choice for use in engines and airframes. However, a thorough surface modification process is needed prior to their usage. Surface modification process produces required specific surface topographies, cleans and roughens the surface in addition to enhancing several surface properties like adhesion in bonding, corrosion resistance, wear resistance etc.² The prominent surface modification methods are mechanical, chemical, sol–gel, anodic oxidation, nitriding and ion implantation techniques.^2,3 In recent years, different surface treatments of various Ti alloys and their effects on surface morphologies and corrosion behaviors in different solutions have been studied.^2–16 Gostin et al. have shown that etching treatments of Ti–45Nb alloy in strongly oxidizing acidic solution like HF : HNO₃ (4 : 1) is found to be very effective for generating a nanopatterned surface topography which is expected to be promising for the stimulation of bone tissue growth.⁸ Acid pickling is a popular surface treatment method followed to clean substrate surface by removing oxides and contamination. A mixture of sulfuric acid and hydrogen peroxide has been used to eliminate surface contaminants and reported to produce a consistent and reproducible titanium oxide surface layer.⁹ Takeuchi and coworkers have studied etching efficiencies of Ti surface using acids such as Na₂S₂O₈, H₂SO₄ and HCl.¹⁰ They have found that HCl is the most effective etching agent among the three because of its capability to dissolve titanium salts easily without weakening Ti surfaces. Alkali treatment by alkali solutions such as NaOH or KOH forms a bioactive porous layer on the substrate materials. Immersion of titanium alloy in a 5–10 M NaOH or KOH solution for 24 h is reported to form a porous layer of sodium titanate on the surface which has aided the formation of bone-like apatite layer.¹⁶

Sol–gel process is one of the popular surface modification methods carried out by depositing thin coatings on various substrates including titanium and its alloys. This technique is capable of producing coatings with controlled composition, microstructure and better chemical homogeneity. Low densification temperature and cost effectiveness are added advantages of this process. This method of surface modification has also been adopted by few researchers on titanium alloys.^17–20 In the work of Liu et al., it has been demonstrated that glycidoxypropyl-trimethoxy-silane (GTMS) molecules are preferred to be adsorbed on β phase particles of Ti–6.5Al–1Mo–1V–2Zr alloy than α phase, especially at lower concentrations of GTMS which is due to the different chemical compositions and variation of the resulting surface hydroxylation in the studied solution.²⁰ However, there is no study reported on the effect of surface modification from the corrosion aspect for titanium β–21S (Ti–15Mo–3Nb–3Al) alloy. Further, little is known about the passivity and corrosion resistance of this alloy in marine environment.²¹ In the present work, the effect of different alkali and acid treatments on the morphology and surface composition of uncoated alloy is presented by surface sensitive probes like 3D-profilometry, field emission scanning electron microscopy (FESEM) and X-ray photoelectron spectroscopy (XPS). The electrochemical behavior of the surface is investigated by potentiodynamic polarization and impedance measurements in 0.6 M NaCl medium. Coated alloy has been characterized by Fourier transform infrared spectroscopy (FTIR) and Raman spectroscopy. Further, the effect of sol–gel coating in obviating the effects of surface treatments on corrosion resistance is presented.

2. Experimental details

2.1. Materials

Titanium β–21S (Ti–15Mo–3Nb–3Al–0.2Si) alloy was cut into small pieces of 1.5 × 1.5 × 0.2 cm³ size. The samples were smoothened by grinding with silicon carbide emery papers of different grit sizes (800, 1000 and 1200) and then polished to mirror-like finish using 0.03 μm alumina powder mixed with distilled water. The polished samples were ultrasonically cleaned in acetone for 15 min before any further use. Sol–gel precursors namely 3-glycidoxypropyltrimethoxysilane (GPTMS) and titanium isopropoxide (TIP) were supplied by Sigma-Aldrich and were used as-received. Acetylacetone, used as a complexing agent for TIP was also from Sigma-Aldrich. Ethanol (anhydrous, 99.5%) was used as the solvent and was from Changshu Yangyuan Chemicals, China. Nitric acid (HNO₃, ∼70%) and sodium hydroxide (NaOH) were procured from Merck. Hydrofluoric acid (48%) was received from Sigma-Aldrich. Milli Q ultra pure water was used for the preparation of solutions.

2.2. Surface treatment

The polished samples were chemically treated with different acidic and alkaline etchants prior to sol–gel coating deposition. Solutions of 20% HNO₃, 0.5 M NaOH and HF–HNO₃ mixture in 2 : 10 volume with 40 mL water were prepared in Milli Q water and used as etchants. Acid pickled surfaces were obtained by immersing the substrates in HNO₃ for 10 min. Similarly, alkaline treated surfaces were prepared by NaOH solution immersion for 10 min. Etching in HF–HNO₃ mixture was limited to only 2 min. Cathodic cleaning was conducted by immersing the sample in 10% NaOH solution under an external bias of 2 V for 60 s. The sample acted as the cathode and a stainless steel strip was used as the anode. Then it was immersed in 10% H₂SO₄ for 30 s at ambient conditions. The treated surfaces were thoroughly washed with Milli Q water. They were then room temperature dried for about 15 min and then subjected to coating deposition.

2.3. Sol synthesis and deposition of coatings

GPTMS–TIP hybrid sol was synthesized by mixing two separately prepared precursor sols. GPTMS was magnetically stirred with ethanol in the presence of 0.36 mL 0.2 M HNO₃ for 1 h. The molar ratio of silane : ethanol : water was 1 : 3 : 2. TIP sol was prepared as the inorganic part by mixing TIP with acetylacetone, ethanol and water. TIP was first stirred with acetylacetone for 15 min in the molar ratio of 1 : 1 prior to the addition of solvent. HNO₃ was again added as the catalyst to promote hydrolysis. The molar ratio of GPTMS to TIP was 2 : 1. GPTMS sol was a clear colorless liquid while the TIP sol was a clear pale yellow liquid. Finally, after completing the hydrolysis, the two sols were mixed together and diluted with equal volume of ethanol. The final mixture was further stirred for 2 h for complete homogenization.

The prepared GPTMS–TIP hybrid sol was coated on to the treated substrates by dip-coating method (single dip coater, model SDC-2007C from Apex Instruments Co. Pvt. Ltd.). The process was carried out at a withdrawal speed of 90 mm min⁻¹ after a residence time of 60 s in the sol. After coating, the samples were first dried at room temperature for 24 h and were subsequently heat-treated at 100 °C for 2 h in an air oven.

2.4. Surface characterization

The surface morphologies of the samples were examined by Carl Zeiss Supra 40 VP FESEM. The 3D surface images and 2D line profiles were obtained by NanoMap 500 LS 3D surface profilometer. Compositions of the samples were obtained by energy dispersive spectroscopy (EDS) using Oxford instruments Inca Penta FET X3 (Oxford Instruments, United Kingdom) attached to the FESEM.

Structural characterization of the coating was studied by FTIR and Raman spectroscopy. FTIR spectrum was recorded using Bruker Alpha-P spectrometer. Raman data were obtained using Labram 010 Model of DILOR-JOBIN-YVON-SPEX Micro Raman spectrometer with 632 nm laser.

XPS of bare, chemically etched and sol–gel coated Ti alloys were recorded with a SPECS spectrometer, Germany using non-monochromatic AlKα radiation (1486.6 eV) as X-ray source operated at 150 W (12 kV, 12.5 mA). The binding energies reported here were referenced with C 1s peak at 284.6 eV with a precision of ±0.1 eV. All the survey spectra were obtained with a pass energy of 70 eV with step increment of 0.5 eV and individual spectra were recorded with a pass energy and step increment of 40 and 0.05 eV, respectively. For XPS analysis, samples were mounted on the sample holders and placed into a load-lock chamber with an ultrahigh vacuum (UHV) of 8 × 10⁻⁸ mbar for 5 h in order to desorb any volatile species present on the surface. After 5 h, samples were transferred into the analyzing chamber with UHV of 5 × 10⁻¹⁰ mbar. Ti 2p, Mo 3d and O 1s core level spectra were curve-fitted into their possible components using Gaussian–Lorentzian peaks after subtracting a Shirley background with CasaXPS program.

2.5. Electrochemical measurements

Corrosion studies on the substrate and coated samples were carried out using CH 604D electrochemical workstation supplied by CH instruments, USA. A conventional three electrode system was used to carry out the potentiodynamic polarization and EIS studies. The test sample was kept as the working electrode; Pt foil and saturated calomel electrode (SCE) connected to a Luggin capillary were used as counter and reference electrodes, respectively. NaCl solution of 3.5% was used as a corrosive medium to test the samples. The reference electrode was kept very close to the surface of the working electrode. The sample was immersed in NaCl solution till the open circuit potential (E_OCP) is established (∼1 h). EIS measurement was carried out in the frequency range of 10 mHz to 100 kHz by applying a 10 mV sinusoidal wave. The impedance data were displayed as Nyquist and Bode plots. The acquired data were curve-fitted and analyzed using ZSimpwin program (Princeton Applied Research, USA) to get suitable equivalent circuit parameters. The quality of the fit was verified by the χ² value. After EIS measurements, the system was allowed to attain OCP and subsequently polarization studies were carried out in the potential range of E_OCP ±200 mV with a sweep rate of 1 mV s⁻¹. The obtained potentiodynamic polarization plot (Tafel plot) was represented as potential vs. log i. The corrosion potential (E_corr) and corrosion current (i_corr) were deduced from the Tafel plot.

3. Results and discussion

3.1. Microstructural studies

The surface microstructural features of the uncoated and coated samples were studied using 3D profilometer and FESEM. The 3D profiles of the uncoated surfaces after different chemical treatments are shown in Fig. 1. Samples in the as-polished condition with only acetone cleaning are used for comparison. The average roughness of the as-polished samples is 75 ± 10 nm. NaOH treated and cathodically cleaned surfaces show a roughness of about 86 nm and 66 nm, respectively, while a roughness of about 70 nm is observed in HNO₃ treated surface. In the case of HF–HNO₃ treated surface, the surface roughness increases to 208 nm. It is very evident from the profiles that surface treatment with NaOH and HNO₃ including cathodic cleaning do not affect the surface roughness much. However, the highest roughness is seen for the HF–HNO₃ treated surface which is increased almost by 2.7 times. The 3D image also shows that the surface morphology is severely changed and the grain structure is clearly seen. Thus, HF–HNO₃ etching has substantially removed the top layers of the alloy including the protective oxide film.

Fig. 1 3D surface profiles of Ti–15Mo–3Nb–3Al alloys after different chemical treatments.

FESEM images of the acetone degreased and HF–HNO₃ etched surfaces are shown in Fig. 2. As can be seen from the figures, acetone degreased sample shows few striations on the surface which must have resulted during surface grinding and polishing (Fig. 2a). The FESEM image of the HF–HNO₃ etched surface at lower magnification (Fig. 2c) shows the grain structure. At higher magnification (Fig. 2b) lots of craters on the surface are seen which might be due to severe dissolution of the alloy in the strong etchant (Fig. 2b and c). Micrographs of the NaOH, HNO₃ and cathodically cleaned surfaces (not shown here) appear very similar to acetone cleaned sample in their surface features. Thus, FESEM studies also support the 3D profilometer results that HF–HNO₃ treatment leads to severe damage of the substrate surface. Fig. 2d shows a representative image of the sol–gel coating on the substrate in the as-prepared condition. The coating of about 4 μm thick covers all the surface features of the chemically treated alloy samples. Since all the samples exhibit smooth and uniform surface, only one image is shown for representation.

Fig. 2 FESEM images of Ti–15Mo–3Nb–3Al samples; (a) acetone degreased, (b) HF–HNO₃ etched surface, (c) HF–HNO₃ etched surface showing the grain boundaries and (d) sol–gel coated surface in the as-prepared condition and (e) EDS of the GPTMS–TIP sol–gel coating.

3.2. EDS studies

Compositional analysis of the substrate in the as-received condition as well as after chemical treatment has been carried out by EDS. Compositional analysis of the coating has also been done. It is to be noted that thickness of the Ti alloy substrate is 1.5 mm and that of the coating is 4 μm. Energy of 10 keV is used for EDS experiment and penetration depth of electrons with this energy is less than 1 μm. Therefore, sampling depth for EDS experiment is much less than sample and coating thickness. The substrate shows Al, Si, Nb and Mo as the alloying elements with their wt% of 2.47, 0.58, 3.19 and 15.74, respectively and the balance being Ti. The alloy also shows an oxygen concentration of 4.7 wt%. The compositions of the treated substrates are as shown in Table 1. The concentrations of Al and Nb are varied marginally on different treatments. There is a slight enrichment in the Mo content. The concentration of oxygen on the samples decreases on the subjection to chemical treatments. It is about 2.5 wt% for the HNO₃ and cathodically cleaned surface and only about 1 wt% for NaOH treated surface. HF–HNO₃ shows no traces of oxygen on the surface. ED spectrum for the sol–gel coated sample is shown in Fig. 2e. The concentration of Si and Ti are 9.5 and 4.8 at% which reaffirms the 2 : 1 molar ratio of GPTMS and TIP.

Table 1 Elemental compositions evaluated from EDS analysis in Ti–15Mo–3Nb–3Al alloys treated with different conditions

Alloys	Composition (wt%)
	Al	Si	Nb	Mo	O	V	Ti
Acetone degreased	2.47	0.58	3.19	15.74	4.73	—	73.29
NaOH treated	2.65	—	2.66	17.19	0.98	0.46	76.06
HNO3 treated	2.71	—	3.69	17.67	2.43	0.36	73.15
HF–HNO3 treated	2.70	—	3.94	16.99	—	—	76.37
Cathodically cleaned	2.57	0.81	2.84	16.16	2.58	—	75.04

---

3.3. FTIR and Raman studies

The chemical characterization of coating has further been performed by vibrational spectroscopy techniques, namely FTIR and Raman. The corresponding spectra of coated alloy are shown in Fig. 3a and b, respectively. In the FTIR spectrum (Fig. 3a), a broad peak at 3470 cm⁻¹ and small peak at 1629 cm⁻¹ correspond to O–H bonds. The vibrational band around 2850 cm⁻¹ is attributed to the C–H vibrations in the glycidoxypropyl group of GPTMS and acetylacetone.^22,23 C=O stretching vibrations of acetylacetone can occur at 1630 cm⁻¹. Characteristic band for epoxide ring vibrations of GPTMS is seen at 1257 cm⁻¹. Strong bands observed at 1100–1130 cm⁻¹ are characteristic of Si–O–Si moieties.²⁴ At lower wavenumbers, asymmetric stretching vibration of Si–O–Si is seen at around 900 cm⁻¹. The data, thus suggest the formation of a hybrid network between GPTMS and TIP.

Fig. 3 (a) FTIR and (b) Raman spectra of GPTMS–TIP coated surface.

In the Raman spectrum (Fig. 3b), the strong peak at 456 cm⁻¹ originates from –(Si–O–Si) bending. The weak signal seen in the spectrum at 644 cm⁻¹ is due to symmetric stretching of –(Si(–O–CH₃)₃) moieties. This shift is characteristic of partially unhydrolyzed trimethoxysilane molecules of GPTMS. The characteristic Raman shift of GPTMS is located at 1264 cm⁻¹ and is assigned to epoxy ring breathing. The peak at 1293 cm⁻¹ is due to (–(CH₂)_n–) wagging. The intensity at 1455 cm⁻¹ is assigned for asymmetric bending modes of δ(CH₃), δ(CH₂) or δ(CH) groups. The vibrations at 1033, 1058 and 1409 cm⁻¹ are related to the C–C skeletal vibration or ν(Si–O) stretching in inorganic network.²⁵ The vibration at 848 cm⁻¹ corresponds to oxirane symmetric deformation.²⁶ The asymmetric stretching of Si–O–Si linkage occurs at 1196 cm⁻¹. A Raman shift at 980 cm⁻¹ is of the Si–O stretching vibration of silanol (Si–OH) groups.

3.4. XPS studies

Detailed XPS characterization has been carried out to understand the surface nature of Ti alloys with different conditions. Typical XPS survey spectra of Ti alloys with different etching conditions are given in Fig. 4. They clearly show the presence of Ti, Mo and O species in all the alloys. Ti, O and Si are observed to be present in sol–gel coated Ti alloy. In Table 2, surface concentrations of Ti, Mo and Nb are given. It has been observed that Mo amount increases in the alloys with etching treatment in HF–HNO₃ and in cathodically cleaned alloy in comparison with bare substrate alloy. However, there is not much change in Mo amounts in NaOH and HNO₃ treated alloys. Ti 2p core level spectra of bare Ti substrate and etched alloys have been recorded and are displayed in Fig. 5. Bare Ti substrate shows Ti 2p_3/2,1/2 peaks at 458.6 and 464.2 eV that correspond to Ti⁴⁺ species.²⁷ Ti is observed to be present in +4 oxidation state in the NaOH, HF–HNO₃ and HF etched and cathodically cleaned alloys. There is a small component peak around 457.5 eV that related to Ti³⁺ species in all alloys. Mo 3d core level spectra of the treated alloys are shown in Fig. 6. Spectral envelops of Mo 3d core levels are broad in nature indicating that Mo is present in different oxidation states and it can be curve-fitted into sets of spin–orbit doublets. Typical curve-fitted Mo 3d core level spectrum of the alloy etched with HF–HNO₃ is given in Fig. 7. Observed Mo 3d_5/2 peaks at 227.6, 229.5, 231.1 and 232.5 eV in all alloys are assigned to Mo⁰, Mo⁴⁺, Mo⁵⁺ and Mo⁶⁺ species, respectively.^28,29 However, amount of metal is found to be less in cathodically cleaned and HNO₃ treated alloys compared to bare substrate alloy, whereas metal concentration is more in HF–HNO₃ etched alloys. Nb is present in +5 oxidation state in all alloys. Binding energies, surface concentrations of component species of Ti and Mo in the treated alloys evaluated from XPS studies are summarized in Table 3. XPS of O 1s core levels of alloys treated with different conditions are broad in nature that can be curve-fitted into several component peaks. Main peak around 530.2 eV is attributed to oxide species present in the alloys and the observed peaks around 531.8 and 533.5 eV correspond to oxygen associated with hydroxyl (OH⁻) species and H₂O.³⁰ A typical curve-fitted O 1s core level spectrum of the alloy etched with HNO₃ is shown in Fig. 8.

Fig. 4 XPS survey spectra of Ti–15Mo–3Nb–3Al alloys with different conditions.

Table 2 Surface concentrations (at%) of Ti, Mo and Nb evaluated from XPS studies in Ti–15Mo–3Nb–3Al alloys treated with different conditions

Alloys	Ti	Mo	Nb
Acetone degreased	87.1	10.9	2.0
NaOH treated	88.4	9.4	2.2
HNO3 treated	87.6	10.3	2.1
HF–HNO3 treated	74.9	22.3	2.8
Cathodically cleaned	84.5	13.3	2.2

---

Fig. 5 Ti 2p core level spectra of Ti–15Mo–3Nb–3Al alloys with different conditions.

Fig. 6 Mo 3d core level spectra of Ti–15Mo–3Nb–3Al alloys with different conditions.

Fig. 7 Curve-fitted Mo 3d core level spectrum of Ti–15Mo–3Nb–3Al alloy surface treated with HF–HNO₃.

Table 3 Binding energies and relative peak areas of different Ti and Mo species in Ti–15Mo–3Nb–3Al alloys with different etching conditions

Alloys	Ti species	Binding energy of Ti 2p3/2 (eV)	Relative peak area (%)	Mo species	Binding energy of Mo 3d5/2 (eV)	Relative peak area (%)
Bare alloy	Ti^3+	457.6	8	Mo^0	227.4	11
	Ti^4+	458.8	92	Mo^4+	229.3	33
				Mo^5+	231.1	28
				Mo^6+	232.4	28
NaOH treated	Ti^3+	457.5	12	Mo^0	227.7	13
	Ti^4+	458.5	88	Mo^4+	229.4	46
				Mo^5+	231.5	19
				Mo^6+	232.7	22
HNO3 treated	Ti^3+	457.7	10	Mo^0	227.6	8
	Ti^4+	458.5	90	Mo^4+	229.5	30
				Mo^5+	231.1	30
				Mo^6+	232.7	32
HF–HNO3 treated	Ti^3+	457.6	13	Mo^0	227.6	15
	Ti^4+	458.6	87	Mo^4+	229.5	28
				Mo^5+	231.1	30
				Mo^6+	232.7	27
Cathodically cleaned	Ti^3+	457.7	8	Mo^0	227.6	8
	Ti^4+	458.8	92	Mo^4+	229.2	15
				Mo^5+	231.0	25
				Mo^6+	232.5	52

---

Fig. 8 Curve-fitted O 1s core level spectrum of Ti–15Mo–3Nb–3Al alloy surface treated with HNO₃.

3.5. Electrochemical studies

3.5.1. Potentiodynamic polarization studies. Potentiodynamic polarization curves of the uncoated and sol–gel coated samples with different surface treatments obtained in 3.5% NaCl solution are shown in Fig. 9a and b, respectively. The corrosion current density (i_corr) has been obtained by extrapolation of the anodic and cathodic branches of the polarization curves and the corresponding Tafel parameters are tabulated in Table 4. The i_corr value obtained for the acetone degreased uncoated sample is 0.04 μA cm⁻² (Fig. 9a). From, the table it is seen that i_corr values increase by almost 2 orders for all the treated samples in comparison with degreased uncoated sample. Thus, corrosion resistance deteriorates drastically on subjection to different chemical treatments. The shift in the E_corr of all treated samples to more negative values compared to acetone degreased surface (−0.4 V from −0.3 V) also suggests the surfaces are susceptible to corrosion. This behavior is due to the partial dissolution or breakdown of the oxide layer on the surface because of interaction with the etchants. The HNO₃ treated surface however, shows an intermediate level of performance with reduction in i_corr by only 1 order of magnitude (0.25 μA cm⁻²). Thus, the treated surfaces show poor corrosion protection.

Fig. 9 Potentiodynamic polarization curves of (a) chemically treated uncoated substrates and (b) sol–gel coated Ti–15Mo–3Nb–3Al alloys in 3.5% NaCl solution.

Table 4 Potentiodynamic polarization data of uncoated and coated samples

Alloys	Uncoated samples		Coated samples
	icorr (μA cm^−2)	Ecorr (V)	icorr (μA cm^−2)	Ecorr (V)
Acetone degreased	0.04	−0.3	0.06	−0.22
NaOH treated	1.2	−0.33	0.08	−0.25
HNO3 treated	0.25	−0.43	0.08	−0.19
HF–HNO3 treated	2	−0.41	2.5	−0.29
Cathodically cleaned	1.6	−0.35	0.02	−0.25

---

Sol–gel coating on acetone degreased surface shows i_corr value of 0.06 μA cm⁻² (Fig. 9b) which is almost same as that in uncoated condition. However, it is worth to note from the figure that the anodic curve for the coated sample is smoother and horizontal compared to the uncoated substrate. This is a clear indication that the passive behavior of the surface is indeed enhanced by the sol–gel coating thereby acting like an additional barrier layer. Surfaces with different chemical treatments also exhibit very low i_corr values (0.02–0.08 μA cm⁻²) on sol–gel coating. Thus, the coating provides a barrier protection to the surface against the ingress of electrolyte and prevents the attack by the chloride ions. The E_corr values for the coated samples fall in a very narrow range of −0.19 to −0.25 V and this further confirms the protection rendered by the coating. But, HF–HNO₃ treated surface does not show any major improvement in spite of sol–gel coating. The defects caused by surface etching can be a reason for this which is understood from their microstructural features and surface composition as discussed previously.

3.5.2. EIS studies. Tafel results are further authenticated by EIS data. EIS allows the separation of responses of various components of a system such as resistances and capacitances of the protective layers and oxide films. The evolution of these parameters can be used to understand the behavior of the coating system. The experiments were performed at the open circuit potential and results are shown in Fig. 10 and 11. They are displayed in both complex impedance (Nyquist diagram) and Bode plots (impedance modulus and phase angle). Since Nyquist plots are relatively featureless, Bode plots are often preferred for indicating the changes in the electrochemical characteristics of the system. The impedance spectra of the chemically treated uncoated surfaces are displayed in Fig. 10 and the spectra of chemically treated and sol–gel coated samples are depicted in Fig. 11. Both the systems are found to exhibit two time constants. They can be assigned to two distinct frequency regions as follows: the time constant resolved in the high-frequency region can be attributed to the impedance characteristics obtained due to the penetration of the electrolyte through a porous film. This describes the properties of the outer porous layer. The time constant at the low-frequency regime accounts for the processes taking place at the substrate/electrolyte interface situated at the base of the porous layer. Such a behavior is typical of a metallic material covered by a porous film which is exposed to an electrolytic environment. In the case of uncoated samples, the porous layer is the treated surface covered with a thin naturally formed oxide layer on the surface. For the coated samples, the porous layer is the sol–gel coating applied on them, which by default contains some pores.

Fig. 10 Electrochemical impedance curves: (a) Bode plot and (b) phase angle plot of chemically treated uncoated Ti–15Mo–3Nb–3Al alloys in 3.5% NaCl solution.

Fig. 11 Electrochemical impedance curves: (a) Bode plot and (b) phase angle plot of sol–gel coated Ti–15Mo–3Nb–3Al alloys in 3.5% NaCl solution.

The corrosion protection efficiency of a system is usually assessed from the impedance values obtained at the low frequency region; higher the value better is the protection. From Fig. 10, acetone degreased substrate shows the highest value of impedance compared to all other chemically treated surfaces. This increased protection is definitely from the naturally formed oxide layer which is intact. In all other cases, the drop in the impedance value indicates the partial breakdown of the protective layer and initiation of corrosion activity at the metal surface. A broad phase angle peak seen for the acetone degreased surface is indicative of the capacitive behavior of protective oxide layer and its distributed pore sizes. Reduction in the peak width, observed for the chemically treated surfaces at ∼11 Hz can imply thinning of the oxide layer. Decrease in the peak height is an indication of deterioration in capacitive behavior of the oxide layer. A second low frequency time constant which is an indication of corrosion attack is well defined for the HF–HNO₃ treated surface. This is because, HF being a strong etchant for the alloy, has destroyed the oxide layer almost completely and thus the surface is immediately attacked in the NaCl solution. Whereas in the other cases, it is observed that although the drop in the impedance modulus is to the same extent, the time constant corresponding to the corrosion activity is not so well defined. Thus, the plots confirm that all the used chemical treatments partially destroy the native oxide film on the metal surface and hence affects it inherent corrosion resistant property.

For the sol–gel coated samples, the variation in the impedance values at low frequency region is not very significant (Fig. 11). This suggests that the sol–gel layer offers uniform improved protection to the surface. In fact, sol–gel coating on HNO₃ treated surface gives the maximum protection to the surface with an impedance modulus of about 1 × 10⁶ Ω cm². However, sol–gel coating on HF–HNO₃ treated surface shows the lowest protection with a value of almost one order less. There is not much difference between the performances of other three surfaces except some minimal changes. The EIS curves obtained for the sol–gel uncoated and coated samples have been suitably fitted using equivalent circuit shown in Fig. 12 and the results obtained are summarized in Tables 5 and 6, respectively. The fitted curves are also shown in the respective impedance plots. Good conformity has been obtained between the fitted and experimental data in all the cases and quality of the fit has been checked by the χ² value. The non-ideal capacitive response of the oxide films has been entailed by using a constant phase element (CPE) instead of a pure capacitance in the fitting procedure. This CPE can be due to difference in the relaxation times as a result of different degrees of surface in-homogeneity, roughness factors and compositions of surface layers. The impedance with the capacitances can be defined as Z_CPE = 1/Q(jω)ⁿ, where Q, j, ω and n are the pseudo capacitance, imaginary function (√−1), angular frequency and the deviation from the ideal behavior of a pure capacitor, respectively. When n = 1, the system behaves like a pure capacitor and Q = C where C is capacitance.

Fig. 12 Electrochemical equivalent circuits used to fit the EIS data of Ti–15Mo–3Nb–3Al alloys with different conditions.

Table 5 EIS data of uncoated samples

Alloys	Qcoat (S s^n cm^−1)	ncoat	Rcoat (Ω cm^2)	Qdl (S s^n cm^−1)	ndl	Rct (Ω cm^2)
Acetone degreased	2.2 × 10^−5	0.94	11 700	1.54 × 10^−5	0.88	2.7 × 10^6
NaOH treated	2.9 × 10^−5	0.92	6922	22 × 10^−5	0.57	2.5 × 10^4
HNO3 treated	2.6 × 10^−5	0.94	7189	17 × 10^−5	0.83	6.6 × 10^5
HF–HNO3 treated	4.1 × 10^−5	0.89	5593	28 × 10^−5	0.8	3.6 × 10^4
Cathodically cleaned	1.8 × 10^−5	0.91	8300	0.67 × 10^−5	0.73	8.7 × 10^4

---

Table 6 EIS data of coated samples

Alloys	Qcoat (S s^n cm^−1)	ncoat	Rcoat (Ω cm^2)	Qox (S s^n cm^−1)	nox	Rox (Ω cm^2)
Acetone degreased	2.72 × 10^−9	0.96	332.8	2.1 × 10^−5	0.91	1.54 × 10^7
NaOH treated	2.30 × 10^−8	0.83	169.0	4.07 × 10^−5	0.92	5.1 × 10^6
HNO3 treated	3.40 × 10^−9	0.98	111.2	1.75 × 10^−5	0.91	1.65 × 10^7
HF–HNO3 treated	4.20 × 10^−6	0.63	46.6	3.67 × 10^−5	0.85	1.59 × 10^5
Cathodically cleaned	1.3 × 10^−5	0.92	107.0	1.95 × 10^−5	0.90	1.45 × 10^7

---

The equivalent circuit shown in Fig. 12a describes the surface of uncoated chemically treated samples and the obtained parameters are summarized in Table 5. Here, R_s is the solution resistance, R_O and Q_O are the resistance and pseudo capacitance associated with the penetration of electrolyte through the porous oxide layer. Similarly, R_ct and Q_dl represent the resistance and pseudo capacitance of the charge transfer reactions happening at the substrate/electrolyte interface. The value of R_O for the as-received acetone degreased sample is 1.2 × 10⁴ Ω cm². The alkaline and HNO₃ treated samples have the porous oxide layer showing resistance values relatively low in the range 7–8 kΩ cm². The HF–HNO₃ treated sample shows the lowest resistance of 5593 Ω cm², against the penetration of electrolyte. This suggests that HF–HNO₃ etchant damages the outer porous protective layer (reduces the thickness) and permits the ingress of electrolyte. R_ct is also seen to be highest for the acetone cleaned sample compared to the other samples.

The impedance data of the coated samples are fitted with the circuit shown in Fig. 12b. The obtained parameters are tabulated in Table 6. In the circuit, R_O and Q_O correspond to the resistance and pseudo capacitance of the porous sol–gel layer and R_in and Q_in to that of the inner dense oxide layer. The dense oxide layer is formed from the interaction between Ti–OH groups of the substrate and Si–OH, Ti–OH groups of the sol, thus forming Ti–O–Si and Ti–O–Ti covalent bonds. Comparing the values of R_O, it is evident that the pore resistance of sol–gel coating on acetone degreased surface is about 2–3 times higher than those on the chemically treated samples. The low values of R_O indicate the attack of the coating surface by chloride ions. The value of R_in is highest for HNO₃ treated sample (2.5 × 10⁷ Ω cm²). R_O values of acetone degreased and cathodically cleaned sol–gel coated samples are in the same order (1.5 × 10⁷ Ω cm²). The coated sample with HF–HNO₃ treatment has the lowest R_O value of 46 Ω cm² and also exhibits the oxide resistance lower by about 2 orders (1.6 × 10⁵ Ω cm²). Thus, the protection to the surface has been rendered by the dense layer of the oxide film and that from the porous top layer is negligible. The value of n₂ being 0.9 (closer to 1) further confirms the intactness of the layer.

As seen from above results among all surface treated alloys, lowest corrosion current is observed for HNO₃ treated samples. HNO₃, being a good oxidizing agent may help passivation of the surface thereby being less detrimental to it. Also Mo is known to show better corrosion resistance behavior. XPS studies have demonstrated that oxidized Mo is observed to be more in HNO₃ etched coating compared to other treated coatings (Table 3). On the other hand, NaOH and HF–HNO₃ treated alloys contain less oxidized Mo according to XPS studies. 3D profilometry studies also show that HF–HNO₃ treatment removes passive oxide layer on the top surface. Hence, corrosion resistance property of HF–HNO₃ treated alloy is low compared to other alloys. Sol–gel coating removes the effects of all the surface treatments and restores the corrosion resistance of the treated surface except in the case of HF–HNO₃ treated one. This may be due to poorer condition of the initial surface.

4. Conclusions

Titanium alloy (Ti–15Mo–3Nb–3Al) has been chemically treated with different reagents including NaOH, HNO₃ and HF–HNO₃ solutions. The morphological study indicates that the surface has severely been affected by the HF–HNO₃ etchant. Other treatments do not bring about any major changes. EDS and XPS studies reveal the partial breakdown of surface oxide layer. HNO₃ treated alloy surface gives a better corrosion protection to the surface because of its passivation nature and HF–HNO₃ treated alloy shows the least protection. Sol–gel coatings on the other hand, provide improved protection to all the surfaces and this is because of the formation of a dense sol–gel oxide layer due to the interaction between Ti–OH groups of the substrate and Si–OH, Ti–OH groups of the sol, thereby forming Ti–O–Si and Ti–O–Ti covalent bonds.

Acknowledgements

Authors wish to thank Director, CSIR-National Aerospace Laboratories for giving permission to publish this work. Help rendered by Mr Siju and Mr Praveen for FESEM and 3D profilometry studies is gratefully acknowledged.

References

1. M. Peters, J. Kumpfert, C. H. Ward and C. Leyens, Adv. Eng. Mater., 2003, 5, 419.
2. S. Izman, M. R. Abdul-Kadir, M. Anwar, E. M. Nazim, R. Rosliza, A. Shah and M. A. Hassan, Titanium Alloys – Towards Achieving Enhanced Properties for Diversified Applications, in Surface Modification Techniques for Biomedical Grade of Titanium Alloys: Oxidation, Carburization and Ion Implantation Processes, ed. Dr A. K. M. N. Amin, InTech, Rijeka, Croatia, 2012.
3. X. Liu, P. K. Chu and C. Ding, Mater. Sci. Eng., R, 2004, 47, 49.
4. A. M. Kumar, P. Sudhagar, S. Ramakrishna, Y. S. Kang, H. Kim, Z. M. Gasem and N. Rajendran, Appl. Surf. Sci., 2014, 307, 52.
5. A. Kazek-Kęsik, M. Krok-Borkowicz, E. Pamuła and W. Simka, Mater. Sci. Eng., C, 2014, 43, 172.
6. H.-H. Huang, C.-P. Wu, Y.-S. Sun, H.-M. Huang and T.-H. Lee, Thin Solid Films, 2013, 549, 87.
7. Y. Sasikumar, M. Karthega and N. Rajendran, J. Mater. Eng. Perform., 2011, 20, 1271.
8. P. F. Gostin, A. Helth, A. Voss, R. Sueptitz, M. Calin, J. Eckert and A. Gebert, J. Biomed. Mater. Res., Part B, 2013, 101, 269.
9. A. Nanci, J. D. Wuest, L. Peru, P. Brunet, V. Sharma, S. Zalzal and M. D. McKee, J. Biomed. Mater. Res., 1998, 40, 324.
10. M. Takeuchi, Y. Abe, Y. Yoshida, Y. Nakayama, M. Okazaki and Y. Akagawa, Biomaterials, 2003, 24, 1821.
11. Y.-L. Zhou and D.-M. Luo, J. Alloys Compd., 2011, 509, 6267.
12. H. Krawiec, V. Vignal, J. Loch and P. Erazmus-Vignal, Corros. Sci., 2015, 96, 160.
13. R. Mythili, S. Saroja and M. Vijayalakshmi, Mater. Charact., 2010, 61, 1326.
14. J. Jayaraj, A. R. Shankar and U. K. Mudali, Electrochim. Acta, 2012, 85, 210.
15. Y. Yang, C. Xia, Z. Feng, X. Jiang, B. Pan, X. Zhang, M. Ma and R. Liua, Corros. Sci., 2015, 101, 56.
16. H.-M. Kim, F. Miyaji, T. Kokubo and T. Nakamura, J. Biomed. Mater. Res., Part A, 1996, 32, 409.
17. S. M. Bhola and B. Mishra, Int. J. Electrochem. Sci., 2013, 8, 7075.
18. R. Roest, A. J. Atanacio, B. A. Latella, R. Wuhrer and B. Ben-Nissan, Mater. Forum, 2007, 31, 160.
19. M. C. Advincula, D. Petersen, F. Rahemtulla, R. Advincula and J. E. Lemons, J. Biomed. Mater. Res., Part B, 2007, 80, 107.
20. J.-H. Liu, Z.-W. Zhan, M. Yu, S.-M. Li, Q. Wang, J. Zhang and M. Wang, Surf. Interface Anal., 2012, 44, 863.
21. D. G. Kolman and J. R. Scully, J. Electrochem. Soc., 1993, 140, 2771.
22. R. V. Lakshmi, T. Bharathidasan and B. J. Basu, Appl. Surf. Sci., 2011, 257, 10421.
23. R. V. Lakshmi, G. Yoganandan, K. T. Kavya and B. J. Basu, Prog. Org. Coat., 2013, 76, 367.
24. A. Jitianu, J. Doyle, G. Amatucci and L. C. Klein, J. Sol-Gel Sci. Technol., 2010, 53, 272.
25. M. Gnyba, M. Keränen, M. Kozanecki, R. Bogdanowicz, B. B. Kosmowski and P. Wroczynski, Opto-Electron. Rev., 2002, 10, 137.
26. B. Riegel, W. Kiefer, S. Hofacker and G. Schottner, J. Sol-Gel Sci. Technol., 2002, 24, 139.
27. M. Sreedhar, I. N. Reddy, P. Bera, D. Ramachandran, K. G. Saravanan, A. M. Rabel, C. Anandan and P. Kuppusami, Appl. Phys. A, 2015, 120, 765.
28. S. I. Castañeda, I. Montero, J. M. Ripalda, N. Díaz, L. Galán and F. Rueda, J. Appl. Phys., 1999, 85, 8415.
29. M. Tordjman, C. Saguy, A. Bolker and R. Kalish, Adv. Mater. Interfaces, 2014, 1, 1300155.
30. P. Bera, H. Seenivasan and K. R. Rajam, Surf. Rev. Lett., 2013, 20, 1350006.

---