Corrosion behaviour of glow discharge nitrided titanium alloys

Corrosion behaviour of glow discharge nitrided titanium alloys

Corrosion Science 45 (2003) 511–529 www.elsevier.com/locate/corsci Corrosion behaviour of glow discharge nitrided titanium alloys S. Rossi a,* , L...

2MB Sizes 1 Downloads 8 Views

Corrosion Science 45 (2003) 511–529 www.elsevier.com/locate/corsci

Corrosion behaviour of glow discharge nitrided titanium alloys S. Rossi

a,*

, L. Fedrizzi b, T. Bacci c, G. Pradelli

c

a b

Department of Materials Engineering, University of Trento, via Mesiano 77, 38050 Trento, Italy Department ICMMPM, University of Rome ‘‘La Sapienza’’, via Eudossiana 18, 00184 Rome, Italy c Department of Mechanical and Industrial Technologies, University of Florence, via S. Marta 3, 50139 Florence, Italy Received 25 April 2001; accepted 13 June 2002

Abstract The aim of this work is the study of the corrosive behaviour of glow discharge nitrided Ti– 6Al–4V alloy, using electrochemical techniques. Potentiodynamic polarisation and electrochemical impedance measurement show the excellent corrosive resistance of titanium alloy after thermochemical treatment, as well as different behaviour between the alloy before and after the nitriding treatment. Also, in very hostile environments, such as 5 wt.% HCl, where the titanium alloy is heavily corroded, the nitrided samples show good resistance. Nevertheless, 750 °C nitrided samples have worse behaviour than those treated at 900 °C. This is probably due to the lower corrosion resistance of the nitride e present in the nitrided layer created at the lower temperature than the nitride d, and the limited thickness of the modified layer obtained at this temperature. Ó 2002 Elsevier Science Ltd. All rights reserved.

1. Introduction The corrosion behaviour of titanium and its alloys is very good in many environments [1,2]. For this reason, and because of its good mechanical properties (such as the high yield stress/density ratio) industrial interest in titanium and its alloys has increased significantly in recent decades. This remains true, despite the high cost of these materials compared with steel, especially under heavy working conditions and

*

Corresponding author. Tel.: +39-0461-882403; fax +39-0461-881977. E-mail address: [email protected] (S. Rossi).

0010-938X/03/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 1 0 - 9 3 8 X ( 0 2 ) 0 0 1 3 9 - 7

512

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

when it is necessary to achieve high reliability (for example, use with aeronautic components and biomaterials). On the other hand, titanium alloy’s wear resistance is inadequate in many applications. In order to improve the tribological behaviour of these alloys, surface thermochemical treatments (such as nitriding in gas or in molten salts to obtain hard surface compounds) have been proposed [3–7]. Glow-discharge ion-nitriding is an interesting technique because of the versatility of process conditions. It is well known that the glow-discharge, ion-nitriding process allows us to obtain, in a short treatment time, very hard, nitrided surface layers on titanium alloys [8–10]. In fact, as a function of the alloy chemical composition and working conditions, the ion-nitriding process produces a continuous surface layer (compound layer) consisting essentially of the phases d (derived from the cubic nitride TiNx , 0:43 < x < 1:08, with a maximum hardness of 2400 HV) and e (a tetragonal nitride with a formula close to Ti2 N and a maximum hardness of 1600 HV) in variable ratios, and an inner diffusion layer characterised by the presence of crystals of a phase (interstitial solid solution of nitrogen in hcp titanium, with a maximum nitrogen content of 23 wt.% at 1323 K) embedded in the a–b metal matrix [11–13]. In many literary works concerning microstructural characterisation, the mechanical and physical–chemical properties [4,10,14–19] and the tribological behaviour of these nitrided alloys are reported [8,9,11,18,20]. Moreover, it is possible to find many articles regarding the corrosive behaviour of physically and chemically vapour deposited titanium nitride films, [21,22] but there is less work concerning the corrosion properties of thermochemical nitrided materials. Rolinski [18] studied the surface properties of plasma, nitrided titanium alloy WT3-1. Using anodic polarisation curves in 15 wt.% H2 SO4 , he noted a corrosion potential increase and decrease of the anodic current. A change in the nitriding temperature (730–1030 °C) did not lead to any essential changes in the shape of the polarisation curves. Other authors [23] investigated the corrosion resistance of the Ti–6Al–4V alloy nitrided by pulsed laser in sulphuric or hydrochloric acid. These authors also observed that nitriding treatment improves the corrosion behaviour of titanium alloys. The aim of the present work is to study the corrosion behaviour characterisation of ion glow discharge nitrided titanium alloy Ti–6Al–4V, using electrochemical techniques such as potentiodynamic curves and electrochemical impedance spectroscopy. Samples nitrided at two temperatures (750 and 900 °C) were studied in different, aggressive environments such as NaCl, Na2 SO4 and HCl. For the sake of comparison, a non-nitrided alloy was also considered.

2. Experimental Disc samples 18 mm in diameter and 2 mm thick, obtained from the same alloy bar (with the following chemical composition: Al 5.98, V 3.25, Fe 0.09, O 0.16, N 0.03, Ti bal.) were studied, polished with emery paper (from 320 to 4000) and 1 lm of diamond paste before the electrochemical test. After polishing, the samples were treated in a laboratory glow discharge, ion nitriding system, where the sample was the cathode, and the anode, the armour of

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

513

the treatment chamber. The following process parameters were used: gaseous mixture composed of 80% nitrogen and 20% hydrogen; pressure of 8 Torr; feed voltage of 400–450 V and current density of 15–20 mA/cm2 ; nitriding temperature of 750 or 900 °C; 6 h of nitriding time. Before nitriding, the surface was cleaned by a cathodic sputtering for 30 min at low pressure (0.1 Torr) and with a voltage of about 1000 V. The nitrided samples were first characterised by optical and electron microscopes. Polarisation curves were carried out on treated and untreated samples using a PAR 273 potentiostat with a potential scan rate of 0.2 mV/s. A three electrode electrochemical cell was used where the counter electrode was a platinum wire, and the reference electrode was a calomel for the chloride containing solutions (SCE þ245 mV vs. NHE) and a sulphate for environments without chlorides (SSE þ615 mV vs. NHE). The samples were introduced in a sample holder which left an exposed test surface of 2.5 cm2 , allowing electrical contact. The employed test solutions were: 2.8 wt.% Na2 SO4 and 3.5 wt.% NaCl (neutral, and at pH ¼ 2, by adding H2 SO4 or HCl respectively). Electrochemical impedance measurements were carried out during immersion in test solutions (for 60 days) using a PAR 273 potentiostat and a Solartron 1255 frequency response analyser. These measurements were carried out at the corrosion potential in a frequency range from 100 kHz to 1.5 mHz with a peak to peak amplitude of 10 mV. For these experiments, the environments selected were: 3.5 wt.% NaCl at pH ¼ 2 (addition of HCl) and 5 wt.% HCl. In the most aggressive environment (5 wt.% HCl), samples nitrided at both 750 and 900 °C were tested, whereas in NaCl solution, only samples nitrided at the higher temperature were evaluated; the untreated titanium alloy was also tested in both environments. Impedance data was analysed by using a fitting program developed by Boukamp [24]. The samples were observed after the electrochemical tests with the electron microscope to highlight the type of corrosive attack.

3. Results and discussion 3.1. Microstructural analysis Fig. 1 shows a metallographic section of the titanium alloy nitrided at 900 °C after metallographic attach with 1 ml HF, 2 ml H2 NO3 , 97 ml H2 O. The typical microstructure of this alloy is evident. Nitriding obtained at a temperature lower than b transus, modifies the surface structure creating a thin, continuous layer of composition made of nitrides d (TiN) and e (Ti2 N) [15,16]; under this layer the diffusion zone of titanum a crystals enriched with nitrogen is present, and finally there are single a crystals enriched with nitrogen immersed in a a þ b matrix. During this process the b phase partially transforms to a phase, so there is the presence of a grains in a b matrix transformed with a Widmanstatten structure.

514

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

Fig. 1. Electron micrography of sample nitrided at 900 °C after etching.

The structure of 750 °C nitrided samples is shown in Fig. 2. There are equiaxed grains of a phase with the b phase in the grain boundaries [17]; in these samples there

Fig. 2. Electron micrography of sample nitrided at 750 °C after etching.

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

515

is a greater presence of the nitride Ti2 N (as proposed in the literature) and the thickness of the surface layer of the nitrides is lower than the one observed after treatment at 900 °C [12,13,17]. The thickness of the composite layer is not completely homogeneous. However, it is about 5 lm for the 900 °C treated sample, and about 1.5 lm for the 750 °C one. In both samples, the nitrided layer does not show defects or cracks passing from the surface to the substrate alloy; only some defects can be noted with a horizontal course at the interface between the composition and diffusion layers [15,16]. The analysis carried out on the nitrided sample surfaces using X-ray diffraction does not show the presence of pure titanium or TiO2 as described in previous works [15,16,25]. This is explicable by the operative conditions of the nitriding treatment, because of the total absence of oxygen in the treatment chamber. The nitrogen that is still not bound as a nitride, tends to bind itself to the titanium atoms left free before diffusing inside. 3.2. Anodic polarization curves Fig. 3 shows the polarization curves obtained in a solution of 2.8 wt.% Na2 SO4 . It is possible to observe different behaviour between treated and untreated samples. The untreated alloy shows an anodic current with an almost constant value (in the order of some lA/cm2 ) determined by the presence of a passive oxide layer, as well known.

Fig. 3. Polarisation curves of treated and untreated samples in 2.8 wt.% Na2 SO4 solution.

516

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

Fig. 4. Polarisation curves of treated and untreated samples in 3.5 wt.% NaCl solution.

To the contrary, the nitrided samples show anodic currents that, at first, are lower by more than one order of magnitude than the non-treated alloy. Nevertheless, these values do not remain constant; they grow according to the anodic polarization potential (see Fig. 4 too). As previously stated, [15,16,25] the treated Ti–6Al–4V alloy only shows titanium bound with nitrogen on the surface. This modified surface layer, made up of two different titanium nitrides phases, have great chemical inertia in several aggressive environments [21]. At the two different temperatures, the nitrided samples both show a rise of the corrosion potential compared with the non-treated alloy and the courses of the polarisation curves are comparable. The same behaviour is evident even in solutions containing chloride (Fig. 4); both the thermally treated samples and the non-treated alloy do not suffer from increased hostility of the corrosive solution. The tested samples show no defects or localised attacks after the corrosion tests.

3.3. Electrochemical impedance measures For comparison, impedance measures were carried out during immersion time in solutions of 3.5 wt.% NaCl with HCl (pH ¼ 2) and 5 wt.% HCl, on the 900 °C nitrided samples and on the non-treated samples as well. Remembering the previous results, and given that impedance measures are conducted under free immersion, we chose these test conditions since we considered them most aggressive. In the more aggressive environment, we also tested the 750 °C nitrided samples.

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

517

Fig. 5. Corrosion potential of samples (untreated and nitrided at 900 °C) in 3.5 wt.% NaCl þ HCl (pH ¼ 2) solution.

3.4. 3.5 wt.% NaCl + HCl (pH 2) Fig. 5 shows the corrosion potential measured during immersion of the treated and untreated samples. It is interesting to observe the greater nobility of the nitrided samples, whose potential remains almost the same for more than 50 days during testing. The behaviour shown by the untreated alloy is completely different: it seems to suffer under the reaction with the aggressive solution. In particular, the break of the passivity promoted by the chlorides, opposes the trend of the potentials towards more noble values, as expected, given the natural passivation of the alloy. This is clear after about five days into immersion and, definitely, after 13 days. This last, sudden variation of the potential towards values between 250 and 300 mV SCE is a symptom of the damage done to the passive titanium oxide film. Figs. 6 and 7 show, in the form of Bode phase, the impedance spectra relevant to Ti–6Al–4V alloy and the 900 °C nitrided sample, respectively. For the untreated samples, the data obtained from the moment of immersion up until about 12 days after, are very different from those obtained during the following period. In fact, after this period, pitting begins with the breaking of the passive film: this fact is well noted from the variation of the phase angle at low frequencies [26] and from the decrease in the impedance value. The nitrided sample shows a completely different behaviour due to the absence of the surface titanium oxide layer and the presence of titanium nitrides with high chemical inertia. After 50 days of immersion in sodium chloride, the impedance measures still show (besides the stability of a noble potential as described before, a

518

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

Fig. 6. Impedance diagrams (Bode phase) in time of untreated alloy Ti–6Al–4V in 3.5 wt.% NaCl þ HCl (pH ¼ 2) solution.

Fig. 7. Impedance diagrams (Bode phase) in time of 900 °C nitrided sample in 3.5 wt.% NaCl þ HCl (pH ¼ 2) solution.

very high total resistance (>5 MX cm2 ) which confirms the low reactivity of the treated alloy. The study of the corrosion behaviour of passivable metals, like stainless steel and titanium, with the formation of a surface oxide layer, is particularly complex. In literature, some physical–chemical models are suggested to explain the formation of

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

519

Fig. 8. Equivalent electrical circuit used to fit the impedance data of the titanium alloy in passivity state.

this passivity layer [27]. As a result, the equivalent electrical circuit, used for data interpretation of impedance measures is complex. Concerning the study of titanium alloy in its passive state, since there is great similarity of behaviour with stainless steel, we can use the equivalent electrical circuit suggested by Jacobs [28] for the alloy Fe–Cr (20%Cr) (Fig. 8). Qox and Rox are representative of the capacity and the resistance of the oxide film; Rct and Qdl describe the electrochemical reactions that occur at the metal-passive layer interface, whereas the time constant (R1 C1 ) describes the phenomena of ionic diffusion in the oxide. This electrical circuit can be representative if it is presumed that with EIS we monitor the anodic process. The use of this equivalent electrical circuit is considered valid for the study of titanium under passivity conditions. Nevertheless, even if some data obtained with the use of the fitting program show that there are three time constants, they are not clear in the measurements carried out. Because of this, we decided to analyse the impedance measurement data through a simplified circuit derived from the one described above with some modifications. Using this simplified circuit, we obtained acceptable fitting errors, (v2  105 ), and it was possible to characterise the evolution of the corrosion behaviour of these materials considering the value variations of the typical parameters. The circuit adopted to understand the impedance data of the untreated alloy when it is in its passivity state is shown in Fig. 9.

Fig. 9. Simplified equivalent electrical circuit used to fit the impedance data of the titanium alloy in passivity state.

520

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

It is made of two electrical circuits (RQ in series) where the first represents the ionic diffusion processes into the oxide layer, and the second represents the corrosion process where Rct is the charge transfer resistance. Qdl is the double layer capacity and R0 represents the resistance of the electrolyte. In this circuit, the time constant, related to the resistance and capacity of the oxide, is neglected. Owing to the elimination of the time constant related to the oxide, it is necessary to reduce the analysis of the impedance diagram, avoiding the high frequency part where there is information regarding the resistance and capacity of the oxide itself, so we can obtain acceptable results. (The analysis was done in a frequency range between 1–10 mHz and 1 kHz.) When there are pitting phenomena, to fit the experimental data, we used use a simplified electrical circuit made up of only one time constant, R0 ðRct Qdl ) with acceptable fitting errors because of the predominance of the circuit part related to the corrosive phenomenon (Fig. 10). In this case, the electrical circuit is short-circuited in the active defect area. To interpret the impedance measurement data obtained on the nitrided samples, we used an equivalent electrical circuit R0 ðRct Qdl ) that represented the condition of a non-passive material, remembering the homogeneous surface conditions of these samples (Fig. 10); this circuit is also confirmed by the low fitting error (v2 ¼ 105 ). Fig. 11 shows the trend of the Rct inversely proportional to the corrosion rate for the nitrided sample and the untreated alloy. The untreated alloy shows high initial values, in the order of 10 MX cm2 , which indicates a very low corrosion rate, certainly due to the presence of the surface titanium oxide layer. The sudden fall, after five days of immersion, followed by a rise in the Rct to initial values, and the ultimate reduction of resistance after 13 days (in the range of 700 kX cm2 for 22 days of immersion) agree perfectly with the course seen for the potential and can be related to a pitting attack promoted by the chlorides present in the testing solution. Pitting attack is confirmed by observation done with an electron microscopy (Fig. 12). Electrochemical impedance spectroscopy vs. time is very sensitive to the pitting corrosion phenomena; this susceptibility was not highlighted by potentiodynamic polarisation, probably because of the need to use a much lower potential scanning rate. The nitrided sample shows high initial values of Rct in the order of 10 MX cm2 ; during the long period of the test, we can only see a slight fall of the Rct (to values of about 3.7 MX cm2 ).

Fig. 10. Equivalent electrical circuit used to fit the impedance data of the titanium alloy in activity state and of nitrided alloy.

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

521

Fig. 11. Rct trend for the nitrided sample and for the untreated alloy in 3.5 wt.% NaCl þ HCl (pH ¼ 2) solution (e:e:c: ¼ equivalent electric circuit showed in Figs. 9 and 10).

Fig. 12. Pitting attack of untreated sample after immersion.

As far as capacitance values are concerned, it is possible to observe values in the range of same dozens of lF/cm2 for the untreated alloys; whereas the values of nitrided samples are in the range of a hundred lF/cm2 . These values confirm the

522

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

absence of a surface oxide. Therefore, the behaviour of the nitrided alloy is that of a material with very low reactivity. Microscopic analysis of the sample after 60 days of testing does not show signs of serious corrosive attack. 3.5. 5 wt.% HCl Fig. 13 shows the corrosion potential trends with the time of the nitrided and nonnitrided alloy in the chloride acid solution. The corrosion potential of the alloy Ti– 6Al–4V shows a rapid decrease towards very active values already in the first hours of immersion (about 700 mV SCE). This indicates that the protective titanium oxide layer does not resist the environmental conditions. The potential of the 900 °C nitrided sample is almost stable and noble (about 200 mV SCE) for about 50 days into the immersion. It is interesting to note that the nitrided alloy potential is similar to the one measured in the sodium chloride solution. However, the 750 °C nitrided samples show a slow initial decrease of the potential value from 255 to 18 mV SCE after 21 days of immersion. With immersion time the potential shows a sudden decrease of about 609 mV after 31 days. These potential values indicate clearly that the substrate alloy is wet by the electrolyte. Electron microscopic analysis of the untreated samples after that tests, shows that intense general corrosion has occurred. Fig. 14 show the impedance diagrams vs. time of the untreated alloy in 5 wt.% HCl (Bode phase representations). Total impedance is much lower, compared to the data obtained from the sodium chloride solution, due to the higher aggressive solution: the values are, at first immersion time, in order of 40 kX cm2 and already, after 20 h of immersion, there is a quick drop in value to about 4 kX cm2 with a

Fig. 13. Corrosion potential of samples (untreated and nitrided at 900 and 750 °C) in 5 wt.% HCl solution.

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

523

Fig. 14. Impedance diagrams (Bode phase) of untreated alloy Ti–6Al–4V in 5 wt.% HCl solution.

reduction in time to about 750 X cm2 at day 16. This phenomenon agrees with the potential values discussed before, and considering the very aggressive environment, it appears that the passive layer of TiO2 is destroyed after the first hours of immersion. Nevertheless, after 24–48 h, a second loop can be seen at low frequencies, probably associated with the presence of intermediate reactions or the formation of corrosion products. Even in the most aggressive solution, that of 5 wt.% HCl, the nitrided samples show better behaviour than the base alloy samples. In Fig. 15, the impedance diagrams of the alloy treated at 900 °C (a) and 750 °C (b) are reported with the representation of Bode phase. The alloy nitrided at 900 °C shows total impedances greater than MX cm2 at first immersion times. As test time goes on, there is a slow fall in value in the order of 350 kX cm2 ; this fact highlights a corrosive process even if it is less important than the one that occurred to the untreated alloy. By electron microscopic analysis, after 60 days immersion, the samples show attack; nevertheless, it does not highlight the localised attack points that permit the electrolyte to reach the substrate alloy. Also, the test solution analysis with the plasma spectroscopy (direct current plasma) shows only the presence of titanium, and not aluminium and vanadium, coming from the substrate. The aggressive environment produces only a light dissolution of the nitride’s layer, and after 60 days of immersion, the base alloy is still protected by the nitride’s layer. This fact is confirmed by the constant value of the corrosion potential in time (Fig. 13). The 750 °C nitrided samples also show good corrosion behaviour; nevertheless, they are worse compared with the previous samples. The total impedances just after immersion are greater than MX cm2 but they decrease in time to values in the order of 200–300 kX cm2 . This reduction can also be noted from the phase angle variation

524

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

Fig. 15. Impedance diagrams (Bode phase representation) in time of the alloy treated at 900 °C (a) and 750 °C (b) in 5 wt.% HCl solution.

at low frequencies in the Bode phase diagram. The electron microscope analysis of the surface and the metallographic sections of the samples, after 60 h immersion in chloride acid also highlight a corrosive attack which is greater than the one observed on the samples nitrided at 900 °C (Fig. 16). The X-ray diffraction analysis, done before and after immersion, confirms these results: the spectra of the 900 °C nitrided sample remains unchanged (Fig. 17); on the contrary, as in the case of the 750 °C treated sample, it is possible to observe an intification of the signal related to the

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

Fig. 16. 750 °C nitrided samples after immersion in 5 wt.% HCl solution.

Fig. 17. XRD spectra of 900 °C nitrided sample after immersion in 5 wt.% HCl solution.

525

526

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

Fig. 18. XRD spectra of 750 °C nitrided sample after immersion in 5 wt.% HCl solution.

substrate (Fig. 18). Regarding the 750 °C nitrided sample, the analysis of the test solution with plasma spectroscopy points out the presence of aluminium, besides titanium, and this confirms the damage to the nitrided layer. We propose that the worst corrosion behaviour shown by the samples nitrided at 750 °C, in comparison to the ones treated at a higher temperature, could be due to the lower corrosion resistance of the nitride e. This nitride is mostly present in the nitrided layer created at the lower temperature, in comparison to the nitride d and limited thickness of the modified layer obtained at this temperature. We can use the equivalent electrical circuit, in the case of an untreated sample in active conditions of a sodium chloride solution, to interpret the impedance data on the untreated alloy samples for the part of the diagram regarding the corrosive phenomenon (Fig. 10). The electrolyte resistance is some X cm2 . For the samples nitrided at 900 °C, the equivalent circuit used is made of a simple circuit, already adopted in the case of the less aggressive solution (Fig. 10). The circuit used for samples nitrided at 750 °C was the same one used for the previous sample up until 27 days of immersion; after this period, observing a decrease in the potential value an equivalent circuit made of two electrical circuits was adopted (Fig. 19). Here, R1 represents the pore resistance, Q1 the capacity of the nitrided layer, and Rct and Qdl the charge transfer resistance and double layer capacity on the substrate alloy. Fig. 20

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

527

Fig. 19. Equivalent electrical circuit used to fit the impedance data of the 750 °C nitrided samples after 27 days of immersion in 5 wt.% HCl solution.

Fig. 20. Rct values in time for the untreated alloy and the nitrided samples (e.e.c. ¼ equivalent electrical circuit showed in Figs. 10 and 19).

represents Rct vs. time, for the untreated alloy and nitrided samples. It is clear that for the non-nitrided samples, the Rct reaches very low values in the order of 400–500 X cm2 just after a few hours; this demonstrates a fast degradation phenomenon. The sample nitrided at 900 °C, shows values of Rct of about 500 kX cm2 , signs of corrosion much slower than the one present on the non-treated alloy (even if they are higher than the ones measured in the sodium chloride solution). With immersion time there are no significant modifications, and this fact leads to the belief that there is a nitrided layer up until the end of the tests; this fact is also confirmed by the constant value of the corrosion potential. In the 750 °C nitrided samples, the Rct also have high values of over 600 kX cm2 . After 27 days immersion, when the equivalent electrical circuit is used with two time constants, the values of Rct always remain in the order of 500 kX cm2 , high for a

528

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

charge transfer resistance of a titanium alloy (see the tests on the non-nitrided alloy). However, it must be remembered that the exposed area factor is lower than the one considered in the tests. The R1 resistance of the nitrided layer defects has values of 3– 14 kX cm2 . Considering the Qdl , in time, the untreated samples show values that rapidly become much greater than those measured in the NaCl solution, with an increasing trend due to the corrosive process that produces both a surface roughness increase and the presence of corrosion products; the capacity goes from 18.9 lF/cm2 of the first immersion times to 700–900 lF/cm2 . These values confirm the absence of protective oxides. Regarding the nitrided samples, the double layer capacity is in the order of 200 lF/cm2 , which confirms the absence of titanium oxides on the surface. However, when several defects start to grow in the nitrided layer, the 750 °C treated sample, after 27 days, values at a capacity of 6–20 lF/cm2 .

4. Conclusions From the electrochemical tests, optimal corrosion resistance properties of Ti–6Al– 4V alloys nitrided by ion glow discharge were characterised. In NaCl solution, all samples show good corrosion resistance. While the untreated alloy shows good corrosion resistance because of the surface passive layer made by TiO2 , in the case of the nitrided samples, good behaviour is due to the chemical inertia of the surface nitrides. Also, in very aggressive environments, such as 5 wt.% HCl, where titanium alloys are heavily corroded, the nitrided samples show good resistance. Nevertheless, the 750 °C nitrided samples have poorer behaviour than those treated at 900 °C. In the case of a lower temperature treatment, it is possible to observe a decrease in the corrosion potential and the presence of localised attack after about 27 days of immersion owing to the degradation of the treated layer. The worse corrosion behaviour shown by samples nitrided at 750 °C. The ones treated at a higher temperature, show lower corrosion resistance because of the nitride e.

References [1] J.E.G. Gonz alez, J.C. Mirza-Rosca, Journal of Electroanalytical Chemistry 471 (1999) 109–115. [2] J. Pan, D. Thierry, C. Leygraf, Electrochimica Acta 41 (1996) 1143–1153. [3] Y.X. Leng, P. Yang, J.Y. Chen, H. Sun, J. Wang, G.J. Wang, N. Huang, X.B. Tian, P.K. Chu, Surface and Coatings Technology 138 (2001) 296–300. [4] A.L. Yerokhin, X. Nie, A. Leyland, A. Matthews, Surface and Coatings Technology 130 (2000) 195– 206. [5] S.M. Johns, T. Bell, M. Samandi, G.A. Collins, Surface and Coatings Technology 85 (1996) 7–14. [6] B.S. Yilbas, M.S.J. Hashmi, Journal of Materials Processing Technology 103 (2000) 304–309. [7] A. Bloyce, P.Y. Qi, H. Dong, T. Bell, Surface and Coatings Technology 107 (1998) 125–132. [8] B.S. Yilbas, M. Sami, S.Z. Shuja, A. Aleem, J. Nickel, A. Coban, Wear 212 (1997) 140–149. [9] B.S. Yilbas, A.Z. Sahin, A.Z. Al-Garni, S.A.M. Said, Z. Ahmed, B.J. Abdulaleem, M. Sami, Surface and Coatings Technology 80 (1996) 287–292.

S. Rossi et al. / Corrosion Science 45 (2003) 511–529

529

[10] T. Wierzcho, A. Fleszar, Surface and Coatings Technology 96 (1997) 205–209. [11] B. Tesi, A. Molinari, G. Straffelini, T. Bacci, G. Pradelli, in: P.A. Blenkinsop, W.J. Evans, H.M. Flower (Eds.), Proceedings of Titanium’95 Science and Technology, Birmingham, 22–26 October 1995, The Institute of Materials, pp. 1983–1990. [12] B. Tesi, T. Bacci, C. Badini, C. Gianoglio, Metallurgia Italiana 81 (1989) 367–374. [13] C. Gianoglio, C. Badini, B. Tesi, T. Bacci, Metallurgia Italiana 80 (1988) 291–296. [14] A. Raveh, A. Bussiba, A. Bettelheim, Y. Katz, Surface and Coatings Technology 57 (1993) 19–29. [15] P. Scardi, B. Tesi, T. Bacci, G. Gianoglio, Surface and Coatings Technology 41 (1990) 83–91. [16] T. Bacci, G. Pradelli, B. Tesi, G. Gianoglio, C. Badini, Journal of Materials Science 25 (1990) 4309– 4314. [17] M. Salehi, T. Bell, P.H. Morton, in: F.H. Froes, I. Caplan (Eds.), Titanium’92 Science and Technology, The Minerals, Metals and Materials Society, 1993, p. 2.127–2.134. [18] E. Rolinski, Materials Science and Engineering A 128 (1989) 37–44. [19] A. Raveh, P.L. Hansen, R. Avni, A. Grill, Surface and Coatings Technology 38 (1989) 339–351. [20] T. Bacci, G. Pradelli, A. Molinari, B. Tesi. Proc. III Congresso AIMETA tribologia, Capri 29–30 Settembre, 1994, pp. 135–142. [21] Y. Massiani, P. Gravier, J.P. Crousier, L. Fedrizzi, M. Dapor, V. Micheli, L. Roux, Surface and Coatings Technology 52 (1992) 159–167. [22] Y. Massiani, A. Medjahed, P. Gravier, L. Argeme, L. Fedrizzi, Thin Solid Film 191 (1990) 806–817. [23] M. Magrini, B. Badan, M. Bianco, La Metallurgia Italiana 84 (1992) 631–634. [24] B. Boukamp, Solid State Ionics 20 (1986) 31–44. [25] T. Bacci, P. Scardi, B. Tesi, G. Gianoglio, Surface Engineering 8 (1992) 141–144. [26] L. Van Leaven, M.N. Alias, R. Brown, Surface and Coatings Technology 53 (1992) 25–34. [27] M.J. Kloppers, F. Bellucci, R.M. Latanision, Corrosion 48 (1992) 229–238. [28] L.C. Jacobs. Ph.D. Thesis, Technische Universiteit Delft (NL) 1995.