Galvanic Corrosion Behaviour of Ti or
Al Metallic Coatings Deposited by PVD Multiarcs Process Onto
Low Alloy Steel
J. Creus, H. Idrissi, M. Mazille
Laboratoire de Physico-Chimie Industrielle,
INSA Lyon, Villeurbanne, France
C. Charrier, P. Jacquot
Innovatique SA (HIT), Chassieu, France
email: mazille@insa.insa-lyon.fr
ABSTRACT :
Keywords : Galvanic corrosion, aluminium coating, titanium coating, PVD, corrosion behaviour
1. Introduction
Different techniques exist to protect a metallic structure against corrosion in aggressive environment. Metallic or organic coatings represent a good choice for the protection of metallic structures. Among them, cadmium coatings are often used. Cadmium coatings onto steel offer good mechanical and electrochemical properties, but cadmium is well knowned for being a polluting substance and since 1991, an european directive strongly regulates its use in the industry. We are at present witnessing of a development of new coating which will be able to replace the cadmium in industry, for instance, electrodeposited Zn and Zn alloys coatings, deposited Al and/or Ti coatings by PVD process.
The low density of titanium associated with a high mechanical resistance of Ti alloys confer good properties for aeronautics and space applications [1]. The corrosion resistance of titanium results from the formation of a protective film, strongly adherent and stable. The oxide, in aqueous solution is typically composed of TiO2. The high temperature oxidation leads to a chemically resistant and cristalline form called Rutile, whereas at low temperature, the amorphous form Anatase is formed [2].
Aluminium is less noble than titanium [3]. When Al is exposed to air, a thin oxide film naturally develops over its surface which usually confers self protection [4]. This oxide film is composed of two layers, a thin layer which is in direct contact with the metal (the barrier film) and a second layer with a hexagonal cell structure and called the porous layer [4-6].
In the present paper, the electrochemical behaviour of titanium and aluminium coatings onto low alloy steel are studied in saline environment.The behaviour of bare metals in the same media is controlled before considering the study of the coating system. Galvanic corrosion current density is often the decisive criterion for using a coating as a protective one against corrosion in an aggressive media. The potential difference between each metal do not provide a quantitative prediction for the galvanic corrosion rate [7]. So, we investigate the effect of galvanic corrosion on the electrochemical behaviour of each coating. Different electrochemical techniques are employed, and the most appropriated would be pointed out.
2. Materials and experimental procedure
2.1. Materials and test solution
Ti coating of 5 mm and Al coating of 10 mm onto low alloy 4135 steel are deposited by PVD multiarcs technique by the Society INNOVATIQUE (HIT). Bare metals : 4135 steel, high purity aluminium and titanium are used as references.
The solution test is a NaCl 30 g/l solution, aerated and stirred with the rotative working electrode at 500 rpm. Usually, the samples are presented as cylindrical pawn of 10 mm diameter, and the exposed surface is fixed to 3 cm2.
2.2. Test method
The coating morphology is observed with a MEB. The experiments are carried out under the control of a EGG 273A potentiostat, using the conventional three electrode technique : the potentials are referred to the saturated calomel electrode SCE, and the counter electrode is in platinum. The electrochemical techniques used to characterize the behaviour of bare metals and coated steels are :
long-lasting immersion test
After degreasing, cleaning and drying, samples are immersed during 48 up to 75 h. The potential versus time is followed with simultaneous measures of polarization resistance performed between 20 mV around free corrosion potential. The speed increment of the polarization is 10 mV/min.
current-voltage curves i(E)
After a potential stabilization of 1 h, the curves i(E) are plotted from -50 mV/open circuit potential in the cathodic side up to 100 mV/OCP in the anodic side.
galvanic current measurements
In order to directly measure the current produced by the coupling of bare metals, a zero resistance ammeter ZRA is used. In fact a small modification of the potentiostat is necessary to convert it into ZRA. The ratio Sa/Sc = 1 is studied during a total immersion duration of 24 h.. The galvanic potential is also followed in the mean time.
electrochemical impedance spectroscopy EIS
Impedance determinations are made with a four channel frequency response analyser, Solartron 1254, over a frequency range from 4 mHz to 64 kHz. The amplitude of sinusoidal signal is chosen at 10 mV around the corrosion potential. Results are interpreted in terms of Nyquist plots.
3 Results and discussion
3.1. Behaviour of bare metals
A) Individual intrinsic behaviour
During long immersion in NaCl 30 g/l, the steel potential decreases quickly during the first hours from -460 to -620 mV/SCE. Then, the potential reachs a steady value around -620 mV/SCE. Mansfeld [8] found the corrosion potential of steel 4130 in NaCl 3 % solution at -598 mV/SCE, which is close to our potential determination.
Rapidly, rust appears on the sample surface. This rust is rather adherent. After several hours, the solution turns to rust colour and a deposit forms on the bottom of the electrolytic cell.
The titanium potential becomes more and more noble with immersion duration, from -420 mV/SCE at the beginning towards -200 mV/SCE after 75h. This increase is related to the oxide film growth. Ayers [9] found steel and Ti potential after 15 days of immersion in natural seawater of, respectively -720 and -150 mV/SCE.
Concerning aluminium, a steady potential is soon reached at about -730 mV/SCE but we noticed slight potential fluctuation around this steady value.
During the first immersion hours, the steel polarization resistance Rp [14] is relatively steady. Then, due to the formation of an adherent rust film, Rp slowly increases.
The values of titanium Rp are quite large, what means that the titanium dissolution is negligible in NaCl 30 g/l solution. The Rp values increase during immersion which corresponds to the growth of the oxide film on titanium.
At last, contrary to the other metals, aluminium Rp decreases.The Al behaviour in saline solution is controlled by the pitting potential which is close to the corrosion potential in saline solution [5].
Zim W
Table 1 summarizes the corrosion potential and current density obtained from the current voltage curves i = f(E). This curves, represented in fig 1, are plotted after a stabilization of the open circuit potential during 1 to 2 hours.
The aluminium potential stabilizes quickly around -750 mV/SCE. This value is in agreement with the value found by Merati [10]. Heurtaux [11] found a dissolution potential near to -740 mV/SCE in aqueous solution of NaCl 30 g/l.
The instantaneous dissolution speed of bare metals can be modified by the formation of a corrosion film on substrate surface. So we notice that steel dissolution is high and depends on the rotating speed of the electrode. Aluminium dissolution rate is smaller, however, it is still too important if we consider thin aluminium coatings. And finally, titanium dissolution is negligible, it presents a high corrosion resistance in NaCl 30 g/l solution.
Electrochemical impedance spectroscopy EIS results
The steel impedance diagrams in fig 2 present only one capacitive loop which is associated to the steel dissolution in NaCl. The loop diameter increases with immersion duration. This is related to the progressive formation of a thin barrier of corrosion product on steel substrate.
fig 2 : Nyquist diagrams of steel for different
immersion period
The Al impedance diagrams present a capacitive loop at high frequencies and an important dispersion of experimental points at low frequencies. The loop diameter decreases quickly and the stabilized value is rather close to those of polarisation resistance Rp.
The titanium impedance diagrams exhibit the beginning of a capacitive loop which is related to an high resistive behaviour of titanium oxide film developped on the surface of titanium.
To sum up, the EIS and stationary electrochemical data are in agreement. The steel has a high corrosion dissolution during the first immersion hours. Then, a barrier film of corrosion products, weakly adherent, reduces the steel dissolution. Aluminium suffers from pitting corrosion in saline solution, whereas titanium corrosion is negligible.
B) Galvanic corrosion between bare metals
To characterize the galvanic corrosion between Al / steel and Ti / steel, the galvanic currents are measured and compared to the results obtained from other electrochemical techniques such as coupling of polarization curves.
The results are reported in table 2.
| Sa/Sc = 1 |
|
|
In support of the above data, Evans diagrams and the monitoring of galvanic current lead to comparable results. For instance, the results obtained while recording ig during a 24 h immersion duration with a ZRA techniques, related to Al / steel couple, confirm that the ig evolution is quite steady around 300 mA/cm2. Weight lost measurement also leads to similar data.
Regarding Ti / steel couple presented in fig 3, the ig evolution is not steady at all. During the first hours, ig is quite low, near to the value reported table 2, then ig strongly increases to a maximum current density of 180 mA/cm2 and then slightly decreases with time.
In aerated chloride aqueous solution, the main cathodic reaction is the reduction of dissolved oxygen. It is also well knowned that Ti has a great tendency to be easily cathodically polarized ; that means that galvanic potential between Ti and another metals is close to the second one's whatever the surface ratio of each coupled metal [12].
This is confirmed by our own results, but still we should ask whether at this highly cathodic potential, the evolution of hydrogen could happen and so induce the formation of titanium hydride. Indeed, cathodically polarized, titanium will absorb hydrogen very easily, and the possible generated hydride may weaken the titanium oxide film [13]. This point is still under investigation.
To sum up, it appears that aluminium coating corrosion mode is basically controlled by galvanic corrosion through pinholes and porosity. While continous Ti coatings will confer an excellent corrosion resistance, but small pores will lead to an accelerated substrate corrosion.
3.2. Behaviour of Al and Ti coating on 4135 steel
A) Morphology
MEB observations have been performed on Al, Ti and Ti / Ti oxidized coatings. Aluminium coatings present different surface defects resulting from the PVD process (condensed dropplets, preferential crystal growth..) (fig. 4). Ti and Ti/TiO2 coatings have similar surface morphology. In this case, the surface is more continuous, and smooth with small white droplets associated to oxide form of titanium.
The XRD diagrams confirm the presence of a cristalline titanium oxide TiO2, with two allotropic structure, mainly rutile and some anatase depending of the experimental conditions used during the PVD process.
B) Electrochemical behaviour
Ti coating behaviour
The potential decreases from the initial potential of -380 mV/SCE, which is close to the corrosion potential of bare titanium, to -550 mV/SCE. Then, the potential evolution is quite similar to the steel one, which confirms the porosity of the coating. Ayers [9] found the potential of a plasma Ti deposit in natural seawater, after 15 days of immersion to be around -560 mV/SCE. This intermediary potential results from a corrosion of steel underneath Ti layer and through porosities. The final potential is close to steel corrosion potential, as it was suggested by Shalaby [12].
Rp data of titanium coating are close to those of bare steel. Rust appears quickly on Ti coating defects and the amount of rust formed is important for long immersion period.
MEB observations show that morphology of the coating is altered by the test. Defects are observed (pinholes, pores...) and seem to increase with steel corrosion beneath the coating. In fact, a local flake off is observed.
The polarisation curves recording on Ti coated samples allow electrochemical characteristics in NaCl solution.
These characteristics depend on the dual behaviour of the outer titanium layer and the steel anodic dissolution through the porosity of the coating. Due to the major contribution of the second process, the shape of the polarization curves of coated samples is close to that obtained on bare steel, still with a definitely much lower current intensity on an evidently much lower anodic area.
Impedance diagram for Ti coating is plotted in fig 5 after a 20 hours immersion test. The EIS is in agreement with the stationary method results. The impedance diagram present two capacitive loops, the first loop is related to the substrate / solution interaction through the porosity of the coating, whereas the low frequency loop characterize a slower phenomenon, perhaps the diffusion of corrosion products through the porosity of titanium coating.
If non porous, Ti coating shows good corrosion behaviour
in saline solution. The inherent porosity of the coating is anyway
disastrous for the substrate due to the very unfavourable surface
ratio. Effectively, a flake off resulting from the steel dissolution
under the coating is noted during long immersion duration.
Al coating behaviour
A 75 hours long immersion test (fig 6) is performed in NaCl solution. The open circuit potential quickly reachs a steady value of -715 mV/SCE. This potential is close to the bare Al metal. During the first 20 hours, the potential increases slightly up to -705 mV/SCE. Then a fast increase is noticed up to a potential of -440 mV/SCE. This value corresponds to the first immersion potential of bare steel in this medium. Then, the potential evolution is similar to the steel potential evolution.
Experimental conditions are quite severe for Al coating. The fig 6 bring us to the question whether the coating is still effective after 20 hours immersion.
It is worth to remember that in saline solution, the pitting potential is close to the Al corrosion potential. An intermediary potential of the coated steel close to pitting potential would explain the high dissolution rate of the coating. MEB observations after immersion show a dense film on steel surface, with an intense crack network.
Impedance diagrams show two capacitive loops in the high and middle range frequencies, and an inductive loop at low frequencies.
In the present work. we observe that the second loop disappears progressively during the immersion duration. So this loop is related to the aluminium film dissolution in this medium, whereas the high frequencies loop is related to the porosity through the Al coating.
4. Conclusion
Ti and Al coating efficiently protect the steel substrate in NaCl 30 g/l solution. The immersion experimental conditions are extremely severe for testing the coating corrosion resistance.
Anyway, the Al dissolution rate is actually important, essentially due to the effect of galvanic corrosion with steel through the porosity. This dissolution should be reduced if this kind of coating has to be employed as protective deposit.
On the other hand, titanium coating should confer a good corrosion resistance. Indeed, its dissolution is negligible but the galvanic corrosion induced is disastrous for steel : - The steel corrosion beneath the coating leads to a flake off. This phenomenon could be reduced by a final oxidation of the coating during its elaboration. In fact, this treatment might reduce the pore size, and so enhance the behaviour of coated steel.
Finally, we have noticed that galvanic current measurement between separated substrate and coating metal or alloy during a long lasting immersion test is a good technique to get meaningful informations about coated metal behaviour subjected to high porosity protective layers.
The authors thank Innovatique (HIT) society for the collaboration during this study. All the people who have contributed to the project are also thanked, especially F. FERRER and A. DELESCAZES for their technical support.
1 - R. TRICOT, Materiaux et technique, 1987, 07/08, p 297-307.
2 - R.W. SCHUTZ, Sixth worldconference on titanium, France 1988, tome 3, 1988, p19717-1922.
3 - J. MASALSKI, J. GLUSZEK, P. FURMAN, J. Mat. sc. letters, 1995, vol 14, p587-588
4 - O.E. BARIA, J.L. CAMARA, O.R. MATTOS, J. appl. Electrochem., 1987, vol 17, p 641 - 647
5 - J. BESSONE, C. MAYER, K. JUTTNER, WJ LORENZ, Electrochim. acta, 1983, vol 28, n°2, p 171-175
6 - B. VANDRILINDEN, H. TERRYN, J. VEREECKEN, J. appli. Electrochem., 1990, vol 20, p 798-803.
7 - W. KAUTEK, Corrosion Sc., 1988, vol 28, n°2, p 1763-199
8 - F. MANSFELD, J.V. KENKEL, Corrosion Sc., 1975, vol 15, p 183 - 198
9 - J.D. AYERS, R.J. SCHAEFER, F.D. BOGAR, E. McCAFFERTY, Corrosion, 1981, vol 37, n°1, p55-57
10 - A. MERRATI, These INSA LYON, 1994, 130p
11 - J. HEURTAUX, Galvano organo, 1990, vol 609, p 797 - 804
12 - L.A. SHALABY, Corrosion Sc. 1971, vol 11, p 767-778
13 - Y. COMBRES, Galvano organo, 1994, vol 645, p 351-354.
14 - J. CREUS, H. IDRISSI, H. MAZILLE (writing)