Go To Review

The Effect of the Thermal Spray Process on the Protective Behaviour of NiCr Alloy in Seawater

M.P.W. Vreijling, E.P.M. van Westing, G.M. Ferrari, P. Svaldi*
F.P.E. Westendorp**, E. Bullock***, J.H.W. de Wit****

TNO Centre for Coatings Research, Department for Corrosion Prevention, PO Box 6034, 2600 AB Delft, The Netherlands.

* University of Trento, Faculty of Materials Engineering, Laboratory for Electrochemistry, via Masiano 77, Italy.

** Ministry of Defence, Directorate Material, Royal Netherlands
Navy, Department of Platform Systems, PO Box 20702, 2500 ES, The Hague, The Netherlands.

*** European Commission, Joint Research Centre, Institute for Advanced Materials, Materials Engineering Unit,
PO Box 2, 1755 ZG, Petten, The Netherlands

**** Delft University of Technology, Section of Corrosion Technology,
Electrochemistry and Atom Spectroscopy, Faculty of Chemical Technology and Materials Science,
PO Box 5025, 2600 GA Delft, The Netherlands.

Email: M.Vreijling@kribc.tno.nl

ABSTRACT

Using metal spray coatings for marine corrosion prevention is widely recognized as a technique with great potential. Both bulk protection using arc or flame spray or the protection of small parts using plasma spray techniques, have received much interest over the last few years.

By contrast, the actual use of these techniques remains modest.

Many sceptics claim that the inherent porosity and presence of contaminants and oxides in the coating will inevitably limit the expected protective performance of these systems.

Also, the role of the spray application technique on the protection mechanism remains a point of discussion.

This paper describes the protective behaviour of a NiCr coating on stainless steel in natural seawater. Both APS (atmospheric plasma spray) and HVOF (high velocity oxygen fuel) sprayed coatings, using various spray powders (size, composition), are examined and compared with the behaviour of NiCr bulk metal.

Potential monitoring during exposure together with static and dynamic polarisation measurements were related to coating microstructure.

Results indicated that the inherent morphology of a sprayed coating can facilitate the passivation reaction, thus improving the resistance to corrosion phenomena as compared to the bulk material.

The presented data are the first results of research directed towards the influence of the thermal spray operation on the corrosion resistance of passivating metals.

Thermal spraying of metallic coatings is not a new phenomenon. Ever since the early 1900s when Schoop started to use the flame spray method for depositing zinc on a steel substrate, the technique has been available. Nonetheless, the actual use of the technique for the protection of steel against corrosion has remained modest as compared to the organic coating market. Despite the potentially much superior protection, organic coatings seem to be favoured for most applications. A possible explanation for this is the relative unfamiliarity of the public with the technique. It is often claimed that "anyone can wield a paint brush" and metal spray applicators need to be specially trained [1]. The result of this can sometimes be witnessed as rather large maintenance costs, but this does not seem to hurt the argument.

A second reason can be that most people are unfamiliar with the protective mechanism. As the protection of an organic coating can easily (but inadequately) be visualised as an isolation layer from the aggressive environment, thermal spray coatings must rely on electrochemical principles.

Thermal spray coatings can, for a given environment be divided into effectively inert coatings (gold, ceramics) and electrochemically active coatings. These metals can be classed as corroding or passivating and when the substrate material is exposed, as cathodic or anodic to their substrate. For each of these situations a different set of considerations needs to be met in order to have the desired protection. The general superior corrosion resistance of passivating alloys in many environments would make them a very suitable candidate to be applied as a coating on lesser resistant metals.

An additional difficulty stems from the morphology of thermal spray coatings. The coating is deposited by (partially) melting of small particles which are projected on to the substrate. A typical coating morphology results from the general lamellar build up, consisting of melted and partially unmelted particles together with oxide inclusions and the inevitable pores. The exact morphology depends on the coating material, the deposition temperature, speed of application etc. which is partially dictated by the choice of the thermal spray method and the applicator's settings and skill.

From both chemical and structural aspects the properties of the sprayed coating can differ considerably from their bulk mass equivalent. For mechanical properties it was shown that the sprayed coating does not need to be inferior, eg: pores can effectively act as crack stoppers thus considerably raising the fracture toughness [2] or favour the integration of the foreign object with the living tissue in spray coatings on medical implants [3]. However, for corrosion protection, the inherent morphology of thermal spray coatings the presence of pores and oxide inclusions is normally regarded as a considerable disadvantage.

This paper shows the first results of research directed towards the influence of the thermal spray operation on the corrosion resistance of passivating alloys.

Materials

Two types of spray coatings were used for these investigations. Both coatings used the same base material, consisting of Ni80Cr20 pre-alloyed powder with spheroidal grains. The size of the grains ranged from 5 µm to 110 µm with a mean size of 30 µm for the HVOF sprayed powder (METCO PEM 43 VF-NS). and from 11 µm to 110 µm with a mean size of 55 µm for the APS coating. (METCO PEM 43 F-NS). Particle sizes were measured using a Malvern Mastersizer.

Samples for measurements on the NiCr mass specimen were obtained from a Ni80Cr20 rod (Goodfellow Cambridge England).

Coating application

Coatings were sprayed on stainless steel AISI 316L coupons of approx. 100x50x3 mm since the NiCr coating were originally intended to be used as an protective intermediate for a ceramic wear resistant top coating. Previous results on this have been published earlier [4] Substrate surfaces were prepared by degreasing in acetone/ultrasonic vibration for 10 minutes, and rinsing in clear acetone. The surfaces were then grit blasted with Al₂O₃ grit to SA 2½, (SY BRANDY SD69 EKW 24), and were air jet blasted for 5 seconds and coated immediately.

Cleaned substrates were mounted, uncooled on a vertical specimen holder and preheated to ca. 100°C using HVOF or APS flame without powder. The coatings were sprayed to thickness between 150 and 200 µm and cooled in still air.

Both APS and HVOF spraying were performed at the Advanced Coating Centre (JRC, ECN) in Petten, The Netherlands.

The APS spray set-up was used in the JRC-IAM standard mode for H₂/N₂ gas spray conditions without special adaptation or optimisation. The HVOF spray series was optimised by variation of H₂ gas flow rate which influences the temperature and velocity of the particles, and therefore the degree of melting and the plasticity of the molten droplets at impact. The tested samples were sprayed at H₂ flow rate of 1300 scfh^-1 (standard cubic feet per hour). Coating morphology was examined using optical microscopy.

Experimental conditions

The electrochemical polarisation measurements were carried out using the Schlumberger 1286 Electrochemical Interface. Potential monitoring was done by a HP 3852A data acquisition control unit.

All polarisation measurements were performed in the Avesta Cell for flat specimens as produced by Bank Elektronik. This cell has a volume of 0.5 litres and a 1cm² working electrode area. Samples were preconditioned at -1000 mV (vs Ag/AgCl) for 60 seconds in the cell to remove any air formed surface oxides.

Polarisation measurements were done in natural seawater and sulphuric acid (H₂SO₄ 5N ca 21%), the latter to eliminate any localised corrosion phenomena.

The polarisation scan in natural seawater ranged from -800 to 1000 mV (vs Ag/AgCl). The range in sulphuric acid was -500 to 1500 mV (vs Ag/AgCl). Both experiments were done with a scan rate of 10 mV per second. During the measurement the electrolyte was stirred with a glass rod at approx. 200 rpm. For static polarisation the sample was polarised at +100 mV (vs Ag/AgCl) for 30 seconds, followed by +800 mV (vs Ag/AgCl) for the same duration. This cycle was repeated 4 times.

Potential monitoring specimens were pre-coated with metal primer and embedded in epoxy resin to insulate the non spray coated faces of the specimen. The electrolyte for the potential monitoring experiments consisted of continuously refreshed seawater from the Den Helder harbour.

Mass specimen samples (both for polarisation and potential monitoring) were polished to grit 1200. All specimens were cleaned in methanol using ultrasonic vibration for 5 minutes.

The effect of the spray application of passivating Ni80Cr20 alloy was investigated using the following experiments:

- Metallographic investigation of cross-sections of the coatings directly after application using optical microscopy;

- Investigation of the passive region/pitting behaviour in natural seawater using potentiodynamic polarisation;

- Investigation of the critical current density in sulphuric acid using potentiodynamic polarisation;

- Evaluation of repassivation after pitting using potentiostatic polarisation in natural seawater;

- Monitoring of the open circuit potential during 300 hours of exposure in natural seawater.

Three types of Ni80Cr20 samples were used in these experiments: atmospheric plasma spray deposited coating (APS), high velocity oxygen fuel sprayed coating (HVOF) and non-sprayed "mass" material.

Metallographic Investigation of the Sprayed Coatings

Due to the inherent differences of the APS coating technique as compared to HVOF application, the resulting coatings show important differences. An overview of the generic differences between the spray techniques can be found elsewhere [5]. In this paper the discussion will be limited to the typical resulting coating structure for the different techniques as shown in figure 1.

Figure 1 Typical micrograph of HVOF coating (left) and APS coating (right) Magnification is approx. 300x

The HVOF coating was deposited at 1300 scfh^-1 H₂ gas flow rate and shows heavy oxidation. The particles had melted well, giving a fine microstructure, but all splat lamellae showed strongly oxidised boundaries - also on the substrate/coating interface.

Samples sprayed with the PEM 43F powder using the standard APS set-up for H₂/N₂ gas mixtures show the coating to be considerably less oxidised but with clearly higher porosity than the HVOF coating. (Surface roughness is also considerably higher which may be advantageous to subsequent ceramic topcoat adhesion, but this is beyond the scope of this paper)

All coatings (both APS and HVOF) contain inclusions of the Al₂O₃ grit blast material at the interface between coating and substrate. These inclusions, together with an incidental substrate interface macro pore, are clearly visible from the left picture. The inclusions represent some 10% of the contact area but are generally well embedded in the substrate and do not appear to influence the coating protective properties.

Image analysis with a resolution of ca. 1 µ showed the surface area multiplication factor for the HVOF coating to be 202 % as compared to 190 % for the APS spray coating. This indicates that although the average roughness of the APS coating is considerably larger, the finer microstructure of the HVOF still leads to a larger surface area.

It is important to note that the largest difference between the APS and the HVOF coating will be the coating density or the amount of porosity. Some variations can be expected in resulting amount of metallic chromium caused by oxidation during spray. A rough estimate of oxide content in HVOF spray coating is 50% as to 10% for the APS coating.

Potentiostatic Polarisation Measurements in Natural Seawater

The scanning of the cathodic and anodic current resulting from a polarisation from -800 mV to +1000 mV (vs Ag/AgCl) in natural seawater is represented in figure 2. The corrosion potential (i=0) is about the same (-500 mV vs Ag/AgCl) for all three samples and corresponds well with literature values. It should be noted that the y-axis in these figures represents an amount of current measured at an exposed working electrode surface of 1 cm². The current values for the spray coatings have been divided by the surface area multiplication factor, discussed in the previous section. Even so, these values can only be regarded as an estimate of the actual current density since the spray coating will contain a certain surface ratio of inert particles (oxides and, in a way also pores). An accurate estimate of their influence proved very difficult to obtain. However, more than 10 percent of shift in current value is not expected.

Most obvious from figure 2 is the very high current values measured on the APS coating. A general multiplication of two decades as compared to the mass specimen can be observed. Also no evidence of significant passivity is present, leading to very poor expectations for the corrosion protection that this type of coating will provide.

The HVOF coating appears much more favourable. A passive region of almost 1000 mV in width, with a passive current below 10 µA/cm² can be considered very suitable for protection purposes. The general criterion for sufficient passivation requires a passive anodic current density equal or less than 100 µA/cm².

A close up of the passive region in figure 3 reveals that for the larger part of the passive region the anodic current density is considerably lower for the sprayed coating as compared to the NiCr mass material. After conclusion of this experiment, both mass specimen and spray coating showed evidence of pitting although the pits are very difficult to distinguish in the surface roughness of the sprayed coating.

Evaluation of Repassivation after Pitting

More information regarding the behaviour after pitting was acquired with the following experiment: 30 seconds of polarisation in the middle of the passive region (+100 mV vs Ag/AgCl) directly followed by 30 seconds at a potential anodic to the pitting potential (+800 mV vs Ag/AgCl) during which the required current was measured.The same potential values were used for both the HVOF and the mass specimen since both corrosion potential and pitting potential (although this value is not as clearly defined for the mass specimen) are not significantly different as can be observed from their polarisation curves (figure 2).

For comparison the results of the same experiment on the APS system are also included in figure 4. This figure shows the typical response of the three systems to the potential jumps.

Coming from their respective "passive" regions (the quotes are intended for the APS material) the current increases to about 5 mA/cm² for each system when the potential is increased to 800 mV (vs Ag/AgCl). The rate with which this increase occurs differs for the HVOF coating significantly from the other two systems. Both the APS sprayed coating and the mass material respond almost instantaneously to the potential jump whereas the HVOF coating requires almost 10 seconds to reach the maximum current value. At reversal of the potential to 100 mV (vs Ag/AgCl), a similar response is witnessed. Both APS coating and mass material jump directly to their respective minimum values. The HVOF coating responds more slowly but the value of the passive current is almost one, and two decades lower than for the mass material and the APS coating respectively.

Potentiodynamic Polarisation Measurements in Sulphuric Acid

The general corrosion behaviour of the NiCr samples was studied in sulphuric acid. The absence of the development and growth of localised corrosion pits can clearly be observed from the polarisation curves in figure 5. The almost perfect duplication of the current - potential relationship after reversal of the scan direction indicates no geometrical change of surface structure (pitting, crevice corrosion etc).

The corrosion potential (i=0) and also the potential at the onset (passivation potential) and end of the passive region (transpassive behaviour) are very similar for the three systems.

Large differences can be observed for the passive current on the APS coating as compared to both other systems. Similar to the results obtained in seawater, the measured passive current is almost two decades higher than the NiCr mass material.

The critical current required for passivation is however quite similar for the APS coating and the mass system. But the corrosion protective quality of the passive layer on the APS coating still remains considerably poorer than of the much denser layers (mass and HVOF).

The HVOF has another remarkable characteristic as can be observed from the magnification in figure 6. The critical anodic current density for this coating remains almost two decades lower than for both the APS and the mass material. The resulting anodic current in the passive region is comparable to the mass specimen. A low critical current density required for passivation is a very favourable characteristic since it limits the risk of corrosion phenomena when the passive layer is damaged. A low critical current ensures rapid passivation, making this coating the preferred choice in this environment.

Potential Monitoring in Natural seawater

During exposure in slow flowing natural seawater the open circuit potential is monitored. The development of this potential during 300 hours can be viewed from figure 7.

Figure 7 Open circuit potential during exposure in slow flowing natural seawater.

Most apparent from this figure is that all systems display an increasing potential to considerably more noble values than that were measured in the polarisation experiments. This must partly be attributed to the non-equilibrium situation during the, relatively fast polarisation measurements, and the development of microfouling slimes which are known to have this effect on the corrosion potential of exposed metals [6,7].

The close up in figure 8 shows the relative stability of the potential signal during exposure.

The APS coating remains fairly stable on a steady decline after the initial increase during the first few hours of exposure. The mass specimen potential increases during exposure to more noble values with the occasional, slow, dip of about 50 mV. The nature of these dips indicating a more stable passive layer [8]. The HVOF coating shows the most activity during exposure. With fairly regular intervals of about 10 hours the potential suddenly drops to values also about 20 to 100 mV below the nominal potential value, only to recover steadily in the following 5 hours.

Combining the results from the static polarisation curves with this system suggests that the slower pitting and repassivation of the HVOF material allows us to see this phenomenon during potential monitoring. The much more rapid pitting-repassivation reactions on the APS and mass sample make the results of the relative low frequency (1 measurement each 10 minutes) potential monitoring on these samples appear much more stable.

Despite the high activity, the potential of the HVOF coated sample remains the most noble of the three, indicating that the high activity does not weaken the corrosion protection properties of the passive film. The slight increase of the potential over time would sooner suggest a strengthening of the passive film, although the interaction of the rough surface with microfouling slime could also have this effect.

A possible explanation for the noble potential could originate from the very high amount of oxides throughout the film. These high temperature oxides will be practically inert in the seawater environment, leading to a relatively small active surface. Judging from the results in the polarisation curves, the local current density rapidly reaches a value high enough for passivation.

During exposure, the local activation is stifled by the presence of the high temperature oxides that may act as a layered shield throughout the coating, or by the influence of the oxides on the formation (nucleation, growth and adhesion) of the passive film on the metallic NiCr core of the splat particles [9,10].

The exact nature of these interactions in passivating thermal spray coatings will be subject of future research.

From the initial results of the electrochemical investigation of the corrosion properties of thermal sprayed Ni80Cr20 coating as compared to the behaviour of "standard" mass material the following conclusions can be drawn:

The atmospheric plasma sprayed coating showed poor passivation in natural seawater which indicated poor protective properties in this environment. This system did show passivity in sulphuric acid but the resulting passive current was almost two decades higher than the mass material.

However, the coating sprayed with the high velocity oxygen fuel technique, although much more oxidised during the spray process, did show good passive behaviour in natural seawater. Also the behaviour of the sprayed coating after the onset of pitting was at least comparable with the mass material. In sulphuric acid the HVOF coating even showed a critical anodic current for passivation that was almost two decades lower than the mass material.

In contrast to popular opinion, the study indicates that the thermal spray application of passivating coatings with a high amount of oxidation may improve the resistance to sea water corrosion. The presence of high temperature oxide particles trapped in the coating seems to contribute to the formation and stability of the passive film. Further research is in progress to establish if it is possible to optimise the thermal spray process to produce coatings with superior properties over mass material of nominally the same composition.

1. The merits of thermal spraying, discussion on the Internet, Ed. M.P.W. Vreijling, http://www.tno.nl/instit/kribc/ca-den_helder/casprdis.html, 1996
2. H. Herman, S. Sampath, Thermal Spray Coatings, http://dol1.eng.sunysb.edu/ thermal/article1.html, 1996
3. AMDRY 918-2 Pure titanium powder for medical applications, SULZER METCO data sheet, 1996
4. E.P.M. van Westing, W.J.H. Wortelboer, G.M. Ferrari, F.P.E. Westendorp and F.P. Ijsseling Progress in the Understanding and Prevention of Corrosion, Ed. J.M. Costa and A.P. Mercer, publ. by the Institute of Materials (1993) 1032-1039. Barcelona 1993
5. L. Pawlowski The science and engineering of thermal spray coatings, Wiley, 1995
6. European federation of corrosion, Working Party on Marine Corrosion, Comparison of Seawater Corrosivity, 1993-1994
7. Subramanian, P. Chandrasekaran ea, Bulletin of electrochemistry 6 (6) June 1990
8. J.H. Gerretsen, J.H.W. de Wit, The passivation of Ni in 0.5M sulfuric acid, chapter 8, thesis TU-Delft 1990
9. M.L. Escudero, J.A. Gonzalez ea., In vitro corrosion behaviour of ceramic sprayed stainless steel prostheses, Br. Corros. J. Vol22, No 3, 1987
10. M.L. Escudero, thesis Universidad Autonoma de Madrid, 1985

• Comments on CR3C2 paper and my own work - Chris Strock 09:33:49 8/07/96 (0)
• Comments on CR3C2 paper and my own work - Chris Strock 09:33:04 8/07/96 (1)
  • Re: Comments on CR3C2 paper and my own work - Mark Vreijling 07:27:45 9/06/96 (0)