Advanced Thermal Spray for Marine Corrosion Protection

Paper presented at Eurocorr 2003 Conference, Budapest, Hungary, 28 Sept - 2 Oct 2003.
Published in: Surface Engineering Vol.20, No.4, p.237-243. August 2004.

The microstructure and aqueous electrochemical corrosion characteristics of stainless steel 316L (UNS S31603) and nickel alloy 625 (UNS N06625) coatings deposited onto a carbon steel substrate using commercial High Velocity Oxy-Fuel(HVOF) spraying systems and a new high velocity wire flame spraying system have been assessed. The electrochemical corrosion behaviour of the as-prepared coatings has been compared against their respective materials in wrought form.

A HVOF sprayed coating of nickel alloy 625 was found to be more corrosion resistant in an aqueous saline environment than a coating of stainless steel 316L. However, the nickel alloy coating did not match the corrosion resistance of the same nickel alloy in wrought form, but may have the ability to offer corrosion resistance in saline aqueous environments approaching that of a wrought stainless steel.

A porosity level of less than 2.5% may not show significant effect on the electrochemical activities of the HVOF sprayed coatings near the free corrosion (rest) potential. However, the level of oxidation depending upon the spraying system was shown to have an effect on the corrosion resistance of the sprayed coatings. The stainless steel and nickel alloy coatings produced using the high velocity wire flame spray system displayed much higher oxides compared to those produced using the HVOF processes and consequently, poorer corrosion resistance.

The authors are in the Metallurgy, Corrosion, Arcs and Surfacing Technology Group, TWI Ltd, Granta Park, Great Abington, Cambridge, CB1 6AL, United Kingdom ( suman.shrestha@twi.co.uk)

1. Introduction

The use of thermal sprayed coatings (TSCs) of corrosion-resistant alloys such as stainless steels or nickel alloys to protect an underlying steel substrate has received much interest over the past few years. Such coatings are believed to be suitable for applications where a barrier layer is needed to protect steel components or structures against corrosion in seawater or corrosive solutions such as mineral or organic acids. This interest is in part due to the expectation that very low porosity coatings of these metallic alloys can be prepared using the High Velocity Oxy-Fuel (HVOF) spraying process. However, HVOF sprayed coatings have different microstructures compared to their wrought counterparts. They consist of deformed particles in a lamellar structure dominated by interpaticle boundaries. And as such they represent complex materials from an electrochemical point of view even with the absence of interconnected porosity. Moreover, loss of the alloying elements in the coating due to possible evaporation or thermal degradation of the starting powder during spraying can result in lowering the coating performance in corrosive environments. Indeed several studies by the authors have demonstrated on a variety of TSCs that the inherent corrosion resistance of the coating is a factor which must be considered when coatings are to be employed in aqueous systems ^[1-4] .

The corrosion performance of coating systems in aqueous environments can be difficult to evaluate. Reliance on immersion methods to compare corrosion behaviour requires a long test duration (often 60 days or longer) and aqualitative visual judgement of any corrosion attack. Electrochemical test techniques are used to provide a quicker and more quantitative tool for evaluating and comparing the corrosion behaviour of wrought alloy materials in aqueous environments. In seawater and some dilute acid environments, localised pitting and crevice attack of corrosion resistant alloys are usually of concern and can lead to the breakdown of corrosion resistance. The ASTM Standard G61describes a procedure for conducting cyclic potentiodynamic polarisation measurements to determine relative susceptibility to localised corrosion of wrought iron or nickel based alloys in chloride-containing environments. More recently, such techniques have also been applied to thermal sprayed metallic coatings ^[1,2,5-7] to provide a relatively quick method to rank their resistance to corrosion.

Although HVOF spraying is often considered for the deposition of high quality coatings, this has some disadvantages. Current HVOF systems are primarily designed for automated use within controlled thermal spray booths and arelimited to powder based consumables which are often more expensive than wire materials. In addition, HVOF systems are usually expensive and are not suitable for on site application. So there exists potential for a new process such asthe high velocity wire flame spraying process due to its portability and low coating deposition costs and hence a need for an evaluation of such a new system.

In this paper, the microstructure and electrochemical corrosion characteristics of stainless steel 316L and nickel alloy 625 coatings produced with three commercial powder based HVOF systems and a new high velocity wire flame spray system have been assessed. The electrochemical corrosion behaviour of the as-prepared coatings have been compared against their respective materials in wrought form using the cyclic potentiodynamic polarisation method in an aqueousNaCl solution. The paper focuses on the electrochemical corrosion characteristics of the coatings in terms of the various spraying systems employed.

2. Experimental procedure

Coating preparation and characterisation

Coatings of stainless steel 316L and nickel alloy 625 were sprayed onto low carbon (0.14% C) steel using the four spraying systems. These were the JP5000 (JP) and TopGun (TG) HVOF systems from Praxair Surface Technologies and the Diamond Jet Hybrid (DJ) HVOF system from Sulzer Metco, all using powder consumables, and the HVw2000 high velocity flame spraying system (HV) from HV Techno, LLC, West Lebanon, NH, USA using a wire consumable. The three HVOF systems differ quite significantly in terms of nozzle design and type of fuel gas used in the high-pressure combustion process. Detailed descriptions of these differences are presented elsewhere. ^[8] Gas atomised powders were used for the HVOF spraying. For the high velocity flame spraying, a 1.6mm diameter wire was used. A description of the high velocity wire flame spraying system can be found elsewhere. ^[9] Details of the spraying systems, fuel types and feed materials are given in Tables 1 and 2.

Table 1 316L stainless steel coatings

Table 2 Nickel 625 coatings

Test pieces of low carbon steel with dimensions 50mm x 50mm were coated to a thickness of about 300µm. Cross sections of each coating were prepared using standard metallographic techniques and examined by optical and scanning electron microscopy. The level of porosity was measured from the images of the cross sections using an image analysis software attached to the optical microscope. Three measurements were made for each coating, and a mean value was calculated. The oxygen content was measured in samples of coating detached from the substrate. The level of oxygen was analysed by the inert gas fusion technique, using Leco TC 136 equipment. Estimates of the oxide level in the coatings were calculated on the assumption that the detected oxygen was associated with the presence of Cr ₂ O ₃ .

Coating adhesion

Adhesion values of the HVOF sprayed coatings were measured in accordance with ASTM C633-79 (Re-approved 1993) 'Standard test method for adhesion or cohesive strength of flame-sprayed coatings'. Adhesion values of the high velocity wire flame sprayed coatings were measured in accordance with ASTM D4541-95 'Standard test method for pull-off strength of coatings using portable adhesion testers'. A high strength structural adhesive (type FM1000) was used to bond the steel dolly to the coating surface. For each coating type, the adhesion strength from five test pieces was measured and the average failure stress calculated.

Electrochemical Testing

The corrosion behaviour of the coated test pieces was evaluated using an electrochemical test method similar to that described in ASTM G61. The coating surface was tested in the as-sprayed condition, immediately following a rinse in alcohol and drying in air. The electrochemical corrosion test comprised cyclic anodic polarisation in a 3.5% NaCl solution of pH 8.2 and at a temperature of 25°C. The test solution was de-aerated continuously by purging with nitrogen.

A coated test piece (50mmx50mm) was clamped to the bottom of an Avesta cell with a 1cm ² area of the coating surface being in contact with a 250ml volume of de-aerated salt solution. The cell seal design allowed flushing with distilled water to minimise crevice corrosion problems associated with the cell edge contacting the coating. This set up was used for the HVOF sprayed coatings. For the high velocity flame sprayed coating, a different set up was used due to its rougher as-sprayed surface, which led to a leakage of the test solution in the Avesta cell. This set up consisted of a square coated specimen encapsulated in a non-conductive epoxy resin with 1cm ² coating surface being exposed. The coating/mounting resin interface was painted with a non-conductive lacquer to prevent penetration of the test solution via this interface to the underlying steel substrate.More details on the preparation of test specimens of the latter type can be found in the literature. ^[1,3] For both set ups the following electrochemical measurements were taken. After a stabilisation period of 2 hours, the free corrosion potential 'E _corr ' of the test surface was measured relative to a reference Saturated Calomel Electrode (SCE). The coated sample was then anodically polarised from its E _corr at a rate of 10mV.min ^-1 , whilst measuring the current density using a platinum auxiliary electrode. On reaching an anodic current density of 10mA.cm ^-2 , the applied potential was reversed back to E _corr . A plot of anodic current density in mA.cm ^-2 was obtained as a function of the applied potential (mV _SCE ). Polarisation plots were also obtained for the low carbon steel substrate and wrought samples of 316L stainless steel and nickel alloy 625, both with similar composition to their respective coating materials. The surface of the wrought alloy samples was abraded prior to testing on a 600-grit SiC paper and then degreased with alcohol, as were the coatings.

In the electrochemical polarisation test, the measurement of free corrosion potential E _corr was taken to indicate the relative activity of the surface exposed in a particular aqueous environment, with a more negative E _corr taken to indicate a more active surface. The anodic current density is a measure of material dissolution from the surface being tested. An immediate increase in anodic current upon polarisation from E _corr was considered as the rapid corrosion of the test surface (active surface). A rapid increase of the anodic current after a stable current density region (passivity) at a potential region further away from theE _corr is considered as the initiation and propagation of localised corrosion attack (such as pitting and crevice). This is believed to be due to the breakdown of the protective oxide film, and the potential value is often known as the breakdown potential 'E _b '. For the low porosity coatings prepared in this work, it was assumed that the measured anodic current originates primarily from the coating.

The corrosion potential at the end of the stabilisation period and the anodic current density at potentials of 0 and 100mV _SCE for the forward scan were used to compare the corrosion performance of the different coatings and wrought alloy materials. Coatings with higher (more noble) corrosion potentials, and lower anodic currents at the selected potentials, were taken to have better resistance to corrosion and expected to provide greater protection to the underlying steel substrate.

3. Results and discussion

3.1 Coating microstructures

Measured porosity and oxide levels for the stainless steel coatings are shown in Table 3 and in Table 4 for the nickel alloy coatings. These results show that coatings were prepared with low levels of porosity below 2.5%. There was a larger variation in the level of oxide in the coatings, ranging from about 26wt% down toless than 1wt% depending on the material type and spray system.

Table 3 Properties of 316L stainless steel coatings

¹ Oxide level calculated from the measured oxygen content and on assumption that the oxide is Cr ₂ O ₃

For the stainless steel coatings, the TG HVOF system produced the coating with the highest oxide level and lowest amount of porosity. A cross section of the stainless steel coating prepared using the TG system is shown in Fig.1a. This image shows that the powder was highly melted during the spraying process to give a lamellar microstructure with oxide stringers (darker contrast phase) aligned parallel to the substrate surface. The stainless steel coatings prepared using the JP and DJ HVOF systems both had much lower levels of oxide, but slightly higher amounts of porosity than the TG coating. Cross sections through these coatings are shown in Figs 1b and 1c. The SEM image of the DJ coating in Fig.1c shows an interfacial separation between the coating and the substrate. This resulted during sectioning of the coated specimen for a metallographic preparation. The coating microstructure obtained with the JP and DJ HVOF systems appears to consist of predominately well-stacked, partially deformed particles. This microstructure type suggests the powder particles were at a lower temperature on impact with the substrate, possibly below their melting point, when sprayed using these two HVOF systems. The coating produced using the wire consumable (HV-316) is shown in Fig.1d, which also shows a dense lamellar coating microstructure having a uniform distribution of bright and light grey ribbon-like layers and was similar in microstructural appearance to the TG coating in Fig.1a. The data in Table 3 shows the wire flame sprayed coating displays the highest level of oxides present for all the stainless steel coatings at 26wt%.

For the nickel alloy, the TG HVOF system again produced a coating with an oxide content higher than the coatings sprayed using the JP and DJ HVOF systems. However, unlike the stainless steel coatings, there was little difference in the porosity levels. Porosity levels of about 2% were obtained with all three HVOF systems. The HV flame sprayed coating again showed an extremely high level of oxide content at 20wt%. Cross sections through nickel alloy coatings are shown in Fig.2. The microstructure of the coating deposited with the TG HVOF system again has a lamellar appearance with many oxide stringers. The microstructures obtained for the coatings sprayed using the JP and DJ HVOF systems appear to consist mostly of well-stacked and partially deformed particles. The DJ HVOF coating also has a noticeable layered structure, with the layers parallel to the substrate separated by a darker contrast phase presumed to be oxide. Each layer is believed to represent one pass of the spray gun over the surface.

The adhesion values measured for the HVOF sprayed stainless steel and nickel alloy coatings, following the ASTM C-633 standard, were considered good and in excess of 70MPa with failure occurring at the coating to substrate interface. The exception was the stainless steel coating prepared using the JP HVOF system, which gave lower adhesion values of about 50MPa with the coating being removed from the substrate. The adhesion values (to ASTM D4541) of the stainless steel and nickel alloy coatings produced by the HV flame spray system were 25-30MPa.

3.2 Electrochemical Corrosion Behaviour

The measured free corrosion potentials and anodic current densities at various potentials are given in Table 3 for the stainless steel coatings, Table 4 for the nickel alloy coatings and Table 5 for the same materials in wrought form. In addition, values are also given for the low carbon steel substrate. The forward scans of the polarization curve for the stainless steel coatings and for wrought stainless steel are shown in Fig.3. Likewise the forward scans for the nickel alloy coatings and wrought nickel alloy are shown in Fig.4. Only the forward scans are shown for clarity.

Table 4 Properties of nickel alloy coatings

Table 5 Corrosion performance of wrought materials

Stainless steel

The HVOF stainless steel coatings in Fig.3 displayed very similar E _corr values ranging between -461 to -514mV _SCE . The JP and DJ HVOF coatings exhibited a stable current density region (passivity) upon anodic polarisation up to a potential value between -60 to 0mV _SCE . Whereas, the TG HVOF and HV flame sprayed coatings displayed a more rapid increase in their current density, suggesting more rapid corrosion was occurring over the surface of the latter two coatings. The passive potential range followed by a rapid increase in the current density near about -60 to 0mV _SCE (breakdown potential 'E') displayed by the JP and DJ HVOF coatings indicates that these two coatings may be susceptible to localised forms of corrosion attack such as pitting. Moreover, the polarisation plots suggest that a coating porosity of less than 2.5% does not appear to have significant effect on the measured free corrosion potentials of the coating/carbon steel combinations during immediate anodic polarisation tests. However it is possible that the effect over longer exposures may reveal the differences in the coatings' barrier capability. Differences in the coatings were shown by the amount of oxides present in the coatings, with those having fewer oxides displaying better corrosion performance as shown by the JP and DJ coatings.

Nickel alloy

The anodic polarisation plots for the nickel alloy coatings are presented in Fig.4, which again displayed similar E _corr values for most of the sprayed coatings except for the TG HVOF coating, which showed a more negative E _corr . Once again the JP and DJ HVOF coatings displayed very similar anodic curves suggesting that these two coatings may have similar anodic activities occurring at the exposed surface. These two coatings also displayed a low current density region over a large potential range (400mV), where this was less than 0.01mA.cm ^-2 and can be considered as passive. This suggests that the coatings are resisting corrosion and protecting the underlying steel substrate during the short-term anodic polarisation tests. Once again the TG and HVflame sprayed coatings displayed a poorer corrosion resistance in the saline aqueous environment. This was shown by the HV flame sprayed coating displaying an increase of the anodic current density upon immediate polarisation from its E _corr and the TG coating displaying a more negative E _corr . It is believed that the corrosion behaviour of the nickel alloy coatings (in a similar manner to the 316L coatings) was affected by the amount of oxides present, with JP and DJ coatings having fewer oxides demonstrating better corrosion resistance during the anodic polarisation tests.

The data in Tables 3 and 4 suggest that for the two coating types, the stainless steel coatings gave much higher anodic current densities, measured up to 5mA.cm ^-2 (at 0mV _SCE ). Lower anodic current densities and less negative corrosion potentials were measured for the nickel alloy coated test pieces. Depending on the spray system, these coatings had anodic current densities of0.001-0.07mA.cm ^-2 (at 0mV _SCE ) and 0.004-0.08mA.cm ^-2 (at 100mV _SCE ), typically two orders of magnitude lower than those for the stainless steel coatings. These results suggest that high quality, HVOF sprayed coatings of nickel alloy with composition similar to alloy 625 can provide significantly better corrosion resistance and consequently better protection to a steel substrate, than similar high quality HVOF coatings of stainless steel.

3.3 Comparison with wrought alloys

In similar electrochemical tests, the wrought 316L stainless steel showed a plot typical for this material ( Fig.3) with a more positive E _corr value of about -200mV compared to about -450mV for the sprayed coatings, suggesting that the wrought materials are nobler than the sprayed coatings. The anodic current density initially increased slowly and did not exceed 0.003mA.cm ^-2 over a large potential range upon polarisation from its E _corr until a breakdown potential (Eb) at about 300mV _SCE . All the HVOF sprayed coatings of stainless steel exhibited a significantly lower rest potential and showed a much more rapid rise in anodic current density as the potential was increased. The coatings(DJ-316 and JP-316) also showed an apparent breakdown potential near 0mV _SCE , considerably lower than that seen for the wrought alloy.

The wrought nickel alloy exhibited a similar E _corr value to that shown by the sprayed coatings. The wrought alloy also displayed a gradual increase in anodic current density as the potential was raised from E _corr . This is shown in Fig.4. All the nickel alloy coatings showed higher anodic current densities at potentials up to about 400mV _SCE than the nickel alloy in wrought form by an order of a magnitude. The more heavily oxidised coatings of this material prepared by the TG HVOF system had a considerably more negative corrosion potential and much higher anodic current density.

A part of the forward scan obtained for one of the better nickel alloy coatings (JP-625) together with those for wrought stainless steel 316L, wrought nickel 625 alloy and for the uncoated carbon steel substrate are shown in Fig.5. These results indicate that the nickel alloy coating on a carbon steel substrate has a much lower anodic current density and consequently can have better corrosion resistance than the uncoated substrate material.However, the corrosion resistance of the nickel alloy coating does not match that of the same alloy in its wrought form. The nickel alloy coating gives a much higher anodic current density than the wrought nickel alloy, and has values slightly higher than those measured for the wrought 316L stainless steel (below its E _b ). A steel substrate with a HVOF sprayed nickel alloy coating is believed to have the ability to offer corrosion resistance in a saline aqueous solution similar to wrought 316L stainless steel, but not that of the wrought nickel alloy.

3.4 Corrosion mechanism

Similar free corrosion potentials displayed by the different HVOF sprayed stainless steel coatings and also by the different HVOF sprayed nickel alloy coatings indicate that for these coatings the effect of coating porosity on the electrochemical activities at the initial stage of aqueous immersion is negligible. It has been reported in the past that the level of porosity in the HVOF sprayed coatings e.g. WC-CoCr ^[3] and NiCrSiB ^[10] is very low and therefore, corrosion by penetration of the aqueous media via interconnected porosity is not the main concern.

Although, not investigated in this study, it is believed that the reduced corrosion resistance displayed by these coatings in contrast to their respective wrought counterparts is due to their complex microchemistry/microstructure.One possible explanation is due to a depletion of the alloying elements in the coating that confer corrosion resistance e.g. chromium. The depletion of chromium in the particle splats of the HVOF sprayed NiCrSiB coatings compared tothe original feed stock powders has been reported in the literature ^[10] . A different mechanism 'macropitting' due to a splat particle removal caused by a preferential corrosion attack at the splat boundary and a further corrosion acceleration of the already attacked splat with respect to the adjacent splats due to complex micro-galvanic interactions is another possibility. ^[10] The 'macropitting' may allow the liquid phase to penetrate through the coating to cause corrosion and coating adherence problems at the coating/substrate interface.

4. Conclusions

1. A steel substrate coated with HVOF sprayed nickel alloy 625 may have the ability to offer corrosion resistance (in saline solution) approaching wrought 316L stainless steel, but not that of the nickel alloy in wrought form.
2. HVOF sprayed coatings of nickel alloy 625 can provide better resistance to corrosion in a saline aqueous environment than coatings of stainless steel 316L.
3. The stainless steel 316L and nickel 625 coatings produced using the high velocity wire flame spray process result in extensive oxidation and consequently lower adhesion to the substrate and poorer corrosion resistance in the saline solution compared to those produced with the HVOF process.

5. References

1. S. Shrestha, A. Neville and T. Hodgkiess: J. Therm. Spray. Technol., 2001, 3, (10), 470-479.
2. S. Shrestha, A. Sturgeon, T. Hodgkiess and A. Neville: Tagungsband Conf. Proc. 2002, (ed. E. Lugscheider), 692-697; DVS Deytscher Verband fur Schweisen.
3. A. Neville and T. Hodgkiess: Surf. Eng., 1996, 4, (12), 303-312.
4. H. Edris, D. G. McCartney and A. J. Sturgeon: J. Mat. Sci., 1997, 32, 863-872.
5. S. Kuroda et al: Thermal Spray - Surface Engineering via Applied Research, C. C. Berndt ed., Pub. ASM International, 2000.
6. C. M. Eminoglu et al: Tagungsband Conf. Proc., E. Lugscheider and P. A. Kammer ed, Pub. DVS, Germany, 1999.
7. A. J. Sturgeon and D. Buxton: Thermal Spray - Surface Engineering via Applied Research. C. C. Berndt ed, Pub. ASM International, 2000.
8. H. Kreye, F. Gartner, A. Kirsten and R. Schwetzke: Proc. 5 ^th HVOF Spraying Colloquium, Erding, Pub. GTS, 2000.
9. H. Kreye, A. Kirsten, F. Gartner, X. Qi and W. Krommer: Proc. ITSC 2001, (ed. C. C. Berndt et al.) 461-466, 2001, Materials Park, OH, ASM International.
10. S. Shrestha: 'Corrosion and erosion-corrosion of HVOF thermal sprayed coatings and stainless steels', PhD Thesis, The University of Glasgow, Scotland, UK, 2000.