Microstructure characteristics and corrosion behaviour of HVOF sprayed metallic coatings

A J Sturgeon
TWI Ltd, Granta Park, Abington, Cambridge, UK

Paper AS153 presented at ITSC 2001 International Thermal Spray Conference, 28-30 May 2001, Singapore

Abstract

The application of HVOF spraying to deposit high quality coatings of corrosion resistant alloys for protecting an underlying steel substrate against corrosion in seawater has received much interest over the past few years. Despitethe attainment of low levels of porosity and oxide, the coatings to not appear to offer the same level of corrosion resistance as the corresponding bulk materials. The aim of the work reported here is to demonstrate the level ofcorrosion performance that can be expected from coatings of corrosion resistant alloys deposited using the HVOF spraying process. Three alloy types are considered, a stainless steel with a composition similar to 316L, a nickel alloywith a composition similar to 625 alloy, and commercially pure titanium.

Introduction

The use of corrosion resistant alloys such as stainless steels, nickel alloys or titanium as coatings to protect an underlying steel substrate has received much interest over the past few years. ^[1,2,3] This is in part due to the expectation that suitably low porosity coatings of these metallic alloys can be prepared using the HVOF process. Such coatings are considered for applications where a barrier layer is needed to protectagainst corrosion in seawater or corrosive solutions such as mineral or organic acids. The presence of porosity or other defects in the coating that provide a path that allows the seawater (or corrosive solution) to reach thesubstrate. Rapid attack of the substrate may then occur at this point. This may also lead to the situation where an electrochemical cell is set up with the coating acting as a large cathode driving rapid corrosion at localised (anode)sites on the steel substrate. The requirement therefore is to deposit coatings with low levels of porosity that is 'closed' in nature. The HVOF spraying process has been shown to deposit coatings of several alloy types, includingstainless steels and nickel alloys with both low levels of porosity (less than 2%) and very low levels of oxide (again less than 2%). This is a consequence of the higher particle velocities and relatively lower particle temperaturesobtained with HVOF spraying compared to most other thermal spraying processes.

However, even with the attainment of such low porosity microstructures, HVOF sprayed coatings do not appear to offer the same level of corrosion resistance as the corresponding bulk materials. This may be due to the moreinhomogeneous microstructures present in the sprayed coatings compared to the same material in bulk form. The coating microstructure is dominated by inter-particle (splat) boundaries, often depleted in alloy elements, and the presenceof thin oxide films at these (splat) boundaries.

The corrosion performance of coating systems in aqueous environments can be difficult to evaluate. Reliance on immersion methods to compare corrosion behaviour requires long test durations and a qualitative visual judgement of anycorrosion. Electrochemical test techniques are used to provide a quicker and more qualitative tool for evaluating and comparing the corrosion behaviour of bulk alloy materials in aqueous environments. In seawater and dilute acidenvironments, 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 G61 describes a procedure for conducting cyclicpotentiodynamic polarisation measurements to determine relative susceptibility to localised corrosion of bulk iron or nickel based alloys in chloride containing environments. More recently such techniques have also been applied tothermal sprayed metallic coatings ^[4,5] to provide a relatively quick method to rank their resistance to corrosion.

The aim of the work reported here is to compare the level of corrosion performance that can be expected from coatings of corrosion resistant alloys deposited onto a steel substrate using the HVOF spraying process. Three alloy typeswere considered, a stainless steel with a composition similar to 316L, a nickel alloy with a composition similar to 625 alloy, and commercially pure titanium (CPT). Coatings of each material type were prepared using two particle sizeranges. These coatings were prepared within a larger activity, not reported here, that optimised their corrosion resistance in aqueous (sea water and dilute acid) environments. The cyclic potentiodynamic polarisation method was used toexamine the corrosion behaviour of these coatings and the same alloys in bulk form.

Experimental procedure

Coating preparation and characterisation

Coatings of stainless steel, nickel alloy and commercially pure titanium were sprayed onto low carbon (0.14% C) steel using the high pressure JP5000 HVOF system (Praxair-Tafa, Concord, USA). Each of the three coating types weresprayed using different parameter settings developed within a 'design of experiment' approach to maximise their corrosion resistance (not reported). For each coating type, two powder size distributions were used. All powders werespherical in shape and prepared using atomisation techniques. The composition and size distributions of these powder consumables are given in Table 1.

Table 1: Powder particle size range and composition

Test pieces of low carbon steel with dimensions 25mm x 25mm were coated to a thickness of about 300µm. Cross sections of each coating were prepared and examined by optical and scanning electron microscopy. The level of porositywas measured from optical images of the cross-section at x400 magnification, using quantitative image analysis equipment. For each coating three measurements were made and a mean value calculated. The oxygen content was measured onsamples of coating detached from the substrate and ground to a powder. 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 theassumption that the detected oxygen was associated with the presence of Cr ₂O ₃ for the stainless steel and nickel alloys, or TiO ₂ for the titanium coating.

Electrochemical testing

The corrosion behaviour of the coated test pieces was evaluated using an electrochemical test method similar to that described by the ASTM standard G61 for bulk iron and nickel based alloys. The coating surface was tested in theas-sprayed condition, immediately following a detergent wash, water rinse, acetone degrease and drying in air. The electrochemical corrosion test comprised of cyclic anodic polarisation in artificial seawater solution (3.5% NaCl)purged with nitrogen, with a pH of 8.2 and at a temperature of 25°C.

A 25x25mm coated test piece was clamped to the bottom of an Avesta-type cell arrangement, with a 1cm ² diameter area of the coating in contact with a 250 ml volume of de-aerated seawater solution. The cell seal design allowed flushing with purified water to minimise crevice corrosion problems, associated with thecell edge contacting the coating. A schematic of the test cell is shown in Figure 1. After a stabilisation period of 2 hours, the corrosion potential was measured relative to a reference saturated Calomel electrode (sce). The coated sample was then anodically polarised from the corrosion potential ata rate of 10 mV.min ^-1, whilst measuring the corrosion current to a platinum counter electrode. On reaching a corrosion current of 10mA, the applied potential was reversed and scanned back down to the corrosion potential. A plot ofcorrosion current in mA from the 1cm ² test area was obtained as a function of the applied potential (in mVsce).

Polarisation plots were also obtained for the low carbon steel substrate and bulk samples of stainless steel (316L type), Ni alloy (625 type), and commercially pure titanium, all with similar composition to the respective coatingmaterials. The surface of the bulk alloys was first roughened using 600 grit SiC paper and then degreased as described above for the coatings.

In the electrochemical polarisation test, the corrosion current density is a measure of material dissolution from the surface being tested. A rapid increase in corrosion current during the test is often associated with theinitiation and propagation of localised corrosion due to the formation of pits or the presence of crevices. For the low porosity coatings prepared in this work it is assumed that the measured corrosion current originates primarily fromthe coating. In an actual service environment the coating may be exposed to anodic potentials that could reach 400mV for natural seawater with the presence of a bio-film, and possibly up to 600mV in a chlorinated seawaterenvironment.

Corrosion potential at the end of the stabilisation period and the anodic current at potentials of 100 and 400mVsce for the forward scan were used as measures to compare the corrosion performance of the different coatings and bulkalloy materials. Coatings with higher (more noble) corrosion potentials, and low corrosion currents at the selected potentials, were taken to have better resistance to corrosion and to provide greater protection to the underlying steelsubstrate. In interpreting the polarisation plots for corrosion resistant alloys in bulk form, the presence of a corrosion current above 0.01 mA.cm ^-2 is often taken to indicate the onset and progression of localised pitting or crevice corrosion attack at the surface of the material being tested.

Results and discussion

Coating microstructures

The measured porosity and oxide levels for coatings are given Table 2. These results show that coatings were prepared with low levels of porosity, at or below 5vol%. The actual levels of porosity are different for the three material types and depend on the powder size distribution usedfor each material. For the three coating materials, the smaller powder size range gave coatings with lower porosity. Similarly, the oxide level in each coating type is very dependent on the powder size range, with the smaller sizegiving a noticeably higher oxide content in the prepared coatings.

Table 2: Coating properties and corrosion test results

The stainless steel coatings had the lowest porosity levels at 0.9 and 1.5 vol% for the two powder size. Oxide levels in these coatings showed a wide range, from 0.8wt% (larger powder size) and up to 2.9wt% (smaller powder size).The measured level of porosity in the nickel alloy coatings was slightly higher at 2.2 and 2.5vol%, while the oxide levels were quite similar to those in the stainless steel coatings. The smaller powder size for the nickel alloy gavean oxide content of about 3.3wt%. The level of porosity in the titanium coatings was 2vol% with the smaller powder size and 5vol% with the larger powder size. The smaller titanium powder size produced a particularly high oxide contentin the coating, measured at about 11.7wt%, compared to 1.4wt% with the larger powder size.

Cross sections of all the coatings are shown Figure 2. These figures illustrate the lamella nature of the coating microstructures, with the deformed powder particles (splats) and inter-particle boundaries clearly visible. The figures also show the presence of oxide inthe coating microstructures (dark contrast phase), and reveal that most of this oxide is located at the inter-particle boundaries. The stainless steel, nickel alloy and titanium coatings prepared with powders having the smaller powdersize contain higher amounts of oxide at the inter-particle boundaries.

Values for the measured corrosion potential and corrosion currents at 100 and 400 mVsce are given in Table 2. Examination of these results reveal that for the three coating types, the stainless steel coated test pieces gave the highest corrosion currents, measured at 4.3 and 7.7 mA.cm ^-2 (100mVsce). The titanium coated test pieces exhibited corrosion currents an order of magnitude lower than those for the stainless steel coatings, measured at 0.15 and 0.69 mA.cm ^-2 (100mVsce). These values for corrosion current are quite high and well above the value of 0.01 mA.cm ^-2 usually taken to indicate the onset of localised attack (pitting or crevice). However, the lowest corrosion currents and least negative corrosion potentials were measured for the nickel alloy coated testpieces. Depending on the powder size used, these coatings had a corrosion current of 0.006 and 0.024 mA.cm ^-2, an order of magnitude lower than those for the titanium coatings. This result is also demonstrated by Figure 3, which shows a proportion of the forward scans for these coatings.

The 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 of a steel substrate, thansimilar high quality coatings of stainless steel or even titanium.

This work also indicates that for the nickel alloy, a smaller size powder gives a coating with noticeably better corrosion resistance. For example, the nickel alloy coating prepared with the smaller sized powder (Ni1) gave acorrosion potential of -125 mVsce and a corrosion current of 0.006 mA.cm ^-2 (100mVsce), compared to -375 mVsce and 0.024 mA.cm ^-2 for the coating prepared with the larger particle size (Ni2). This improvement in corrosion performance is achieved despite the smaller powder size giving a much higher oxide content in the deposited coating.This is an interesting observation because it is often argued that higher levels of oxide formation are detrimental to good corrosion resistance. The oxide formation removes alloying elements, such as Cr, from the alloy material makingit more susceptible to corrosive attack. Measured values of porosity are similar for the nickel alloy coatings prepared with the small and larger powder sizes. It appears that a smaller sized powder is able to produce a coatingmicrostructure more resistant to localised corrosion attack, possible due to differences in the nature of porosity in the coatings or in the nature of the inter-particle boundaries.

Titanium in bulk form would be expected to show better resistance than nickel alloy 625 to localised corrosion in seawater. The results reported here show that a titanium coating (Ti1) with a reasonably low porosity of about 2%,achieved using the smaller size powder, did not produce a more corrosion resistant coating than a nickel alloy coating with similar porosity levels (Ni1). This may be due to the titanium coating containing much higher levels of oxide.For pure titanium, the formation of oxide will clearly not deplete the surrounding material of alloying elements (they are not present). The occurrence of the titanium oxide itself must be detrimental to the corrosion resistance of thealloy. When the level of oxide in the titanium coating was reduced, by using a larger powder size, the porosity in the coating (Ti2) increased to 5vol%. This coating showed an increase in measured corrosion current, possibly as aresult of the higher porosity level that may have exposed the underlying substrate.

Comparison with bulk alloys

It is worthwhile comparing the corrosion performance of the best coating achieved in this work (Ni1) with that for the three coating materials in bulk form. Measured values for corrosion potential and corrosion current can be foundin Table 2. A proportion of the forward scans obtained for the nickel alloy coating Ni1 and for stainless steel, nickel alloy and titanium in their bulk form, together with that for the uncoated steel substrate are shown in Figure 4. These results show that a Ni alloy coated steel substrate has a much lower corrosion current and consequently is believed to have better corrosion resistance than the uncoated substrate material. But, the corrosionresistance of the nickel alloy coating does not match that of the same nickel alloy in its bulk form. The nickel alloy coating gives much higher corrosion currents than the bulk nickel alloy, and has values slightly higher than thosemeasured for bulk stainless steel (below its pitting potential). A steel substrate with a HVOF sprayed nickel alloy coating should be considered as having the ability to offer corrosion resistance similar to bulk stainless steel, butnot that of bulk nickel alloy.

Fig. 4. Forward polarisation scans for Ni alloy coating Ni1 and bulk alloys

Summary and conclusions

HVOF sprayed coatings of nickel alloy provided significantly better resistance to corrosion in a seawater environment than HVOF sprayed coatings of stainless steel or titanium.

For the nickel alloy coating, a smaller powder size range of 15-45µm gave a coating with better resistance to corrosion than a coating prepared using a larger powder size, despite the coating having a higher oxide content.

A steel substrate coated with HVOF sprayed nickel alloy may have the ability to offer corrosion resistance (in a seawater environment) similar to bulk stainless steel, but not that of the nickel alloy in bulk form.

Acknowledgements

The author would like to thank BP-Amoco, Shell, Petrobras, Marathon Oil, US and UK Navies, IHI, and Sulzer Metco for their support of this work.

References

1. S. Kuroda et al: Microstructure and corrosion resistance of HVOF sprayed 316L stainless steel and Ni base alloy coatings, Thermal Spray - Surface Engineering via Applied Research, Ed. CC Berndt, Pub. ASM International, (2000)
2. A.H. Dent et al, J. Thermal Spray Technol., 8, 399-404 (1999)
3. K. Ishikawa et al, Thermal Spray Technol., 8, 273-278 (1999)
4. C. M. Eminoglu et al: Potentiodynamic corrosion testing of HVOF sprayed stainless steel alloy, Tagungsband Conference Proceedings, Ed. E. Lugscheider, P.A. Kammer, Pub. DVS, Germany (1999)
5. A.J. Sturgeon and D. Buxton: The electrochemical behaviour of HVOF sprayed coatings, Thermal Spray - Surface Engineering via Applied Research, Ed. CC Berndt, Pub. ASM International, (2000)