Microstructure formation and corrosion behaviour in HVOF-sprayed Inconel 625 coatings

Abstract

The nickel-based alloy Inconel 625 was thermally sprayed by two different variants of the high velocity oxy-fuel process. In this study, coatings deposited by a liquid-fuelled gun were compared with those produced by a gas-fuelled system; in general, the former generates higher particle velocities but lower particle temperatures. Investigations into the microstructural evolution of the coatings, using scanning electron microscopy and X-ray diffraction, are presented along with results on their aqueous corrosion behaviour, obtained from salt spray and potentiodynamic tests. It is inferred from coating microstructures that, during spraying, powder particles generally comprised three separate zones as follows: fully melted regions; partially melted zones; and an unmelted core. However, the relative proportions formed in an individual powder particle depended on its size, trajectory through the gun, the gas dynamics (velocity/temperature) of the thermal spray gun and the type of gun employed. Cr₂O₃ was the principal oxide phase formed during spraying and the quantity appeared to be directly related to the degree to which particles were melted. The salt spray test provides a sensitive means of determining the presence of interconnected porosity in coatings and those produced with the liquid-fuelled gun exhibited reduced interconnected porosity and increased corrosion resistance compared with deposits obtained from the gas-fuelled system. In addition, potentiodynamic tests revealed that passive current densities are 10–20 times lower in liquid-fuel coatings than in those sprayed with the gas-fuelled gun.

Introduction

In recent years there has been a growing interest in the use of high velocity oxy-fuel (HVOF) thermal spraying to deposit protective overlay coatings onto the surfaces of engineering components to allow them to function under extreme conditions [1]. In the HVOF process powder particles, typically in the size range 10–63 μm, are injected into a high temperature, high speed gas jet within a specially designed gun [2], [3], [4]. The jet is produced through the combustion of a fuel with oxygen at high pressures and flow rates within the gun. The powder particles typically attain velocities of 300–800 ms⁻¹ at the substrate to be coated whilst reaching temperatures which allow them to be molten or semi-molten prior to impact [5], [6]. By scanning the gun across the substrate, a coating layer of low porosity is built up from the impact, bonding and solidification of successive particles. In first generation HVOF systems the fuel is normally a gas (e.g. hydrogen or propylene) and the gun has a parallel sided barrel as shown schematically in Fig. 1(a). The operation of this type of gun is as follows. Powder is fed axially into the rear of the combustion chamber where oxygen and propylene are mixed and burn. The hot gas and entrained powder are then accelerated through the combustion chamber and down the parallel sided nozzle, which is approximately 120 mm long, before emerging as a free jet. The gas dynamics of such a system are described in detail elsewhere [7].

Second generation systems now being used are designed with a converging-diverging throat between the combustion chamber and the nozzle to substantially increase the gas velocity [2] and in some systems a liquid-fuel, namely kerosene, is employed. A schematic illustration of a liquid-fuelled gun is shown in Fig. 1(b). In this gun, kerosene and oxygen are fed into the combustion chamber where kerosene is vapourised, mixed with oxygen and the mixture burns before passing through the converging-diverging throat which accelerates the gas to a Mach Number between 1.5 and 2. Powder is fed radially into the gas stream through two ports which are located downstream from the throat. The gas and entrained powder then flow along a parallel sided nozzle, typically 100–200 mm long, before emerging as a free jet. In the nozzle of a liquid-fuelled system, the gas velocity is estimated to be ∼1700 ms⁻¹ and the gas temperature ∼2500 K [8]. These compare with gas velocity and temperature values of approximately 1300 ms⁻¹ and 3000 K respectively which have been calculated to develop in the nozzle of the gas-fuelled gun shown in Fig. 1(a) [7]. Thus, in comparison with a first generation gas-fuelled system, a liquid-fuelled gun is capable of generating higher particle velocities but without an attendant excessive rise in particle temperature [9]. In this paper, coatings deposited by a gas-fuelled gun will be compared with those produced by a liquid-fuelled one and so, for brevity, these processes will be referred to as HVOGF and HVOLF respectively.

HVOGF and HVOLF thermal spraying are both widely used to deposit cermet coatings (e.g. WC–Co) which provide enhanced wear resistance. However, increasing attention is now being paid to the spraying of high grade metallic alloys (e.g. Inconel, Hastelloy, Stellite or stainless steel types) [10]. The coatings deposited are designed to be used for protection of substrates in aqueous environments where electrochemical corrosion can occur. In a coating that is not galvanically sacrificial, e.g. an Inconel alloy on a mild steel substrate, it is believed that interconnected porosity and intersplat oxide formation are the principal microstructural features affecting the corrosion resistance [10]. Briefly, interconnected porosity can lead to penetration of the corrosive medium to the substrate causing localized attack and eventual debonding of the coating. The role of intersplat oxides, commonly Cr₂O₃ or a Cr-containing mixed oxide in these high grade corrosion resistant alloys, remains unclear although localized Cr-depletion of the metal matrix has been proposed as one cause of increased reactivity in addition to oxide-matrix interfaces.

To date, a number of studies have been undertaken on the corrosion resistance of Inconel, Hastelloy, stainless steel and other alloy coatings sprayed with HVOGF systems [11], [12], [13], [14], [15], [16], [17], [18], [19]. An early study [11] used potentiodynamic polarisation tests to investigate corrosion of Inconel 625 coatings both on mild steel and after detachment from the substrate. This showed that the corrosion current density of a coating on a substrate was significantly affected by interconnected porosity and that microstructural changes, arising from different spray parameters, also influenced corrosion rates. Studies by Dorfman and DeBarro [12] and Ishida [14] on various stainless steels using electrochemical and other tests also confirmed that interconnected porosity played an important role in determining overall corrosion rates in coating-substrate combinations. However, little attention was given to detailed aspects of coating microstructures, although Dorfman and DeBarro [12] did suggest that oxide particles or interparticle boundaries were likely sites of enhanced anodic action. Harvey et al. [13] and Normand et al. [15], [18] came to contradictory conclusions regarding the significance of oxide formation in influencing the corrosion rates of Inconel coatings as determined by potentiodynamic methods. The former concluded that oxide contents should be minimised to reduce corrosion rates whereas the latter demonstrated lowest passive current densities in a highly oxidised coating. However, in both studies important microstructural details such as type of oxide formed and the extent of interconnected porosity in different coatings were not reported. A general feature of the above studies and others [16], [17], [19] is that passive current densities of coatings on mild steel substrates in solutions such as 0.5 M H₂SO₄, as determined through direct current (DC) potentiodynamic tests, are 10²–10³ times greater than in equivalent wrought alloy samples. However, much remains to be investigated regarding the relative contribution to this increase of porosity and microstructure. Additionally, it is recognised that short-term, DC potentiodynamic testing is not well suited to evaluating the extent of substrate corrosion arising from corrodent penetration through interconnected porosity in the coating and that complementary techniques need to be employed to evaluate this.

Detailed studies on the corrosion of HVOLF-sprayed coatings, where powder particles acquire substantially more kinetic energy, without excessive thermal energy (i.e. temperature rise), are relatively scarce.

In a limited study, Dvorak and Heimgartner [20] found that the passive current density of a HVOLF-sprayed Inconel 625 coating on mild steel was around 100 times lower than that of a HVOGF-sprayed deposit. However, this was still approximately 10 times the value measured for the wrought alloy. Neville et al. [21] also studied a HVOLF-sprayed Inconel 625 coating and investigated the mechanism of attack in seawater. They reported an initial attack at particle boundaries and also within particles at regions with varying compositions on a 1–5 μm scale. Corrosion at the substrate by penetration of solution through interconnected porosity was reported to be not significant. However, Neville et al. [21] did not attempt to investigate the linkages between spray conditions, coating microstructure and corrosion performance, nor did they report passive current density values in relation to wrought material. Although the HVOLF spray process appears to show considerable promise in producing coatings with improved corrosion resistance much remains to be investigated, particularly the relationship between microstructure and corrosion behaviour.

Therefore, the principal aim of the present study was to examine the effect of feedstock powder size range on microstructure formation and the resultant corrosion behaviour in HVOLF-sprayed coatings of Inconel 625 using a fixed set of spray parameters. Two different commercially available powders, one of nominal size range (−45 to +15 μm) and the other of nominal size (−63 to +15 μm), were employed in the experiments. For comparison purposes, and to establish a baseline in corrosion behaviour, the HVOGF process was used to spray the (−63 to +15 μm) powder, again with a fixed set of process parameters. The resultant coatings were subjected to both detailed microstructural investigation and corrosion evaluation using salt spray tests and electrochemical methods.