Improvement of corrosion resistance of HVOF thermal sprayed coatings by gas shroud Jin Kawakita, Takeshi Fukushima, Seiji Kuroda, and Toshiaki Kodama National Research Institute for Materials Science Abstract The corrosion resistance of the coating, obtained by HVOF spraying with the shroud mechanism by the N 2 gas, was evaluated by the electrochemical techniques. This system could increase in flight velocity of spray particles and suppress their oxidation simultaneously, leading to attainment of both the smaller porosity and the lower oxide content of the resulting coating. The resistance of the HVOF sprayed coating against seawater corrosion was improved considerably by attachment of the gas shroud. Keywords: HVOF, HastelloyC, gas shroud, corrosion resistance, seawater Introduction In our national research project named as Ultra Steel for 21st Century, one of the subjects is to coat the offshore structural steels by anti-corrosion alloys with the thermal spray technique. Targets of the spray coating are the tidal and the splash zones of these structures, which are known as the severely corrosive environments. In such environments, a significantly high corrosion resistance is essential for protection of the structural steels. For this reason, a nickel base alloy, HastelloyC, was selected as the coating material because of its appropriate melting point its high resistance against both crevice and pitting corrosion. As for the thermal spraying, the high velocity oxy-fuel (HVOF) method was adopted so that this method can deposit a denser and a less oxidized coating than the conventional method such as the plasma spraying. The HastelloyC coating, prepared under the standard condition recommended by the apparatus manufacturer, had an open-porosity below the limit of detection by mercury intrusion porosimetry [1]. After the immersion test in artificial seawater, the HastelloyC coating itself showed no evidences of the corrosion damage while small rusts with no easy visibility appeared on the surface when this coating was deposited on a mild steel, SS400. This result indicated that the existence of a slight amount of through-pores, along which seawater could permeate the coating, leading to formation of the galvanic cell at the interface between the electrochemically noble coating and the less-noble substrate. Then, the various spray conditions were examined to establish the guideline for decrease in through-porosity of the coating down to zero. At the same time, the method for the higher precision detection of through-pores was devised and developed [2]. By raising the flame energy upon spraying, through-porosity of the HastelloyC coating was found to attain zero because no evidences for the existence of through-pores were detected even by using the mentioned-above method with the higher sensitivity by two orders of magnitude than mercury intrusion porosimetry. Simultaneously, however, the higher flame energy brought about the higher oxide content of the coating, leading to the decrease in corrosion resistance of the coating itself. In order to realize the compatibility of both the high flame energy and the suppression of oxidation, a shroud mechanism by an inert gas was devised and attached to the end of HVOF gun [3], as illustrated schematically in Fig. 1. The principle of its mechanism is that the flame is surrounded by the inert gas in the longer distance than as usual, and its advantages are 1) to accelerate and to cool flying spray particles and 2) to prevent the flying particles from
contacting the ambient air, in particular oxygen. In this paper, the HastelloyC coating deposited by HVOF spraying with the gas shroud was investigated in terms of its physical and its chemical properties. The corrosion resistance of this Fig. 1 Schematic illustration of gas shroud coating was examined with the attachment with HVOF spray gun. electrochemical techniques. In addition, this coating was compared with other coatings obtained before. Experimental Coating on the SS400 substrate was carried out by HVOF spraying (TAFA JP-5000) with kerosene as the fuel. The primary spray condition for spraying was listed in Table 1. The substrate with 2t 50 100 mm in size was blasted using alumina grits and was degreased ultrasonically in acetone before spraying. The feedstock was the powders of HastelloyC and SUS316L stainless steel. The latter was used for comparison. Their chemical compositions were presented in Table 2. The shroud gas of nitrogen was injected at the flow rates of 2.5 m 3 min -1 from upstream and of 0.3 m 3 min -1 from downstream. The oxygen content of the coating was obtained by the inert-gas fusion method. Its open-porosity was determined by mercury intrusion porosimetry. Its through-porosity was evaluated by Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES) of the eluted ion derived from the substrate while the coated specimen was immersed in 0.5 N HCl [2]. The corrosion resistance of the specimen was evaluated by the electrochemical methods with the three-electrode system using both the polarization and the alternating current (AC) Table 1 Spray conditions. Parameter Unit Range Fuel flow rate dm 3 min 1 0.25 ~ 0.49 Oxygen flow rate dm 3 min 1 760 ~ 1080 Combustion pressure MPa 0.43~0.86 Fuel/oxygen ratio* 0.7~1.2 Barrel length mm 102, 204 Powder feed rate g min 1 60 Torch velocity mm s 1 700 Spray distance* mm 482 Powder feed gas Nitrogen (N 2 ) Film thickness. µm 50~400 *1.0 corresponds to stoichiometric mixture ratio. *From exit of combustion room. Table 2 Chemical compositions of feedstock powders. Material Composition HastelloyC (TAFA1286F) Ni bal, Mo 16.95, Cr 16.57, Fe 6.2, W 4.5, Mn 0.72, Co 0.31, Si 0.73 SUS316L (TAFA 1236F) Fe bal, Cr 16.8, Ni 10.8, Mo 2.05, N 0.131, O 0.026
impedance measurements. The sample electrode was prepared as follows: 1) the coated specimen was cut into pieces in a size of 2.5 cm square and was cleaned ultrasonically in acetone and ion-exchanged water, repeatedly, 2) the stainless lead was connected with the opposite side of sprayed face of the specimen, and 3) the sprayed area of 2 cm 2 left exposed and the rest of the specimen surface was insulated with the silicon resin. The reference electrode was the Ag/AgCl electrode in the saturated KCl solution. The electrolyte was artificial seawater of ph8.3 at 300K. The salt bridge of agar containing KCl was used in addition to the reservoir of artificial seawater to avoid mixing artificial seawater and the KCl solution. From the polarization measurement, the potential-current curve was obtained by measuring the current value when the electrode potential of the sample was scanned at the rate of 10 mv s 1 using a potentiostat with a function generator (Hokuto Denko, HAB-151). The sample electrode was immersed in the electrolyte for 24 hours to reach the steady state. In this case, the counter electrode was a platinum plate with a dimension of 0.2t 100 100 mm, and continous nitrogen bubbling was carried out for deaeration of the electrolyte. From the impedance measurement, the polarization resistance of the sample was determined by subtracting one impedance value at the frequency of 10 khz from another at 100 mhz with the corrosion monitor (Riken Denshi, Model CT-5). In this case, the counter electrode was the same thing as the sample one, and continuous air bubbling was carried out for aeration of the electrolyte. This polarization resistance was monitored every 10 minutes for 3 days in addition to the corrosion potential. Results and discussion In general, the open-porosity of the HVOF sprayed coatings was inversely related to their oxygen content, as shown in Fig. 2. Some plots concerning a kind of material in this figure indicates values obtained under different spray conditions. This relation seems to be a trade-off, and both properties were difficult to improve (i.e. decrease) together in the HVOF spray process. However, this problem was overcome by attaching the gas shroud. The open-porosity below 0.1 vol% was attained accompanying with 0.2 wt% of oxygen content. Such a small porosity was not detectable by mercury intrusion porosimetry, then it could be evaluated precisely by our method, as mentioned above. As seen in Fig. 3, both the decreasing tendency and the detection of eluted Fe ion from the SS400 substrate indicated that the standard HastelloyC () coating with open-porosity below 0.1 vol% retained through-pores even at 400 µm in thickness. The through-porosity of the on-shroud HastelloyC (+SH) Open-porosity / vol% 5.0 4.0 3.0 2.0 1.0 0 : +SH : : 316L 0.5 1.0 Oxygen content / wt% Fig. 2 Relation between open-porosity and oxygen content of HVOF sprayed coatings. -2 h -1 Elution rate (substrate) / mol cm 10-5 10-6 10-7 10-8 10-9 10-10 Detection limit SS400 plate +SH 0 100 200 300 400 Coating thickness / µm 0.1 0.01 0.001 0.0001 Fig. 3 Dependence of elution rate of Fe ion from SS400 substrate on thickness of coating. 10 1 mm year -1
a b c Fig. 4 Optical microscopic images of cross section of HVOF sprayed coatings, (a) SUS316L, (b) HastelloyC and (c) on-shroud HastelloyC. coating was supposed to become zero over 300 µm in thickness because its elution rate was almost the same value as the detection limit of this method. The decrease in through-porosity of the coating by attachment of gas shroud was confirmed also in the cross sectional view of the coated specimens, as shown in Fig. 4. The on-shroud coating had no obvious voids and pores, compared to the SUS316L and the HastelloyC coatings by HVOF spraying without the shroud. In addition, borders between the deposited particles of the on-shroud coating became more ambiguous than the other two coatings, indicating the decrease of the surface oxide of the particles. The corrosion resistance of the HVOF sprayed coatings was discussed in this paragraph. All the specimens for the sample electrodes were about 400 µm in thickness of the coating. As seen in Fig. 5, in the polarization curve of the HVOF SUS316L coating, an anodic current was much larger in the passive region and the pitting potential was much lower than the usual SUS316L bulk material. This was due to overlapping of corrosion of the substrate caused by the presence of through-pores in addition to the poor corrosion resistance of the coating itself. Also in the case of the HastelloyC coating without the gas shroud, an anodic current in the passive region was comparatively large while the pitting corrosion did not take place. Taking into account the characteristics of its bulk material, this result suggested the simultaneous occurrence of both corrosion of the substrate and formation of the oxide on the coating surface. Compared to these two, the on-shroud HastelloyC coating had a lower anodic current in the passive region. Fluctuations of the anodic current with the scanned potential, however, were observed, suggesting the localized corrosion caused by the microscopic ununiformity of the deposited particles. As shown in Fig. 6, the rank of polarization resistance of the coatings had a similar Current density / A cm -2 10-2 10-3 10-4 10-5 10-6 10-7 10-8 316L +SH 10-9 -500 0 500 1000 Potential / mv vs. Ag/AgCl Fig. 5 Polarization curves of HVOF sprayed coatings on SS400 in deaerated artificial seawater at 300K. Polarization resistance / Ω cm 2 10 6 10 5 10 4 10 3 10 2 +SH 316L 10 1 0 20 40 60 Immersion time / hour Fig. 6 Variation in polarization resistance of HVOF sprayed coatings on SS400 with time of immersion in aerated artificial seawater at 300K.
tendency to the result of their polarization measurement. The initial drops in polarization resistance of the SUS316L and the HastelloyC coatings by HVOF spraying without the gas shroud corresponded to the progress of corrosion of the substrate, depending on the through-porosity. After these drops in resistance, almost the same value indicated the occurrence of the steady galvanic corrosion. On the other a Fig. 7 Photographs of HVOF sprayed coatings immersed in aerated artificial seawater at 300K for 3 days, (a: SUS316L, b: standard HastelloyC and c: on-shroud HastelloyC). hand, such a dropping phenomenon was not observed in the on-shroud HastelloyC coating, stemmed from the absence of through-pores. Furthermore, its polarization resistance was comparable to that of bulk material of HastelloyC. Its fluctuations in resistance could be explained by the reason mentioned above. After immersion in artificial seawater for three days, many reddish-brown corrosion products could be seen on the surface the SUS316L coating, indicating corrosion of the substrate, as shown in Fig. 7. The HastelloyC coating without the gas shroud had a significantly small black spot on its surface, indicating corrosion of the substrate caused by remaining of through-pores. The on-shroud HastelloyC coating had some black spots on its surface, suggesting occurrence of the pitting corrosion on the coating surface. Conclusions Addition of the shroud mechanism by the N 2 gas to HVOF spraying could decrease both the through-porosity and the oxide content of the HastelloyC coating. In particular, its through-porosity attained to be zero. Such decreases were allowed to improve the corrosion resistance of coated SS400 up to that of the bulk material of HastelloyC in seawater. A similar reaction to localized corrosion might take place the surface of this coating. References [1] S. Kuroda, T. Fukushima, J. Kawakita, T. Kodama, Proc. ITSC2002, Essen, Germany (2002) pp.819. [2] J. Kawakita, S. Kuroda, T. Kodama, Proc. ITSC2002, Essen, Germany (2002) pp.681. [3] T. Fukushima, H. Yamada, J. Kawakita, S. Kuroda, Proc. ITSC2002, Essen, Germany (2002) pp.912. b c