Hot corrosion studies of HVOF sprayed Cr3C2–NiCr and Ni–20Cr coatings on nickel-based superalloy at 900 °C

Hot corrosion studies of HVOF sprayed Cr3C2–NiCr and Ni–20Cr coatings on nickel-based superalloy at 900 °C

Surface & Coatings Technology 201 (2006) 792 – 800 www.elsevier.com/locate/surfcoat Hot corrosion studies of HVOF sprayed Cr3C2–NiCr and Ni–20Cr coat...

925KB Sizes 1 Downloads 65 Views

Surface & Coatings Technology 201 (2006) 792 – 800 www.elsevier.com/locate/surfcoat

Hot corrosion studies of HVOF sprayed Cr3C2–NiCr and Ni–20Cr coatings on nickel-based superalloy at 900 °C T.S. Sidhu ⁎, S. Prakash, R.D. Agrawal Metallurgical and Materials Engineering Department, Indian Institute of Technology Roorkee, Roorkee-247 667, India Received 15 November 2005; accepted in revised form 20 December 2005 Available online 7 February 2006

Abstract The high velocity oxy-fuel (HVOF) process is widely adopted by many industries due to its flexibility, cost effectiveness, handiness for use and the superior quality of the coatings produced. The present work is a comparative study of HVOF sprayed Cr3C2–NiCr and Ni–20Cr coatings on a Ni-based superalloy in a molten salt environment of Na2SO4–60%V2O5 at 900 °C under cyclic conditions. The thermogravimetric technique was used to establish the kinetics of corrosion. X-ray diffraction, scanning electron microscopy/energy-dispersive analysis and electron probe microanalysis techniques were used to analyse the corrosion products. The hot corrosion resistance of Ni–20Cr coating was better than Cr3C2– NiCr coating. The hot corrosion resistance of both coatings may be attributed to the formation of oxides and spinels of nickel and chromium. These oxides might have blocked the pores and splat boundaries, and acted as diffusion barriers to the inward diffusion of oxidizing species. © 2006 Elsevier B.V. All rights reserved. Keywords: Cr3C2–NiCr; Ni–20Cr; HVOF coating; Hot corrosion; Superalloy; Protective coatings

1. Introduction HVOF process has the advantage of being a continuous and most convenient process for applying coatings to industrial installations at site [1]. Several HVOF sprayed coatings have been subjected to corrosion testing in seawater, including cermets [2–4] and anti-corrosion alloys [5–7]. These studies concluded that the HVOF method produced coatings with higher corrosion resistance when compared with other spraying technologies such as flame spraying, arc spraying, and plasma spraying. Therefore, in this study, the HVOF process is used to deposit the coatings. This research paper is in continuation to an earlier publication of the author [8]. In the present investigation, an attempt has been made to comparatively evaluate the hot corrosion behaviour of HVOF sprayed Cr3C2–NiCr and Ni– 20Cr coatings on a nickel-based superalloy, namely Superni 600 (similar grade Inconel 600) in a corrosive environment of molten salt Na2SO4–60%V2O5 at 900 °C under cyclic

⁎ Corresponding author. Tel.: +91 1332 285731; fax: +91 1332 285243. E-mail address: [email protected] (T.S. Sidhu). 0257-8972/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2005.12.030

conditions. The substrate was provided by Mishra Dhatu Nigam Limited, Hyderabad (India). The objective of the research was to find suitable protective coatings to prolong the lifetime of superalloys for higher temperatures applications in corrosive environments. 2. Experimental 2.1. Substrate material and coating formulation The Superni 600 (15.5Cr–10Fe–0.5Mn–0.2C-balance Ni) was used as a substrate material in the present study. The specimens with dimensions of approximately 20 mm × 15 mm × 5 mm were grinded with SiC-papers down to 180 grit and subsequently grit blasted with alumina powders (Grit 45) before spraying of the coatings by HVOF Process for developing better adhesion to the substrate. Specimens were prepared manually and all care was taken for any structural changes in the specimens. Commercially available Cr3C2–NiCr powder and Ni–20%Cr wire were used as feedstock alloys in the study. The details of these feedstock alloys are given in Table 1. The coatings were sprayed at M/S Metallizing Equipment Co. Pvt. Ltd., Jodhpur (India) by using two types of commercial

T.S. Sidhu et al. / Surface & Coatings Technology 201 (2006) 792–800

793

Table 1 Composition of the feedstock alloys, coating thickness and porosity Coating alloys

Chemical composition (wt.%)

Particle size and shape

Average coating thickness (ìm)

Porosity (% age)

Microhardness, Hv (Vickers hardness)

Cr3C2–NiCr Powder (LA-6875), Blend of 75% LA-6304 and 25% LA-7319 Ni–20Cr wire

75Cr3C2–25 (Ni–20Cr)

−45 μm + 5 μm, Irregular

290

b1.5%

870–950

Ni–20Cr

Wire diameter 3.17 mm

236

b1%

600–630

2.2. Characterisation of the coatings and corrosion products The details regarding characterisation of the coatings and the corrosion products have been reported elsewhere [8,9]. The average thickness of the coatings, porosity and microhardness values of the coatings are given in Table 1. 2.3. Molten salt corrosion tests Hot corrosion studies were performed in a molten salt (Na2SO4–60%V2O5) for 50 cycles under cyclic conditions. Each cycle consisted of 1 h heating at 900 °C in a silicon carbide tube furnace followed by 20 min cooling at room temperature. The studies were performed for uncoated as well as coated specimens for comparison. The specimens were mirror polished down to 1 μm alumina on a wheel cloth polishing machine. A salt of Na2SO4–60%V2O5 (wt.%) was properly mixed in distilled water. After washing with acetone, the specimens were then heated in an oven to about 250 °C. The heating of the specimens was found necessary for proper adhesion of the salt layer. Thereafter, a layer of Na2SO4–60%V2O5 mixture was applied uniformly on the warm polished specimens with the help of a camel hair brush. Amount of the salt coating was kept in the range of 3.0–5.0 mg/cm2. The salt coated specimens as well as the alumina boats were then kept in the oven for 3–4 h at 100 °C. Then they were again weighed before exposing to hot corrosion tests in the tube furnace. During hot corrosion runs, the weight of boats and specimens was measured together at the end of each cycle with the help of a thermogravimetrical balance model 06120 (Contech, India) with a sensitivity of 1 mg. The spalled scale was also included at the time of measurements of weight change to determine total rate of corrosion. Efforts were made to formulate the kinetics of corrosion. XRD, SEM/EDAX, and EPMA techniques were used to analyse the corrosion products.

3. Results 3.1. Thermogravimetric analysis and visual observations Weight change (mg/cm2) vs. time (expressed in number of cycles) plots for the bare and coated superalloy Superni 600 subjected to hot corrosion in Na2SO4–60%V2O5 environment at 900 °C up to 50 cycles are shown in Fig. 1. The weight gain values for the coated Superni 600 superalloy are smaller than those for bare Superni 600. The coated Superni 600 shows higher weight gains during the early cycles of the study, and thereafter the weight gain is nearly gradual. The Ni–20%Cr coating provided relatively better protection than the Cr3C2– NiCr coating and reduced the weight gain by 60% of that gained by bare superalloy, whereas the Cr3C2–NiCr coating reduced the weight gain by 45%. Both the HVOF coatings under study deposited on Superni 600 superalloy almost follow the parabolic behaviour up to 50 cycles as can be inferred from square of weight change (mg2/cm4) vs. number of cycles plots shown in Fig. 2. Hot corroded bare Superni 600 shows visible deviation from the parabolic rate law. The parabolic rate constants (kp in 10− 10 g2 cm− 4 s− 1) for the bare Superni 600 is calculated as 13.7, whereas for Cr3C2–NiCr and Ni–20%Cr coated Superni 600 are found as 3.7 and 3, respectively. It can be inferred from the thermogravimetric study that the necessary protection against hot corrosion is provided by the HVOF coatings under study. A brownish grey scale appeared on the surface of uncoated superalloy during the initial cycles which turned to dark grey after third cycles. The cracks were developed in the scale and 22

Weight gain/Area (mg/cm2)

high velocity oxy-fuel thermal spray system. A Hipojet-2100 HVOF system was used for powder spraying and a Hijet-9600 for wire spraying. The spray parameters employed for Hipojet2100 system were oxygen flow rate, 250 LPM; fuel (LPG) flow rate, 60 LPM; air flow rate, 900 LPM; spray distance, about 20 cm; fuel pressure, 6 kg/cm2; oxygen pressure, 8 kg/cm2 ; air pressure, 6 kg/cm2. The spray parameters employed for Hijet9600 system were also the same except oxygen flow rate, 200 LPM; fuel (LPG) flow rate, 50 LPM; fuel pressure, 4 kg/cm2 and oxygen pressure, 6 kg/cm2. The specimens were cooled with compressed air jets during and after spraying.

20

Bare Superni 600

18

Cr3C2-NiCr Coated

16

Ni-20Cr Coated

14 12 10 8 6 4 2 0 0

5

10

15

20

25

30

35

40

45

50

Number of cycles

Fig. 1. Weight gain vs. number of cycles plot for coated and uncoated Superni 600 subjected to hot corrosion for 50 cycles in Na2SO4–60%V2O5 environment at 900 °C.

794

T.S. Sidhu et al. / Surface & Coatings Technology 201 (2006) 792–800

(Weight gain/area)2 mg2 cm-4

300 Bare Superni 600 250 Cr3C2-NiCr Coated 200

Ni-20Cr Coated

150 100 50 0 0

5

10

15

20

25

30

35

40

45

50

-50

Number of Cycles

Fig. 2. (Weight gain/area)2 vs. number of cycles plot for coated and uncoated superalloy Superni 600 subjected to hot corrosion for 50 cycles in Na2SO4–60% V2O5 environment at 900 °C.

spalling was observed during the course of study. Initially, the spalling was confined to edges and corners, but after 12th cycle, spalling also occurred from the surface of the uncoated superalloy. The spalling tendency of the scale formed on uncoated superalloy can be seen in Figs. 3a and 5a. The Cr3C2–NiCr coated Superni 600 showed minor cracks at the edges and corners, and spalling from these regions occurred from 12th cycle. The colour of the scale which was dark grey during the earlier cycles turned to blackish green with the progress of the study. In the case of Ni–20%Cr coated Superni 600, a blackish grey scale was formed after first cycle which gradually turned to dark black during next few cycles. Subsequently after 15th cycles, a very shining silver grey scale appeared which lasted up to the end of study. The scale was smooth and intact, and spalling of the scales was negligible. Mostly marginal spalling of the coatings was observed near and/ or the edges. BSEIs across the smallest cross-section of scales formed on coated Superni 600 are shown in Fig. 3 which shows the formation of continuous, adherent and compact scale.

scale (Fig. 5a). The thermal cyclic test condition of Na2SO4– 60%V2O5 at 900 °C induces basic fluxing. SEM images of bare superalloy shown in Figs. 3a and 5a seem to indicate well fluxing action of the salt. The EDAX analysis of the scale shows NiO to be the predominant phase. This NiO detected on bare superalloy surface may be porous NiO due to reprecipitaion by fluxing action. The surface scale developed on Cr3C2–NiCr and Ni–20%Cr coated Superni 600 is homogeneous and continuous. The EDAX analysis of the scale of Cr3C2–NiCr coated Superni 600 revealed the formation of Cr2O3 and NiO as principal phases along with small quantities of MnO and V2O5 (Fig. 5b), whereas NiO is detected as the principal phase in the scale of Ni–20% coated Superni 600 together with little quantity of Cr2O3, Fe2O3 and V2O5 (Fig. 5c). The existence of MnO and Fe2O3 on the surface of hot corroded Cr3C2–NiCr (Fig. 5b) and Ni–20Cr (Fig 5c) coated Superni 600, respectively, indicates that these elements have diffused from the substrate to the uppermost part of the scale during hot corrosion of the specimen at 900 °C.

3.2. X-ray diffraction analysis The XRD profiles for the scale of bare and HVOF coated Superni 600 superalloy after hot corrosion in Na2SO4–60% V2O5 environment for 50 cycles at 900 °C are shown in Fig. 4. The major and minor phases detected at the surface of the specimens with the XRD analysis are given in Table 2. The existence of MnO2 and Fe2O3 phases on the surface of hot corroded Cr3C2–NiCr and Ni–20%Cr coated Superni 600, respectively, indicates the diffusion of Mn and Fe from the substrate during hot corrosion of the specimens at temperature about 900 °C. 3.3. SEM/EDAX analysis of the scales 3.3.1. Surface analysis SEM micrographs along with EDAX analysis at some selected sites of interest of the hot corroded bare and HVOF coated Superni 600 superalloy are shown in Fig. 5. The micrograph of the bare Superni 600 superalloy noticeably shows the presence of cracks as well spalling behaviour of the

Fig. 3. SEM back scattered images for the bare and HVOF coated Superni 600 superalloy subjected to hot corrosion in Na2SO4–60%V2O5 environment at 900 °C for 50 cycles. (a) Bare superalloy; (b) Cr3C2–NiCr coated; (c) Ni–20% Cr coated.

T.S. Sidhu et al. / Surface & Coatings Technology 201 (2006) 792–800

μ λ Σ

Fe2O3 NiO FeV MnO2 Cr2O3

μ η

Intensity (arbitrary unit)

η λ

λ Σ η

η η

795

NiCr2O4 FeV2O4 Ni (VO3)2 CrVO4

η

η

η

η

Bare Superni 600

Σ

η

Ση η

Σ η

Ni-20%Cr Coated Superni 600

Σ η

η

η

η η

η

Cr2C3-NiCr Coated Superni 600 20

30

40

50

60

70

80

90

100

110

Diffraction Angle (2θ)

Fig. 4. X-ray diffraction patterns for the bare and coated superalloy Superni 600 subjected to hot corrosion in Na2SO4–60%V2O5 environment at 900 °C for 50 cycles.

3.3.2. Cross-sectional analysis EDAX analysis was carried out at different point of interest along the cross-section of the corroded bare and HVOF coated Superni 600 and the results are given in Fig. 6. A non-uniform oxide scale with variable thickness is formed on bare Superni 600 (Fig. 6a). The presence of about 30 wt.% oxygen at point 1 indicates that the oxygen has penetrated through the oxide scale formed on the surface of Superni 600. The EDAX analysis shows that the contrast grey phase (point 2) mainly consists of Cr, and Ni decreases substantially in this phase. The existence of significant quantity of oxygen (40 wt. %) points out the possibility that this grey phase may be rich with Cr2O3. The absence of oxygen at point 3 and its presence at point 1 show that the growth of oxide scale is irregular. The top surface of the scale (point 5) contains mainly oxides of Ni, Cr and Fe. Cross-section BSEI of hot corroded Ni–20%Cr coated Superni 600 (Fig. 6b) shows that the top surface of the coating mostly has a homogeneous and adherent oxide scale of about 45 μm thick and cracks observe to be formed in the scale. The remaining portion of the coating has maintained its lamellar structure similar to the as-sprayed conditions. EDAX analysis reveals that light grey layer in the upper part of the scale (point 5) is rich with oxides of Ni. The white phase (point 2) is identified as unoxidised Ni-rich splats which are uniformly spread in the scale, and oxides are formed mainly at the boundaries of these Ni-rich splats. EDAX analyses at

point 3 and 4 show that these oxide stringers are mainly rich with oxides of Cr. Therefore, the coating contains chromium oxides uniformly distributed in a Ni-based matrix. Very little amount of oxygen is found to be present near the coating– substrate interface (point 1) suggesting that some oxygen might have penetrated during initial cycles of hot corrosion runs. Thereafter, the splat boundaries and pores present in the coating are clogged due to formation of oxides, which might have blocked the penetration of reacting species towards the base alloy. EDAX analysis for Fe and Mn elements revealed outward diffusion tendency of these elements from the substrates at temperature about 900 °C. These elements reached almost to surface of the scale. Diffusion of Fe is found to be more than Mn near the coating–substrate interface (point 2), whereas Mn shows more diffusion near top of scale (point 4).

Table 2 Major and minor phases identified by XRD analysis for the hot corroded bare and coated Superni 600 Description

Major phases

Minor phases

Uncoated Superni 600 Cr3C2–NiCr coated Ni–20Cr coated

Superalloy NiO, Fe2O3, NiCr2O4, Ni(VO3)2, FeV, and FeV2O4 Cr2O3, NiO, NiCr2O4, Fe2O3, and Ni(VO3)2 NiO, NiCr2O4, Fe2O3, and Ni(VO3)2

CrVO4 MnO2 Cr2O3

796

T.S. Sidhu et al. / Surface & Coatings Technology 201 (2006) 792–800

a 88% NiO 03% Cr2O3

96% NiO

07% Fe2O3

02% Cr2O3

01% V2O5

01% Fe2O3

b 63% Cr2O3 67% Cr2O3

35% NiO

31% NiO

01% MnO

01% V2O5

c 98% NiO

96% NiO

01% Cr2O3

02% Cr2O3

01% V2O5

01% Fe2O3

Fig. 5. SEM/EDAX analysis showing elemental composition (Wt.%) for the bare and coated Superni 600 subjected to hot corrosion in Na2SO4–60%V2O5 environment at 900 °C for 50 cycles: (a) Bare Superni 600; (b) Cr3C2–NiCr coated: (c) Ni–20%Cr coated.

3.4. EMPA analysis BSEI and EPMA elemental mappings for the coated Superni 600 after cyclic oxidation at 900 °C in Na2SO4–60%V2O5 environment for 50 cycles are shown in Figs. 7 and 8. The scale formed on the Cr3C2–NiCr coated Superni 600 exhibits dense structure with layered morphology in which Crrich and Ni-rich splats are presents at an alternate position (Fig. 7). The density of the Cr-rich splats is more than the Ni-rich Splats. The elemental maps for Cr, Ni, Fe and O showed that the scale formed is rich in Cr2O3 with small amount of NiO and Fe2O3. The Ni-rich splats remain mostly in an unoxidised state. The Cr, Ni and O coexist at some places suggesting the formation of NiCr2O4. The elemental mapping for Fe showed that outward diffusion from the substrate to the coating occurred and the diffusion seems to have reached the surface. It is very interesting to note that concentration of Cr remains the same through the scale and its concentration did not get reduced by diffusion. In the O elemental map the higher concentration was noticed at the interface. The islands present at the interface do not match with the elements of both the substrate and coating. These islands are identified as aluminum oxides possibly due to the incorporation of alumina powder during polishing. Similar

islands of aluminum oxides due to retained alumina powder during polishing were also reported by Sundararajan et al. [10]. The S penetrates through the coating and reaches up to the scale–substrate interface. BSEI for the cross-section of hot corroded Ni–20Cr coated Superni 600 (Fig. 8) shows the appearance of thin featureless layer in the upper part of the scale, whereas the rest of the coating retains lamellar structure similar to as-sprayed coating. X-ray mappings reveal that this continuous, compact and adherent topmost scale is rich in nickel. Underneath of topmost layer, the nickel-rich and chromium-rich splats are present at an alternate position. Iron diffuses from the substrate into the scale and diffusion is significant near the substrate–scale interface. Manganese also diffuses from the substrate to surface of the scale. The presence of sodium, sulfur and vanadium throughout the scale indicates the penetrating behaviour of these corroding species through the coating. 4. Discussions Fig. 1 shows that the weight of bare superalloy increases continuously, whereas weight gain of the coated specimens is relatively high during the first 5–6 cycles of hot corrosion, but

T.S. Sidhu et al. / Surface & Coatings Technology 201 (2006) 792–800 100

120

Ni V

O Mn

Cr S

Fe Na

Ni

100

Weight %

80

Weight %

797

60 40 20

Cr

Fe

O

Mn

80 60 40 20

0

0 1

2

3

4

5

1

2

3

4

a

b

4

3

4

5

2 1 3

2 10μm

5

Point of Analysis

Point of Analysis

5

50μm

1

Fig. 6. Oxide scale morphology and variations of elemental composition across the cross-section of HVOF coated Superni 600 superalloy hot corroded in Na2SO4– 60%V2O5 environment at 900 °C for 50 cycles. (a) Bare Superni 600; (b) Ni–20%Cr coated.

BSEI

Epoxy

O-Kα

200μm Substrate

Cr-Kα

Ni-Kα

Fe-Kα

S-Kα

Fig. 7. Composition image (BSEI) and X-ray mappings of the cross-section of the Cr3C2–NiCr coated Superni 600 superalloy subjected to hot corrosion at 900 °C in Na2SO4–60%V2O5 environment for 50 cycles.

798

T.S. Sidhu et al. / Surface & Coatings Technology 201 (2006) 792–800

Ni-Kα BSEI

Cr-Kα

Na-Kα

V-Kα

Fe-Kα

Mn-Kα

S-Kα

Fig. 8. Composition image (BSEI) and X-ray mappings of the cross-section of the Ni–20%Cr coated Superni 600 superalloy subjected to hot corrosion at 900 °C in Na2SO4–60%V2O5 environment for 50 cycles.

subsequently increase in weight is gradual. The initial high oxidation rate of the coated specimens might be ascribed to the rapid formation of oxides at the splat boundaries and within open pores due to the penetration of the oxidizing species. The EDAX analysis at the cross-section of the coating (Fig. 6b) shows the formation of oxides at the splat boundaries. The formation of oxides at the splat boundaries is supported by EPMA analysis (Fig. 7). These oxides might have blocked the pores and splat boundaries, and acted as diffusion barriers to the inward diffusion of oxidizing species. As a consequence, the growth of the oxides becomes limited mainly to the surface of the specimens. Therefore, the steady state of oxidation might have reached with the progress of exposure time. The rapid

increase in weight gain during the earlier cycles can also be attributed to the possible formation of NaVO3 (m.p. ≈ 610 °C) due to reaction of Na2SO4 and V2O5 at 900 °C [11]. This NaVO3 serves as an oxygen carrier which will lead to the rapid oxidation of the coatings to form the protective oxide scale during initial cycles [12]. There may also be simultaneously dissolution of Cr2O3 in the molten salt due to the reaction Cr2O3 + 4NaVO3 + 3/2O2 → 2Na2CrO4 + 2 V2O5 [13,14]. This Na2CrO4 gets evaporated as a gas [15]. Therefore, higher weight gain during initial cycles is due to rapid oxidation but slower increase in weight gain during the subsequent cycles is probably due to the growth of oxides, and simultaneously dissolution of Cr2O3.

T.S. Sidhu et al. / Surface & Coatings Technology 201 (2006) 792–800

Both the coatings deposited on Superni 600 followed nearly parabolic oxidation behaviour up to 50 cycles in the given environment, whereas the uncoated Superni 600 shows deviations from the parabolic law. The parabolic rate constant for the uncoated superalloy is greater than the coated superalloys. Therefore, it can be inferred that the coatings showed hot corrosion resistance in the given environment at 900 °C. The bare Superni 600 shows the formation of fragile scale with number of cracks and significant spalling (Figs. 3a and 5a). The oxide scale is penetrating deeply into the substrate (Fig. 6a). Higher spalling of the scale of the bare superalloy can be attributed to severe strain developed due to the precipitation of Fe2O3 from the liquid phase during cooling period of thermal cycles, and interdiffusion of intermediate layers of iron oxide [16]. Further, the presence of 07 different phases in the bare Superni 600 might impose severe strain on the surface thin layer during cooling period which may result in cracking and exfoliation of the scale. The cracks may allow the aggressive liquid phase to reach the metal substrate [16]. The coated specimens have shown little spallation of the scale due to the formation of thick protective layers of Cr2O3 and NiO, as indicated by EPMA analysis (Figs. 7 and 8). Continuous and compact surface oxide scales are formed on the coated Superni 600 with very little spallation (Fig. 5b and c). The BSEIs show that the scale is adherent and the substrate is not affected by internal oxidation (Fig. 3b and c). This is further confirmed by the cross-sectional EDAX analysis for the Ni–20Cr coated Superni 600 (Fig. 6b), which shows that the oxygen is not penetrated into the substrate alloys. The formation of oxides along the splat boundaries of the coatings might have acted as diffusion barrier to the inward diffusion of corrosive species (Fig. 6b, point 3 and point 4). EPMA results further support the BSEI observations and EDAX analysis (Fig. 7). In terms of weight gains, the Ni–20Cr coating showed better resistance to hot corrosion and provided best protection to the base alloy as it decreased the overall weight gain by 60% approximately. This may be attributed to the presence of protective oxides of nickel and chromium, and their spinels as detected by the surface XRD analysis. The presence of these phases is supported by the surface/cross-sectional EDAX analysis. Selective oxidation of chromium along the nickelrich splats boundaries, as detected by cross-sectional EDAX analysis (Fig. 6b), might have acted as diffusion barrier to the inward diffusion of corrosive species. These results are in good agreement with the findings of Longa-Nava et al. [17] and Calvarin et al. [18]. Calvarin et al. [18] have established that the oxide scale formed after high temperature oxidation of Ni–20Cr foils at 900 °C consists of an outer NiO layer and an intermediate layer composed of oxides rich in nickel and chromium. The better hot corrosion resistance of the coating can also be partly attributed to the uniform fine grain microstructure of as sprayed Ni–20Cr wire coating [8]. Formation of small amounts of Fe2O3 and MnO indicated by EDAX/XRD analysis and supported by EPMA results may be due to their diffusion from the substrate to the coating at higher temperature.

799

In terms of weight gains, the Cr3C2–NiCr coated Superni 600 showed less hot corrosion resistance than the Ni–20Cr coated Superni 600. This can be attributed to the presence of higher amounts of chromium which may lead to the formation of thick oxide layer rich in chromium. The elemental mappings show that chromium and oxygen coexist suggesting the formation of chromium-rich thick oxide layer (Fig. 7). Whereas in case of Ni–20Cr coated Superni 600, the chromium-rich oxides are formed only at the boundaries of nickel-rich splats. Hence, the Cr3C2–NiCr coated alloy shows higher weight gain than the Ni–20Cr coated Superni 600. The presence of Cr2O3, NiO, and NiCr2O4 surface oxides, as detected by XRD and EDAX analysis and supported by EPMA analysis, might have contributed to the hot corrosion resistance for Cr3C2–NiCr coated superalloy. Formation of small amounts of oxides of iron and manganese as indicated by XRD/EDAX/EPMA analysis may be due to the diffusion from the substrate to the coating. Some minor spalling of the oxide scale of coated superalloys especially on the edges and corners [19] during cooling periods of the thermal cycles may be due to different values of thermal expansion coefficients of the coatings, substrate and oxides [20–22]. 5. Conclusions 1. The uncoated Superni 600 showed less resistance to hot corrosion in Na2SO4–60%V2O5 molten salt environment at 900 °C due to spalling behaviour of the scale. The cumulative weight gain of bare Superni 600 was more than the coated Superni 600 during hot corrosion study. 2. HVOF sprayed Cr3C2–NiCr and Ni–20Cr coatings improve the hot corrosion resistance of the Superni 600 in the given conditions. The formation of oxides along the splat boundaries and within open pores of the coatings might have acted as diffusion barrier to the inward diffusion of corrosive species. The oxides formed at the surface may be porous due to reprecipitaion by fluxing action. The HVOF sprayed Cr3C2–NiCr coatings has formed thick layer of Cr2O3, whereas Ni–20Cr coating showed the formation of alternative Ni- and Cr-rich layers. 3. The Ni–20Cr wire coating has better hot corrosion resistance than the Cr3C2–NiCr powder coating. The Ni–20Cr coating reduced the weight gain by 60% of that gained by bare Superni 600, whereas Cr3C2–NiCr coating reduced the weight gain by 40%. The presence of oxides of nickel and chromium and their spinels NiCr2O4, and uniform fine grain microstructure of the as sprayed coating [8] might have contributed for better hot corrosion resistance of HVOF sprayed Ni–20Cr wire coating. 4. The hot corrosion resistance of Cr3C2–NiCr coating might be due to the formation of protective phases like NiO, Cr2O3 and NiCr2O4. A dense oxide scale formed on the Cr3C2– NiCr coated Superni 600 is rich in chromium. 5. The marginal spalling observed in case of Cr3C2–NiCr coated Superni 600 and negligible spalling of Ni–20Cr coated Superni 600 might be due to the different values of

800

T.S. Sidhu et al. / Surface & Coatings Technology 201 (2006) 792–800

thermal expansion coefficients of the coatings, substrate and oxides. 6. Some elements of the substrate such as Fe and Mn showed a tendency of outward diffusion from the substrate to the coating. References [1] T.S. Sidhu, S. Prakash, R.D. Agrawal, Mar. Technol. Soc. J. 39 (2) (2005) 55. [2] J.M. Guilemany, J. Fernandez, J.M. de Paco, J. Sanchez, Surf. Eng. 14 (1998) 133. [3] A. Collazo, X.R. Novo, C. Perez, Electrochim. Acta 44 (1999) 4289. [4] K. Tani, M. Adachi, A. Nakahira, Y. Takatani, Proc. 1st Int. Thermal Spray Conf., Montréal, Québec, Canada, ASM International, May 8–11 2000, p. 1025. [5] P. Gu, B. Arsenault, J.J. Beaudoin, J.G. Legoux, B. Harvey, J. Fournier, Cem. Concr. Res. 28 (1998) 321. [6] A.J. Sturgeon, D.C. Buxton, Proc. 1st Int. Thermal Spray Conf., Montreal, Canada, ASM International, May 8–11 2000, p. 1011. [7] D. Harvey, O. Lunder, R. Henriksen, Proc. 1st Int. Thermal Spray Conf., Montreal, Canada, ASM International, May 8–11 2000, p. 991. [8] T.S. Sidhu, S. Prakash, R.D. Agrawal, Characterisation of NiCr Wire Coatings on Ni- and Fe-Based Superalloys by the HVOF Process, Surface and Coatings Technology, in press (available online).

[9] T.S. Sidhu, S. Prakash, R.D. Agrawal, International Conference on Surfaces, Coatings and Nanostructured Materials, University of Aveiro, Portugal, September 6–9 2005, Paper No. FO-17. [10] T. Sundararajan, S. Kuroda, T. Itagaki, F. Abe, ISIJ Int. 43 (1) (2003) 95. [11] G.A. Kolta, I.F. Hewaidy, N.S. Felix, Thermochim. Acta 4 (1972) 151. [12] Gitanjaly, Role of Inhibitors on Hot Corrosion of Superalloys in Na2SO4– V2O5 Environment, Ph.D. Thesis, Met. Mat. Engg. Dept., Indian Institute of Technology Roorkee, Roorkee, India., 2003. [13] M. Seiersten, P. Kofstad, High Temp. Technol. 5 (3) (1987) 115. [14] J. Swaminathan, S. Raghavan, S.R Lyer, Trans. Indian Inst. Met. 46 (3) (1993) 175. [15] G.C. Fryburg, F.J. Kohl, C.A. Stearns, J. Electrochem. Soc. 131 (2) (1984) 2985. [16] K. Sachs, Metallurgia (Apr. 1958) 167. [17] Y. Longa-Nava, Y.S. Yang, M. Takemoto, R.A. Rapp, Corrosion 52 (9) (1996) 680. [18] G. Calvarin, R. Molins, A.M. Huntz, Oxid. Met. 53 (1–2) (2000) 25. [19] Q.M. Wang, Y.N. Wu, P.L. Ke, H.T. Cao, J. Gong, C. Sun, L.S. Wen, Surf. Coat. Technol. 186 (3) (2004) 389. [20] P. Niranatlumpong, C.B. Ponton, H.E. Evans, Oxid. Met. 53 (3–4) (2000) 241. [21] H.E. Evans, M.P. Taylor, Oxid. Met. 55 (1–2) (2001) 17. [22] B.S. Sidhu, S. Prakash, Oxid. Met 63 (3–4) (2005) 241.