- Email: [email protected]
Loss of Y from NiCrAlY powder during air plasma spraying
Surface & Coatings Technology 280 (2015) 277–281
Contents lists available at ScienceDirect
Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat
Loss of Y from NiCrAlY powder during air plasma spraying Panadda Niranatlumpong ⁎, Chalermchai Sukonkhet, Kittichai Ninon National Metals and Materials Technology Center, Pathumthani, Thailand
a r t i c l e
i n f o
Article history: Received 28 January 2015 Revised 14 July 2015 Accepted in revised form 6 September 2015 Available online 11 September 2015 Keywords: Plasma spraying NiCrAlY In-flight oxidation Y depletion
a b s t r a c t This study involves changes in microstructures and chemical compositions of NiCrAlY particles that occurred during an air plasma spraying of the powder. In this work the spraying particles were rapidly cooled and collected at a spraying distance just before the coating deposition in order to study the mid-flight transformation. Microscopy analysis of the particle cross-sections reveals that different structural and chemical transformations took place presumably depending on the flight path in relation to the plasma plume. Particle A group has a tendency to show some increase in Al and Y contents, possibly due to an evaporation of γ-(Ni solid solution) and β-(NiAl). Accompanying β grain growth can sometimes be observed. Particle B group often forms a large amount of Al–Y rich oxide due to an increase in oxygen activity during flight. This results in a significant reduction of Y in the particle. The coating produced using the same set of spraying parameter shows a heterogeneous structure with respect to the distribution of Y. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Yttrium has been added to NiCrAl alloy as a reactive element to produce a starting material for a bondcoat used extensively as part of a thermal barrier coating (TBC) for land-based gas turbine Ni alloy components [1,2]. Various techniques can be employed for TBC deposition including plasma spraying, physical vapour deposition and chemical vapour deposition [3–7]. In plasma spraying processes, the coating materials in powder form are deposited in a molten or semi-molten condition on a substrate to form a spray deposit. During a coating deposition, the spraying material is subjected to a temperature higher than 1000 °C for a short time. The function of the bondcoat is to protect the component from oxidation. The oxidation resistance is achieved mainly by a formation of a continuous protective Al2O3 scale during service [8, 9]. The scale helps to reduce the extent of corrosion of the underlying substrate thus allowing the engine to operate at a higher efficiency. After the deposition of the bondcoat, ZrO2 topcoat stabilised with other oxides such as Y2O3 and CeO2 can also be deposited to create TBC [10,11]. The exact role of Y in NiCrAlY has received attention in recent years. Previous research works have shown that Y can increase the adhesion of the thermally-grown oxide scale formed during TBC service at high temperature via sulphur gettering or oxide pegging mechanisms, thus enhancing the service life of the coating [12,13]. Scarcity of rare earth metals including Y has sent their prices soaring since before 2011. Their global demands continue nevertheless, encouraging more exploration and development of new mines. More mining of ⁎ Corresponding author. E-mail address: [email protected] (P. Niranatlumpong).
http://dx.doi.org/10.1016/j.surfcoat.2015.09.009 0257-8972/© 2015 Elsevier B.V. All rights reserved.
rare earth metals unavoidably leads to more environmental degradation and human health hazards [14]. It is vital then that the usage of the metals is justified and this requires a good understanding of their functions. This work investigates the loss of Al and Y from NiCrAlY during its inflight oxidation. This was achieved by rapidly cooling the spraying powder mid-flight. The powder was then characterised in order to obtain a better insight into its oxidation behaviour. The result was compared with that of a coating produced using the same spraying parameter. The findings will form a basis to the understanding of the mechanism of Y in prolonging the life of NiCrAlY coating at high temperature. 2. Experimental procedure An air plasma spraying experiment is carried out with an aim to collect the spraying bondcoat powder mid-flight at a deposition distance. The equipment setup is shown in Fig. 1. The bondcoat powder used in this work is Amdry 962 from Sulzer Metco with a particle size of 53–106 μm. The chemical composition is Ni–(21–23)Cr–(9– 11)Al–(0.8–1.2)Y in weight percent. The topcoat powder is Amperit 827.7 from H.C. Stark with a particle size of 45–90 μm. The chemical composition is ZrO2–7Y2O3 (wt.%). The bondcoat powder was sprayed using a 3MBII plasma spray gun (Sulzer Metco) and a set of parameter as shown in Table 1. A metal screen with a 5 mm opening was placed at a distance of 125 mm in front of the spray gun in order to allow a small amount of spraying powder through. As the spraying powder passed through the opening, an Argon jet of 0.1 MPa pressure was used as a medium to rapidly cool the powder. An earlier experiment was also carried out using
278
P. Niranatlumpong et al. / Surface & Coatings Technology 280 (2015) 277–281
3. Results and discussion 3.1. As-received powder The cross-sectional structure of the as-received powder is shown in Fig. 2. The particle surface is free of continuous oxide layer. The back scattered electron image shows that the alloy contains 2 major phases, γ (Ni solid solution) and β (NiAl). This is in accordance with the isothermal section of a Ni–Cr–Al ternary phase diagram as proposed by Gleeson et al. (1993) [15]. There is also Y segregation in γ phase, seen as light-coloured patches in the particle, see Fig. 3. The as-received powder shows structural consistency from one particle to another. Examples of the chemical composition of the individual particles using EDS are shown in Table 2. It should be noted that Y content is slightly higher than the powder specification. Fig. 1. Experiment setup.
3.2. Ar cooled powder
water as a cooling medium. Even though the water cooling process was simpler but the collected particles were found to be enveloped in thick crusts of oxide. This is believed to be because the collecting distance was specified at 125 mm which was close to the plasma jet thus some of the water may be dissociated while in contact with the particle, resulting in a rapid oxidation of the powder particle. Further experiment was therefore carried out using Ar as the cooling medium. The screen served multiple functions. It helped to shield the spray plume from interference from Ar jet which could cause an abnormal turbulence. Also Ar jet can cool the small amount of spraying powder more effectively and propel it into a cooling tube where the powder can be collected. It was originally intended that the screen will also act as a sampling window, allowing the powder collection from different sections across the plume. When the powder was collected after spraying however, it was found that the distinction in average powder appearances across the cross section of the plume was not obvious. This is because there is an arc fluctuation in a plasma spraying process making it difficult to predict the flight path of the particles. However, individual powder particles exhibited differences in appearance and composition. The bondcoat powder specimens were indicated as “as-received” for the starting powder and “Ar-cooled” for the sprayed powder, cooled and collected at 125 mm spraying distance. Accuraspray-g3 (Tecnar, Sulzer Metco) plume diagnostic equipment was used to determine the temperature of the powder at 125 mm spraying distance. The powder temperature was measured through thermal emission from the semimolten particles. This technique may produce inaccurate results due to its reliance on a database on the emissivity of materials. The result should be treated only as a guideline. Coating specimens were also produced with both bondcoat and topcoat deposition using the parameters in Table 1. The specimens were cross-sectioned. Scanning electron microscope and quantitative energy-dispersive X-ray spectroscopy (EDS) (JEOL6301) were used to characterise the specimens.
As the powder was passed through a high temperature H2/N2 plasma plume, each particle was subjected to different environments according to its flight path in relation to the plume. For example, if Particle A travels through the centre of the plume, it will be subjected to the highest temperature of that plasma plume in a reducing atmosphere until it leaves the centre or travels further from the nozzle where the temperature begins to drop and oxygen partial pressure increases [16]. Whereas Particle B travels along the edge of the plasma plume, it will be subjected to a lower temperature throughout the flight in a more oxidising atmosphere. Even though the plasma gas does not contain oxygen, there are turbulences at the edge of the plasma plume resulting in air entrainment. The presence of oxygen together with the high temperature of the plasma results in a highly oxidising environment [17]. Plume analysis reading indicates that the average temperature of the powder surface at the centre of the plume and at a spraying distance of 125 mm is 3078 ± 23 °C and the average powder velocity is 424 ± 4 m/s. At this high temperature it can be expected that Particle A and B will exhibit a different microstructure. The comparison between Particle A and B is an extreme example where the two particles take very different paths. It is more likely however that a particle take a non-linear path, for example, it may starts off at the centre of the plume where the temperature is high and very quickly moves to the fringe of the plume where oxygen is abundance. In this experiment, the powder was sprayed and collected at a spraying distance of 125 mm which is a distance generally used for this set of spraying parameter. In order to “freeze” the structure at the distance of 125 mm, an argon jet of 0.1 MPa pressure was used to intersect the powder in mid-flight and propel it into a cooling tube. A temperature sensor indicated that on average the ambient temperature
Table 1 Parameters for plasma spraying.
Hydrogen pressure Hydrogen flow rate (SCFm) Nitrogen pressure Nitrogen flow rate (SCFm) Current Potential difference Powder feed rate Spraying distance Hopper pressure
Bondcoat NiCrAlY
Topcoat ZrO2-Y2O3
0.34 MPa 1.2 × 10−4 m3 s−1 0.69 MPa 7.9 × 10−4 m3 s−1 500 A 70–80 V 0.78 g s−1 125 mm 0.55 MPa
0.34 MPa 1.2 × 10−4 m3 s−1 0.69 MPa 7.9 × 10−4 m3 s−1 500 A 65–70 V 0.71 g s−1 125 mm 0.9 MPa
Fig. 2. Back scattered micrograph of an as-received powder particle.
P. Niranatlumpong et al. / Surface & Coatings Technology 280 (2015) 277–281
279
Fig. 3. Back scattered micrograph of an as-received powder particle at ×10000 magnification and EDS line scanning result.
of the cooling tube is below 50 °C. Therefore in this work, we assume that the structure of the powder was successfully captured at the moment it was intersected by the argon jet. The sprayed powder shows a variety in physical changes including a small amount of coalescing and tearing of the particles which is not the focus of this paper and will not be discussed. The particles can be coarsely grouped into 2 groups namely Particle A and Particle B according to their microstructures and the oxide appearance. There are also some particles that fall between the two groups with a mixed microstructure. A particle would be classed as Particle A if there was no oxide or only a thin oxide layer on the particle surface. The thin oxide layer of generally less than 0.1 μm thick cannot be easily observed using SEM. It was suspected that Particle A travels near the centre of the plume so even though it was subjected to high temperature, the oxygen activity is presumably too low for it to form a significant amount of oxide [16], see Fig. 4. Table 2 shows the average chemical composition of the particle. Many particles show no observable chemical changes from the asreceived powder. A small number of particles however show a large change in chemical composition, in particular, increases in Al and Y. An increase in Y content is likely to be due to the selective evaporation of γ and β during spraying, leaving a larger proportion of Y in the particle. This is because Y has a negligible solubility in Ni and a large proportion of Y exists as a separate phase [18], see Fig. 3. Fig. 4 also shows β-NiAl grain growth. Generally the diameter of β-NiAl increases from 1–2 μm in the as-received powder to up to 4 μm in Particle A, varying from one particle to another. This may occur due to shifting in the chemical composition. As Y content increases the alloy becomes more brittle. Cracking can occasionally be observed on collected particles. The average measurements of the particle diameter indicate a slight decrease after spraying although this is not significant due to the large variation in the particle size in the as-received powder. A particle would be classed as Particle B if there was a large amount of oxide on the powder surface. This is very distinct on an SEM image. Particle B forms a thick oxide scale on one side of the particle. The oxide can remain intact on the collected particles or may break off during the rapid cooling of the particles, see Fig. 5. An EDS study shows that the oxide scale consists of 2 phases; dendritic Al-rich oxide and Y-rich oxide. Consequently, the proportions of Al and Y in the particle decrease due to the outward diffusion for the oxide formation.
The average chemical composition of the particle is shown in Table 2. The standard deviations are higher than those of the as-received powder indicating a large variation in the chemical composition from particle to particle. The result generally indicates a depletion of Y. The microstructure of the powder also reflects the change in its composition. The structure consists mostly of γ phase with β at grain boundaries. The segregation of Y is no longer observed, see Fig. 5. The average measurements of the particle diameter show a significant reduction in the particle size after spraying. The measurement on Particle B does not include the oxide scale. The large reduction may indicate evaporation of the particle during flight. 3.3. As-sprayed coating Microstructure of the coating produced using the same set of parameter as the cooled powder is shown in Fig. 6. The structure shows a large variation from one splat to another which is in agreement with the large variation found in the Ar cooled powder in Table 2. An overall composition of the coating indicates that the coating has slightly lower Y content than the starting powder. The deposited splats can be coarsely separated into two groups according to their chemical compositions and the microstructures; Splat A and Splat B. It is reasonable to assume that a collected Ar cooled powder of Particle A group, if instead allowed to deposit as a coating, will form a splat in the Splat A group. Splat A and Particle A generally has similar amounts of Al and Y as the as-received powder. Occasionally, Al and Y contents in Splat A and Particle A are much higher than those of the as-received powder resulting in high average values. These splats often show multi-phase structures consisting of β-NiAl grains in a matrix containing a varying amount of Ni, Cr and Y. Small cracks are observed in Splat A although infrequent. The presence of crack is possibly due to an increase in its brittleness as a result of a large proportion of Y. On the other hand, a powder from Particle B group if allowed to deposit as a coating will result in a splat depleted of Y. These splats, belonging to Splat B group, show a single phase splat containing significantly reduced percentage of Y when compared to the as-received powder, see Table 2 and Fig. 6. Al–Y rich oxide with a dendritic structure which was previously observed in Fig. 5 however was rarely present in the coating. Fig. 7 shows
Table 2 Average chemical composition of powder particles and coating splats. Average wt.%
Ni
Cr
Al
Y
Average particle diameter (μm)
Analysed area
As-received powder Ar-cooled powder
66.3 ± 0.3 62.3 ± 2.3 67.5 ± 1.4 66.8 ± 0.5 63.2 ± 2.3 67.2 ± 2.4 6.1 ± 3.4
22.1 ± 0.1 21.2 ± 1.6 21.5 ± 1.0 22.2 ± 0.3 20.9 ± 0.8 22.4 ± 0.7 6.9 ± 5.4
10.0 ± 0.3 14.8 ± 2.4 10.8 ± 1.8 9.7 ± 0.5 14.6 ± 2.4 9.8 ± 2.6 50.8 ± 12.5
1.5 ± 0.1 1.7 ± 0.6 0.2 ± 0.2 1.25 ± 0.1 1.4 ± 0.5 0.6 ± 0.3 36.2 ± 13.8
88.7 ± 8.4 81.4 ± 10.9 56.8 ± 9.5
Particle Particle A Particle B Overall coating Splat A Splat B Splat Oxide
As-sprayed coating
Each average result is based on 10 measurements, with an exception of the results for Splat Oxide which are averaged from 3 measurements.
280
P. Niranatlumpong et al. / Surface & Coatings Technology 280 (2015) 277–281
Fig. 4. Particle A2, back scattered electron micrograph ×2500 mag.
an infrequent occurrence of the dendritic oxide inclusion within the coating. It may be that a thicker oxide with an observable dendritic structure had broken off in flight or had spalled off during coating deposition due to a thermal expansion coefficient mismatch with the splat. This can also explain the reduction in Y content in the coating. More frequently detected is a thinner Al–Y rich oxide around the splat, see Fig. 6. This Al–Y rich oxide does not exhibit a dendritic structure. It may be an oxide formed during the flight or formed in a second stage after deposition and before being covered by the next depositing splat. A small amount of Ni-rich oxide is also detected. The formation of the Ni-rich oxide suggests an increase in the oxygen partial pressure during the high temperature exposure.
Fig. 6. As-sprayed coating, back scattered electron micrograph ×200 mag.
Even though the as-sprayed coating contains only slightly lower amounts of Y with the as-received powder, some are in the form of an intersplat oxide which may not be available for protective oxide scale formation during service at high temperature. Some Y is contained within Splat A and can be used for oxide formation. Further study however is needed in order to determine whether the localised availability of Y affects the scale formation. 4. Conclusion This work has shown that the structure of an as-received NiCrAlY plasma spraying powder consists of γ (Ni solid solution) and β (NiAl) phases with Y segregation within the γ phase. During spraying, powder particles underwent different transformations depending on their flight path. Particle A group exhibits an increase in Al and Y content due to an incongruent evaporation. Accompanying β grain growth can sometimes be observed. There is no significant change in the particle size. Particle B group forms Al–Y rich oxide. This results in a depletion of Y in the particle. The particle in this group also shows a reduction in the particle size. The resulting coating thus shows a heterogeneous structure with respect to the distribution of Y. Acknowledgements This research is funded by the National Metal and Materials Technology Center, the National Science and Technology Development Agency, Thailand.
Fig. 5. Particle B1, back scattered electron micrograph ×3200 mag.
Fig. 7. Dendritic Al–Y rich oxide in as-sprayed coating, back scattered electron micrograph ×3200 mag.
P. Niranatlumpong et al. / Surface & Coatings Technology 280 (2015) 277–281
References [1] M.J. Pomeroy, Coatings for gas turbine materials and long term stability issues, Mater. Des. 26 (2005) 223–231. [2] M. Okazaki, High-temperature strength of Ni-base superalloy coatings, Sci. Technol. Adv. Mater. 2 (2001) 357–366. [3] M. Boidot, S. Selezneff, D. Monceau, D. Oquab, C. Estournès, Proto-TGO formation in TBC systems fabricated by spark plasma sintering, Surf. Coat. Technol. 205 (2010) 1245–1249. [4] J.R. Vargas Garcia, T. Goto, Thermal barrier coatings produced by chemical vapor deposition, Sci. Technol. Adv. Mater. 4 (2003) 397–402. [5] U. Schulz, K. Fritscher, M. Peters, D. Greuel, O. Haidn, Fabrication of TBC-armored rocket combustion chambers by EB-PVD methods and TLP assembling, Sci. Technol. Adv. Mater. 6 (2005) 103–110. [6] M. Karger, R. Vaßen, D. Stöver, Atmospheric plasma sprayed thermal barrier coatings with high segmentation crack densities: spraying process, microstructure and thermal cycling behaviour, Surf. Coat. Technol. 206 (2011) 16–23. [7] Y.C. Jung, T. Sasaki, T. Tomimatsu, K. Matsunaga, T. Yamamoto, Y. Kagawa, Y. Ikuhara, Distribution and structures of nanopores in YSZ-TBC deposited by EBPVD, Sci. Technol. Adv. Mater. 4 (2003) 571–574. [8] D. Strauss, G. Muller, G. Schumacher, V. Engelko, W. Stamm, D. Clemens, W.J. Quaddakers, Oxide scale growth on MCrAlY bond coatings after pulsed electron beam treatment and deposition of EBPVD-TBC, Surf. Coat. Technol. 135 (2001) 196–201.
281
[9] M. Seraffon, N.J. Simms, J. Sumner, J.R. Nicholls, The development of new bond coat compositions for thermal barrier coating systems operating under industrial gas turbine conditions, Surf. Coat. Technol. 206 (2011) 1529–1537. [10] S.A. Tsipas, Effect of dopants on the phase stability of zirconia-based plasma sprayed thermal barrier coatings, J. Eur. Ceram. Soc. 30 (2010) 61–72. [11] M.P. Schmitt, A.K. Rai, R. Bhattacharya, D. Zhu, D.E. Wolfe, Multilayer thermal barrier coating (TBC) architectures utilizing rare earth doped YSZ and rare earth pyrochlores, Surf. Coat. Technol. 251 (2014) 56–63. [12] K.A. Unocic, B.A. Pint, Oxidation behavior of co-doped NiCrAl alloys in dry and wet air, Surf. Coat. Technol. 237 (2013) 8–15. [13] B.G. Mendis, K.J.T. Livi, K.J. Hemker, Observations of reactive element gettering of sulfur in thermally grown oxide pegs, Scr. Mater. 55 (2006) 589–592. [14] R. Cho, Rare Earth Metals: Will We Have Enough?, State of the Planet Blog, The Earthe Institute, Columbia University, 2012. [15] B. Gleeson, W.H. Cheung, D.J. Young, Corros. Sci. 35 (1993) 639. [16] G. Espie, A. Denoirjean, P. Fauchais, J.C. Labbe, J. Dubsky, O. Schneeweiss, K. Volenik, In-flight oxidation of iron particles sprayed using gas and water stabilized plasma torch, Surf. Coat. Technol. 195 (2005) 17–28. [17] A. Vardelle, P. Fauchais, N.J. Themelis, Advances in thermal spray science and technology, in: C.C. Berndt, S. Sampath (Eds.),ASM International, Materials Park, OH, USA 1995, p. 175. [18] Phase diagrams of binary nickel alloys, in: P. Nash (Ed.), ASM International, Materials Park, OH, USA 1991, pp. 374–378.
Contents lists available at ScienceDirect
Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat
Loss of Y from NiCrAlY powder during air plasma spraying Panadda Niranatlumpong ⁎, Chalermchai Sukonkhet, Kittichai Ninon National Metals and Materials Technology Center, Pathumthani, Thailand
a r t i c l e
i n f o
Article history: Received 28 January 2015 Revised 14 July 2015 Accepted in revised form 6 September 2015 Available online 11 September 2015 Keywords: Plasma spraying NiCrAlY In-flight oxidation Y depletion
a b s t r a c t This study involves changes in microstructures and chemical compositions of NiCrAlY particles that occurred during an air plasma spraying of the powder. In this work the spraying particles were rapidly cooled and collected at a spraying distance just before the coating deposition in order to study the mid-flight transformation. Microscopy analysis of the particle cross-sections reveals that different structural and chemical transformations took place presumably depending on the flight path in relation to the plasma plume. Particle A group has a tendency to show some increase in Al and Y contents, possibly due to an evaporation of γ-(Ni solid solution) and β-(NiAl). Accompanying β grain growth can sometimes be observed. Particle B group often forms a large amount of Al–Y rich oxide due to an increase in oxygen activity during flight. This results in a significant reduction of Y in the particle. The coating produced using the same set of spraying parameter shows a heterogeneous structure with respect to the distribution of Y. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Yttrium has been added to NiCrAl alloy as a reactive element to produce a starting material for a bondcoat used extensively as part of a thermal barrier coating (TBC) for land-based gas turbine Ni alloy components [1,2]. Various techniques can be employed for TBC deposition including plasma spraying, physical vapour deposition and chemical vapour deposition [3–7]. In plasma spraying processes, the coating materials in powder form are deposited in a molten or semi-molten condition on a substrate to form a spray deposit. During a coating deposition, the spraying material is subjected to a temperature higher than 1000 °C for a short time. The function of the bondcoat is to protect the component from oxidation. The oxidation resistance is achieved mainly by a formation of a continuous protective Al2O3 scale during service [8, 9]. The scale helps to reduce the extent of corrosion of the underlying substrate thus allowing the engine to operate at a higher efficiency. After the deposition of the bondcoat, ZrO2 topcoat stabilised with other oxides such as Y2O3 and CeO2 can also be deposited to create TBC [10,11]. The exact role of Y in NiCrAlY has received attention in recent years. Previous research works have shown that Y can increase the adhesion of the thermally-grown oxide scale formed during TBC service at high temperature via sulphur gettering or oxide pegging mechanisms, thus enhancing the service life of the coating [12,13]. Scarcity of rare earth metals including Y has sent their prices soaring since before 2011. Their global demands continue nevertheless, encouraging more exploration and development of new mines. More mining of ⁎ Corresponding author. E-mail address: [email protected] (P. Niranatlumpong).
http://dx.doi.org/10.1016/j.surfcoat.2015.09.009 0257-8972/© 2015 Elsevier B.V. All rights reserved.
rare earth metals unavoidably leads to more environmental degradation and human health hazards [14]. It is vital then that the usage of the metals is justified and this requires a good understanding of their functions. This work investigates the loss of Al and Y from NiCrAlY during its inflight oxidation. This was achieved by rapidly cooling the spraying powder mid-flight. The powder was then characterised in order to obtain a better insight into its oxidation behaviour. The result was compared with that of a coating produced using the same spraying parameter. The findings will form a basis to the understanding of the mechanism of Y in prolonging the life of NiCrAlY coating at high temperature. 2. Experimental procedure An air plasma spraying experiment is carried out with an aim to collect the spraying bondcoat powder mid-flight at a deposition distance. The equipment setup is shown in Fig. 1. The bondcoat powder used in this work is Amdry 962 from Sulzer Metco with a particle size of 53–106 μm. The chemical composition is Ni–(21–23)Cr–(9– 11)Al–(0.8–1.2)Y in weight percent. The topcoat powder is Amperit 827.7 from H.C. Stark with a particle size of 45–90 μm. The chemical composition is ZrO2–7Y2O3 (wt.%). The bondcoat powder was sprayed using a 3MBII plasma spray gun (Sulzer Metco) and a set of parameter as shown in Table 1. A metal screen with a 5 mm opening was placed at a distance of 125 mm in front of the spray gun in order to allow a small amount of spraying powder through. As the spraying powder passed through the opening, an Argon jet of 0.1 MPa pressure was used as a medium to rapidly cool the powder. An earlier experiment was also carried out using
278
P. Niranatlumpong et al. / Surface & Coatings Technology 280 (2015) 277–281
3. Results and discussion 3.1. As-received powder The cross-sectional structure of the as-received powder is shown in Fig. 2. The particle surface is free of continuous oxide layer. The back scattered electron image shows that the alloy contains 2 major phases, γ (Ni solid solution) and β (NiAl). This is in accordance with the isothermal section of a Ni–Cr–Al ternary phase diagram as proposed by Gleeson et al. (1993) [15]. There is also Y segregation in γ phase, seen as light-coloured patches in the particle, see Fig. 3. The as-received powder shows structural consistency from one particle to another. Examples of the chemical composition of the individual particles using EDS are shown in Table 2. It should be noted that Y content is slightly higher than the powder specification. Fig. 1. Experiment setup.
3.2. Ar cooled powder
water as a cooling medium. Even though the water cooling process was simpler but the collected particles were found to be enveloped in thick crusts of oxide. This is believed to be because the collecting distance was specified at 125 mm which was close to the plasma jet thus some of the water may be dissociated while in contact with the particle, resulting in a rapid oxidation of the powder particle. Further experiment was therefore carried out using Ar as the cooling medium. The screen served multiple functions. It helped to shield the spray plume from interference from Ar jet which could cause an abnormal turbulence. Also Ar jet can cool the small amount of spraying powder more effectively and propel it into a cooling tube where the powder can be collected. It was originally intended that the screen will also act as a sampling window, allowing the powder collection from different sections across the plume. When the powder was collected after spraying however, it was found that the distinction in average powder appearances across the cross section of the plume was not obvious. This is because there is an arc fluctuation in a plasma spraying process making it difficult to predict the flight path of the particles. However, individual powder particles exhibited differences in appearance and composition. The bondcoat powder specimens were indicated as “as-received” for the starting powder and “Ar-cooled” for the sprayed powder, cooled and collected at 125 mm spraying distance. Accuraspray-g3 (Tecnar, Sulzer Metco) plume diagnostic equipment was used to determine the temperature of the powder at 125 mm spraying distance. The powder temperature was measured through thermal emission from the semimolten particles. This technique may produce inaccurate results due to its reliance on a database on the emissivity of materials. The result should be treated only as a guideline. Coating specimens were also produced with both bondcoat and topcoat deposition using the parameters in Table 1. The specimens were cross-sectioned. Scanning electron microscope and quantitative energy-dispersive X-ray spectroscopy (EDS) (JEOL6301) were used to characterise the specimens.
As the powder was passed through a high temperature H2/N2 plasma plume, each particle was subjected to different environments according to its flight path in relation to the plume. For example, if Particle A travels through the centre of the plume, it will be subjected to the highest temperature of that plasma plume in a reducing atmosphere until it leaves the centre or travels further from the nozzle where the temperature begins to drop and oxygen partial pressure increases [16]. Whereas Particle B travels along the edge of the plasma plume, it will be subjected to a lower temperature throughout the flight in a more oxidising atmosphere. Even though the plasma gas does not contain oxygen, there are turbulences at the edge of the plasma plume resulting in air entrainment. The presence of oxygen together with the high temperature of the plasma results in a highly oxidising environment [17]. Plume analysis reading indicates that the average temperature of the powder surface at the centre of the plume and at a spraying distance of 125 mm is 3078 ± 23 °C and the average powder velocity is 424 ± 4 m/s. At this high temperature it can be expected that Particle A and B will exhibit a different microstructure. The comparison between Particle A and B is an extreme example where the two particles take very different paths. It is more likely however that a particle take a non-linear path, for example, it may starts off at the centre of the plume where the temperature is high and very quickly moves to the fringe of the plume where oxygen is abundance. In this experiment, the powder was sprayed and collected at a spraying distance of 125 mm which is a distance generally used for this set of spraying parameter. In order to “freeze” the structure at the distance of 125 mm, an argon jet of 0.1 MPa pressure was used to intersect the powder in mid-flight and propel it into a cooling tube. A temperature sensor indicated that on average the ambient temperature
Table 1 Parameters for plasma spraying.
Hydrogen pressure Hydrogen flow rate (SCFm) Nitrogen pressure Nitrogen flow rate (SCFm) Current Potential difference Powder feed rate Spraying distance Hopper pressure
Bondcoat NiCrAlY
Topcoat ZrO2-Y2O3
0.34 MPa 1.2 × 10−4 m3 s−1 0.69 MPa 7.9 × 10−4 m3 s−1 500 A 70–80 V 0.78 g s−1 125 mm 0.55 MPa
0.34 MPa 1.2 × 10−4 m3 s−1 0.69 MPa 7.9 × 10−4 m3 s−1 500 A 65–70 V 0.71 g s−1 125 mm 0.9 MPa
Fig. 2. Back scattered micrograph of an as-received powder particle.
P. Niranatlumpong et al. / Surface & Coatings Technology 280 (2015) 277–281
279
Fig. 3. Back scattered micrograph of an as-received powder particle at ×10000 magnification and EDS line scanning result.
of the cooling tube is below 50 °C. Therefore in this work, we assume that the structure of the powder was successfully captured at the moment it was intersected by the argon jet. The sprayed powder shows a variety in physical changes including a small amount of coalescing and tearing of the particles which is not the focus of this paper and will not be discussed. The particles can be coarsely grouped into 2 groups namely Particle A and Particle B according to their microstructures and the oxide appearance. There are also some particles that fall between the two groups with a mixed microstructure. A particle would be classed as Particle A if there was no oxide or only a thin oxide layer on the particle surface. The thin oxide layer of generally less than 0.1 μm thick cannot be easily observed using SEM. It was suspected that Particle A travels near the centre of the plume so even though it was subjected to high temperature, the oxygen activity is presumably too low for it to form a significant amount of oxide [16], see Fig. 4. Table 2 shows the average chemical composition of the particle. Many particles show no observable chemical changes from the asreceived powder. A small number of particles however show a large change in chemical composition, in particular, increases in Al and Y. An increase in Y content is likely to be due to the selective evaporation of γ and β during spraying, leaving a larger proportion of Y in the particle. This is because Y has a negligible solubility in Ni and a large proportion of Y exists as a separate phase [18], see Fig. 3. Fig. 4 also shows β-NiAl grain growth. Generally the diameter of β-NiAl increases from 1–2 μm in the as-received powder to up to 4 μm in Particle A, varying from one particle to another. This may occur due to shifting in the chemical composition. As Y content increases the alloy becomes more brittle. Cracking can occasionally be observed on collected particles. The average measurements of the particle diameter indicate a slight decrease after spraying although this is not significant due to the large variation in the particle size in the as-received powder. A particle would be classed as Particle B if there was a large amount of oxide on the powder surface. This is very distinct on an SEM image. Particle B forms a thick oxide scale on one side of the particle. The oxide can remain intact on the collected particles or may break off during the rapid cooling of the particles, see Fig. 5. An EDS study shows that the oxide scale consists of 2 phases; dendritic Al-rich oxide and Y-rich oxide. Consequently, the proportions of Al and Y in the particle decrease due to the outward diffusion for the oxide formation.
The average chemical composition of the particle is shown in Table 2. The standard deviations are higher than those of the as-received powder indicating a large variation in the chemical composition from particle to particle. The result generally indicates a depletion of Y. The microstructure of the powder also reflects the change in its composition. The structure consists mostly of γ phase with β at grain boundaries. The segregation of Y is no longer observed, see Fig. 5. The average measurements of the particle diameter show a significant reduction in the particle size after spraying. The measurement on Particle B does not include the oxide scale. The large reduction may indicate evaporation of the particle during flight. 3.3. As-sprayed coating Microstructure of the coating produced using the same set of parameter as the cooled powder is shown in Fig. 6. The structure shows a large variation from one splat to another which is in agreement with the large variation found in the Ar cooled powder in Table 2. An overall composition of the coating indicates that the coating has slightly lower Y content than the starting powder. The deposited splats can be coarsely separated into two groups according to their chemical compositions and the microstructures; Splat A and Splat B. It is reasonable to assume that a collected Ar cooled powder of Particle A group, if instead allowed to deposit as a coating, will form a splat in the Splat A group. Splat A and Particle A generally has similar amounts of Al and Y as the as-received powder. Occasionally, Al and Y contents in Splat A and Particle A are much higher than those of the as-received powder resulting in high average values. These splats often show multi-phase structures consisting of β-NiAl grains in a matrix containing a varying amount of Ni, Cr and Y. Small cracks are observed in Splat A although infrequent. The presence of crack is possibly due to an increase in its brittleness as a result of a large proportion of Y. On the other hand, a powder from Particle B group if allowed to deposit as a coating will result in a splat depleted of Y. These splats, belonging to Splat B group, show a single phase splat containing significantly reduced percentage of Y when compared to the as-received powder, see Table 2 and Fig. 6. Al–Y rich oxide with a dendritic structure which was previously observed in Fig. 5 however was rarely present in the coating. Fig. 7 shows
Table 2 Average chemical composition of powder particles and coating splats. Average wt.%
Ni
Cr
Al
Y
Average particle diameter (μm)
Analysed area
As-received powder Ar-cooled powder
66.3 ± 0.3 62.3 ± 2.3 67.5 ± 1.4 66.8 ± 0.5 63.2 ± 2.3 67.2 ± 2.4 6.1 ± 3.4
22.1 ± 0.1 21.2 ± 1.6 21.5 ± 1.0 22.2 ± 0.3 20.9 ± 0.8 22.4 ± 0.7 6.9 ± 5.4
10.0 ± 0.3 14.8 ± 2.4 10.8 ± 1.8 9.7 ± 0.5 14.6 ± 2.4 9.8 ± 2.6 50.8 ± 12.5
1.5 ± 0.1 1.7 ± 0.6 0.2 ± 0.2 1.25 ± 0.1 1.4 ± 0.5 0.6 ± 0.3 36.2 ± 13.8
88.7 ± 8.4 81.4 ± 10.9 56.8 ± 9.5
Particle Particle A Particle B Overall coating Splat A Splat B Splat Oxide
As-sprayed coating
Each average result is based on 10 measurements, with an exception of the results for Splat Oxide which are averaged from 3 measurements.
280
P. Niranatlumpong et al. / Surface & Coatings Technology 280 (2015) 277–281
Fig. 4. Particle A2, back scattered electron micrograph ×2500 mag.
an infrequent occurrence of the dendritic oxide inclusion within the coating. It may be that a thicker oxide with an observable dendritic structure had broken off in flight or had spalled off during coating deposition due to a thermal expansion coefficient mismatch with the splat. This can also explain the reduction in Y content in the coating. More frequently detected is a thinner Al–Y rich oxide around the splat, see Fig. 6. This Al–Y rich oxide does not exhibit a dendritic structure. It may be an oxide formed during the flight or formed in a second stage after deposition and before being covered by the next depositing splat. A small amount of Ni-rich oxide is also detected. The formation of the Ni-rich oxide suggests an increase in the oxygen partial pressure during the high temperature exposure.
Fig. 6. As-sprayed coating, back scattered electron micrograph ×200 mag.
Even though the as-sprayed coating contains only slightly lower amounts of Y with the as-received powder, some are in the form of an intersplat oxide which may not be available for protective oxide scale formation during service at high temperature. Some Y is contained within Splat A and can be used for oxide formation. Further study however is needed in order to determine whether the localised availability of Y affects the scale formation. 4. Conclusion This work has shown that the structure of an as-received NiCrAlY plasma spraying powder consists of γ (Ni solid solution) and β (NiAl) phases with Y segregation within the γ phase. During spraying, powder particles underwent different transformations depending on their flight path. Particle A group exhibits an increase in Al and Y content due to an incongruent evaporation. Accompanying β grain growth can sometimes be observed. There is no significant change in the particle size. Particle B group forms Al–Y rich oxide. This results in a depletion of Y in the particle. The particle in this group also shows a reduction in the particle size. The resulting coating thus shows a heterogeneous structure with respect to the distribution of Y. Acknowledgements This research is funded by the National Metal and Materials Technology Center, the National Science and Technology Development Agency, Thailand.
Fig. 5. Particle B1, back scattered electron micrograph ×3200 mag.
Fig. 7. Dendritic Al–Y rich oxide in as-sprayed coating, back scattered electron micrograph ×3200 mag.
P. Niranatlumpong et al. / Surface & Coatings Technology 280 (2015) 277–281
References [1] M.J. Pomeroy, Coatings for gas turbine materials and long term stability issues, Mater. Des. 26 (2005) 223–231. [2] M. Okazaki, High-temperature strength of Ni-base superalloy coatings, Sci. Technol. Adv. Mater. 2 (2001) 357–366. [3] M. Boidot, S. Selezneff, D. Monceau, D. Oquab, C. Estournès, Proto-TGO formation in TBC systems fabricated by spark plasma sintering, Surf. Coat. Technol. 205 (2010) 1245–1249. [4] J.R. Vargas Garcia, T. Goto, Thermal barrier coatings produced by chemical vapor deposition, Sci. Technol. Adv. Mater. 4 (2003) 397–402. [5] U. Schulz, K. Fritscher, M. Peters, D. Greuel, O. Haidn, Fabrication of TBC-armored rocket combustion chambers by EB-PVD methods and TLP assembling, Sci. Technol. Adv. Mater. 6 (2005) 103–110. [6] M. Karger, R. Vaßen, D. Stöver, Atmospheric plasma sprayed thermal barrier coatings with high segmentation crack densities: spraying process, microstructure and thermal cycling behaviour, Surf. Coat. Technol. 206 (2011) 16–23. [7] Y.C. Jung, T. Sasaki, T. Tomimatsu, K. Matsunaga, T. Yamamoto, Y. Kagawa, Y. Ikuhara, Distribution and structures of nanopores in YSZ-TBC deposited by EBPVD, Sci. Technol. Adv. Mater. 4 (2003) 571–574. [8] D. Strauss, G. Muller, G. Schumacher, V. Engelko, W. Stamm, D. Clemens, W.J. Quaddakers, Oxide scale growth on MCrAlY bond coatings after pulsed electron beam treatment and deposition of EBPVD-TBC, Surf. Coat. Technol. 135 (2001) 196–201.
281
[9] M. Seraffon, N.J. Simms, J. Sumner, J.R. Nicholls, The development of new bond coat compositions for thermal barrier coating systems operating under industrial gas turbine conditions, Surf. Coat. Technol. 206 (2011) 1529–1537. [10] S.A. Tsipas, Effect of dopants on the phase stability of zirconia-based plasma sprayed thermal barrier coatings, J. Eur. Ceram. Soc. 30 (2010) 61–72. [11] M.P. Schmitt, A.K. Rai, R. Bhattacharya, D. Zhu, D.E. Wolfe, Multilayer thermal barrier coating (TBC) architectures utilizing rare earth doped YSZ and rare earth pyrochlores, Surf. Coat. Technol. 251 (2014) 56–63. [12] K.A. Unocic, B.A. Pint, Oxidation behavior of co-doped NiCrAl alloys in dry and wet air, Surf. Coat. Technol. 237 (2013) 8–15. [13] B.G. Mendis, K.J.T. Livi, K.J. Hemker, Observations of reactive element gettering of sulfur in thermally grown oxide pegs, Scr. Mater. 55 (2006) 589–592. [14] R. Cho, Rare Earth Metals: Will We Have Enough?, State of the Planet Blog, The Earthe Institute, Columbia University, 2012. [15] B. Gleeson, W.H. Cheung, D.J. Young, Corros. Sci. 35 (1993) 639. [16] G. Espie, A. Denoirjean, P. Fauchais, J.C. Labbe, J. Dubsky, O. Schneeweiss, K. Volenik, In-flight oxidation of iron particles sprayed using gas and water stabilized plasma torch, Surf. Coat. Technol. 195 (2005) 17–28. [17] A. Vardelle, P. Fauchais, N.J. Themelis, Advances in thermal spray science and technology, in: C.C. Berndt, S. Sampath (Eds.),ASM International, Materials Park, OH, USA 1995, p. 175. [18] Phase diagrams of binary nickel alloys, in: P. Nash (Ed.), ASM International, Materials Park, OH, USA 1991, pp. 374–378.