Influence of ceramic top coat and thermally grown oxide microstructures of air plasma sprayed Sm2SrAl2O7 thermal barrier coatings on the electrochemical impedance behavior

Accepted Manuscript Influence of ceramic top coat and thermally grown oxide microstructures of air plasma sprayed Sm2SrAl2O7 thermal barrier coatings on the electrochemical impedance behavior

T. Baskaran, Shashi Bhushan Arya PII: DOI: Reference:

S0257-8972(18)30309-8 doi:10.1016/j.surfcoat.2018.03.058 SCT 23238

To appear in:

Surface & Coatings Technology

Received date: Revised date: Accepted date:

2 November 2017 2 March 2018 22 March 2018

Please cite this article as: T. Baskaran, Shashi Bhushan Arya , Influence of ceramic top coat and thermally grown oxide microstructures of air plasma sprayed Sm2SrAl2O7 thermal barrier coatings on the electrochemical impedance behavior. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Sct(2017), doi:10.1016/j.surfcoat.2018.03.058

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Influence of ceramic top coat and thermally grown oxide microstructures of air plasma sprayed Sm2SrAl2O7 thermal barrier coatings on the electrochemical impedance behavior T. Baskarana, Shashi Bhushan Aryaa,* a

Department of Metallurgical and Materials Engineering

PT

National Institute of Technology Karnataka, Surathkal, Mangaluru-575 025, India

RI

*Corresponding author is at: Department of Metallurgical and Materials Engineering,

SC

National Institute of Technology Karnataka, Surathkal, Mangaluru-575 025, India E-mail address: [email protected] (Shashi Bhushan Arya)

AC

CE

PT E

D

MA

NU

Tel: +91-0824-2473754 and (Mobile): +91-8762526500

1

ACCEPTED MANUSCRIPT

Abstract Electrochemical impedance spectroscopy (EIS) technique is used to examine the top coat and thermally grown oxide (TGO) microstructures of Samarium Strontium Aluminate (SSA) thermal barrier coatings (TBCs) after exposed to pre-oxidation and oxidation at 1050 and 1100 ºC,

PT

respectively. EIS spectra showed that the three relaxations frequencies in Bode plot corresponded to SSA top coat, TGO, and TGO-bond coat interface. A significant reduction in polarization resistance

RI

of SSA top coat and increase in capacitance for different pre-oxidation times of 10, 20 and 30 h are

SC

being found due to increasing the defects (pores and cracks) of about 13, 29 and 45%, respectively at 1050 ºC. The reduction in bi-axial residual stresses was calculated to be about 57% as a

NU

consequence of decrement in top coat resistance. The growth of SSA TGO was found from 10 to 20

MA

h of pre-oxidation treatment caused to decrease the capacitance which indicates the presence of highly enriched α-Al2O3 at the bond coat-top coat interface. The highest charge transfer resistance

D

and lowest capacitance were found to be about 0.48×106 Ohm cm2 and 1.1 nF cm-2 respectively for

PT E

20 h of pre-oxidation which could reflect the overall impedance kinetics of TBC system (SSA top coat and TGO) at the TGO-bond coat interface.

CE

The impedance responses of SSA top coat, TGO, and TGO-bond coat interface were reduced drastically after oxidation at 1100 ºC for 10 h pre-oxidized specimen as compared to 20 and

AC

30 h due to the compositional change of pure α-Al2O3 based TGO into more conductive NiCr2O4. The lowest diffusion coefficient, DCr 3 in the NiO lattice which reduced the formation of metal ion vacancies at the TGO-top coat interface caused to exhibit higher TGO resistance for 20 h preoxidized specimens over 10 and 30 h after oxidation at 1100 ºC. Keywords: Oxidation, Thermally grown oxide, Impedance, Charge transfer resistance, Diffusion coefficient 2

ACCEPTED MANUSCRIPT

1. Introduction Developing oxidation resistance thermal barrier coatings (TBCs) is one of the most important subjects in today’s gas turbine industry. In general, TBCs composed of bi-layers including MCrAlY bond coat (M=Ni or Co) and ceramic top coat (usually Yttria stabilized

PT

zirconia-YSZ) [1, 2]. A thermally grown oxide (TGO) is formed at the interface of ceramic top coat and bond coat in the oxygen-rich atmosphere at a higher temperature (900-1200 ˚C). TGO mainly

SC

RI

consists of aluminum oxide which reduces the rate of oxidation and hot corrosion [3, 4]. The microstructure of air plasma sprayed ceramic layer consists of un-melted particles

NU

(grains), completely melted splats (grain boundary), intersplat pores and cracks. The microstructure of top coat is mainly controlled by the formation of thermal stresses, porosity, and cracks at

MA

elevated temperature (500-1300 ˚C) [5]. The presence of pores and cracks in the top coat microstructure were found to be significantly increased the strain tolerance of TBCs and thereby,

D

reduced the residual thermal stresses [6]. A formation of stable, a dense α-Al2O3 TGO is beneficial

PT E

for protection against oxygen diffusion but mixed oxides (NiO, Cr2O3, and NiCr2O4) with a discontinuous α-Al2O3 exhibited a weak protection to the substrate [7]. The TBCs pre-oxidized at

CE

1050 ºC under an oxygen partial pressure of 1.01325×10-9 to 1.01325×10-10 Pa. which showed lowest mixed oxides content in TGO and longest thermal cyclic life [8]. The diffusion of bond coat

AC

metals (M2+) and oxygen (O2-) ions would be much faster when there is a formation of metal excess and oxygen deficit compounds at the bond coat and ceramic interface. Particularly, the existence of p-type oxides such as NiO and Cr2O3 at the interface increases the diffusion of O2- ions through the vacancies in the oxide lattice at higher temperature [9]. Along with these mixed oxides, sintering of ceramic top coat and residual stresses between layers led to delamination of ceramic layer [10]. In literature, there are several techniques have been 3

ACCEPTED MANUSCRIPT

proposed to access the damage and life of TBCs namely, infrared thermography [11, 12], acoustic emission [13], and photoluminescence piezospectroscopy [14]. In addition to these techniques, the Electrochemical impedance spectroscopy (EIS) has also been found to be very useful in evaluating the structural changes of TBCs such as thickness reduction, the formation of porosities, TGO

PT

growth, defects initiation, and coating delamination as the conductive electrolyte can easily penetrate into former defects [15,16]. These degradations can be measured non-destructively using

RI

an impedance kinetic (resistance and capacitance) changes of both the ceramic top coat and TGO

SC

layer.

NU

Evaluation of microstructure of YSZ TBCs after thermal cycling was studied by Song et al. using EIS technique and they found the resistivity of alumina-based TGO scale increases with

MA

increasing oxidation time, due to increasing TGO growth at the interface [17]. Jayaraj et al. have reported EIS study of YSZ TBCs after isothermal and cyclic oxidation test at 1100 ºC. It was

D

concluded that large-scale spallation of top coat and TGO layer which could provide more

PT E

electrolyte contact (0.01 molar K3Fe(CN)6/K4Fe(CN)6.3H2O) with metallic bond coat surface, caused to an abrupt increase in the capacitance of top coat and TGO [18]. Zhang et al. have

CE

developed correlation between impedance kinetics such as pore resistance, ceramic resistance, and ceramic capacitance with the morphological structures of YSZ TBCs. The electrical resistance of

AC

top coat is directly proportional to its thickness, the presence of porosity, pore size, and shape distributions [19]. EIS analysis of 7YSZ TBCs during high-temperature oxidation test at 950 ºC was examined by Chen et al. They have found that at the beginning of oxidation time (<50 h) YSZ grain boundary resistance increased due to increasing cracks. Further increasing oxidation time (>150 h) showed decreasing grain boundary resistance mainly due to sintering of ceramic top coat [20]. A correlation between Young’s modulus of the ceramic top coat and electrochemical impedance was reported by Gomez-Garcia et al. The YSZ top coat crack resistance (Rcrack) was increased up to a 4

ACCEPTED MANUSCRIPT

critical value 19×103 Ohm after exposure to isothermal oxidation test at 1050 ºC for 75 h. A reduction in Rcrack was observed to be about 26% after exposure to 1050 ºC for 336 h. A linear relationship has been found between Rcrack and Young’s modulus as a consequence of sintering [21]. Moreover, at higher temperatures (more than 1200 ºC), the accelerated sintering of YSZ

PT

would degrade the strain tolerance of coatings which are not beneficial for TBC applications [22]. Hence, a search for alternate TBCs candidate material with low thermal conductivity, higher

RI

melting point, corrosion resistant and phase stability over conventional YSZ has been motivated in

SC

the last decade [23].

NU

Among the several ceramics that have been proposed as an alternate TBC material, most recently, the perovskite-type single phase Sm2SrAl2O7 (SSA) ceramic was also developed and

MA

explored as a top coat material over conventionally used YSZ due their excellent thermo-physical properties and phase stability [[24]; [25]; [26]]. The pre-oxidized SSA TBCs showed significant

D

oxidation resistance about 65% over YSZ after oxidation test at 1100 ºC in the air [27]. The

PT E

developed SSA based TBCs considerably reduced the oxidation of underlying substrate due to the presence of compositionally pure and stable α-Al2O3 TGO at the interface. However, the failure of

CE

SSA TBCs was also observed at the interface of the ceramic top coat and TGO due to the formation of spinel oxides and SmAlO3. The EIS tool was used in this study to examine the role of top coat

AC

microstructure and the formation of mixed oxides at the TGO-top coat interface of SSA TBCs after oxidation test at 1100 ºC in the air. The present research mainly deals with coated SSA based TBCs by air plasma spray process, role of pre-oxidized and oxidized top coat and TGO microstructures on the impedance kinetics. The presence of defects such as pores and cracks in the top coat on the bi-axial residual stress formation and the formation of non-stoichiometric oxides on the diffusion of oxygen ions at the interface were also analyzed with respect to impedance behavior. 5

ACCEPTED MANUSCRIPT

2. Experimental procedure 2.1.

Substrate preparation, spray feed stocks, pre-oxidation and oxidation study

The specimens were cut into 20 mm × 20 mm × 5 mm size from a wrought Ni-based super alloy (Inconel 718) with nominal composition (wt.%) of Ni-17.65Fe-17.60Cr-5.10Nb-2.98Mo-

PT

0.80Ti-0.48Co-0.24Cu-0.30Mn. All the samples are ground properly and subjected to alumina grit

RI

(Grit 45) blasting in order to improve the adherence of the bond coat with the substrate. The average

SC

surface roughness (Ra) and maximum peak height profile (Rt) of the substrates were found about 6±0.1µm and 17±0.1µm respectively for grit blasting (Grit 45) alumina. NiCrAlY (wt.%: 68.29 Ni,

NU

21.41 Cr, 10.27 Al, and 0.93 Y) bond coat and SSA top coat were sprayed on Inconel 718 superalloy substrate using atmospheric plasma spray (APS) technique using a 55 kW F4-MB

MA

plasma gun. The bond coat and top coat thickness are around 50 and 100 µm, respectively. The precise plasma spray parameters for both NiCrAlY bond coat and SSA ceramic top coat were

D

reported in Table 1. The standard plasma spray parameters of conventional YSZ TBCs was taken

PT E

from the literature as a reference and used for the deposition of SSA TBCs [27, 28]. The particle size distributions of NiCrAlY and SSA powders were found about -112/+53 and -75/+2 µm

CE

respectively. The particle size of bond coat NiCrAlY powder (PAC 9620 AM) was provided by M/s. Powder Alloy Corporation, Ohio, USA. The particle size of synthesized SSA top coat powder

AC

was determined by sieve analysis according to ASTM B-214. The particle morphology of bond coat NiCrAlY and top coat SSA powders were shown in Fig. 1. The spherical and cuboidal shape of NiCrAlY and SSA respectively were observed. The pre-oxidation test specimens was performed on SSA TBCs at 1050 ºC for 10, 20, and 30 h in argon (99.99% purity) atmosphere to get TBCs with different TGO thicknesses prior to oxidation test under air. Isothermal oxidation study was conducted at 1100 ºC for 15 h in air using a resistance controlled tubular furnace with ±1 ºC accuracy. 6

ACCEPTED MANUSCRIPT

2.2.

Characterization of coating before and after oxidation test

The oxidized specimens are mounted in epoxy resin to prevent the spalling of ceramic top coat from the substrate. Further, it was subjected to section by low-speed diamond saw to retain the TBCs integrity during the cutting process. A detailed phase analysis of the oxidized TBC samples

PT

was carried out by X-ray diffractometer (JEOL DX GE-2P, Japan) with Cu Kα (λ=1.514 Å)

RI

radiation at an accelerating voltage of 30 kV and 20 mA. Microstructures of SSA TBCs specimens

SC

were examined using JEOL 6380 SEM and High-resolution SEM (EVO 18, Carl Zeiss, Germany) equipped with an energy dispersive X-ray spectroscope (EDS).

Electrochemical impedance measurement of SSA TBCs

NU

2.3.

MA

The conventional three electrodes system namely, reference electrode (saturated calomel electrode-SCE), a platinum (counter electrode) and working electrode (Oxidized TBC specimen)

D

were employed for EIS measurement. Oxidized TBC specimen was prepared in the following

PT E

manner. One side of the oxidized specimen was mechanically polished to remove the TBC layer and acting as one electrode. The TBC side of the specimen was immersed in an aqueous electrolyte

CE

containing 0.01 molar K3Fe(CN)6/K4Fe(CN)6.3H2O. Particularly, this electrolyte was taken because of its highly reversible exchange current density and minimal interference with the system [21]. EIS

AC

test was conducted using potentiostat with an impedance analyzer (Biologic, SP-150, France), which is computer controlled. Spectral analysis or curve fitting was performed using EC-Lab impedance analysis software attached to the instrument to get the electrical properties (resistance and capacitance) of TBC system. In the EIS measurement, the AC amplitude of 20 mV and frequency in the range of 10-3 Hz to 105 Hz was employed.

7

ACCEPTED MANUSCRIPT

2.4.

Residual stress measurement of the SSA TBCs

The residual stresses of the SSA TBCs were analyzed using micro-Raman spectroscopy (Jobin-Yvon) with an argon ion laser (532 nm) and exciting light source with 4 mW power. The residual stresses were calculated based on peak shifts of Raman bands due to stresses and

PT

conversion of these band shifts to stress values by the use of tensor coefficients [29, 30]. The

Average stress tensor σ = -b∆Ω

SC

(1)

RI

relation between the bi-axial residual stress and peak shift of a given Raman band by

NU

Where ∆Ω =Ω ̶ Ωo and Ωo are the position of selected Raman line for the stress-free sample. In the present work, the strong Raman line at 320 cm-1 was used to determine the residual stress and

MA

variation of ∆Ω at 320 cm-1 was used to indicate the variation of stress in the TBCs.

3. Results and discussion

D

Surface and cross-sectional microstructures of the as-sprayed SSA coatings

PT E

3.1.

The surface and cross-sectional SEM images of as sprayed SSA coating are displayed in

CE

Fig. 2 (a and b). The presence of inter-splat cracks, semi-melted, and un-melted particles could be observed on the surface of top coat due to the incomplete melting of large particles during the

AC

plasma spraying process [31]. Globular pores can also be visibly seen in the cross-sectional image resulting from the dissolved gasses in the molten particles. The ceramic top coat microstructures containing defects such as pores and micro-cracks have a great influence on thermal insulation and thermal shock resistance of the TBCs at higher temperature [32]. Similarly, the bond coat microstructure also contains pores but comparatively lesser than ceramic top coat due to complete melting of NiCrAlY particles during the deposition process. The complete melting of NiCrAlY

8

ACCEPTED MANUSCRIPT

alloy can be achieved easily over SSA ceramic because the melting point of metallic NiCrAlY (~1400 ºC) is quite lower than SSA ceramic (~1950 ºC) [26].

The EDS elemental analysis was carried out in the selected regions of SSA top coat and NiCrAlY bond coat to know the chemical composition after coating. The selected region in Fig. 2

PT

(a) composed of Sm, Sr, Al and O (with composition of 18.54 at.% Sm, 8.08 at.% Sr, 16.79 at.% Al

RI

and 56.59 at.% O) and region in Fig. 2 (b) composed of Ni, Cr, Al and O (with composition of

SC

57.84 at.% Ni, 17.68 at.% Cr, 14.73 at.% Al and 9.75 at.% O). Therefore, EDS analysis confirmed the atomic ratio of SSA (Sm2SrAl2O7) ceramic which is followed in the required range of 2:1:2:7.

NU

The considerable amount oxygen was also found in the bond coat region (~10 at. %) due to the

3.2.

MA

oxidation of metallic NiCrAlY during atmospheric plasma spraying process [33].

Microstructure and electrical property relation of pre-oxidized SSA TBCs

D

3.2.1. Electrochemical impedance spectra of pre-oxidized SSA TBCs

PT E

EIS study is usually represented in Nyquist and Bode plots, in which Nyquist plot consists of a real part and imaginary part of impedances. Bode plot, represents modulus of impedance and

CE

phase angle with respect to frequency. In order to find out the electrical property changes of a different layer, it is necessary to design a suitable AC equivalent circuit model to fit the generated

AC

experimental spectra. Since the electrode surface is not uniform; an ideal capacitor cannot be used to describe the capacitive behavior of SSA TBCs. In such case, constant phase element (CPE) can be used to describe the non-ideal capacitance behavior [[34]; [35]; [36]]. For non-ideal capacitors, the capacitance (C) can be calculated by [37, 38] C = R (1-n)/n A1/n

(2)

9

ACCEPTED MANUSCRIPT

Where, R is the resistance and A is the parameter of CPE. The CPE behaves like a pure resistor and capacitor, when n lies between 0 and 1, respectively.

Nyquist and Bode plots of SSA specimens subjected to different pre-oxidation treatment are represented in Fig. 3 (a and b) and (c), respectively. In Fig. 3 (a and b), each plot consists of two

PT

semicircles with different diameters, the first semicircle corresponds to grain, grain boundary of

RI

SSA top coat, TGO and second semicircle correspond to electrode effect due to TGO-bond coat interface charge transfer resistance. Fig. 3 (c) represents Bode plot which is classified as three

SC

different zones such as high frequency of 105-102 Hz, medium frequency of 100-102 Hz and low

NU

frequency of 10-1-10-3 Hz corresponds to grain and the grain boundary of SSA TBCs, TGO layer and charge transfer response at the TGO-bond coat interface respectively. These phenomenons have

MA

been observed in many literatures on an impedance spectroscopy study of the high-temperature

D

oxidation of thermal barrier coatings [[20]; [21]; [22]].

PT E

A clearly two relaxation process is observed for as deposited SSA top coat due to the absence of TGO layer at the junction of top coat and bond coat. Significantly higher impedance was noticed for as deposited SSA TBCs over pre-oxidized specimens due to the absence of

CE

interconnected cracks and pores [Fig. 3 (a)]. A magnified view of high frequency plot of Fig. 3 (a)

AC

is shown in Fig. 3 (b). The SSA TBCs pre-oxidized for 20 h at 1050 ºC showed the largest impedance than 10 and 30 h treated specimens. According to Fig. 3 (c), the pre-oxidation treatment led appreciable changes in the frequency response of the SSA TBCs. In contrast to as-deposited SSA TBCs, the pre-oxidized SSA TBCs showed good frequency distribution in the region from 105 to 101, 101 to 100 and 100 to 10-3 Hz. It is also evident that both fitted and measured Nyquist impedance results of as deposited and pre-oxidized SSA TBCs are very well consistent [Fig. 3 (a)]. The obtained fitting error of impedance simulation was less than 10 %. The impedance plot was 10

ACCEPTED MANUSCRIPT

fitted based on a chosen model of 3 R/CPE in the series connection, which corresponds to the SSA top coat, TGO and electric charge transfer resistance as shown in Fig. 4 (a-c). However, two R/CPE model was used for fitting of as deposited SSA TBCs due to the absence of TGO layer [39].

PT

3.2.2. Role of micro-structural defects in SSA top coat on EIS

The cross-sectional SEM microstructure of SSA top coat after exposure to 1050 ºC for

RI

different pre-oxidation duration is presented in Fig. 5 (a-c). It is clearly revealed that microstructure

SC

consists of a ceramic top coat, TGO layer (black dark region) followed by the bond coat and substrate. Some cracks and porosities can be clearly seen in the microstructure of SSA top coat

NU

owing to the formation thermal stresses due to TGO growth during pre-oxidation treatment at high

MA

temperature. TGO layer has formed due to the oxidation of metallic NiCrAlY bond coat and other top coat property changes at 1050 ºC. Cracks were exhibited at 10 h is found to be lower [Fig. 5 (a)]

D

over the higher exposure time of 30 h [Fig. 5 (c)]. The cross sectional SEM images (processed by

PT E

Image J software) of pre-oxidized SSA TBCs are shown in Fig. 6 (a-c). The porosity measurement shows the typical pore size distributions in the range of 0.1 - 4 µm [Fig. 6 (d-f)]. The fine pores less than 1 µm correspond to incomplete contact between splats and the micro-cracks. The large pores

CE

(size < 1 µm) arises by imperfect layer built-up by the molten droplet during the deposition process

AC

[40]. More fine pores and cracks could be seen for 30 h treatment as compared to 10 and 20 h, it may be due to higher pre-oxidation duration. It was also observed that the number of cracks and pores increased and a number of small cracks and pore merge together to form larger cracks in the top coat after 30 h of exposure. The calculated defects in the form of pores and cracks have found to be 13, 29, and 45% respectively for 10, 20, and 30 h of pre-oxidation at 1050 ºC. From Nyquist impedance plot [Fig. 3 (a and b)], it is found that the diameter of the first semicircle is gradually increased from 10 to 20 h and then decreased due to increasing percentage of defects in the top coat. 11

ACCEPTED MANUSCRIPT

The Bode plot [(Fig. 3 (c)] shows a shift of phase angle (towards low frequency) in the range 101100 Hz; it indicates impedance increases with increasing pre-oxidation time from 10 to 20 h and a significant reduction at higher exposure time of 30 h. A similar type of results was reported by Song et al. that impedance increased gradually from 15 to 100 cycles and then to 250 cycles of heating

PT

and cooling at 1030 ºC and 200 ºC, respectively. But the specimen subjected to 400 cycles, the impedance decreased drastically due to degradation of TGO in terms of changes of microstructures

SC

RI

and compositions [17].

The electrochemical impedance kinetic parameters were obtained from fitting of

NU

experimental EIS data using aforesaid equivalent circuits (Fig. 4). The electrical properties such as polarization resistance (RSSA) and capacitance (CSSA) of SSA top coat at the different pre-oxidation

MA

times are reported in Table 2. It clearly showed that gradual decrease in RSSA from 10 to 30 h due to increase in percentage of defects in the ceramic top coat. At the same time, a gradual decrease in

D

CSSA is observed up to 20 h, and then there is an abrupt increment indicates failure of top coat

PT E

occurs only after exposure to 30 h. Byeon et al. have studied EIS of multilayered YSZ TBCs by atmospheric plasma spray and electron-beam physical vapor deposition (EB-PVD) techniques.

CE

They found that an increasing oxidation time led the initial increase in resistance of YSZ and TGO. But further increase in time caused to decrease the resistance of YSZ and TGO with an abrupt

AC

increase in capacitance due to generation of defects. These results are good agreement with the present results obtained for SSA TBCs. The resistivity of SSA top coat can be evaluated by [16]

R

 S

(3)

Where R is the resistance; ρ is resistivity; δ is TGO thickness and S is the area of the electrode. The calculated resistivity values of pre-oxidized SSA top coat were found to be 1.4×105, 6.3×104 and 12

ACCEPTED MANUSCRIPT

5×103 Ohm cm, for 10, 20, and 30 h respectively. In general, the electrochemical impedance response of thermal barrier coatings depends on the penetration of conductive electrolyte through the pores of ceramic top coat layer [41]. The effective resistivity of SSA coatings during impedance measurement can be expressed as ρeff =ρSSAVSSA + ρelectrolyteVelectrolyte

PT

(4)

RI

Where, ρSSA and ρelectrolyte are the resistivities of SSA ceramic and electrolyte (~10-2-10-3 Ohm cm),

SC

respectively [19]. VSSA and Velectrolyte denote the volume fraction of SSA and electrolyte, respectively. The resistivity gradually decreased from 10 to 30 h due to the higher volume of

NU

conductive electrolyte penetration through more pores and cracks containing SSA top ceramic layer. After 30 h, SSA top ceramic layer becomes porous and contributes to lower effective

MA

resistivity with the decrease in volume fraction of SSA according to Eq. (4).

D

3.2.3. Growth kinetics and compositional changes of SSA TGOs

PT E

The growth of TGO and its composition is most important for the understanding of oxidation characteristics of TBCs [42]. The continuous TGO layer could be seen at the interface

CE

between NiCrAlY bond coat and SSA top coat after exposure to 1050 ºC for different pre-oxidation times of 10, 20, and 30 h in the argon atmosphere [Fig. 5]. The thickness of TGO was measured

AC

using SEM micrographs and found to be about 4±1, 5±1 and 6±2 µm. It is also observed that the growth of oxide layers increased with an increasing pre-oxidation time which is more consistent with the previous investigations [43, 44]. The growth of TGO layer was found to be dependent on the oxidation temperature and duration due to the phase transformations from γ to θ to α-Al2O3 phases (γ-Al2O3 → θ-Al2O3 → α-Al2O3) at 1000-1200 ºC [45]. In this work, an instantaneous increment of TGO thickness was found to be about 4 µm in the onset of pre-oxidation treatment corresponded to nucleation and growth of α-Al2O3 at different sites of bond coat and top coat 13

ACCEPTED MANUSCRIPT

interface. This process was continued until a thin layer of TGO covering the entire bond coat surface up to 10 h. Once the TGO layer is formed at the interface, it protects the bond coat from further oxidation. Only a slight increment in TGO thickness from 4 to 5 µm was observed at 20 h due to the presence of protective Al2O3 scale. When the oxide scale thickens from 5 to 6 µm at the

PT

end of 30 h, the growth stresses caused to form defects such as porosity and cracks inside the TGO

RI

layer.

SC

The EDS analysis was done at TGO-top coat interface of TBCs to evaluate the compositional changes during pre-oxidation treatment as shown in Fig. 5. In all the cases the atomic

NU

percentage of Al and O were found to be similar to pure α-Al2O3; however, Ni content varied in the TGO-top coat interface as 7.05, 1.17 and 1.30% for 10, 20 and 30 h of pre-oxidation, respectively.

MA

Similarly, the atomic percentage of Cr also varied from 2.27 (10 h) to 0.53 (30 h). Hence, it is indicating that the increasing the TGO growth reduces the diffusion of Ni and Cr by 80 and 76%,

D

respectively from 10 to 30 h of pre-oxidation. The presence of Ni and Cr is lower for 20 and 30 h

PT E

treated specimens as compared to 10 h of pre-oxidation, which means that the purity of TGO of SSA TBCs is quite superior for 20 and 30 h than 10 h of pre-oxidation. It is interesting further to

CE

see that the role of TGO treated for 20 and 30 h of pre-oxidation on the oxidation performance of

AC

SSA TBCs during extended thermal exposure in air. The shifting of relaxation frequency zone (100-101 Hz) for TGO layer is observed in Fig. 3 (c), which indicates the impedance of TGO layer increases with increasing pre-oxidation time from 10 to 20 h. In other words, an increasing pre-oxidation duration increases the TGO thickness which caused to increase impedance from 10 to 20 h according to Eq. 3. But further increasing the time to 30 h caused to reduce the impedance behavior due to microstructure and compositional changes of

14

ACCEPTED MANUSCRIPT

TGO layer. The relation between the maximum of the absolute values of phase angles (Ф) and respective frequencies (f) with polarization resistance (Rp) are given by the following Eqs. (5 and 6)

1 2 CRp

(5)

Rp

(6)

1/2

2  Rs ( Rs  R p ) 

RI

max  arctan

 Rp  1    Rs 

PT

f max 

SC

According to Eq. 5, it is observed that an increase of Rp at constant C and Rs moves fmax to lower frequency. Similarly, an increase of Rp at constant Rs moves Фmax to higher phase angle [46].

NU

In the present study, the obtained modulus of impedance and corresponding phase angle values were plotted with respect to frequency as shown in Fig. 7 (a). The highest modulus of impedance

MA

was found at 20 h for both TGO and charge transfer region as compared to 10 and 30 h of preoxidation, which indicates a reduction in Rp from 20 to 30 h caused to shifts the frequency from fb2

D

to fb3. The variation in phase angle (ФTGO) with respect to polarization resistance (RTGO) were also

PT E

seen at the frequency of 100 Hz and 10-3 Hz in TGO zone and TGO-bond coat interface, respectively which is reported in Table 3. The highest phase angle was found for 20 h (Ф = 27º)

CE

showed RTGO is about 4.3×104 Ohm cm2 due to the presence of pure Al2O3 protective layer at the interface. The lowest phase angle was found for 30 h (Ф = 14º) showed RTGO is about 6.1 × 103

AC

Ohm cm2 mainly due to the presence of conductive oxides and defects in the TGO. In general, the shift of phase angle towards 90º indicates the presence of the resistive type of oxide film and vice versa. Zhang et al. have investigated anodization of two industrial grades aluminium alloys (AA 6082 and 7075) using EIS technique. The anodization result showed as two layer structures on Al alloys which consist of a barrier inner layer and porous outer layer. The EIS fitting was done using CPE1/R1 and CPE2/R2 series model which represents outer layer close to the electrolyte and inner layer close to the metal substrate, respectively. The inner layer exhibited high capacitance as 15

ACCEPTED MANUSCRIPT

compared to outer layer due to thin resistive anodic oxide [47]. So, based on these observations it is understood that specimen pre-oxidized at 20 h exhibited the pure resistive type of TGO and capacitive type was observed in the case of 10 and 30 h. The variation in resistance and capacitance of the SSA TGOs are also reported in Table 2 as

PT

a function of pre-oxidation time. Capacitance behavior of SSA TGOs decreases with increasing time and an abrupt increase in capacitance for 30 h indicates there may be a failure or compositional

RI

changes in the TGO region. The resistivity of SSA TGOs was found to be 1.5 × 108, 1.7 × 108, and

SC

2.05 × 107 Ohm cm for 10, 20, and 30 h, respectively. The maximum charge transfer resistance (Rct)

NU

was found at 20 h over 10 and 30 h of pre-oxidation. Higher capacitance for the charge transfer effect (Cct) was noticed for 10 and 30 h, where the failure actually occurred as reported in Table 2.

MA

A close observation of Table 2 indicates that the resistance from charge transfer effect changes in the same order of resistance from SSA top coat and TGO. Hence, the resistance from charge

PT E

D

transfer effect can be used to reflect the changes of overall resistance in the TBCs components.

The thickness of TGO was calculated using the capacitance values obtained from EIS. The

 o S 

(7)

AC

C

CE

capacitance of TGO with their thickness can be related by the following formula [18]

Where ε is dielectric constant; εo is vacuum dielectric constant; δ is the TGO thickness and S is the area of the electrode. The calculated values of TGO thickness is shown in Fig. 7 (b). So it is evident that the thickness of TGO layer increased with pre-oxidation time and followed a parabolic kinetics. An empirical relationship between TGO thickness and pre-oxidation time (t) can be obtained by polynomial fitting from Fig. 7 (b) and given as Eq.(8). δ = -0.0025t2 + 0.097t + 0.74

(8) 16

ACCEPTED MANUSCRIPT

It shows that a calculated value of TGO thickness from EIS is smaller than the TGO thickness which is obtained from SEM measurements. This is probably due to the presence of Ni and Cr based oxides, which are more conductive than pure Al2O3, leads to increasing the capacitance. From the above Eq. 8, it can be said that the time increased from t to t + 1 (h), the TGO

PT

growth changes by 0.097 - 0.0025 × (2t + 1) µm. An increasing trend of TGO thickness was found

SC

3.2.4. Breakpoint frequency of pre-oxidized SSA TBCs

RI

up to 20 h and followed by reduction at 30 h due to the increasing capacitance behavior [(Fig. 7(b)].

The breakpoint frequency method has been widely used to evaluate the performance of

NU

coatings [48]. The breakpoint frequency is defined as the frequency at which phase angle 45º and it

Ad A

(9)

D

f 45  K

MA

can be expressed as,

PT E

Where Ad and A are the delaminated area or reactive area and total exposed area of the sample, respectively. K is the material constant for the coating under study and is equal to 1/2ԑԑ0ρ0. Where ԑ

CE

is the dielectric constant of the coating, ԑ0 is the vacuum permittivity and ρ0 is the ionic resistance of the coating. Since the exposed area is same for all the tested samples, it can be said that the reactive

AC

area is directly proportional to the breakpoint frequency at 45º. In the present study the breakpoint frequency is calculated from frequency vs. phase angle plot of SSA TBCs after pre-oxidation at 1050 ºC [shown in Fig. 7 (c)]. After pre-oxidation treatment, the lowest breakpoint frequency was observed at 20 h than 10 and 30 h indicating more reactive arc for 30 h followed by 10 h [Fig. 3 (a)]. The 10 and 30 h pre-oxidized samples showed more delaminated and reactive area than 20 which is also supported by porosity and EDS results (Figs. 5 and 6).

17

ACCEPTED MANUSCRIPT

3.2.5. Residual stresses in the SSA top coat after pre-oxidation treatment The Raman spectra obtained from the SSA top coat after pre-oxidation treatment at 1050 ºC for different exposure time is shown in Fig. 8 (a). In general, Raman spectroscopy is very sensitive to the structure and metal-oxygen stretching modes of oxides. The spectrum indicates that TBC

PT

layer is a single phase BCT structure as confirmed by XRD pattern [Fig. 8 (b)]. No Raman study of SSA compound has been reported to date, but for the doped SrAl2O4 ceramic has been reported by

RI

several previous researchers [49, 50]. The spectrum exhibited peaks at 108, 195, 224 (low intense)

SC

320 and 465 cm-1 (strong intense) attributed to the Sr-Al-O bond in the SSA crystal structure. After

NU

pre-oxidation treatment, the frequency shift was observed from 320 to 314.8, 315.4 and 317.9 cm-1 for 10, 20, and 30 h of treatment at 1050 ºC. The spectral shift (∆Ω) between stress-free and pre-

MA

oxidized SSA TBC samples is shown in Table 4. It could be seen that the positive shift increased with increasing pre-oxidation duration. So, it is understood that the residual tensile stresses

D

decreased in the SSA top coat with increasing pre-oxidation time at 1050 ºC. Generally, TGO layer

PT E

posses a large residual compressive stresses due to thermal expansion mismatch with the ceramic top coat when its cools down to atmospheric temperature [10]. Tomimatsu et al. have studied the

CE

residual stress distribution of EB-PVD deposited YSZ TBCs by Raman spectroscopy. It was proved that stress in the TBC layer has a distribution (compressive and tensile) and maximum stresses were

AC

found at the bond coat and top coat interface; however, it gradually decreased towards top coat [30].

It is also obvious that an increasing TGO growth induces higher residual stresses in the top coat. But in the present study, the 30 h treated SSA TBC samples showed almost 59% lower tensile stress as compared to 10 and 20 h due to the presence of more porosities and cracks in the SSA top coat (Fig. 6). The presence of defects in the SSA top coat was helped to mitigate the residual stresses for 30 h of pre-oxidized specimens. A relationship between the polarization resistance and residual stress formation of SSA top coat as a function of different pre-oxidation duration is shown 18

ACCEPTED MANUSCRIPT

in Fig. 9. As discussed above, a decreasing trend of RSSA and residual stresses was observed with an increase in pre-oxidation times of 10, 20 and 30 h which mainly due to increasing porosities of the top coat (13, 29 and 45%). It was well proved that the ceramic top coat defects such as pores and micro-cracks play a vital role on thermal insulation and thermal shock resistance of the coatings at

PT

the higher temperature.

RI

3.2.6. Microstructure of SSA top coat and TGO after oxidation test in air

SC

Fig. 10 shows the SEM cross-sectional micrographs and respective EDS micro-region analysis of oxidized SSA TBCs after exposed to 1100 ºC in the air. Severe coating damages were

NU

observed in the 10 h pre-oxidized SSA TBCs [Fig. 10 (a)] than that of 20 and 30 h [Fig. 10 (b and c)]. The top coat microstructure of 30 h pre-oxidized specimens showed more interconnected

MA

vertical cracks as compared to 20 h due to sintering of SSA ceramic at 1100 ºC. The thickness of TGO varies as 9.5 ± 2, 6 ± 2, and 8 ± 3 µm, for 10, 20, and 30 h pre-oxidized specimens,

D

respectively. The lowest growth of TGO was observed at 20 h than 10 and 30 h because of stable

PT E

and high pure α-Al2O3 TGO obtained during pre-oxidation treatment at 1050 ºC. EDS analysis was conducted near to TGO-top coat interface of all pre-oxidized specimens, which revealed the

CE

presence of Al and O elements with an atomic ratio of (Al: O = 0.299; 10 h), (Al: O = 0.666; 20 h)

AC

and (Al: O = 0.486; 30 h) for Al2O3 TGO scale. The suitable atomic ratio for Al2O3 was found at 20 h, indicates higher chemical stability of TGO at 20 h. In addition to Al and O elements, the other elements such as Ni, Cr, and Sm are also found at the interface between TGO and top coat.

Fig. 11 (a and b) shows the Nyquist and Bode impedance plots of SSA TBCs after exposed to 1100 ºC in the air. With an exception of the charge transfer effect in the low-frequency zone, the overall resistance of TBCs characterized by the intercept of the real part impedance, which is higher for 20 h pre-oxidized specimens followed by 30 and 10 h. The Nyquist plot was fitted based on a 19

ACCEPTED MANUSCRIPT

model of 3 R/CPE in the series connection as shown in Fig. 11 (a). It is evident that measured impedance results are well matched with fitted results. The typical relaxation frequency for the TGO layer is about 100-101 Hz and this response can be clearly seen in the case of 20 h than 30 and 10 h due to the presence of continuous TGO layer at the interface between bond coat and top coat

PT

[Fig. 11 (b)]. The obtained electrical properties of pre-oxidized SSA TBCs after oxidation test are reported in Table 5. The 20 h pre-oxidized specimen exhibited highest top coat and TGO resistance

RI

as compared to 10 and 30 h. It is also indicated the resistance changes from charge transfer effect

SC

are in the same order of changes in resistance of SSA top coat and TGO. Table 5 also revealed highest charge transfer resistance for 20 h pre-oxidized samples which caused by the presence

NU

Al2O3 insulator at the interface. But in other cases such as 10 and 30 h, the presence of conductive

MA

NiO, Cr2O3, and NiCr2O4 which act as a p-type semiconductor, increases the charge transfer due to the presence of more vacancy sites. The highest capacitance of SSA top coat, TGO and charge

D

transfer effect were noticed for 10 h pre-oxidized specimens (Table 5), at which the failure of TBCs

PT E

occurred as shown in [Fig. 10 (a)]. According to Eq. 7 the capacitance of TGO is inversely proportional to thickness. Even though the specimens subjected to 10 and 30 h pre-oxidation exhibits much higher TGO thickness than 20 h, it showed higher capacitance due to the degradation

CE

of protective Al2O3 TGO into conductive mixed oxides at the interface.

AC

Fig. 12 (a-f) shows the schematic representation of electrochemical redox reactions of SSA TBCs after exposed to pre-oxidation and oxidation treatment at 1050 [Fig. 12 (a-c)] and 1100 ºC [Fig. 12 (d-f)], respectively. During the initial stages of the pre-oxidation process, the metallic aluminum in bond coat reacted with atmospheric oxygen to form the aluminum oxide at 1050 ºC. The formed oxide ligands grow in the lateral direction and form continuous TGO layer at the bond coat and top coat interface. Further, degradation of TGO layer in the microstructure and composition led diffusion of Ni and Cr from metallic bond coat to TGO-top coat interface. The 20

ACCEPTED MANUSCRIPT

higher percentage of Ni (7.05 at.%) and Cr (2.27%) were seen for the pre-oxidation treatment of 10 h followed by 30 h (1.30 at.% Ni and 0.53 at.% Cr) and 20 h (1.17 at.% Ni and 0.00 at.% Cr) at the TGO-top coat interface due to the absence of high purity TGO after pre-oxidation test at 1050 ºC [Fig. 12 (a-c)]. The presence of NiO and Cr2O3 play an important role in the formation of mixed

PT

oxides after oxidation test at 1100 ºC for 15 h in the air which directly affects diffusivity of metallic ions and oxygen.

RI

In presence study, 20 h pre-oxidized sample exhibited significantly the lowest percentage of

SC

Cr (2.58 at.%) at the top coat-TGO interface than 10 h (5.91 at.%) and 30 h (4.01 at.%) due to the presence of highly enriched α-Al2O3 TGO after oxidation test at 1100 ºC for 15 h in air. The

NU

variation of Ni was also seen in the TGO-top coat interface as 1.57, 6.34, and 9.92 at.% for the pre-

MA

oxidation treatment of 10, 20, and 30 h respectively after oxidation test. It seems low Ni availability for 10 h; however, the highest presence of Cr was found for the same pre-oxidized specimens after oxidation test at 1100 ºC. So, it is interesting to understand the role of Cr3+ ions diffusion into the

PT E

D

NiO lattice after exposure to oxidation test in the air. It is well known that the minimum 3Cr3+ ions require to form an extra metal vacancy sites in the NiO lattice [51] and it can be represented by the

CE

Kroger-Vink notation (Eq. 10 ) [Fig. 12 (d-f)]. 3Cr2O3  3NiO  VNi  6OOx  NiNi

AC



(10) x

Where, VNi is a nickel vacancy with positive charges, OO an oxygen ion sitting on an oxygen lattice, with a neutral charge and

 NiNi

is a nickel ion sitting on a nickel lattice site, with single positive

charge.

21

ACCEPTED MANUSCRIPT

The value of diffusion coefficient of Cr3+ ions into NiO lattice through the protective TGO layer at the interface can be calculated using Wagner’s oxidation theory [52]. The parabolic rate constant can be directly related to diffusion coefficient as follows [45]:

 m   DCr 3 M Al2O3  Al2O3 dcNi2  t    jAm  A    2

PT

(11)

RI

Where (∆m/A)2 / t is the parabolic rate constant (g2 cm-4 s-1), DCr 3 is the diffusion coefficient (m2

SC

s-1) of Cr metal ions, M and ρ are the molar mass (102 g mol-1) and density (3.97 g cm-3) of α-

NU

Al2O3, respectively, dc is the concentration of Ni2+ vacancies (g m-3), j is the no of atoms of O2-

MA

required to form 1 mole of Al2O3 and Am is the atomic mass of oxygen. In our previous study the parabolic rate constant (kp) was calculated to be 6.08×10-11,

D

3.90×10-11 and 4.55×10-11 mg2 cm-4 s-1 for 10, 20, and 30 h of pre-oxidation treatment after exposure

PT E

to oxidation test at 1100 ºC for 15 h in the air [27]. These values were substituted into Eq. 11 and the corresponding values of the DCr 3 ions in NiO lattice were calculated and reported in Table 6.

CE

The lowest DCr 3 was found at 20 h of pre-oxidation (2.28×10-15 m2 s-1) over 10 and 30 h due to the presence of highly enriched α-Al2O3 TGO layer at the interface or the presence of lowest

AC

amounts of Cr2O3 (5.6 mol.%). A slightly higher DCr 3 observed at 10 (4.0×10-15 m2s-1) followed by 30 h (3.1×10-15 m2s-1) due to the formation of higher amounts of 8.9 mol. % and 7.8 mol.% Cr2O3 respectively. Hence, the lowest DCr 3 in the NiO which reduced the formation of more metal ion vacancies after exposure in the air at 1100 ºC for 15 h caused to exhibit highest oxidation resistance (lowest kp value) for 20 h of pre-oxidized specimens as compared to 10 and 30 h [27]. However, in the case of 10 and 30 h pre-oxidized specimens exhibited significantly higher kp values 22

ACCEPTED MANUSCRIPT

due to the greater DCr 3 values which increased the formation of more metal vacancies. As a result, the 10 and 30 h treated SSA TGO allowed more oxygen diffusion towards the metallic bond coat which caused higher oxidation at 1100 ºC for 15 h.

PT

Fig. 12 (g) shows the Nyquist plot of the sample exposed to 20 h of pre-oxidation at 1050 ºC followed by oxidation test at 1100 ºC for 15 h in the air. The plots clearly revealed that pre-oxidized

RI

sample showed higher polarization resistance than those of oxidized one which is mainly due to the

SC

presence of insulator type Al2O3 in the TGO. After oxidation test, the formed conductive mixed oxides of Ni and Cr at the TGO-top coat interface make TGO more conductive than pure Al2O3.

NU

These observations clearly revealed that an increasing metal ions vacancy sites in the mixed oxides

MA

decreased the impedance response after exposure to 1100 ºC. At the TGO-top coat interface, the Sm2O3 reacted with Al2O3 to form SmAlO3 when exposed to the higher temperature at 1100 ºC (Eq.

NiO + Cr2O3 → NiCr2O4

(12)

PT E

Sm2O3 + Al2O3 → 2SmAlO3

D

12) [53, 54].

(13)

CE

Subsequently, the rate of diffusion of Ni and Cr from the metallic bond coat might be quite

AC

faster through already degraded TGO layer and formed NiO and Cr2O3 just beneath the top coat ceramic layer. These oxides reacted themselves to form NiCr2O4 at the interface of TGO-top coat (Eq. 13) [32]. A plot of polarization resistance of TGO and diffusion coefficient of Cr3+ ions in the NiO at the TGO-top coat interface of pre-oxidized SSA TBCs after exposed to oxidation at 1100 ºC is shown in Fig. 12 (h). An increase in polarization resistance of TGO is about 94% being found from 10 to 20 h of pre-oxidized SSA TBCs due to the lowest DCr 3 in the NiO at the TGO-top coat interface. This observation clearly indicated that the presence of lowest content of mixed oxides in 23

ACCEPTED MANUSCRIPT

the TGO (at 20 h) considerably increased the polarization resistance of TGO after oxidation at 1100 ºC. From these observations, it is understood that the total resistance of SSA TGO is more susceptible to compositional changes rather than thickness during oxidation at 1100 ºC in the air. Based on the impedance study of oxidized SSA TBCs, it can be concluded that 20 h pre-oxidation

PT

time would be sufficient enough to provide defect-free protective TGO layer at the interface for

RI

extended thermal exposure in air at the higher temperature.

SC

4. Conclusions

The SSA TBCs has been developed on NiCrAlY bond coated Inconel 718 superalloy substrate

NU

using an air plasma spray process and their microstructural and compositional changes have been studied after pre-oxidation and oxidation processes with the help of EIS. Based on the experimental

MA

observations the following conclusions are made:

D

1. After pre-oxidation there is an initial decrease in resistance of SSA top coat (RSSA) and increase

PT E

in capacitance were observed from 10, 20 and 30 h due to increase in the percentage of defects of about 13, 29 and 45, respectively at 1050 ºC.

CE

2. The increase in pre-oxidation duration caused a gradual increase in TGO growth. The higher

AC

resistivity of TGO (1.72×108 Ohm cm) was found at 20 h of pre-oxidation. 3. The growth of SSA TGO during pre-oxidation and oxidation treatment was inversely proportional to the capacitance in the presence of highly enriched α-Al2O3 at the interface. 4. The impedance response of SSA top coat and TGO vanishes after oxidation at 1100 ºC for 10 h of pre-oxidized specimens as compared to 20 and 30 h due to the compositional change of pure α-Al2O3 TGO to more conductive NiCr2O4.

24

ACCEPTED MANUSCRIPT

5. The resistance changes from charge transfer effect directly reflect the overall resistance changes of the TBCs components. 6. The polarization resistance of TGOs is more sensitive to compositional changes at the TGO-

PT

ceramic top coat interface during oxidation at 1100 ºC.

RI

Acknowledgements

Mr. T. Baskaran would like to thank Ministry of Human Resources Development (MHRD),

SC

Government of India, for a research fellowship. The authors are thankful to M/s. Metallizing

NU

Equipment Co. Pvt. Ltd, Jodhpur, India for coating the specimens. The authors also would like to thank Dr. B. Karthikeyan, NIT, Tiruchirappalli, India for extending the facility to record Raman

MA

spectra.

D

References

PT E

[1]. R.A. Miller, Current status of thermal barrier coatings-an overview, Surf. Coat. Technol. 30 (1987) 1–11. [2]. N.P. Padture, M. Gell, E.H. Jordan, Thermal barrier coatings for gas-turbine engine applications, Science, 296 (2002) 280–284.

CE

[3]. Y.N. Wu, M. Qin, Z. C. Feng, Y. Liang, C. Sun, F. H. Wang, Improved oxidation resistance of NiCrAlY coatings, Mat. Let. 57 (2003) 2404-2408.

AC

[4]. A. Keyvani, M. Saremi, M.H. Sohi, An investigation on oxidation, hot corrosion and mechanical properties of plasma-sprayed conventional and nanostructured YSZ coatings, Surf. Coat. Technol. 206 (2011) 208-216. [5]. J.L. Smialek, Moisture-induced TBC spallation on turbine blade samples. Surf. Coat. Technol. 206 (2011) 1577-1585. [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]. L.Y. Ni, C. Liu, H. Huang, C.G. Zhou, Thermal cycling behavior of thermal barrier coatings with HVOF NiCrAlY bond coat, J. Therm. Spray Technol. 20 (2011) 1133-1138. 25

ACCEPTED MANUSCRIPT

[8]. M. Matsumoto, K. Hayakawa, S. Kitaoka, H. Matsubara, H. Takayama, Y. Kagiya, and Y. Sugita, The effect of pre-oxidation atmosphere on oxidation behavior and thermal cycle life of thermal barrier coatings, Mat. Sci. Eng. A, 441 (2006) 119-125. [9]. N. Perez, Electrochemistry and corrosion science, Kluwer Academic publisher, India, 2004. [10]. X. Xie, H. Guo, S. Gong, H. Xu, Lanthanum-titanium-aluminum oxide: a novel thermal barrier coating material for applications at 1300 C, J. Eur. Ceram. Soc. 31 (2011) 1677-1683.

PT

[11]. B. Jayaraj, V.H. Desai, C.K. Lee, Y.H. Sohn, Electrochemical impedance spectroscopy of porous ZrO2-8 wt.% Y2O3 and thermally grown oxide on nickel aluminide, Mat. Sci. Eng. A, 372 (2004) 278-286.

SC

RI

[12]. S. Harada, Y. Ikeda, Y. Mizuta, Y. Sugita, A. Ito, The inspection of delaminated defects in ZrO2 coated materials by infrared thermography techniques, Proceedings of the 1st symposium on thermographic NDT and E techniques, 1995, 47-52.

NU

[13]. C.C. Berndt, Acoustic emission evaluation of plasma sprayed thermal barrier coatings, J. Eng. Gas Turbine and Power, 107 (1985) 142-146.

MA

[14]. K.S. Murphy, K.L. More, M.J. Lance, As sprayed mixed zone in thermally grown oxide beneath a thermal barrier coating, Surf. Coat. Technol. 146-147 (2001) 152-161.

D

[15]. C. Amaya, W. Aperador, J.C. Caicedo, F.J. Espinoza-Beltran, J. Munoz-Saldana, G. Zambrano, P. Prieto, Corrosion study of alumina/yttria stablilized zirconia (Al2O3/YSZ) nanostructured thermal barrier coatings (TBC) exposed to high temperature treatment, Corros. Sci. 51 (2009) 2994-2999.

PT E

[16]. J.W. Byeon, B. Jayaraj, S. Vishwewaraiah, S. Rhee, V.H. Desai, Y.H. Sohn, Nondestructive evaluation of degradation in multi-layered thermal barrier coatings by electrochemical impedance spectroscopy, Mat. Sci. Eng. A, 407 (2005) 213-225.

CE

[17]. S.H. Song, P. Xiao, L.Q Weng, Evaluation of microstructural evolution in thermal barrier coatings during thermal cycling using impedance spectroscopy. J. Eur. Ceram. Soc. 25 (2005) 1167-1173.

AC

[18]. B. Jayaraj, S. Vishweswaraiah, V.H. Desai, Y.H. Sohn, Electrochemical impedance spectroscopy of thermal barrier coatings as a function of isothermal and cyclic thermal exposure, Surf. Coat. Technol. 177-178 (2004) 140-151. [19]. J. Zhang, V. Desai, Evaluation of thickness, porosity and pore shape of plasma sprayed TBC by electrochemical impedance spectroscopy, Surf. Coat. Technol. 190 (2005) 98-109. [20]. W.L. Chen, M. Liu, J.F. Zhang, Impedance analysis of 7YSZ Thermal barrier coatings during high-temperature oxidation. J. Therm. Spray. Technol., 25 (2016) 1596-1603. [21]. J. Gomez-García, A. Rico, M.A. Garrido-Maneiro, C.J. Munez, P. Pozam, V. Utrilla, Correlation of mechanical properties and electrochemical impedance spectroscopy analysis of thermal barrier coatings. Surf. Coat. Technol. 204 (2009) 812-815. 26

ACCEPTED MANUSCRIPT

[22]. D. Zhu, R.A. Miller, Sintering and creep behavior of plasma sprayed zirconia and hafnia based thermal barrier coatings, Surf. Coat. Technol. 108-109, (1998) 114-120. [23]. R. Vaßen M.O. Jarligo, T. Steinke, D.E. Maik, D. Stöver, Overview an advanced thermal barrier coatings, 205 (2010) 938-942. [24]. C. Wan, T.D. Sparks, W. Pan, D.R. Clarke, Thermal Conductivity of the Rare-Earth Strontium Aluminates, J. Amer. Ceram. Soc. 93 (2010) 1457-1460.

PT

[25]. J. Feng, B. Xiao, R. Zhou, W. Pan, D.R. Clarke, Anisotropic elastic and thermal properties of the double perovskite slab-rock salt layer Ln2SrAl2O7 (Ln=La, Nd, Sm, Eu, Gd or Dy) natural superlattice structure, Acta Mater. 60 (2012) 3380-3392.

SC

RI

[26]. I. Zvereva, Yu. Smirnov, V. Gusarov, V. Popova, J. Choisnet, Complex aluminates RE2SrAl2O7 (RE=La, Nd, Sm-Ho): Cation ordering and stability of the double perovskite slabrocksalt layer P2/RS intergrowth, Solid state Sci. 5 (2003) 343-349.

NU

[27]. T. Baskaran, S.B. Arya, Role of Thermally grown oxide and oxidation resistance of samarium strontium aluminate based air plasma sprayed ceramic thermal barrier coatings, Surf. Coat. Technol. 326 (2017) 299-309.

MA

[28]. F.H. Yuan, Z.X. Chen, Z.W. Huang, Z.G. Wang, S.J. Zhu, Oxidation behavior of thermal barrier coatings with HVOF and detonation-sprayed NiCrAlY bond coats, Corros. Sci. 50 (2008) 1608-1617.

D

[29]. Y. Li, Y. Xie, L. Huang, X. Liu, X. Zheng, Effect of physical vapor deposited Al 2O3 film on TGO growth in YSZ/CoNiCrAlY coatings, Ceram. Int. 38 (2012) 5113-5121.

PT E

[30]. T. Tomimatsu, Y. Kagawa, S.J. Zhu, Residual stress distribution in electron beam-physical vapor deposited ZrO2 thermal barrier coating layer by Raman spectroscopy, Metall. Mater. Trans. A, 34 (2003) 1739-1741.

CE

[31]. F. Zhou, Y. Wang, Z. Cui, L. Wang, J. Gou, Q. Zhang, C. Wang, Thermal cycling behavior of nanostructured 8YSZ, SZ/8YSZ and 8CSZ/8YSZ thermal barrier coatings fabricated by atmospheric plasma spraying, Ceram. Int. 43 (2017) 4102-4111.

AC

[32]. H.B. Guo, R. Vaβen, D. Stӧver, Atmospheric plasma sprayed thick thermal barrier coatings with high segmentation crack density, Surf. Coat. Technol. 186 (2004) 353-363. [33]. A.C. Karaoglanli, K.M. Doleker, B. Demirel, A. Turk, R. Varol, Effect of shot peening on the oxidation behavior of thermal barrier coatings, Appl. Surf. Sci. 354 (2015) 314-322. [34]. J.R. MacDonald, Impedance spectroscopy-emphasizing solid materials and systems, Wiley, New York, 1987. [35]. J.G. Fletcher, A.R. West, J.T.S. Irvine, The AC-impedance response of the physical interface between yttria-stabilized zirconia and YBa2Cu3O7-x, J. Electrochem. Soc. 142 (1995) 2650-2654.

27

ACCEPTED MANUSCRIPT

[36]. S.T. Amaral, I.L. Muller, Effect of silicate on passive films anodically formed on iron in alkaline solution as studied by electrochemical impedance spectroscopy, Corrosion 55 (1999) 17-23. [37]. J. Feig, The grain boundary impedance of random microstructures: Numerical simulations and implications for the analysis of experimental data, Solid State Ionics 150 (2002) 181-193.

PT

[38]. X. Wang, J. Mei, P. Xiao, Determining oxide growth in thermal barrier coatings (TBCs) non-destructively using impedance spectroscopy, J. Mat. Sci. Lett. 20 (2001) 47-49. [39]. H. Huang, C. Liu, L. Ni, C. Zhou, Evaluation of micrsotructural evolution of thermal barrier coatings exposed to Na2SO4 using impedance spectroscopy, Corr. Sci. 53 (2011) 1369-1374.

SC

RI

[40]. H. Du, J. H. Shin, S. W. Lee, Study on porosity of plasma sprayed coatings by digital image analysis method, J. Therm. Spray. Technol. 14 (2005) 453-461. [41]. J. Zhang, V. Desai, Determining thermal conductivity of plasma sprayed TBC by electrochemical impedance spectroscopy, Surf. Coat. Technol. 190 (2005) 90-97.

NU

[42]. W.R. Chen, X. Wu, B.R. Marple, P.C. Patnaik, The growth and influence of thermally grown oxide in a thermal barrier coating, Surf. Coat. Technol. 201 (2006) 1074-1079.

MA

[43]. S. Nath, I. Manna, J.D. Majumdar, Kinetics and mechanisms of isothermal oxidation of compositionally graded yttria stabilized zirconia (YSZ) based thermal barrier coating, Corros. Sci. 88 (2014) 10-22.

PT E

D

[44]. S. Ahmadian, E.H. Jordan, Explanation of the effect of rapid cycling on oxidation, rumpling, micro-cracking and lifetime of air plasma sprayed thermal barrier coatings, Surf. Coat. Technol. 244 (2014) 109-116. [45]. A.C. Fox, T.W. Clyne, Oxygen transport by gas permeation through the zirconia layer in plasma sprayed thermal barrier coating, Surf. Coat. Technol. 184 (2004) 311-321.

CE

[46]. P. Marcus, F. Mansfeld, Analytical methods in corrosion science and engineering, CRC press, Florida, 2006.

AC

[47]. F. Zhang, J. Evertsson, F. Bertram, L. Rullik, F. Carla, M. Langberg, E. Lundgren, J. Pan, Integration of electrochemical and synchrotron-based X-ray techniques for in-situ investigation of aluminium anodization, Electrochim. Acta, 241 (2017) 299-308. [48]. Z. Feng, G.S. Frankel, Evaluation of coated Al alloys using the break point frequency method, Electrochim. Acta, 187 (2016) 605-615. [49]. M. Ayvacıklı, A. Ege, S. Yerci, N. Can, Synthesis and optical properties of Er3+ and Eu3+ doped SrAl2O4 phosphor ceramic, J. Lumin. 131 (2011) 2432-2439. [50]. P. Escribano, M. Marchal, M.L. Sanjuan, P.A. Gutierrez, B. Julian, E. Cordoncillo, Lowtemperature synthesis of SrAl2O4 by a modified sol–gel route: XRD and Raman characterization, J. Solid State Chem. 178 (2005) 1978-1987. 28

ACCEPTED MANUSCRIPT

[51]. C. Greskovich, Kinetics of NiCr2O4 formation and diffusion of Cr3+ ions in NiO, J. Amer. Ceram. Soc. 53 (1970) 198-502. [52].

J.J. Moore, Chemical Metallurgy, Butterworth and Co, 1981.

[53]. T.J. Nijdam, L.P.H. Jeurgens, J.H. Chen, W.G. Sloof, On the microstructure of the initial oxide grown by controlled annealing and oxidation on a NiCoCrAlY bond coating, Oxid. Met. 64 (2005) 355-377.

AC

CE

PT E

D

MA

NU

SC

RI

PT

[54]. H. Zhao, M.R. Begley, A. Heuer, R.S. Moshtaghin, H.N.G. Wadley, Reaction, transformation and delamination of samarium zirconate thermal barrier coatings, Surf. Coat. Technol. 205 (2011) 4355-4365.

29

ACCEPTED MANUSCRIPT

List of Tables Table 1: The precise plasma spray parameters of bond coat NiCrAlY and top coat SSA. NiCrAlY

SSA

Arc current (A)

500

525

Arc voltage (V)

67

68

PT

Parameter

1000

Spray distance (mm)

110

100

2.28 0.48

0.6

Working gas, Ar (m3/h)

2.28

0.30

0.28

45

30

6

6

90

90

1.8

1.8

NU

Working gas, H2 (m3/h)

Feed rate (g/min) Anode-nozzle internal diameter (mm)

AC

CE

PT E

Injector internal diameter (mm)

D

Injector position (deg.)

MA

Carrier gas, N2 (m3/h)

30

1000

SC

RI

Torch traverse speed (mm/s)

ACCEPTED MANUSCRIPT

Table 2: Resistance and capacitance of SSA top coat, TGO and TGO-bond coat interface as a function of pre-oxidation time after exposed to 1050 ºC in argon atmosphere. Pre-oxidation time (h) Electrical properties 20

30

RSSA, Ohm cm2

697

315

25

RTGO, Ohm cm2

30238

43074

Rct, Ohm cm2

322714

480711

CSSA, µF cm-2

0.274

0.285

CTGO, µF cm-2

2.84

2.18

Cct, µF cm-2

0.06

18877

SC

RI

6150

0.582

0.0011

0.0016

NU MA

D PT E CE AC

31

PT

10

2.42

ACCEPTED MANUSCRIPT

Table 3: Fitting data from EIS spectra of the pre-oxidized SSA TBCs after pre-oxidation at 1050 ºC.

Pre-oxidation

Region

time (h)

resistance, Rp (Ohm cm2)

22

3.0×104

TGO/BC interface

33

3.2×105

TGO

27

TGO/BC interface

39

TGO

14

TGO/BC interface

13

PT

TGO

AC

CE

PT E

D

MA

NU

30

Polarization

4.3×104

RI

20

(deg.)

SC

10

Phase angle, Ф

32

4.8×105 6.2×103 1.9×104

ACCEPTED MANUSCRIPT

Table 4: Spectral shift ∆Ω of the band at 320 cm-1 in the SSA top coat. Bi-axial residual stress

Pre-oxidation time (h)

5.57

20

4.96

30

2.39

AC

CE

PT E

D

MA

NU

SC

RI

10

PT

(GPa)

33

ACCEPTED MANUSCRIPT

Table 5: Resistance and capacitance of SSA top coat, TGO and TGO-bond coat interface after isothermal oxidation as a function of different pre-oxidation time after exposed to oxidation at 1100 ºC for 15 h in the air. Pre-oxidation time (h)

RSSA, Ohm cm2

519

962

RTGO, Ohm cm2

1162

20460

Rct, Ohm cm2

4544

16551

CSSA, µF cm-2

61.4

0.062

CTGO, µF cm-2

3.39

Cct, µF cm-2

26000

RI

20

AC

CE

PT E

D

MA

NU

SC

10

PT

Electrical properties

34

30 704 16129 14342 56.3

0.4

2.06

50

165

ACCEPTED MANUSCRIPT

Table 6: Diffusion coefficients of Cr3+ in NiO through TGO layer at the interface after exposure to 1100 ºC in the air.

Oxidation time (h)

10

1100

15

6.08

20

1100

15

30

1100

15

Diffusion coefficient (m2 s-1)×10-15

4.0

3.90

5.6

2.3

4.55

7.8

3.1

SC

8.9

NU MA D PT E CE AC

35

Cr2O3 in NiO (mol.%)

PT

Oxidation temperature (ºC)

RI

Preoxidation time (h)

Parabolic rate constant kp (mg2 cm4 -1 s ) ×10-5

ACCEPTED MANUSCRIPT

List of figure captions Fig. 1: SEM morphology of NiCrAlY and synthesized SSA powder.

Fig. 2: SEM images of SSA TBCs: (a) surface morphology; and (b) cross section of as deposited

PT

coatings.

Fig. 3: (a) Nyquist impedance; (b) Magnified high frequency Nyquist impedance plot; and (c) Bode

RI

phase angle plot of pre-oxidized SSA specimens after exposed to 1050 ºC for different exposure

SC

time. Note: The fitting of measured Nyquist impedance is also shown in 3 (a).

NU

Fig. 4: Schematic illustrations of SSA TBCs components with AC equivalent circuit model (a) 10;

MA

(b) 20; and (c) 30 h pre-oxidized at 1050 ºC in argon atmosphere.

Fig. 5: Cross sectional SEM microstructure of pre-oxidized SSA TBCs and respective EDS analysis

PT E

D

after exposure to (a) 10; (b) 20; and (c) 30 h at 1050 ºC in argon atmosphere.

Fig. 6: Image J software processed higher magnification cross sectional SEM images of SSA coatings after exposed to pre-oxidation at 1050 ºC for (a) 10; (b) 20; (c) 30 h and (d-f) porosity

CE

distributions of (a-c); respectively.

AC

Fig. 7: (a) Frequency dependence of –Im (Z) and Ф; (b) thickness of SSA TGOs as a function of pre-oxidation time calculated from EIS measurement; and (c) Breakpoint frequency vs. preoxidation time of SSA TBCs after exposure to 1050 ºC.

Fig. 8: (a) Raman spectra of the SSA top coat after pre-oxidation treatment for different exposure time; and (b) XRD pattern of 30 h pre-oxidized SSA top coat at 1050 ºC in argon atmosphere.

36

ACCEPTED MANUSCRIPT

Fig. 9: Polarization resistance and bi-axial residual stress of SSA top coat after exposed to preoxidation at 1050 ºC for different time. Fig. 10: Cross-sectional SEM micrographs and corresponding EDS micro-region analysis of preoxidized SSA TBCs for (a) 10 h; (b) 20 h; and (c) 30 h after exposed to oxidation test at 1100 ºC for

PT

15 h in the air.

RI

Fig. 11: (a) Nyquist; and (b) Bode impedance plots of pre-oxidized SSA specimens after exposed to

SC

oxidation test at 1100 ºC for 15 h in the air.

Fig. 12: Schematic illustration of formation of TGO after pre-oxidation at 1050 ºC for (a) 10; (b)

NU

20; (c) 30 h; (d, e and f) defect oxide structure of TGO after oxidation at 1100 ºC for 15 h in the air, (g) Nyquist impedance plot of 20 h pre-oxidized specimen after oxidation in the air; and (h)

MA

Polarization resistance of TGO and diffusion coefficient of Cr3+ ions in the NiO at the TGO-top

AC

CE

PT E

D

coat interface of pre-oxidized SSA TBCs after exposed to oxidation test at 1100 ºC in the air.

37

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

Fig. 1

SC

RI

PT

Figures

38

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Fig. 2

39

Fig. 3

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

40

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

RI

Fig. 4

41

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 5

42

AC

CE

PT E

D

MA

NU

Fig. 6

SC

RI

PT

ACCEPTED MANUSCRIPT

43

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 7 44

AC

CE

PT E

D

MA

NU

Fig. 8

SC

RI

PT

ACCEPTED MANUSCRIPT

45

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Fig. 9

46

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 10

47

AC

CE

PT E

D

MA

SC

NU

Fig. 11

RI

PT

ACCEPTED MANUSCRIPT

48

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 12

49

ACCEPTED MANUSCRIPT

Highlights

A variation in polarization resistance and capacitance of TGO were observed.



The higher resistivity of TGO is being found at 20 h of pre-oxidation at 1050 ºC.



The growth of TGO is inversely proportional to capacitance in the presence of α-Al2O3.



Increasing content of Cr3+ ions increases more vacancy sites at the interface.



Resistance of SSA TGOs inversely proportional to diffusion coefficient of Cr3+ in the NiO.

AC

CE

PT E

D

MA

NU

SC

RI

PT



50