Combustion Turbine Hot Section Coating Life Management

The purpose of this project was to enhance the capability of COATLIFE to handle spallation life prediction for thermal barrier coatings TBCs that are being used in advanced turbines manu

Combustion Turbine Hot Section Coating Life

Management COATLIFE for Advanced Metallic Coatings and TBCs

1011593

Combustion Turbine Hot Section

Coating Life Management

COATLIFE for Advanced Metallic Coatings and

TBCs

1011593

Technical Update, March 2005

Cosponsor

U.S Department of Energy

National Energy Technology Laboratory

626 Cochrans Mill Road

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC (EPRI) NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED

OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT

ORGANIZATION(S) THAT PREPARED THIS DOCUMENT

Southwest Research Institute

Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc

Copyright © 2005 Electric Power Research Institute, Inc All rights reserved.

CITATIONS

This report was prepared by:

Southwest Research Institute

The report is a corporate document that should be cited in the literature in the following manner:

Combustion Turbine Hot Section Coating Life Management: COATLIFE for Advanced Metallic Coating and TBCs, EPRI, Palo Alto, CA, and U.S Department of Energy, Pittsburg, PA: 2005

1011593

iii

REPORT SUMMARY

COATLIFE is a coating life model that predicts the oxidation life of overlay and diffusion

coatings The purpose of this project was to enhance the capability of COATLIFE to handle spallation life prediction for thermal barrier coatings (TBCs) that are being used in advanced turbines manufactured by major domestic OEMs, and to broaden COATLIFE capability to cover

a broader range of MCrAlY coatings for oxidation and thermomechanical fatigue (TMF) life prediction of turbine blades

Background

COATLIFE has been designed to predict the oxidation life of combustion turbine coatings and TMF life of coated blade alloys under variable plant operation conditions The algorithms in COATLIFE take into account all the degradation mechanisms involved during long-term service

of the coated blades and TMF life of the blades The specific physical degradation mechanisms considered in the model include oxide formation kinetics, spallation of the protective oxide (Al2O3 layer), and interdiffusion of aluminum from the coating into the superalloy blade

substrate This approach can account for the contribution of time in service, number of startup and shutdown cycles, and variable temperature operation (that is, part load operation)

Equipment manufacturers also commonly use two other coatings on the blades and vanes of current engines of advanced turbines: NiCoCrAlY and CoNiCrAlY These coatings are similar in composition; General Electric uses NiCoCrAlY (GT33), and Siemens-Westinghouse uses

CoNiCrAlY (trade name CT102) In order to be more effective, the COATLIFE code needed to

be enhanced to handle these widely used coatings

Approach

The MCrAlY coating selected for the evaluation was CT102 Siemens-Westinghouse uses this coating on combustion turbine (CT) turbine blades and vanes The chemical composition of CT102 coating is similar to the nominal chemistry of GE’s proprietary NiCoCrAlY coating GT33, which is used on the blades of GE’s F, G, and H class turbines

Cyclic oxidation tests were conducted on the NiCoCrAlY-coated specimens Coating life

diagrams for the coating were computed as a function of temperature Isothermal oxidation tests

at three temperatures were conducted on the TBC-coated GTD-111 and IN-738 specimens with

or without a platinum interlayer between the bond coating and TBC with or without a platinum interlayer between the bond coating and TBC Thermal cycling tests were also conducted at two peak temperatures on the TBC-coated GTD-111 and IN-738 specimens with two coatings to determine the constants for COATLIFE model Isothermal oxidation tests on the TBC-coated specimens were performed at three different temperatures: 1850°F (1010°C), 1900°F (1038°C), and 1950°F (1066°C) Cyclic oxidation testing of the TBC-coated specimens was performed at the two peak temperatures of 1950°F (1066°C) and 1850°F (1010°C)

Results

Thermally grown oxide (TGO) thickness between the bond coat and the TBC was determined to

be a function of time and temperature for four bond coat/substrate systems Oxidation life and TMF life equations for the NiCoCrAlY coating have been incorporated in COATLIFE 4.0 In addition, the graphical user interface (GUI) was revised to allow NDE input of aluminum content and/or volume percent of the beta phase for computation of coating life using the NDE data

A model for TBC life was developed, and the constants for the model were determined from the long-term testing of the TBC-coated GTD-111 and IN-738 materials The TBC life model was validated with the laboratory data that have not been used to determine the model constants The TBC life equations have been incorporated into an upgraded version of the COATLIFE software EPRI Perspective

Performance and durability of coating systems are prime life-limiting factors for hot section components Turbine blades are the most critical and expensive parts of these components

because the reliability and availability of a turbine often depends on blade life, which depends on the coating life A reliable method for more accurately predicting the life of all coatings

commonly in use today is crucial to assessing and extending blade service life

Keywords

Gas turbines

Thermal barrier coatings

Turbine life management

Life assessment

Turbine blades

Thermomechanical fatigue

vi

of aluminum content and/or volume percent of the beta phase for computation of coating life using the NDE data The software has been upgraded to COATLIFE 3.5 A model for TBC life has been developed, and the constants for the model are determined from the long-term testing of the TBC-coated GTD-111 and IN-738 materials The TBC life model was validated with the laboratory data that have not been used to determine the model constants The TBC life

equations have been incorporated into an upgraded COATLIFE 4.0 software, and a users manual

for COATLIFE 4.0 has been prepared

vii

EXECUTIVE SUMMARY

This report addresses the results of two tasks of a major project entitled, “Combustion Turbine (CT) Hot Section Coating Life Management.” The two tasks include COATLIFE for Advanced Metallic Coatings and TBCs (Task 2 of the project) and Field Validation of COATLIFE and NDE (Task 4 of the project)

The objective of Task 2 was to develop the capability of COATLIFE to handle spallation life prediction for thermal barrier coatings (TBCs) that are used in advanced turbines manufactured

by major domestic OEMs, and to enhance COATLIFE to cover a broader range of MCrAlY coatings for oxidation life prediction The MCrAlY coating selected for the evaluation was CT102 Siemens-Westinghouse uses this coating on CT turbine blades and vanes The chemical composition of CT102 coating is similar to the nominal chemistry of GE’s proprietary

NiCoCrAlY coating GT33, which is used on the blades of GE’s F, G, and H class turbines The COATLIFE code needed to be enhanced to handle the widely used NiCoCrAlY coating The objective of Task 4 was to validate the predictive capabilities of COATLIFE and provide the required metallurgical data to correlate eddy current NDE results obtained on service-run blades

Cyclic oxidation tests were conducted on the NiCoCrAlY-coated specimens Coating life

diagrams for the coating were computed as a function of temperature Oxidation and

thermomechanical fatigue (TMF) life equations and coating life diagrams were incorporated into COATLIFE 3.0 The graphical user interface (GUI) has been revised to allow NDE input of aluminum content and/or volume fraction of the remaining beta (β) phase in the coating for computation of coating life using the NDE data The software has been upgraded to COATLIFE 3.5

The COATLIFE model treats coating degradation mechanisms that are applicable to TBC

spallation and bond coat degradation resulting from loss of aluminum Isothermal oxidation tests

at three temperatures were conducted on the TBC-coated GTD-111 and IN-738 specimens with

or without a platinum interlayer between the bond coating and TBC with or without a platinum interlayer between the bond coating and TBC Thermal cycling tests were also conducted at two peak temperatures on the TBC-coated GTD-111 and IN-738 specimens with two coatings to determine the constants for the COATLIFE model

Isothermal oxidation tests on the TBC-coated specimens were performed at three different

temperatures, 1850°F (1010°C), 1900°F (1038°C), and 1950°F (1066°C) The tests were

performed for up to 18,000 hours The TBC on the IN-738 specimens exhibited longer life than

on the GTD 111-specimens The presence of a platinum interlayer increased the TBC life on both substrate alloys Following testing, metallurgical samples were prepared from the exposed specimens, and the samples were examined in optical and scanning electron microscopes for

ix

coating degradation Thermally grown oxide (TGO) thickness between the bond coat and the TBC was determined as a function of time and temperature for each system The TGO thickness

on all coating systems increased with increasing exposure temperature and/or time The results showed no significant variation in the TGO thickness among the four-coating/substrate systems These results were used to determine the oxidation kinetics for the TBC life model

Delamination cracks were observed at the TGO/TBC interface on all samples after isothermal exposure The extent of cracking and crack size in all coating systems increased with the

exposure temperature and time The TGO-induced interface cracking is implicitly assumed to be directly related to TBC spallation or external cracking

Cyclic oxidation testing of the TBC-coated specimens was performed at the two peak

temperatures of 1950°F (1066°C) and 1850°F (1010°C) The results showed no significant difference in time to cracking among the four coating/substrate systems investigated These results, in conjunction with the time to cracking under isothermal conditions, were used to

determine the critical TGO thickness for TBC cracking and/or spallation A 24-hour hold time testing at the peak temperature of 1950°F (1066°C) was also conducted to validate the TBC life model

On all coating systems, the presence of TBC accelerated the kinetics of TGO formation and bond coating degradation The β phase in the bond coating was consumed after a short time exposure

at three temperatures The β phase led to localized (pitting-like) attack and internal oxidation of the bond coating The extent of internal oxidation of the bond coating is directly related to

exposure time and temperature The bond coating with the platinum interlayer on IN-738

specimens exhibited the highest resistance for internal oxidation The NiCoCrAlY coating with

or without platinum on the GTD-111 substrate showed the least resistance The results also showed that the TBC failure mechanism depends on the exposure time and temperature The TBC failure mechanism changes from the TGO growth-controlled mechanism to bond coating oxidation after long-term exposure The TBC failure location also changes from the TGO/TBC interface to the bond coat/substrate interface

Considering the physical degradation mechanisms, a mechanistic model has been developed for predicting the remaining service life of TBCs A generic TBC life model was developed for both air plasma spray (APS) TBC and electron-beam physical vapor deposition (EB-PVD) TBC The predictive capability of the TBC life model was demonstrated for the APS TBC using both literature data and laboratory data generated in this program The predictive capability of the model for the EB-PVD TBC was illustrated via literature data The predictive capabilities of the model are demonstrated by comparing the model predictions against the results published in the literature for APS TBCs The model was validated using the 24-hour hold time data These results indicated that the model predictions were in good agreement with the experimental data The TBC life equations have been incorporated into COATLIFE 4.0, and a user’s manual for COATLIFE has been issued

For validation of COATLIFE and NDE, two service-run GE Frame 7FA blades (#7 and #57) and six mounts prepared from the 50% airfoil height of 7FA blade (C) were received for

metallurgical evaluation following NDE evaluation Blades 7 and 57 had seen 8286 hours of operation with 670 start-stop cycles and 2000 hours of operation with 219 start-stop cycles,

x

respectively Blade C had operated for 6156 hours with 272 starts All three blades had operated under similar conditions

The blades had GT33+, over-aluminized NiCoCrAlY coating For COATLIFE validation, three transverse sections at the 25%, 50%, and 75% blade height locations were removed from Blades

7 and 57 The mounts were prepared from these sections and examined in an optical microscope and scanning electron microscope (SEM) to determine the location and the extent of TMF

cracking and coating degradation at various locations on the blades The remaining aluminum content in the coating at various locations in the blades was also determined These results were used to correlate with the NDE measurements by EPRI NDE Center

The metallurgical examinations showed that Blade 57 was more severely cracked than Blade 7 Blade 7 had a uniform top aluminide (0.002-inch (50-µm) thick) layer and NiCoCrAlY coating

on the convex and concave sides of the airfoil Severe TMF cracking was also noted at the 50% airfoil height of Blade C The coating thickness on Blade 57 and Blade C was found to vary from location to location Both blades had about 0.004-inch (100-µm) thick aluminide top coating In addition, the NiCoCrAlY/substrate interfaces of these two blades were severely contaminated with grit particles In addition, the aluminum content in the NiCoCrAlY coating on these two blades C was significantly higher than that of the coating on Blade 7 Thicker aluminide coating and the higher aluminum content in the coating on Blade 57 and Blade C were presumably responsible for extensive TMF cracking

Though the coating on the leading edge of Blade 7 at the 75% airfoil height was severely

degraded, the coating away from the leading edge of Blade 7 was in good condition The coating

on Blade 57 and Blade C was also found to be in good condition Consistent with the

metallography results, the coating at all locations where eddy current NDE measurements were made contained a significant amount of remaining aluminum (over 10 wt %), indicating that the coating had not significantly degraded at these locations

These data have been used to verify COATLIFE-4 for oxidation and the TMF life of GT 33+ coated blades The COATLIFE-predicted TMF and oxidation lives of GT 33+ coated blades are

in good agreement with the metallography results

xi

CONTENTS

1 ADVANCED METALLIC COATINGS – DEGRADATION 1-11.1 Experimental Procedure 1-11.1.1 Material, Coatings, and Test Specimens 1-11.1.2 Cyclic Oxidation Tests 1-31.1.3 Metallography 1-41.2 Results And Discussions 1-41.2.1 Microstructure of As-Coated Specimens 1-41.2.2 Cyclic Oxidation Test Results 1-51.2.3 Oxide Scale Characterization 1-71.2.4 Coating Microstructure After Exposure 1-151.3 Conclusions 1-201.4 References 1-20

2 THERMAL BARRIER COATINGS (TBCS) – DEGRADATION 2-12.1 Experimental Procedure 2-22.1.1 Material and Coatings 2-22.1.2 Isothermal Tests 2-62.1.3 Thermal Cycling Tests 2-62.1.4 Burner-Rig Tests 2-62.1.5 Metallography 2-62.2 Results and Discussions 2-82.2.1 Microstructure of As-Coated Specimens 2-82.2.2 Isothermal Exposure Testing 2-132.2.3 Thermal Cycling Test Results 2-172.2.4 Burner-Rig Test Results 2-202.2.5 Microstructure of Exposed Specimens 2-202.2.5.1 Coating Degradation 2-202.2.5.2 TGO Thickness 2-33

xiii

2.2.5.3 TGO/Bond Coat Interface Cracking 2-372.2.5.4 Internal Oxidation of Bond Coating 2-462.3 Conclusions 2-592.4 References 2-60

3 COATING LIFE ALGORITHM DEVELOPMENT 3-13.1 NiCoCrAlY Life Prediction Algorithm Development 3-23.1.1 Evaluation of Model Constants 3-23.1.2 Coating Life Diagrams for NiCoCrAlY 3-103.1.3 Thermomechanical Fatigue Life (TMF) Relation for NiCoCrAlY 3-143.1.4 Coating Life Prediction Based On NDE Input 3-153.1.5 COATLIFE – Version 3.5 3-193.2 TBC Life Prediction Algorithm Development 3-243.2.1 Failure Mechanisms 3-243.2.2 TBC Life Model 3-26Life-Prediction Equation 3-27Mechanical Strain Range 3-27Oxidation Kinetics 3-28Parametric Calculations 3-28Model Validation Using Literature Data of APS TBC and EB-PVD TBC 3-303.2.3 TBC Life Algorithm Development for Current Coating Systems 3-323.2.4 TBC Model Validation 3-413.3 Conclusions 3-433.4 References 3-44

4 FIELD VALIDATION OF COATLIFE AND NDE 4-14.1 Experimental Procedure 4-14.1.1 Material, Coatings, and Test Specimens 4-14.1.2 Metallography 4-14.2 Results and Discussions 4-34.2.1 Blade Materials and Coatings 4-34.2.2 TMF Cracking 4-44.2.3 Coating Quality 4-84.2.4 Coating Degradation 4-104.2.5 Chemical Analysis of the Coating 4-14xiv

4.3 COATLIFE and NDE Validation 4-264.4 Conclusions 4-294.5 References 4-29

xv

LIST OF FIGURES

Figure 1-1 Dimensions of cyclic oxidation specimens .1-2Figure 1-2 Microstructure of as-deposited NiCoCrAlY coating .1-5Figure 1-3 Cyclic oxidation behavior of GT33-like coating at 1950°F (1066°C) 1-6Figure 1-4 Cyclic oxidation behavior of GT33-like coating at 1850°F (1010°C) 1-7Figure 1-5 SEM backscattered electron images of the oxide scale on the GT33-like

coating after exposure for (a) 600 cycles, (b) 1500 cycles, and (c) & (d) 1750 cycles

between room temperature and 1950°F (1066°C) .1-8Figure 1-6 SEM backscattered electron images of the oxide scale on the GT33-like

coating after exposure for (a) 800 cycles, (b) 1500 cycles, (c) after 2000 cycles, d)

3486 one-hour cycles, and (e) & (f) 4500 one-hour cycles between room

temperature and 1850°F (1010°C) 1-9Figure 1-7 EDS results obtained from a continuous dark-appearing oxide scale on the

coating after (a) 600 one-hour thermal cycles between room temperature and

1950°F (1066°C) and (b) 800 one-hour thermal cycles between room temperature

and 1850°F (1010°C) (c) EDS spectrum obtained from area 3 marked in Figure

1-6(a) containing precipitates light in color Note that the

dark-appearing scale is aluminum-rich oxide presumably Al2O3 The light precipitate

particles in the Al2O3 are yttrium-rich particles 1-11

Figure 1-8 EDS results obtained from a continuous light-appearing oxide scale on the

coating after (a) 1500 one-hour thermal cycles & (b) 1750 one-hour thermal cycles

between room temperature and 1950°F (1066°C), and (c) area 3 in Figure 1-6(b) &

(d) area 4 in Figure 1-6(b) 1500 one-hour thermal cycles between room

temperature and 1850°F (1010°C) Note that the scale is a mixture of Ni, Co, Cr,

and Al oxides 1-13Figure 1-9 Microstructure of GT33-like coating on GTD-111 specimens showing variation

of the β phase in the coating as a function of thermal cycles between room

temperature and 1950°F (1066°C): (a) after 280 cycles, (b) after 1000 cycles, (c)

after 1500 cycles, and (d) after 2000 cycles .1-17Figure 1-10 Optical micrographs of transverse section removed from the exposed

specimens after (a) 2700 cycles, (b) 3486 cycles, and (c) 4500 cycles between

room temperature and 1850°F (1010°C) Note thermal fatigue cracks in the coating

on the specimens after 3486 and 4500 thermal cycles exposure The crack in the

coating penetrated through the interdiffusion zone in the specimen after 4500

thermal cycles .1-18

xvii

Figure 2-1 Photograph of an as-coated specimen (a) before bolt removal and (b) after

application of alumina on the ends The dotted line on photograph (a) points to the

location where the specimen was sectioned to remove the bolt head .2-4Figure 2-2 Dimensions of a hollow burner rig test specimen The TBC coating was

applied on the surface between the arrowheads .2-5Figure 2-3 Photographs of (a) thermal cycling furnace with a computer-controlled moving

arm and cooling system, (b) specimens immediately after their removal from the

furnace, and (c) forced-air-cooled specimens 2-7Figure 2-4 Photograph of the NRC high-velocity burner rig system 2-8Figure 2-5 Optical micrographs of as-coated TBC on an NiCoCrAlY-coated GTD-111

specimen The arrows point to delamination cracks at the TBC/bond coat interface 2-9Figure 2-6 Optical micrographs of as-coated TBC on an NiCoCrAlY-coated IN-738

specimen Arrows point to delamination cracks at the TBC/bond coat interface .2-10Figure 2-7 Optical micrographs of as-coated TBC on a platinum-plated NiCoCrAlY

coating on a GTD-111 specimen Arrows point to delamination cracks 2-11Figure 2-8 Optical micrographs of as-coated TBC on a platinum-plated NiCoCrAlY

coating on an IN-738 specimen Arrows point to delamination cracks at the

TBC/bond coat interface .2-12Figure 2-9 Condition of the coated specimens after exposure at 1850°F (1010°C), (a)

about 12,030 hours showing no cracking in the TBC, and (b), (c), (d) & (e) about

18,000 hours of exposure Arrows point to cracks that were not initiated from the

end faces of the specimen .2-14Figure 2-10 Condition of the coated specimens after exposure at 1900°F (1038°C)

Arrows point to cracks that were not initiated from the end faces of the specimen 2-15Figure 2-11 Photographs of (a) condition of TBC on all four specimens, and (b) cracks in

the TBC on MCrAlY-coated GTD-111 specimen after about 2925 hours, (c) cracks in

the TBC on MCrAlY-coated IN-738 specimen after about 3517 hours, (d) cracks in

the TBC on Pt-plated MCrAlY-coated GTD-111 specimen after about 3767 hours,

and (e) cracks in the TBC on Pt-plated MCrAlY-coated IN-738 specimen after about

4079 hours of exposure at 1950°F (1060°C) 2-16Figure 2-12 Photographs of TBC (a) on MCrAlY-coated GTD-111 specimen after about

2480 thermal cycles, (b) on MCrAlY-coated IN-738 specimen after about 2320

thermal cycles, (c) on Pt-plated MCrAlY-coated GTD-111 specimen after about

2490 thermal cycles, and (d) on Pt-plated MCrAlY-coated GTD-111 specimen after

about 2789 thermal cycles between the peak temperature of 1950°F (1060°C) and

room temperature Arrows point to cracks that did not initiate from the end faces .2-19Figure 2-13 Condition of the burner-rig test specimens after 373 one-hour thermal cycles

between the peak temperature of 1950ºF (1066ºC) and room temperature .2-21Figure 2-14 Optical micrograph of the TBC/NiCoCrAlY + Pt interface of a coated GTD-

111 specimen after 600 hours of exposure at 1950°F (1066°C) along with EDS

results obtained from the bond coating Note the Pt peak in the EDS spectrum .2-22Figure 2-15 Optical micrograph of the TBC/NiCoCrAlY + Pt interface of a coated IN-738

specimen after 600 hours of exposure at 1950°F (1066°C) along with EDS results

obtained from the bond coating Note the Pt peak in the EDS spectrum .2-23

xviii

Figure 2-16 Optical microstructure of NiCoCrAlY bond coating on the GTD-111

specimens Note that the β phase in the bond coating is completely consumed after

2015 hours of isothermal exposure at 1850°F (1010°C) 2-24Figure 2-17 Optical microstructure of the bond coating (a) on NiCoCrAlY/GTD-111-

coated specimen after 2480 cycles, (b) on NiCoCrAlY/IN-738-coated specimen

after 2320 cycles, (c) on NiCoCrAlY + Pt /GTD-111-coated specimen after 2490

cycles, and (d) on NiCoCrAlY + Pt /IN-738-coated specimen after 2789 cycles

between the peak temperature of 1850°F (1010°C) and room temperature Note

that the β phase in the bond coating is completely consumed in all four coating

systems .2-25Figure 2-18 EDS spectra obtained from the bond coating (a) on NiCoCrAlY/GTD-111-

coated specimen after 2480 cycles, (b) on NiCoCrAlY/IN-738-coated specimen

after 2320 cycles, (c) on NiCoCrAlY + Pt /GTD-111-coated specimen after 2490

cycles, and (d) on NiCoCrAlY + Pt /IN-738-coated specimen after 2789 cycles

between the peak temperature of 1850°F (1010°C) and room temperature .2-27Figure 2-19 Optical micrographs of oxide scale on TBC-coated GTD-111 samples after

thermal exposure: (a) TGO on NiCoCrAlY coating, (b) TGO on NiCoCrAlY coating,

NiCoCrAlY+Pt Note the growth of TGO into the TBC 2-30

Figure 2-20 Backscattered electron micrographs of mixed oxides in the TGO on the (a)

NiCoCrAlY-TBC-coated IN-738 and (b) NiCoCrAlY +Pt –TBC-coated IN-738

specimens Arrows on the micrographs point to the mixed oxides and (c) & (d) EDS

spectra obtained from the mixed oxides shown in (a) & (b), respectively .2-31Figure 2-21 Optical micrographs showing variation of TGO thickness on the NiCoCrAlY-

coated GTD-11 specimens after (a) 1510 hours of exposure and (b) 2785 hours of

exposure at 1950°F (1066°C) .2-35Figure 2-22 Optical micrographs showing variation of TGO thickness on the NiCoCrAlY-

coated GTD-111 specimen after (a) 2015 hours of exposure, (b) 5015 hours of

exposure, and (c) 9850 hours of exposure at 1850°F (1010°C) .2-36Figure 2-23 Optical micrographs of the TBC/NiCoCrAlY interface of GTD-111 specimens

after (a) 2015 hours of exposure, (b) 5015 hours of exposure, (c) 9850 hours of

exposure, and (d) 12030 hours of exposure at 1850°F (1010°C) showing variation

of the interface cracking .2-38Figure 2-24 Optical micrographs of the TBC/NiCoCrAlY interface of IN-738 specimens

after (a) 2015 hours of exposure, (b) 5015 hours of exposure, (c) 9850 hours of

exposure, and (d) 12030 hours of exposure at 1850°F (1010°C) showing variation

of the interface cracking .2-40Figure 2-25 Optical micrographs of the TBC/NiCoCrAlY+Pt interface of IN-738

specimens after (a) 2015 hours of exposure, (b) 5015 hours of exposure, (c) 9850

hours of exposure, and (d) 12030 hours of exposure at 1850°F (1010°C) showing

variation of the interface cracking .2-42Figure 2-26 Optical micrographs of the TBC/NiCoCrAlY+Pt interface of IN-738

specimens after (a) 1005 hours of exposure and (b) 2015 hours of exposure at

1950°F (1066°C) showing variation of the interface cracking .2-44

xix

Figure 2-27 Variation of delamination crack length at the TGO/TBC interface on the

NiCoCrAlY-coated GTD-111 specimens as a function of exposure time at 1850°F

(1010°C) 2-45Figure 2-28 Variation of the average delamination crack length at the TGO/TBC interface

among the three coating systems as a function of exposure time at 1850°F

(1010°C) 2-45Figure 2-29 Optical microstructure of the NiCoCrAlY bond coating on the GTD-111

specimens after (a) 2015 hours, (b) 5015 hours, (c) 9850 hours, and (d) 12,000

hours of exposure at 1850°F (1010°C) showing the onset of internal oxidation .2-47Figure 2-30 Optical microstructure of the NiCoCrAlY bond coating on (a) a GTD-111

specimen after 7470 hours, (b) an IN-738 specimen after 7290 hours of exposure

and microstructure of the NiCoCrAlY+ Pt bond coat on (c) a GTD-111 specimen

hours, and (d) an IN-738 specimen after 11,580 hours of exposure at 1900°F

(1038°C) Note the internal oxidation of the bond coating and delamination cracking

at the bond coat/substrate interface Also note the onset of internal oxidation of the

bond coating with the Pt interlayer on the IN-738 specimen (d) 2-49

Figure 2-31 Optical microstructure of the coating after 18,000 hours (a) NiCoCrAlY on

GTD-111, (b) NiCoCrAlY on IN-738, (c) NiCoCrAlY+Pt on GTD-111, and (d)

NiCoCrAlY +Pt on IN-738 Note the delamination cracking at the bond

coat/substrate interface (a) after 18,000 hours of exposure at 1850°F (1010°C) .2-51Figure 2-32 Optical microstructure of the NiCoCrAlY bond coating on GTD-111 after (a)

2925 hours of exposure at 1950°F (1066°C), (b) 6470 hours of exposure at 1900°F

(1038°C), and (c) 18,000 hours of exposure at 1850°F (1010°C) Note the onset of

bond delamination and void coalescence at the bond coat/substrate interface .2-53Figure 2-33 Optical micrographs of the NiCoCrAlY bond coating/GTD-111 interface after

(a) 600 hours of exposure, (b) 1510 hours of exposure, (c) 2785 hours of exposure,

hours of exposure at 1950° (1066°C) Note the variation of the interface voids and

the void coalescence with the exposure time 2-54

Figure 2-34 Optical micrographs of the NiCoCrAlY bond coating/IN-738 interface after

(a) 600 hours of exposure, (b) 1510 hours of exposure, and (c) 2785 hours of

exposure at 1950°F (1066°C) Note the variation of the interface voids and the void

coalescence with the exposure time .2-56Figure 2-35 Optical micrographs of the NiCoCrAlY+ Pt bond coating/IN-738 interface after

(a) 600 hours of exposure, (b) 1510 hours of exposure, and (c) 2785 hours of

exposure at 1950°F (1066°C) Note the variation of the interface voids and the void

coalescence with the exposure time .2-57Figure 2-36 Optical micrographs of the NiCoCrAlY bond coating/GTD-111interface after

(a) 2015 hours of exposure and (b) 8155 of hours of exposure at 1850°F (1010°C)

Note the variation of the interface voids with the exposure time .2-58Figure 3-1 Schematics of degradation mechanisms treated in the COATLIFE model

From Chan et al [1] .3-1Figure 3-2 A comparison of computed and measured Al content in GT33-coated

GTD-111 specimens as a function of thermal cycles The coated specimens were

xx

subjected to 1-hour thermal cycles between 77°F (25°C) and 1950°F (1066°C) The

model was fitted to the experimental data of Al content 3-3Figure 3-3 A comparison of computed and measured volume percent β in GT33-coated

GTD-111 specimens as a function of thermal cycles The coated specimens were

subjected to 1-hour thermal cycles between 77°F (25°C) and 1950°F (1066°C) 3-4Figure 3-4 A comparison of computed and measured weight change data of

GT33-coated GTD-111 as a function of thermal cycles The coated specimen was

subjected to 1-hour thermal cycles between 77°F (25°C) and 1950°F (1066°C)

Arrows indicate specimens whose oxide-scale compositions were determined using

the EDS technique .3-4Figure 3-5 Microstructure and composition of the oxide scale formed on GT33-like

coating after 600 one-hour cycles between 77°F and 1950°F: (a) alumina scale with

dispersed Y2O3, and (b) EDS result shows a high Al content in the outer oxide scale

(Location A) 3-5Figure 3-6 SEM micrograph shows that three oxides (A, C, and Y) are present in the

scale formed on the GT33-like coating after 1500 one-hour thermal cycles between

77°F (25°C) and 1950°F (1066°C) .3-6Figure 3-7 Composites of Oxides A, C, and Y: (a) Al-rich Oxide A, (b) Cr-rich Oxide C,

and (c) yttrium-rich Oxide Y .3-7Figure 3-8 Comparison of computed and measured weight changes data of NiCoCrAlY

(GT33-like coating) at 1850°F (1010°C) Deviations of computed and measured

weight changes for thermal cycles greater than 2000 cycles were caused by the

formation of alumina, yttria, and chromia in the experimental data, rather than a

continuous layer of alumina as assumed in the model 3-8Figure 3-9 Al depletion compared against experimental data for NiCoCrAlY (GT33-like

coating) at 1850°F (1010°C) .3-9Figure 3-10 Computed volume percent of β phase compared against experimental data

for NiCoCrAlY (GT33-like coating) at 1850°F (1010°C) .3-9Figure 3-11 Computed coating life diagram for GT33-like coating at 1950°F (1066°C) .3-11Figure 3-12 Computed coating life diagram for GT33-like coating at 1850°F (1010°C) .3-11Figure 3-13 Computed coating life diagram for GT33-like coating at 1750°F (954°C) .3-12Figure 3-14 Computed coating life diagram for GT33-like coating at 1650°F (899°C) .3-12Figure 3-15 Values of a and b as a function of temperature for GT33 coating: (a) value of

a, and (b) value of b in the coating life expression Ns = 10a τ .3-13Figure 3-16 TMF strain-life relations developed based on experimental data of

NiCoCrAlY-coated GTD-111DS .3-14Figure 3-17 A comparison of the TMF strain-life relation for NiCoCrAlY/GTD-111DS

generated in this program against those of NiCoCrAlY+/GTD-111DS

and NiCoCrAlY/IN-738 from previous EPRI programs [9,10] 3-15Figure 3-18 Comparison of experimental Al content against model calculations for

GT29+: (1) a linear fit to the experimental data, (2) COATLIFE, and (3) the linear

damage (Al depletion) model used in COATLIFE for variable temperature

conditions .3-16

xxi

Figure 3-19 Comparison of experimental data of volume percent β phase against model

calculations for GT29+: (1) a linear fit to the experimental data, (2) COATLIFE, and

(3) the linear damage (β depletion) model used in COATLIFE for variable

temperature conditions 3-17Figure 3-20 Comparison of experimental Al content against model calculations for

GT33+: (1) a linear fit to the experimental data (2) COATLIFE, and (3) the linear

damage (Al depletion) model used in COATLIFE for variable temperature

conditions .3-17Figure 3-21 Comparison of experimental data of volume percent β phase against model

calculations for GT33+: (1) a linear fit to the experimental data, (2) COATLIFE, and

(3) the linear damage (β depletion) model used in COATLIFE for variable

temperature conditions 3-18Figure 3-22 Comparison of experimental Al content against model calculations for GT33:

(1) a linear fit to the experimental data, (2) COATLIFE, and (3) the linear damage

(Al depletion) model used in COATLIFE for variable temperature conditions 3-18Figure 3-23 Comparison of experimental data of volume percent β phase against model

calculations for GT33: (1) a linear fit to the experimental data, (2) COATLIFE, and

(3) the linear damage (β depletion) model used in COATLIFE for variable

temperature conditions 3-19Figure 3-24 Revised graphical user interface (GUI) of COATLIFE that allows data input

in terms of number of startup cycles, as well as NDE input of Al content and volume

percent of β phase for remaining life calculation .3-20Figure 3-25 Data input and predicted coating life for GT33/GTD-111DS subjected to

thermal cycling at a peak temperature of 1800°F (982°C) and a cycle time of 24

hours for 200 startup cycles The TMF strain ranges are 200 cycles at 0.55% The

predicted oxidation life, percent life consumed, and the remaining life are 728.08

cycles (or 17,474 hours), 27.47%, and 528.08 cycles, respectively, while the

predicted remaining TMF life is 529.29 cycles .3-21Figure 3-26 Coating life diagram predicted for GT33/GTD-111DS at 1800°F (982°C) 3-21Figure 3-27 Data input and predicted coating life for GT33/IN-738 subjected to thermal

cycling at a peak temperature of 1750°F (954°C) and a cycle time of 1000 hours for

40 startup cycles The TMF strain ranges are 30 cycles at 0.35%, 5 cycles at

0.45%, and 5 cycles at 0.55% The predicted coating life, percent life consumed,

and the remaining life are 26.423 cycles (or 26,423 hours), 151.38%,

and -13.577 cycles, respectively .3-22Figure 3-28 Predicted coating life diagram for GT33/IN-738 at 1750°F (954°C) .3-23Figure 3-29 Data input of temperature (1800°F (982°C)), cycle time (200 hours), and

NDE input of 20% β phase, together with the predicted coating life (97.614 cycles),

percent life consumed (50%), and the remaining life (48.807 cycles) for

GT33/GTD-111DS .3-23Figure 3-30 Data input of temperature (1800°F (982°C)), cycle time (200 hours), and

NDE input of Al content in atomic percent (11 at.%), together with the predicted

coating life (97.614 cycles), percent life consumed (103.85%), and the remaining

life (-3.7544 cycles) for GT33/GTD-111DS A negative value for the remaining life

means that the coating has failed by oxidation and Al depletion 3-24Figure 3-31 Degradation mechanisms in APS TBCs 3-25xxii

Figure 3-32 Cracking and spallation of TBC: (a) isothermal oxidation (502 hours at

1975°F (1079°C)), and (b) cyclic oxidation (395 one-hour cycles at a peak

temperature of 1950°F (1066°C)) The arrow indicates a TBC crack formed by

thermal cycling .3-25Figure 3-33 Summary of the TBCLIFE code 3-26Figure 3-34 Effect of peak temperature on the calculated TBC life .3-29Figure 3-35 Effect of the TBC fatigue strength coefficient on the calculated TBC life .3-29Figure 3-36 Effect of sintering and time-dependent fatigue strength on the TBC life .3-30Figure 3-37 Effect of substrate curvature on the calculated TBC life 3-30Figure 3-38 Calculated TBC life compared with burner-rig and furnace data from the

HOST program [14] 3-31Figure 3-39 Computed TBC life diagram for EB-PVD TBCs compared with experimental

data from Tolpygo et al [23], Leyens et al [24], and Jordan and Gell [25] 3-32Figure 3-40 Plot of TGO thickness, δ, as a function of t1/2, where t is the time of oxidation

in hours The solid line is a least squares fit to the experimental data .3-33Figure 3-41 A comparison of measured and calculated TGO thickness at various times

of oxidation 3-33Figure 3-42 A comparison of measured and computed TGO thickness at various times of

oxidation 3-34Figure 3-43 An Arrhenius plot of the parabolic constant, kp, versus the reciprocal

temperature 3-34Figure 3-44 Comparison of computed and measured values of oxide thickness as a

function of time of oxidation and the experimentally observed failure mechanisms

for three TBC systems: (1) APS TBC/NiCoCrAlY/GTD-111DS, (b) APS

TBC/NiCoCrAlY/IN-738, and (c) APS TBC/NiCoCrAlY + Pt/IN-738 3-35Figure 3-45 Determination of the critical oxide thickness at TBC cracking for APS TBC/

NiCoCrAlY/GTD-111DS: (a) TBC cracking was observed after 5000 hrs at 1900°F

(1038°C), and (b) the onset of bond coat internal oxidation was observed after

abo

Figure 3-46 Computed TBC life diagram compared with experimental data for APS

TBC/NiCoAlY/GTD-111DS at 1950°F (1066°C) 3-37Figure 3-47 Computed TBC life diagram compared with the experimental data of four

TBC systems at 1950°F (1066°C) 3-37Figure 3-48 Comparison of computed TBC life boundaries with experimental data: (a)

1900°F (1038°C), and (b) 1850°F (1010°C) 3-38Figure 3-49 Parameters of the TBC life boundary (Ns = 10aτCb) as a function of

temperature: (a) parameter a, and (b) parameter b .3-39Figure 3-50 Flash screen of COATLIFE-4.0, which incorporates a life prediction

capability for APS TBCs 3-40Figure 3-51 The graphical user interface (GUI) of COATLIFE-4.0 with TBC life prediction

capability .3-40

xxiii

Figure 3-52 Predicted TBC life diagram showing the current status of an APS TBC after

100 startup cycles at 1700°F (927°C) and 100 hours/cycle .3-41Figure 3-53 Coating life diagram for APS TBC for four different TBC/bond coat/substrate

systems compared with TBCLIFE and COATLIFE .3-42Figure 3-54 Verification of COATLIFE prediction against burner-rig tests: (a) COATLIFE

prediction of a TBC life after 914 cycles, and (b) predicted coating life diagram for

the APS TBC showing the TBC being protective after 373 one-hour thermal cycles 3-42Figure 4-1 Photograph showing grids on the airfoil sections of Blades 7 and 57 where

NDE measurements were taken Note the smaller squares on Blade 57 .4-2Figure 4-2 Photographs showing metallurgical sample locations in Blades 7 and 57 .4-2Figure 4-3 Photograph of six mounts received for analysis .4-3Figure 4-4 Optical micrographs of TMF cracks at different blade heights on Blade 7 .4-5Figure 4-5 Optical micrographs of TMF cracks at different blade heights on Blade 57 .4-6Figure 4-6 Optical micrographs of TMF cracks at 50% airfoil height of Blade C 4-7Figure 4-7 Typical microstructure of the coating at different locations on Blade 7 4-8Figure 4-8 Optical micrographs of coating on Blade 57 showing the microstructure and

coating thickness and the grit particles at the coating/substrate interface .4-9Figure 4-9 Optical micrographs of the coating on Blade C showing the microstructure

and coating thickness and the grit particles at the coating/substrate interface .4-10Figure 4-10 Optical micrographs showing the variation of coating degradation at the 75%

airfoil height of Blade 7 Note that the coating is completely degraded on the suction

side near the leading edge (b) and that the coating showed no evidence of

degradation on the suction side of the airfoil 1.75" from the trailing edge (d) Also

note SS and PS on the micrographs denote suction and pressure sides,

respectively .4-11Figure 4-11 Optical micrographs showing the variation of coating degradation at the 50%

airfoil height of Blade 7 Note the degradation of coating completely at the LE (a)

and note that the coating showed no evidence of degradation on the suction side of

the airfoil, 1.75 inches away from the trailing edge .4-12Figure 4-12 Microstructure of the coating on pressure side (PS) of the airfoil at the 75%

and 50% blade heights 4-13Figure 4-13 Microstructure of the coating on the suction side at the 50% airfoil height of

Blade C .4-14Figure 4-14 Comparison of the COATLIFE prediction of oxidation life for GT33+ against

field data for the leading edge (75% blade height) of Blade 7 in a 7FA machine after

8286 hours and 670 startups: (a) oxidation failure predicted by COATLIFE-4, and

(b) metallographic section showing β-depleted coating at the leading edge tip of

Blade 7 .4-31Figure 4-15 Comparison of the COATLIFE prediction of oxidation life for GT33+ against

field data for the trailing edge (50% blade height) of Blade 7 in a 7FA machine after

8286 hours and 670 startup cycles: (a) COATLIFE prediction of 42.73% life

consumed and a coating life of 1567.9 startup cycles (see Table 3-1), and (b)

metallographic section showing the GT33+ coating at the trailing edge (50% blade

height) being protective and in good condition 4-32

xxiv

Figure 4-16 Verification of the COATLIFE prediction against field data of GT33+ coated

GTD111 DS blade: (a) COATLIFE prediction of TMF failure at 25% BH of Blade 7 in

a 7FA machine, and (b) metallographic section showing a TMF crack penetration

into the substrate 4-27Figure 4-17 Verification of the COATLIFE prediction against field data of GT33+ coated

GTD111 DS blade: (a) COATLIFE prediction of TMF failure at 50% blade height of

Blade C in a 7FA machine, and (b) metallographic section showing a TMF crack

penetration into the substrate and exposed blade due to coating spallation 4-28

xxv

LIST OF TABLES

Table 1-1 Semi-quantitative chemical composition of the GTD-111 blade, wt % .1-3Table 1-2 Chemical composition of CT102 coating powder, wt % 1-3Table 1-3 Semi-quantitative chemical composition of NiCoCrAlY coating, wt % .1-5Table 1-4 Average aluminum content and volume of fraction of the β phase in the

coating as a function of thermal cycles between room temperature and 1950°F

(1066°C) 1-19Table 1-5 Average aluminum content and volume of fraction of the β phase in the

coating as a function of thermal cycles between room temperature and 1850°F

(1010°C) 1-19 Table 2-1 Chemical composition of GTD-111 and IN-738 test materials (wt %) 2-2Table 2-2 Chemical composition of NiCoCrAlY bond coating powder (wt %) 2-3Table 2-3 Chemical composition of ceramic coating powder (wt %) 2-3Table 2-4 Time for TBC cracking or spallation after isothermal exposure .2-13Table 2-5 Thermal cycling between 1950ºF (1066ºC) and room temperature test results .2-17Table 2-6 Thermal cycling between 1850ºF (1010ºC) and room temperature test results .2-18Table 2-7 Thermal cycling test results at the peak temperature of 1950°F (1066°C) with

24-hour hold time .2-20Table 2-8 Aluminum content in the bond coating after thermal exposure 2-29Table 2-9 Influence of exposure temperature and time on thermally grown oxide (TGO)

scale thickness on TBC-coated specimens .2-33Table 3-1 Material constants for GT33-like coatings .3-10Table 4-1 Semi-quantitative chemical composition of top aluminide and MCrAlY coating

on the blades, wt % 4-3Table 4-2 Summary of TMF cracking on Blade 7 and Blade 57 .4-4Table 4-3 Chemical composition of the coating at various locations on Blades 57 and 7,

wt % .4-15Table 4-4 Verification of COATLIFE-4 predictions against field data for GT33+ in 7FA

machines 4-33

xxvii

1-1

1

ADVANCED METALLIC COATINGS – DEGRADATION

Under EPRI sponsorship, Southwest Research Institute (SwRI) has developed the COATLIFE code for life prediction of coatings on gas turbine blades The algorithms in COATLIFE take into account all the degradation mechanisms [1-4] involved during long-term service of the coated blades and thermo-mechanical fatigue (TMF) life of the blades The specific physical degradation mechanisms considered

in the model include oxide formation kinetics, spallation of the protective oxide (Al2O3 layer), and interdiffusion of aluminum from the coating into the superalloy blade substrate This approach is able

to account for the contribution of time in service, number of startup and shutdown cycles, and variable temperature operation (that is, part load operation) The model was evaluated and validated for

aluminide and duplex (GT29 + and GT 33+ type) coatings used in the industry

The equipment manufacturers also widely use either NiCoCrAlY or CoNiCrAlY coatings on the blades and vanes of current engines These coatings are similar in composition; the former NiCoCrAlY coating, GT33, is used by General Electric, and the latter coating, CoNiCrAlY (trade name CT102), is used by Siemens-Westinghouse The COATLIFE code requires enhancement to handle this widely used coating The objective of this task was to establish the coating degradation mechanisms for this widely used MCrAlY coating and to generate the necessary laboratory data to determine the model constants for COATLIFE

1.1 Experimental Procedure

1.1.1 Material, Coatings, and Test Specimens

Flat rectangular specimens were machined from the shank section of a retired GTD-111 blade using an electro-discharge machining (EDM) process It is well known that the shank section of a blade operates at a significantly lower temperature than the blade airfoil section As a result, the blade material at the shank section does not degrade due to long-term service exposure The microstructure and mechanical properties

at this location represent the original, as heat-treated condition of the blade [5,6] The dimensions of the specimens are shown in Figure 1-1 The edges of all specimens were beveled, and the flat faces were ground and polished to remove the EDM-induced recast layer prior to coating application

Advanced Metallic Coatings – Degradation

The chemical composition of the blade was determined by using energy-dispersive spectroscopy (EDS) The semi-quantitative chemical composition of a specimen is shown in Table 1-1

composition of CT102 powder is similar to GE’s proprietary GT33 coating, the coating is

hereafter referred to as NiCoCrAlY or GT33-like coating To facilitate rotation of the specimen during coating application on all sides of the specimen, a small bolt was tack-welded to one of the width edges of each specimen After application of the coating, the bolt was removed, and the tack-welded region was ground prior to testing All specimens were given a vacuum

diffusion (or partial solution) heat treatment at 2050°F (1121°C) for two hours and an aging treatment at 1550°F (843°C) for 24 hours

Table 1-2

Chemical composition of CT102 coating powder, wt %

1.1.2 Cyclic Oxidation Tests

Cyclic oxidation tests were conducted using a facility designed and fabricated at SwRI The test facility consists of a furnace, a forced-air cooling system, a coated superalloy frame for

suspending the test specimens, and a computer-controlled moving arm that transfers the

specimens in and out of the furnace and to the cooling system For cyclic oxidation testing, coated specimens were inserted into the furnace, which was maintained at a desired peak

temperature, and held at that temperature for 55 minutes prior to moving them into the cooling system After the specimens were cooled for 5 minutes to room temperature, they were then automatically reinserted back into the furnace Cyclic oxidation tests were performed at two peak temperatures 1850°F (1010°C) and 1950°F (1066°C) The tests were interrupted at

predetermined intervals to monitor weight change of the specimens as a function of thermal cycles At predetermined intervals, a specimen was removed from testing for post-exposure metallographic evaluation

1-3

Advanced Metallic Coatings – Degradation

1.1.3 Metallography

For metallographic evaluation, a transverse section through the centerline was removed from an as-coated specimen to collect the baseline data Transverse sections were also removed from the exposed specimens to determine the coating degradation as a function of thermal cycles All sections were mounted in a conductive mounting media, polished using standard metallographic techniques, and examined under optical and scanning electron microscopes (SEMs) to

characterize the coating structure and to determine the chemical composition of the coating either in the as-coated or post-exposed condition The composition of the phases in the coating was determined using energy-dispersive spectroscopy (EDS) and the volume fraction of the beta (β) phase in the coating was determined using quantitative metallographic techniques

1.2 Results And Discussions

1.2.1 Microstructure of As-Coated Specimens

The metallurgical mount prepared from a specimen in the as-coated, pre-exposed condition was examined under optical and scanning electron microscopes to characterize the coating structure These examinations showed that the coating was dense and exhibited a duplex structure

consisting of β-phase (aluminum-rich) particles in a matrix of γ (solid solution of Ni, Cr, and Co), as shown in Figure 1-2 EDS analysis was performed on several β-phase particles The results showed that a β-phase particle contains about 17 wt % aluminum, while γ contains about

3 wt % aluminum The coating on all faces was relatively uniform in thickness The thickness of the coating varied from 8.5 to 10 mils (216 to 254 µm) In isolated areas, a few fine oxide

particles and voids or pores were observed in the coating The porosity/oxide content in the coating was measured to be less than 0.5% No significant contamination was observed at the coating/substrate interface The interface contamination was found to be less than 0.5% The microstructure of the coating on the specimen was considered normal for the LPPS process [5-7] The coating exhibited irregular surface topography, which is also considered normal for an LPPS-processed coating

1-4

Advanced Metallic Coatings – Degradation

Figure 1-2

Microstructure of as-deposited NiCoCrAlY coating

The chemical composition of the coating was determined by performing EDS analysis on the coating at multiple locations The EDS analysis gives the semi-quantitative composition of the coating The typical results of the analysis are shown in Table 1-3 The chemical composition of the coating is consistent with the composition of the powder given in Table 1-2

Table 1-3

Semi-quantitative chemical composition of NiCoCrAlY coating, wt %

1.2.2 Cyclic Oxidation Test Results

The weight change results at the two peak temperatures, 1950°F (1060°C) and 1850°F (1010°C), are presented in Figures 1-3 and 1-4 as a function of thermal cycles At both temperatures, the

1-5

Advanced Metallic Coatings – Degradation

specimens initially gained weight, due to formation of oxide scale with thermal cycles, up to approximately 300 thermal cycles, and then the specimens lost weight due to the domination of oxide scale spallation with increasing thermal cycles Similar results were reported in EPRI report TR-113899 [8] In the initial stages of exposure, up to approximately 300 cycles, the weight gain data showed no significant scatter The scatter in the weight loss data increased with increasing thermal cycles at both temperatures after about 500 thermal cycles It is clear from the figures that the scatter in the data was extremely large after longer exposure (≥ 1000 one-hour thermal cycles) at both temperatures This suggests that longer exposure promotes formation of relatively non-adherent oxide scale in localized regions The variation of the non-adherent oxide scale formation and spallation among the specimens is, in part, responsible for this scatter

As described previously, a small bolt was tack welded to one of the width edges to facilitate rotation of the specimen during the coating application The tack-welded region on the

specimens was ground prior to cyclic oxidation testing Variation of the uncoated area on the tack-welded edge among the specimens is also, in part, responsible for this scatter Variation of mixed oxide scale spallation from these ground and uncoated regions among the specimens can result in pronounced weight changes

Advanced Metallic Coatings – Degradation

Cyclic oxidation behavior of GT33-like coating at 1850 F (1010 C)

1.2.3 Oxide Scale Characterization

Metallurgical mounts were prepared from all the specimens exposed at the peak temperatures of 1950°F (1066°C) and 1850°F (1010°C) The mounts were examined to characterize the oxide scale Types of oxide scale formed at these peak temperatures varied with thermal cycles

Typical variation of the type of oxide scale formed on the coated specimens is illustrated in Figures 1-5 and 1-6 as a function of thermal cycles In backscatter electron images, the phases lighter in color are rich in heavy elements, such as yttrium, chromium, nickel, cobalt, etc The darker phases are rich in light elements, such as aluminum The variation in contrast is an

indication that the scale contains different type of oxides The specimens exposed for up to approximately 1000 thermal cycles at both peak temperatures exhibited predominantly dark-appearing, presumably aluminum-rich oxide, Al2O3 In isolated areas, fine light-appearing, presumably yttrium-rich particles, were observed in the alumina scale, Figure 1-5(a) The dark-appearing aluminum-rich scale is uniform The presence of a uniform adherent aluminum-rich scale is most likely responsible for minimum scatter observed in the weight change data among the specimens in the initial stages of thermal exposure

1-7

Advanced Metallic Coatings – Degradation

Dark Al-Rich Oxide

Dark Al-Rich Oxide

Yttria or Yttrium

Light Phase Mixed Oxides

SEM backscattered electron images of the oxide scale on the GT33-like coating after

exposure for (a) 600 cycles, (b) 1500 cycles, and (c) & (d) 1750 cycles between room

temperature and 1950 F (1066 C)

1-8

Advanced Metallic Coatings – Degradation

SEM backscattered electron images of the oxide scale on the GT33-like coating after

exposure for (a) 800 cycles, (b) 1500 cycles, (c) after 2000 cycles, d) 3486 one-hour cycles, and (e) & (f) 4500 one-hour cycles between room temperature and 1850 F (1010 C)

1-9

Advanced Metallic Coatings – Degradation

SEM backscattered electron images of the oxide scale on the GT33-like coating after

exposure for (a) 800 cycles, (b) 1500 cycles, (c) after 2000 cycles, d) 3486 one-hour cycles, and (e) & (f) 4500 one-hour cycles between room temperature and 1850 F (1010 C)

EDS analysis was performed at several locations on the scale on each specimen to characterize the dark-appearing and light-appearing oxides EDS measurements showed that the uniform and continuous dark-appearing oxide scale that formed on the coating within the first few hundred cycles of exposure at both peak temperatures was predominantly aluminum-rich oxide,

presumably Al2O3 A typical EDS spectrum obtained from a continuous dark-appearing oxide scale shown in Figures 1-5(a) and 1-6(a) is presented in Figure 1-7 The small light-appearing precipitate particles in the dark aluminum-rich oxide were found to be a yttrium-rich phase Figure 1-8 shows the EDS spectra obtained from the light-appearing scale denoted as a mixed oxide in Figures 1-5(b) and 1-6(b) The scale is found to contain Ni, Co, Cr, Ti, and Al,

suggesting that it is a mixture of oxides of these elements It is well known that a continuous

Al2O3 scale exhibits excellent spallation resistance The mixed-oxide scale is less resistant to spallation than a stable and adherent aluminum-rich oxide, Al2O3 [9,10] The presence of Ti in the oxide lowers the spallation resistance of the scale

It is apparent from the backscattered electron micrographs that the scale on the specimens

exposed for about 1500 cycles at both peak temperatures is a mixed oxide (see Figures 1-5(b) and 1-6(b)) Unlike Al2O3, the mixed-oxide scale was not found to be uniform, but it was

observed only in isolated areas Under the mixed-oxide scale, a thin layer of aluminum-rich oxide was observed in a majority of the samples The metallographic results further showed that the amount of mixed-oxide scale on the coated samples increased with increasing thermal

exposure In the specimens exposed for 1750 cycles at 1950°F (1066°C) and 4500 cycles at 1850°F (1010°C), no aluminum-rich oxide was found under the mixed-oxide scale in isolated areas as shown in Figures 1-5(d) and 1-6(f) In isolated locations after long-term exposure, a continuous light-appearing mixed-oxide scale was also observed as shown in Figure 1-5(d) Variation in formation and spallation of these mixed oxides among the samples during long-term

1-10