DE1ERMINA nON OF DEGREE OF 1HERMAL BARRIER COATING DIS BONDING

BY TIME-RESOLVED INFRARED RADIOMElRY (TRIR)

J.W. Maclachlan Spicer, W.D. Kerns, L.C. Aamodt and lC. Murphy

Center for NDE, Applied Physics Laboratory The Johns Hopkins University Johns Hopkins Road, Laurel, MD 20723-6099, USA

IN1RODUCITON

We have previously described the time-resolved infrared radiometry (TRIR) technique and demonstrated how both measurement of coating thickness and detection of coating disbonding can be made within the same measurement [1]. We have also reponed indications of the sensitivity of the lRlR technique to differing degrees of coating disbonding [2]. In the present work we examine this question in detail and compare TRIR experimental results with both an analytical multilayer theory and a destructive analysis of the coating-substrate interface. While the material system studied in this work is a zirconia thermal barrier coating on a superalloy substrate, the methodology and analytical basis of the technique are applicable to a wide variety of materials and components including printed circuit boards and composite materials. The capability of the TRIR technique for characterization of these specific systems will depend on details of the thermal properties and layer thickness in these specimens.

EXPERIMENTAL 1ECHNIQUE

The time-resolved infrared radiometry (1RIR) technique differs from many other pulsed radiometry techniques [3,4,5] in that we monitor the increase in specimen surface temperature during the application of a step heating pulse of specified duration rather than following the decay in temperature after application of a shon heating pulse. The experimental system consists of an infrared scanner (Mikron 6T61) used to monitor the time development of the surface temperature of the specimen. Synchronizing electronics and an acousto-optic modulator are used to gate a heating beam from an argon ion laser (Coherent Innova 90-5) at a fixed time with respect to the frame rate of the scanner. For most of the images presented here, a line heating source was employed. However a variety of heating beam patterns on the specimen can be produced by use of various lenses and an acousto-optic deflector. A positioning system (Burleigh CE-2000) is used to translate the specimen with respect to the line source. Operation of the infrared scanner, synchronizing electronics and positioning system is all under control of a microcomputer (Macintosh II) over the IEEE-488 bus.

With this system it is possible to produce area maps of the time development of the surface temperature with a time resolution equivalent to the scan rate of the horizontal galvanometer (8.3 msec). The surface temperature along one horizontal line on the infrared scanner is monitored as the laser heating pulse is applied to the sample. A position-time temperature image is generated for this location on the sample. The sample is then translated venically so that the line heating source and the infrared viewing line have been shifted to a new horizontal line on the sample and a new position-time image is then recorded for the new location. After the desired number of positions on the sample have been recorded, a lRIR X-

Review of Pro!!ress in Quantitative Nondestructive Evaluation. Vol. lOB Edited by D.O. Thompson and D.E. Chimenti. Plenum Press. New York. 1991

1193

Y image such as those shown in Fig. 4 is generated for any particular time after the turning on of the heating pulse. This is done by extracting the line corresponding to the desired time interval from each of the position-time images and stacking them in the order in which they were acquired to generate a full-field image of the temperature of the specimen at the desired time.

While the TRIR X ~ Y images at different time intervals reveal spatial variations in the temperature response across the sample, it is necessary to examine the full temperature-time development at a single location in order to determine coating parameters such as thickness and bonding integrity. Plots of temperature versus time such as those shown in Fig. 6 can be made for any location on the specimen by extracting a vertical line from the appropriate position-time image for the desired location.

MULTll...A YER THEORY

While multilayer systems had been studied in an earlier work using fmite element analysis [6], previous analytical work has focussed on single coating layers on semi-infinite substrates. We have now developed an analytical theory which is applicable to multilayer systems and have used this to describe experimental results on an air-zirconia-superalloy­zirconia-air specimen [7]. In the present work we use this model to predict the effect on the TRIR response of varying the thickness of an air layer between a zirconia coating and a superalloy substrate.

A schematic diagram of the sample configuration under consideration is shown in Fig. 1. The zirconia layer is 250 microns thick and the thickness of the air layer is varied from 0 to 100 microns. The multilayer theory was used to calculate the expected variation of surface temperature as the specimen is heated and these plots are given in Fig. 2. Note that the temperature is plotted against the square root of time since the surface temperature of a heated semi-infinite specimen increases as the square root of time. This then provides a convenient reference frame for evaluating the temperature development of layered specimens. The semi­infinite response is shown as a straight line on the temperature versus square root time plots in Fig. 2 and Fig. 6.

Figure 1

1194

One-dimensional excitation

Zirconia (KZ' pz ,Cz )

Superalloy (Ks, Ps, Cs)

Schematic diagram showing the sample configuration for the multilayer analysis.

Figure 2

1.0 r;:==:::::r:========::;T--~i-__ A ~

0.8

0.6

0.4

0.2

Air Layer 1 2 3 4 5 6 7 8 9 10

Thickness 100 microns 70 microns 50 microns 30 microns 10 microns 7 microns 5 microns 3 microns 1 micron Omicrons

l l I !

.f ......................... ..1 ......... . : :

I i , ~

• .:0 .................. .

i L-____ ~------~------~!

~ Semi-infinite -----ji~~t:

--o-----r------r------- . ........................ :..................... . I

I -I-------r--------r-------0~~~~_4_4-+-+~~~~~~+_+_+_+_~

o 0.2 0.4 0.6 0.8 1.0

Square Root Time (sec 1/2 )

Analytical results for variation of surface temperature as a function of square root of time for a multilayered specimen of a zirconia coating of thi~l<:ness 250 microns on a thermally-thick superalloy substrate with an intervening air layer of thicknesses ranging from 0 to 100 microns.

The temperature development for the case of an air layer thickness of 0 microns gives the same results as the well-bonded case in the earlier single-layer coating model [8]. Here the temperature rises proportional to the square root of time until a characteristic thermal transit time is reached when the presence of the substrate is sensed. This thermal transit time is approximately equal to 0.36 VIa. where L is the coating thickness and a. is the through­thickness thermal diffusivity of the coating. Since the substrate is more thermally conductive than the coating, the rate of temperature increase drops at this time. At the other end of the range of air thicknesses, the result for a 100 micron layer shows a dramatic rise in the rate of temperature increase when the thermal transit time is reached. This is due to the thermally insulating nature of this thick air layer. Of interest for the evaluation of coating-substrate separation is the temperature-time development for air layer thicknesses between these two extremes. Fig. 2 shows that there is a continuous range of temperature-time responses depending on the thickness of the air layer. This provides the potential for determining the coating-substrate separation using the TRIR technique.

1195

EVALUATION OF TIIERMAL BARRIER COATING DISBONDING

Time-Resolved Infrared Radiometry Measurements

A specimen of zirconia thennal barrier coating on a superalloy rod was examined using the TRIR technique. This specimen was obtained from Pratt and Whitney and had been tested in a burner rig to produce localized spallation of the coating. TRIR X -Y images were performed at six locations around the circumference of the rod. These locations are indicated on the sketch of the rod specimen in Fig. 3 and the images are shown in Fig. 4. Each image is for a time slice at 1 second after the beginning of the heating pulse and covered an area 18 mm wide by 40 mm long. The decrease in signal strength near the edges of each image is due to the heating beam profile.

The location of the spalled area of coating where the bare substrate is visible is indicated by the "S" in the TRIR X-Y images. Note that the temperature increase here is negligible due to the high thermal conductivity of the substrate material. The temperature increase is also small in areas of the coating at the bottom and top of the images where the coating is well bonded to the substrate. In the vicinity of the spalled region, however, the temperature response shows a wide range of values, with the maximum response occurring at the periphery of the spalled region. It is important to note that this structure was not evident in either white light viewing of the sample or in passive infrared imaging of the sample. This structure appeared only in the TRIR X-Y images.

Figure 3

• I

i,,~ i -: : j

I I I ! I 3

~ ! i

I i

! I

r-A+B+C+D+E,+F~

Linescan Region

Sketch of thermal barrier coating rod specimen indicating positions of TRIR X­Y images performed around the circumference of the rod.

In order to obtain quantitative information on the nature of the coating-substrate interface as a function of location in the TRIR X-Y images, the temperature versus square root time response was analyzed for a series of locations around the circumference of the rod. These locations are indicated in the cross-sectional view of the sample shown in Fig. 5. The location of the cross-section is indicated by the horiwntalline in Fig. 3. The TRIR temperature versus root-time linescans for these positions are shown in Fig. 6 along with the response for a w.ell-bonded region of coating near the top of the sample and a straight line indicating the theoretical response expected for a semi-inf'mite specimen of zirconia. The curves show the same range of responses as predicted by the multilayer analytical model for a range of different air gap thicknesses.

1196

rJ:)

-.J

Fig

ure

4 TR

IR X

-Y im

ages

(tim

e sl

ice

at 1

sec)

of t

henn

al b

arrie

r rod

spec

imen

for

6 po

sitio

ns a

roun

d th

e ci

rcum

fere

nce

of th

e ro

d. T

he "

S" in

dica

tes

the

posi

tion

of th

e sp

alle

d re

gion

rev

ealin

g ba

re s

ub­

stra

te m

ater

ial.

Zirconia Coating

Figure S

Figure 6

1198

Air Separation

Superal\oy Substrate

8 6 7

Cross-sectional view of thermal barrier rod specimen indicating locations at which the temperature vs. root-time linescans shown in Fig. 6 were performed.

14 rr==:======~====~--;-~--:---~~ Angular

12

10·

8 .

6

4

Linescan

1 2 3 4 S 6 7 8 9 10 11 12

2 ········ .............. ·· .. T .... ·· .. ··

o 0.0 0.2

Position

o degrees 30 degrees 60 degrees 90 degrees 120 degrees ISO degrees 180 degrees 210 degrees 240 degrees 270 degrees 300 degrees 330 degrees

0.4

Well Bonded

0.6 0.8

Square Root Time (sec 1/2 )

1.0

Experimental surface temperature vs. square root time linescans for 12 positions around the circumference of the thermal barrier rod specimen as shown in Fig. S. Also shown for comparison is the response for a well-bonded region of coating and the square-root-time response expected for a semi-infinite zirconia specimen.

Comparison with Destructive Analysis

After the TRIR measurements were completed the sample was returned to Pratt and Whitney for destructive analysis. The specimen was mounted in epoxy, sectioned along the horizontal line indicated in Fig. 3 to reveal the cross-section at the location of the TRIR linescans shown in Fig. 6, and polished to allow evaluation of the coating-substrate interface. Micrographs were made of the interface and the coating-substrate separation was measured.

A comparison of the destructive analysis results with the TRIR measurements is shown in Fig. 7. The open squares show the measured coating-substrate separation as a function of angular position around the specimen. The solid squares show the TRIR ratio for the same positions. This ratio is used to provide a measure of the variation of the TRIR response after the thermal transit time. It is defmed as the ratio of the temperature at 1.0 sec after turning on the heating pulse (well after the thermal transit time of the coating) to the temperature at 0.2 sec (slightly before the thermal transit time). While there is not total point-to-point agreement between the two curves, they do demonstrate similar trends. The TRIR response is largest near 00 where the coating separation is 115 microns and is smallest near 1500 where the coating separation is 40 microns. The TRIR responses for intervening values of coating separation fall between these two extremes and follows the same trends as shown in the multilayer analysis of Fig. 2.

Although the theoretical and experimental results exhibit the same general trends as a function of coating-substrate separation, exact agreement between the measured coating separation and the air layer thickness used in the theoretical model is not obtained. The theoretical air layer thickness which provides the best fit to the experimental TRIR results is of the order of 10 times smaller than the measured coating separation. This discrepancy is due to the use of a simple air layer model rather than taking into consideration the presence of oxidation products and local points of attachment on the thermal transfer through the disbond region. Further work is in progress to identify the effect of local variation in disbonding over a scale finer than the resolution of the TRIR measurement on the TRIR response.

• ~ ~ :3 N

c:::i E-o E-o

r;; .--CIS ~

~ ~

~ E-o

Figure 7

14 120 0 -----'Il

100 = 0 12 100 t.J .-

80 e '-'

10 = 0 60 .--CIS

8 100 CIS

40 Q. ~

rr.J

6 20 I:J)

= .--CIS

4 0 0

-15 75 165 255 345 U

Angular Position Comparison of TRIR ratio with the measurement of coating separation after destructive analysis as a function of position around the circumference of the thermal barrier rod specimen.

1199

CONCLUSIONS

Comparison of TRIR measurements with destructive analysis and with the results of a multilayer, analytical theory demonstrate the capability of the technique to detennine the amount of coating separation in thennal barrier coating specimens. Similar results have also been obtained in other materials systems such as the decohesion of aluminum heat sink from printed circuit boards. Work in progress involves evaluation of the nature of the disbond including use of lateral heat flow to monitor the local structure of the bond.

ACKNO~GEMENTS

Research supported by the Center for Advanced Nondestructive Evaluation for the Air Force Wright Patterson Aeronautical Labs/Materials Laboratory under Sub-Contract SC90-185 with the Ames Laboratory.

REFERENCES

1. J.W. Maclachlan Spicer, W.D. Kerns, L.C. Aamodt and J.C. Murphy, J. Nondestr. Eval. 8., No.2, 107-120 (1989).

2. J.W. Maclachlan Spicer, W.D. Kerns, L.e. Aamodt and J.e. Murphy, in Review of Progress in Ouantitative NDE, edited by D.O. Thompson and D.E. Chimenti (Plenum Press, New York, 1990) Vol. 9, pp. 1169-1176.

3. P. Cielo, J. Appl. Phys. 22, No.1, 230-234 (1984). 4. R.E. Imhof, D.J.S. Birch, F.R. Thomley, J.R. Gilchrist and R.A. Strivens, J. Phys.

D:Appi. Phys. 18, Ll03-Ll06(1985). 5. D.L. Balageas, J.C. Krapez and P. Cielo, J. Appl. Phys. 59, No.2, 348-357 (1986). 6. J.W. Maclachlan, L.C. Aamodt and J.C. Murphy, in Review of Progress in

Ouantitative NDE, edited by D.O. Thompson and D.E. Chimenti (Plenum Press, New York, 1989) Vol. 8, pp. 1297-1304.

7. L.C. Aamodt, J.W. Maclachlan Spicer and J.C. Murphy, J. Appl. Phys., to be published.

8. J.W. Maclachlan Spicer, L.e. Aamodt and J.e. Murphy, Ceram. Eng.Sci. Proc. 2., No. 9-10, 1181-1188 (1988).

1200