TMS ONLINE | TMS PUBLICATIONS | SITE MAP An Article from the January 2003 JOM-e: A Web-Only Supplement to JOM

VISIT THE JOM COVER GALLERY JOM MENU

INTRODUCTION

Continuous-fiber-reinforced ceramic-matrix composites (CFCCs) are promising materials for high-temperature structural applications.^1–8 They provide good corrosion resistance and a good balance of strength and toughness at temperatures above 1,000°C, and they are lightweight compared to metallic materials. Due to those advantages, CFCCs have been widely used for high-temperature structural and aerospace applications such as heat exchangers, gas turbines, and space shuttle engine components.
To encourage wide applications of CFCCs, nondestructive evaluation (NDE) is essential not only for assuring manufacturing quality and lifetime service applications of CFCCs, but also as a characterization means for the research and development of advanced materials. Nondestructive evaluation techniques such as ultrasonic testing (UT) and infrared (IR) imaging can be powerful methods to investigate fracture behavior and defect information in the composites.

Several NDE techniques are available for assessing the integrity of composite materials.^9,10 For example, the UT method has become one of the most commonly used inspection techniques to determine the quality and internal structure of composite materials.^9–20 The attenuation of ultrasonic signals through the material, which results from the interactions between ultrasonic waves and defects or microstructural features in the composite, can be determined by measuring transmitted ultrasonic amplitudes. Ultrasonic testing has been successfully applied to characterize composites based on variations in transmitted ultrasonic amplitudes.^9–20 The transmitted ultrasonic waves can be related to the presence of defects and anomalies in the composites.

Infrared (IR) thermography, another NDE method for the characterization of CFCCs, is based on the concept that after applying a uniform heat pulse to a sample surface, a localized disruption of the heat flow will occur when defects and/or flaws are present. The change in heat flow translates into temperature differences on the material surface,^21–26 which can be used to create thermographic images in terms of either temperature difference or thermal diffusivity. In this study, IR thermography was employed to map the through-thickness thermal diffusivity of CFCC specimens and relate the results to the UT-transmitted amplitude. Infrared thermography can also be used for in-situ measurements of temperature changes during mechanical testing.

However, relatively little work has been performed relating the NDE results to understanding the mechanical behavior of CFCCs. Therefore, the main objectives of this research are to assure the quality and structural integrity of Nicalon/SiC composites using NDE techniques; to perform NDE using UT and IR thermography methods for the analyses of defect distributions that may affect mechanical properties; to investigate tension behavior of Nicalon/SiC composites with the aid of NDE methods; and to provide fracture and NDE information to aid in the fabrication, development, and selection of Nicalon/SiC composites for structural applications.

EXPERIMENTAL PROCEDURES

The Materials Systems

Figure 1. A schematic illustration of the chemical vapor infiltration process.27

Two types of continuous Nicalon™ (Nippon Carbon Company Ltd., Japan) fiber-reinforced SiC ceramic matrix composites were used for this study: plain-weave Nicalon/SiC composites and 0°/90° cross-ply Nicalon/SiC composites. Nicalon, an amorphous/crystallite fiber, is predominantly SiC, with a diameter of approximately 10 mm to 15 mm and a chemical composition of 59% Si, 31% C, and 10% O₂ by weight. Both composites had a fiber volume content of 40% and were fabricated by the chemical vapor infiltration (CVI) technique²⁷ (Figure 1).

Two types of tensile specimens were machined from plain-weave Nicalon/SiC composites and 0°/90° cross-ply composites, respectively. The reinforcements were two-dimensional (2-D) plain-weave laminates that were interlaced by an insertion of the 90° yarns into 0° yarns, and thus, made with fabrics of woven Nicalon fibers, as shown in Figures 2a and 2b. The reinforcement fabrics were stacked in a graphite mold and then densified with a silicon-carbide matrix by the CVI technique. Each yarn consisted of approximately 500 Nicalon fibers. The main defects in plain-weave composites included the surface flaws induced by the CVI process and the intersection voids between the 0° fill (transverse yarn) and 90° warp (lengthwise yarn) with eight layers in about 3 mm thickness.

The 0°/90° cross-ply laminate preform was composed of alternate layers of 0° and 90° unidirectional Nicalon fibers. The 24-ply laminates ([0/90]₁₂) in approximately 3 mm thickness were chemically bonded together with CVI techniques. The potential defects of the cross-ply composites included interlaminar porosity, as shown in Figures 3a and 3b.

Figure 2. The microstructures of plain-weave Nicalon/SiC composites.
Figure 3. The microstructures of cross-ply Nicalon/SiC composites.

Using the CVI method, the two types of laminated Nicalon fiber fabric preforms were deposited in an inductively heated susceptor cavity such that the preforms and the methyltrichlorosilane (CH₃SiCl₃) gas were maintained at a uniform temperature of approximately 1,000°C. Silicon carbide was deposited from the decomposition of the methyltrichlorosilane gas at the uniform temperature.

NDE of Composites before Mechanical Testing

Nondestructive evaluation techniques were used to characterize the defect distribution and to provide defect information for composites before mechanical testing. Nondestructive characterization with as-received specimens could provide inherent defect information and/or distributions, which may affect the fracture behavior during mechanical testing. At first, UT was applied to CFCCs specimens to provide a 2-D defect distribution as a function of ultrasonic transmitted amplitudes. The other NDE technique, IR thermography, was employed for thermal-diffusivity mapping of CFCCs specimens.

Figure 4. A schematic of the through-transmission ultrasonic testing setup.

Ultrasonic Testing

Ultrasonic NDE is the most widely used technique for the detection and characterization of composite materials.^9–20 Ultrasonic testing (UT) is a nondestructive method in which beams of high-frequency sound waves are introduced into materials for the detection of both surface and internal flaws.^11–20 The UT sound waves travel through the material with some attendant loss of energy and are deflected at interfaces and/or defects. The deflected beam can be displayed and analyzed to assess the presence and location of flaws or discontinuities. Most ultrasonic inspection is done at frequencies between 0.1 MHz and 25 MHz. In this research, ultrasonic amplitude measurements were performed using a through-transmission C-scan mode at a frequency of 15 MHz, in an immersion tank (Figure 4).

Figure 4 shows a schematic diagram of the through-transmission ultrasonic (TTU) setup using a pair of focused transducers. The transmitting transducer (pulser) and receiving transducer (receiver) are aligned with each other. The beam path is always kept perpendicular to the test specimen during the scan and the specimen is placed at the focal zone. In this work, the transducers have a diameter of 1.27 cm and a focus length of 5.08 cm.

The reason for selecting the TTU geometry is partly due to the high attenuation of the material and the difficulty of obtaining and interpreting pulse-echo signals. In the TTU approach, a time-gate is applied to encompass the transmitted signal and to record its amplitude during an UT C-scan.

Moreover, the C-scan image was developed by capturing and displaying the transmitted signals in a raster scan along the X-Y plane of the test sample (Figure 4). Because the specimen was fairly porous and was immersed in water, some of the surface-breaking voids were filled by water. This phenomenon can have an effect on the scan results, depending on the relative volume percent of surface-breaking voids and blind voids. For the specimens used in this work, this effect was not a major factor. However, further work is underway to eliminate the problem of water intake.

The amplitude of the transmitted ultrasonic beam is measured by the receiving transducer, which is also focused and positioned symmetrically with respect to the transmitting transducer. The amplitude of the TTU signal is quite sensitive to the presence of internal defects, such as voids and delaminations, and to variation of the internal structure, including the undulation of fiber tows. Any defects or internal material conditions that attenuate, scatter, or block the transmitting ultrasonic beam will result in a low TTU signal in the C-scan image.

Figure 5. A schematic illustration for mapping thermal diffusivity of Nicalon/SiC samples using a xenon-flash method.

Infrared Thermography

Infrared thermography is a powerful NDE technique for the characterization of composite materials.^9,28,29 Since the composite materials show relatively high emissivities, composites are suited for examinations with or without surface treatments.⁹

In this study, thermography was used to measure the thermal diffusivity. Figure 5 shows the xenon-flash thermal-diffusivity method. A xenon-flash lamp provided a short heat pulse (2,400 W in this investigation) to the front surface, and the IR camera was used to record the temperature rise at the back surface after the pulse. The system uses Parker's method³⁰ to calculate thermal diffusivity. Parker's method assumes no heat loss during the test. Although this assumption can generate a small systematic error (3–5%) in thermal diffusivity, this method was chosen because of its simplicity, and more importantly, because the main focus of the present research was on the variation from point to point. Infrared images were acquired using a personal computer and thermal diffusivity was calculated pixel by pixel.

The theoretically predicted back-surface temperature, T, as a function of time, t, and specimen thickness, L, according to Parker³⁰ is given by

(1)

where t_0.5 is the half-rise time.^29,30

Tensile Testing

Monotonic tensile tests were conducted using a computer-controlled Material Test System 810 servohydraulic frame equipped with hydraulic grips. The tensile tests were performed at room temperature under displacement control at a cross-head speed of 0.5 mm/min. In this study, dog-bone type flat specimens were used to investigate the damage evolution during mechanical tests. The IR camera was employed for in-situ monitoring of the temperature change during the tests. Aluminum tabs were attached with epoxy glue at each side of the shoulder section of the coupon in order to avoid the rupture of the grip parts during loading the sample.

RESULTS AND DISCUSSION

NDE before Tensile Testing

In order to obtain a defect distribution and/or information for as-received samples, UT C-scans were performed on CFCC samples and a xenon-flash technique was used to obtain thermal-diffusivity maps. Figures 6a and 6b present the NDE results prior to mechanical testing (i.e., the UT C-scan and thermal-diffusivity map for the plain-weave Nicalon/SiC composite, respectively).

Figure 6. A qualitative relation between (a-left) a UT C-scan and (b-right) a thermal-diffusivity map obtained for the plain-weave Nicalon/SiC sample before mechanical testing. Note that all NDE results were obtained from the gage section of each dog-bone type flat bar tensile specimen.

The UT C-scan result of the plain-weave Nicalon/SiC sample presents relatively low values of UT-transmitted amplitudes through the whole gage section of the sample. Most areas of the sample are composed of yellow and red colors, which indicate relatively low values of UT amplitudes at the UT scale bar, as shown in Figure 6a. Due to the attenuation and/or reflection of UT sound waves at the presence of pores and/or defects, relatively low values of UT amplitudes have been obtained in the defect-containing areas. Generally, blue or green colors indicate higher UT amplitude values and defect-free areas, while red or yellow colors represent defect-containing regions. Figure 6b exhibits a thermal-diffusivity map for the plain-weave Nicalon/SiC sample with a thermal-diffusivity range of 0.005 cm²/s to 0.02 cm²/s. Since the presence of defects retards the rate of the heat flow, relatively lower thermal diffusivity can be obtained in the areas around defects with red or yellow colors.

The qualitative consistency between the UT-transmitted amplitude and thermal diffusivity is shown in Figures 6a and 6b. The end part of the left-hand side and some part of the right-hand side in the gage section of the plain-weave sample show much higher values of UT-transmitted amplitudes (i.e., more green areas than red and yellow areas), as shown in Figure 6a. The same areas in the thermal-diffusivity map show a similar trend (i.e., higher thermal diffusivity with sky-blue areas as compared with green areas), as shown in Figure 6b. Therefore, a qualitative agreement can be found between the UT C-scan and thermal-diffusivity map for the plain-weave sample, as shown in Figures 6a and 6b.

Like the plain-weave composite, a similar agreement has been found in the case of the cross-ply Nicalon/SiC composite, as shown in Figures 7a and 7b, for the UT C-scan and thermal-diffusivity map, respectively. In Figure 7a, the UT C-scan result for the cross-ply composite showed overall lower UT amplitude signatures at the left-hand side of the gage section of the sample as compared with the plain-weave composite (Figure 6a) (i.e., there are more red areas in the gage section in Figure 7a). In Figure 7b, the thermal-diffusivity map for the cross-ply composite presented variations of thermal diffusivity through the gage section of the sample. Qualitative agreement between the two images of the UT C-scan and thermal-diffusivity map seems to be present because very low values (red or yellow colors) of UT amplitudes can be seen on the left-hand side of the gage section in the cross-ply sample (Figure 7a), and the corresponding area on the thermal-diffusivity map (Figure 7b) represents low values (red or yellow colors) of thermal diffusivity.

Figure 7. A qualitative relation between (a-left) a UT C-scan and (b-right) a thermal-diffusivity map obtained for the 0°/90° cross-ply Nicalon/SiC sample before mechanical testing. Note that all NDE results were obtained from the gage section of each dog-bone type flat bar tensile specimen.

After the investigation of the qualitative correlation between the UT amplitude and thermal-diffusivity, quantitative analyses were conducted on Nicalon/SiC samples based on both kinds of NDE results. The current image analyses were performed on a Macintosh computer using the public domain NIH Image program [developed at the U.S. National Institutes of Health (NIH) and available on the Internet at http://rsb.info.nig.gov/nih-image/] to establish a quantitative relationship between the UT amplitude and thermal diffusivity. Since both the UT C-scan and thermal-diffusivity images presented more variations along the length direction of the specimen, a line profile was drawn along the length direction with the center of the specimen.

The NIH Image software is very effective for analyzing an image containing the digitized data. Fundamentally, it uses a scaling system for pixels, with a one-to-one correspondence to the darkness of an image. The fundamental concept of the NIH Image analyzer is based on the densitometric analysis, which depends on color-density measurements between pixels of an image. Through these analyses, the results of the relative density calculations in the different color density areas can be obtained. On the NIH Image analyzer, regions containing the lower UT amplitude or lower thermal diffusivity are assumed to be of light color (red or yellow colors) on both images, whereas the dark regions (green or blue colors) are regarded as the higher UT amplitude or higher thermal-diffusivity areas. A relative color distinction between the light and dark regions made possible the quantitative calculations of the UT amplitude and thermal diffusivity on both images.

The NIH Image software provides a plot profile for a given line profile, and the plot profile [pixels (y-axis) vs. the length of the line profile (x-axis)] can be converted into data values. Then, the relative pixel values for the subsequent image analysis could be obtained. The same image analyses were performed for plain-weave and cross-ply Nicalon/SiC composites. Since there were differences in resolution (pixel size) between the images (i.e., the thermal-diffusivity map showed about 380 to 400 pixels along the length direction of the sample while the UT C-scan image contained 220 to 250 pixels along the length direction of the sample) with different samples. Twenty data points, equally spaced along the length direction of each sample, per each image were carefully selected to perform the direct comparison between the two images. Each data point was obtained by taking an average of the values from each portion among 20 portions.

Figures 8 and 9 show the relation between the UT amplitude and thermal diffusivity for plain-weave and cross-ply Nicalon/SiC composites, respectively. As predicted in previous images (Figures 6a and 6b), the plain-weave sample presents relatively lower values of UT amplitudes with 50% to 60% of the relative UT-transmitted amplitude.

Figure 8. Thermal diffusivity increasing with UT amplitude for the plain-weave Nicalon/SiC composite.		Figure 9. Thermal diffusivity increasing with UT amplitude for the cross-ply Nicalon/SiC composite.
Figure 10. The dotted lines indicate actual fracture positions; the final ruptures occurred along the lowest ultrasonic signatures for both composites.

Also, it is understandable that there is a correlation between the thermal diffusivity and UT amplitude. In other words, the UT amplitude generally increases as thermal diffusivity increases, as exhibited in Figure 8, for plain-weave composites. In cross-ply composites, it was observed that the thermal diffusivity increases with increasing the UT amplitude, as shown in Figure 9.

Although a relationship seems to exist between the two images of the UT C-scan and thermal-diffusivity map, it is difficult to say whether it is a linear, exponential, or polynominal function. Further study is needed to predict the exact trend between the two images.

Figure 10 represents the positions of actual fracture after tensile testing. Nondestructive evaluation results before mechanical testing showed a defect distribution in terms of UT-transmitted amplitudes. The data also indicated the possibility of the presence of the defects and/or porosity. Tensile testing found that the actual failures occurred at the defect-rich areas or lower UT-transmitted amplitudes, as shown in Figure 10 for both plain-weave and cross-ply Nicalon/SiC composites. In this respect, NDE before mechanical testing is important to provide defect information of the virgin material as well as to predict the final rupture position. Moreover, NDE can be a powerful characterization means to obtain this kind of information.

Figure 11. (a-top left, c-bottom left) IR camera images and (b-top right, d-bottom right) temperature evolutions during tensile testing for the plain-weave Nicalon/SiC composite.		Figure 12. (a-top left, c-bottom left) IR camera images and (b-top right, d-bottom right) temperature evolutions during tensile testing for the cross-ply Nicalon/SiC composite.

Figure 13. The tensile stress-strain behavior of plain-weave and cross-ply Nicalon/SiC composites.

Figure 14. SEM micrographs showing the fiber pullout of the 90° warp followed by severe matrix cracking of the 0° fill with (a-top) relatively higher magnification, and (b-bottom) lower magnification in the plain-weave Nicalon/SiC composite (9°C temperature increase at the time of fracture).

Figure 15. Severe delaminations after failure of the 0° and the 90° laminates in the cross-ply Nicalon/SiC composite.		Figure 16. A cross-sectional view of the fracture surface in the cross-ply Nicalon/SiC composite (5°C temperature increase at the time of fracture).

ACKNOWLEDGEMENTS

This work is supported by the National Science Foundation, the Combined Research-Curriculum Development Program under contract No. EEC-9527527, the Division of Design, Manufacture, and Industrial Innovation, under contract No. DMI-9724476, and the Integrative Graduate Education and Training Program, under contract No. DGE-9987548 with M.F. Poats, D.R. Durham, W. Jennings and L. Goldberg as contract monitors, respectively. Also, the research is sponsored by the assistant secretary for energy efficiency and renewable energy, Office of Transportation Technologies, as part of the High-Temperature Materials Laboratory User Program, Oak Ridge National Laboratory, managed by Lockheed Martin Energy Research Corporation for the U.S. Department of Energy under contract number DE-AC05-96OR22464. The authors appreciate H. Wang of Oak Ridge National Lab for his excellent technical support on IR thermography.

References

1. R.L. Lehman, S.K. El-Rahaiby, and J.B. Wachtman, Handbook on Continuous Fiber-Reinforced Ceramic Matrix Composites (Westerville, OH: The Am. Ceram. Soc., 1995), p. 495.
2. P.K. Liaw, “Fiber-Reinforced CMCs: Processing, Mechanical Behavior and Modeling,” JOM, 47 (10) (1995), p. 38.
3. P.K. Liaw et al., Acta Metal. et Mater., 44 (5) (1996), p. 2101.
4. N. Miriyala et al., J. Nuclear Materials, 253 (1998), p. 1.
5. N. Miriyala and P.K. Liaw, “CFCC Fatigue: A Literature Review,” JOM, 49 (7) (1997), p. 59.
6. N. Miriyala and P.K. Liaw, “The Monotonic and Fatigue Behavior of CFCCs,” JOM, 48 (9) (1996), p 44.
7. W. Zhao, P.K. Liaw, and N. Yu, J. Nuclear Materials, 253 (1998), p. 10.
8. J.H. Miller, P.K. Liaw, and J.D. Landes, Mater. Sci. and Eng. A, 317 (1-2) (2001), p. 49.
9. S. Ness, and C.N. Sherlock, technical eds., P.O. Moore and P.M. McIntire, eds., Nondestructive Testing Handbook, vol. 10, Nondestructive Testing Overview, 2nd ed. (Columbus, OH: American Society for Nondestructive Testing, 1996).
10. ASM Handbook, vol. 17, Nondestructive Evaluation and Quality Control (Metals Park, OH: ASM Int., 1992), p. 231.
11. D.K. Hsu et al., Mat. Res. Soc. Symp. Proc., 365 (Pittsburgh, PA: MRS: 1995), p. 203.
12. P.K. Liaw et al., J. Nuclear Materials, 219 (1995), p. 93.
13. D.C. Kunerth, L.A. Lott, and J.B. Walter, Ceram. Eng. Sci. Proc., 11 (9-11) (1990), p. 1685.
14. J. Kim and P.K. Liaw, JOM-e, 50 (11) (1998), (http://www.tms.org/pubs/journals/JOM/9811/Kim/Kim-9811.html).
15. J. Kim et al., Nondestructive Evaluation of Continuous Nicalon Fiber Reinforced Silicon Carbide Composites, Nondestructive Evaluation and Materials Properties III, ed. P.K. Liaw et al. (Warrendale, PA: TMS, 1997), p. 55.
16. J. Kim et al., “Nondestructive Evaluation of Nicalon/SiC Composites by Ultrasonics and X-ray Computed Tomography,” Ceramic Engineering and Science Proceedings, 18 (4) (1997), p. 287.
17. P.K. Liaw et al., “Nondestructive Evaluation of Woven Fabric Reinforced Ceramic Composites,” Nondestructive Evaluation of Ceramics, Ceramic Transactions, vol. 89, ed. C. Schilling et al. (Westerville, OH: Amer. Cer. Soc., 1998), p. 121.
18. J. Kim et al., “Thermal and Mechanical Characterization of Ceramic Matrix Composites by Nondestructive Evaluation (NDE) Techniques,” Ceramic Transactions, vol. 124, ed. J.P. Singh, Narottam P. Bansal, and E. Ustundag (Westerville, OH: Amer. Cer. Soc., 2001), p. 241).
19. J. Kim et al., “Damage Assessment of Ceramic Matrix Composites by Nondestructive Evaluation Techniques,” Fatigue and Fracture Behavior of High Temperature Materials, ed. P.K. Liaw (Warrendale, PA: TMS, 2000), p. 59.
20. W. Zhao et al., Metallurgical and Materials Transactions A, 31A (3A) (2000), p. 911.
21. P.K. Liaw et al., Scripta Mater., 42 (2000), p. 2000.
22. H. Wang et al., Metallurgical and Materials Transactions, 31A (4) (2000), p. 1307.
23. L. Jiang et al., Metallurgical and Materials Transaction A, 32 (9) (2001), p. 2279.
24. B. Yang et al., Mater. Sci. and Eng., A314 (2001), p. 131.
25. L. Jiang et al., Transactions of Nonferrous Metals Society of China, 12 (2002), p. 734.
26. H. Wang et al., Metallurgical and Materials Transactions A, 33 (4) (2002), p. 1287.
27. T.M. Besmann et al., Science, 253 (September 1991), p. 1104.
28. S. Ahuja et al., “Thermal and Mechanical Test Methods and Behavior of Continuous-Fiber Ceramic Composites,” ASTM STP 1309, ed. M.G. Jenkins et al. (Philadelphia, PA: American Society for Testing and Materials, 1997).
29. H. Wang, R.B. Dinwiddie, and P.S. Gaal, Thermal Conductivity, 23 (1996), p. 119.
30. W.J. Parker et al., J. Applied Physics, 32 (9) (1961), p. 1679.

For more information, contact P.K. Liaw, University of Tennessee, Department of Materials Science and Engineering, 427-B Dougherty Engineering Building, Knoxville, TN 37996-2200; (865) 974-6356; fax (865) 974-4115; e-mail pliaw@utk.edu.