ANGLE BEAM ULTRASONIC SPECTROSCOPY SYSTEM FOR QUANTITATIVE. Laszlo Adler', Stanislav Rokhlin2, Christophe Mattei', Gabor Blaho' and Qiang Xie2

Transcription

1 ANGLE BEAM ULTRASONIC SPECTROSCOPY SYSTEM FOR QUANTITATIVE INSPECTION OF ADHESIVE BONDS. Laszlo Adler', Stanislav Rokhlin2, Christophe Mattei', Gabor Blaho' and Qiang Xie2 'Adler Consultants, Inc Kinnear Road Columbus, OH The Ohio State University Edison Joining Technology Center 1248 Arthur E. Adams Drive Columbus, OH INTRODUCTION Adhesive joining of materials has found widespread use in industry because of the simplicity and versatility of the process. However, adhesive applications in critical structural joints are still limited by the absence of reliable, rapid industrial NDE methods. Weak bonds remain hidden from conventional inspection methods capable of detecting only gross defects such as voids and open delamination. Weak bonds can exhibit substantially reduced strength without apparent defects. This work focuses on the use of ultrasonic spectroscopy [1,2] to evaluate the integrity of adhesive joint bonded composite materials. Since in general ultrasonic wave interaction with materials is frequency dependent, ultrasonic spectroscopy has advantages for nondestructive evaluation over conventional ultrasonic inspection. In principle, ultrasonic spectroscopy makes use of the conventional A-scan ultrasonic technique by using broadband pulses and frequency spectra to obtain the spectral signature of the material in question. Ultrasonic spectroscopy has been applied to a variety ofn.d.e. problems [3]. Most of the early work was directed toward defect characterization, i.e. determining flaw size, orientation, etc. In addition, the technique has been applied to evaluate mechanical (cohesive) properties of adhesive materials and material properties of composites. We have developed a scanning ultrasonic technique for quantitative evaluation of adhesive bond integrity using Angle Beam Ultrasonic Spectroscopy (ABUS). The novelty Review of Progress in Quantitative Nondestructive Evaluation, Vol. 18 Edited by Thompson and Chimenti, Kluwer Academic/Plenum Publishers,

2 5l -10 '0 IE 8 iij20 ~ c"o o ~ ~ Poorbord 0:: a::: (a) Frequency (M-lz) 7 8 c 0_\, ' \.' -10, (b) --,.. /... '0 IE :.,:., ! : 5 ~ v.' '.'. ~ '. -30 '. ~ --Good bond a::: Poor bond 4 0 ~ Frequency (MHz) Figure 1: Bondline frequency response at normal (a) and oblique (b) incidence. of our approach is that it utilizes an obliquely incident ultrasonic beam on the bond line and measures the frequency response of the reflected ultrasonic signals. The oblique wave introduces shear stress on the bond line and allows discrimination of kissing or poor bonds from good bonds. Also, the bond line thickness is determined from the normal incidence ultrasonic signature and thus the thickness variation is accounted for in data analysis. THEORETICAL BACKGROUND To select the optimal experimental conditions and to relate ultrasonic signature to joint quality we use the spring model of the interface [4,5] for the angle-beam technique. Normal Kn and transverse Kt springs simulate conditions at the interface. The relation between the values of interface springs depends on the physical conditions of the poor interface. Here as an example, we use a single shear spring to describe a damaged interface between the adhesive layer and the adherents. When the shear spring stiffness K t is infinitely large, the interfacial bond is perfect; when Kt is infinitely small, the interfacial bond has no resistance to shear stress, i.e. it degenerates into an ideal slip bond (total disbond). Thus K t can be used as a quantitative parameter to describe the extent of interface damage. In this model the adherents (composites) are considered as solid semispaces; thus interference of ultrasonic waves in the adherent is excluded. This is appropriate when short ultrasonic pulses are used. Using this model we can evaluate critical parameters used in angle-beam spectroscopic inspection techniques, such as angle of incidence, selection of transducer frequency and position of the spectral minima. Figure I a shows reflection spectra from the bondline at normal incidence for poor and good interlace conditions. The results are identical, indicating that normally incident waves are insensitive to the bond quality. Figure 1 b shows the same but for obliquely incident waves. In that case, a significant shift occurs for a shear disbond (solid line) from that off a perfect bond (dashed line). THE ANGLE BEAM ULTRASONIC SPECTROSCOPY (ABUS) In our approach, we combine obliquely and normally incident ultrasonic beam spectroscopy [6]. The two angle measurements allow decoupling the effects of the bond line thickness. To do this we developed a transducer head schematically shown in figure 2. The head accommodates transducers for the normal and oblique measurements. The ultrasonic wave excited by the transducer is reflected from the interface toward a reflector and returns back to the transducer. 1554

3 y-axis motor x-axis motor Scanning Head Adhesive Joint Figure 2: Transducer head. The procedure used to scan the quality of the thin adhesive layer embedded between two composite plates is twofold. At each point of the scan, the nonnal and angle beam time domain signals reflected from the layer are recorded. Then they are analyzed in the Fourier domain and processed using our algorithms to obtain the quality of the bond line. The time signals are recorded using the transducer head. The transducer head is moved over the sample from point to point and each acquisition is recorded. The scan and data acquisition are controlled by a C++ program which synchronizes the transducer head movements and the time domain signal acquisition. The collected time domain signals are stored in the fonn of a 2-D array corresponding to the sampled scan area. After acquisition of the time signals, the second step is analysis in the Fourier domain of these signals to get the relevant parameters for the layer properties detennination. The following procedure is applied for each point: - Gating of the useful part of the time domain signal (shown in Figure 3) - Computation of the Fast Fourier Transfonn > E '-' 0 0 '"0 :::l.-:: E «-200 o -,A ~ V 1 2 A ~ " Vvv 3 Time (/ls) w " 4 5 Figure 3: Typical time domain signal at nonnal incidence and the window used to compute the reflection coefficient 1555

4 -8-10 CO -12 ~ -14 "0.E Q. -16 E <t: Frequency (MHz) Figure 4: Normal incidence reflection coefficient. Cross: experimental result, plain line: reconstructed reflection spectrum. - Gating of the useful part of the frequency domain signal - Determination of the resonance position (in Mhz) and width (shown in Figure 4) - Computation of the layer properties at the corresponding point. The final result (bond line quality reconstruction for every scanned points) is displayed as a bond line quality image in color or gray-level. Figure 5 shows a schematic of the system developed by Adler Consultants, Inc. A three axis mechanical system set on top of a immersion tank is controlled by a PC computer via a motion controller. An ND converter board mounted in the PC and the pulser/receiver are synchronized with the motion of the transducer head. EXPERIMENTAL RESULTS In order to demonstrate the ability of the Angle Beam Ultrasonic Spectroscopy technique to characterize bond line quality, we performed a set of measurements with the Hardware control Data processing Reconstruction Display Stortlge PC Computer \ Figure 5: ABUS Scanner system 1556

5 ABUS Scanner for adhesive quality lap shear joints of different quality and compared the ultrasonic results to joint strength obtained by mechanical testing. Lap shear joints are prepared using 1.6mm thick 2024-T3 aluminum alloy or graphite-epoxy composites and FM-73 adhesive film. The adhesive prepeg FM-73, manufactured by American Cyanamid, is a modified epoxy adhesive film supplied with a polyester knit fabric. The single lap joint has dimensions as described in ASTM D with I inch width and 0.5 inch overlap length. The thickness of the adhesive layer is usually from 140 to m. Part of the prepared samples were aged using a proprietary environmental aging process previously developed at Ohio State University for another purpose. This process reduces the strength of the joint by 30% to 50% and leads to poor adhesion between the adhesive and substrate changing the failure mode from cohesive for standard samples to interfacial for modified samples. The images of the effective moduli for a treated and an untreated adhesive joint samples obtained with the ABUS system are given in Figures 6 and 7. The untreated sample (Fig. 6) presents a fairly homogeneous pattern for both longitudinal and transverse effective stiffness. On the other hand, the stiffness for the treated sample (Fig. 7) presents some important variations over the scanned area and very low stiffness values are measured in some areas. This effect is especially strong for the transverse stiffness which is the more sensitive to interface properties. Failure load: 1708 Ib Figure 6: ABUS bond quality scan and picture of the bond line after mechanical testing for the reference sample Failure load: 720lb Figure 7: ABUS bond quality scan and picture of the bond line after mechanical testing for the treated sample. Darker area represents lower values of the bond quality factor. 1557

6 The lap-shear strengths of treated and untreated samples were measured using an MTS machine. The failure load of the untreated (reference) sample was found to be 1708 lb. The failure load of the treated sample was 720 lb. As shown on Figures 6 and 7, the failure mode is mixed: partially coherent for the good sample (failure occurs mostly in the bulk of the adhesive) and purely interfacial for the treated sample (failure occurs at the interface adhesive/adherent) which is in good agreement with the ABUS image. The reconstructed thickness of the adhesive layer is in very good agreement with the actual thickness measured with a micrometer. Figures 8 and 9 summarize the results obtained for Aluminum and uni-directionnal composites samples. A good correlation between joint strength and the effective moduli obtained by averaging over the joint area is found. CONCLUSION Adler Consultants, Inc. has developed and tested a prototype of the Angle Beam Ultrasonic Spectroscopy (ABUS) scanner. This system performs a simultaneous ro 1.2 :6 (J) 1.1 : : : : l ~ :::: ~ 01'0 C 0.Q E '0 > N:P := 0 C\l ~ 0. 9 z 0.8 o Bond strength reduction (%) Figure 8: Normalized longitudinal effective modulus versus bond strength for Aluminum adhesive bonds. 1.2, (J)1.1 > ~ :.e:::o :J 0'0 & 0 E al ~ 0.9 ::c. ~ ro ::::l E :g, 0.8 o c z.q o Bond strength reduction (%) Figure 9: Normalized longitudinal effective modulus versus bond strength for unidirectional composite adhesive bonds. 1558

7 measurement of the normal and oblique incidence reflection spectra. Determination of the bond integrity is then performed from the reflection spectra at each measurement point, allowing a mapping of the bond line quality. ABUS Scans have been performed on reference and weakened (treated) samples. The ABUS scans show a lower effective stiffness of the bond line for the treated sample and give some important information about the non-homogeneity of the bond line quality. A linear relation between the effective moduli determined with the ABUS technique and the measured strength of the bonds is observed. These tests show the ability of the ABUS technique to provide a parameter proportional to the bond strength. The ABUS scanner is a promising new technique for rapid and reliable evaluation of adhesive bond integrity. ACKNOWLEDGEMENTS This work is supported by Naval Air Warfare Center Aircraft Division, SBIR Phase II program. Technical monitor: Dr Ignacio Perez. REFERENCES 1. D.W. Fitting" and L. Adler" Ultrasonic Spectral Analysis/or Nondestructive Evaluation, Plenum Press, New York (1981). 2. L. Adler" V. Cook, and W.A. Simpson, "Ultrasonic Frequency Analysis," Research Techniques in Nondestructive Testing J., 1-49 (1977). 3. L. Adler, K. V. Cook, R.W. McClung, and H.L. Whaley, "Applications of Frequency Analysis in Ultrasonic Testing," Mater. Ev. 33, 19 (1975). 4. J. M. Baik, and R.B. Thompson, "Ultrasonic scattering from imperfect interface: quasi-static model," J. Nondestr. Eval.1, (1984). 5. A. I. Lavrentyev and S.1. Rokhlin, "Models for ultrasonic characterization of environmental degradation of interfaces in adhesive joints," J. Appl. Phys. 76, (1994). 6. W. Wang and S. I Rokhlin, "Evaluation of interfacial properties in adhesive joints of aluminum alloys using angle-beam ultrasonic spectroscopy," J. Adhesion Sci (1991). T e c h n o l. ~, 7. A. I. Lavrentyev and S.I Rokhlin, "Determination of elastic moduli, density attenuation and thickness ofa layer using ultrasonic spectroscopy at two angles," J. Acoust. Soc. Am, 102, (6), (1997). 1559