Laser-ultrasonic surface wave dispersion measurements on surface-treated metals

Transcription

1 Ultrasonics 42 (2004) Laser-ultrasonic surface wave dispersion measurements on surface-treated metals Alberto Ruiz, Peter B. Nagy * Department of Aerospace Engineering and Engineering Mechanics, University of Cincinnati, Baldwin Hall 745, ML0070, Cincinnati, OH , USA Abstract Surface acoustic wave (SAW) velocity spectroscopy has been long considered to be one of the leading candidates for nondestructive characterization of surface-treated metals because of its ability to probe the material properties at different penetration depths depending on the inspection frequency. We developed a high-precision laser-ultrasonic technique to study the feasibility of SAW dispersion spectroscopy for residual stress assessment on shot-peened metals. This technique is capable of measuring SAW dispersion with a relative error of 0.1% over a frequency range from 2 to 15 MHz. Our experimental results obtained from shotpeened aluminum 2024-T351 samples indicate that the dispersion of the surface wave is a superposition of different effects of surface treatment in the material, including surface roughness, compressive residual stress, and cold work. Although the surface roughness induced component is often the dominating part of the overall dispersion, the experimental results also indicate that it is feasible to observe a perceivable change in the dispersion of the SAW when the specimen is heat-treated at different temperatures, which has no perceivable effect on the surface roughness. The part of the dispersion, which changes during annealing via thermal relaxation, is due to near-surface residual stresses and the decay of texture, although at high frequencies nonuniform grain coarsening could also play a significant role. Ó 2004 Elsevier B.V. All rights reserved. Keywords: Surface acoustic waves; Laser-ultrasonics; Dispersion; Nondestructive evaluation 1. Introduction Surface wave dispersion measurements can be used to nondestructively characterize shot-peened, laser shock-peened, low-plasticity burnished, and otherwise surface-treated metals. In recent years, there have been numerous efforts to separate the contribution of surface roughness, which is not affected by thermo-mechanical relaxation, from those of near-surface material variations, i.e., the primary residual stress effect and the secondary cold work effects (such as anisotropic texture and increased dislocation density), which all significantly decay during relaxation. State-of-the-art laserultrasonic scanning and sophisticated digital signal processing methods allow us to measure the velocity on rough shot-peened specimens with high accuracy [1]. Generally, the measured dispersion of the surface wave * Corresponding author. Tel.: ; fax: / address: peter.nagy@uc.edu (P.B. Nagy). arises from three different sources. First, there is an apparent dispersion due to the diffraction of the surface acoustic wave (SAW) as it travels over the surface of the specimen. This dispersion effect is on the order of 0.1%, which is significantly higher than the experimental error associated with the measurement and comparable to the expected velocity change produced by near-surface compressive residual stresses in metals below their yield point. Second, there is a real but spurious dispersion caused by SAW scattering on the rough surface, which is an adverse geometrical byproduct of certain surface treatment procedures such as shot peening. Finally, there is the principal dispersion caused by a number of material effects of the surface treatment, including the primary compressive residual stress effect and the secondary cold work effect. The cold work effect itself can be further divided into at least three main contributions, namely anisotropic texture, increased dislocation density, and grain refinement/coarsening. Generally, all these effects must be taken into consideration and properly accounted for in order to quantitatively X/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved. doi: /j.ultras

2 666 A. Ruiz, P.B. Nagy / Ultrasonics 42 (2004) characterize the effect of surface treatment on the fatigue life of fracture-critical components. The use of SAWs for the characterization of surface roughness, surface residual stress, and coating thickness measurements has been studied by several authors [2 6]. In addition, the effect of surface cracks on the attenuation and dispersion of Rayleigh waves was explored in great depth [7 12]. These studies have found that the total change in the Rayleigh wave velocity is typically less than 1%. One of the main challenges, of course, is to measure the Rayleigh wave velocity with sufficiently high accuracy. In this paper we will demonstrate that laser-ultrasonic measurements allow us to measure the velocity on rough shot-peened specimens with better than 0.1% accuracy. 2. Laser interferometric dispersion measurements Laser interferometry employs the principle of optical interference to recover the sought acoustic information from the light reflected from or scattered by a surface subject to ultrasonic vibration. Its noncontacting nature makes laser probing the preferred alternative to contact methods in investigating the spatial evolution of surface waves, their dispersion, diffraction and damping on inherently rough surface-treated metals. Fig. 1 shows a schematic diagram of the experimental arrangement used in our surface wave dispersion measurements [1]. A wedge transducer was mounted on the specimen and the surface wave was detected by a LUIS35 Fabry-Perot interferometer. The ultrasonic signal detected by the interferometer was digitized and averaged by a LeCroy 9310 oscilloscope and then sent to a computer for further processing. The specimen was mounted on a Velmex translation table and the relative position between the wedge transmitter and the laser spot was changed by a computer accessible stepping motor controller. The same LabView program controlled both the data acquisition and the scanner. In order to assure the absolute accuracy of our velocity measurements, the temperature of the specimen was stabilized at 26.6 C within ±0.2 C (temperature variations are not expected to lead to dispersion unless the surface temperature of the specimen is significantly different from that of the interior). Spatial averaging was achieved via lateral scanning normal to the wave propagation direction to reduce the incoherent scattering by surface roughness and the inhomogeneous microstructure as well as coherent diffraction effects, especially in the near-field of the transmitter. A slower stepwise scanning was performed in the axial direction, i.e., parallel to the wave propagation, in order to map the phase of the surface vibration as a function of the propagation distance. The LabView program gradually changed the delay time of the digital oscilloscope with respect to the trigger signal by using the programmable digital delay and highly accurate quartz master clock of the oscilloscope. At each step, the new delay time was calculated from the scanning position using a nominal tracking velocity. In most cases, the recorded signal moved only very slightly either forward or backward within the data window depending on whether the chosen tracking velocity was a little higher or lower, respectively, than the actual surface wave velocity. In an ideal case, the tracking velocity is very close to the actual one, and the broadband pulse does not move appreciably at all, though its shape slightly changes as a result of dispersion. The recorded data was spectrum analyzed by a Discrete Fourier Transform algorithm that determined the phase of the signal at different frequencies depending on the bandwidth of the transmitter (the bandwidth of the interferometer is essentially flat from 2 to 100 MHz). In order to verify that the measurement system is sufficiently accurate to study the rather weak dispersion exhibited by ultrasonic surface waves propagating on surface-treated metals, first a couple of 2024-T351 aluminum bars of 50.8 mm (width) 203 mm (length) 12.7 mm (height) were carefully polished with 2000-grade sandpaper parallel to the length of the specimen to optical fiber Fabry-Perot interferometer digital oscilloscope averaged signal computer tracking trigger pulser objective lens wedge transducer SAW Nd:YAG laser primary lens laser beam specimen translation table stepping motor driver Fig. 1. Experimental setup for laser-ultrasonic SAW velocity measurement.

3 A. Ruiz, P.B. Nagy / Ultrasonics 42 (2004) Normalized Phase [10 o /div] 4.7 MHz 4.5 MHz 4.3 MHz 4.1 MHz 3.9 MHz 3.7 MHz 3.5 MHz 3.3 MHz 3.1 MHz 2.9 MHz 2.7 MHz 2.5 MHz 2.3 MHz 2.1 MHz 1.9 MHz 1.7 MHz Propagation Distance [mm] Fig. 2. Example of the measured normalized phase versus propagation distance curves on an Almen 4 shot-peened specimen at 16 different frequencies. Velocity [mm/µs] minimize scattering by surface irregularities. Then, the specimens were heat treated at 345 C for 1 h. The objective of this initial heat treatment was to remove all near-surface material variations that might be present in the as-received cold-rolled bar stock material. After heat treatment, the samples were shot peened over a mm 2 square area at their center using 0.3- mm-diameter MI-110-H shots and 100% coverage. Three series of specimens were prepared using 4A, 6A, and 10A Almen intensities. The unpeened smooth sides of the specimens were used as reference (0A). As an example, Fig. 2 shows the normalized phase as a function of the propagation distance obtained from a 4A-intensity shot-peened specimen using a 3.5-MHz transducer. At each frequency, the phase velocity of the SAW can be calculated from the overall slope of the normalized phase versus distance curve and the tracking velocity chosen to minimize the observed total phase variation over the fairly long propagation distance [1]. In Fig. 2, the low-frequency slopes are slightly negative, which means that the low-frequency components propagated at a velocity slightly higher than the tracking velocity. In comparison, the high-frequency slopes are positive, which indicates that they propagated slower than the tracking velocity, i.e., the velocity decreased with increasing frequency over the bandwidth of the transmitter. Fig. 3 shows the results of our surface wave velocity measurements on three different aluminum specimens using three different transmitters of 2.25, 5, and 10 MHz nominal center frequencies. Interestingly, a perceivable dispersion effect is observed even on the smooth unpeened side. This apparent dispersion effect is mainly caused by the diffraction of the slightly divergent SAW as it propagates over the surface of the specimen [1]. On the other hand, the two shot-peened specimens exhibit much stronger dispersion which increases with peening intensity. This effect is caused partly by surface roughness induced scattering [13 16] and partly by the other previously mentioned material effects of shot peening. 3. Thermal relaxation smooth Almen 6 Almen 10 Fig. 3. Surface wave velocity measurements on shot-peened aluminum specimens. Two of the main effects produced by shot peening, namely residual stress and cold work, can significantly change during service as a result of thermo-mechanical relaxation, which can be simulated under laboratory conditions by appropriate heat treatments that cause thermally activated stress release and recrystallization. A series of subsequent heat treatments were conducted on the shot-peened specimens, starting from 150 C up to 325 C in steps of 25 C. The specimens were held for 1 h at each temperature in a protective nitrogen atmosphere. After each heat treatment, the velocity of the SAW was measured again for the same frequency range on the treated and smooth sides of the sample. The objective of measuring the velocity on the smooth side was to verify that the surface wave velocity does not change significantly as a result of heat treatment. Fig. 4 shows the results of the velocity measurements on an Almen 10 sample after a series of heat treatments. The SAW velocity on the unpeened smooth side does not change significantly during the heat treatments and the single data set shown in this figure is the average of the eight individual measurements. These results illustrate that heat treatment at 325 C reduces the dispersion by roughly 2/3 compared to the untreated condition. In addition, it can be observed that initially most of the velocity change occurs at high frequencies. As the heat treatment temperature increases, the dispersion becomes gradually weaker at high frequencies while continues a little longer at lower frequencies. Typically, the penetration depth of the excess cold work is only about one-third of the penetration depth of the residual stress. Therefore, the thermally induced relaxation and recrystallization processes occur much

4 668 A. Ruiz, P.B. Nagy / Ultrasonics 42 (2004) Velocity [mm/µs] smooth 325 C C C C 225 C C C 150 C 2.89 untreated 2.88 Fig. 4. Surface wave velocity measurements on a shot-peened (Almen 10) aluminum specimen after different levels of thermal relaxation. faster directly below the surface than at larger depths. As a result, initially the near-surface region is much more affected than the material at larger depths, which explains why the low-frequency components of the interrogating SAW are initially much less sensitive to annealing than the high-frequency components. As the temperature increases, the relative velocity difference between the peened and unpeened specimens further decreases at low frequencies, which indicates that the relaxation continues at larger depths. However, the velocity does not change any further at high frequencies, which indicates that very close to the surface the relaxation and recrystallization has been essentially completed. Fig. 5 shows the same results relative to the baseline velocity spectrum measured on the unpeened reference surface. To obtain these curves, for every heat treatment level, we subtracted the average phase velocity measured Relative Velocity Change [%] roughness C C C 250 C C C C C -1.8 untreated -2 Fig. 5. Relative SAW velocity on a 10A shot-peened aluminum specimen after increasing levels of thermal relaxation at eight different temperatures. Theoretical predictions for the surface roughness induced dispersion are also shown (dotted line, 7.5 lm rms roughness and 60 lm correlation length). over the smooth part of the specimen, which includes the effects of diffraction and grain scattering induced dispersion, from the data obtained over the peened part. By doing this, it is possible to better observe the combined effects of surface roughness, residual stress, and cold work on the SAW velocity. The relative velocity curves gradually approach the expected level of the surface roughness induced dispersion, which was calculated using the mean-field perturbation approach [13,14]. In the weak roughness approximation, this method can yield analytical solutions for both the attenuation and dispersion of the surface wave. In these calculations, the rms roughness and correlation length of the shot-peened surface was assumed to be 7.5 and 60 lm, respectively, which are consistent with white light interference microscopic measurements. These results clearly indicate that the thermal relaxation process significantly reduces the dispersion of ultrasonic surface waves on shot-peened aluminum surfaces. To quantify how much of this change is actually due to residual stress relaxation and how much to the recrystallization of the cold-worked microstructure, it would be necessary to destructively measure the residual stress and cold work by X-ray diffraction at each heat treatment temperature. At this point, such measurements have not been done yet; therefore, we do not know the exact behavior of residual stress and cold work during heat treatment. Based on independent acousto-elastic measurements and assuming that the maximum residual stress before relaxation is on the order of the yield strength of the material, we expect that the surface wave velocity in Al2024-T351 should increase by approximately 0.4%. However, the SAW measurements revealed an overall negative dispersion of about 1%, which clearly shows that the residual stress effect is not dominant in the dispersion of the surface wave. This conclusion is consistent with earlier findings for aluminum alloys [3,4]. Therefore, it appears that the effect of residual stress is significant, but not dominant in the dispersion of the SAW. 4. Conclusions Laser-interferometric surface wave velocity measurements allow us to observe the small, but very important changes caused by different heat treatments in shot-peened aluminum samples. The experimental results indicate that SAW dispersion in shot-peened aluminum specimens is similar to the theoretically predicted behavior for surface roughness scattering, but the magnitude of the dispersion is larger than expected based on the rms roughness and correlation length of the surface. The excess dispersion is due to the competing cold work and residual stress effects. The shallow sub-surface region exhibits a higher rate of relaxation and recrystalli-

5 A. Ruiz, P.B. Nagy / Ultrasonics 42 (2004) zation than deeper regions. Close to the surface, recrystallization starts at lower temperatures and occurs faster than at larger depths and usually the surface residual stress and cold work are essentially gone after the first heat treatment. This explains why the high-frequency components change faster than the low-frequency components which are affected by the material properties at larger depths. At higher annealing temperatures, the lowfrequency dispersion further decreases as a result of continuing partial relaxation and recrystallization deeply below the surface, but at high frequencies, the velocity starts to decrease again since the recrystallized grains start to grow in size. Despite the considerable complexity of the problem, the results indicate that ultrasonic evaluation of near-surface material variations in shot-peened metals is feasible. Although SAW dispersion is mainly sensitive to cold work effects and much less to residual stress effects, it still can be a very useful NDE tool since quantitative assessment of the level and distribution of cold work in surface-treated metals is of primary importance from the point of thermo-mechanical stability of the beneficial residual stresses. Acknowledgements This work was supported by the Air Force Research Laboratory, Metals, Ceramics, and NDE Division, under Contract no. F C The authors wish to thank to the Universidad Michoacana de San Nicolas De Hidalgo and CONACYT-MEXICO for their support of Alberto Ruiz Marines during his Ph.D. studies. The authors would also like to acknowledge valuable discussions with Dr. Mark P. Blodgett of AFRL/MLLP. References [1] A.M. Ruiz, P.B. Nagy, J. Acoust. Soc. Am. 112 (2002) 835. [2] F. Lakestani, J.F. Coste, R. Denis, NDT&E Int. 28 (1995) 171. [3] A.I. Lavrentyev, P.A. Stucky, W.A. Veronesi, in: Rev. Progr. Quant. NDE, vol. 19, AIP, Melville, 1999, pp [4] C. Glorieux, W. Gao, J. Appl. Phys. 88 (2000) [5] M. Duquennoy, M. Ouaftouh, M.L. Qian, F. Jenot, M. Ourak, NDT&E Int. 34 (2001) 355. [6] A.I. Lavrentyev, W.A. Veronesi, in: Rev. Progr. Quant. NDE, vol. 20, AIP, Melville, 2001, pp [7] C. Zhang, J.D. Achenbach, J. Acoust. Soc. Am. 88 (1990) [8] P.D. Warren, C. Pecorari, O.V. Kolosov, S.G. Roberts, G.A.D. Briggs, Nanotechnology 7 (1996) 295. [9] C. Pecorari, J. Acoust. Soc. Am. 100 (1996) [10] C. Pecorari, J. Acoust. Soc. Am. 103 (1998) [11] C. Pecorari, Ultrasonics 38 (2000) 754. [12] C. Pecorari, Wave Motion 33 (2001) 259. [13] A.G. Eguiluz, A.A. Maradudin, Phys. Rev. B 28 (1983) 728. [14] V.V. Kosachev, A.V. Shchegrov, Ann. Phys. 240 (1995) 225. [15] A.P. Mayer, M. Lehner, Waves Rand. Med. 4 (1993) 321. [16] V.V. Krylov, Z.A. Smirnova, Sov. Phys. Acoust. 36 (1990) 583.