外文原文-使用扫描霍尔显微镜观察由于应力或疲劳过程在奥氏体不锈钢的磁性图像的变化

Journal of Nondestructive Evaluation, Vol. 20, No. 3, September 2001 (q 2001) Observation of Changes in Magnetic Images Due to a Strain or Fatigue Progress in Austenite Stainless Steels using a Scanning Hall-Sensor Microscope A. Oota,1,* K. Miyake,1D. Sugiyama,2and H. Aoki2 Received September 14, 1999; Revised September 15, 2001 Using a scanning Hall-sensor microscope with an active Hall element of area 50mm 3 50mm, we measured two-dimensional magnetic images of spontaneous magnetization on a surface of a number of 304 stainless plates in a paramagnetic austenite phase. Stainless plates with a yield point of 31 kg/mm2were subjected to a strain or pulling-fatigue process at room temperature. In a strain state, most plates show the same progressive change in magnetic images resulting from a strain-induced martensite transformation, but others produce additional and/or different change although the origin remains unsolved. A pulling-fatigue process at a stress amplitude of 28 kg/mm2also produces some differences in the images from the strain process, at the early stage of fatigue progress. The study should provide a new horizon to elucidate the complex destructive-progress in strain or fatigue process of304 stainlesssteels that areused asstructural materials inelectric powerstations, chemical plants and medical equipment. KEYWORDS:ScanningHall-sensormicroscope;magneticimages;spontaneousmagnetization;non-destructive evaluation; stainless steel; austenite phase; strain-induced martensite transformation; fatigue process. tum interference devices (SQUID)(2,3)and micro-HallI. INTRODUCTION sensors(5,6)provides new and promising techniques for NDE of magnetic materials, because of many advantagesInvoking of social securities and environmental over conventional techniques such as eddy current, ultra-problems demands higher levels of reliability and stabil- sound and x-ray imaging, etc. Magnetic imaging tech-ity for structural materials such as iron and stainless nique using a SQUID has enabled a non-destructivesteels supporting infra-structuring in industries. Much detection of Lu ders bands in mild steels subjected toeffort has been made to develop non-destructive evalua- a strain process.(3)A scanning Hall-sensor microscopetion(NDE) techniquesto investigatethe destructivepro- (SHM) with an active Hall element of area 50mm 3 gress under fatigue and/or strain processes in struc- 50mm has also enabled a magnetic detection of small tural materials.(14) cracks (,10 mm long and ,0.1 mm wide) in mild steels Recent progress in magnetometers utilizing highly- caused by a fatigue process,(6)so that SHM can be used sensitivemagneticsensorssuchassuperconductingquan- as a simple, economic and conventional tool for NDE of magnetic materials. We present here the magnetic images of spontaneous magnetization on a surface of 304 stain- * Corresponding author. E-mail: ootaeee.tut.ac.jp less steels subjected to strain or fatigue process at room 1Toyohashi University of Technology, Tempaku-cho, Toyohashi, Aichi temperature observed using this microscope, and the 441-8580, Japan. influence of the destructive progress on the images of 2Technical Research and Development Laboratories, Topy Industries Ltd., Akemi-cho, Toyohashi 441-8510, Japan.304 stainless steels. 87 0195-9298/01/0900-0087/0 q 2001 Plenum Publishing Corporation 88Oota, Miyake, Sugiyama, and Aoki 2. EXPERIMENTALprogressive way. Measurements were repeated after observing the samples to higher levels and reducing the load to zero. In addition, a number of the plates A were 2.1 Stainless Plates subjected to a pulling-fatigue test using a conventional hydraulic-servo fatigue testing machine at a stress ampli-Commercial 304 stainless steels with a yield point of 31 kg/mm2and a tensile strength of 68 kg/mm2weretude of 28 kg/mm2and a frequency of 29.2 Hz, to eluci- date a destructive progress under a fatigue process.machinedintoa3.8mmthickplate.Theyhavethefollow- ing chemical composition except for iron: Cr, 18.2 wt%;Magnetic measurements were made as a parameter of stress cycles in a progressive manner by stopping theNi, 8.3 wt%; C, 0.05 wt%; Si, 0.43 wt%; Mn, 0.89 wt%; P, 0.03 wt%; S, 0.05 wt%. As can be seen from Fig.machine and taking away the plate, until micro-cracks appeared in the sample.1, two types of stainless plates A and B with different geometries were used for the study. There are notches on both sides in the former, but not in the latter.2.2 Measurements Toinduceaplasticdeformationatroomtemperature, Magnetic images of spontaneous magnetization anumberofstainlessplatesoftypesAandBwereuniaxi- were measured on a surface of stainless plates under zero ally strained using a conventional Amsler-type testing external fields using SHM with an active area 50mm 3 machine at a strain rate of approximately 0.001 s21along 50 mm, of which the details have been published else- the length of the sample. Special attention was paid to where.(6)The SHM was equipped with a micro-Hall sen- the environment around the machine so as not to magne- sor on a movable x-y stage using a stepping motor and tize the sample during the test. After reducing the load a sample holder facing the sensor. The Hall sensor (ARE- to zero, the plates in a strain state were subjected to POC: HHP-VP) consisting of epitaxial GaAs films was magnetic measurements as a parameter of strain in a used to measure magnetic fields normal to the sample surface. The sensitivity was 3.39 mV/G at a Hall current of 10mA. Hall voltages were measured using a nano- voltmeter. The Hall sensor on the x-y stage was scanned at a finite step (minimum step being 0.05mm) in a two- dimensional grid pattern, at a fixed distance (,0.5mm) from the sample surface. The two-dimensional distribu- tions of magnetic fields from the sample were obtained through a conversion procedure from voltages to mag- netic fields by a personal computer. Special attention was paid to a drift of null voltages causing a zero-point shift of magnetic fields during the whole measuring run. The scannedareasonasurfaceofstainlessplatesareindicated in Fig. 1. 3. RESULTS AND DISCUSSION A large number of 304 stainless plates A and B were subjected to a strain or fatigue process at room temperature, andthe magnetic images dueto spontaneous magnetizationweremeasuredonthesurfaceofthesample using SHM. In this section, we present the magnetic images ofthe plates A andB subjected toa strain process, and then move onto the results for the plates A subjected to a pulling-fatigue process. Fig. 1. The geometry and dimensions in mm for 304 stainless plates 3.1 Change in Magnetic Images for Stainless Plateswith (a) type A having notches at both ends and (b) type B without notches. Note that a load line for application of a strain or a pulling A and B Under a Strain Progress fatigue is directed to the horizontal direction in the figures. Also shown Most of the plates A exhibited similar changes in are scanned areas on a sample surface using a scanning Hall-sensor microscope.the magnetic images under a strain progress. As can be Changes in Magnetic Images Due to a Strain or Fatigue Progress in Austenite Stainless Steels89 (a) (b) (c) Fig. 2. Magnetic images of spontaneous magnetization on a surface of stainless plate A subjected to a strain progress: (a) unstrained; (b) 2.5%- strained; (c) 7.1%-strained; (d) 11.0%-strained. Note that a load line for the strain is directed parallel to the y-axis. Also shown in the frame is the gray scale corresponding to approximate values of the magnetic field in Gauss to guide eyes for a reader. 90Oota, Miyake, Sugiyama, and Aoki (d) Fig. 2Continued. seen from Fig. 2, the images of unstrained plate show a clear boundary line with the shape of the notches at both ends. An application of plastic strain causes significant change in the images after loading and converts them to the distributions with two peaks, positioning at around the notches along a direction at a finite angle from the load line along which the sample was strained. In spite ofnomovementsofpeakpositionsintheimages,adegree of magnitude for the change in the images becomes more significant with an increase in the plastic strain. In an attempt to interpret the change in the images in a strain state, it is worth mentioning that 304 stainless steels undergo a transformation from the paramagnetic austenite phase to the ferromagnetic martensite phase under a plastic strain at room temperature,(710)which is called the strain-induced martensite transformation. To address thistransformation, the value ofVickers hardness Hv at a load level of 5 kg was measured from place to place on the surface of the plate. The result shows a strong correlation between the distribution of Hv and the magnetic images for the strained plate A shown in Fig. 2. Incidentally, the value of Hv for the 11%-strained plate showsamaximumof320360aroundtheplacesatwhich two peaks appear in the images as shown in Fig. 2(d). The same correlation also stands between the distribution of magnetic permeability and the magnetic images on a surface of the strained plate A. Furthermore, as shown in Fig. 3, there are many slip lines in microstructures near the notches of the strained plate A, in comparison with the unstrained sample. All these facts suggest that the change in the magnetic images shown in Fig. 2 can be ascribed to the strain-induced martensite transformation. Fig. 3. Microstructures of stainless plate A near a notch when it was: In contrast to the result mentioned above, several (a) unstrained; (b) 11.0%-strained. Note that the sample is the same as that shown in Fig. 2.(but not exceptional number) of the plates A showed Changes in Magnetic Images Due to a Strain or Fatigue Progress in Austenite Stainless Steels91 different and anomalous changes in the images under atrigger for circular patterns) in a strain progress, and revealed the complexity in the destructive progress of astrain progress, although the imagesin an unstrained state showed no notable difference from Fig. 2(a). Such exam-strain process, accompanied by the strain-induced mar- tensite transformation.ples are shown for the 5.5%-strained plate in Fig. 4(a) and also for the 0.7%-strained plate in Fig. 4(b). AsIn comparison, it is worth mentioning the results for the plates B, where there is relatively more evencan be seen from the figures, the latter shows additional circular patterns in the images between both notches,distribution of stress under a plastic strain in both the width and length directions because of the sampletogether with the signals ascribed to the martensite trans- formationmentionedabove.Ontheotherhand,theformergeometry. Figure 5 shows change in the images during a strain progress. Note that in order to check the repro-has further anomalies in the images that are made of a few circular patterns, masking the signals resulting fromducibility, we took the magnetic images from several plates B due to a strain process and obtained nearly thethe martensite transformation around the notches. To examine the origin of the anomalies, the plates A withsame results from them. As can be seen from Fig. 5, an application of plastic strain to the plates B producestheimagesshowninFig.4weresubjectedtoconventional metallurgical studies by an X-ray diffraction analysis andthe complexity in the images that are made of many circularpatterns,wherethenumberofpatternsincreasesa scanning electron microscope with energy dispersive X-ray spectroscopy. However, the results show neitherwith a strain progress. Although there are many slip lines in microstructures ascribed to the martensite trans-any difference from the plate A shown in Fig. 2 nor any evidence of uneven distribution of composition in theformation in some places on the strained plate B, the influence of this transformation on the images stillplate.Alltrialsresultedinfailuresinexplainingtheorigin of the anomalies in the images (i.e., the kernel and/or theremains unsolved. Fig. 4. Different magnetic images of spontaneous magnetization on a surface of stainless plate A subjected to a strain process: (a) 5.5%-strained; (b) 0.7%-strained. Note that there is no notable difference in the images for unstrained state from the result shown in Fig. 2(a). Also shown in the frame is the gray scale corresponding to approximate values of the magnetic field in Gauss to guide eyes for a reader. 92Oota, Miyake, Sugiyama, and Aoki (c)(a) Fig. 5Continued. of 28 kg/mm2and a frequency of 29.2 Hz. Note that the sample showed small cracks (,5mm long and ,0.1mm wide) near the top of the notch (i.e., x,12mm and y,12mm in Fig. 6(d) after stress cycle N of 61,300. The result in the virgin state exhibits a clear boundary line of the sample including shapes of notches in the middle of the plate. A fatigue progress causes changes in the images, although the stress amplitude is less than the yield point (530 kg/mm2). In comparison with the results for a strain process (see Fig. 2), the images have a single peak at near the top of the notch. An increase in N enhances the peak amplitude in the images (i.e., the stray field strength) and finally leads to small cracks (,5mm long and ,0.1mm wide) near the notch for N 5 61,300. Note that the place at which the cracks occur correspondstotheplaceatwhichtheimageshaveasingle (b) peak near the notch. This fact suggests the possibility to Fig. 5. Magnetic images of spontaneous magnetization on a surface of foresee the places of strain localization at which micro- stainless plate B subjected to a strained progress: (a) unstrained; (b) cracks occur in the plate ina fatigue progress. In addition, 0.8%-strained; (c) 13.8%-strained. Note that a load line for the strain there is a strong meandering line at around x,30 mm is directed parallel to the y-axis. Also shown in the frame is the gray alongthey-axisintheimagesattheinitialstageoffatigue scale corresponding to approximate values of the magnetic field in Gauss to guide eyes for a reader. progress (i.e., low cycle of N 5 20,000). This line never smears out even when small cracks occur at N 5 61,300. Note that such a line does not appear in the images under astrainprogressmentionedinsection3.1.Sinceapulling- 3.2. Change in Magnetic Images for Stainless Plates fatigue test is made with a load level below a yield point, A Under a Pulling-Fatigue Progress thefatigue progressshouldaffect microstructureswithout macroscopic deformation and produce localized micro-Figure 6 shows progressive change in magnetic images caused by a fatigue progress in stainless platesstrains.Thismaycausesomedifferenceinthemechanism of martensite transformation so that it leads to differentA, subjected to a pulling fatigue test at a stress amplitude Changes in Magnetic Images Due to a Strain or Fatigue Progress in Austenite Stainless Steels93 (a) (b) (c) Fig. 6. Magnetic images of spontaneous magnetization on a surface of stainless plate A subjected to a pulling-fatigue test at: (a) virgin state; (b) 20,000 cycles; (c) 60,000 cycles, (d) 61,300 cycles. Note that a load line for a pulling fatigue test is directed parallel to the y-axis. Also shown in the frame is the gray scale corresponding to approximate values of the magnetic field in Gauss to guide eyes for a reader. 94Oota, Miyake, Sugiyama, and Aoki (d) Fig. 6Continued. changesin theimages inthe destructiveprogress betweenACKNOWLEDGMENT a strain process and a fatigue process. WewouldliketothankDr.K.Kawanoforhistechni- cal assistance in measurements and also for valuable dis- cussion. 4. SUMMARY REFERENCES We succeeded in visualizing complex changes in 1. D. E. Bray and D. McBride (eds), Non-destr