Development of Observation Method for Tempered Martensite Microstructure Using Chemical Mechanical Polishing Technique*

Transcription

1 Materials Transactions, Vol. 46, No. 11 (2005) pp to 2448 #2005 The Japan Institute of Metals Development of Observation Method for Tempered Martensite Microstructure Using Chemical Mechanical Polishing Technique* Masao Hayakawa 1, Saburo Matsuoka 2, Yoshiyuki Furuya 1 and Yoshinori Ono 1 1 Materials Information Technology Station, National Institute for Materials Science (MITS/NIMS), Tsukuba , Japan 2 Faculty of Engineering, Kyushu University, Fukuoka , Japan A new observation method was developed for visualizing microstructure on a chemical mechanically polished (CMP) surface for tempered martensite of JIS-SCM440, a medium-carbon steel. The CMP and an electropolished (EP) surfaces were observed using an atomic force microscope (AFM) and a field emission type-scanning electron microscope (FE-SEM), respectively. AFM images, FE-secondary and backscattered electron images and electron backscattered patterns (ESP) were obtained for the CMP and EP surfaces. The AFM and FE-secondary electron images of the EP surface clearly visualized martensite blocks and cementite particles, since unevenness corresponding to blocks and cementite particles was created by electropolishing. On the other hand, the AFM images of the CMP surface revealed that the CMP process produced a very smooth surface with unevenness not exceeding 10 nm. The FE-backscattered electron images of the CMP surface only visualized the crystal misorientation of the martensite matrix microstructure, since the images are not influenced by surface unevenness. The CMP surface is more appropriate than the EP surface for ESP measurements, since the CMP surface is smoother than that of the EP surface. locks with a high-angle boundary exceeding 15 could be recognized by ESP mapping, but laths with a low-angle boundary below 3 could not. (Received June 27, 2005; Accepted September 7, 2005; Published November 15, 2005) Keywords: tempered martensite, medium-carbon steel, chemical mechanical polishing, electropolishing, atomic force microscopy, scanning electron microscopy, electron backscattered patterns 1. Introduction Tempered lath martensite has a multi-scale structure consisting of, in descending order of size, prior austenite () grains, packets, blocks, laths and precipitates. 1,2) Prior grain boundaries, packet boundaries and block boundaries are high-misorientation-angle boundaries, and lath boundaries are low-misorientation-angle boundaries. Medium- or high-content carbon steels have extremely small blocks. 2,3) It is difficult to discriminate structural components smaller than blocks using optical or scanning electron microscopy (SEM) on a chemical-etched surface. The authors showed that blocks and precipitates in tempered martensite microstructure can be successfully visualized by observing the electropolished (EP) surface of a medium-carbon low-alloy steel with an atomic force microscope (AFM). 4) An AFM with high vertical resolution proved able to discriminate minute level differences below 10 nm among the blocks and precipitates on the EP surface, formed by differences in the rate of electropolishing, which depends on the crystal orientation and phase on the surface. Moreover, the authors also successfully identified prior grain boundaries on the EP surface by superimposing an AFM image of the EP surface and an AFM image of the same area after it had been treated by prior grain boundary etching. 5 7) In this study, new methods for observing tempered martensite microstructure are reported, including those that use EP treatment, which were reported in our previous papers, 4 7) and those that use chemical mechanical polishing (CMP), a technique used to smooth the surfaces of silicon *This Paper was Originally Published in Japanese in J. Jpn. Inst. Met. 68 (2004) wafers and other semiconductor materials, 8) and is a commonly used method for smoothing silicon surfaces to the atomic level without introducing the strain of polishing to the surface. This CMP technique was applied to treat the surface of the tempered martensite microstructure of medium-carbon steel covered with cementite (Fe 3 C) precipitates. To examine the effectiveness of these treatments, the EP and CMP surfaces were observed under an AFM. Secondary electron images and backscattered electron images of the surfaces were obtained using a field-emission-type scanning electron microscope (FE-SEM), and electron backscattered patterns (ESP) were measured. The ESP mappings had been obtained for microstructures of low-carbon steels 9,10) and misorientation angle distributions have been under discussion in detail. However, for the tempered martensite of medium-carbon steel, it has been particularly hard to obtain the ESP mapping, since the structure is very fine and covered with cementite precipitates. 2. Experimental Procedure 2.1 Specimen The specimen was JIS-SCM440, a medium-carbon lowalloy steel (mass%: 0.40% C, 0.24% Si, 0.81% Mn, 1.03% Cr, and 0.16% Mo) with the same chemical compositions and subjected to the same thermal treatment conditions as those used in the previous paper. 4,5) The specimen used for thermal treatment was a bar 12 mm in diameter and 80 mm in length. The bar was austenized at 1153 K for 2.7 ks and then oil-quenched to form the lath martensite microstructure. To acquire a sufficient amount of cementite precipitation, the specimen was tempered at 723 K for 5.4 ks and cooled in water.

2 2444 M. Hayakawa, S. Matsuoka, Y. Furuya and Y. Ono (a) (b) Fig. 2 TEM image of the tempered lath martensite. 4) Fig. 1 Optical microscope images of the 3% nital- and picric acid-etched surfaces in (a) and (b), respectively. 4) Optical microscopic images of the surface etched with 3% nital and an aqueous solution of picric acid are shown in Figs. 1(a) and (b), respectively. 4) The contrast shown in (a), which was produced by nital etching, corresponds to the various boundaries. However, it was difficult to identify structural components that were smaller than blocks. In (b), only the prior grain boundaries were selectively corroded. The nominal grain size of a prior grain is 17 mm. A transmission electron microscopic (TEM) image is shown in Fig. 2, which shows lath martensite structure formation. 4) The tensile strength of the specimen was 1400 MPa, and the Vickers hardness Hv (load of 490 N) was Surface preparation Specimens for microscopic observations were prepared by excising 7-mm long, 5-mm wide and 2-mm thick rectangular plates from the bar such that the observation surfaces were along the bar axis. The observation surfaces were polished to mirror finish using diamond particles 1 mm in diameter and were then treated with the following methods Electropolishing (EP) The solvent used for electropolishing consisted of 8 vol% perchloric acid, 10 vol% butoxyethanol, 70 vol% ethanol and 12 vol% distilled water. EP was conducted at 273 K at a voltage of 40 V for 10 s Chemical mechanical polishing (CMP) Specimens were immersed in 500 cm 3 of an aqueous solution that contained 200 g of silica micro-powder in a container with an abrasive cloth attached to the bottom. The pressure between the specimen and the abrasive disk was set at about 10 kpa, and the abrasive disk was slowly rotated at 1 rps. The optimum duration was 1.8 ks. A shorter duration resulted in unevenness remaining on the surface. The smoothing of single-crystal materials such as semiconductors depends only on time, with the correct duration of processing creating the required smooth surfaces. However, the surfaces of martensitic steels corrode if polished for too long. After Treatments (1) or (2), the specimens were ultrasonically cleaned in ethanol. 2.3 Observation method Atomic force microscopy AFM has atomic-level vertical resolution and can measure heights quantitatively ) AFM observations were conducted in the atmosphere in tapping mode at 300 khz resonance frequency and a scanning rate of 0.3 Hz. The maximum scanning range of the scanner used was 90 mm lengthwise and 5 mm heightwise. The cantilever was made of silicon and was designed to have a spring constant of 20 N/ m, resonance frequency of 300 khz, and probe tip diameter of 5 10 nm FE-secondary electron and backscattered electron microscopy Secondary electron images were used to analyze the forms of unevenness on specimen surface, since contrast mainly results from the surface configuration. 14) On the other hand, backscattered electron images show contrast formed by chemical composition, crystal misorientation and surface unevenness of the specimen, and are mainly used to emphasize the differences in contrast attributable to differences in chemical composition or crystal misorientation. 14) FE-secondary electron and backscattered electron images were obtained at a working distance of 7 mm. The acceleration voltage was set low, at 5 kv, to obtain clear images of

3 Development of Observation Method for Tempered Martensite Microstructure Using Chemical Mechanical Polishing Technique 2445 the microstructure on the specimen surface by reducing the depth of electron penetration. 14) ESP measurement ESP identifies crystal orientation by analyzing the pattern of Kikuchi lines obtained from backscattered electrons. 15) Thus, the observed surface must be smooth and free of polishing-induced strain. In ESP measurements, observation surfaces are usually inclined 70 to the incident angle of electrons to increase the yield of backscattered electrons at the detectors. In this experiment, an FE-type electron gun was used. The measurement area was about 20 mm 20 mm, and the measurement intervals were 0.04 mm. A 80nm 40nm 0 3. Results and Discussion 3.1 AFM observation of microstructure An AFM image of the EP surface and its cross-sectional profile (height indication: 50 nm) is shown in Fig. 3. The black and white contrast in the image shows the level differences on the surface attributable to differences in crystal orientation. Therefore, each strip of uniform brightness corresponds to a martensite block that has a high-misorientation-angle boundary. Cementite particles, which project out from the martensite matrix, are shown as white spots. High-misorientation-angle boundaries act as preferential sites for precipitation, and are thus densely covered with cementite particles. As shown in the cross-sectional profile, Vertical Distance,Dv/nm A Horizontal Distance,DH/ µ m Fig. 4 AFM image of the chemical mechanically polished surface and the cross-sectional profile through line A in the AFM image. Vertical Distance,Dv/nm 80nm 40nm 0 A A Horizontal Distance,DH/ µ m Fig. 3 AFM image of the electropolished surface and the cross-sectional profile through line A in the AFM image. the level difference on the surface was a maximum of only 50 nm. Thus, an AFM, which has high vertical resolution, is ideal for identifying cementite particles and high-angle block boundaries on the EP surface, which have small level differences. An AFM image of the CMP surface and its cross-sectional profile (height indication: 50 nm) are shown in Fig. 4. The height indication is the same as that in Fig. 3. The CMP treatment produced a very smooth surface with level differences smaller than 10 nm. Polishing residues are shown as white particles. Except for these, the largest level difference was 1/5 smaller than that in the EP surface. The contrast in the AFM image shows only the unevenness attributable to imperfect CMP treatment, so the unevenness does not reflect the martensite microstructure. This highly smooth CMP surface was very useful for obtaining the FE-backscattered electron image observations and ESP measurements described in the following section. 3.2 FE-SEM observation of microstructure An FE-secondary electron and a backscattered electron image of the EP surface are shown in Figs. 5 and 6, respectively. The secondary electron image shows stripes of uniform brightness corresponding to blocks. Cementite particles projecting from the martensite matrix are shown as white spots. Dense precipitations of cementite particles can be observed on boundaries. Thus, the FE-secondary electron image provided similar information to that of the AFM image. In other words, structural analysis that involves

4 2446 M. Hayakawa, S. Matsuoka, Y. Furuya and Y. Ono Fig. 5 FE-secondary electron image of the electropolished surface. Fig. 7 FE-secondary electron image of the chemical mechanically polished surface. as the AFM image (Fig. 4). Indeed, secondary electron images of average contrast were very similar to the AFM images. However, when the contrast was boosted to maximum (Fig. 7), the secondary electron image became very similar to its corresponding backscattered electron image. This was likely because the CMP surface was very smooth and the secondary electron image emphasized only the backscattered electron components, which depend on crystal misorientation. Fig. 6 FE-backscattered electron image of the electropolished surface. observation of unevenness on the EP surface can be similarly performed using an FE-secondary electron image or an AFM image. On the other hand, the backscattered electron images of the EP surface show both the contrast types characteristic of crystal misorientation and surface unevenness. Thus, the two sets of information cannot be discriminated in the backscattered electron image of the EP surface, and so almost no beneficial information was obtained. An FE-secondary electron image and backscattered electron image of the CMP surface are shown in Figs. 7 and 8, respectively. Since the CMP surface is smooth, their backscattered electron images were expected to clearly show the crystal misorientation. As expected, the image in Fig. 8 was very different from the backscattered electron image of the EP surface (Fig. 6), which showed unevenness of the surface, and was rather similar to the TEM image in Fig. 2, which showed lath structures. Laths can be observed in backscattered electron images of the CMP surface. CMP treatment is easier than preparing thin film specimens for TEM, and CMP specimens are easier to handle than TEM specimens. Moreover, the range that can be observed is larger in CMP specimens than in TEM specimens. On the other hand, if a secondary electron image reflects only the morphology of the surface, it should look the same 3.3 ESP measurement ESP mappings of the EP and CMP surfaces are shown in Figs. 9 and 10, respectively. Crystal orientations are shown as colored regions. The high smoothness of the CMP surface considerably assisted the ESP measurement, and crystal orientation could be analyzed over larger regions of the CMP surface than in the EP surface. This was likely because the rate of incident electrons blocked by cementite particles and level differences among blocks was smaller in the CMP surface than in the EP surface, since the surfaces were inclined 70. In the mappings, regions that were uniform in color and have high-angle boundaries not smaller than 15 in misorientation were blocks. Patterns probably attributable to cementite particles were observed in many blocks. Thus, laths, which have low-misorientation-angle boundaries not larger than 3, could not be identified in the blocks. For future ESP measurements of tempered martensite structures of medium-carbon steel, smoother CMP surfaces should be prepared to enable expansion of their measurable range and identification of laths. Finally, combining EP and CMP surfaces provides several advantages. locks and cementite particles can be discriminated in FE-secondary electron images as well as in AFM images of the EP surfaces. Observation of smooth CMP surfaces enables clear discrimination of blocks by orientation analysis using ESP. FE-backscattered electron images of CMP surfaces can visualize laths. These observation methods, surface treatment methods, and targets of observation are summarized in Table 1. Selection of appropriate combinations will enable multi-scale structures of tempered martensite to be analyzed.

5 Development of Observation Method for Tempered Martensite Microstructure Using Chemical Mechanical Polishing Technique 2447 (a) (b) Fig. 9 ESP mapping of the electropolished surface as the same area of the AFM image in Fig. 3. (1 step = 0.04 mm) 001 Fig. 8 Low- and high-magnification FE-backscattered electron images of the chemical mechanically polished surface in (a) and (b), respectively Conclusions AFM images, FE-secondary and backscattered electron images, and ESP mappings were obtained for the EP and CMP surfaces of the tempered martensite microstructure of JIS-SCM440 medium-carbon low-alloy steel. Our resultant conclusions are: (1) AFM images, which reflect surface unevenness, and FE-secondary electron images are appropriate for identifying cementite particles and blocks on the EP surface. However, backscattered electron images, which reflect both the crystal misorientation states and surface unevenness, provide no beneficial information on the EP surface. (2) The CMP treatment produces a smooth surface with no unevenness greater than 10 nm. Microstructures could Fig. 10 ESP mapping of the chemical mechanically polished surface. (1 step = 0.04 mm) not be identified from level differences on AFM images of the CMP surface. (3) ackscattered electron images of the smooth CMP surface clearly show crystal misorientation of the martensite matrix, free from the effects of unevenness of the surface attributable to cementite particles. The images are similar to TEM images of the lath structure. (4) ESP mapping is possible over larger areas of the CMP surface than of the EP surface. locks that have highangle boundaries not smaller than 15 in misorientation Table 1 Summary of combinations with surface-polishing techniques and microscopes for the tempered martensite microstructure. Surface EP : CMP Uneven (Non-flat) : Even (Flat) Secondary electron (SEM) : Topography : lock, Carbide : X ackscattered electron (SEM) : Crystal-misorientation : X : Lath ESP : Orientation : X : lock AFM : Topography : lock, Carbide : X (X; Inappropriate)

6 2448 M. Hayakawa, S. Matsuoka, Y. Furuya and Y. Ono could be identified. However, laths with low-misorientation-angle boundaries up to 3 in the blocks could not be identified. Acknowledgements The authors would like to thank Mr. Isao Suzuki and Mr. Takashi Konosu of Yokohama Science International, Ltd. for providing valuable advice. REFERENCES 1) A. R. Marder and G. Krauss: Trans. ASM 60 (1967) ) T. Maki, K. Tsuzaki and I. Tamura: Trans. ISIJ 20 (1980) ) T. Maki and I. Tamura: Tetsu-to-Hagane 67 (1981) ) M. Hayakawa, T. Hara, S. Matsuoka and K. Tsuzaki: J. Jpn. Inst. Met. 64 (2000) ) M. Hayakawa, S. Matsuoka and K. Tsuzaki: Mater. Trans. 43 (2002) ) M. Hayakawa, S. Terasaki, T. Hara, K. Tsuzaki and S. Matsuoka: J. Jpn. Inst. Met. 66 (2002) ) M. Hayakawa, S. Matsuoka and Y. Furuya: J. Jpn. Inst. Met. 67 (2003) ) R. Schmolke, R. Deters, P. Thieme, R. Pech, H. Schwenk and G. Diakourakis: J. Electrochem. Soc. 149 (2002) G257 G265. 9) A. F. Gourgues, H. M. Flower and T. C. Lindley: Mater. Sci. Technol. 16 (2000) ) H. Nakashima: Tetsuo-to-Hagane 90 (2004) ) G. innig, C. F. Quate and Ch. Gerber: Phy. Rev. Lett. 56 (1986) ) T. R. Albrecht and C. F. Quate: J. Appl. Phys. 62 (1987) ) S. Alexander, L. Hellemans, O. Marti, J. Schneir, V. Elings and P. K. Hansma: J. Appl. Phys. 65 (1989) ) Y. Tanabe et al.: asic and Application of Scanning Electron Microscopy, ed. by The Kanto ranch of The Japanese Society of Microscopy (Kyoritsu Shuppan Co. Ltd., Tokyo, 1983) pp ) J. A. Venables and C. J. Harland: Philos. Mag. 27 (1973)