Solid- and Liquid-Solid Reactions in Aluminum-Coated Titanium Substrate Fabricated by Using Explosive Energy

Transcription

1 Materials Transactions, Vol. 52, No. 12 (2011) pp to 2183 #2011 The Japan Institute of Metals Solid- and Liquid-Solid Reactions in Aluminum-Coated Titanium Substrate Fabricated by Using Explosive Energy Yasuhiro Morizono, Tomofumi Fukuyama*, Mitsuhiro Matsuda and Sadahiro Tsurekawa Department of Materials Science and Engineering, Graduate School of Science and Technology, Kumamoto University, Kumamoto , Japan A Ti plate coated with an Al foil with a thickness of 100 mm was prepared by using explosive energy. The bonding was momentarily achieved without breaking the Al foil. This Al-coated Ti plate was heated between 898 and 1273 K in the air to investigate interfacial reactions in solid Al/solid Ti and liquid Al/solid Ti states from the viewpoint of diffusion coating for Ti materials. At 898 K, a continuous Al 3 Ti layer was formed along the interface. Although the Al foil was almost replaced by Al 3 Ti in a holding time of more than 32.4 ks, many defects such as void and crack were observed in the Al 3 Ti layer. It is noteworthy that the thickness of the Al 3 Ti layer increases in proportion to the holding time in the early stage, and then, to the square root of the holding time. On the other hand, the specimen, which was heated at a temperature above the melting point of Al, showed a porous structure of Al 3 Ti via an Al + Al 3 Ti duplex state. The holding time to produce the porous structure became short with increasing heating temperature. It is expected that the Al + Al 3 Ti duplex structure is more beneficial as a surface coating layer formed by the interfacial reaction. To replace the Al foil region by such a duplex structure, continuous heat treatment was performed between 1023 and 1073 K. There were no defects in the Al + Al 3 Ti duplex region on the Ti plate, and its maximum Vickers hardness was approximately 3 times higher than that of the Ti plate. [doi: /matertrans.m ] (Received July 27, 2011; Accepted October 3, 2011; Published November 25, 2011) Keywords: titanium, aluminum foil, explosive coating, heat treatment, Al 3 Ti 1. Introduction Explosive welding is basically a solid state welding process. One of its positive aspects is the capability for the formation of substantial bond between two materials, which cannot be joined by conventional methods like diffusion bonding. Various clad materials have been fabricated by explosive welding and used in the industrial fields such as chemical and electric power plants. In the clad materials, excellent mechanical and chemical properties of a cladding metal are added to a base metal. This also serves to save raw materials and their product cost. When a metal foil is used as the cladding metal, the bonding process can be regarded as a surface coating. However, explosive welding is difficult to apply to the metal foil, because there is a risk that the foil is broken by the detonation and the collision in the process. The authors previously attempted the bonding of a Ag-Cu filler alloy sheet to a steel plate by explosive welding. 1) The alloy sheet was bonded to a supporting plate of copper by using organic glue, and this combined plate was used as the cladding metal. The bonding of the alloy sheet to the steel plate was momentarily achieved by explosive energy and, simultaneously, the supporting plate separated from the alloy sheet. We referred to this process as explosive coating. Recently, titanium aluminides, which are typified by Al 3 Ti, TiAl and Ti 3 Al, are expected to be useable as hightemperature structural materials in the aerospace and automobile industries. However, their practical uses are limited owing to poor ductility at room temperature, low fracture toughness and poor workability. To overcome these disadvantages, the alloy design and the microstructure control have been studied extensively in relation to the processing techniques used in industrial applications. In addition, *Graduate Student, Kumamoto University. Present address: Nippon Yakin Kogyo Co., Ltd., Kawasaki , Japan titanium aluminides have a possibility to use as a surface coating material because of their excellent oxidation resistance. This is enhanced with increasing Al content in titanium aluminides, and it has been reported that Al 3 Ti shows relatively excellent oxidation resistance at 1473 K in the air. 2) The simplest way to form the titanium aluminide layer on metal substrates is the application of reactive diffusion between Al and Ti. In the present study, a Ti plate was coated with an Al foil with a thickness of 100 mm by using explosive energy, and then heat-treated in the range from 898 to 1273 K in the air. The solid- and liquid-solid reactions between finite Al and semi-infinite Ti were investigated from the viewpoint of diffusion coating for Ti materials. 2. Experimental Procedure Commercially pure Ti (TP340H) was used as a substrate. This was machined into a plate with dimensions of 90 mm 40 mm 5 mm and then annealed at 973 K for 5.4 ks in a vacuum of less than 3 mpa. This bonding surface was finished with #1200 emery paper. The thickness of an Al foil whose purity is over 99 mass% was 100 mm. The foil was also cut into a rectangular shape with dimensions of 90 mm 40 mm. These materials were degreased in acetone using an ultrasonic cleaner and dried with hot air. The coating operation was carried out using an experimental apparatus shown in Fig. 1(a). The Al foil was combined with a supporting plate of steel with a thickness of 2 mm to play a role as a flyer plate. These bonding were performed by using organic glue. This combined plate was fixed parallel to the Ti plate. The distance between the Al foil and the Ti plate was 5 mm. The powdery explosive consisting mainly of ammonium nitrate was set on the combined plate. The amount of the explosive whose detonation velocity is about 2400 ms 1 was 43.2 g. The flyer plate velocity during

2 Solid- and Liquid-Solid Reactions in Aluminum-Coated Titanium Substrate Fabricated by Using Explosive Energy 2179 Fig. 1 (a) Schematic illustration of experimental assembly for coating. (b) Stacking sequence of diffusion-bonded joint. coating was estimated to be approximately 290 ms 1. 3) The bonding of the Al foil to the Ti plate was momentarily achieved and, simultaneously, the supporting plate separated from the Al foil. The Ti plate coated with the Al foil was cut into a square shape with dimensions of 5 mm 5 mm. The specimens were heated between 898 and 1273 K for 0.6 to 90 ks in the air to investigate the interfacial reaction between finite Al and semi-infinite Ti. After a holding step, they were cooled in the furnace to room temperature. To verify the experimental results for the Al-coated specimen, diffusion bonding of Al to Ti was carried out at 898 K for 3.6 ks in a vacuum of less than 3 mpa. Figure 1(b) shows the stacking sequence of a diffusion-bonded specimen. The Al sheet with a thickness of 800 mm was located between two Ti cubes, which were 5 mm on a side. The obtained joints were also heated at 898 K for 3.6 to ks in an evacuated silica tube. The specimens were cut perpendicular to the interface to reveal the microstructures. The cross sections were ground with emery papers and then finished with alumina powders of 50 nm in particle size. They were examined by using an optical microscope and a scanning electron microscope (SEM) equipped with an energy dispersive X-ray spectroscope (EDX). A scanning ion microscope (SIM) in a focused ion beam (FIB) system and an electron probe X-ray micro analyzer (EPMA) were also used in some cases. X-ray diffraction (XRD) and a Vickers hardness test were carried out for the Al-coated specimens. In the hardness test, a load of 0.49 N was applied at room temperature for 15 s. 3. Results and Discussion 3.1 Solid state reaction in Al-coated Ti plate Figure 2(a) shows an external view of the Ti plate coated with the Al foil with a thickness of 100 mm. The sound specimen was obtained without breaking the Al foil. The Al Fig. 2 (a) External view of as-coated specimen. (b) Optical micrograph of the collision interface between Al foil and Ti plate. (c) SIM image of the collision interface finished with FIB milling. foil was attached firmly to the Ti plate and there were no defects like void in the collision interface, as shown in Fig. 2(b). Figure 2(c) shows an SIM image of the cross section in the as-coated specimen after FIB milling. Two kinds of reaction regions, indicated by the single and double arrows, were observed at the collision interface. The region indicated by the single arrow was composed of fine grains, and another region had a different feature from the Al and Ti matrices. These are considered to be the trace of melting and subsequently rapid solidification at the contact surfaces of Al and Ti. 4,5) Such a microstructure would contribute to substantial bond at the interface. Figure 3 shows a hardness distribution in the vicinity of the collision interface in the Al-coated Ti plate. The dashed line in the figure is the position of the interface. The hardness of the Al foil and the Ti plate increased slightly by explosive coating, although the marked increase of the hardness due to high-speed collision was not recognized at the interface. To form titanium aluminides on the Ti plate, the as-coated specimen was heated at a temperature of less than melting point of Al (933 K). Figure 4(a) shows an optical micrograph of the interface in the specimen heated at 898 K for 14.4 ks in the air. A continuous reaction layer was observed between Al and Ti. The reaction layer was identified as Al 3 Ti by using SEM-EDX. XRD patterns of the surface on the Al side after heating at 898 K are presented in Fig. 5. Diffraction peaks of Al 3 Ti became clear with increasing holding time. The Al foil was almost replaced by Al 3 Ti in a holding time of 32.4 ks.

3 2180 Y. Morizono, T. Fukuyama, M. Matsuda and S. Tsurekawa However, there were many voids and cracks in the Al 3 Ti layer, as shown in Fig. 4(b). Although the Al-coated Ti plate was heated at 898 K for up to 90 ks, TiAl and Ti 3 Al compounds in the Al-Ti binary system were not observed between Al 3 Ti and Ti. Such a situation has been also reported in several articles. 6 9) This Fig. 3 Vickers hardness in the vicinity of the collision interface in Ti plate coated with Al foil. would be attributed to the lowest free energy of formation of Al 3 Ti in the Al-Ti binary system. 10) The reactive diffusion in the Al/Ti couple was studied in the temperature range between 789 and 915 K by van Loo and Rieck. 6) They have reported that the growth of the Al 3 Ti layer was roughly proportional to the holding time. The similar tendency has been also reported by Enjyo et al. 7) On the other hand, there are reports that the growth of Al 3 Ti obeys the parabolic time dependence for the Al/Ti couple. 8,9,11) Therefore, the agreement on the layer growth of Al 3 Ti formed between Al and Ti does not reach even now. Figure 6 shows the thickness of the Al 3 Ti layer in the Alcoated Ti plate after the heat treatment at 898 K for various periods. The dashed line denotes the results reported by van Loo and Rieck. 6) In the early stage, the growth of the Al 3 Ti layer accorded with the dashed line. However, a significant increase of the layer thickness was observed in the specimen heated at 898 K for 14.4 ks. To ascertain the truth of such a behavior, the diffusion-bonded Al/Ti couple was also prepared. The total holding time of the bonding and heat treatments is shown in the horizontal axis in Fig. 6. The thickness of the Al 3 Ti layer in the diffusion-bonded specimen also increased in proportion to the holding time in the early stage. After heating at 898 K for 90 ks, however, the Fig. 4 (a) Optical micrograph of the interface in Al-coated Ti plate after heat treatment at 898 K for 14.4 ks. (b) Optical micrograph of the surface on the Al side of Al-coated Ti plate after heat treatment at 898 K for 57.6 ks. Fig. 5 XRD patterns of the surface on the Al side of Al-coated Ti plates after heat treatments at 898 K for (a) 3.6 ks, (b) 7.2 ks, (c) 14.4 ks, (d) 32.4 ks and (e) 57.6 ks. Fig. 6 Relationship between holding time and thickness of Al 3 Ti layer formed between Al and Ti at 898 K. The Al/Ti couples were prepared by explosive coating and diffusion bonding. Dashed line denotes the results reported by van Loo et al. 6)

4 Solid- and Liquid-Solid Reactions in Aluminum-Coated Titanium Substrate Fabricated by Using Explosive Energy 2181 Fig. 7 Relationship between holding time and thickness of Al 3 Ti layer plotted on the basis of turning point (t 0, d 0 ) in Fig. 6. significant increase of the layer thickness occurred, as well as the Al-coated specimen. These findings indicate that there is a turning point in the growth process of the Al 3 Ti layer formed at the Al/Ti interface. Based on Fig. 6, the holding time (t o ) and the thickness (d o ) at the turning point were determined to be t o ¼ 10:8 ks and d o ¼ 5:9 mm for the Alcoated specimen, and t o ¼ 72:0 ks and d o ¼ 19:4 mm for the diffusion-bonded specimen. The differences from these turning points were plotted in Fig. 7. The horizontal axis is expressed by the square root of time. It should be noted that the growth of the Al 3 Ti layer showed the liner relationship to the square root of the holding time. This means the diffusion-controlled growth of the Al 3 Ti layer. Nonaka et al. have reported that the impurity elements in the Al and Ti matrices, such as O, Fe and Si, exert a great influence on the layer growth of Al 3 Ti. 9) They have also pointed out the formation of Al oxide between Al and Al 3 Ti. It is known that the stable oxide film exists on the surfaces of Al and Ti, and is responsible for the difficulty in bonding. In addition, the Al foil used in the present study contained the impurity like Fe, as mentioned later. Therefore, the growth of the Al 3 Ti layer in the early stage in Fig. 6 is thought to be affected by the oxide film and the impurity element in the vicinity of the Al/Ti interface. As their influence weakens, its growth may satisfy the parabolic law, as shown in Fig. 7. Figure 6 also indicates that the Al 3 Ti layer grows rapidly for the shorter holding time in the Al-coated Ti plate. The purities of Al and Ti used in the Al-coated specimen were equal to those used in the diffusion-bonded specimen. Thus, the holding time for the marked growth of the Al 3 Ti layer would be affected by the bonding method, although the atmosphere in the heat treatment was also different. The principle of the coating technique in the present study is identical with that of explosive welding. In explosive welding, it is known that the surface layer like oxide film is removed from metal plates by metal jet just before bonding. The metal plates, which are cleaned by the jet action, are joined each other by high pressure and establish a metallurgical bond. Consequently, the cleanliness at the interface in the Al-coated Ti plate would be superior to that achieved by diffusion bonding. Fig. 8 Optical micrographs of the interface in Al-coated Ti plates after heat treatments at 1073 K for (a) 0.6 ks and (b) and (c) 3.6 ks. The specimen in (c) was prepared to determine the position of initial interface. As mentioned above, the Al foil on the Ti plate was mostly replaced by Al 3 Ti after the heat treatment at 898 K for more than 32.4 ks in the air. This Al 3 Ti layer is inadequate to use as a coating layer for the Ti materials because of the occurrence of voids and cracks. 3.2 Liquid-solid reaction in Al-coated Ti plate The Al-coated Ti plate was heat-treated at a temperature of more than melting point of Al (933 K) to investigate the interfacial reaction between molten Al and Ti. Figures 8(a) and 8(b) show optical micrographs of the interface in the specimens heated at 1073 K for 0.6 and 3.6 ks, respectively. The molten Al foil did not spread on other Ti surfaces without coating, since the heat treatment was conducted in the air. In Fig. 8(a), two kinds of reaction products were observed in the solidified Al region. From the results of composition analyses, the granular products indicated as I were identified as Al 3 Ti, while Al and Fe were mainly detected in the needle-shaped products indicated as II. The former products would be produced by the diffusion of Ti into molten Al, and the latter products are considered to be caused by the enrichment of impurities in the Al foil. As the holding time at 1073 K was increased, the microstructural aspect changed drastically. As indicated by III in Figs. 8(b) and 8(c), the Al-coated Ti plate heated at 1073 K for more than 1.2 ks had a porous structure consisting of

5 2182 Y. Morizono, T. Fukuyama, M. Matsuda and S. Tsurekawa Al 3 Ti particles. Such a microstructure was observed in the specimens heat-treated between 973 and 1273 K, and the holding time for its formation decreased with increasing heating temperature. However, the needle-shape products indicated as II in Fig. 8(a) were not recognized in the porous structure. Figure 8(c) shows an optical micrograph of the Al 3 Ti/Ti interface in the purposely-prepared specimen after the heat treatment at 1073 K for 3.6 ks. Two white arrows in the figure denote the Ti surface without coating. This region corresponds to the position of the initial interface, and thus the interface is thought to migrate to the Ti side by the preferential diffusion of Ti into molten Al. Accordingly, the porous structure would result from the consecutive formation of the Al 3 Ti particles in molten Al and the simultaneous consumption of molten Al. The similar microstructure has been also observed in the heattreated Al/Ti multilayer composite material by Fukutomi et al. 12) 3.3 Application to coating material The bonding technique using explosive energy is available for various combinations of metals. In addition, the bonding strength is higher than that of the joint fabricated by any other bonding method. We therefore coated the Ti plate with the Al foil by using explosive energy, and then performed the heat treatment from the viewpoint of diffusion coating for Ti materials. The Al 3 Ti layer formed at 898 K had many voids and cracks, as shown in Fig. 4(b). On the other hand, the porous Al 3 Ti layer was eventually produced by the heat treatment at melting temperatures of Al. This was very brittle, and some Al 3 Ti particles dropped out during grinding and polishing for microstructural observations. In both cases, it is difficult to utilize the Al 3 Ti layer as a coating material. However, the Al + Al 3 Ti duplex structure like Fig. 8(a) is expected to be useable as a coating layer. To disperse the Al 3 Ti particles in the whole region of molten Al, various process flows of the heat treatment were tried as a prior-experiment. An optical micrograph of the acceptable Al + Al 3 Ti duplex state is shown in Fig. 9(a). The duplex state is also confirmed in Fig. 9(b). This microstructure was accomplished with 2-step heat treatment for the Al-coated Ti plate. In the first step, the specimen was heated at 1073 K for 0.6 ks in the air, and then cooled in the furnace to 1023 K. Subsequently, it was kept at 1023 K for 0.6 ks and cooled in the furnace to room temperature. There were no defects like void and crack in the duplex region. In addition, the ratio of Al to Al 3 Ti increased in the vicinity of the surface. Figure 10 shows a hardness distribution of the cross section in the duplex structure as a function of the distance from the surface. The hardness considerably increased from the surface toward the Ti matrix side. This reflects the increase in the ratio of Al 3 Ti in the duplex structure, as shown in Fig. 9(a). The maximum hardness in the Al + Al 3 Ti region was approximately 3 times higher than that of the Ti plate. Although high melting point and excellent oxidation resistance of Al 3 Ti is inefficient owing to the existence of Al in the coating material, the Al + Al 3 Ti duplex state would be beneficial as a surface coating layer formed by the interfacial reaction. Fig. 9 (a) Optical micrograph of the interface in Al-coated Ti plate after continuous heat treatment between 1023 and 1073 K. The Al-coated specimen was heated at 1073 K for 0.6 ks and then furnace-cooled to 1023 K. Subsequently, it was kept at 1023 K for 0.6 ks. (b) XRD pattern of the surface in (a). Fig. 10 Hardness distribution of Al + Al 3 Ti duplex region formed on Ti plate in Fig Conclusions The Ti plate was coated with the Al foil with a thickness of 100 mm by using explosive energy, and then heated between 898 and 1273 K for 0.6 to 90 ks in the air. The main results obtained are as follows. (1) At 898 K, Al 3 Ti was formed along the Al/Ti interface. The thickness of this layer increased in proportion to the holding time in the early stage, and then to the square root of

6 Solid- and Liquid-Solid Reactions in Aluminum-Coated Titanium Substrate Fabricated by Using Explosive Energy 2183 the holding time. Although the Al foil was almost replaced by Al 3 Ti in a holding time of more than 32.4 ks, many voids and cracks were observed in the Al 3 Ti layer. (2) At melting temperatures of Al, the Al-coated Ti plate showed a porous structure consisting of Al 3 Ti particles via the Al + Al 3 Ti duplex state. The holding time to produce the porous structure became short with increasing heating temperature. (3) The Al + Al 3 Ti duplex state was expanded in the whole region of the Al foil by continuous heat treatment between 1023 and 1073 K. There were no defects like void and crack in the duplex region, and the maximum hardness of this region was approximately 3 times higher than that of the Ti plate. Acknowledgements The authors would like to express their appreciation to Dr. T. Yamamuro of Kumamoto University for his kind assistance in the experiment. The present study was supported by JSPS Grant-in-Aid for Scientific Research (C) (No ) and we greatly appreciate their support. REFERENCES 1) Y. Morizono and A. Mizobata: ISIJ Int. 50 (2010) ) D. E. Mikkola, J. P. Nic, S. Zhang and W. W. Milligan: ISIJ Int. 31 (1991) ) B. Crossland: Explosive Welding of Metals and its Application, (Clarendon Press, Oxford, 1982) ) M. Nishida, A. Chiba, Y. Honda, J. Hirazumi and K. Horikiri: ISIJ Int. 35 (1995) ) Y. Morizono, Y. Hirokawa and M. Nishida: Sci. Technol. Energ. Mater. 70 (2009) ) F. J. J. van Loo and G. D. Rieck: Acta Metall. 21 (1973) ) T. Enjyo, K. Ikeuchi, M. Kanai and T. Maruyama: J. Jpn. Weld. Soc. 46 (1977) ) T. Shimozaki, T. Okino, M. Yamane, Y. Wakamatsu and M. Onishi: Defect and Diffusion Forum (1997) ) K. Nonaka, H. Fujii and H. Nakajima: J. Jpn. Inst. Metals 64 (2000) ) U. R. Kattner, J.-C. Lin and Y. A. Chang: Metall. Trans. A 23A (1992) ) J. Tardy and K. N. Tu: Phys. Rev. B 32 (1985) ) H. Fukutomi, M. Nakamura, T. Suzuki, S. Takagi and S. Kikuchi: Mater. Trans. JIM 41 (2000)