APPLIED PHYSICS LETTERS 92, 194101 共2008兲

Dynamic amorphization and recrystallization of metals in kinetic spray process Yuming Xiong, Kicheol Kang, Gyuyeol Bae, Sanghoon Yoon, and Changhee Leea兲 Kinetic Spray Coating Laboratory (NRL), Division of Materials Science and Engineering, College of Engineering, Hanyang University, Seoul 133-791, Republic of Korea

共Received 19 March 2007; accepted 23 April 2008; published online 12 May 2008兲 We reported dynamic amorphization and recrystallization processes of metals upon impact of micron-scaled particles at a high strain rate 共109 s−1兲 combining adiabatic heating with rapid cooling 共1010 K s−1兲 in a kinetic spray process. At the interface of the particle/substrate, an amorphous zone with a thickness of about 3 nm was observed after individual particle impact. It is consistent with the mechanism of amorphous shear lamella and adiabatic shear instability characteristics in kinetic spray process. At the interface of coating/substrate, a rapid phase transition from unstable amorphous to crystalline helps the formation of ductile joints of coatings. © 2008 American Institute of Physics. 关DOI: 10.1063/1.2928218兴 Recently, a kinetic spray process has been of fundamental interest for formation of metallic coatings at low temperatures. High speed 共300– 1200 m / s兲 and high strain rate 共up to 109 / s兲 impact of micron-scaled particles 共around 50 ␮m兲 is the main feature of this process.1–3 Although many coatings 共i.e., pure metals, alloys, polymers, composites, and bulk metallic glasses兲 have been obtained, the bonding mechanisms of coatings are not precisely known. A prevalent opinion is that bonding of kinetic spray coatings depends on the occurrence of adiabatic shear instability at the impact interface, based on finite element analysis 共FEA兲.4,5 Moreover, dynamic recrystallization and/or recovery are considered the main microstructural characteristics with respect to the stacking fault energy of impact materials.6 Similar to the shock compaction process of powder, the strain rate of particles during kinetic spray impact can reach up to 109 / s, which is close to the critical values 共5 ⫻ 1010 / s and 7 ⫻ 1010 / s兲 determined by molecular dynamics 共MD兲 simulations for strain rate-induced amorphization of metallic monocrystal nanowire.7,8 Meanwhile, extremely rapid heating and cooling 共up to 2 ⫻ 1010 K / s兲 共Ref. 4兲 within the zone of adiabatic shear instability are also features of this lowtemperature impact process which might lead to amorphization of metals 共even pure metals兲 through rapid quench with cooling rates in excess of 1010 K / s, as mentioned by Ashby and Jones.9 Therefore, the bonding process of kinetic spray coatings seems to also involve an amorphization process of metals in some cases. Although strain- and strain rateinduced amorphization of many alloys and minerals have been reported this decade,10–14 amorphizations of pure metals are hardly experimentally observed due to inadequate quench rates using current technologies.15 Also, few papers on amorphization have addressed the emergent kinetic spray process. One possible reason is that heat generated during coupled mechanical deformation destabilized amorphous structures thereby concealing the preamorphization process. In this letter, an individual aluminum 共Al兲 particle impact and the absence of long-term heating through successive impacts and hot process gas, makes it possible to determine amorphization at the impact interface. The interface between a兲

Electronic mail: [email protected].

a full nickel 共Ni兲 coating and Al-6061-T6 alloy substrate 共Al-0.5Si-0.28Cu-1.0Mg-0.2Cr, at. %兲 was also investigated to elucidate the bonding process of coatings from amorphization to recrystallization. A kinetic spraying system 共KINETIC3000, Germany兲 was used. For individual impact test, spherically shaped Al powder 共purity ⬎99%兲 with a mean size of 68 ␮m was used as feedstock. Also, full Ni coating was deposited using spherically shaped Ni powder with a mean size of 35 ␮m. Helium was used as process gas to achieve high impact speed of incident particles and resultant hyperhigh strain rate. The detailed experimental setup and coating process are described elsewhere.16 According to an empirical equation,17 the velocities of Al particles 共accelerated by 300 ° C-2.4 MPa gas兲 and of Ni particles 共accelerated by 400 ° C-2.5 MPa gas兲 are 875 and 784 m s−1, respectively. The localized strain rate and cooling rate, estimated by FEA using ABAQUS, were up to 1.36⫻ 109 / s and 1.83 ⫻ 1010 K / s upon impact, respectively. Due to the advanced microsampling technology of focused ion beam 共FIB兲, the transmission electron microscopy 共TEM兲 samples of individual particle impact interfaces can be conveniently prepared on site. Specimens were sliced and thinned by FIB across the central interface between an individual Al particle and 1050 Al alloy substrate 共⬎99.5% Al兲 关Fig. 1共a兲兴 for high resolution TEM observation. In the case of the full nickel coating, the cross-sectional TEM samples were prepared through a painstaking procedure. Two polished Ni coating surfaces 共Ra ⬍ 1.5兲 were tightly bonded face to face using Mbond 610 adhesive, and solidified for 2 h at 120 ° C. Subsequently, the specimens were sliced normal to the bonded interface to a thickness of 1.5 mm, polished to ⬍40 ␮m, mounted on a copper grid with a diameter of 3 mm, dimpled, and ion milled for thinning 共perforating at the bonded interface兲 for field emission-TEM observation at an applied excitation voltage of 300 kV. As shown in Fig. 1共b兲, a bonding zone with a thickness of around 10 nm between Al particle and substrate is observed. A higher magnification image 关Fig. 1共c兲兴 at the bonding zone 关taken from boxed region in Fig. 1共b兲兴 reveals the presence of an amorphous zone with a thickness of around 3 nm upon impact. Fast fourier transforms 共FFTs兲 taken from

0003-6951/2008/92共19兲/194101/3/$23.00 92, 194101-1 © 2008 American Institute of Physics Downloaded 12 May 2008 to 166.104.133.131. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

194101-2

Xiong et al.

FIG. 1. 共a兲 TEM sample preparation for individual impact interface by FIB. 共b兲 HREM images of the individual particle impact interface between Al particle and substrate. 共c兲 Close observation at the selected box region 共insets for FFT patterns of selected area兲.

within box-1 contain a diffuse halo, which is a typical character of an amorphous phase. A survey of high resolution electron microscopy 共HREM兲 images at the impact interface indicates that the length of the amorphous zone is ⬎300 nm. On the side of the substrate, no evidence of defects including dislocations, stacking faults, or twins, is observed in regions away from the amorphous zone. Evidence of a sharp truncation of pre-existing grains along the 共0 0 2兲 crystallographic orientation reveals an interruption of the fringes of the Al grain. On the side of the particle, a blurry transient boundary from amorphous to crystalline 共A-C兲 can be detected. From the amorphous zone, the disorder degree of atomic alignment decreased toward the particle interiors. In addition, an angular mismatch about 1.5° between the lattice images along the 共−111兲 crystallographic orientation on both sides of the amorphous zone was measured and determined with FFT patterns. The collapse of the Al crystalline structure for amorphization seems not to occur along specific crystallographic planes of each nanosized grain at the impact interface because of heterogeneous localized deformation in the kinetic spray process. These observations might be consistent with the mechanism of the A-C phase transition being the development of shear instability, interpreted as amorphous shear lamellae, in which amorphous nucleation occurs among the weak planes with the highest shear strain in the crystalline structure.18 Therefore, the amorphization of aluminum seems to be associated with the occurrence of adiabatic shear instability, which is considered the bonding mechanism of kinetic spray coatings. Only within those

Appl. Phys. Lett. 92, 194101 共2008兲

zones of shear instability can the strain rate and strain approach the critical values for amorphization. It is worth noting that the strain rate in our case is lower than the lower limit of 5 ⫻ 109 / s for amorphization of Ni nanowire estimated by the MD simulation.7 Although the limit for strain rate-induced amorphization of different materials is inevitably affected by several factors, e.g., the deformability and glass forming ability, the adiabatic heating and rapid cooling at the impact interface cannot be ignored when considering the amorphization mechanism of Al in our impact case. The presence of a blurry boundary between the amorphization zone and the particle implies that rapid quenching during impact would also synchronously stimulate the amorphization process. As shown in Fig. 1共b兲, the amorphous zone has a flow-alike feature. Mechanical melting and subsequent rapid quench might have occurred, thereby resulting in an amorphization process at the impact interface since the simulated maximum interface temperature 共836 K兲 approaches the melting points of Al 共933 K兲 and its alloys 共⬃855 K兲 in terms of FEA. Monte Carlo simulations suggested that stress-induced melting of fcc elemental crystals can occur at temperatures Tm共␴兲 well below the thermodynamic melting temperature Tm共␴ = 0兲.19 However, an ab initio MD simulation indicated that pure Al can be amorphized through an “undermelt quench” approach.20 Whether melting occurs or not at impact interfaces is still one of the controversial topics so far in the field of kinetic spray. Hence, further work combining MD-type simulations with specifically designed individual particle impact tests is needed to fully understand the complex amorphization process in kinetic spray process. It is very difficult to maintain and experimentally observe the amorphization of pure metals because of rapid crystallization at longer time scales 共even several nanoseconds for impact time in our cases兲 even at near zero Kelvin. The amorphization of Al in our cases may also be stimulated and stabilized by some absorbable impurities. Although the surface oxides and melts 共if melting兲 could be mostly jetted out upon impact, oxygen and other absorbable impurities could be trapped in the remaining melted layer at the impact interface. Moreover, a rough calculation indicates that elemental diffusion distance in Al melts is around 1 nm within impact time 共⬃45 ns兲. Thus, the thickness 共⬃3 nm兲 of amorphization zone in this letter seems to be in good agreement with the effect of impurities stabilization. Precise elemental composition analysis within amorphization zone is needed to clarify the characteristics. The full coating case is quite different from the individual particle impact in a kinetic spray process. The rise of temperature within coatings due to coupled mechanical deformation and heating by hot gas during spraying may last several minutes, thereby accelerating the subsequent recrystallization and/or recovery of the preformed amorphous structures. Figure 2共a兲 shows a typical low-magnification TEM image of the interface between two continuously impacted Ni particles and a substrate. Figure 2共b兲 shows the FEA simulated interface temperature distribution after successive impact. The detailed FEA modeling process was described elsewhere.5 Dependent on deformability, the kinetic energy of the injecting Ni particle is mostly converted into plastic deformation and mechanical heating of relatively soft Al alloy substrate. FEA implies that the strain rate of the substrate 共1.36⫻ 109 / s兲 in case of Ni-to-Al is higher than

Downloaded 12 May 2008 to 166.104.133.131. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

194101-3

Appl. Phys. Lett. 92, 194101 共2008兲

Xiong et al.

FIG. 2. 共a兲 Low-magnification TEM image of interfacial microstructures between as-sprayed Ni coatings and Al-6061-T6 alloy substrate in a kinetic spray process. 共b兲 FEA results for interface temperature under successive impacts; and HREM images of boxed region-1 共c兲 and region-2 共d兲 in 共a兲 共insets for FFT patterns of respectively selected areas兲.

that 共1.03⫻ 109 / s兲 in case of Al-to-Al. In terms of the critical strain rate for amorphization,7 the interfacial preamorphization process at the side of the severely deformed substrate may have taken place prior to the successive impact of the second Ni particle. As shown in Fig. 2共c兲, remaining amorphous structure is observed at the interface according to HREM image and FFT patterns taken from the boxed region-i. It implies that the strain rate-induced amorphous structure might be stable to some extent against heating by hot process gas within spraying time. As reported by Zhang and Zhao,15 the pressure-induced amorphization of pure zirconium shows reversible and irreversible features within different pressure ranges. An irreversible transition along closepacked hexagonal, hexagonal, and body-centered cubic into amorphous structures occurs under medium pressure load. A theoretical prediction for the kinetic spray process reveals that the impact pressure of incident materials ranges in between 3 and 35 GPa depending on the impact velocity and locations.21 Thus, the transformation C-A of Al metals presented in this letter might also involve both the features within different strain-affected regions although irreversible transformation zones were extremely narrow with a length of ⬍15 nm. In fact, most interface of the Ni/ Al shows crystallographic transition from particle to substrate. As shown in Fig. 2共d兲, a close observation of boxed region-2 in Fig. 2共a兲 关maximum shear strain and temperature rise according to FEA-Fig. 2共b兲兴 reveals that no evidence of the existence of an amorphous structure was found, apart from some discontinuous disordered microstructures 共marked with an ellipse兲. This might be due to incomplete transformation A-C under heating. Interestingly, the interface between coating and substrate exhibits an ideal transition of atomic alignment. Fresh Al grains prefer being oriented along the specifically close-packed planes of the preexisting neighboring Ni lattice 关e.g., preferable 共200兲 Al 共2.0245 Å兲 orientation due to the smallest difference in interplanar distance with neighboring 共111兲 Ni 共2.0334 Å兲兴. The preferably oriented nucleation and growth of Al grains from the amorphous phase helps to release internal stress to form ductile bonding of kinetic spray coatings. However, within the substrate interiors away from the boundary of Ni/ Al, polycrystalline grains form with random orientations. It may result from dynamic recovery of aluminum during impact at lower strain rate than the critical one for amorphization.6 In summary, interfacial amorphization of Al metals during high strain rate impact of micron-scaled particles in a

kinetic spray process was observed in this letter. This might result from synergistic effects of strain rate induction and rapid quenching. Also, absorbable impurities 共e.g., oxygen兲 and interdiffused elements were trapped at the impact interface to stabilize amorphous phase. Nucleation and growth of fresh grains from amorphous due to heating by hot process gas and coupled impacts during spraying are helpful to form ductile joints of coatings. This work was supported by the Korea Science and Engineering Foundation 共KOSEF兲 grant funded by the Korean government 共MOST兲 共No. 2006-02289兲. 1

A. P. Alkhimov, V. F. Kosarev, N. I. Nesterovich, and A. N. Papyrin, Method of applying coatings, Russian Patent No. 1,618,778 共8 September 1990兲. 2 A. P. Alkhimov, A. N. Papyrin, V. F. Kosarev, N. I. Nesterovich, and M. M. Shushpanov, U.S. Patent No. 5,302,414 共12 April 1994兲. 3 W. B. Choi, L. Li, V. Luzin, R. Neiser, T. Gnäupel-Herold, H. J. Prask, S. Sampath, and A. Gouldstone, Acta Mater. 55, 857 共2007兲. 4 T. Schmidt, F. Gärtner, H. Assadi, and H. Kreye, Acta Mater. 54, 729 共2006兲. 5 H. Assdi, F. Gärtner, T. Stoltenhoff, and H. Kreye, Acta Mater. 51, 4379 共2003兲. 6 C. Borchers, F. Gartner, T. Stoltenhoff, and H. Kreye, J. Appl. Phys. 96, 4288 共2004兲. 7 H. Ikeda, Y. Qi, T. Cagin, K. Samwer, W. L. Johnson, and W. A. Goddard, Phys. Rev. Lett. 82, 2900 共1999兲. 8 P. S. Branicio and J. P. Rino, Phys. Rev. B 62, 16950 共2000兲. 9 M. F. Ashby and D. R. H. Jones, Engineering Materials 2, An Introduction to Microstructures, Processing and Design, 2nd ed. 共Pergamon, Oxford, 1998兲. 10 J. Li, Z. Q. Jin, J. P. Liu, Z. L. Wang, and N. N. Thadhani, Appl. Phys. Lett. 85, 2223 共2004兲. 11 T. Irifune, K. Kuroda, N. Funamori, T. Uchida, T. Yagi, T. Inoue, and N. Miyajima, Science 272, 1468 共1996兲. 12 P. R. Okamoto, J. K. Heuer, N. Q. Lam, S. Ohnuki, Y. Matsukawa, K. Tozawa, and J. F. Stubbins, Appl. Phys. Lett. 73, 473 共1998兲. 13 M. Chen, J. W. McCauley, and K. J. Hemker, Science 299, 1563 共2003兲. 14 E. Ma, Scr. Mater. 49, 941 共2003兲. 15 J. Zhang and Y. Zhao, Nature 共London兲 430, 332 共2004兲. 16 S. H. Yoon, C. Lee, H. S. Choi, and H. G. Jo, Mater. Sci. Eng., A 415, 45 共2006兲. 17 J. Wu, H. Fang, S. Yoon, H. Kim, and C. Lee, Appl. Surf. Sci. 252, 1368 共2005兲. 18 P. Cordier, J. C. Doukhan, and J. Peyronneau, Phys. Chem. Miner. 20, 176 共1993兲. 19 R. M. Lynden-Bell, J. Phys.: Condens. Matter 7, 4603 共1995兲. 20 A. A. Valladare, R. M. Valladares, and A. Calles, MRS Symposia Proceedings No. 848 共Materials Research Society, Warrendale, PA, 2005兲, FF9.29.1. 21 J. Vlcek, L. Gimeno, H. Huber, and E. Lugscheider, J. Therm. Spray Technol. 14, 125 共2005兲.

Downloaded 12 May 2008 to 166.104.133.131. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp