Deposition behavior of multi-particle impact in cold spraying process

International Journal of Minerals, Metallurgy and MaterialsVolume 17, Number 5, October 2010, Page 635DOI: 10.1007/s12613-010-0367-8Deposition behavior of multi-particle impact in cold spraying processXiang-lin Zhou, Xiang-kun Wu, Hui-hua Guo, Jian-guo Wang, and Ji-shan ZhangState Key Laboratory for Advanced Metals and Materials, University of Science and Technology Bejing, Beiing 100083, China(Received: 10 December 2009; revised: 20 January 2010; accepted: 25 January 2010)Abstract: In the practical cold-spraying process, a number of particles impact onto a substrate and then form a coating. To study the defor-mation behavior and multi-particle interactions, single-particle, two-particle, and three-particle impacts were simulated using the AN-SYS/LS-DYNA version 970. A copper coating was prepared and scanning electron microscopy (SEM) was employed to analyze the micro-structures of the powders and the coating. Numerical results reveal that the critical deposition velocity is 600 m/s for a copper particle/coppersubstrate. The particles deform more fully due to multi-particle interactions, such as tamping, interlocking, and extrusion effects. The com-multi-particle morphology and compression ratio in the experiment are consistent with those of simulation results. Based on these results, thecoating of high performance can be prepared through selecting appropriate parameters and suitable pre-treatment processes.Keywords: spraying; deposition; interfaces; particle interactions; bonding; simulation[This work was financially supported by the National Natural Science Foundation of China (Nos. 50871019 cand 50874009) cand the NationalSuperiority Discipline Innovation Platform Project (SKL AMM-CS).]1. Introductionintimate conformal contact, which is beneficial to physicalbonding by Van der Waals forces. However, it is very diffi-Cold spraying is a new coating technology. Comparedcult to find experimental proofs of this type of bondingwith thermal spraying [1-2], some processes affected bymode, so far, the physical bonding mode just remains as athermally activation such as oxidation, decarburization,presumption [8]. For the metallurgical bonding, some re-phase transition, and micro-structural changes can besearchers [13-14] hold that the adiabatic temperature shouldavoided. In recent years, investigations on the cold-sprayingreach the melting point at the contact interface to achievetechnology have primarily focused on the optimization ofmetallurgical combination. This usually occurs in conditionsthe cold-spraying system [3-4] and processing parametersof a higher impact velocity or materials with lower melting[5], the development of new coating materials [6], applica-temperatures. For the mechanical bonding, Gruijicic et al.tions in specific fields [7], particle deposition behavior, and[9-10] proposed that jet-like metals may produce interfacialbonding mechanisms [8-12], etc. Numerical and experi-roll-ups and vortices through the Kelvin-Helmholtz effect,mental researches on the bonding mechanisms and deforma-which cause particles to mechanically combine on a sub-tion behavior have also been reported [8-15]. A number ofstrate. However, this view only derives from the qualitativedifferent bonding mechanisms in the particle/substrate con-tact interface, induced by adiabatic shear instability, haveanalysis and presumption.been proposed. For physical bonding, the formation ofFor the reason of the very short duration of particle im-jet-like metals at the local intensively deformed zones maypact, it is very difficult to observe the whole deformationdisrupt thin surface films such as oxides, and provideprocess of particles. Therefore, the most effective methodsCorresponding author: Xiang-lin Zhou E-mail: bkdzxl@sina.com◎University of Science and Technology Bejing and Springer-Verlag Berlin Heidelberg 2010中国煤化工包SpringerMYHCNM HG.636Int. J. Miner. Metall. Mater, Vol.17, No.5, Oct 2010in investigating the bonding mechanism between particleswhereTr is the reference temperature and Tm the meltingand substrates are numerical simulation or theoretical analy-temperature. The material parameters are taken from John-sis associated with relative experiments. Modeling of a sin-son and Cook's research [17].gle particle impacting onto a substrate using finite-elementAnalyses were performed using the models of particlemethods has been mentioned in a few studies, but few at-impact with 4-node shell elements, hourglass control, andtempts have been made to explore multi-particle effectstwo-dimensional automatic single surface contact. The La-though a number of particles impacting onto a substrate andgrangian algorithm was used for solutions based on the ba-then forming a coating in the practical cold-spraying process.sic mass, momentum, and energy conservation equations.Also, these effects have an important influence on the for-The initial kinetic energy of particles was the dominant fac-mation of a cold-spraying coating. The theory of adiabatictor in deposition behavior. Therefore, the effect of the initialshear instability and the formula of critical deposition veloc-temperature was not considered, and the substrate and parti-ity [9, 12] were presented based on numerical simulation.cle temperatures were set to 298 K.Additionally, the criterion of effective bonding is based onThe cold-sprayed particles were assumed to be spherical,the adiabatic shear instability of the particle [8-9]. In fact, ithas been confirmed experimentally that when the hardnessand the substrate was a semi-infinite thick plate. The singleof a substrate is much higher than that of the impacting par-particle model was performed to analyze the deposition be-ticles, adiabatic shear instability in the substrate is negligible.havior. A two-particle numerical model was used to studyConsequently, effective bonding is rarely achieved, even ifthe tamping effect, where the particle diameter was 20 um,the distance between the two particles was 10 um, and theobvious adiabatic shear instability occurs in the particleaverage meshing size at the contact interface was 1 um. To[15-16].describe multi-particle interactions and their impacting be-The adiabatic shear instability in both of the particle andhavior on a substrate, a typical three-particle model wasthe substrate was regarded as a criterion for effective bond-presented, in which particles were 20 pum in diameter, theing in the present study. Numerical simulation, confirmeddistance between the particles was 10 um, and the averagewith actual experiments, was performed to determine themeshing size at the contact interface was 1 pum. The sin-critical velocity for particle bonding. A two-particle modelgle-particle impact and two-particle impact were analyzedand a three-particle model were then explored to studyusing the two-dimensional axisymmetric geometric model,multi-particle interactions, such as tamping, interlocking,while the three-particle model was performed using theand extrusion effects.two-dimensional plain strain model.2.2. Cold gas-spraying process2. Research methodsPowders used for cold spraying were commercially pure2.1. Numerical modelingcopper powders (99.8%) composed of spherical 0Deformation of the particles impacting onto a substratenear-spherical particles ranging from 5 to 45 pum in diameter,was analyzed using the finite-clement program ANas shown in Fig. 1. The substrates were copper and alumi-SYS/LS-DYNA version 970. The plastic response of metalnum with a thickness of 10 mm. A de Laval round-typematerials was assumed to comply with the Johnson-Cooknozzle was used. The acceleration and powder carrier gasplasticity model, and a linear Mie-Gruneisen equation ofwas nitrogen. The stagnation pressure and temperaturestate (EOS) was employed. The model can be written asσ=(A+ BEp )(1+clne*)(l-T*")(1where σ is the flow stress，εp the equivalent plastic strain,ε* the equivalent plastic strain rate normalized with re-spect to a reference strain rate, constants A, B, n, and m aredetermined by material properties, and T* is calculated by1 100 umT*=T-T(2)Fig. 1. Scanningmorphology ofTm-Tthe copper powde中国煤化工TYHCNMH G.X.L. Zhou et al, Deposition behavior of multi-particle impact in cold spraying process637of the acceleration gas were 2.8 MPa and 500°C, respec-Consequently, metallurgical bonding is not possible.tively. A ZESSSUPRA 55 field emission scanning electronThe critical deposition velocity of Cu particles on a Cumicroscope with energyemployed to analyze the powder morphology, microstruc-substrate was calculated, taking into account the adiabaticshear instability of the monitored element both in the parti-tures of the coating, and element ditribution in the crosscle and the substrate as a criterion for effective bonding be-section of the cold-sprayed coatings.tween the particle and the substrate. From Fig. 2(a), obvi-3. Results and discussionously metallic jets of the particle and the substrate areformed at a velocity of 600 m/s, so the critical deposition3.1. Single-particle impactvelocity is determined to be 600 m/s. This value is consis-Fig. 2 shows the simulated impacting behavior of a cop-tent with that of Assadi [8], who reported a calculated criti-per particle on a copper substrate. It is obvious that the de-cal deposition velocity of 550-580 m/s and an experimen-formation of particles is more severe, and the crater on thetally observed value of 570 m/s. The critical deposition ve-substrate surface spreads out further and deeper at a higherlocity for a Cu particle impacting onto aluminum is 500 m/s,impact velocity. As shown in Fig. 2(a), a jet-type flow of thewhich is similar to Grujicic's results [9].material at the interface generates as the impact velocity3.2. Two-particle impactreaches 500 m/s. This flow jet becomes pronounced as theThe Tamping effect of subsequent particles impactingvelocity reaches 600 m/s, where adiabatic shear instabilityonto the initially deposits remarkably affects the growth andalso occurs. A metallic jet was also observed in the experi-properties of a cold-spraying coating in practicalmental run [18]. The distributions of strain and temperaturecold-spraying processes. The simulation result of aat 50 ns after impacting at the velocity of 600 m/s are showntwo-particle impact at a velocity of 500 m/s is shown in Fig.in Figs. 2(b) and (C), respectively. Severe but uniform de-3. Particle A first impacts onto the substrate and both theformation primarily occurs at the particle/substrate interface.particle and the substrate deform. Particle B begins to im-The temperature ditribution is similar to the strain. As canpact onto particle A at 50 ns, causing particle A to furtherbe seen from Fig. 2, a higher temperature region occursdeform and flatten. This phenomenon is the tamping effect.along the particle/substrate interface, with the highest tem-Meanwhile, particle B deforms further and then stops at 100perature in the region of the metallic jet rather than at thens. As shown in Fig. 4, particle I observed in theinitial impact area. The highest temperature increases up tocross- section of the cold-spraying Cu coating is an exampleabout 580 K, which is below the material melting point.of particle A in simulation.(国Impact velocity300 m/s400 m/s500 m/s600 m/s(bPlastic strain (C)Temperature/K6.791e+005.822e+026.112e+00-5.540e+02-5.433e+00-5.258e+02-4.754e+004,975e+02-A4.075e+00/4.693e+02-3.395e+00-' 4.411e+02-2.716c+00、4.129e+02-2.037e+00-3.847e+02-1.358e+00-3.564e+02-6.791e -013.282e+02-0.000e+00-3.000e+02-Fig. 2. Simulated impact of a copper particle onto a copper substrate: (a) deformation behavior for various initial impact velocities;(b) plastic strain and (c) temperature at 50 ns at a velocity of 600 m/s.中国煤化工MHCNMH G.638Int. J. Miner. Metall. Mater., Vol.17, No.5, Oct 2010Ons50ns80ns100 nsB●Fig. 3. Simulation results for two copper particles impacting the copper substrate at a velocity of 500 m/s.velocity with and without tamping effect is shown compara-tively in Fig. 5. The compression ratio increases from 40%to 70% at a velocity of 500 m/s due to the tamping effect.The experimental value of particle I, based on the SEMmeasurement and calculation from Eq. (3), is 69.3%, asshown in Fig. 4. This is consistent with the simulated result.Therefore, the tamping efect plays an important role in par-ticle deformation.|10μm3.3. Three-particle impactFig. 4. SEM morphology of a cold-sprayed Cu coating.As shown in Fig. 6(a), particle B first impacts on theThe compression ratio (R.) of the particles is often used70to indicate the particle deformation level. It is unaffected by65 一Without tamping efetthe meshing size and can be calculated by the following? 60-equation.i sstR。=dp-hpPx 100%(350dpwhere dp is the diameter of original particles and calculated35by dp=引/hp?bp in the experiment, hp the height of the30Eflattened particles along their impacting directions, and bp300350400 450 50( )the width of the flattened particles in the cross-section, asVelocity / (m:s-)shown in Fig.4.Fig. 5. Influence of tamping effect on the particle compressionThe relationship of particle compression ratio and impactratio.(a)16nlet_5 ns) ns50 ns24 ns60nsInterlock中国煤化工Fig. 6. Muti-particle interaction for an impacting velocity of 600 m/s: (a) in.MYHCNMH G efet..X.L. Zhou et al, Deposition behavior of multi-particle impact in cold spraying process639substrate at a velocity of 600 m/s. The plastic deformation in of the impacted particles, which create a mechanical forcethe particle and in the substrate gradually becomes severe,on the adjacent particles, as shown in Fig. 7. Grujicic et al.and consequently, metallic jets form in the particle/substrate10] hypothetically studied the possibility of mechanicalinterface at 16 ns. Then particles A and C begin to impact onbonding in cold-spraying coatings and believed that a me-the substrate with similar evolution of plastic deformations,chanical mixing layer below nanometer or micrometer sizewhich interact with the jet induced by particle B at 20 ns.might be a principle combination mechanism. Through theThe jet of particle B inlays into particles A and C to causearea scanning elemental analysis around the interface of ainterlocking at 24 ns. The jets of particles A and C form atCu coating with an Al substrate, it is found that elements Cuand Al mingle with each other, possibly due to the mixture35 ns, as shown in Fig. 6(b), and further enh:nance the inter-locking effect of particle B and another two particles.of metallic plastic flows shown in Fig. 8. Consequently, theMeanwhile, the plastic deformation and spreading out ofmulti-particle interaction helps to form the mechanicalparticles A and C increase until it is 60 ns. Particles A and C .bonding in cold-spraying coatings. Similar morphologies arealso observed with other cold-spraying coatings, indicatingextrude particle B, which restricts particle B, preventing itthat mechanical bonding extensively occurs in thefrom rebounding. This is beneficial for deposition ancold-spraying coating process.bonding. As shown in Fig. 4, particle II observed in coatingis a typical example of particle B in the simulation. Fig. 7also shows a close-up image that is similar to the numericalresults.The cold-sprayed copper coating onto copper and alumi-num substrates was performed experimentally. Fig. 4 showsthat copper particles deform severely and form lens-like orsplat shapes, and the numerical simulation results are similarto this. When the trace of the interaction effect is caught on1 pumthe suface of the cold-spraying coating through SEM ob-servation, metallic jets are seen to have formed at the fringeFig. 7. SEM close-up image of a cold-sprayed Cu coating.65535 AIK0中71 CuK0■，「(a)| (b)10um10 umfig. 8. SEM analysis of a cold-sprayed Cu coating on an Al substrate: (a) SEM cross section; (b) area scanning of element Al in theCu coating/AI substrate interface; (C) area scanning of element Cu in the Cu coating/Al substrate interface.tamping effect substantially improves the particle compres-4. Conclusionsion ratio from 40% to 70% at a velocity of 500 m/s, furtherThe calculated critical velocity of impacting particles,enhances the deformation and flattening, and improves thestress, strain, and the temperature around the contact inter-density of the cold-spraying coating. Multi-particle interac-face of the particle/substrate enable an accurate descriptiontions such as interlocking and extrusion are helpful in estab-of the deposition behavior and bonding character in thelishing the mechanical bonding mode, therefore improvingcold-spraying process, using the criterion of adiabatic shearthe deposition and bonding efficiency of cold-spraying par-instability both in the particle and the substrate. The criticalicles. Lens-lik中国煤化seved fom thevelocity of Cu particles on a Cu substrate is 600 m/s. Across-section ofaning elementalMYHCNMHG.640Int. J. Miner. Metall. Mater, Vol.17, No.5, Oct 2010analysis around the interface of a Cu coating with an AlAdiabatic shear instability based mechanism for parti-substrate shows that elements Cu and Al mingle with eachcles/ substrate bonding in the cold-gas dynamic-spray process,other. The multi- particle morphology and compression ratioMater. Des, 25(2004), No.8, p.681.are consistent with those of the simulation results.[10] M. Grujicic, JR. Saylor, D.E. Beasley, et al, Computationalanalysis of the interfacial bonding between feed-powder par-Referencesticles and the substrate in the cold-gas dynamic-spray process,1] C. Xia, X. Peng, and J. Li, Behavior of NiCrAIY coating onAppl. Surf. Sci, 219(2003), No.3-4, p.211.the TC6 titanium alloy, J. Univ. Sci. Technol. Beijing,[11] T. Schmidt, F. Gartner, H. Assadi, and H. Kreye, Develop-15(2008), No.2, p.167.ment of a generalized parameter window for cold spray depo-2] Z.s. Fan, D.B. Sun, H.Y. Yu, et al, Preparation of iron basesition, Acta Mater, 54(2006), No.3, p.729.amorphous and nanocrytalline alloy coatings by plasma[12] K. Yokoyama, M. Watanabe, S. Kuroda, et al, Simulation ofspraying, J. Univ. Sci. Technol. Bejing (in Chinese),solid particle impact behavior for spray processes, Mater.27(2005), No.5, p.582.Trans, 47(2006), No.7, p.1697.3] J. Pttion, S. Clotto, R. Morgan, et al, Cold gas dynamic[13] W.Y. Li, C. Zhang, X. Guo, et al, Study on impact fusion atmanufacturing: A non-thermal approach to freeform fabrica-particle interfaces and its efct on coating microstructure intion, Int. J. Mach. Tools Manuyf, 47(2007), No.3-4, p.627.cold spraying, Appl. Surf. Sci, 254(2007), No.2, p.517.4]B. Jodoin, P. Richer, G. Berube, et al, Pulsed-gas dynamic[14] CJ. Li, W.Y. Li, and Y.Y. Wang, Formation of metastablespraying: process analysis, development and selected coatingphases in cold-sprayed soft metallic deposit, Suurf. Coat.examples, Surf. Coat. Technol, 201(2007), No.16-17, p.7544.Technol, 198(2005), No.1-3, p.469.5] T. Stoltenhoff, H. Kreye, and HJ. Richter, An analysis of the15] X.L. Zhou, X.Y. Su, H. Cui, et al, Effect of material proper-cold spray process and its coatings, J. Therm. Spray Technol,ties of cold-sprayed particles on its impacting behavior, Acta11(2002), No.4, p.542.Metal. Sin, 44(2008), No.11, p.1286.6] J.S. Kim, Y.S. Kwon, O.I. Lomovsky, et al, Cold spraying of[16] HB. Zhang, J.B. Zhang, L. Wei, et al, Effecet of substratein situ produced TiB2-Cu nanocomposite powders, Compos.hardness on impact and deposition of cold-sprayed Cu alloySci. Technol, 67(2007), No.11-12, p.2292.particles, [in] Proceedings of Yearly Progress of 2006 on7] S.V. Raj, C. Barett, J. Karthikeyan, and R. Garlick, Com-Material Science and Engineering- -2006 Beijing Interma-parison of the cyclic oxidation behavior of cold sprayedtional Materials Week, Bejjing, 2006, p.417.CuCrAl-coated and uncoated GRCop-84 substrates for space17]G.R. Johnson and W.H. Cook, A constitutive model and datalaunch vehicles, Surf. Coat. Technol, 201(2007), No.16-17,for metals subjected to large strains, high rates and high tem-p.7222.peratures, [in] Proceedings of the 7th Intermational Sympo-8] H. Assadi, F. Gartner, T. Stoltenhoff, and H. Kreye, Bondingsium on Ballistict, Hague, 1983, p.541.mechanism in cold gas spraying, Acta Mater, 51(2003),[18] HH. Guo, X.IL. Zhou, H. Cui, et al, Preparation and simula-No.15, p.4379.tion of cold sprayed copper coatings on metals subtrate,9] M. Grujcic, C.L. Zhao, W.S. DeRosset, and D. Helfritch,Trans. Mater. Heat Treat, 30(2009), No.6. p.158.中国煤化工MYHCNMH G.