Al nanocomposite using copper interlayer

Al nanocomposite using copper interlayer

Materials and Design 53 (2014) 275–282 Contents lists available at SciVerse ScienceDirect Materials and Design journal homepage: www.elsevier.com/lo...

4MB Sizes 3 Downloads 130 Views

Materials and Design 53 (2014) 275–282

Contents lists available at SciVerse ScienceDirect

Materials and Design journal homepage: www.elsevier.com/locate/matdes

Microstructure and mechanical properties of transient liquid phase bonding of Al2 Op3 /Al nanocomposite using copper interlayer S.S. Sayyedain, H.R. Salimijazi ⇑, M.R. Toroghinejad, F. Karimzadeh Department of Materials Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran

a r t i c l e

i n f o

Article history: Received 31 October 2012 Accepted 29 June 2013 Available online 12 July 2013 Keywords: Mechanical properties Microstructure Characterization Nanostructured materials Mechanical alloying Sintering

a b s t r a c t Transient liquid phase (TLP) bonding is one of the state-of-the-art joining processes. It is used for welding composites and advanced materials. Microstructure and mechanical properties of TLP bonding depends on the bonding time and temperature. In the current study, the effect of bonding time on the microstructure and bonding strength of the TLP diffusion bonded of Al2 Op3 /Al nanocomposite were investigated. A thin layer of copper deposited by electroplating was used as interlayer. With increasing the bonding time from 20 to 60 min, at a constant bonding temperature of 580 °C, the isothermal solidification was completed and the final joint microstructure consisted of soft a-Al phase with dispersed CuAl2 precipitated particles. Decreasing the amount of brittle eutectic structures in the joint seam by increasing the bonding time was the main reason for improvement of the joint shear strength. The maximum joint shear strength was achieved at 580 °C for 60 min which was about 85% of the shear strength of the base material. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Aluminum matrix composites are widely used in many industrial applications such as aerospace, automotive and electronics due to their good properties such as high strength to weight ratio, low density, high hardness, and good oxidation and wear resistance. These properties can be superior when the composites structures become nanostructure. Despite the unique properties of nanocomposite materials, the lack of a reliable joining process has limited their full potential in engineering applications [1–4]. Fusion welding processes are inefficient for joining of advanced materials due to forming undesirable interfacial reactions between the matrix and reinforcement, poor flowability of the liquid welding pool, and producing defects such as porosity in the bonding area [5–7]. Solid state bonding processes has been widely developed for joining composite materials. Amongst solid state bonding processes, transient liquid phase (TLP) diffusion bonding has been widely used for joining of Al matrix composites [8,9]. TLP diffusion bonding is a joining process which involves formation of metallurgical bonding. In this process a thin continuous layer of liquid at the joint interface can be formed due to the eutectic reaction formed between the interlayer and the base metal. This reaction also has the advantage of been able to remove surface oxides, which is useful for Al- based composites that have tenacious oxide

⇑ Corresponding author. Tel./fax: +98 311 391 5769. E-mail address: [email protected] (H.R. Salimijazi). 0261-3069/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.matdes.2013.06.074

layer on their surfaces [10,11]. The liquid film wets the contacting metallic substrates and isothermal solidification takes place due to the inter-diffusion of atoms at a constant bonding temperature. It is generally considered that TLP diffusion bonding consists of four different stages [12–16]: (1) (2) (3) (4)

Dissolution of the interlayer. Homogenization of the interlayer. Isothermal solidification. Homogenization of the bond region.

The ceramic particles in the composites are incorporated into the bond region either by using a particle reinforced filler material [17] or by melting the substrate material [18] during TLP process. Early two stages are controlled by the liquid diffusion and may last from less than a minute to few minutes. The last two stages are controlled by the solid diffusion and may need several hours for completion. The third stage is completed when the whole joint area is solidified isothermally. Effective parameters in the third stage are bonding time and temperature. The solid/liquid interface moves towards the bond centerline. If the conditions for completion of the isothermal solidification are not fulfilled, the residual liquid at the bond centerline is solidified during cooling [19,20]. In this study, the TLP diffusion bonding process for joining of aluminum matrix nanocomposite reinforced by nanometric alumina particles was studies. A thin layer of copper electroplating was used as an interlayer between two parts. The effect of bonding time

276

S.S. Sayyedain et al. / Materials and Design 53 (2014) 275–282

on the microstructural development across the joint and its subsequent effect on the shear strength and failure mechanism were investigated.

2. Materials and methods 2.1. Materials Al/10 wt.% Al2O3 nanocomposite powders were produce by mechanical alloying of nano Al2O3 (20–30 nm) and pure aluminum powders under argon atmosphere and room temperature conditions after 7 h milling. In order to produce bulk nanocomposite samples, nanocomposite powders were hot pressed at 500 °C using 450 MPa pressure.

Fig. 2. Effect of bonding temperature on total time required for the isothermal solidification of different eutectic systems [13].

2.2. Sample preparation and bonding process The specimens were prepared for bonding by cutting to a dimension of (20  5) mm2 with 5 mm thickness. The bonding surfaces of the base materials were grinded to 800 grit SiC sand papers and then cleaned in an acetone bath. Prior to bonding, a 25 lm thick layer of copper coating was deposited on one piece of each couple joining surfaces of the nanocomposite by electro deposition technique. Each two specimens were assembled to each other at room temperature (Fig. 1). Since the base metal is nanocomposite, to prevent growing of the matrix grain, the bonding temperature was selected so that isothermal solidification performed in minimum time (Fig. 2). Thus, the bonding temperature (580 °C) was selected which is about 30 °C above the eutectic temperature of Al–Cu (TE = 548 °C). Thus, TLP diffusion bonding was carried out at 580 °C for different durations of 20, 40 and 60 min in a vacuum furnace under a vacuum of 102 mbar (1 Pa) at constant pressure of 0.086 MPa. The specimens were brought to the joining temperature, held at joint temperature followed by cooling down to room temperature at the furnace. Bonded samples were cut perpendicular to the bond-line by an abrasive saw to four equal parts. Three parts were used to evaluate the shear strength of the bonded joints and other part was used for microstructural analyses. Cross-sections of the joints were prepared for metallographic analysis by standard polishing techniques according to ASTM-B253. For examination of microstructures, specimens were etched by Keller etchant. Optical microscopy, scanning electron microscopy, and transmission electron microscopy were employed to study the microstructures of the bonding areas as well as bulk nanocomposite material. The shear strength of the bonded joints was evaluated by means of a specially designed fixture as illustrated in Fig. 3, in a tension testing machine (HOUNSFIELD/H50KS) with a loading speed of 1 mm/ min. The shear strength was calculated by dividing the maximum load by the bonded area. For each bonding condition, three specimens were tested and the average value was reported as the shear strength (bond strength). Finally, evaluations of the joint fracture surfaces were carried out by scanning electron microscopy (SEM).

Fig. 3. Schematic of the supplemental fixture for shear test of joints.

3. Results and discussion 3.1. Microstructure of TLP bonded joints Fig. 4a and b shows TEM and SEM images of the Al/10 wt.% Al2O3 nanocomposite microstructure used as the parent material for TLP welding. The alumina reinforcement particles, with a particle size of about 20 to 30 nm, well dispersed in the soft matrix of aluminum. Within the matrix of metal, dislocations can be seen as interconnected networks. In some regions, cells and sub-grains with sizes less than 100 nm can also be observed. These areas are dislocation-free spaces that are separated by walls of dislocations. The nanometeric alumina particles can also be observed in these walls

Fig. 1. Specimens assembling before the joint process.

S.S. Sayyedain et al. / Materials and Design 53 (2014) 275–282

277

Fig. 5. SEM image of Al/10 wt.% Al2O3 nanocomposite powder microstructure.

Fig. 4. Microstructure of the Al/10 wt.% Al2O3 metal–matrix nanocomposite. (a) TEM image. (b) SEM image.

of cells and sub-grains. The nanocomposite has some pores (Fig. 4b) which came from two sources: porosities produced during hot pressing and sintering process of powder particles and lamellar porosities produced during mechanical grinding process. In the mechanical grinding process, the soft aluminum powder particles were flatted and cold worked to form layered structures after impacting by bullets. Due to sequential processes of the cold welding and fracture within each of the powder particles, the lamellar structures were formed. The layered structure within a particle of nanocomposite powder can be seen in Fig. 5. Fig. 6 shows optical micrographs of the joint microstructure at 580 °C for different bond times. The joint seams cannot be seen easily and in some cases it has been disappeared. Some phases with brighter color from the base metal can be distinguished in the joint seam and inside the pores of the base metal, close to the joint seam. 3.1.1. Mechanism of TLP bonding in Al matrix nanocomposite In the bonding of aluminum based metals with copper interlayer, copper atoms diffuse from the interlayer through the base metal and aluminum atoms diffuse from the base metal into the copper interlayer. In fact, diffusion takes place in the real contact area between the base metal and interlayer. By proceeding the diffusion in the contact areas, the base metal composition reaches to CaL, based on Al–Cu phase diagram shown in Fig. 7. By continuing the diffusion of copper atoms into the base metal, the composition of the contact regions can reach to more than CaL and these regions start to be melted, resulting in forming a layer of molten materials between the base metals. This interlayer be-

comes wider, gradually. Since the base metal has a porous structure (Fig. 4b), the melt created in the joint seam quickly flows into the interconnected pores due to the capillary force. If the base metal has micronic reinforcement particles, while the liquid phase was created in the joint seam, the reinforcement particles that have been trapped in the composite matrix, were and agglomerated in the joint seam. Agglomeration of micronic reinforcement particles to the bond region has been reported in TLP bonding of TiCp/AZ91D [14,15] and Al MMCs [21]. The agglomerated particles in the joint seam reduced the bond strength, significantly [7]. If the base metal has nanoparticles such as materials used in this study, with dissolution of the nanocomposite matrix in the area around the joint seam, alumina nanometeric particles that already were scattered in the aluminum matrix entered to the melt. But since their particles dimensions are very small (around 20–30 nm), these nanoparticles could penetrate into the pores easily by the molten material. Then, the copper interlayer and a part of the base metal that was already in contact with the interlayer are completely dissolved and the liquid in the joint seam reaches to the chemical composition between CLa and CLb (Fig. 7). This liquid contains more copper atoms than the base metal that is in contact with the liquid. The surfaces of the base metal that is in contact with liquid (the joint seam and the porous surfaces), has CaL chemical composition and the liquid which is in contact with the surfaces has CLa homogeneous chemical composition. From this moment onwards, the isothermal solidification is started by nucleation of the solid embryos on the base metal surfaces. These nucleuses have a chemical composition equal to CaL (Fig. 7) following by entering into the growing step. The movement of the solid/liquid interface tries to maintain a thermodynamically balance between CLa of liquid and CaL of solid, but since the establishment of such as balance needs solid state diffusion, the speed is very low and this step is very time consuming. It is expected that the isothermal solidification rate is high due to its nanostructures and high diffusion coefficient. If time for the isothermal solidification is sufficient, the melt completely solidified isothermally and the final structure will be solid-phase with CaL composition. But, if the bonding time is not sufficient, while the melt is at the bonding temperature, it continues to isothermally solidify and when time ended and the bonding temperature drops, the melt is solidified conventionally. The final structure consists of a solid (isothermal solidification stage), primary a solid due to melt cooling with CLa composition from the bonding temperature to the eutectic temperature and the eutectic solid of CuAl2 + a from the eutectic solidification of the melt at the eutectic temperature. This structure consists of platelet of brittle

278

S.S. Sayyedain et al. / Materials and Design 53 (2014) 275–282

Fig. 6. Optical micrographs of the TLP bonded microstructure at 580 °C: (a) 20 min, (b) 40 min, and (c) 60 min.

Fig. 7. Al–Cu phase diagram [21].

CuAl2 at the center of the joint line, which reduces the bonding strength. After completion of the isothermal solidification, homog-

enization will happen. In this stage, the solid composition that was formed in the isothermal solidification stage (CaL), reduces due to

S.S. Sayyedain et al. / Materials and Design 53 (2014) 275–282

Fig. 8. The structure of the solidified liquid in porosities of the base metal.

279

will be similar to the matrix. Since the homogenization stage is very time consuming, the bonding time is not sufficient for complete homogenization. Therefore, the composition of the solid phase in the pores and joint areas remains around CaL or even less than CaL. After finishing the bonding process, if the amount of copper in the a-solid phase is higher than its solubility limit at ambient temperature, CuAl2 precipitated particles will emerge in the context of the solid phase. It should be noted that nanometeric alumina particles that are in the context of a-Al solid phase, can have an important role in the nucleation of this precipitated particles and accelerate the nucleation and improve the distribution of the precipitated particles. The phases observed with the brighter color in the joint seam and inside the porosities, are the isothermally solidified solid phase. The melt created in the joint seam after being sucked into the pores and cavities that have been connected to the bonding surfaces, was solidified isothermally and aluminum solid solution containing copper is produced. The copper content depends on the bonding temperature and time). This phase can be seen in more detail at higher magnification in Fig. 8. The CuAl2 precipitated particles can be observed in the matrix of a-Al phase. Layer structure of the base metal (the left side of Fig. 8) can be seen in bright colors which is known as a-Al solid phase. 3.1.2. The effect of bonding time on joint microstructures The bonding time can strongly affect the final joint microstructure. If the isothermal solidification time is not sufficient, the final microstructure consists of a eutectic composition of CuAl2 + a

Fig. 9. Eutectic solidification between the solidified phases in the isothermal solidification.

Fig. 10. The Effect of bonding duration on the joint shear strengths at 580 °C.

the diffusion of copper atoms into the base metal. Thus, the chemical composition of the produced solid and its melting temperature

Fig. 11. SEM micrographs from the fractured surfaces of the base metal (a) and high magnified image (b).

280

S.S. Sayyedain et al. / Materials and Design 53 (2014) 275–282

Fig. 12. SEM micrographs from the fractured surfaces of joints TLP bonded at 580 °C: (a) 20 min; (b) 40 min; (c) 60 min.

Fig. 13. XRD patterns of the fracture surface of the joint TLP bonded at 580 °C and 60 min.

between solids. For example, in the specimens bonded at 580 °C for 40 min, the retained melt was solidified eutectically between the solid phases. This eutectic solidification is shown in Fig. 9. Thus, at bonding temperature of 580 °C, 40 min time is not sufficient to complete the isothermal solidification. Formation of these compounds by eutectic reaction during cooling of joint has been reported in literature [14,15,21]. By increasing the bonding time to 60 min, no significant eutectically solidified structures can be observed, so the bonding time of 60 min probably has enough to complete isothermal solidification. According to literature

[16,21], enough bonding time at a constant bonding temperature led to elimination of intermetallic compounds during TLP bonding. 3.2. Mechanical properties of TLP bonded joints 3.2.1. Shear strength of TLP bonded joints The shear strength versus bonding time at bonding temperature of 580 °C is showed in Fig. 10. The shear strength increased with increasing the bonding time. There was no significant difference between the bonding

S.S. Sayyedain et al. / Materials and Design 53 (2014) 275–282

strength of the joints at 20 and 40 min bonding time, but with increasing the bonding time to 60 min, a significant increase can be observed. At bonding time of 20 and 40 min, the time required for isothermal solidification of molten was not sufficient. The melt remained in the joint seam, eventually participated by the eutectic reaction, and a network of the brittle eutectic compounds in the joint seam is formed (Fig. 9). The failure of the bonding is happened through the brittle eutectic phases. By increasing the bonding time to 60 min, the bonding strength increased, significantly due to the absence of the brittle eutectic structure in the joint area. At this bonding condition, the time required to complete the isothermal solidification is enough and the melt completely solidified isothermally. The required time for TLP bonding of the aluminum composites reinforced by micro size particles are much higher than that for this nanocomposite [21]. Reduction in the required time for the isothermal solidification in the nanocomposite structures can be attributed to two main reasons; (i) penetration of the molten material into the porosities increases the contact surfaces between the base material and molten metal and decreases the thickness of the molten layer and (ii): base metal structure consists of nanometeric grains, high density of grain boundaries and nanometric alumina particles which increases the copper diffusion coefficient in the base metal. High diffusion coefficient of copper in the base metal can also reduce the time required for isothermal solidification. The effect of diffusion coefficient and thickness of the molten metal on the completion time of the isothermal solidification can be expressed in Eq. (1) [13].



W 2 p C La DaCu C aL

!

281

4. Conclusions Microstructure and mechanical properties of the TLP bonded Al2 Op3 /Al nanocomposite by using copper interlayer were investigated, comprehensively. The base metal consisted of some porosities which came from two main sources; (i) porosities which are produced in the hot pressing process and sintering of powder particles and (ii) layer porosities which are produced by mechanical grinding process. Al2 Op3 /Al nanocomposites were bonded successfully by TLP diffusion bonding using Cu interlayer. The bonding time was a most important factor that had significant effects on the final microstructure and mechanical properties of the joints. By dissolution of the nanocomposite matrix in the area around the joint seam, alumina nanometeric particles entered into the molten material and penetrated into the porosities. The bonding times of 20 and 40 min were not sufficient to complete the isothermal solidification of the melt. By increasing the bonding time to 60 min, the required time for completing the isothermal solidification was sufficient and the final microstructure consisted of soft a-Al phase. Because of the absence of the brittle eutectic structures in the joint seam, the bonding strength increased, significantly. The maximum joint shear strength was achieved in the sample bonded at 580 °C for 60 min, which was about 85% of the shear strength of the base material. The fracture surfaces of the joints bonded for 20 and 40 min were ductile with shear dimples in some areas. Crack progress path was through the brittle fracture regions. The fractured surface of the joints bonded for 60 min was ductile and shear dimples clearly observed. Some CuAl2 dispersed precipitated particles were seen around the shear dimples.

ð1Þ References a

where W is thickness in contact with the molten metal, DCu is the copper diffusion coefficient in the base metal and t is the time required to completing isothermal solidification.

3.2.2. Fracture surface of TLP bonded area In order to compare the fracture surfaces of the bonded areas to fracture surface of the base metal and identify the mechanism of failure, initially the fracture surface of the base metal was evaluated. Fig. 11 shows the fracture surfaces of the base metal prior to the joining. Crack growth was happened between the sintered powder particles and also through the layers of the mechanically alloyed particles (shown with arrows). Some microcracks and porosities can also be observed. The fracture surfaces of the joints area in the different bonding times are shown in Fig. 12. The fracture surfaces of the joints were ductile in some areas and shear dimples can be observed. Brittle fracture can also be observed in some areas. Brittle fracture regions were probably due to the failure of the eutectic phase (see Fig. 9). In the joints bonded for 20 and 40 min, due to the insufficient time for completing the isothermal solidification, networks of eutectic brittle structures formed in the joint seam. Fig. 12(c) shows the SEM image of the fractured surfaces of the joint bonded for 60 min. The fracture surface was ductile and shear dimples can be seen clearly. Moreover, the CuAl2 precipitated particles can be seen around the shear dimples (arrows). When the shear force applied to the joint seams, the soft a-Al phase in the joint seams, was stretched around the CuAl2 particles (which were scattered in the context of Al-a) and finally ruptured. Fig. 13 shows X-ray diffraction pattern performed on the fracture surfaces. Aluminum phase from the base metal and a-Al and CuAl2 phases (precipitated particles) are clearly identified.

[1] Moya JS, Esteban SL, Pecharroma C. The challenge of ceramic/metal microcomposites and nanocomposites. J Prog Mater Sci 2007;52:1017–90. [2] Tabandeh Khorshid M, Jenabali Jahromi SA, Moshksar MM. Mechanical properties of tri-modal Al matrix composites reinforced by nano- and submicron-sized Al2O3 particulates developed by wet attrition milling and hot extrusion. J Mater Des 2010;31:3880–4. [3] Ellis MBD. Joining of metal–matrix composites: a review. J TWI 1997;6: 69–128. [4] Kumar Pal T. Joining of aluminum metal–matrix composites. J Mater Manuf Proc 2005;20:717–26. [5] Muratoglu M, Yilmaz O, Aksoy M. Investigation on diffusion bonding characteristics of aluminum metal matrix composites (Al/SiCp) with pure aluminum for different heat treatments. J Mater Proc Technol 2006;178: 211–7. [6] Arik H, Aydin M, Kurt A, Turker M. Weldability of Al4C3–Al composites via diffusion welding technique. J Mater Des 2005;26:555–60. [7] Zhang G, Zhang J, Pei Y, Li S, Chai D. Joining of Al2 Op3 /Al composites by transient liquid phase (TLP) bonding and a novel process of active-transient liquid phase (A-TLP) bonding. J Mater Sci Eng A 2008;488:146–56. [8] Huibin X, Jiuchun Y, Zhiwu X, Baoyou Z, Shiqin Y. Interface structure changes during vibration liquid phase bonding of SiCp/A356 composites in air. J Composites: Part A 2006;37:1458–63. [9] Xu Z, Yan J, Wu G, Kong X, Yang S. Interface structure and strength of ultrasonic vibration liquid phase bonded joints of Al2 Op3 /6061Al composites. J Scri Mater 2005;53:835–9. [10] Gale WF, Butts DA. Transient liquid phase bonding. Sci Technol Weld Join 2004;9:283–300. [11] Zhou Y, Gale WF, North TH. Modeling of transient liquid phase bonding. Int Mater Rev 1995;40:181–96. [12] Macdonald WD, Eager TW. Transient liquid phase bonding processes minerals. J Met Join Soc 1992:93–101. [13] Tuah-Poku I, Dollar M, Massalski TB. A study of the transient liquid phase bonding process applied to a Ag/Cu/Ag sandwich joint. J Trans A 1988;19A: 675–86. [14] Gu XY, Sun DQ, Liu L, Duan ZZ. Microstructure and mechanical properties of transient liquid phase bonded TiCp/AZ91D joints using copper interlayer. J Alloys Compd 2008;476:492–9. [15] Gu XY, Sun DQ, Liu L. Transient liquid phase bonding of TiC reinforced magnesium metal matrix composites (TiCp/AZ91D) using aluminum interlayer. J Mater Sci Eng A 2008;487:86–92. [16] Pouranvari M, Ekrami A, Kokabi AH. Effect of bonding temperature on microstructure development during TLP bonding of a nickel base superalloy. J Alloys Compd 2009;469:270–5.

282

S.S. Sayyedain et al. / Materials and Design 53 (2014) 275–282

[17] Huang J, Dong Y, Zhang J, Wan Y, Zhou G. Reactive diffusion bonding of SiCp/Al composites by insert powder layers with eutectic composition. J Mater Sci Technol 2005;5:21–5. [18] Shirzadi AA, Wallach ER. New approaches for transient liquid phase diffusion bonding of aluminum based metal–matrix composites. J Mater Sci Technol 1997;13:135–42. [19] Shirzadi AA, Assadi H, Wallach ER. Interface evolution and bond strength when diffusion bonding materials with stable oxide films. J Surf Interf Anal 2001;31: 609–18.

[20] Padron T, Khan TI, Kabir MJ. Modeling the transient liquid phase bonding behavior of a duplex stainless steel using copper interlayers. J Mater Sci Eng A 2004;385:220–8. [21] Nami H, Halvaee A. Transient liquid phase diffusion bonding of Al/Mg2Si metal matrix composite. J Mater Des 2011;32:3957–65.