Fracture Variation of Welded Joints at Various Temperatures in Liquid-Phase-Pulse-Impact Diffusion Welding of Particle Reinforcement Aluminum Matrix Composites

Abstract

The fracture variation of liquid-phase-pulse-impact diffusion welding (LPPIDW) welded joints of aluminum matrix composites (ACMs: SiCp/A356, SiCp/6061Al, and Al2O3p/6061Al) was investigated. Results show that under the effect of pulse-impact (i) initial pernicious contact state of reinforcement particles changes from reinforcement (SiC, Al2O3)/reinforcement (SiC, Al2O3) to reinforcement (SiC, Al2O3)/matrix/reinforcement (SiC, Al2O3) and (ii) the fracture of welded joints with optimal processing parameters is the dimple fracture. Meanwhile, scanning electron microscope (SEM) of the fracture surface shows some reinforcement particles (SiC, Al2O3) in the dimples. Moreover, the slight reaction occurs at the interfaces of SiCp/6061Al, which is propitious to improve the property of welded joints because of the release of internal stress caused by the hetero-matches between the reinforcements and matrix. Consequently, aluminum matrix composites (SiCp/A356, SiCp/6061Al, and Al2O3p/6061Al) were welded successfully.

1. Introduction

Aluminum matrix composites (AMCs) have a wide application in the fields of aerospace, automobile, structural components, heat-resistant-wearable parts in engines, and so on due to their high specific strength, rigidity, wear resistance, corrosion, and good dimensional stability [1–5]. The reinforcements in AMCs may be either in the form of particulates or as short fibers, whiskers, etc. [5, 6]. These discontinuous natures create several problems to their joining techniques for acquiring their high strength and good quality welded joints.

The high specific strength, good wearability, and corrosion resistance of aluminum matrix composites (AMCs) attract substantial industrial applications. Typically, AMCs are currently used widely in automobile and aerospace industries, structural components, heat-resistant-wearable parts in engines, etc. [7–10]. The particles of reinforcement elements in AMCs may be either in the form of particulates or as short fibers, whiskers, and so forth [10, 11]. These discontinuous natures create several problems to their joining techniques for acquiring their high strength and good quality welded joints. Typical quality problems of those welding techniques currently available for joining AMCs [12–20] are as elaborated below.

(1) The distribution of particulate reinforcements in the weld.

As properties of welded joints are usually influenced directly by the distribution of particulate reinforcements in the weld, their uniform distribution in the weld is likely to give tensile strength higher than 70–80% of the parent AMCs. Conglomeration distribution or the absence (viz., no reinforcement zone) of the particulate reinforcements in the weld generally degrades markedly the joint properties and subsequently resulted in the failure of welding.

(2) The interface between the particulate reinforcements and aluminum matrix.

High welding temperature in the fusion welding methods (typically TIG, laser welding, electron beam, etc.) is likely to yield pernicious Al₄C₃ phase in the interface. Long welding time (e.g., several days in certain occasions) in the solid-state welding methods (such as diffusion welding) normally leads to (i) low efficiency and (ii) formation of harmful and brittle intermetallic compounds in the interface.

To alleviate these problems incurred by the available welding processes for welding AMCs, a liquid-phase-pulse-impact diffusion welding (LPPIDW) technique has been developed [21–23]. This work aims at providing some specific studies that influence the pulse-impact on the fractography variation of welded joints. Analysis by means of scanning electron microscope (SEM), transmission electron microscope (TEM), and X-ray diffraction (XRD) allows the micro-viewpoint of the effect of pulse-impact on LPPIDW to be explored in more detail. Also, temperature distribution in heated sample also calculates.

2. Experimental material and procedure

2.1. Material

Stir-cast SiC_p/A356, P/M SiC_p/6061Al, and Al₂O_3p/6061Al aluminum matrix composite, reinforced with 20%, 15% volume fraction SiC, Al₂O₃ particulate of 12 μm, 5 μm mean size, were illustrated in Figure 1.

Figure 1.

Microstructure of aluminum matrix composites.

2.2. Experimental procedure

The quench-hardened layer and oxides induced by wire-cut process, on the surfaces of aluminum matrix composite specimens, were removed by polishing with 400 # grinding paper carefully. The polished specimens were then properly cleaned by acetone and pure ethyl alcohol so as to remove any contaminants off its surfaces. A DSI Gleeble^®-1500D thermal/mechanical simulator with a 4 × 10⁻¹ Pa vacuum chamber was subsequently used to perform the welding.

The microstructures and the interface between the reinforcement particle and the matrix of the welded joints were analyzed by SEM, TEM, and XRD.

2.3. Operation of LPPIDW

Figure 2 shows a typical temperature and welding time cycle of a LPPIDW. It basically involves with (i) an initially rapid increase of weld specimens, within a time of t_a, to an optimal temperature T_a at which heat was preserved constantly at T_a for a period of (t_b-t_a); (ii) at time t_c, a quick application of pulse-impact to compress the welding specimens to accomplish an anticipated deformation δ within 10⁻⁴–10⁻² s, whilst the heat preservation was still maintained at the operational temperature T_a; and (iii) a period of natural cooling to room temperature after time t_b.

Figure 2.

Schematic diagram of liquid-phase-pulse-impact diffusion welding.

3. Results and discussion

3.1. Temperature distribution calculation in heated sample

The schematic diagram of temperature distribution in heated sample is shown in Figure 3, where L is the length of sample between two holders.

Figure 3.

Schematic diagram of temperature distribution in heated sample.

T₀ is the initial temperature. When the sample is heated, the heat transfer at the distance of x in Δt is

$Q=-\mathit{\alpha A}\frac{\partial T}{\partial x}\mathit{\Delta t}$ E1

where α is the coefficient of conduction and A is the cross-sectional area.

At the same time, the heat transfer at the distance of (x + Δx) is

$Q+\mathit{\Delta Q}=-\mathit{\alpha A}\left[\frac{\partial T}{\partial x}+\frac{\partial }{\partial x}\left(\frac{\partial T}{\partial x}\right)\mathit{\Delta x}\right]\mathit{\Delta t}$ E2

Supposedly, the heat input into the sample is W per volume unit, and the loss due to irradiation and convection is neglected.

Therefore, the energy obtained in per length at Δx is

$\mathit{\Delta Q}+\mathit{\text{WAΔxΔt}}=\mathit{\alpha A}\frac{{\partial }^{2}T}{\partial x^2}\mathit{\text{ΔxΔt}}$ E3

Consequently, the rate of temperature variation is

$\frac{\partial T}{\partial x}=\frac{1}{\mathit{\rho c}}\left(W+\alpha \frac{{\partial }^{2}T}{\partial x^2}\right)$ E4

where ρ is the density of sample and c is the specific heat capacity.

When the processing is stable, $\frac{\text{∂}T}{\text{∂}x}=0$ , then

$W=-\alpha \frac{{\partial }^{2}T}{\partial x^2}$ E5

After integration,

$\frac{\partial T}{\partial x}=-\frac{W}{\alpha }x+b$ E6

When $x=\frac{1}{2}L$ , $\frac{\text{∂}T}{\text{∂}x}=0$ , it obtains

$b=\frac{\mathit{WL}}{2\alpha }$ E7

Integratedly, then

$T=-\frac{W}{2\alpha }{x}^2+\frac{\mathit{WL}}{2\alpha }x+b^{\prime }$ E8

At x = 0 and T = T₀, so b^′ = T₀; therefore,

$T=-\frac{W}{2\alpha }{x}^2+\frac{\mathit{WL}}{2\alpha }x+T_0$ E9

or

$T-T_0=\frac{W}{2\alpha }\left(L-x\right)x$ E10

At $x=\frac{1}{2}L$ and T = T_max, so

$T_{\max }-T_0=\frac{{\mathit{WL}}^2}{8\alpha }$ E11

Assumption: T = T_max − ΔT, then

$T_{\max }-\mathit{\Delta T}-T_0=4\frac{\left(T_{\max }-T_0\right)}{L^2}\left(L-x\right)x$ E12

Therefore,

$x=\frac{L}{2}\left[1\pm \sqrt{\frac{\mathit{\Delta T}}{T_{\max }-T_0}}\right]$ E13

As a result, the length between T_max and T_max − ΔT is

$\mathit{\Delta x}=L\sqrt{\frac{\mathit{\Delta T}}{T_{\max }-T_0}}$ E14

According to Eq. (14), it reveals that with the increment of temperature difference, the area between the solid phase and liquid phase increases simultaneously. However, if the temperature is too high, it will make the welding failure because the area between the solid phase and liquid phase enlarges. As a result, when the samples are impacted, the relative sliding of samples occur [21–23]. Meanwhile, the grain size and microstructure of AMCs will be too larger and coarser to decrease the property of the welded joints.

Furthermore, considering the practical situations during the welding process, the resistance between the two welded pieces due to the heterogeneous materials at the welding area is definitely higher than that of the calculation on the basis of the theory. Consequently, Eq. (14) will be

$\mathit{\Delta x}=\mathit{\eta L}\sqrt{\frac{\mathit{\Delta T}}{T_{\max }-T_0}}$ E15

where η is the coefficient of the influence of hetero-resistance at the welding area.

3.2. Microstructure of welded joint at various temperatures

The fractographs of SiC_p/A356 are shown in Figure 4. It illustrates that when the welding temperature is 563°C, the initial morphology of substrate is still obviously detected, and some sporadic welded locations appear together with some rather densely scattering bare reinforcement particles as shown in Figure 4a. With the welding temperature increasing to 565°C, more liquid phases form. Under the effect of pulse-impact, some wet locations in the joint excellently weld, and the condition of the aggregated solid reinforcement particles is improved. However, the bare reinforcement particles still distribute on the fractographic surface. It indicates that substrates do not weld ideally and it consequently results in a low-strength joint (Figure 4b).

Figure 4.

Fractographs of SiC_p/A356 at various temperatures. (a) 563°C, (b) 565°C, (c) 570°C and (d) 575°C.

Figure 4c shows the fractograph of welded joint at 570^°C. It illustrates that the fracture is dimple fracture. Moreover, SEM image of the fracture surface shows some reinforcement particles (SiC) in the dimple. In order to confirm the state of these reinforcement particles, particles itself and matrix near to these particles were analyzed by energy-dispersive X-ray analysis (EDX). The result is expressed in Figure 5. It indicates that reinforcement particles (SiC) are wet by matrix alloy successfully suggesting that the reinforcement particles have been perfectly wet and the composite structure of reinforcement/reinforcement has been changed to the state of reinforcement/matrix /reinforcement.

Figure 5.

Energy-dispersive X-ray analysis of the fracture surface of SiC_p/A356.

As welding temperature increases to 575^°C, it leads to more and more liquid-phase matrix alloy distributed in the welded interface. Meanwhile, more liquid-phase matrix alloy reduces the effect of impact on the interface of the welded joints; subsequently, the application of transient pulse-impact causes the relative sliding of the weldpieces that jeopardizes ultimately the formation of proper joint as shown in Figure 4d.

The relevant fractographies of SiC_p/6061Al and Al₂O_3p/6061Al at various welding temperatures are shown from Figures 6 to 7. It shows that the fracture surfaces under the effect of pulse-impact are similar to that of SiC_p/A356. The fractures are all dimple fractures with some reinforcement particles (SiC, Al₂O₃) in the dimple.

Figure 6.

Fractographs of SiC_p/6061Al at various temperatures. (a) 620°C, (b) 623°C and (c) 625°C.

Figure 7.

Fractographs of Al₂O_3p/6061Al at various temperatures. (a) 641°C, (b) 644°C and (c) 647°C.

SEM results of the fracture surface show that the reinforcement particles have been perfectly wet and the composite structure of reinforcement/reinforcement has been changed to the state of reinforcement/matrix /reinforcement. XRD pattern of the fracture surfaces (Figure 8) does not show the existence of any harmful phase or brittle phase of Al₄C₃. This suggests the effective interface transfers between reinforcement particles and matrix in the welded joint that subsequently provides favorable welding strength [16–18].

Figure 8.

XRD pattern of the fracture surfaces. (a) SiC_p/A356, (b) SiC_p/6061Al and (c) Al₂O_3p/6061Al.

3.3. Effect of pulse-impact on SiCp/6061Al LPPIDW on the interface reaction

During LPPIDW, SiC_p/6061Al welding temperature is so high that the reaction between SiC particles and aluminum matrix occurs as follows:

$\text{3SiC}\left(\mathrm{s}\right)+\text{4Al}\left(\mathrm{l}\right)\to {\mathrm{Al}}_4{\mathrm{C}}_3\left(\mathrm{s}\right)+\text{3Si}\left(\mathrm{s}\right)$ E16

The relevant free energy [24] is

$\mathrm{\Delta G}\left(\mathrm{J}·{\mathrm{mol}}^{-1}\right)=113900-12.06\text{TlnT}+8.92\times {10}^{-\text{3}}{T}^2+7.53\times {10}^{-4}{T}^{-1}+21.5T+3\mathit{RT}\text{ln}{\alpha }_{\left[\mathrm{Si}\right]}$ E17

where α_[Si] Si is the activity in the liquid of aluminum.

In the viewpoint of thermodynamics, ΔG > 0 during LPPIDW SiC_p/6061Al. Therefore, the reaction of Eq. (16) does not occur. However, in accordance with binary alloy-phase diagrams of Al-Si as shown in Figure 9 [11], Si will dissolve in the aluminum matrix during the welding, and the activity of Si in the liquid aluminum varies at various temperatures. The relationship between free energy and welding temperature in the interface after considering Si dissolution in the aluminum matrix is shown in Figure 10, which indicts that Eq. (16) will occur in the welding.

Figure 9.

Binary alloy phase diagrams of al-Si [25].

Figure 10.

Relationship between free energy and welding temperature in the interface.

It is well known that the interfaces between the reinforcement (particles) and matrix are playing the extremely vital role in AMCs, especially in adjusting their matches. The slight reaction at the interfaces is propitious to improve the property of welded joints due to the release of internal stress, which is caused by the hetero-matches between the reinforcements and matrix. The distribution of dislocations at the interface is shown in Figure 11. It illustrates that the density of dislocations decreases remarkably, especially compared with that of its nearing area, which elucidates that the effective reaction at the interface occurs and releases the internal stress due to the hetero-matches between the reinforcement/particle (SiC) and matrix. Consequently, the property of welded joints improves compellingly and convincingly.

Figure 11.

Distribution of dislocations at the interface.

Also, the results shown in Figure 6 of SiC_p/6061Al are better than that of Al₂O_3p/6061Al shown in Figure 7 just due to this interfacial reaction between the reinforcement and matrix, which releases the thermal mismatch stress to an acceptable extent between the reinforcement and matrix to allow load transfer from the matrix to reinforcement successfully. As a result, it has advantageous effect of improving the strength of welded joints further [18].

4. Conclusion

The fractography results of liquid-phase-pulse-impact diffusion welding of particle reinforcement aluminum matrix composites (SiC_p/A356, SiC_p/6061Al, and Al₂O_3p/6061Al) show that:

(1) The area of the solid-phase and liquid-phase coexistence in the welding can be calculated by $\mathit{\Delta x}=\mathit{\eta L}\sqrt{\frac{\mathit{\Delta T}}{T_{\max }-T_0}}$

(2) The fracture of welded joints with optimal processing parameters is the dimple fracture with some reinforcement particles (SiC, Al₂O₃) in the dimples.

(3) Distinctly clear interface between reinforcement particle and matrix overcomes some diffusion problems normally encountered in conventional diffusion welding and prevents the formation of harmful microstructure or brittle phase in the welded joint.

(4) The slight reaction occurs at the interfaces of SiC_p/6061Al, which is propitious to improve the property of welded joints because of the release of internal stress caused by the hetero-matches between the reinforcements and matrix.