Microstructure and mechanical properties of Al-Cu alloy during wire and arc additive manufacturing by adding micron TiB₂ particles

ABSTRACT

Micron TiB₂ particles were added to the 2319 aluminum alloy using a surface coating method during the additive process. The study focused on the formation, microstructure, and properties in order to achieve deposited components with a uniform microstructure and enhancing strength and plasticity. The critical migration velocity (V_cr= 335 μm/s) of with micron-sized TiB₂ particles addition was larger than the solidification rate, which caused particles to be pushed by the solid-liquid interface and ultimately reside at the grain boundary position. The micron-sized TiB₂ particles reduced the activation energy for nucleation of the precipitation of the θ’ phase, lowered the energy barrier for nucleation, and introduced dislocations to create lattice distortion channels. This enabled rapid diffusion of solute atoms and facilitated the precipitation and growth of the θ’ phase. Finally, the transverse strength and elongation were higher, reaching 339.07 MPa and 21.3%, which increased by 45% and 60% respectively.

KEYWORDS:

• 2319Al
• WAAM
• TiB2 particles
• θ' phase

1. Introduction

Al-Cu alloy is widely used in aerospace and other fields due to its high specific strength, specific modulus, good fatigue performance, and corrosion resistance [1, 2]. With the rapid development of the aerospace industry, there is a growing demand for large-scale and integrated aluminum alloy manufacturing. Wire and arc additive manufacturing (WAAM) has been widely used in the manufacturing of large-sized and complex-shaped aluminum alloy additive components [3]. Due to defects such as uneven structure and pores in the process of WAAM, the directly deposited state of aluminum alloy additive components exhibits performance degradation and anisotropy [2–6]. These defects significantly limit the usability of such components.

Several researchers have improved the strength and plasticity of aluminum alloy additive components through various methods. The most common process for enhancing the performance of heat-treatable Al alloys through additional components is heat treatment. Gu et al. [7] found that the mechanical properties of the entire Cold Metal Transfer (CMT) WAAM 2319Al alloy, including the substrate, exhibited consistent characteristics after undergoing the T6 heat treatment process. The average ultimate tensile strength (UTS) exceeded 430 MPa, representing a 65% increase, while the plasticity remained comparable to that of the substrate. Bai et al. [8] proposed a long-term solution for the aging process. They observed that the average tensile strength increased to 391 MPa. However, pore defects resulted in a disparity between the transverse and longitudinal tensile strength. Gu et al. [9] studied the microstructure and mechanical properties of as-deposited and heat-treated WAAM alloys during the preparation of an Al-Cu-Mg aluminum alloy using CMT additive manufacturing with a specific Al-Cu-Mg welding wire. The average grain size increased by 10%. Additionally, the average diameter, sphericity, and volume fraction of micropores increased by 9%, 7%, and 11%. Moreover, the average yield strength (YS) and UTS increased by 116% and 66% respectively. However, the elongation decreased. Several scholars have utilised mechanical processing methods as a supplementary technique to effectively eliminate pores and alter the microstructure in the WAAM process. The method of interlayer hammering was used in the WAAM process of 2319Al alloy, as described by Fang et al. [10]. The slender sub-grains were obtained through hammering, resulting in a 60.6% increase in yield strength and a 13% increase in tensile strength. Gu et al. [11, 12] used interlayer rolling to assist CMT-WAAM of 2319 Al alloy. As the interlayer rolling load increased, the grains became elongated and refined, and the pores were effectively eliminated. After subjecting the sample to a rolling force of 45 kN, the strength of the sample increased by 19.4% and the microhardness increased by 49.3%. The use of auxiliary processes has different effects on the performance of additive components. It can refine grains, improve the uniformity of the microstructure, and enhance the yield strength and ultimate tensile strength. However, it may impair the plasticity and toughness of materials.

In the process of laser additive manufacturing of aluminum alloy, many scholars have mixed Al alloy particles with reinforcements to prepare composite materials, resulting in improved strength and plasticity of deposited samples [13, 14]. Ceramic particles, such as TiB₂ [15–18], TiC [19, 20], SiC [21, 22], and Al₂O₃ [23], are commonly used as reinforcements in the fabrication of Al matrix composites. Li et al. [17] used AlSi₁₀Mg alloy powder, which was mixed with nano-TiB₂, to perform Laser selective melting (SLM). The columnar crystals in the microstructure disappeared, and submicron-sized grains formed. Nano-scale Si was evenly distributed within the grains, and the interface was strongly bonded, thereby providing a reinforcing effect as a second phase. The tensile strength was 530 MPa, the elongation was 15.5%, and the microhardness was 191 HV. Li et al. [14] used SLM to prepare Al-Si-Mg-Ti/TiC material. Benefiting from the role of non-uniform nucleation and limited particle growth, the microstructure consisted mainly of fine equiaxed grains. The ultimate tensile strength (UTS) increased by 48%, and the elongation at break (El) increased by 109%. Ceramic particles were also used as reinforcements in the arc welding process of aluminum alloys. Sokoluk [24] used a pre-prepared TiC/7075 aluminum alloy filler rod to weld the 7075Al alloy. The hot cracking tendency of the welded joint was reduced. The typical second phase Mg (Zn, Cu, Al)₂ was finer and more randomly distributed. The number and size of eutectic structures were also significantly reduced. The strength of the nano-treated joints also increased significantly. After the heat treatment, the joint strength reached 551 MPa, which is 93% of the strength of the forged AA7075-T6 sheet. Additionally, the elongation increased to 5.21%. The addition of particles alters the grain growth mode, resulting in refined grains and a homogenised structure that eliminates defects such as cracks. This, in turn, enhances the strength of joints and additive parts. In the process of solidification, the particles may either be engulfed or repelled by the growing grains [25–31]. Additionally, the particles have the effect of refining the grains and promoting the precipitation of the second phase in the molten pool [32, 33]. The performance improvement of particle-reinforced composites is influenced by the distribution and effect of particles. Dong et al. [34] prepared TiB₂ reinforced Al-Si-Cu alloy by casting. It is found that the addition of TiB₂ particles can reduce the precipitation temperature of the Al2Cu phase, promote the rapid diffusion of copper element, and promote the precipitation of θ’ phase. θ’ phase is a critical strengthening phase of Al-Cu alloy, and its large amount of precipitation provides reliable strength for applying alloy components [35].

As such, in this work, TiB₂ particles with high thermal stability were selected as reinforcements and added during the WAAM process of the 2319 aluminum alloy. The effects of particle content on additive samples’ formation, microstructure, and properties were investigated. The distribution of particles, the effect of particles on grain refinement, and θ’ phase precipitation were studied. Additionally, the change in performance was examined in conjunction with the existing strengthening mechanism.

2. Experimental procedure

The substrate material used 1060 Al alloy in the experiment. The length and thickness of the substrate were 150 and 8 mm, respectively. The Al-Cu alloy wire (ER2319) was used with a diameter of 1.2 mm. Table 1 shows the chemical composition of the 1060 Al alloy and Al-Cu alloy wire. The substrate was treated by mechanical grinding to remove the oxide film and oil from the surface before additive manufacturing. Micron TiB₂ particle was chosen as the filling material because of its excellent performance, such as extremely high hardness, specific strength, high thermal stability, and low thermal expansion coefficient. The morphology and size distribution of micron TiB₂ particles using scanning electron microscopy (SEM) are shown in Figure 1. Image J software was used to calculate the size of TiB₂ particles. TiB₂ particles were mainly bulk, as shown in Figure 1 (a). The average size of micron TiB₂ particles was about 7.12 μm, as shown in Figure 1 (b). Micron TiB₂ particles were evenly dispersed into ethanol via ultrasonic vibration before adding to prepare an ethanol-based paint. Different qualities of micron TiB₂ particles were coated onto the surface of the deposited layer. To ensure the percentage of added particles, a balance was used to weigh the additive sample before and after particle addition. The mass fraction was controlled by the coating speed to make that TiB₂ particles uniformly coat on the aluminum alloy surface. The weight percentage of painted micron TiB₂ particles ranged from 0.3–1.2 wt.%. Taking the pure argon (Ar) gas as protected atmosphere with a gas flow rate of 15 L/min during the additive manufacturing process.

Figure 1. SEM image and size distribution of the TiB₂ particles and schematic diagrams of the device (a) SEM image and (b) size distribution of the TiB₂ particles, (c) Schematic diagrams of the additive manufacturing device, (d) sampling locations and tensile sample sizes.

Table 1. Chemical compositions of the1060 Al alloy and Al-Cu alloy wire (wt. %).

Material	Mg	Mn	Si	Cu	Ti	Fe	Al
1060Al	0.03	0.03	0.25	0.05	0.03	0.35	Bal.
ER2319	≤0.2	0.2-0.4	≤0.2	5.8-6.8	0.1-0.2	≤0.3	Bal.

The additive manufacturing equipment was a self-built system based on a TIG power source (Magic Wave 5000 Job G/F), as shown in Figure 1(c). It consists of four main parts, which are welding power source, wire feeding system, self-built welding platform moving system and fixture. The testing parameters were that the deposition current (110 A), the distance between the tungsten electrode and the deposited layer (5 mm), wire feeding speed (1.9 m/min) and deposition speed (170 mm/min), cooling time (about 10 min), respectively. A long cooling time was used to the deposited sample could be cooled to room temperature.

After the additive manufacturing process, the metallurgical and tensile samples were cut to observe microstructure and testing performance, which sampling locations and sizes are shown in Figure 1 (d). The inlaid Metallographic samples with 50 mm length and 10 mm thickness were ground and mechanical polished, and then were etched with Keller reagent (HNO₃: HCl: HF: H₂O = 2.5: 1.5: 1: 95) for 10–15 s. A DSX510 optical microscope (OM) was used to observe the metallographic microstructure of the deposited sample. Microstructures and fractures were observed by field emission scanning electron microscopy (SEM, MERLIN Compact) in the secondary electron mode, and the distribution and chemical compositions of precipitated phases analysis were performed by energy dispersive spectrometer (EDS) using point and line scans. The electron backscatter diffraction (EBSD) system was used to analyze the texture and grain size of the deposited samples. The polished samples were electro-polish with perchloric acid ethanol solution (HClO₄: C₂H₅OH = 1:9) for map scanning. The EBSD testing step was set to 4 μm. The X and Y dimensions were 1380 and 1046.16 μm. The operating voltage for SEM and EBSD was set to 20 kV. The OIM software was used to analyze the EBSD raw data. The microstructure and phase analysis were conducted using a transmission electron microscope (TEM) of JOEL-2100. The samples were made into discs with a diameter of 3 mm by punching mechanism and ground to a thickness of about 20 μm. The samples were thinned by the ion thinning machine (Gatan) to obtain the desired thin zone. The Gatan Digital Micrograph software was utilised to process the TEM data. The tensile sample size was designed according to GB/T228.1-2010 and tensile tests were performed at room temperature with a tensile rate of 2 mm/min, using an Instron 5967 30 kN universal material test machine. To reduce the error, three repeated tests were performed for each deposited sample, and the strength and elongation were separately calculated.

3. Results

3.1. Macro forming

The macroscopic forming of single-channel single-layer as-deposited 2319 aluminum alloy with 0.3 wt.%−1.5 wt.% TiB₂ particle addition was displayed in Figure 2. With the increase in TiB₂ particles content, the wettability between the aluminum alloy and the substrate gradually decreased, and the weld formation began to deteriorate. When the particle content was lower than 1.5 wt%, the surface formed well without obvious defects. However, when the particle content reached 1.5 wt%, the particles reduced the wettability between the substrate and the deposited metal, and the surface of the deposited layer appeared a discontinuity hump, which hindered subsequent deposition. Therefore, the content of particles should not be more than 1.2 wt.% to obtain a better weld shape and reduce the performance degradation caused by defects. Figure 3 shows the macroscopic forming of single-channel multi-layer as-deposited 2319 Al thin-walled parts with micron TiB₂ addition of 0.3–1.2 wt.%. With the increase in the particle content, the wettability of the liquid Al-Cu alloy to the deposited layer became worse due to the particles, which hindered the transition and deposition of droplets in the subsequent process. Finally, the deposited layer appeared slight snake-shaped weld defects.

Figure 2. Macroscopic forming images of single-channel single-layer as-deposited 2319 aluminum alloy with different contents of micron TiB₂ addition 0.3 wt.%−1.5 wt.%.

Figure 3. Macroscopic forming images of single-channel multi-layer as-deposited 2319 Al alloy thin-walled parts with different contents of micron TiB₂ addition (a) 0.3 wt.%, (b) 0.6 wt.%, (c) 0.9 wt.%, (d) 1.2 wt.%.

3.2. Grain morphology

Figure 4 shows the metallographic structure of the deposited sample under an optical microscope with 0.3-1.2wt.% micron TiB₂ particles. According to previous studies [18, 36–38], the microstructure without TiB₂ particles was divided into two regions: inner-layer and inter-layer. The coarse columnar grains and coarse equiaxed grains existed in the inner-layer, while the inter-layer microstructure consisted of mainly fine equiaxed grains and coarse dendrites. With the addition of 0.3wt. % TiB₂ particles, the grain morphology in the inner-layer structure changed from columnar to equiaxed, and the fine dendrites at the inter-layer position replaced the coarse dendrites, and the grains were refined, as shown in Figure 4 (a). With the further increase in particle content, the proportion of dendrites in the interlayer region gradually decreased, as shown in Figure 4 (b-c). When the particle addition ratio reached 1.2wt.%, the uniformity of the microstructure was significantly improved, and the thickness of the dendrite region was reduced to several grain sizes, as shown in Figure 4 (d).

Figure 4. OM images of deposited 2319Al microstructure with micron TiB₂ particles: (a-d) 0.3, 0.6, 0.9, and 1.2 wt.% micron TiB₂ particles.

In order to further study the effect of the content of micro TiB₂ particles on the grain size and grain morphology of the microstructure, EBSD analysis was used. Figure 5 shows the EBSD images of the deposited samples with 0.3-1.2wt.% micron TiB₂ particles. When 0.3 wt.% of TiB₂ particles were added, the coarse columnar grains were eliminated, the equiaxed grains were also refined, and the uniformity of the microstructure was improved. However, there was still a fine equiaxed grain band, and the grain boundaries between the grains were primarily composed of high-angle grain boundaries. The grains grow in different directions, which hindered the directional growth trend of columnar grains, as shown in Figure 5 (a). With the increase in particle content, more particles were introduced as sites for heterogeneous nucleation, which led to an increase in the number of nucleation events, resulting in enhanced grain refinement. As a result, the coarse columnar crystals in the microstructure were eliminated. The microstructure consisted of uniform and fine equiaxed grains, as shown in Figure 5 (b-d). Figure 5 (a1-d1) display the grain size distribution of the deposited samples with the addition of 0.3-1.2 wt.% micron TiB₂ particles. After adding TiB₂ particles, the grains are refined. With the increase in particle content, the average grain size gradually decreased. When the particle content increased to 1.2 wt.%, the grain size decreased to 44.07 μm. Compared with the grain size without the addition of particles [18, 36–38], the grain size decreased by 54.33%. Figure 5 (a2-d2) display the pole figure of the aluminum matrix and CuAl₂ in 2319 aluminum alloy deposited samples with 0.3–1.2wt.% TiB₂ particles. After incorporating micron-sized TiB₂ particles, the Al matrix still exhibited a clear orientation along the (001) and (111) basal planes. The maximum pole density gradually decreased with the increase in TiB₂ particle content. After adding 1.2 wt.% TiB₂ particles, the maximum pole density decreased to 1.254. Additionally, the texture features became more uniform not only along the deposition direction (A2) but also along the direction perpendicular to the deposition direction (A1). As a result, the texture characteristics were compromised.

Figure 5. EBSD diagram of 2319Al deposited samples microstructure：(a-d) inverse pole figure with 0.3-1.2 wt.% TiB₂ particles, (a1-d1) average grain size with 0.3-1.2 wt.% TiB₂ particles, (a2-d2) PF of matrix Al with 0.3-1.2 wt.% TiB₂ particles.

Figure 6 shows the microstructure of the equiaxed region in the deposited sample layer with a weight percentage of 0.3-1.2 micron-sized TiB₂ particles. Figure 7 displays the results of line scanning for the grain boundary position in the equiaxed region. Compared to the microstructure of the additive sample without TiB₂ particles, the proportion of Cu element atoms in the grain boundary region of the equiaxed grain was significantly reduced, resulting in a notable enhancement in segregation. When the content of TiB₂ particles was between 0.3-0.9 wt.%. The atomic ratio of Cu element at the grain boundary was between 10–12 at.%. The microstructure near the grain boundary consisted mainly of an Al-Cu eutectic structure with a thin band. The width of the grain boundary ranged from 0.3-1.7 μm, as shown in Figure 7 (a-c). When the addition ratio of TiB₂ particles reached 1.2 wt.%. As shown in Figure 6 (d), the grain boundary contrast appeared in two colours: light and dark. The highest proportion of Cu atoms in the dark grain boundary was only 5.3 at.%. When adding 0.6 wt.% TiB₂ particles, the colour contrast of different grains was not consistent under the same conditions, as shown in Figure 6 (b). Compared with adding of 0.3 wt.%. TiB₂ particles, some grains appeared silver gray, and the colour depth increased closer to the grain boundary, becoming closer to silver-white. Additionally, the higher the addition ratio of TiB₂ particles, the greater the number of silver-gray grains. The EDS line scan from the black grains to the linear region of silver-gray grains is shown in Figure 7 (c). The proportion of solute Cu atoms in the black grains was ranged from 1.0–2.5 at.%. The proportion of solute Cu atoms in the silver-white grains was generally between 2.5-4.0 at.%. When the addition ratio of TiB₂ particles is 1.2 wt.%, as shown in Figure 7 (d), the silver-white area accounted for a larger proportion.

Figure 6. SEM images of deposited 2319Al microstructure with micron TiB₂ particles: (a-d) SEM images with 0.3, 0.6, 0.9, and 1.2 wt.% micron TiB₂ particles.

Figure 7. EDS spectrum analysis of deposited 2319Al microstructure with micron TiB₂ particles: (a-d) EDS spectrum analysis with 0.3, 0.6, 0.9, and 1.2 wt.% micron TiB₂ particles.

3.3. The second phase

To conduct a more detailed analysis of the second phase, high-power scanning electron microscopy (SEM) and transmission electron microscope (TEM) were used. The second phase of the deposited samples was mainly composed of micron-sized θ-CuAl₂ phase and submicron-sized θ'-CuAl₂ phase, as shown in Figure 8. The morphology of θ-CuAl₂ phase is mainly ellipsoidal and rod-shaped, but there are pores in the ellipsoidal θ-CuAl₂ phase, which presents a network-like feature, as shown in Figure 8 (d-e). After the particles were added, a large number of lamellar second phase-θ’ phase precipitated near the grain boundary, with a size of hundreds of nanometres, as shown in Figure 8 (f). The lamellar θ’ phase exhibited a cross-distribution, and its needle-like and disc-like morphology could be observed under the electron microscope due to the limited observation angle, as shown in Figure 8 (a). With the increase in particle content, the number of precipitated lamellar θ’ phases also gradually increased. Inhibited the segregation of the Cu element at the grain boundary and reduced the formation of the grain boundary network phase, as shown in Figure 8 (b-c).

Figure 8. SEM and TEM images of deposited 2319Al microstructure with micron TiB₂ particles: (a) 0.3 wt.%, (b) 0.6 wt.%, (c) 1.2 wt.%, (d) TEM image of network θ-CuAl₂, (e) TEM image of rod-like θ-CuAl₂, (f) TEM image of θ'-CuAl₂.

In order to further study the effect of the content of micro TiB₂ particles on the second phase, EBSD analysis was used. Figure 9 shows the EBSD images of the deposited samples with 0.3-1.2wt.% micron TiB₂ particles. Figure 9 (a1-d1) show the CuAl₂ phase in the microstructure of the additive samples with varying ratios of TiB₂ particles. However, the step size of EBSD was set to 4 μm in order to increase the test speed and observe more grains. Therefore, such the small nano-scale precipitated θ’ phase cannot be recognised by EBSD. In Figure 9, the CuAl₂ phases were only micro-scale spherical precipitates of θ phases. The content of the θ-CuAl₂ phase was 14.4% without particles. However, the proportion of the θ-CuAl₂ phase decreases with an increase in particle content. When adding 1.2 wt.% TiB₂ particles the proportion of the θ-CuAl₂ phase was 3.1%, which had been reduced by 75%. The decrease in θ-CuAl₂ phase content also confirms the increase in θ'-CuAl₂ phase content. The θ-CuAl₂ phase was dispersed throughout the microstructure, but there was an uneven distribution of the θ-CuAl₂ phase in certain localised areas. The θ-CuAl₂ primarily existed in the eutectic or hypoeutectic structure formed with α-Al near the grain boundary and within the grain interior. It was distributed in a dotted or long rod-like pattern, but it was almost nonexistent in dendritic grains. After the addition of TiB₂ particles, the dendritic CuAl₂ phase, which was distributed at the grain boundary, disappeared due to changes in grain morphology and size. Instead, the long rod-like CuAl₂ phase is intermittently distributed near the grain boundary and point-like CuAl₂ phases existed within the grains. After adding micron TiB₂ particles, the maximum pole density of the CuAl₂ phase decreased as the content of TiB₂ particles increased. When the content of TiB₂ particles reached 1.2 wt.%, the maximum pole density of the CuAl₂ phase decreased to 2.942. However, the preferred orientation of low-index surfaces, such as (001), (101), (111), and (112), has not yet been eliminated, and texture characteristics still mainly existed the direction perpendicular to the deposition direction (A1).

Figure 9. EBSD diagram of 2319Al deposited samples microstructure：(a-d) CuAl₂ phase with 0.3-1.2 wt.% TiB₂ particles, (a1-d1) PF of CuAl₂ phase with 0.3-1.2 wt.% TiB₂ particles.

3.4. Mechanical property

The mechanical properties of the deposited samples were tested using a universal material testing machine. The results are shown in Figure 10. When TiB₂ particles were not added, the transverse tensile strength of the deposited sample was 233.6 MPa, and the elongation was 13.3%. The longitudinal tensile strength was 226.96 MPa, and the elongation was 12.7%. With an increase in the content of micron-sized TiB₂ particles, the properties of the deposited samples also improve. After adding 1.2wt.% micron-sized TiB₂ particles, the transverse tensile strength of the deposited sample increased to 339.07 MPa, and the elongation increased to 21.3%, indicating a 45% and 60% increase, respectively. The longitudinal tensile strength was 322.3 MPa, and the elongation was 19.84%, which has increased by 42% and 56%, respectively.

Figure 10. Tensile strength and plasticity of micron TiB₂ particles reinforced specimens: (a) transverse tension, (b) longitudinal tension.

In order to further investigate the strength and plasticity of 2319Al alloy deposited structural parts, the cross sections of the transverse and longitudinal tensile specimens after fracture were observed. Figure 11 shows the SEM images of the transverse tensile fracture of the deposited sample after adding different weight percentages of micron-sized TiB₂ particles. It could be observed that in the absence of particles, as shown in Figure 11 (a-a2), numerous pores and brittle compounds were present in the transverse fracture, and the number of dimples was also small. There was a tendency for brittle fracture. The pore size measured approximately 50 μm or larger, with some reaching up to 100 μm. During stretching, the pores reduced the contact area of the deposited sample, resulting in weak mechanical properties of the material. After the adding of particles, the number of pores was significantly reduced. According to the study of Cong et al [39], a competitive relationship was found between the nucleation of pores and the nucleation of grains. The small grains reduced the distance over which hydrogen elements in liquid metal could diffuse, making it more challenging for nucleation and growth. With adding 0.3 wt.% of micron-sized TiB₂ particles, the pores are primarily distributed in the interlayer position, as shown in Figure 11 (b-b2). However, there were still some pores present in the intralayer region. When the content of micron-sized TiB₂ particles reached 1.2 wt.%, a significant number of uniformly distributed and fine dimples can be observed, along with a small number of pores distributed in the interlayer position, as shown in Figure 11 (c-c2). The number of pores was significantly reduced. As a result, the strength has been significantly improved.

Figure 11. The transverse tensile fracture surface SEM images of the micron TiB₂ adding sample: (a-a1) without TiB₂ particles, (b-b2)0.3 wt.% TiB₂ particles, (c-c2)1.2 wt.% TiB₂ particles.

Figure 12 shows the SEM images of the longitudinal tensile fracture of the 2319Al alloy deposited samples with different weight percentages of micron-sized TiB₂ particles. Without TiB₂ particle addition, the longitudinal tensile sample broken in the interlayer position with a higher concentration of pores, as shown in Figure 12 (a-a2). In addition, there was a significant presence of brittle compounds. After the particles were added, the number of pores decreased significantly, and the size of pores also reduced, as shown in Figure 12 (b-c). The number of pores decreased as the particle content increased, while the dimples exhibited the opposite trend to the pores, with their number increasing as the particle content increased, as shown in Figure 12 (a1-c1). And then, the number of brittle compounds decreased too, with the addition of TiB₂ particles, as shown in Figure 12 (a2-c2).

Figure 12. The longitudinal tensile fracture surface SEM images of the micron TiB₂ adding sample: (a-a1) without TiB₂ particles, (b-b2) 0.3 wt.% TiB₂ particles, (c-c2) 1.2 wt.% TiB₂ particles.

However, there were more longitudinal pores than transverse pores, resulting in a relatively lower longitudinal tensile performance compared to the transverse tensile performance. Furthermore, the disparity in the number of pores between the transverse fracture and the longitudinal fracture was more significant when TiB₂ particles were present than when they are absent. Therefore, the difference between the transverse tensile and longitudinal tensile properties becomes more apparent after adding particles.

4. Discussion

4.1. Particle distribution and grain refinement

For the force analysis of TiB₂ particles, the force analysis model is shown in Figure 13. Firstly, the TiB₂ particles are considered to be regular spheres with a diameter of d_p. During the solidification process, TiB₂ particles move at a speed in the liquid metal due to gravity, Stokes drag force, and buoyancy. At the same time, the growth of the grains causes the solid-liquid interface to move at a growth velocity V, which is known as the solidification velocity. When the velocity of particle movement exceeds the velocity of movement of the solid-liquid interface, the particles will be pushed by the solid-liquid interface. When the velocity of particle movement is slower than the velocity of movement of the solid-liquid interface, the particles are engulfed by the growing solid, potentially resulting in the intragranular distribution of the particles. During the solidification process, the pushing action causes particles to be rejected at the solid-liquid interface and ultimately distributed along the grain boundary. The relative the velocity between the particles in the liquid metal and the solid-liquid interface determines whether the particles are engulfed or repelled by the solid-liquid interface. Therefore, the velocity of particles in liquid metal is defined as the critical migration velocity, V_cr. Previous researchers established a prediction model to forecast the relationship between the critical migration velocity and the particle size in the Al-TiB₂ composite system [25–31]. It has been found that the critical migration velocity is closely related to the particle diameter. In other words, the spatial distribution of the particles is closely linked to their size.

Figure 13. Schematic diagram of force analysis of particles in liquid metal.

Firstly, TiB₂ is influenced by gravity and buoyancy in liquid metal. Both of them are affected by the volume of particles, so the combined force of the two is called the body force F_G, which is expressed in Eq.1 [26, 40]. (1) $F_G={\displaystyle \frac{4}{3}\pi r^3\mathrm{\Delta \rho g}}$(1) Where r is the radius of TiB₂ particles (μm), g is the gravitational acceleration (m/s²), Δρ is the density difference between TiB₂ particles and liquid metal Al-Cu (kg/m³). The direction of this force depends on the sign of Δρ, that is, the relative density of TiB₂ particles compared to the liquid metal Al-Cu.

During the solidification process, the flow of liquid around the particles generates resistance, which hinders the particles from moving away from the solid-liquid interface. Considering the Stokes’ equation in the regime of fluid motion, the drag force F_D could be expressed as Eq.2. (2) $F_D=6\mathrm{\pi \eta rV}$(2) Where η is the viscosity of the liquid Al-Cu alloy (N/m), V is the rate of solidification front (m/s). When particles approach the solid phase, the solid-liquid interface does not exhibit complete flatness and often exhibits some curvature. To account for this, the interface shape factor α is introduced for correction. In many cases, α is considered to be the ratio of the thermal conductivity of the particles k_TiB2 to the thermal conductivity of the melt (k_Al). The modified equation are as follows: Eqs.3 and 4 [26, 40]: (3) $F_D=6\mathrm{\pi \eta V}{\displaystyle \frac{r^2}{d}{\alpha }^2}$(3) (4) $\alpha ={\displaystyle \frac{k_{\mathrm{TiB}2}}{k_{\mathrm{Al}}}}$(4) In the process of particles being engulfed by the solid-liquid interface, the interfacial energy of the solid-liquid interface will inevitably change, resulting in a force F_L that repels the solid-liquid interface and particles. This force can be expressed as Eqs.5-6 [41]. (5) $F_L=2\mathrm{\pi r}\mathrm{\Delta }{\sigma }_0\alpha {\left({\displaystyle \frac{a_0}{a_{0+d}}}\right)}^2{\left({\displaystyle \frac{d_0}{d}}\right)}^4$(5) (6) $\mathrm{\Delta }{\sigma }_0={\sigma }_{\mathrm{ps}}-{\sigma }_{\mathrm{pl}}-{\sigma }_{\mathrm{ls}}$(6) Where a₀ is the atomic spacing (m), Δσ₀ is the interface energy difference (J/m²), σ_ps is the interfacial energy between TiB₂ particles and solid Al (J/m²), σ_pl is the interfacial energy between TiB₂ particles and liquid Al (J/m²), and σ_ls is the interfacial energy between liquid Al and solid Al (J/m²).

The critical migration rate V_cr of solid-liquid interface with micron-size TiB₂ particles in the liquid Al-Cu alloy is solved by balancing the forces(F_G, F_D and F_L) of particles in the liquid metal, which is expressed as follows Eq.7. (7) $V_{\mathrm{cr}}={\displaystyle \frac{\mathrm{\Delta }{\sigma }_0{d}_0{a}_0^2}{3\mathrm{\eta r\alpha }{(a_0+d_0)}^2}-{\displaystyle \frac{2d_0\mathrm{gr}\mathrm{\Delta }\rho }{9\eta {\alpha }^2}}}$(7) The parameters used to calculate V_cr are shown in Table 2. The change in critical migration rate with particle radius is shown in Figure 14. With the increase in TiB₂ particle size, the critical migration rate tends to decrease exponentially. When the radius of TiB₂ particles is in the range of 2–10 μm, the critical migration rate ranges from 100 to 600 μm/s. In this study, micron-sized TiB₂ particles with an average radius of 3.56 μm were selected. After performing calculations, the critical migration rate was determined to be 335 μm/s.

Figure 14. Plots of critical movement rate (V_cr) versus radius of TiB₂ particles (r).

Table 2. The parameters used to calculate V_c [33, 42, 43].

a0，m	Δσ0，J/m^2	η，N/m	d0，m	Δρ，g/cm^3	α
2.0 × 10^−10	0.4	1.1470× 10^−3	1 × 10^−8	1.77	0.374

The migration rate of the solid-liquid interface is calculated as the solidification rate (V) which could be estimated by the grain growth rate. According to previous research [37, 44], it has been found that the solidification rate of 2319 Al alloy would be lower than around 100 μm/s in the WAAM process. It has been found that the critical migration rate of micron-sized TiB₂ particles is significantly higher than the solidification rate of liquid 2319 aluminum alloy. Therefore, the TiB₂ particles will be repelled by the moving solid-liquid interface and distributed at the grain boundary after solidification. The distribution of micron-sized TiB₂ particles in the microstructure was observed using SEM, and the results are shown in Figure 15. It could be observed that the majority of micron-sized TiB₂ particles are located at the grain boundary. This finding aligns with the calculations of the stress model and the critical migration rate V_cr. However, it was still found that a very small amount of micron-sized TiB₂ particles existed within the grain structure. It was speculated that the particles acted as heterogeneous nucleation sites and were quickly enveloped by the newly formed grains during the initial crystallization process.

Figure 15. The images of SEM and EDS point scan results of micron TiB₂ particle distribution: (a-b) 0.3 wt%; (c-d) 1.2wt.%.

As the heterogeneous nucleation phase in α-Al, TiB₂ particles facilitate grain nucleation and reduce the undercooling necessary for grain growth. This makes it easier to meet the energy conditions required for both grain nucleation and growth. At the same time, TiB₂ particles, which are not utilised as nucleation particles, will be influenced by the interplay of repulsive and attractive forces. They will be pushed by the solid-liquid interface, hindering the movement of grain boundaries. This will limit the growth of grains, resulting in the formation of fine and uniformly shaped equiaxed grains.

4.2. Precipitation of θ’ phase

After the addition of TiB₂ particles, the precipitated θ’ phase increased significantly. Therefore, the effect of the added TiB₂ particles on the precipitation of the second phase was investigated. The precipitation of the second phase also involves nucleation and growth processes. When the second-phase particles are uniformly precipitated in the aluminum alloy, the minimum activation energy barrier $\mathrm{\Delta }{G}_{\mathrm{Hom}}^{\ast }$ can be expressed using the following Eq.8 [32, 45]. (8) $\mathrm{\Delta }{G}_{\mathrm{hom}}^{\ast }={\displaystyle \frac{16\pi {\sigma }_{\mathrm{\alpha \beta }}^3}{3{(\mathrm{\Delta }{G}_V-\mathrm{\Delta }{G}_S)}^2}}$(8) Where，σ_αβ and rG_V are the interface energy difference (mJ/m²) and the free energy difference (J) between the Al matrix (α) and the second phase (β), rG_S is the mismatch strain energy (J) of the precipitated phase per unit volume. Similarly, for the heterogeneous precipitation at the TiB₂/Al interface, the equation for the minimum activation energy barrier $\mathrm{\Delta }{G}_{\mathrm{Het}}^{\ast }$ is [32]: (9) $\mathrm{\Delta }{G}_{\mathrm{het}}^{\ast }=-{\displaystyle \frac{4\pi {\sigma }_{\mathrm{\alpha \beta }}^3}{3{(\mathrm{\Delta }{G}_V-G_S)}^2}[2-3\cos \theta +{(\cos \theta )}^3]}$(9) where θ is the contact angle (°) between the interface phase nucleus and the TiB₂ substrate. According to Eqs. 8 and 9, the relationship between $\mathrm{\Delta }{G}_{\mathrm{Hom}}^{\ast }$ and $\mathrm{\Delta }{G}_{\mathrm{Het}}^{\ast }$ can be expressed by the following Eq. 1(1) $F_G={\displaystyle \frac{4}{3}\pi r^3\mathrm{\Delta \rho g}}$(1) 0: (10) ${\displaystyle \frac{\mathrm{\Delta }{G}_{\mathrm{het}}^{\ast }}{\mathrm{\Delta }{G}_{\mathrm{hom}}^{\ast }}={\displaystyle \frac{1}{4}[2-3\cos \theta +{(\cos \theta )}^3]=f(\theta )}}$(10) If the contact angle between the precipitated second phase and the TiB₂ particles is less than 180°, the calculated f(θ) will always be less than 1. This means that the energy barrier for the heterogeneous precipitation of the second phase at the TiB₂/Al interface is lower than that for the homogeneous nucleation of the second phase in the Al matrix.

Previous studies have shown that the precipitation of the second phase in Al-Cu alloys is a diffusion-controlled phase transition, which depends on the diffusion of solute atoms. The addition of TiB₂ particles will cause dislocations in the interface area between TiB₂ and Al, as shown in Figure 16 (a). The lattice distortion near the dislocation core creates a pathway for the rapid diffusion of solute atoms, effectively reducing the activation energy for solute atom diffusion and facilitating the precipitation and growth of the second phase [46].In addition, solute atoms, vacancies, and dislocations at the TiB₂/Al interface promote the formation of interphase precipitates [47]. The enhancement factor of the growth rate of the precipitated second phase is expressed as Eq. 11 [48]. (11) $f=1+{\displaystyle \frac{b\sqrt{2}}{\pi r_{\beta }}\sqrt{f(t^{\mathrm{\prime }}){\displaystyle \frac{\sqrt{D_p}}{\sqrt{D}}}}}$(11) Where D is the volume diffusion coefficient (m²/s) expressed in units, D_p is the diffusion rate of the Cu element along the dislocation core (m²/s), and b is the Berkshire vector (m). If the core width of the dislocation is assumed to be equal 2b, f(t) can be expressed by Eq. 1(1) $F_G={\displaystyle \frac{4}{3}\pi r^3\mathrm{\Delta \rho g}}$(1) 2: (12) $f(t^{\mathrm{\prime }})=\underset{0}{\overset{\mathrm{\infty }}{\int }}{\displaystyle \frac{e^{(-t^{{}^{\mathrm{\prime }}{x}^2})}}{x[J_0^2(x)+Y_0^2(x)]}\mathrm{dx}}$(12) Where J₀(x) and Y₀(x) are the first and second zero-order Bessel functions, respectively. It can be found that the enhancement factor f is closely related to the diffusion rate D_p of the Cu element along the dislocation core and the radius r_β of the second phase nucleus precipitated in the interface region. The larger the D_p is, the larger the f is, and the smaller the r_β is, the larger the f is.

Figure 16. TEM and SEM images of the TiB₂ particle and θ’ phase: (a) TEM image of the TiB₂ particle (a1) electron diffraction patterns for TiB₂ particle, (b) SEM image of the TiB₂ particle and θ’ phase.

On one hand, the introduction of TiB₂ particles reduces the activation energy barrier of the second phase and the size of the precipitated phase nucleus. On the other hand, the addition of TiB₂ particles induces dislocations, which in turn subsequently the lattice distortion and the diffusion rate of the Cu element. The combined effect of the two increases the precipitation rate of the second phase and promotes the precipitation of the θ’ phase. Therefore, many θ’ precipitates can be found near TiB₂ particles and dislocations, as shown in Figure 16 (b).

4.3. Strengthening mechanism and fracture

The microstructure is closely related to the properties. The tensile test results show that the addition of particles improves the strength and plasticity of the material to a certain extent. Therefore, the strengthening mechanism of TiB₂ particle addition is studied in combination with the changes in microstructure. After the addition of TiB₂ particles, the grains are refined to produce fine grain strengthening, which is denoted as Δσ_H-P. Additionally, the dispersion distribution of the particles produces Orowan dispersion strengthening, denoted as Δσ_Orowan. The precipitation strengthening occurs due to a large number of θ’ phases precipitated at the grain boundary, denoted as Δσ_P. Furthermore, the interface connection between the Al matrix and the TiB₂ particles during loading generates load transfer strengthening, denoted as Δσ_Load. The overall strengthening increment of the additive component can be expressed by the following Eq.13 [17, 44]. (13) $\mathrm{\Delta }\sigma =\mathrm{\Delta }{\sigma }_{H-P}+\mathrm{\Delta }{\sigma }_{\mathrm{Orowan}}+\mathrm{\Delta }{\sigma }_P+\mathrm{\Delta }{\sigma }_{\mathrm{Load}}$(13) According to the previous study, after adding TiB₂ particles, the strengthening effect is mainly reflected in the precipitation strengthening. The strength increment of precipitation strengthening Δσ_P is shown in Eq.14 [41, 48]: (14) $\mathrm{\Delta }{\sigma }_P=M{\displaystyle \frac{0.4G_{\alpha }b}{\pi {(1-v)}^{{\displaystyle \frac{1}{2}}}}{\displaystyle \frac{\ln (2\overline{r_{\beta }}/b)}{{\lambda }_{\beta }}}}$(14) Where υ is the Poisson's ratio with a value of 0.33, M is the orientation factor, with a value of 3.06, λ_β is the spacing (m) of the precipitated θ’ phase, and the average radius (m) of the precipitated θ’ phase. The average radius of the θ’ phase precipitated by adding micron-sized TiB₂ is approximately 164.29 nm, and the spacing of the θ’ phase is around 236 nm. The calculated Δσ_P value is about 107.7135 MPa. However, the precipitation of the θ’ phase is not uniform, and the actual strengthening effect due to precipitation should be slightly less effective than the theoretical calculation effect.

After the addition of particles, cracks initiate in the brittle θ-CuAl₂ phase at the grain boundary and within the grain, as shown in Figure 17 (a-b). However, the further growth of the grain boundary cracks is impeded by the precipitation of the θ’ phase near the grain boundary, resulting in a halt in crack propagation. The resistance to crack propagation within the grain is low, and eventually, it penetrates the Al matrix. The cracks in the grain interior and grain boundaries were observed using higher magnification scanning electron microscopy, as shown in Figure 17 (c-d). The grain boundary cracks and intragranular cracks were elongated. However, the trend of crack propagation was not observed, indicating that the precipitation of θ and θ’ phases near the grain boundary hindered the propagation of grain boundary cracks. This phenomenon increases the resistance of crack propagation, requiring more tensile stress and consuming more energy. Additionally, the decrease in grain boundary Cu concentration strengthened the grain boundary. The fracture mode changes from intergranular fracture to transgranular fracture. The strength and plasticity of the material are improved to a certain extent, which confirms the effectiveness of precipitation strengthening after adding particles.

Figure 17. SEM image of the fracture process of the deposited sample: (a) SEM image without TiB₂, (b) SEM image with TiB₂, (c) cracks in the grain interior, (d) cracks in the grain boundaries.

5. Conclusion

The study investigated the influence of TiB₂ particle content on the formation, microstructure, and mechanical properties of the deposited samples. Combined with the analysis of microstructure and mechanical properties, the mechanisms of microstructure changes and strengthening were investigated, leading to the following conclusions:

1. TiB₂ particles hindered the wetting of the liquid metal on the deposition layer. This limited the increase in the concentration of particles that were added.

2. Adding micron-sized TiB₂ particles improved microstructure uniformity in the deposition layer, reducing the number of pores. Grain composition primarily consisted of uniformly sized equiaxed particles. Precipitation of sub-micron lamellar θ’ phases near grain boundaries occurred, inhibiting Cu segregation, and narrowing grain boundary width.

3. The addition of TiB₂ particles reduced the solidification rate, leading to particles distributing at grain boundaries, which hindered grain boundary migration, and inhibited grain growth, and refined grains. TiB₂ particles facilitated the nucleation and growth of the second phase (θ’ phase) by reducing activation energy and introducing dislocations for lattice distortion channels.

4. The primary strengthening mechanism is the precipitation strengthening caused by the large precipitation of θ’ phase. The transverse strength of the additive sample is 339.07 MPa, and the elongation is 21.3%. The longitudinal strength is 322.3 MPa, and the elongation is 19.84%.

Credit author statement

Huisheng Ren: Writing  – original draft, Writing  – review & editing.

YiBo Liu: Funding acquisition, Conceptualisation, Visualisation.

Haoyu Kong: Data curation.

Chenyu Song: Investigation, Methodology.

Yifan Xie: Formal analysis.

Qingjie Sun: Funding acquisition, Project administration.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Additional information

Funding

This work was supported by National Science and Technology Major Project: [Grant Number No.J2019-VII-0004-0144]; The National Natural Science Foundation of China: [Grant Number No.52175299].