AA4043 aluminium wire (1.2 mm diameter) was used as the feedstock material [3][5][22], deposited onto a T-shaped AA6061-T6 substrate (10 mm × 25 mm × 150 mm). Prior to deposition, the substrate surface was cleaned with alkaline water, air-dried, and mechanically abraded to remove oxide layers and ensure adequate bonding of the initial WAAM layer.

The WAAM deposition was conducted using a Gas Metal Arc Welding (GMAW) system with a KUKA robot model KR 4C and EWM 352 Alpha Q welding as heat source, under an argon shielding gas flow (≥98% purity, 20 L/min) (Figure 1).

Figure 1. KUKA robot model KR 4C.

Three process parameters—arc current (80 - 140 A), arc voltage (17 - 20 V), and travel speed (300 - 500 mm/min)—were varied according to a Taguchi L9 orthogonal array generated in Minitab, producing nine parameter combinations (Table 1). Bead height (BH) and bead width (BW) were measured for each sample and jointly optimised using the mean-of-means approach to identify the combination yielding the most consistent and dimensionally favourable bead geometry.

Upon completion of WAAM deposition, a single post-deposition FSW pass was applied across the top surface of the fully-built wall using a Jinfeng Welding and Cutting FSW machine [9]; FSW was not applied between individual deposited layers in this study. A custom aluminium jig (155 mm × 82 mm × 20 mm, secured with four M6 screws) was fabricated via CNC machining to hold the small WAAM samples securely during processing. Two parameter sets—900 RPM/30 mm/min and 1100 RPM/90 mm/min—were trialled; 900 RPM with 30 mm/min travel speed was selected for further characterisation, as higher rotational and travel speeds produced overheating and inconsistent weld quality under the available equipment and operating expertise. Detailed FSW (Figure 2) tool specifications (tool material, shoulder and pin geometry, tilt angle, plunge depth, dwell time, and axial force) and additional WAAM parameters (wire-feed speed, contact-tip-to-work distance, torch angle, deposition path, interpass temperature, and interlayer cooling time) were not systematically logged during this exploratory study; this is acknowledged as a limitation for full reproducibility and is addressed under Future Work.

Figure 2. Jinfeng welding and cutting (FSW).

Samples were sectioned, mechanically ground (240 - 800 grit), polished with diamond lubricant, and etched using Keller’s reagent for microstructural revelation. Field Emission Scanning Electron Microscopy (FE-SEM) was used to examine grain morphology, and Energy Dispersive X-ray Spectroscopy (EDX) was used to map elemental distribution (Al, Si) before and after post-deposition FSW. Vickers microhardness was measured using a 0.1 HV load with a 10 s dwell time at five locations per sample, for both as-deposited WAAM and WAAM + post-deposition FSW conditions.

It should be noted that the five reported hardness values for each condition were obtained from five locations on a single fabricated specimen rather than from independently fabricated replicate specimens; consequently, the reported standard deviation reflects within-specimen spatial variation rather than between-specimen reproducibility, and no formal statistical hypothesis test was performed to compare the two conditions.

The Taguchi L9 results (Table 1) show that bead height and bead width varied non-linearly with arc current, voltage, and travel speed, reflecting the coupled influence of heat input and deposition rate on melt pool behaviour [7][23]. Applying the mean-of-means approach to the dual bead height/width response identified 140 A, 20 V, and 400 mm/min as the optimum WAAM parameter set, corresponding to Sample 9, which produced the most dimensionally consistent bead geometry among the nine trials. For post-deposition FSW, 900 RPM and 30 mm/min travel speed were adopted as the working parameters (Table 2), as this combination provided adequate plastic deformation and heat generation for grain refinement while avoiding overheating and excessive tool wear observed at higher RPM/travel-speed settings [9][24].

Table 1.Taguchi L9 orthogonal array for WAAM parameters and resulting bead geometry.

Table 2.Optimised WAAM and post-deposition FSW process parameters.

FE-SEM examination of the as-deposited WAAM samples revealed coarse columnar α-Al dendrites aligned along the build direction, consistent with the directional solidification and thermal gradients inherent to WAAM’s rapid melting-cooling cycles [2][6]. A Mg2Si-rich eutectic phase was observed segregated along dendrite boundaries, together with isolated microporosity [15][25]—features expected to reduce ductility and produce anisotropic, spatially inconsistent mechanical properties (Figure 3).

Figure 3. FE-SEM micrograph of the as-deposited WAAM AA4043 sample, showing coarse columnar α-Al dendrites, Mg2Si-rich eutectic segregation, and isolated microporosity.

Following post-deposition FSW, the microstructure transformed into fine, equiaxed grains, indicating that the severe plastic deformation and localised frictional heating from the rotating tool promoted dynamic recrystallisation [11]-[13][26]-[28]. The Mg2Si-rich phase was more evenly dispersed after FSW, and no porosity was visually apparent in the stirred zone, consistent with closure of gas pockets and improved interlayer bonding [11][25][29][30] (Figure 4).

EDX mapping (Figure 5 and Figure 6) corroborated these observations: aluminium and silicon distributions were markedly more homogeneous after post-deposition FSW than in the as-deposited condition, supporting the interpretation that the stirring action redistributes alloying elements more uniformly across the matrix [12].

Vickers microhardness measurements (Figure 7, Figure 8 and Table 3) showed a mean hardness of 66.3 HV (range 61.8 - 74.5 HV) for the as-deposited WAAM condition, consistent with the coarse-grained, porosity-containing microstructure. After post-deposition FSW, mean hardness increased to 95.3 HV (range 81.7 - 116.3 HV), an improvement of approximately 44%. This increase is attributed to grain refinement—via the Hall-Petch relationship, finer grains provide greater resistance to dislocation motion—together with reduced porosity and improved interlayer bonding [10][11][13][29][31].

Figure 4. FE-SEM micrograph of the WAAM + post-deposition FSW AA4043 sample, showing fine, equiaxed grains and homogenised Mg2Si-rich phase distribution.

Figure 5. EDX elemental maps (Al, Si) of the as-deposited WAAM AA4043 sample.

Figure 6. EDX elemental maps (Al, Si) of the WAAM + post-deposition FSW AA4043 sample, showing more homogeneous elemental distribution.

Figure 7. Vickers microhardness profile of the as-deposited WAAM AA4043 sample across five measurement points.

Figure 8. Vickers microhardness profile of the WAAM + post-deposition FSW AA4043 sample across five measurement points.

Table 3. Statistical summary of Vickers microhardness before and after post-deposition FSW.

The larger standard deviation observed after FSW (13.4 HV, versus 4.8 HV for as-deposited WAAM) reflects localised variation associated with non-uniform precipitate/phase redistribution during stirring [12][17], suggesting that further refinement of FSW parameters (e.g., tool geometry, plunge depth) could improve hardness uniformity.